U.S. Department of Commerce Volume 104 Number 1 January 2006 Fishery Bulletin U.S. Department of Commerce Carlos M. Gutierrez Secretary National Oceanic and Atmospheric Administration Vice Admiral Conrad C. Lautenbacher Jr., USN (ret.) Under Secretary for Oceans and Atmosphere National Marine Fisheries Service William T. Hogarth Assistant Administrator for Fisfieries ^'i^^'°"=o^^ % ^ATSS 0» *■ 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 01,5700, Seattle, WA98115-007d. Periodicals postage is paid at Seattle, WA, and at additional mailing offices. POST- MASTER: Send address changes for sub- scriptions to Fishery Bulletin, Superin- tendent of Documents, Attn.: Chief, Mail List Branch, Mail Stop SSOM, Washing- ton, DC 20402-9373. Although the contents of this publica- tion have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has deter- mined that the publication of this peri- odical is necessary according to law for the transaction of public business of this Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $5.5.00 domestic and $68.75 foreign. Cost per single issue: $28.00 domestic and $35.00 foreign. See back for order form. Scientific Editor Adam Moles, Ph.D. Technical Editor Elizabeth Calvert 11305 Glacier Highway Juneau, Alaska 99801-8626 Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C15700 Seattle, Washington 981 15-0070 Editorial Committee Harlyn O. Halvorson, Ph.D. Ronald W. Hardy, Ph.D. Richard D. Methot, Ph.D. Theodore W. Pietsch, Ph.D. Joseph E. Powers, Ph.D. Harald Rosenthal, Ph.D. Fredric M. Serchuk, Ph.D. George Walters, Ph.D. University of Massachusetts, Boston University of Idaho, Hagerman National Marine Fishenes Service University of Washington, Seattle National Manne Fisheries Service Universitat Kiel, Germany National Marine Fisheries Service National Marine Fisheries Service Fishery Bulletin web site: www.fishbull.noaa.gov The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. It began as the Bulletin of the United States Fish Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46: the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions. State and Federal agencies, and in e.xchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 104 Number 1 January 2006 Fishery Bulletin Contents The conclusions and opinions ex- pressed in Fishery Bulletin are solely those of the authors and do not represent the official position of the National Marine Fisher-ies Ser- vice (NOAA) or any other agency or institution. The National Marine Fisheries Service (NMFSl does not approve, recommend, or endorse any proprie- tary 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, recom- mends, or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. Articles MBLWHOI Library JAN 1 0 2006 Mas: 1 Rochet, Marie-Joelle, Jean-Francois Cadiou, and Verena M. Trenkel Precision and accuracy of fish length measurements obtained with two visual underwater methods 10 Witteveen, Briana H., Robert J. Foy, and Kate M. Wynne The effect of predation (current and historical) by humpback whales (Megaptera novaeang/iae) on fish abundance near Kodiak Island, Alaska Companion papers 21 Weinberg, Kenneth L., and David A Somerton Variation in trawl geometry due to unequal warp length 35 Kotwicki, Stan, Kenneth L. Weinberg, and David A. Somerton The effect of autotrawl systems on the performance of a survey trawl 46 Zeidberg, Louis D., William M. Hamner, Nikolay P. Nezlin, and Annette Henry The fishery for California market squid f.Loligo opa/escens) (Cephalopoda: Myopsida), from 1981 through 2003 60 Criales, Maria M. John D. Wang, Joan A. Browder, Michael B. Robblee, Thomas L. Jackson, and Clinton Hittle Variability in supply and cross-shelf transport of pink shrimp (Farfantepenaeus duorarum) postlarvae into western Florida Bay 75 Grandcourt, Edwin M., Thabit Z. Al Abdessalaam, Ahmed T. Al Shamsi, and Franklin Francis Biology and assessment of the painted sweetlips (Diagramma pictum (Thunberg, 1792)) and the spangled emperor (Lethrinus nebulosus (Forsskal, 1775)) in the southern Arabian Gulf Fishery Bulletin 104(1) 89 Porch, Clay E., Anne-Marie Ekiund, and Gerald P. Scott A catch-free stoack assessment model with application to goliath grouper (Epinephelus itajara) off southern Florida 102 Tuckey, Troy D., and Mark Dehaven Fish assemblages found in tidal-creek and seagrass habitats in the Suwannee River estuary 118 Clarke, Lora M., Alistair D. M. Dove, and David O. Conover Prevalence, intensity, and effect of a nematode (Philometra saltatnx) in the ovaries of bluefish (Pomatomus saltatnx) 125 Edwards, Elizabeth F. Duration of unassisted swimming activity for spotted dolphin (Stenella attenuata) calves; implications for mother-calf separation during tuna purse-seme sets 136 Saillant, Eric, and John R. Gold Population structure and variance effective size of red snapper (Lutianus campechanus) in the northern Gulf of Mexico Note 149 Friedland, Kevin D., Lora M. Clarke, Jean-Denis Dutil, and Matti Salminen The relationship between smolt and postsmolt growth for Atlantic salmon (Salmo salar) in the Gulf of St. Lawrence 156 Erratum 157 Guidelines for authors Abstract — During the VITAL cruise in the Bay of Biscay in summer 2002, two devices for measuring the length of swimming fish were tested: 1) a mechanical crown that emitted a pair of parallel laser beams and that was mounted on the main camera and 2 1 an underwater auto-focus video camera. The precision and accuracy of these devices were compared and the various sources of measurement errors were estimated by repeatedly measuring fixed and mobile objects and live fish. It was found that fish mobility is the main source of error for these devices because they require that the objects to be measured are perpendicular to the field of vision. The best performance was obtained with the laser method where a video- replay of laser spots (projected on fish bodies) carrying real-time size information was used. The auto-focus system performed poorly because of a delay in obtaining focus and because of some technical problems. Precision and accuracy of fish length measurements obtained with two visual underwater methods Marie-Joelle Rochet Ldboratoire MAERHA, IFREMER Rue de nie d'Yeu, B.P 21105 44311 Nantes, Cedex 03, France E-mail address mirochetifiifremerfr Jean-Francois Cadiou IFREMER (DNIS/SM/IM)-Centre Mediterranee B P 330 83507 La Seyne, France Verena M. Trenkel Laboratoire MAERHA, IFREMER Ruede llle d Yeu, BP 21105 44311 Nantes, Cedex 03, France Manuscript submitted 22 July 2003 to the Scientific Editor's Office. Manuscript approved for publication 20 April 2005 by the Scientific Editor. Fish. Bull. 104:1-9(20061. Visual sampling of marine systems by SCUBA divers and underwater vehi- cles is increasingly used to estimate animal abundances, to observe natu- ral behavior and response behavior to fishing gear in situ, and to assess community interactions (e.g., Bublitz, 1996; Auster et al., 1997; Davis et al., 1997; Uiblein et al., 2002; Trenkel et al., 2004). Visual methods also allow estimates of population-size struc- tures without the bias caused by the size selectivity of fishing gear. Visual techniques have been used in the wild for measuring the length of animals by SCUBA divers (e.g., Yoshihara, 1997; Pfister and Goulet, 1999; Harvey et al., 2002a) or by submersibles (Love et al., 2000; Yoklavich et al., 2000). They have also been employed for esti- mating the length frequency of the catch of live tuna to be fattened after capture (Harvey et al., 2003), and in aquaculture to estimate the size range offish (Petrell et al., 1997). Until now these techniques were mainly used in shallow waters or tanks. Because of the optical characteristics of sea water — its turbidity, the variations in light intensity with depth and water movements and fish movements, these methods are subject to measurement errors. Estimating the order of mag- nitude of this measurement error has been the focus of many studies (van Rooij and Videler, 1996; Yoshihara, 1997; Harvey et al., 2001, 2002a, 2002b, 2003). Efficient methods for measuring fish length in situ can also be used in deeper waters not accessible to divers. Parallel laser projected from a video camera onto the seafioor or fish bodies permit accurate measure- ments (Love et al., 2000; Yoklavich et al., 2000). Albert et al. (2003) mea- sured fish lengths on a video screen and then transformed these mea- surements into real length knowing the distance of the camera from the ground, its tilt angle, and the hori- zontal opening angle of the camera. If fish are not on or close to the bottom, it is necessary to know their distance off the bottom to apply this method. Auster et al. (1997) and Norcross and Mueter (1999) measured fish size on a video screen when the fish appeared between the skids of their ROV. The screen measurement is then related to the known distance of the skids. This method relies on the fish and skids being in the same horizontal plane and on the fish being perpen- dicular to the axis of the camera. Krieger (1992) used a submersible to estimate the size of rockfish. Two methods were tested during the VITAL cruise in the Bay of Bis- cay, in late August and early Septem- Fishery Bulletin 104(1) 232 mm 232 mm Figure 1 The four laser pointers mounted on a crown around the main camera in front of the ROV Victor 6000. The inner circle is the camera lens. Figure 2 Laser spots (indicated by arrows) visible on and under a fish. These laser spots documented on videotape provided size infor- mation both in real time and during video replay. ber 2002.^ Victor 6000, a remotely operated vehicle (ROV) equipped with several video cameras and re- corders, was operated at depths ranging from 1100 to 1500 m. Fish size was measured both by using a pair of parallel laser beams, and an auto-focus video camera linked to software for estimating object size based on the focal distance of the object in focus. In this study three sources of measurement variability were investi- gated: 1) systematic errors inherent to each method; 2) variability due to observer differences; 3) variability due to continuous fish movements and horizontal body orien- tation. To estimate these error components separately, rigid and articulated artificial objects ("artificial fish") of known size were measured repeatedly by several independent observers. Individuals belonging to several deep-sea fish species were also repeatedly measured. Mixed-effects models and heteroscedastic error models were fitted to the resulting measurements to compare the magnitude of errors due to different sources. Materials and methods Measurement devices Both measurement devices were installed close to each other on the ROV Victor. The laser-beam crown was mounted on the main camera, which was itself attached 1 Trenkel, V. M., N. Bailly. O. Berthele, O. Brosseau, R. Causse, F. de Corbiere, O. Dugornay, A. Ferrant, J. D. M. ordon, D. Latrouite, D. Le Piver, B. Kergoat, P. Lorance, S. Mahevas, B. Mesnil, J.-C. Poulard, M.-J. Rochet, D. Tracey, J.-P. Vach- erot, G, Veron, and H. Zibrowius. 2002. First results of a quantitative study of deep-sea fish on the continental slope of the Bay of Biscav: visual observations and trawling. ICES CM 20b2/L:18, 2002, 15 p. to the pan and tilt unit. The METRAU© (SONY, model FCB-IX 47P) autofocus camera was mounted on the same pan and tilt unit. Laser-beam pointers Four red laser pointers (10 mW, 635 nanometers [nm]) were mounted around the main camera housing (Fig. 1). The distance between each two opposite lasers was 232 mm. Red light is strongly attenuated by water but because of the relatively high power of the laser light-emitting diode (LED), a range of up to 7 m is reachable in clear waters. To measure the fish and objects, the laser beams were projected on the target (Fig. 2). The laser spots, visible on the video, give size information both in real time and during video replay. The principle is simple, but several limitations exist. First, the measurement is correct only for an object located in a plane perpendicular to the laser axis. Second, the target should be large enough to be reached by at least two laser beams; the more laser impacts that are seen on the object to be measured, the easier the measurement. For the measurement to be accurate, there must be a strict parallelism between the laser beams. This is complicated by the fact that the laser component itself (the diode with its optic lens) does not necessarily have a beam parallel to the axis of the component package. Further, designing an accurate alignment mechanism that is compatible with offshore and deep underwater operating conditions is difficult. The residual error af- ter alignment is about 0.15°, which entails an error of 10 mm for the distance between two opposite spots at a distance of 4 meters (i.e., 4% of size). METRAU camera The METRAU system is based on the autofocus video camera. The imaging device is an original equipment manufacturer (OEM) camera module similar to those used in off-the-shelf camcorders. The Rochet et al : Precision and accuracy of fish length measurements obtained with two visual underwater methods " .'*^ ms' ..Jg^l 3:02:26 -*•- ■. y ^ Figure 3 A METRAU video camera image with overlaid grid. camera has a built-in automatic focus unit which adjusts lens settings to provide a sharp image. The camera is remotely controlled by a RS323 digital link and sends data back over this link, including zoom position and focal value (see details in Cadiou et al., 2004). A previ- ous calibration in air and in a test tank provided cor- relation rules between raw data and field angle or focal distance. When the system is operated, the data received by the computer are processed in real time. The object distance and the field angle are computed and a scale is overlaid on the video image (Fig. 3). According to optical laws, the depth of focus decreases as the focal length increases. This means that in order to obtain an accurate measure of focal distance, a nar- row field angle is required. In addition, the depth of focus increases when the focal distance moves towards infinity. Consequently, a domain of validity of the mea- surement can be defined. With the METRAU camera, there must be a target distance under three meters and a field angle of less than 6°. These constraints have to be combined with the following conditions: a steady image that would allow the automatic focal servo to stabilize; and avoidance of scenes with several image planes. In turbid waters, particles can create disturbing focal planes and affect the measurement process. Measurement experiments Artificial objects and live fish Objects of known size were used to estimate the potential bias in the length measurements obtained with the two devices. Three rigid objects — a can, a bottle, and a plastic tube mea- suring respectively 13, 30, and 66 cm — were repeatedly measured to evaluate device performance and observer- induced variability in the absence of errors induced by fish movement and variations in horizontal observation angles. Fish movement makes the horizontal observation angle vary continuously. As a result, it is difficult to judge if and when an individual fish is perpendicular to the measurement axis. Further, fish seldom lie in a straight plane. Some species continuously flex their tail, others bend their whole body. To mimic the mobil- ity of a real fish, a mobile object was built consisting of several pieces of Ertalyte (Quadrant Engineering Plastic Products, Bridgeport. CT) plates linked together with rope rings. This artificial fish was designed to be neutrally buoyant so that it could be moved by water currents and undulate like a real swimming fish. The "artificial fish" had three distinctively colored parts. Thus depending on how many parts were measured, a small (13 cm), medium (17 cm), or large (41 cm) "artifi- cial fish" was the result. The real size of rigid objects and of the artificial fish was unknown to the observers throughout the measurement experiment. In addition to measuring each of the rigid objects and the artificial fish, 351 individuals belonging to 21 deep-sea fish species were measured with both methods. The body sizes of these species ranged from 5 to 110 cm. Each individual fish was measured up to nine times. Altogether 2373 measurements were carried out. Real time and postoperation measurements While the ROV was in operation, four to five observers were able to watch the video images. Real time measurements were performed by estimating sizes directly from the screen, without using any measuring instrument. Each observer was asked to write down his or her length estimate without announcing it, so that independent measurements were obtained. All artificial objects were measured by both trained and novice observers; real fish were measured only by trained observers, namely scien- tists and ROV pilots. All objects and fish were measured at distances of 2 to 5 meters. Postoperation measurements were also performed on registered videos and digital images. For the laser method, the video tape was replayed. The tape was stopped when the image with an object or fish seemed to be in the best possible position. The fish or object was then measured with a ruler on the still video image. Postoperation measurements made with a ruler were also performed on digital snapshots taken from the videos in real time for both the laser and the METRAU method. A ruler was used rather than computer image analysis because it was easy and cost-efficient and it was felt appropriate for this trial appraisal of measure- ment methods. The bias introduced by this method was assumed to be negligible compared to observer-induced and fish-movement-induced errors. Operational constraints prevented a full factorial design where all observers could use all methods and measure all objects. Data analysis Variance components for observers and fish movements The measurement variability due to observer differences Fishery Bulletin 104(1) and fish movements was estimated from the measure- ments of the artificial objects, for which the true size was known. Because the measurement variance was expected to be larger for the mobile artificial fish than for the rigid objects, an extended linear mixed-effects model (Pinheiro and Bates, 2000) was used to account for this expected heteroscedasticity. The model included true length as a fixed effect and observer as a random effect. Fixed objects and mobile objects (artificial fish) were allowed separate variances. The resulting model was L,f= 1.1 + pL* + o, + f^ -t-f,,. (1) where L, ^ = length measured by observer / of an object of class J=|fixed, mobile! and of true size L*; o, ~ iV(0,a(o)); and e,j ~ N(0,a), whereas /", ~ N(0, a,(/;)). The model was fitted to the data from the eight observers who had measured all objects with the laser method. Accuracy and precision for artificial objects An extended linear model including heteroscedastic variance terms was used to compare the precision and accuracy of the two methods. The model included fixed effects for true length and the measurement method. The esti- mated fixed effects allow assessment of the potential measurement bias of each method. The measurement errors for fixed and mobile artificial objects were mod- eled separately for each method. This allowed us to compare the precision of the methods. Thus, the fitted model was measurements. The model included a fixed individual fish effect (each fish had a different, unknown size) and a random observer effect; and fish species were al- lowed heteroscedastic variances to account for species behavior differences (Lepidion versus Bathypterois). The stationary species, B. dubius, is easier to measure compared to the more lively L. eqiies. The model was ^,,.,.1 = ^'n +0, +S, + £„,, (3) where Z>,, , ^ = the length measurement obtained by ob- server / for individual fish n belonging to species /; o, ~ MO, a(o)); and f,„, ~ MO, a), whereas s, ~ N(o, a, (s,)). Unfortunately, it was not possible to carry out a direct comparison between the precision of size estimates of fish and artificial objects because the latter were measured seven to 11 times, whereas the former were measured only two to seven times. Random subsamples could be carried out to obtain comparable sample size; unfortunately, subsamples from large samples would still have a larger variance than small samples. All models were fitted by using Splus 6.0 for Unix (MathSoft, Seattle, WA). For heteroscedastic models, be- cause of identifiability constraints, the fitting algorithm provided estimates of the ratio between the standard deviations of each class in relation to the standard deviation of a specified class instead of the full set of standard deviations. Results H + PL* + f,k + f,*. (2) where k = the measurement method and j the object class as before. As in model 1, f^,, ~ N(0, o). In contrast, f^i. ~ N(0, o^f^if^i.)) allowed for separate variances for each object-type and method pair. Only two trained observ- ers used both measurement methods for all objects. Because there was no significant difference between their measurements, and in order to reduce the number of parameters to be estimated, no observer effect was included in this model. Precision of fish measurements The precision of fish- length estimates was compared for two species, Bathyp- terois dubius and Lepidion eques. These species were selected for this analysis because they are abundant and relatively easy to measure, compared to other species that move faster or flex their body more often. Twenty- four individuals belonging to these two species were measured repeatedly by up to five observers using the real-time laser measurement method. Because true fish size was unknown, measurement accuracy could not be estimated. For estimating the pre- cision of fish measurements, an extended linear model with heteroscedastic errors was fitted to the fish length Precision and accuracy of measurements varied among objects and methods (Table 1). The best precision was obtained with the video-replays of laser measurements, whereas METRAU generally did not perform very well, especially on snapshots. The precision was generally much lower for mobile objects than for rigid objects. Mea- surement bias was generally low for the laser method, whereas the METRAU method systematically under- estimated the size of objects. A variety of fish species with various sizes were measured. CVs for individual fish measurements varied from 3% to 23% (Table 2). Species were grouped according to their motion behavior (l=sitting on bottom motionless, 2 = station holding or drifting, 3 = slow swimming, 4=fast swimming [Lorance and Trenkel-]). CVs were found to differ between groups, increasing with mobility (mean CV in group 1: 8.9%; group 2: 9.7%; group 3: 12.9%; group 4 was excluded because there was only one individual, P<10"''). - Lorance, P., and V. Trenkel. In preparation. Natural behaviour and reaction to an approaching ROV of large mid-slope species. IFREMER, Centre do Brest, B.P. 70, 29280 Plouzane. France. Rochet et al.: Precision and accuracy of fish length measurements obtained with two visual underwater methods Table 1 Summary of length measurements for artificial objects by five visual methods. Lengths (in cml are given along with their coefficient of variation i'i) and number of observations in parentheses, a.f = artificial fish. The laser method entailed viewing laser points (that permit accurate length data) projected on fish. The METRAU method is based on an autofocus video camera. Object True size Laser Laser + video Laser + snapshot METRAU METRAU + snapshot Can 13 14.31 (13'7f, 26) 13.76 (3%, 5) 13.55 (4%, 5) 11.25(7%, 10) 10.96 (7%, 5) Bottle 30 30.87(7%, 26) 31.90(2%, 5) 31.47 (2%, 5) 26.90(18%, 10) 32.05(41^5,5) Tube 66 65.62 (9%, 26) 66.40(2'/,, 5) 66.66(1%, 5) 44.25 (14%, 8) 40.75(25'?, 4) Short a.f 13 13.30 (13%. 211 13.32 (6%, 3) 22.10(23%, 2) 11.50(6%, 2) 7.55 (33%, 6) Medium a.f 17 16.49 (14%, 21) 17.11 (5%, 3) 24.86(25%, 2) 11.50(18%, 2) 9.25(32%, 6) Large a.f 41 40.80 (17%, 21) 44.87 (14%, 3) 63.23(25%, 2) 30.00 (14%, 4) 23.21(31%, 6) Table 2 Summary offish length measurements, n = number of individual fis h measui •ed for each species, in = total number of measure- ments. m In = mean number of ob servatior s per indiv idual. Mean length = average individual fish length (cm) per species. Mean CV = average individual coefficient of variation per species. Group = behavioral group of the species (l=sitting -m bottom motionless, 2=station holding or drifting, 3 = slow swimming, 4=fast swimming). Species n in in In Mean length Mean CV Group Alepocephalus bairdi 2 4 2.0 48.4 8.2% 2 Balhypterois 87 586 6.7 19.2 8.9% 1 Breviraja caerulea 3 13 4.3 29.6 4.8% 2 Caelorinchus labiatus 15 73 4.9 28.7 10.2% 2 Cataetyx latyceps 1 5 5.0 24.1 12.4% 2 C him a era monstrosa 6 23 3.8 91.4 9.2% 3 Coelorhyncus labiatus 2 8 4.0 25.4 14.9% 2 Coryphaenoides rupestris 21 77 3.7 45.4 12.1% 2 Cottunculus thomsoni 2 17 8.5 29.1 16.1% 1 Galeus melastomus 1 4 4.0 30.7 2.8% 4 Hoplostethus atlanticus 6 27 4.5 32.6 9.4% 2 Hydrolagus affinis 1 6 6.0 74.9 23.3% 3 Hydrolagus mirabilis 3 6 2,0 78.4 11.9% 3 Lepidion eques 161 814 5.1 26.6 9.6% 2 Mora rnoro 3 8 2.7 60.0 7.9% 2 Nezumia aequalis 10 22 2.2 32.8 8.9% 2 Notacanthus 2 5 2.5 44.6 10.6% 2 Phycis blenoides 1 4 4.0 44.8 15.3% 2 Syphobranchus kaupii 7 23 3.3 27.2 16.2%' 3 Trachyrincus in u rrayi 4 7 1.8 38.9 3.5% 2 Ti-achyscorpia crintulata echinata 14 67 4.8 40.7 8.1% 1 Measurement variance for observers and fish movements The standard deviation of the observer random effect, aio), amounted to approximately 20% of the residual standard deviation. The standard deviation of the random effect for mobile objects, On,obih'^fmobih^^ was 40'7c higher than that for rigid objects, indicating that the measurement variability was lower for fixed than for mobile objects, as expected. The slope and intercept of the fixed effects did not significantly differ from 1 and 0, respectively. Thus the laser method is unbiased and performs equally well for all sizes in the range tested (Table 3). Accuracy and precision of measurements for artificial objects Size measurements carried out using the laser beams in real time or on registered videos were found to be Fishery Bulletin 104(1) Table 3 Estimated fixed-effect coefficients and standard devia- tions for model 1 for the measurements of rigid and mobile objects by eight independent observers using the laser method. SE = standard error; CL =confidence limit. Coefficient Estimate SE P-value ii ft 1.53 0.976 0.97 0.023 0.12 <0.0001 Standard deviation Lower CL Estimated LIpper CL oio) 0.296 1.126 4.435 a 4.369 5.535 7.014 On,jL,,,^lo,„oH.iM.nMc'' 0-523 0.710 0.964 ^ Standard deviations were estimated in relation to the standard deviation for mobile objects. unbiased, whereas the METRAU-based measurements underestimated the true length of objects by as much as 7 cm for real-time measurements and 10 cm for time- delayed measurements (Table 4, Fig. 4). In addition, the variance of METRAU measurements was systemati- cally larger than the corresponding laser measurements (Table 5: ratios larger than 1). For rigid objects, laser- based video-replays had a lower variance than real time measurements. This lower variance for postoperational measurements was due to the allowance of videos to be replayed as many times as necessary in order to select the best image where an object was perpendicular to the optical axis. Use of a ruler also improves the mea- surement. The high estimation variance obtained for mobile objects measured with the laser method on digi- tal snapshots was partially due to one outlier (Fig. 4B). The object was measured at a relatively great distance in somewhat turbid water. The outlier was not removed because these kinds of errors are to be expected under common measurement conditions in the field. Generally the most precise results were obtained for video-replays with the laser beam method. Table 4 Estimated coefficients for model 2 for measurements | of rigid and mobile objects by five varia nts of the two methods. Coefficient Estimate SE P-value f'/oser 0.106 0.934 0.91 I^METRAU -7.198 1.621 <0.0001 f-^hser+video 1.466 0.937 0.15 ' laser+snapsliol 1.348 0.932 0.19 >'METRAU*snap -hot -9.516 1.806 <0.0001 /' 0.976 0.023 <0.0001 a 4.592 Precision of fish measurements The variance of the random effect for B. dtibius was about 66% of the variance estimated for L. eques. For live fish, the standard deviation of the observer random effect was approximately 16% of the residual standard deviation (Table 6). This standard deviation is lower than that obtained for objects of known size because the residual variance was larger owing to the small number of repeated measurements obtained for each fish. It was not easy to repeatedly measure fish because of escapement behavior. In addition, only trained observers took part in this experiment, which reduced observer variability. Discussion Table 5 Estimates of the standard deviations of length mea- surements for rigid and mobile objects obtained by five variants of the two methods, from model 2. Number of measurements are given in parentheses. Estimates sig- nificantly different from 1 are in bold font. Method Rigid objects Mobile objects Laser 1.12(21) 1.00' (91 Laser + video 0.21(15) 1.03(91 Laser + snapshot 0.19(151 3.43(6) METRAU 2.11(201 1.06(81 METRAU + snapshot 3.16(14) 1.58(181 All standard deviations are relative to the standard deviation for laser measurements of mobile objects. The potential sources of errors and variability in visual fish length measurements are 1) the design and calibra- Table 6 Estimates and 9b''i confidence limits for the standard deviations of the components of model 3 for fish size mea- surements obtained for two species by five independent observers using the laser method. Standard deviation Lower Esti- mate Upper aio) o '^B.dubius^^Bdub,us''l "L.eques'^L.eques^ 0.068 1.096 0.41 0.278 1.130 1.702 2.642 0.66 1.06 Standard deviations provided in relation to the standard devia- tion (or Lepuiion measurements. Rochet et al Precision and accuracy of fish length measurements obtained with two visual underwater methods 16 —r- Can 25 ^ m 20 D 2^°^' ■ 14 12 10 1 = 2 15 r^ ^ ^ g 10 ^^ (=: S M+S L M L+V L+S M+S L M L+V L+S 50 — 30 „ Bottle B 25 E Medium ■ 40 30 20 20 ■ » ■.... B M+S = B " f - L M L+V L+S M+S L M L+V L+S 70 60 50 40 30 ^ Tube C.H ^ s 70 Large s s M+S ■ 60 ~ — 50 ~ i i '° M H 30 H 20 = i L M L+V L+S M+S L M L+V L+S Figure 4 Boxplots for the distributions of length measurements (cm) obtained by five measure- ment methods for each of three fixed (left) and three mobile (right) objects Boxes = interquartile range, white line = median, whiskers = extremes (excluding outliers). Dotted line is the true size of the object. L = laser method; M = METRAU method; L+V = laser video-replay; L + S = laser snapshot measurement; M + S = METRAU j snapshot measurement. tion of the measurement devices, 2) differences among observers, 3) orientation and position offish in relation to the camera, and 4) swimming motion. We investigated each of these featuress. Among the two methods tested during this study, the METRAU method performed poorly. It has several disadvantages. First, METRAU system needs the fish brought into focus when it is perpendicular to the field of vision. This may take time during which the fish can escape. Second, the registered video images do not in- clude the superimposed calculated scales. Thus, unlike the laser method, it is not possible to replay the video to identify the best image. Postoperational measurements can be performed only by using the digital snapshots registered in real time. Third, this method had a higher variance than the laser method, probably because of the technical constraints just mentioned. Fourth, during the VITAL cruise the estimates were systematically biased downwards. Measurements carried out after the cruise in a laboratory pool confirmed this systematic underes- timation to be -20% of the real sizes. This result may be due to errors in the software which processes the output of the camera. Although the errors could prob- ably be fixed, and the hardware improved to generate a video signal with the overlaid scale for recording, the other disadvantages are more difficult to eliminate, be- cause of the intrinsic limitations of the system. Hence the method is not promising for estimating the size of fish in the wild. By contrast, the laser beam method performed rath- er well, at least for rigid objects. We obtained CVs of 7-13% for rigid objects in real time and 1-4% with image postprocessing (Table 1), both of which compare well with CVs for silhouette measurements obtained with a single camera placed in a laboratory pool (with scale bars placed on the bottom of the pool) and with computer image processing (1%, Harvey et al., 2002b), or even with stereo-video measurements of silhouettes (0.6-7.5%, Harvey and Shortis, 1996). Length mea- surements were always unbiased and postoperational measurements on video images reached a high precision for rigid objects and for small- to medium-size mobile objects. Thus the method seems suited for measuring the size of animals of low mobility, like invertebrates, along visual observation transects. The variance due to differences between observers was about 20% of the residual variance. This variance was reduced to 16% when measurements were per- formed by trained observers. This is true for real-time measurements. For video-replays of the laser-beam data, the variance due to observers was very small because of the use of a ruler instead of subjective extrapolation of Fishery Bulletin 104(1) the known laser distance. The use of automated analy- sis of video images may further reduce the observer error source. The major difficulty in measuring fish length /;; situ is caused by fish mobility, which causes them to be in variable orientations and positions in relation to the camera, and also to be flexed. We addressed both variance components together by comparing measure- ments of mobile objects ("artificial fish") and rigid objects. With the laser beam method, the measurement standard deviation of rigid objects was estimated to be 20% of the standard deviation of mobile objects (95% confidence limits: 6-75%). These components may be of the same order of magnitude as those for fish measurements, although fish measurements could not be estimated in our study because the true size of the fish was unknown. An attempt to disentangle both components is provided by estimating the differ- ence between the precision obtained for species with contrasting behaviors. Bathypterois dubius individu- als lie motionless on the bottom and seldom move, but because they stand on their fins they are never exactly perpendicular to the camera. By contrast, L. eques swims close to the bottom and tends to escape when the ROV is approaching too closely. This species continuously moves its tail; therefore it is very diffi- cult to obtain an image with the whole body properly orientated and straight. The standard deviation of B. dubius length measurements was estimated to be 66% of that of L. eques. This difference is smaller than the difference between rigid and mobile objects above; therefore we conclude that the major part of variance is due to the orientation of the fish in relation to the camera. Similarly, the estimated CVs of 21 species grouped by motion behavior differed only slightly. This is consistent with previous studies which have shown that relative errors of single-camera or stereo-video measurements of silhouettes or frozen fish could reach 10% to 30%, depending on the distance to the camera, when the angle to the camera was increased from 0° to 60°, whereas the measurement CVs increased fourfold (Harvey and Shortis, 1996; Petrell et al., 1997; Harvey et al., 2002b). By contrast, error due to tail flexion and muscle contractions during swimming motions was estimated at -5% in a comparison of "linear" to "sinusoidal" length of dorsally photographed sharks (Klimley and Brown, 1983) and at 0.5% for repeated stereo-video measurements of swimming tunas (Har- vey et al., 2003). In conclusion, the major source of measurement error for live fish may be their orientation and position in relation to the camera. For animals that are sessile or lying immobile on the ocean floor, this would be much reduced if the camera and laser beams were mounted vertically instead of obliquely. Thus the laser-beam method may be potentially useful for measuring ben- thic animals. For mobile animals, however, stereo-video methods (Harvey et al., 2001; Harvey et al., 2002a; van Rooij and Videler, 1996) may be more promising, and are continuously improving (Harvey et al., 2003). Acknowledgments We thank all observers for taking part in the experi- ment, and the ROV pilots for their willingness and skill at pursuing fish and in obtaining positions suitable for being measured. An anonymous referee gave very helpful comments on a previous version of this manuscript. Literature cited Albert, O. T., A. Harbitz, and A. S. Koines. 2003. Greenland halibut observed by video in front of survey trawl: behaviour, escapement, and spatial patterns. J. Sea Res. 50:117-127. Auster, P.. R. Malatest, and C. Donaldson. 1997. Distributional responses to small-scale habitat variability by early juvenile silver hake, Merluccius bilniearis. Environ. Biol. Fish. 56:195-200. Bublitz, C. 1996. Quantitative evaluation of flatfish behaviour during capture by trawl gear. Fish. Res. 25:293-304. Cadiou, J.-F., V. Trenkel, and M.-J. Rochet 2004 Comparison of several methods for in situ size measurements of moving animals. In Proceedings of the fourteenth (2004) international offshore and polar engineering conference; Toulon, France, May 23-2S, 2004, p. 438-444. Int. Soc. Offshore Polar Engineers, Danvers, MA. Davis, C. L., L. Carl, and D. Evans. 1997. Use of a remotely operated vehicle to study habitat and population density of juvenile lake trout. Trans. Am. Fish. Soc. 126:871-875. Harvey, E., M. Cappo, M. Shortis, S. Robson, J. Buchanan, and P. Speare. 2003. The accuracy and precision of underwater measure- ments of lentgh and maximum body depth of southern bluefin tuna iTIiunnus maccoyii) with a stereo-video camera system. Fish. Res. 63:315-326. Harvey, E., D. Fletcher, and M. Shortis. 2001. A comparison of the precision and accuracy of esti- mates of reef-fish lengths determined visually by divers with estimates produced by a stereo-video system. Fish. Bull. 99:63-71. 2002a. Estimation of reef fish length by divers and by stereo-video. A first comparison of the accuracy and precision in the field on living fish under operational conditions. Fish. Res. 57:255-265. Harvey, E.. and M. Shortis. 1996. A system for stereo-video measurement of sub-tidal organisms. Mar. Technol. Soc. J. 29:10-22. Harvey, E.. M. Shortis, M. Stadler, and M. Cappo. 2002b. A comparison of the accuracy and precision of mea- surements from single and stereo-video systems. Mar. Technol. Soc. J. 36:38-49. Klimley, A. P., and S. T. Brown. 1983. Stereophotography for the field biologist: mea- surement of lengths and three-dimensional positions of free-swimming sharks. Mar. Biol. 74:175-185. Krieger, K. J. 1992. Distribution and abundance of rockfish determined from a submersible and by bottom trawling. Fish. Bull. 91:87-96. Love, M. S., J. E. Caselle, and L. Snook. 2000. Fish assemblages around seven oil platforms in the Santa Barbara Channel area. Fish. Bull. 98:96-117. Rochet et a\. Precision and accuracy of fish length measurements obtained with two visual underwater methods Norcross, B., and F.-J. Mueter. 1999. The use of an ROV in the study of juvenile flat- fish. Fish. Res. 39:241-251. PetreU, R. J., X. Shi, R. K. Ward, A. Naiberg, and C. R. Savage. 1997. Determining fish size and swimming speed in cages and tanks using simple video techniques. Aquacult. Engineer. 16:63-84. Pfister. R. D., and D. Goulet. 1999. Nonintrusive video technique for in situ sizing of coral reef fishes. Copeia 1999:789-793. Pinheiro, J. C, and D. M. Bates. 2000. Mixed-effects models in S and S-PLUS, 528 p. Springer, New York, NY. Trenkel, V. M., P. Lorance, and S. Mahevas. 2004. Do visual transects provide true population den- sity estimates for deep-water fish? ICES J. Mar. Sci. 61:1050-1056. Uiblein, F., P. Lorance, and D. Latrouite. 2002. Variation in locomotion behaviour in northern cut- throat eel iSynaphobranchus kaupi) on the Bay of Biscay continental slope. Deep-Sea Res. 49l 11:1689-1703. van Rooij, J. M., and J. J. Videler. 1996. A simple field method for stereo-photographic length measurement of free-swimming fish: merits and constraints. J. E.xp. Mar. Biol. Ecol. 195:237-249. Yoklavich, M. M., H. G. Greene, G. M. Cailliet, D. E. Sullivan, R. N. Lea, and M. S. Love. 2000. Habitat associations of deep-water rockfishes in a submarine canyon: an example of a natural refuge. Fish. Bull. 98:625-641. Yoshihara, K. 1997. A fish body length measuring method using an underwater video camera in combination with laser discharge equipment. Fish. Sci. 63:676-680. 10 Abstract — Humpback whales iMegap- tera novaeangliae) are significant marina consumers. To examine the potential effect of predation by hump- back whales, consumption (kg of prey daily) and prey removal (kg of prey annually) were modeled for a current and historic feeding aggregation of humpback whales off northeastern Kodiak Island, Alaska. A current prey biomass removal rate was modeled by using an estimate of the 2002 hump- back whale abundance. A historic rate of removal was modeled from a prewhaling abundance estimate (pop- ulation size prior to 1926). Two pro- visional humpback whale diets were simulated in order to model consump- tion rate. One diet was based on the stomach contents of whales that were commercially harvested from Port Hobron whaling station in Kodiak, Alaska, between 1926 and 1937. and the second diet, based on local prey availability as determined by fish surveys conducted within the study area, was used to model consumption rate by the historic population. The latter diet was also used to model consumption by the current popula- tion and to project a consumption rate if the current population were to grow to reach the historic population size. Models of these simulated diets showed that the current population likely removes nearly 8.83x10*' kg of prey during a 5-month humpback whale feeding season, which could include around 3.26 x lO*" kg of juve- nile pollock (Theragra chalcogramma). 2.55 X 10'' kg of capelin iMalloti/s vil- losiis). if these species are consumed in proportion to their availability. The historic humpback whale population may have removed over 1.76 x 10'' kg of prey annually. The effect of predation (current and historical) by humpback whales (Megaptera novaeangliae) on fish abundance near Kodiak Island^ Alaska Briana H. Witteveen Robert J. Foy Kate M. Wynne School of Fisheries and Ocean Sciences University of Alaska Fairbanli :4'(r\\ i53"3'o"w is: 4:ii"\\ 1 1 i522r(rw 1^2 ri)-\\ ' ' \ N A Afognak Island • • • ', ^■--^x' •\ Mannot Bay + + ^u^_ ■ r'u'l + y >^-'' • + ', * • • •• ' 4- • -' - '" t •■ "^ / .•. - > 1 Chiniak Ba\ Kodiak Island • • • 1 ' 0 5 10 20 ' 1 30 40 1 1 1 Figure 2 A close-up of the study area showing locations of humpback whale (Megaptera novaeangliae) sightings and prey tows ( + ) for 2001 and 2002. Only mid-water trawl locations are shown. occurrence within the diet. Thus, diet B simulated a weighted availability of prey species based on temporal and spatial overlap between prey surveys and humpback whale sightings within the study period (Fig. 2). Consumption rate A seasonal consumption rate was estimated for both the current humpback whale population and the pre- whaling humpback whale population. The prewhaling consumption rate was estimated by using diet A only. Diet B was used to estimate the consumption rate by the current humpback whale population and to project the consumption rate by a humpback whale population at the prewhaling abundance level. The active metabolic rate (kcal/day) of feeding humpback whales was estimated in this study as £ = 192A/o '5 where Kleiber's (1961) model for basal metabolic rate (BMR; E=70M"'^) was modified by us- ing average oxygen consumption estimates for feeding baleen whales, where M is average body weight (kg) (Wahrenbrock et al., 1974: Sumich, 1983; Perez and McAlister, 1993). Daily prey consumption was then estimated as 1 K 1,000 where / = total prey consumption (kg/day); E = estimated daily energy requirements (kcal/ day); and K = the estimated energy density (kcal/gram wet weight) of presumed prey. The average body mass for humpback whales (Mi was set equal to 30,408 kg (Trites and Pauly. 1998). The total energy density iK) of each diet was calculated by multiplying the average seasonal energy density of each prey species sampled in the study area by the percentage of that species within each diet and summing across all species. Values of A' for individual prey species came from proximate compositions that were determined from prey collected during 2002 trawl surveys for all months within the study period (Foy*). For each month, energy density was calculated by multiplying percent lipid by 9.4 kcal/g and percent protein by 4.3 kcal/g, which are conversion factors based on heat produced during metabolism of food (Schmidt-Nielson, 1997). Carbohydrates were considered to be bound and not available for nutrition (Gaskin, Fishery Bulletin 104(1) 1982). The average seasonal energy density of each prey species was calculated by summing all values of energy dens.ty and dividing by the number of months in the study period. Previously published proximate composition values for surf smelt {Hypomesus pretious) and energy density data for euphausiids (Thysanoessa spp.) were used (Davis et al., 1997; Payne et al., 1999). Seasonal prey consumption for the population was es- timated by multiplying 1 by estimates of abundance (A^ ) and the total number of days in the humpback whale feeding season. Consumption estimates were calculated for both the upper and lower 95% confidence limits on the abundance estimates to show a possible range of consumption. The length of the feeding season was pre- sumed to be 152 days (Perez and McAllister, 1993). Results Table 1 Number of sightings of humpback whales iMegaptera novaeangliae) for 2001 and 2002 by subarea and month in the Kodiak Island study area. Area 2001 and 2002 June July August September Total 1 10 7 10 3 30 2 29 89 22 3 143 3 20 8 0 12 40 4 0 3 0 7 10 Nearshore 5 14 0 0 19 Total 64 121 32 25 242 Analysis of sightings showed that humpback whales were not uniformly distributed within the study area (Table 1). Occurrence of humpback whales within sub- areas was variable, indicating within-season shifts of habitat use. Peak humpback whale sightings occurred in subarea 2 in July of both years. No humpback whales were sighted in the nearshore area after the month of July in either year. Only two prey items were identified in the 27 stom- achs that contained appreciable quantities of prey of 39 stomachs analyzed by Thompson (1940). Surf smelt were found in 21 of 27 (78%) stomachs and euphausiids were found in 6 of 27 (22%) stomachs (Table 2). These percentages represent diet A. Energy densities of these two species combined to give a total energy density of 1.31 kcal/gram (Table 3). The fish species in areas used by humpback whales in 2001 and 2002, as shown by mid-water trawl sur- veys, were pollock (36.96%), capelin (28.89%), eula- chon (7.60%), Pacific sandlance (4.44%), Pacific sandfish (0.08%), and Pacific herring (0.03%) (Table 2). These percentages represent diet B. Calculated energy densi- ties of prey species ranged from a high (eulachon) of 2.52 kcal/gram to a low of 1.12 kcal/gram (juvenile pollock). The total energy density for diet B was 1.19 kcal/gram (Table 3). Based on energetic content of the above diets, the model indicated that each humpback whale in the study area would consume 338 kg/day on diet A and 370 kg/ day on diet B. Using a prewhaling estimate of 343 (95% CI=331-376) animals in the study area, we determined that humpback whales feeding on diet A prior to 1927 would have removed an estimated 1.76x10" kg of prey annually (95% CI= 1.70x10" to 1.93x10"), including nearly 3.87 xlO^ (3.74x10^ to 4.24x106) kg of euphausi- ids and approximately 1.37x10" (1.32x10" to 1.50x10") kg of surf smelt (Table 4). If diet B accurately reflects prey selection by the estimated 157 (95% 01 = 114-241) Table 2 Composition and relative occurrence of prey species represented in simulated humpback whale (Megaptera novaeangliae) diet A (historic) and diet B (current). Diet Prey species Common name Percent of total diet A Hypomesus pretious Thysanoessa spp. surf smelt euphausiid spp. Total 78.00% 22.00% 100% B Theragra chalcogramma Mallotus villosus Thysanoessa spp. Thaleichthys pacificus Ammodytes hexapterus Trichodon trichodon Clupea harengus pallasi walleye pollock capelin euphausiid spp. eulachon Pacific sandlance Pacific sandfish Pacific herring Total 36.96% 28.88% 22.00% 7.60% 4.44% 0.08% 0.03% 100% WItteveen et al Effect of prey removal by Megaptera novoeong/iae on fisfi abundance Table 3 Monthly and average energy densities kcal/g "am 1 of prey species represented in simulated humpbat k whale {Megaptera novuc- angliae) diets A and B based on lipid and protein composition. Energy densities were used to estimate consumption by humpback whales. Average values in parentheses have been adjusted to reflect standard deviations of lipid and protein composition. N/A = | not available. Species Energy densities (kcal/gram) June July August September Average Capelin 1.1285 1.2632 1.1956 1.4298 1.2542(1.1665, 1.3755) Pacific sandlance 1.4179 1.4179 1.4179 1.4179 1.4179(1.3211, 1.5590) Pacific sandfish 0.8661 1.2126 1.1165 1.1165 1.0779(1.0449, 1.1300) Eulachon 2.1582 2.5218 2.6758 2.7424 2.5245(2.3761,2.6860) Herring 2.0999 2.0999 1.9454 2.1205 2.0664(1.9432,2.2942) Juvenile pollock 1.0144 1.0657 1.1380 1.2461 1.1160(0.9730, 1.29941 Euphausiids N/A N/A N/A N/A 0.7430 Surf smelt N/A N/A N/A N/A 1.4698 Table 4 Daily and annual (over a 152-day feeding season) consumption of prey from two different diets off northeastern Kodiak Island by humpback whales ^Megaptera novaeangUae) at two levels of population abundance: the current population of 157 and the historic population of 343 (also presumed to be the carrying capacity to which the current population will recover). Diet A is the simulated diet of the historic population through analysis of stomach contents of 39 whales in 1937; Diet B is the simulated diet of the historic and current population based on currently available prey of suitable size for consumption. Prey species Daily prey removal (kg) Annual prey removal (kg) Mean Lower limit Upper limit Historic population Diet A Surf smelt 90,301 13,725,715 13,245,515 15,046,264 Euphausiids 25,469 3.871,355 3,735,914 4,243,818 Total 115,770 17.597,070 16.981.429 19,290,083 Diet B Euphausiids 27,934 4,246,006 4.097.458 4,654,514 Walleye pollock 46,924 7,132,406 6.882.876 7,818,614 Capelin 36,671 5,573,943 5,378.936 6,110,211 Eulachon 9,652 1,467,057 1,415,731 1,608,202 Pacific sandlance 5,635 856,475 826,511 938,876 Pacific sandfish 98 14,907 14,386 16,342 Pacific herring 33 5,038 4,862 5,523 Total 126,974 19,300,028 18,624,808 21,156,882 Current population Diet B Euphausiids 12,786 1.943,507 1,411,209 2,983,345 Walleye pollock 21,478 3,264,687 2,370,537 5,011,399 Capelin 16,785 2,551,.338 1,852,564 3,916,385 Eulachon 4,418 671,510 487,593 1.030.789 Pacific sandlance 2,579 392,031 284,659 601,780 Pacific sandfish 45 6,824 4,9.55 10,474 Pacific herring 15 2,306 1,675 3,540 Total 58.119 8,834,124 6,414,587 13,560,661 16 Fishery Bulletin 104(1) humpback whales currently feeding in the study ar- ea, these whales would be removing nearly 8.83x10^ (6.41x106 to 1.36x10") kg annually, including 3.26xl0« {2.37x10*5 to 5.01x10*5) kg of pollock, nearly 2.55 xlO*^ (1.85x10*5 to 3.92x10*5) kg of capelin, and 6.71x105 (4.88x105 to 1.03x10*5) kg of eulachon. If the same diet were consumed by a population of humpback whales al- lowed to return to prewhaling abundance, the projected population would remove 1.9x10" (1.86x10" to 2.12x10") kg of prey annually, including approximately 7.13x10*5 (6.88x10*5 to 7.82x10*5) kg of pollock, 5.57x10*5(5.38x10*5 to 6.11x10*5) kg of capelin, and 4.25x10*5 (4.10x10*5 to 4.65x10*5) kg of euphausiids (Table 4). Discussion Consumption rate Estimating the energy requirements of large cetaceans is inherently difficult and values presented in the present study may be subject to substantial uncertainty. Previ- ous studies in which consumption rates for cetaceans were estimated have used a range of values to adjust BMR (£=70M0 '5) for active metabolism. These values generally range from approximately 1.5 to 3 times BMR (Hinga, 1979; Lockyer, 1981; Sigurjonsson and Vikings- son, 1998). Our value of 192 is 2.7 times larger than 70 and is, therefore, a reasonable estimate because it fits within this range and is based on the observed oxygen consumption rates of baleen whales. However, the con- sumption estimates are highly sensitive to perturbations of model input; a 5% error in this value would cause deviation of the same percentage (5%) in final consump- tion values. Further, all values in our consumption model are assumed to be constant when body mass, physiologi- cal status, and assimilation efficiency are likely subject to large seasonal fluctuations (Innes et al., 1987; Perez and McAlister, 1993, Kenney et al.. 1997; Trites et al., 1997; Sigurjonsson and Vikingsson, 1998). Our model, however, did account for seasonal changes in the energy density of local prey sources; previous models, on the other hand, did not account for these changes (Perez and McAlister, 1993). Further research is necessary to obtain reliable field estimates of metabolic rates if model uncertainty is to be reduced. The historic prevalence of surf smelt in diet A could imply a dramatic change in surf smelt availability, misidentification, or an overestimation of smelt found in stomachs. Thompsons (1940) analysis resulted from "samples of stomach contents" obtained from catcher vessels; therefore, these samples may have completely missed less prevalent species. Further, stomach samples may have only reflected the most recent meal of the whale and therefore be biased toward a single species. This potential bias, however, could have been minimized by sampling stomachs throughout the season (May 30- August 09) (Thompson, 1940). Diet B was dominated by walleye pollock, a species not present in historic diet A. The increased importance of juvenile pollock in contem- porary humpback whale diet B could reflect changes in prey species availability and use, foraging selectivity, or reflect our diet reconstruction method. Diet B is considered provisional for two reasons. First, it is assumed that humpback whales eat prey species in proportion to their availability within foraging ar- eas. Humpback whales select preferred prey species and consumption, therefore, may be disproportional to availability. That is, they may be selectively forag- ing from all available prey sources. Previous foraging studies have described humpback whale distribution as being correlated with areas of capelin (Whitehead and Carscadden 1985; Piatt et al. 1989) and sandlance abundance (Payne et al. 1986; Kenney et al. 1996) and this correlation may indicate a possible preference for small forage fish species. Given that in the decades since whaling, the Gulf of Alaska has shifted from a system dominated by forage fish to one dominated by pollock and other groundfish (Merrick 1997; Anderson and Piatt 1999; Benson and Trites 2002), a shift in prevalence from surf smelt in the historic diet to pol- lock in the current diet is not unexpected. Pollock have been shown to be a dominant prey source of humpback whales harvested in Russia (Klumov, 1963). Addition- ally, humpback whales in southeastern Alaska have been observed near schools of juvenile pollock and are believed to eat pollock to an unknown, but potentially large, extent in some years (Gabriele^). The second source of uncertainty in diet B stems from the assumption that our mid-water trawl surveys provide unbiased samples of all available prey. Because these surveys were not designed to sample zooplankton, they may have produced a biased estimate of euphausiid availability. This bias may not be significant, however, because the 22% value we used in diet B was based on historic usage and falls within the range of euphausiid consumption (5-30% of the total diet) estimated in other humpback whale studies (Perez and McAlister, 1993; Kenney et al., 1997). Further, diet B was constructed from the results of mid-water trawl surveys that may underestimate the availability of some forage fishes, particularly Pacific sandlance. Pacific sandlance are often small enough to swim through the meshes in the net or are found in benthic habitats and cannot be captured by mid-water trawl methods. To minimize this potential sampling bias, we supplemented our trawl surveys with purse seine sampling in the nearshore subarea. Despite this effort we may have underestimated the prevalence of Pacific sandlance in the area because it was found to dominate the diets of other coastal piscivores; stomach contents of 34 coho salmon (Oncorhynchus kisutch) and Pacific halibut iHippoglossus stenolepis) in 2002 (Wit- teveen^) and regurgitants from blacklegged kittiwakes = Gabriele, C. 2001-2002. Personal commun. Glacier Bay National Park. P.O. Box 140. Gustavus, AK 99826-0140. '5 Witteveen, B. H. 2002. Unpubl. data. Fishery Industrial Technology Center, University of Alaska Fairbanks, Kodiak, AK 99615. Witteveen et al : Effect of prey removal by Megaplera novaeonglioe on fisfi abundance 17 in 2001 (n=96) and 2002 (72=147) were dominated by Pacific sandlance (Murra et al., 2003). Ecological effects from humpback whale prey consumption Although estimates of consumption are highly dependent on estimates of population abundance and metabolic rates, these values indicate that humpback whales were, and still are. significant predators within the Kodiak Island ecosystem. Historic commercial whaling reduced the population in our study area to an estimated low of 27 animals by 1938 (Witteveen, 2003). The removal of so many large consumers likely had significant impacts on the surrounding ecosystem. As modeled, reducing historic consumption to that of current levels would release nearly 10,000 tons of prey within the study area in a single feeding season. Such a release could have caused a trophic cascade effect. Cetacean removals in the Southern Ocean have dem- onstrated how trophic cascades can affect marine eco- systems through removal of large marine predators, including whales (Laws, 1985). It has been hypothesized that a similar reorganization of the marine community may have occurred in the Bering Sea and Gulf of Alas- ka, although the mechanisms of such a cascade are not well understood (Merrick, 1997; Trites, 1997; Springer et al. 2003). Removal of whales during commercial harvest reduced predation on certain fish, cephalopod, and zooplankton species, which were then available to other consumers. This large number of unconsumed prey, when combined with environmental factors such as the 1977 regime shift, may have contributed to the growth of sympatric marine predator populations from the late 1940s to late 1970s. It is hypothesized that whale stock resurgence, coupled with the 1977 regime shift that favored the proliferation of groundfish species, may have reduced prey availability to other piscivores in the system and may have led to declines seen in har- bor seal (Phoca vitulina), Steller sea lion, northern fur seal iCallorhinits iirsinus), common murre (Ur-ia aalge), thick-billed murre iU. lomvia). and red-legged kittiwake {Rissa brevirostris) populations (Merrick, 1995, 1997; NRC, 1996; Trites, 1997). The Gulf of Alaska and Ber- ing Sea ecosystems may still be affected by changes caused by baleen whale removals and their recovery (NRC, 1996). Assuming that the Kodiak Island study area was similarly affected by this trophic reorganization, an es- timate of the current consumption by humpback whales would help elucidate the role that a humpback whale recovery is playing in ecosystem dynamics. If our diet composition and subsequent consumption estimates are accurate, our results indicate that the diet of hump- back whales in Kodiak waters directly overlaps those of sympatric piscivores and the biomass that is removed may be substantial. The top species modeled in the humpback whale diet represent important sources of energy for multiple higher-trophic-level species and are known to be significant dietary species for Steller sea lions (Wynne^), harbor seals (Jemison**). tufted puffins (Fratercula cirrhata) (Piatt et al., 1997), blacklegged kittiwakes (Murra et al., 2003), adult pollock. Pacific halibut, and arrowtooth flounder (Livingston, 1993; Yang, 1995; Merrick, 1997; Best and St. Pierre^). Our model indicates that humpback whales within the study area may currently be consuming a signifi- cant amount of fish, including over 3.26 x 10'^ kg of juve- nile pollock and nearly 3.62 x 10'' kg of small forage fish, such as capelin, eulachon and Pacific sandlance, during a 152-day feeding season. In comparison, tufted puffins consume less juvenile pollock (6.40x10'' kg) between mid-July and mid-September, but this amount still accounts for one-tenth of the age-0 pollock stock in the Gulf of Alaska during early July (Hatch and Sanger, 1992). In addition, gadid removal by Steller sea lions in 1998 was estimated to be 1.79 xlO» kg, or 12% of the total gadid biomass that is removed by commercial fisheries for that year (Winship and Trites, 2003). This amount, although nearly 55 times the amount of pol- lock removal due to consumption by humpback whales, includes all gadid (not only pollock) species removals in all Alaskan waters. More importantly, these fish are likely larger (>60 cm vs. ^30 cm) than fish targeted by humpback whales. Although humpback whales generally feed on smaller age classes than are targeted by commercial fisheries or Steller sea lions (Perez and McAlister'; Kenney et al., 1997), consumption of younger age classes may affect future recruitment into the fishery. Barrett et al. (1990) stated that consumption of young cod (Gadus morhua) and saithe (PoUachius virens) by shags (Phalacrocorax aristotelis) and cormorants (P. carbo) in the Northeast Atlantic could be a limiting factor in recruitment in years of low stock size, even if consumption of these species was overestimated by an order of magnitude. Thus, it is noteworthy that the removal by humpback whales of an estimated 3.26x10'' kg of pollock (age 0-2) equals 30% of the 2002 commercial pollock harvest of 1.09x10'' kg (ages 3 to 8) for the entire Kodiak Island management area and 2.1% of the 2002 spawning bio- mass of pollock for the entire Gulf of Alaska, which was estimated at 1.58x108 kg (NMFSi"; NPFMCi'^^), These comparisons are based on mean estimates of prey removal and do not take into account model uncer- ' Wynne, K. M. 2002. Unpubl. data. Fishery Industrial Technology Center, Univ. Alaska Fairbanks, Kodiak, AK 99615. * Jemison, L. A. 2001. Summary of harbor seal diet data collected in Alaska from 1990-1999. In Harbor seal inves- tigations in Alaska iR. J. Small, ed.), p. 314-22. Ann. Rep. NOAA Grant NA 87Fx0300. Alaska Departmart of Fish and Game, P.O. Box 240020, Douglas, AK 99824. ^ Best, E. A., and G. St. Pierre. 1986. Pacific halibut as predator and prey. International Pacific Halibut Commis- sion Technical Report 21, 27 p. Website: http://www.iphc. washington.edu/halcom/pubs/techrep/tech0021.pdf [Accessed on 31 May 2003.1 10. n. 12 ggg next page for footnote text. 18 Fishery Bulletin 104(1) tainty. When uncertainty is considered, comparison to even the lower end of estimates of prey removal are still of note. For example, assuming that removal of juvenile pollock is equal to the lower estimate, or 2.37x10'' kg, the removal of pollock by humpback whales could still equal 21,7% of the 2002 commercial pollock catch and 1.5% of 2002 spawning biomass. Thus, it follows that if true consumption is actually closer to the upper esti- mates, the impact of prey removal by humpback whales would likely increase. The humpback whale represents only one of a myriad of marine consumers within the Kodiak Island ecosys- tem whose ecological role cannot be determined without sophisticated multispecies models and an analysis of ecosystem interactions. This study was designed to provide essential baseline data and a model for estimat- ing prey removal by foraging humpback whales. Our results show that the potential for biomass removal due to consumption by humpback whales is significant and that the foraging strategies of these whales warrant further investigation. Continued research efforts can improve estimates of biomass removal by identifying target prey, determining the degree of prey selectivity, and assessing variable foraging efficiency. Acknowledgments The authors are grateful for the insight and guidance provided by Janice M. Straley, Terry J. Quinn II, Bren- dan P. Kelly, and Chris Gabriele. Field and labora- tory assistance was provided by Lisa Baraff and Katie Brenner. Financial support for this project was provided by the Rasmuson Fisheries Research Center and NOAA Grant no. NA16FX1270 to the University of Alaska Gulf Apex Predator-Prey Project. All research was conducted until provisions of NMFS Scientific Research Permit 473-1433. Literature cited Anderson, P. J., and J. F. Piatt, 1999. Community reorganization in the Gulf of Alaska following ocean climate regime shift. Mar. Ecol. Prog. Ser. 189:117-123. 10 NMFS 2002. 2002. Gulf of Alaska groundfish quotas and preliminary catch in round metric tons. NMFS/AKR Fish Management, Juneau, AK 99802. " North Pacific Fisheries Management Council (NPFMC). 2003. Stock assessment and fishery evaluation (SAFE) report for the groundfish resources of the Gulf of Alaska, 846 p. Website: http://www.afsc.noaa.gov/refm/stocks/His- toric_Assess.htm (Accessed on 24 February 2004.] 12 NPFMC. 2002. SAFE report for the groundfish resources of the Gulf of Alaska, Assessment of Walleye Pollock Stock in the Gulf of Alaska, 90 p. Website: http://www.afsc. noaa.gov/refm/stocks/Historic_Assess.htm [Accessed on 24 February 2004.] Baker, C. S., and L. M. Herman. 1987. Alternative population estimates of humpback whales (Megaptera novaeangliae) in Hawaiian waters. Can. J. Zool. 65:2818-2821. Barrett. R. T., N. Rov, J. Loen, and W.A. Montevecchi. 1990. Diets of shags Phalacrocorax aristotelis and cor- morants P. carbo in Norway and possible implications in gadoid stock recruitment. Mar. Ecol. Prog. Ser. 66:205-218. Bax, N. J. 1991. A comparison of the fish biomass flow to fish, fish- eries, and mammals in six marine ecosystems. ICES Mar. Sci. Symp. 193:217-224. Benson, A. J., and A. W. Trites. 2002. Ecological effects of regime shifts in the Bering Sea and eastern North Pacific Ocean. Fish Fish. 3:95-113. Croll, D. A., B. R. Tershy, R. P. Hewitt, D. A. Demer, P. C. Fiedler, S. E. Smith, W. Armstrong. J. M. Popp, T. Keikhefer, V. R. Lopez, J. LTrban, and D. Gendron. 1998. An integrated approach to the foraging ecology of marine birds and mammals. Deep-Sea Res. 45: 1353-1371. Davis, N. D., K. W. Meyers, and Y. Ishida. 1997. Caloric value of high-seas salmon prey organ- isms and simulated salmon ocean growth and prey consumption. North Pacific Anadromous Fish Com- mission iNPAFCi 1:146-162. Estes, J. A. 1994. Top-level carnivores and ecosystem effects: ques- tions and approaches. In Linking species and ecosystems (C. G. Jones and J. H. Lawton, eds.) p. 151-158. Chap- man Hall, New York, NY. Gaskin, D. E. 1982. The ecology of whales and dolphins, 459 p. Heine- mann, London. Hairston, N. G.. and N. G. Hairston. 1993. Cause-effect relationships in energy flow, trophic structure, and interspecific interactions. Am. Nat. 142:379-411. Hatch, S. A., and G. A. Sanger. 1992. Puffins as samplers of juvenile pollock and other forage fish in the Gulf of Alaska. Mar. Ecol. Prog. Ser. 80:1-14. Hinga, K. R. 1979. The food requirements of whales in the Southern Hemisphere. Deep-Sea Res. 26:569-577. Innes, S., D. M. Lavigne, W. M. Eagle, and K. M. Kovacs. 1987. Feeding rates of seals and whales. J. Anim. Ecol. 56:115-130. Katona, S., P. Baxter, O. Brazier, S. Kraus, J. Perkins, and H. Whitehead. 1979. Identification of humpback whales by fluke photographs. In Vol. 3: behavior of marine animals (H. E. Winn and B. L. 011a, eds.), p. 33-44. Pages Plenum Press, New York, NY. Kenney, R. D., P. M. Payne, D. W. Heinmann, and H. E. Winn. 1996. Shifts in northeast shelf cetacean distributions relative to trends in the Gulf of Maine/Georges Bank finfish abundance. In The northeast shelf ecosys- tem: assessment, sustainability, and management (K. Sherman, N. A. Jaworski, and T. J. Smayda, eds.). p. 169-197. Blackwell Science, Cambridge, MA. Kenney, R. D., G. P. Scott, T. J. Thompson, and H. E. Winn. 1997. Estimates of prey consumption and trophic impacts Witteveen et al.: Effect of prey removal by Megaptera novaeangliae on fisfi abundance 19 of cetaceans in the USA Northeast continental shelf ecosystem. J. Northwest Atl. Fish. Sci. 22: 55-171. Kleiber, M. 1961. The fire of life; an introduction to animal energet- ics, 454 p. John Wiley and Sons, New York, NY. Klumov, S. K. 1963. Food and helminth fauna of whalebone whales in the main whaling regions of the world ocean. Trudy Instituta Okeanologii 71:94-194. 1966. Plankton and the feeding of the whalebone whales (Mystacoceti). Trans. Inst. Oceanol. 51:142-156. Krieger, K. J., and B. L. Wing. 1984. Hydroacoustic surveys and identification of hump- back whale forage in Glacier Bay, Stephens Passage, and Frederick Sound, Southeastern Alaska, Summer 1983. NOAA Tech. Memo. NMFS F/NWC-66, 60 p. Auke Bay Laboratory National Marine Fisheries Service P.O. Box 210155 Auke Bay, Alaska 99821. 1986. Hydroacoustic monitoring of prey to determine humpback whale movements. NOAA Tech. Memo. NMFS F/NWC-98, 62 p. Auke Bay Laboratory National Marine Fisheries Service P.O. Box 210155 Auke Bay, Alaska 99821. Laevastu, T., and H. A. Larkins. 1981. Marine fisheries ecosystems, its quantitative evalu- ation in management, 167 p. Fishing News Books, Farnham, Surrey. LTK. Laws, R. M. 1977. Seals and whales in the Southern Ocean. Phil. Trans. R. Soc. of London Series B Biol. Sci. 279:81-96. 1985. The ecology of the Southern Ocean. Am. Sci. 73:26-40. Livingston, P. A., A. Ward, G. M. Lang, and M-S. Yang. 1993. Groundfish food habits and predation on com- mercially important prey species in the eastern Bering Sea from 1987 to 1989. NOAA Tech. Memo. NMFS F/NWC-54, 192 p. Lockyer, C. 1981. Estimation of the energy costs of growth, main- tenance and reproduction in the female minke whale (Balaenoptera acutorostrata), from the Southern Hemis- phere. Rep. Int. Whal. Comm. 31:337-343. Markussen, N. H., M. Ryg, and C. Lydersen. 1992. Food consumption of the NE Atlantic minke whale (Balaenoptera acutorostrata) population estimated with a simulation model. ICES J. Mar. Sci. 49:317-323. Merrick, R. L. 1995. The relationship of the foraging ecology of Steller sea lions (Eumetopias jubatusi to their population decline in Alaska. Ph.D. diss., 172 p. Univ. Washington, Seattle, WA. 1997. Current and historical roles of apex predators in the Bering Sea ecosystem. J. Northwest Atl. Fish. Sci. 22:343-355. Mitchell, E. 1973. Draft report on humpback whales taken under spe- cial scientific permit by eastern Canadian land stations, 1969-1971. Rep. Int. Whal. Comm. 23:138-154. Murra, K. A., C. L. Buck. S. D. Kildaw, J. B. Gamble, C. T. Williams. 2003. Forage location, diet, and productivity of black- legged kittiwakes in 2001 and 2002 in Chiniak Bay, Kodiak Alaska (Abstract). In Marine science in the Northeast Pacific: science for resource dependent com- munities joint science symposium (book of abstracts); January 13-17, 2003, Anchorage, Alaska. NRC (National Research Council). 1996. The Bering Sea ecosystem, 307 p. National Acad- emy Press, Washington D.C. Nemoto, T. 1957. Foods of baleen whales in the northern Pacific. Sci. Rep. Whales Res. Inst. Tokyo 12:33-89. 1959. Foods of baleen whales with reference to whale movements. Sci. Rep. Whales Res. Inst. Tokyo 14:244- 290. 1970. Feeding patterns of baleen whales in the ocean. In Marine food chains (J. H. Steele, ed.) p. 241-252. Oliver and Boyd, Edinburgh, Scotland. Nord0y, E. S., L. P. Falkov, P.-E. Matensson, and A. S. Blix. 1995. Food requirements of northeast Atlantic minke whales. In Whales, seals, fish, and man (A. S. Blix, L. Wall0e, and 0. Ulltang, eds.), p. 307-317. Elsevier Science, Amsterdam, The Netherlands. Overholtz, W. J., S.A. Murawski and K.L. Foster. 1991. Impacts of predatory fish, marine mammals, and seabirds on pelagic fish ecosystems of the northeastern USA. In Multispecies models relevant to management of living resources (N. Daan and M.P. Sissenwine, eds.) p. 198-208. ICES Mar. Sci. Symp. 193. International Council for the Exploration of the Sea (ICES), Copen- hagen V, Denmark. Pascual, M.. R. Hilborn, L. Fritz, H. Xi, and J. Moss. 1993. Modeling the trophic relationships between fish and marine mammal populations in Alaskan waters. In Is it food? Addressing marine mammal and seabird declines, p. 30-44. Alaska Sea Grant Rep. 93-01. Alaska Sea Grant Program, Univ. Alaska, Fairbanks, AK. Payne, S. A., B. A. Johnson, and R. S. Otto. 1999. Proximate composition of some northeastern Pacific forage fish species. Fish. Oceanogr. 8:159-177. Payne, P. M., J. R. Nicolas, L. O'Brien, and K. D. Powers. 1986. The distibution of the humpback whales Megap- tera novaeangliae on Georges Bank and in the Gulf of Maine in relation to densities of the sand eel Amniodytes americanus. Fish. Bull. 84:271-277. Payne, P. M., D. N. Wiley, S. B. Young, S. Pittman, P. J. Clapham. and J. W. Jossi. 1990. Recent fluctuations in the abundance of baleen whales in the southern Gulf of Maine in relation to changes in selected prey. Fish. Bull. 88:687-696. Perez, M. A., and W .B. McAlister. 1993. Estimates of food consumption by marine mam- mals in the eastern Bering Sea. NOAA Tech. Memo. NMFS-AFSC-14, 36 p. National Marine Fisheries Ser- vice Alaska Fisheries Science Center 7600 Sand Point Way N.E., Building 4, Seattle, WA 98115. Perry, S. L., D. P. DeMaster, and G. K. Silber. 1999. The great whales: History and status of six species listed as endangered under the Endangered Species Act of 1973. Mar. Fish. Rev. 61(l):l-74. Piatt, J. F., D. A. Methven, A. E. Burger, R. L. McLagan. V. Mercer, and E. Creelman. 1989. Baleen whales and their prey in a coastal environment. Can. J. Zool. 67:1523-1530. Piatt, J. F., D. D. Roby, L. Henkel, and K. Neuman. 1997. Habitat use, diet, and breeding biology of tufted puffins in Prince William Sound, Alaska. Northwest. Nat. 78:102-109. Rice, D. W. 1978. The humpback whale in the North Pacific: distribu- tion, exploitation and numbers. Appendix 4 in Report on a workshop on problems related to humpback whales 20 Fishery Bulletin 104(1) (Megaptera novaeangliae) in Hawaii (K. S. Norris and R. R. Reeves, eds.), p. 29-44. U.S. Dep. Commer., Nat. Tech. Info. Serv. PB-280 794, Springfield, VA. Schmidt-Nielson, K. 1997. Animal physiology: adaptation and environment, S"' ed., 607 p. Cambridge Press, Cambridge, UK. Sigurjonsson, J., and G. A. Vikingsson. 1998. Seasonal abundance of and estimated food consump- tion by cetaceans in Icelandic and adjacent waters. J. Northwest Atl. Fish. Sci.. 22:271-287 Springer, A. M., J. A. Estes, G. B. van Vilet, T. M. Williams, D. F. Doak, E. M. Banner, K. A. Forney, and B. Pfister. 2003. Sequential megafaunal collapse in the North Pacific Ocean: An ongoing legacy of industrial whaling? Proc. Acad. Nat. Sci. 100:1222:3-12228. Sumich, J. L. 1983. Swimming velocities, breathing patterns, and esti- mated costs of locomotion in migrating gray whales, Eschrichtius robustus. Can. J. Zool. 61:647-652. Thompson, R.J. 1940. Analysis of stomach contents taken during the years 1937 and 1938 from the North Pacific. M.Sc. thesis, 82 p. Univ. Washington, Seattle. WA. Tomilin, A. G. 1954. Adaptive types in the order Cetacea (the prob- lem of ecological classification of Cetacea). Zool. Zh. 33:677-692. Trites, A. W. 1997. The role of pinnipeds in the ecosystem. In Pin- niped populations, eastern north Pacific: status, trends, and issues. A symposium of the 127"^ annual meeting of the American Fisheries Society (G. Stong, J. Goebel, and S. Webster, eds.), p. 31-38. New England Aquarium, Conservation Department, Boston, MA. Trites, A. W. and D. Pauly 1998. Estimating mean body masses of marine mam- mals from maximum body lengths. Can. J. Zool. 78:886-896. Trites, A. W., V. Christensen, and D. Pauly. 1997. Competition between fisheries and marine mam- mals for prey and primary production in the Pacific Ocean. J. Northwest Atl. Fish. Sci. 22:173-187. van Franeker, J. A. 1992. Top predators as indicators for ecosystem events in the confluence zone and marginal ice zone of the Weddell and Scotia seas, Antarctica, November 1988 to January 1989 (EPOS Leg 2). Polar Biol. 12:93-102. Wahrenbrock, E. A., G. F. Maruschak, R. Eisner, and D. W. Kenney. 1974. Respiration and metabolism in two baleen whale calves. Mar. Fish. Rev. 36:1-9. Whitehead, H., and J. E. Carscadden. 1985. Predicting inshore whale abundance — whales and capelin off the Newfoundland coast. Can. J. Fish. Aquat. Sci. 42:976-981. Winship, A. J., and A. W. Trites. 2003. Prey consumption of Steller sea lions iEunietopias jubatus) off Alaska: how much prey do they require? Fish. Bull. 101:147-163. Witteveen, B. H. 2003. Abundance and feeding ecology of humpback whales (Megaptera novaeangliae! in Kodiak. Alaska. M.Sc. thesis, 109 p. Univ. Alaska Fairbanks. Fairbanks, AK. Yang, M. S. 1995. Food habits and diet overlap of arrowtooth flounder (Atheresthes stomias) and Pacific halibut iHippoglossus stenolepis) in the Gulf of Alaska. In Proceedings of the international symposium on North Pacific flatfish, p. 205-223. Alaska Sea Grant Program Report 95-04, Univ. of Alaska, Fairbanks, AK. 21 Abstract — Survey standardization procedures can reduce the variabil- ity in trawl catch efficiency thus producing more precise estimates of biomass. One such procedure, towing with equal amounts of trawl warp on both sides of the net, was experimentally investigated for its importance in determining optimal trawl geometry and for evaluating the effectiveness of the recent National Oceanic and Atmospheric Adminis- tration (NOAA) national protocol on accurate measurement of trawl warps. This recent standard for measuring warp length requires that the differ- ence between warp lengths can be no more than 49f of the distance between the otter doors measured along the bridles and footrope. Trawl perfor- mance data from repetitive towing with warp differentials of 0. 3, 5, 7, 9, 11, and 20 m were analyzed for their effect on three determinants of flatfish catch efficiency: footrope distance off-bottom, bridle length in contact with the bottom, and area swept by the net. Our results indi- cated that the distortion of the trawl caused by asymmetry in trawl warp length could have a negative influ- ence on flatfish catch efficiency. At a difference of 7 m in warp length, the NOAA 4'7f threshold value for the 83- 112 Eastern survey trawl used in our study, we found no effect on the acous- tic-based measures of door spread, wing spread, and headrope height off- bottom. However, the sensitivity of the trawl to 7 m of warp offset could be seen as footrope distances off-bottom increased slightly (particularly in the center region of the net where flatfish escapement is highest), and as the width of the bridle path responsible for flatfish herding, together with the effective net width, was reduced. For this survey trawl, a NOAA threshold value of 4% should be considered a maximum. A more conservative value (less than 4%) would likely reduce potential bias in estimates of relative abundance caused by large differences in warp length approaching 7 m. Variation in trawl geometry due to unequal warp length Kenneth L. Weinberg David A. Somerton National Marine Fisheries Service Alaska Fisheries Science Center 7600 Sand Point Way N E Seattle, Washington 98115 0070 E-mail address (for K L, Weinberg): l0 at zero offset, the tangential velocity was not significantly different than zero it-tesi. P=0.71). This result indicates that the alignment of the experimental tows in relation to the pre- vailing current was sufficient to reduce the cross current to negligible levels. Net and door measurements At zero offset, the mean door spread obtained was 61.9 m. The mean wing spread was 17.1 m, and the mean headrope height was 2.0 m. Differ- ences in the means of all three quantities were not apparent at lower offsets, however headrope height and wing spread were more sensitive to changes in large offsets because both were signifi- cantly (P<0.05l greater than the zero offset means when offset was increased to 11 m, whereas door spread did not differ significantly until the offset was approximately 14 m (Fig. 6). Bridle and footrope distance off-bottom by position Bridle and footrope off-bottom distance varied consider- ably with position, not only with respect to the mean value but also with respect to the sensitivity of the mean to changes in offset. At zero offset, the mean off-bottom distance of the bridle declined from 12.3 cm at 50 m from the wing tips, to 3.2 cm at 40 m and 2.0 cm at 25 m (Fig. 7). Mean off-bottom distance remained small along the footrope, varying from 1.7 cm at 1 m behind the wing tip to 2.5 cm at the corner and 1.9 cm at the center. The mean response to changes in offset varied greatly by position. Along the bridles, the most sensitive location was at 50 m, where off-bottom distance increased on the short side and decreased on the long side with increasing offset (Fig. 7). At 40 m. a similar pattern was repeated, but for most offsets on the long side, the off-bottom dis- tance was near the minimum recorded, indicating that the bridle was resting on the bottom. At 25 m, the bridle was nearly always in contact with the bottom and off- bottom distance was insensitive to variations in warp offset. Along the footrope, the most sensitive position was the corner where off-bottom distance increased greatly with offset, particularly with positive offsets due to the relaxation in warp tension. At the center of the footrope, off-bottom distance was also sensitive to warp offset, re- sponding almost identically on the long and short sides. At 1 m behind the wing tip, sensitivity to warp offset was quite low and the off-bottom distance indicated that Figure 5 Current velocity (in m/sec) measured at the center of the hea- drope for each tow is shown separated into the component per- pendicular to the headrope (O) and the component tangential to the headrope ( + ). The solid and dashed lines connect the means at each offset increment. Note, for clarity, that the offsets are incremented by plus or minus 0.1 m for the tangential and perpendicular components. the footrope was in contact with the bottom except for large offsets on the long side. An alternate method of assessing the sensitivity of geometry of the 83-112 Eastern trawl to changes in offset is to determine if the mean off-bottom distance at 7 m, the maximum offset allowed under the NOAA protocols for the 83-112 Eastern trawl, differs statisti- cally from the mean off-bottom distance at zero offset. Based on the bootstrapped confidence intervals (Fig. 7), off-bottom distance is significantly different from what it is at zero offset at the 50-m and 40-m bridle positions and at the center and corner footrope positions but is not significantly different at the wing and the 25-m bridle position. Bridle shape and herding area To understand better how the change in tension that accompanies offsets in warp leads to changes in bridle shape, we show the mean off-bottom distances plotted against the BCS positions on the wing and bridles for both the short side and long side of the trawl. From this perspective it is clear that as the tension is increased, off-bottom distance increases on the forward part of the bridle. Likewise, as the tension is reduced, the off-bottom distance decreases (Fig. 8). For flatfish, the effect of these changes in off-bottom distance is a change in the area subjected to herding stimuli. For the case where the reaction height is 1 cm, the bridle contact length is determined by the intersection of the line depicting Weinberg and Somerton Variation in trawl geometry due to unequal warp length 27 the reaction height with the lines depicting the bridle shape at each offset (Fig. 8). The change in these lengths on the short and long sides of the trawl is asymmetric with changes in warp offset (Fig 9). For the long side, bridle contact length increases linearly with positive offset. However, for the short side, bridle contact length decreases nonlinearly with warp offset — the greatest changes occurring with small offsets. This difference likely leads to a change in the total width of the herding area with changes in warp offset. If, for example, it is assumed that the angle-of- attack (a) is the same for the long and short sides of the trawl, then the width of the herded area declines to a minimum at about 8 m offset, at which the herded area is reduced by W.SVr compared to that at zero offset. Headrope shape and effective net width With increasing difference in warp length, the model we used to describe headrope shape predicts three distinct changes in shape. First, the headrope is distorted so that the wing tip on the short side of the trawl precedes that on the long side in the direction of travel (Fig. 10). The difference in the forward position of the wing tips, however, is much less than the warp offset. For example, an 11-m difference in warp length resulted in an offset in the position of the wing tips of only 2-3 m. This dif- ference occurs because the increased tension on the short warp changes the catenary in both the bridles and the warps (i.e., both become effectively longer as the sag is reduced). Second, the headrope is distorted so that its center is increasingly displaced away from the midpoint between wings and toward the short side of the trawl. When this displacement occurs, the perpendicular at the center of the headrope is no longer aligned with the direction of travel. Third, the headrope is distorted so that the effective width of the net (i.e., the wing spread projected to the line perpendicular to the towing direc- tion) becomes increasingly shorter than the distance measured by the acoustic net sensors. The difference between the effective and the measured net width is negligible for offsets up to 7 m but rapidly increases at greater offsets (Fig. 11). Footrope shape viewed from In front of the net The distance of the footrope off-bottom, when viewed from a position in front of the net, increases with increasing offset; however, the location of the maximum off-bottom distance in relation to the midpoint between wings, shifts slightly with increasing offset (Fig. 12). With off- sets of 9 m or less, the position of maximum off-bottom distance is at the corner of the footrope on the long side of the trawl. However, with increasing offset, the shift in the position of the footrope corner changes because of the rotation of the trawl in relation to the direction of travel; and at a 20-m offset the footrope corner on the long side of the trawl is positioned, when viewed on a plane perpendicular to the direction of travel, almost exactly midway between the wing tips. 14 16 18 20 10 12 Id 16 18 20 i ^ + + ^ pqqrt 10 12 14 16 18 Offset (m) Figure 6 Mean door spread, wing spread, and headrope height are shown ( + ) plotted against offset increment. The means for all values of offset increment were fitted with a cubic spline function (solid curve). Bootstrapped 959^ confidence bounds are shown with shading. Also shown are the mean door spread, wing spread, and headrope height for treatments with zero offset (solid horizontal line) and the corresponding bootstrapped 95% confidence bounds Idashed horizontal lines). 28 Fishery Bulletin 104(1) Bridle 50 m forward of wing tip "H :l4 i * * :'i^/- : i~ -20 -15 -10 -5 0 5 10 15 20 18 16 14-^ 10 Bridle 40 m forward of wing tip "--t - I t £ ^ T . m -i -20 -15 -10 10 15 20 Bridle 25 m forward of wing tip 3.5- 10 9 8 7 6 5 4 3 2 1 0 30 25 20 15 Footrope 1 m aft of wing tip Footrope corner 20 18 16 14- 12 10 8 6 4 2 Footrope center / -20 -15 -10 10 15 Offset (m) Figure 7 Mean bridle and footrope distance at each bottom contact sensor position is shown ( + i plotted against offset increment. The means for all values of offset increment were fitted with a cubic spline function (solid curve). Bootstrapped 95':r confidence bounds are shown with shading. Also shown is the mean distance off-bottom for only treatments with zero offset (solid horizontal line) and the bootstrapped 95% confidence bounds (dashed horizontal lines). Weinberg and Somerton Variation in trawl geometry due to unequal warp length 29 Long 25- 20- 15 - /° 10- /.. 5 ^^^^^ 0- 10 20 30 40 10 20 30 40 Distance from wing tip (m) 50 Shon ,20 25 Ir 20- r 15 - i: 10- J 1/ 5- ^^. / 0 - 50 Figure 8 Mean off-bottom distance is shown plotted against the distance measured from the wing tip to the positions of the wing and the three bridle bottom contact sensors. This approximation to the shape of the bridle when viewed laterally is shown for each of the offset incre- ments for both the side with the longer warp and. the side with the shorter warp. The dashed line represents the hypothetical reaction height of a fish. The intersection of the dashed line with the solid line for each configura- tion defines the bridle length that is sufficiently close to the bottom to elicit a herding response. 50 - , - - -° ' _ 45 - B) 1 40- o--o--°-"" .0- ' .- O' 03 \ § 35- \ 03 \ ^ 30 - o\ 1^ '-' — D 25 - 0 5 10 15 20 24.5 - \ 1 - 24.0 - Q. \ P 23.5 - \ 1 JD \ 1 0) \ 1 % 23.0 - \ 1-10.3% y 1 22.5- \ ^y^ 22.0 - ^ — "^^^ 0 5 10 15 20 Offset (m) Figure 9 The length of the bridle with the off-bottom distance <1 cm is shown plotted against the offset increment in meters for both the short warp (solid line) and long warp (dashed line) sides of the trawl (upper panel). In both cases the lines are represented using cubic spline smoothing functions. The width of the swept bridle path as a function of warp offset is represented in the lower panel. Discussion There are two distinct approaches forjudging whether a difference in warp length between the sides of a survey trawl will lead to a significant bias in estimates of rela- tive abundance. In both approaches we focused on the adequacy of the maximum 7-m offset allowed for the 83-112 Eastern trawl under NOAA trawl survey pro- tocols. In the first approach, we simply asked whether, given the sampling effort used in the experiment, any of the measured dimensions at 7-m offset were statis- tically different from zero offset. In our experiment, none of the three standard measures of trawl geometry (i.e., door spread, wing spread, and headrope height) differed from mean values at zero offset. This finding indicates that either these dimensions are fairly robust to changes in warp offset or that the acoustic measure- ment of these dimensions was insufficiently precise to detect a difference. Off-bottom distance, however, was significantly different at the two forward positions on the bridles and along the footrope at the center and corner positions. From the perspective of trawl survey standardiza- tion, however, the detectability of changes in geometry is not of primary importance; these changes, however, may produce a significant effect on estimates of relative abundance. Bias in these estimates could result either because the change in trawl geometry leads to an inac- 30 Fishery Bulletin 104(1) 30 25 20 15 10 5 0 30 25 20 15 10 5 0 30 25 20 15 10 5 0 3 m offset 5 m offset -15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15 7 m offset 9 m offset -15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15 1 1 m offset 30 ■ 20 m offset 25- 20 - 15 ,.'] 10 ■ K j 5 \ I 0- \ — r -15 -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15 Meters Figure 10 Estimated net shape (curve i and position of headrope center (O) at varying levels of warp offset. The mean distance between wing tips that was acoustically measured during the experimental tows is indicated with a dashed line. The ma.ximum lateral dimensions of the net are indicated with solid vertical lines. The distance between these lines is the effective net width needed for estimat- ing the area swept by the net. 5 7 9 Offset (m) Figure 11 The mean distance between wing tips mea- sured acoustically lOl during the experimental tows (net width) and the calculated effective net width (-t-) are shown plotted against the offset increment in meters. Note that there is little difference between the two measures of net width until the offset increment is increased to 9 m. curate measurement of swept area or because it leads to a change in catch efficiency. Relative abundance indices produced for the eastern Bering Sea shelf survey are based on catch per area swept between the trawl wings. As the net distorts on account of differential warp length, the effective net width will become increasing less than the width that is acousti- cally measured during the survey. Thus, net width will become increasingly overestimated and relative abundance of fish species therefore will be underestimated. At 7-m offset, however, the measured net width differed from the ef- fective net width by only 0.5%, and therefore this source of error is unlikely to contribute to bias in the swept area estimates. However, the difference between measured and effective net width increases rapidly at greater offsets and could present a problem if a less restrictive threshold value of offset were used. Catch efficiency of the 83-112 Eastern trawl depends primarily on 1) herding by the bridles, doors, and the mud clouds they create; 2) the escapement under the footrope; and 3) the es- capement through the mesh in the body of the net. The relative importance of these three processes, however, will vary for the major spe- cies groups that are targeted in the surveys. Gadoids (primarily walleye pollock [Theragra chalcogramma] and Pacific cod [Gadus macro- cephalus] appear to have little or no herding response to the 83-112 Eastern trawl (Somer- ton, 2004) and rarely pass under the footrope (Somerton, unpubl. data). However, both Pa- Weinberg and Somerton; Variation in trawl geometry due to unequal warp length 31 20 15 ■ 10 ig sh )rt 10 ■ 5 U \ " ~ r 5 10 15 7 m offset long 5 m offset 20 15 long sh )r1 10 ■ 5 U 1 , ^ 5 10 9 m offset loiig 15 0 Distance from wing tip (m) Figure 12 Mean off-bottom distances (in cm) at the five bottom contact sensor positions along the footrope (circles) are shown for each warp offset. The positions are projected onto the wing tip to wing tip plane to depict the footrope as it would appear if one were looking into the net from the direction of travel. The vertical solid lines indicate the positions of the wing tips on the long warp and short warp sides of the trawl. The dashed line indicates the midpoint between wing tips. Note that the projection considers the reduction in effective net width with increasing offset. cific cod and walleye pollock are found gilled in the body of the net; therefore some mesh escapement may occur, especially if any distortion of the net results in altered water flow through the meshes. Video observa- tions of crabs (snow and Tanner crabs [Chionoecetes sp.] and king crabs [Paralithodes sp.] show that they also exhibit little or no herding response to the 83-112 Eastern trawl (Weinberg, unpubl. data). However, both 32 Fishery Bulletin 104(1) Chionoecetes species (Somerton and Otto, 19991 and red king crab (Weinberg et al., 20041 do escape under the footrope. Flatfishes (including yellowfin sole [Limanda aspera], flathead sole [Hippoglossoides elassodon], and rock sole [Lepidopsetta biliueata] display a strong herd- ing response to the 83-112 Eastern trawl (Somerton and Munro, 2001); as much as 49% of the catch con- sisted of fish that were herded by the bridles into the net path. Likewise, flatfishes are readily capable of escaping under the footrope and, for species such as yellowfin sole, at least 25% of the largest individuals escape in this manner (Munro and Somerton, 2002). Thus, the capture efficiency of the trawl is species- specific. The change in catch efficiency due to warp offset is likely minimal for species not captured as a function of herding and footrope escapement behaviors; however, because flatfish are susceptible to herding and are adept at footrope escapement, their catch rate could potentially be affected most by warp offsets. Bridle efficiency (i.e., the fraction of fish in the area between the wing tips and doors that are herded into the path of the net) for flatfish catch is strongly influ- enced by the size of the herding area or area swept by the bridles because flatfish are stimulated to herd by the close approach or direct contact of the lower bridle. Although we were able to measure the off-bottom dis- tance along the bridle and thereby predict the shape of the bridle, there is still considerable uncertainty as to the exact size of the herding area because the reaction height of a fish will vary with species, size, physiological state, state of arousal to the approaching bridle, viewing conditions for the fish, and, perhaps, other variables. Additionally, there is uncertainty in the estimate of the size of the herding area because it is based on the as- sumption of symmetry in the bridle angle-of-attack — a symmetry that is increasingly untenable with increasing offset. Despite this uncertainty, it is likely that the loss of herding area on one side of the trawl is not countered by an increase on the other side; thus some overall loss of herding efficiency is to be expected. In the hypotheti- cal case chosen in our study, the reduction in herded area was 10.3%, which when applied to a strong herding flatfish such as rock sole (the herded component of the catch has been estimated to be about 49%, Somerton and Munro, 2001), the expected reduction in catch with an 8-m offset would be roughly 5%. Flatfish escapement under the footrope will be influ- enced not only by the increase in off-bottom distance but also by the location along the footrope where the increase occurs. At a 7-m offset, the footrope off-bottom distance is about 2 cm higher in the footrope corner on the long side of the trawl and approximately 1 cm higher at the center and opposing corner than at zero offset (Fig. 12). Footrope off-bottom distances increased appreciably with greater offsets. Weinberg et al. (2002) demonstrated that flatfish escapement can increase with similar increases in footrope off-bottom distance; however their study focused on a different trawl and considered escapement for the entire footrope rather than by position along the footrope. Although we are unaware of any studies that quantify escapement rate by position along the footrope, our video observations indicate that fiatfish are less likely to escape under the footrope near the wings than in the center (Somerton, unpubl. data). Because it is the center portion of the foo- trope where most of the increases in off-bottom distance occur in the 83-112 trawl, flatfish escapement is likely increased and potentially could represent a significant, but presently unquantifiable, loss in catch and a source of bias in estimates of relative abundance. In conclusion, most aspects of the 83-112 Eastern trawl geometry were significantly degraded by warp offset differences equal to or greater than 7 m compared to zero offset. More importantly, the locations where the detectable differences occurred could affect catch effi- ciency; therefore a NOAA threshold value of 4% should be considered a maximum value for the 83-112 Eastern trawl and perhaps a more conservative value (less than 4%) would be prudent. However, given today's standard- ized survey procedures for measuring warp and for real- time monitoring of warp offset, the probability of warp offsets even approaching 7 m is highly unlikely on our surveys when locked-winches are used. Likewise, we argue that any appreciable differences in warp lengths between sides due to stretching are unrealistic because AFSC charter vessels use large diameter, compressed, solid-core wire. In fact, a review of the 413 hauls made during the 2004 EBS survey revealed that only three tows had a recorded maximum 1-m length difference between sides (Weinberg, unpubl. data). Acknowledgments We would like to thank Captain Brad Lougheed and the FV Vesteraalen crew, Niles Griffen, Waldemar Janezak, Van Ngo, and Todd Becker for their outstanding profes- sionalism, positive attitudes, and constant attention to detail; AFSC scientists Stan Kotwicki and Dennis Benja- min for their assistance at sea; reviewers Guy Fleischer and Henry Milliken for their helpful comments; and the AFSC editorial staff, including Gary Duker, Jim Lee, and Kama McKinney for their help during the in-house review and manuscript preparation process. Literature cited Efron, B., and R. Tibshirani. 1993. An introduction to the bootstrap, 436 p. Chapman and Hall, New York, NY. Fridman, A. L. 1969. Theory and design of commercial fishing gear. (Transl. from Russian by Israel Program Sci. Transl., Jerusalem) 1973, 489 p. [Available as TT 71-50129 from Natl. Tech. Inf. Serv., Springfield, VA.] Munro, P. T, and D. A. Somerton. 2002. Estimating net efficiency of a survey trawl for flatfishes. Fish. Res. 55:267-279. Somerton, D. A. 2004. Do Pacific cod (Gadus macrocephalus) and wall- Weinberg and Somerton: Variation in trawl geometry due to unequal warp length 33 eye pollock (Theragra chalcogramrna) lack a herding response to the doors, bridles and mudclouds of survey trawls? ICES J. Mar. Sci. 61:1186-1189. 2003. Bridle efficiency of a survey trawl for flatfish: measuring the length of the bridles in contact with the bottom. Fish. Res. 60:273-279. Somerton, D. A., and P. T. Munro. 2001. Bridle efficiency of a survey trawl for flatfish. Fish. Bull. 99:641-652. Somerton, D. A., and R. S. Otto. 1999. Net efficiency of a survey trawl for snow crab, Chionoecetes opilio, and Tanner crab, C. hairdi . Fish. Bull. 97:617-625. Somerton, D. A., and K. L. Weinberg. 2001. The affect of speed through the water on footrope contact of a survey trawl. Fish. Res. 53:17-24. Stauffer. G. 2004. NOAA protocols for groundfish bottom trawl sur- veys of the nation's fishery resources. NOAA. Tech. Memo. NMFS-F/SPO-65, 205 p. Alaska Fisheries Science Center, 7600 Sand Point Way N.E., Seattle, WA, 98115. Venables, W. N., and B. D. Ripley. 1994. Modern applied statistics with S-plus. 462 p. Springer-Verlag, New York, NY. Weinberg, K. L., D. A. Somerton, and P. T. Munro. 2002. The effect of trawl speed on the footrope capture efficiency of a survey trawl. Fish. Res. 58:303-313. Weinberg, K. L., R. S. Otto, and D. A. Somerton. 2004. Capture probability of a survey trawl for red king crab iParalithodes camtschaticus). Fish. Bull. 102:740-749. Appendix 1: Estimating headrope and footrope shape when the warps differ in length If a trawl headrope has the same shape as a flexible twine under a uniformly distributed load, then the shape of the headrope can be approximated as a quadratic (parabolic) function (Fridman, 1969; p. 84) as y = cx (1) where c is a constant controlling the shape (Fig. Al). As the headrope is distorted by a differential in warp length, not only does the value of c change, but the headrope is displaced along the path of the parabola, so that its center is no longer aligned with the vertex of the parabola. A unique solution to the shape of the headrope when it is distorted in this manner can be determined from three types of data: the total headrope length (L), the measured distance between the wing tips (W), and the measured slope (tangent) of the parabola at the center of the headrope (tan). The third quantity can be obtained from the V (perpendicular to the footrope) and U (tangential to the footrope) velocities measured by the headrope speed sensor as the quotient U/V. With these quantities, the solution can be obtained as follows. A small length interval measured along the headrope can be expressed as X(tn) Figure Al Shape of a trawl headrope as described by a parabola. The total length of the headrope (shown with a solid line) is equal to L. The measured width of the trawl (shown with a dashed line) is equal to W. The circle indicates the center of the headrope where a speed sensor is located. The speed sensor measures water speed both perpendicu- lar and parallel to the headrope. ds = (dy +dx (2) Which, after substitution of the derivative of Equation 1, is ds = [l + (2cx)^fdx. (3) The length of any segment of the headrope, measured from the port end (->;/„„ ^r'' i^ then S= \[l + (2cx)^y ds. (4) A segment equal to the total length of the headrope is obtained by integrating up to the starboard end The solution is approximated numerically in two stag- es. First, for a trial value of c. Equation 4 is integrated from trial values of Ji:,^,,^,^ up to the value of x at which S=L/2 (i.e., -Vjjjj^^/p). The tangent at this position is then evaluated as 2cx,. nldlr (based on the derivative of Eq. 1). This process is then repeated iteratively to find the value of .T:,„„.j,r for the specified value of c at which the calculated tangent equals the tangent value determined from the headrope speed sensor. The value of -v, is 34 Fishery Bulletin 104(1) then determined by integrating Equation 4 from Xi„,^.^,^ to the value of .v at which S=L. The wing tip to wing tip distance IJ,,,,,,^,,^ is then calculated as 1 D. •ingtip -yio ,)^+(a:„ (5) where .v,,^^^. and yi„,^,^,, are obtained from Equation 1. In the second stage, c is varied and the above process is repeated iteratively until the value of O,, ,„„„^j is found that is closest to the measured net spread. At this point the calculated values of Z),,,„„„^, and the tangent at the headrope center will equal the measured values. The headrope-shape model for each offset was used to project the off-bottom distances measured at the five positions along the footrope onto a plane orientated perpendicular to the direction of travel to depict the shape of the footrope as it would appear from a position in front of the trawl. To do this, a shape function was developed for the footrope. Assuming that the coordi- nates of the endpoints of the footrope (-v,,,,, ^,^, -v^^,^^,,^, >'/o„,pr' y ^) were the same as the headrope. Equation 5 was iteratively integrated with varying values of c until the estimated value of the footrope length (S) equaled the true length (34.1 m): S= \ [l + (2cx)-)' dx. (6) Once c is determined, Equation 4 is integrated to find the value of .r associated with the value of S at each BCS (bottom current sensor) position on the footrope. 35 Abstract — Three aspects of a survey bdttdin trawl performance — 1) trawl geometry (i.e., net spread, door spread, and headrope height); 2) footrope dis- tance off-bottom; and 3) bridle dis- tance off-bottom — were compared among hauls by using either of two autotrawl systems (equal tension and net symmetry) and hauls conducted with towing cables of equal length and locked winches. The effects of environmental conditions, vessel heave, crabbing (i.e., the difference between vessel heading and actual vessel course over ground), and bottom current on trawl performance with three trawling modes were investigated. Means and standard deviations of trawl geometry mea- sures were not significantly different between autotrawl and locked-winch systems. Bottom trawls performed better with either autotrawl system as compared to trawling with locked winches by reducing the variance and increasing the symmetry of the footrope contact with the bottom. The equal tension autotrawl system was most effective in counteracting effects of environmental conditions on footrope bottom contact. Footrope bottom contact was most influenced by environmental conditions during tows with locked winches. Both of the autotrawl systems also reduced the variance and increased the symmetry of bridle bottom contact. Autotrawl systems proved to be effective in decreasing the effects of environmental factors on some aspects of trawl performance and, as a result. have the potential to reduce among- haul variance in catchability of survey trawls. Therefore, by incorporating an autotrawl system into standard survey procedures, precision of survey estimates of relative abundance may be improved. The effect of autotrawl systems on the performance of a survey trawl Stan Kotwicki* Kenneth L. Weinberg* David A. Somerton National Marine Fisheries Service Alaska Fisheries Science Center, 7600 Sand Point Way N.E. Seattle, WA 98115 E-mail address (for K L Weinberg, contact author) ken weinbergig'noaa gov 'Equal authorship Manuscript submitted 27 October 2004 to the Scientific Editor's Office. Manuscript approved for publication 6 June 2005 by the Scientific Editor. Fish. Bull. 104:35-45 12006). Bottom trawl survey operating pro- cedures are standardized in order to reduce the variability of catch per unit of effort (CPUE) estimates. Many of the current standardization proce- dures address the efficiency of the trawl gear and the maintenance of constant catchability among samples and over time. Despite these efforts, variability in trawl catchability can be exacerbated by uncontrollable envi- ronmental conditions. Variables such as surface and bottom currents, sea state, wind direction, varying sub- strate types and inclinations, and depth of tow may all contribute to dif- ferences in gear efficiency by influenc- ing the area swept by the net (Rose and Nunnallee, 1998), the herding efficiency of the bridles (Somerton and Munro, 2001; Somerton, 2003), and escapement beneath the footrope (Weinberg et al., 2002). Many bottom trawl surveys con- ducted by the National Marine Fish- eries Service, such as the Alaska Fisheries Science Center's (AFSC) eastern Bering Sea (EBS) shelf sur- vey, operate with trawl winch brakes, set or locked, and tows are made with equal amounts of towing cable (warp) on both sides of the vessel. Other than by controlling towing speed and direction, these surveys are unable to compensate for changing environ- mental conditions. In contrast, auto- trawl systems are widely used by the commercial fleet and are purported to improve fishing performance by stabilizing trawl geometry over vary- ing environmental conditions, such as rough weather when vessel heave produces an upward lift on the trawl door resulting in loss of ground shear and wing spread, or over rough bot- tom when doors and nets have a greater probability of snagging. If autotrawl systems are able to reduce some of the variability in gear effi- ciency that is due to environmental variability, such as sea state and cur- rents, then including the use of auto- trawl systems as a standard survey bottom trawl procedure may improve the precision of survey results. In simple terms, autotrawls are dy- namic systems that operate on the principle of ensuring that the trawl is being towed in a direction perpen- dicular to the center of the footrope and headrope in order to optimize its performance. We are aware of two styles of autotrawl systems current- ly marketed. The first is a tension- controlled system that reacts to the difference in warp tension between winches by equalizing hydraulic pres- sure (equal tension). When the ten- sion on either side exceeds that of the other side (a user-defined threshold) due to factors such as increased drag, currents, sediments, or steep slopes, the system lengthens that warp to equalize the pressure between the two winches. Conversely, when the tension decreases on one warp, the system compensates by shortening that warp to equalize pressure be- tween the two winches. The second autotrawl style is a symmetry-con- trolled system that actively adjusts warp length in response to cross flow 36 Fishery Bulletin 104(1) signals from a sensor mounted on the trawl headrope. This system operates on the principle that net skewing can be caused by a crosscurrent. If the net is pulled square to the direction of flow then, its geometry will be symmetrical and trawl performance will be optimized. In the late summer of 2003, the AFSC conducted an experiment to examine the effect of these two types of autotrawl systems on the geometry of a survey trawl, comparing them to towing with equal amounts of warp on each side with the winches locked. The study consid- ers three aspects of trawl performance: 1) the factors of trawl geometry influencing the area and volume swept by the trawl (door spread, wing spread, and headrope height); 2) the bottom-tending performance of the foo- trope; and 3) the bottom-tending performance of the lower bridles. Materials and methods Operations and instrumentation The experiment was conducted during 19-25 September 2003 aboard the chartered 38-m-long commercial stern trawler FV Vesteraalen on smooth, relatively level bottom in 115 m of water at a site approximately 70 km north of Unimak Pass in the Bering Sea (55°10'N, 166°15'W). The Vesteraalen is powered by a single 1725-hp engine and is equipped with split Rapp Hydema (Rapp Hydema AS, Bod0, Norway) trawl winches carrying 2.5 cm (1") diameter, compacted, solid-core trawl warp. The winches are controlled by a Scantrol 2000 (Scantrol, Bergen, Norway) winch control system capable of quickly switch- ing to different towing modes as requested by the vessel operator. For this experiment towing was performed with the codend open and with three different winch control modes; locked winches with equal amounts of warp on the port and starboard side (locked); a tension- controlled autotrawl, which maintains equal tension on both warps by adjusting warp length based on the drag forces on each side (tension); and a symmetry-controlled autotrawl, which adjusts warp length according to side current forces in order to "optimize" water flow through the net (symmetry). The symmetry-controlled system requires a real-time speed sensor capable of detecting the direction of water flow across the headrope. We used an acoustically linked Scanmar (Scanmar, Asgardstrand, Norway) trawlspeed sensor that transmits flow data at 24-sec intervals both perpendicular and tangential to the headrope at its center. For this experiment and when in symmetry mode, the Scantrol system adjusted warp length at 30-sec intervals in response to changes in tangential velocity. The experiment was conducted with the AFSC stan- dardized trawl for the BBS shelf survey, the 83-112 Eastern bottom trawl. The 83-112 Eastern is a low- rise, 2-seam flatfish trawl designed to fish on smooth, soft bottom. The nylon net is constructed of 10.1-cm stretch mesh in the wing and body, 8.9-cm mesh in the intermediate, and double 8.9-cm mesh lined with Figure 1 Schematic diagram of the 83-112 Eastern bottom trawl and rigging shown from above. Mean door spread (68.0 m) and mean wing spread (17.8 ml were calculated from e.xperimental tows made at a depth of 115 m with the winches locked and 366 m of trawl warp out on each side. Bottom contact sensor units, shown as oversized triangles along the bridle and footrope, are labeled by position, as discussed in the text. 3.1-cm mesh in the codend. It is towed behind a pair of 1.8x2.7 m steel "V" doors, weighing approximately 816 kg apiece, which are attached to the net by two 3-m-long, 1.6-cm long-link chain door legs, a 12.2-m- long, 1.9-cm diameter stranded-wire door leg exten- sion, and a pair of 55-m-long, 1.6-cm diameter bare stranded-wire bridles on each side (Fig. 1). The 25.5- m-long (83 feet) headrope has forty-one evenly spaced, 20.3-cm diameter floats providing 116.4 kg of total lift. The 34.1-m-long (112 feet), 5.2-cm diameter footrope is constructed of 1.6-cm diameter stranded-wire rope that is protected with a single wrap of both 1.3-cm diameter polypropylene line and split rubber hose. The footrope is weighted with 51.8 m of chain (0.8-cm proof-coil) at- tached at every tenth link, forming 168 loops to which the netting is hung. An additional 0.6-m-long, 1.3-cm long-link chain extension connects each lower bridle to the trawl wing tips to help keep the footrope close to the bottom. Kotwicki et al.: Effect of autotrawl systems on tfie performance of a survey trawl 37 Prior to experimental towing, trawl warps were measured and marked at 366 m, the amount of warp used on the EBS survey when stations are fished at a depth of 115 m, the depth of our study site. Warps were measured and marked in ac- cordance with AFSC protocol (Stauffer, 2004) by using in-line wire counters (Olympic 750-N, Vashon, WA) while at the same time, calibration of the geometric winch counters associated with the autotrawl system was performed. A haul consisted of towing at a vessel speed of 3 knots while steering a steady course over ground. Tow direction was selected by attempt- ing to expose the trawl to the maximum amount of side current. Vessel speed and position were measured at 2-sec intervals with satellite navi- gation. Each haul consisted of two treatment sets in which three 15-min towing treatments (locked winches [locked], tension-controlled au- totrawl [tension], symmetry-controlled autotrawl [symmetry]) were conducted, allowing at least two minutes between treatments for the net to equilibrate. Randomizing the order of treatments within each treatment set reduced the influence of treatment order on trawl performance intro- duced by sea state, wind, and tidal currents. Likewise, towing at the same site for the dura- tion of the experiment eliminated bias that could be attributed to varying substrate. Wing spread, door spread, and headrope height were measured acoustically with Scanmar sensors at 4-sec intervals to the nearest 0.1 m. Footrope distance from the sea floor (cm) was measured at 0.5-sec intervals and averaged over 1.5-sec periods at five positions along the footrope si- multaneously by placing bottom contact sensors (BCS) at the center, at both trawl corners (located 3 m to either side of the center), and on each wing 1 m aft of the wing tips (Fig. 1). These sensors are self-contained units consisting of a tilt meter capable of measuring angle to the nearest 0.5° and a data logger housed in a watertight stainless steel container that fits inside a steel sled (Somerton and Weinberg, 2001). One side of the sled clips into a clamp that pivots freely on the trawl footrope and the other end drags along the bottom (Fig. 2). Changes in the distance of the footrope from the bottom produce changes in the recorded tilt angle. Conversion from tilt angle to distance off-bottom was accomplished with a calibration function determined for each BCS unit by fitting a quadratic function to data derived from a separate calibration experiment in which tilt angles were recorded when the footrope clamp was elevated set distances from a hard surface. The BCS unit extended out from the footrope 44 cm and its combined weight (consisting of BCS, sled, and footrope clamp) was 8.9 kg in seawater. The thickness of the clamp beneath the footrope was 2 cm. Because underwater video equipment was unavailable for this experiment, the extent to which this clamp penetrates into variable substrates was not estimated. Figure 2 Bottom contact sensors mounted to the footrope (top) and the bridle (bottom). Bridle distance from the bottom was measured at three positions simultaneously on both port and star- board sides by placing BCS units 25, 40, and 50 m forward of each wing tip. The BCS units and sled were mounted on triangular frames designed to hold them perpendicular to the bridle (Fig. 2, Somerton, 2003). The triangular frame measured 49 cm in its longest dimension. The combined weight of a BCS unit and frame was 8.7 kg in seawater. In addition to trawl mensuration, data were also collected on certain environmental variables during the different towing modes. Three variables were studied; 1) vessel heave measured at the trawl block; 2 ) the relative degree of offset of the warps from the heading of the ves- sel (crabbing); and 3) bottom current velocity both paral- lel and perpendicular to the direction of the vessel. The effect of sea state on the vertical displacement and attitude of the vessel is transmitted to the footrope from the trawl blocks through the trawl warps and 38 Fishery Bulletin 104(1) likely causes variable bridle and footrope contact with the bottom. Vessel heave at the starboard trawl block was used as a proxy for sea state. Heave, pitch, and roll data were collected at 1-sec intervals with a heave sensor (VT TSS, DMS-25, Watford, UK) mounted in the bridge along the centerline of the vessel. Heave data at the starboard block were then predicted, given the X, y, z coordinates of the block from the heave sensor, as distance (cm) from its equilibrium position. In the analyses, the standard deviation of the heave was used as the index of sea state. Net crabbing was subjectively assessed on a four-point scale by a single observer, then numerically coded as follows; 1) none — the net trailed straight behind the vessel; 2) slight — the warp could be seen entering the water between the side rail and the aft gantry; 3) mod- erate— the point of entry of the warp into the water was blocked from view by the aft gantry; and 4) severe — the warp was observed entering the water behind the stern ramp. Conditions usually remained constant and one observation was made per towing-mode treatment. How- ever, in some instances conditions changed rapidly and warranted more than one code. In such cases the aver- age of the observation codes was used in the analyses. Bottom current direction and velocity (cm/sec) were measured at 10-sec intervals with an oceanographic current meter (Nobska, MAVS-3, Woods Hole, MA) moored in the vicinity of our trawling activity three meters from the bottom. The current data were parti- tioned into two directional components, one parallel to the course of the vessel and the other perpendicular, and averaged for each sample treatment. Data analyses Comparison of the means and standard deviations among treatments Our null hypothesis, that the three treat- ments had the same effect, was tested with a Krus- kal-Wallis one-way ANOVA for all measures of trawl performance. We selected this test because of the skewed nature of the data. To describe the effect of the three treatments on trawl geometry features the mean and standard deviation (SD) were calculated for the wing spread, door spread, and headrope height off-bottom from each treatment in each haul. To describe the bottom-tending performance of the footrope we calculated the following statistics for each treatment: 1 mean footrope distance off-bottom — the sum of mean distances off-bottom along the footrope, from five footrope BCS units; 2 standard deviation of the footrope distance off- bottom — the sum of standard deviations along the footrope, from five footrope BCS units; and 3 symmetry of the footrope distance off-bottom — the sum of the absolute difference between the means of the two wing positions and the absolute differ- ence between the means of the two corner positions on the footrope. To describe the bottom-tending performance of the lower bridle, we considered only the BCS unit positioned 40 m from the wing tip. The variability in bottom-tend- ing performance of the bridles 40 m from the wing tip may have the highest impact on the catchability of the trawl because the BCS at the 25 m position was always on the bottom and the BCS at the 50 m position was always off the bottom. Performance of the bridles at the 40 m position was characterized by 1 mean bridle distance off-bottom — the sum of the mean distances off-bottom from the BCS units lo- cated 40 m forward of the wing tip on both lower bridles; 2 standard deviation of the bridle distance off-bot- tom— the sum of the standard deviations from the BCS units located 40 m forward of the wing tip on both lower bridles; and 3 symmetry of the bridle distance off-bottom — the absolute difference between means from the BCS units located 40 m forward of the wing tip on both lower bridles. Assessment of the effect of environmental factors If differences among towing mode treatments were found by the ANOVA (P<0.05), then the effects of heave, crabbing, and bottom current on gear performance within each treatment were explored. Multiple regres- sion analyses were performed for each of the treatment statistics with environmental variables as dependents. At the start of the analyses, the models included all four dependent variables (heave, crabbing, current par- allel, and current perpendicular to the tow direction). Variance inflation factors (VIFs) were calculated for all variables to test for multicollinearity (Neter et al., 1996). Dependent variables in all models had VIFs ranging from 1 to 1.4, indicating that no serious multi- collinearity existed among dependent variables. Models were simplified (backward deletion) until all P-values of the individual slopes were lower than 0.05. This procedure enabled us to establish which variables had a statistically significant impact on performance of the trawl within each treatment. Results A total of 21 successful hauls were performed during the experiment, yielding 42 possible treatment sets for analyses. The number of successful treatment sets included in the analyses varied for each of our perfor- mance statistics as a result of either malfunctions or poor survey trawl performance caused by conflicts with abandoned fishing gear. Comparison of the means and standard deviations among treatments Trawl geometry A total of 41 treatment sets were used to analyze the six trawl geometry statistics (means and Kotwicki et al.: Effect of autotrawl systems on the performance of a survey trawl 39 Table 1 Means of the trawl geometry statistics (ml and P-values of the Kruskal-Wallis treatments. Locked=lockeci winches; symmetry= symmetry-controlled autotrawl test for differences between three towing-mode and tension=tension-controlled autotrawl. Statistic Locked Symmetry Tension P-value Wing mean 17.76 17.77 17.85 0.6952 SD 0.69 0.70 0.73 0.9157 Door mean 68.03 67.67 67.71 0.8380 SD 1.87 1.73 2.14 0.0977 Height mean 1.82 1.81 1.81 0.7618 sn 0.21 0.20 0,2(1 0.8618 Table 2 Means of the footrope and bridle statistics (cm) and P-va ues of the Kruskal-Wallit test for differences between three tow ing-mode treatments. Locked=locked winches symmetry: =symmetry-controlled autotrawl; and tension=tension-controlled autotrawl. Statistic Locked Symmetry Tension P-value Footrope mean 13.98 13.09 12.08 0.0441 SD 6.39 5.67 5.24 0.0391 symmetry 1.95 1.44 1.45 0.0554 Bridle mean 7.50 6.69 6.28 0.0286 SD 4.91 4.03 3.37 0.0046 symmetry 1.59 0.93 0.88 0.0030 standard deviations of the wing spread, door spread, and net height). No significant differences were detected among the three towing modes (Table 1); consequently, no environmental variables were tested for their influ- ence on trawl geometry. Footrope distance off-bottom Analyses based on 33 suc- cessful treatment sets produced varying results among the three measures describing the bottom tending perfor- mance of the footrope. Mean footrope distance off-bottom differed significantly among the three towing modes (Table 2, Fig. 3). Footrope distance was lowest for the tension treatment followed by the symmetry treatment and locked winches. The greatest observed difference occurred between the locked (13.98 cm) and the tension treatment (12.08 cm). Standard deviation in footrope distance off-bottom also differed significantly among treatments. Differences were similar to those observed for the mean, in that the lowest SD was observed for the tension treatment (5.24 cm) and the highest SD was observed for the locked treatment (6.39 cm). Because the variance equals the square of the standard deviation, this seemingly small reduction in standard deviation corresponds to a fairly large reduction in the variance (-30%). Symmetry in the footrope off-bottom distance did not differ significantly (P=0.0554) among the three towing modes. Bridle distance off-bottom A total of 34 successful treatment sets were used to analyze the bridle distance off-bottom data. Mean bridle distance off-bottom differed significantly among treatments (Table 2, Fig. 4); it was lowest for the tension treatment (6.28 cm) and highest with the winches locked (7.50 cm). Standard deviation of bridle distance off-bottom also differed significantly among the three towing modes. SD values were signifi- cantly lower in the autotrawl towing modes, being lowest in the tension treatment (3.37 cm) and highest in the locked treatment (4.91 cm). Bridle distance off-bottom was most symmetrical in the symmetry and tension towing modes. Symmetry and tension means were simi- lar (0.93 cm and 0.88 cm, respectively) and significantly lower than that observed for the locked-winches treat- ment (1.59 cm). Assessment of the effect of environmental factors Footrope distance off-bottom The effect of environmen- tal factors on our three measures describing the bottom tending performance of the footrope produced varying results. Mean distance off-bottom was significantly affected by heave only, during all treatments (Table 3, Fig. 5). The standard deviation of the footrope distance off-bottom was also significantly affected by heave, crab- bing, and bottom current parallel to the direction of the 40 Fishery Bulletin 104(1) 15 h Mean 1.T ^ I -p 12 - 11 : 2,2 - Symmetry 2 - " 18 '. i- 1 6 - J T 1 4 - 1 ? _ -*- Figure 3 Means and standard errors of the footrope distance off-bottom, standard deviation of the footrope distance off-bottom, and footrope symmetry for the locked winches (L), symmetry iS). and tension iT) treatments. 8.7 r Mean 58 SD 8.3 - 5.4 r ~r 7.9 -p 5 >^ E 75 u 7.1 6.7 6.3 5 9 n I E 4.6 u 4.2 3.8 3.4 3 L i i L S T L S T 19 17 15 - Symmetry _ )( E o 1.3 1.1 0.9 n 7 - -r iS L S T Figure 4 Means and standard errors of the bridle distance off-bott 3m, standard deviation of the bridle distance off-bottom, and bridle symmetry for the locked winches (L), symmetry (S), | and tension (T) treatments tow during some treatments (Table 3, Fig. 6). Heave had a greater effect on footrope SD during the locked- winches treatment (slope=0. 06101 than during the sym- metry treatment (slope = 0.0381). Similar effects were also detected for crabbing (slopes = 1.2355 and 0.8493, respectively). Current speed in the direction of the tow affected footrope SD only during the symmetry mode. Heave, crabbing, and current velocity had no effect on Kotwicki et al : Effect of autotrawl systems on the performance of a survey trawl 41 Table 3 Multiple regression model slope coefficients with P-value <0.05. for locked winches, symmetry-controlled, and tension-controlled towing modes (y=bottom current parallel to the direction of the tow, AbsX=absolute value of the crosscurrent). Statistic Treatment Heave Crabbing y mean locked 0.0.572 — symmetry 0.0713 — — tension 0.0.544 — — SD locked 0.0610 1.2355 — symmetry 0.0381 0.8493 -0.0446 tension — — — symmetry locked — — 0.0255 symmetry — — — tension — — — mean locked 0.0680 — — symmetry — — — tension 0.0491 — — SD locked 0.0796 0.9536 — symmetry — — — tension 0.0411 0.3023 — symmetry locked — — — symmetry — — — tension — AbsX ff2 P-value Footrope Bndle — 9.04 0.0495 — 15.64 0.0227 — 16.14 0.0205 — 58.93 <0.0001 — 33.96 0.0066 — — >0.05 — 15.42 .0238 — — >0.05 — — >0.05 — 19.66 0.0086 — — >0.05 — 14.88 0.0242 — 60.30 <0.0001 — — >0.05 — 33.83 0.0017 0.0556 26.58 0.0018 — — >0.05 0.0361 13.25 0.0343 footrope SD during the tension mode. The sym- metry in footrope distance off-bottom was sig- nificantly affected by the current parallel to the direction of the tow during the locked-winches treatment only (Table 3, Fig. 7). Bridle distance off-bottom Mean bridle distance off-bottom was affected by heave only during some of the treatments (Table 3, Fig. 8). Heave had a greater impact on the mean bridle distance off-bottom during the locked-winches treatment (slope = 0.0680) than during the tension treat- ment (slope = 0.0491). The standard deviation of the bridle distance off-bottom was affected by heave and crabbing during two of the treatments (Table 3, Fig. 9). Heave had an effect on the SD of the bridle distance off-bottom during both locked (slope = 0.0796) and tension (slope = 0.0411) treat- ments. A significant effect due to crabbing was also detected during these same two treatments (slopes 0.9536 and 0.3023, respectively). During the symmetry mode no effect due to any of the envi- ronmental conditions was detected, even though SD was almost always higher than that observed in the tension mode, with the exception of extreme heave conditions. Symmetry in the bridle distance off-bottom was affected only by crosscurrent (Table 3, Fig. 10). Crosscurrent had an effect on the symmetry during the locked {slope = 0.0556) and tension (slope=0.0361) treatments. 00 - * AL □ S oT ^ Q •- 1H - 14 - ° /^ !3-^*^^^^r ....•■■•• ■" ,-.-iA- 10 r-."-"- "-'< 0 6 -, , . 0 20 40 60 80 100 Heave (cm) Figure 5 Comparison of regression lines illustrating the relationship between the mean footrope distance off-bottom and heave for locked winches (L), symmetry (S), and tension (T) treatments. Treatments in which heave had a statistically significant effect on the mean are identified with an asterisk (*). Discussion The objective of this study was to identify towing modes that could potentially reduce variance in the catchability of the 83-112 Eastern bottom trawl among hauls. Three separate aspects of trawl performance affecting catch- ability were investigated: trawl geometry, bottom tend- 42 Fishery Bulletin 104(1) 40 60 Heave (cm) 15 12 Q CO oa Aft) -8-9° 0° ifipt 0% I il Oo On o 0 ti_^ -30 -20 -10 0 10 20 30 Current parallel to tow direction (cm/s) Figure 6 Comparison of regression lines illustrating the relationship between the standard deviation (SD) of the footrope distance off-bottom and environmental variables for locked winches (L), symmetry (S), and tension (T) treatments. Treatments in which the environmental variable had a statistically significant effect on the SD are identified with an asterisk (*). 5 4 0 AL DS oT 1 a A f D *A A C) >. 3 - D o A a A (1) A D r+J . L (- __ — F ■^ - A 9 n O ■f^ _A_£__-- AO rn ; A Q- 3 ^ A ""^^"a^ -^i^,e. _^ A □ - T 1 - A 0 8 CO o A* D 0 ^AAQ o o in o -- 4- D 4j a 0 - , . , , , D -30 -20 -10 0 10 20 30 Current parallel to tow direction (cm/s) Figure 7 Comparison of regression lines illustrating the relationship between the symmetry of the footrope distance off-bottom and current parallel to the direction of the tow for locked winches (L), symmetry (S), and tension IT) treatments. Treatments in which the current had a statistically significant effect on the symmetry are identified with an asterisk (*). ing performance of the footrope, and bottom-tending performance of the lower bridles. Trawl geometry Because wing and door spread and net height vari- ability all influence catchability of a trawl (Rose and Nunnallee, 1998), surveys would stand to benefit from autotrawl systems if variances in the trawl geometry features were reduced. Our study showed that neither autotrawl system improved trawl geom- etry over the conventional locked-winches towing mode. Footrope Previous studies have demonstrated that escape- ment under the footrope is an important factor determining the catchability of the 83-112 Eastern survey trawl (Somerton and Otto 1999; Munro and Somerton, 2002; Weinberg et al., 2004). Periodic separation between the footrope and the bottom contribute to variability in catchability and have Kotwicki et al : Effect of autotrawl systems on ttie performance of a survey trawl 43 been shown to be influenced by a variety of factors, such as net speed through the water (Somerton and Weinberg, 2001) that can increase footrope distances off-bottom. Weinberg et al. (2002), using a different survey bottom trawl, found capture probability of some benthic species to decrease as a result of footrope separations with the bottom. These findings likely apply to the 83-112 Eastern trawl as well. Because capture probability is gener- ally size dependent (Munro and Somerton, 2002), an unstable footrope may not only increase the variance of overall biomass estimates, but may bias the size distribution data generated by a survey as well. If autotrawl systems can help maintain footrope contact with the bottom, then the use of an autotrawl during surveys may be warranted. In our experiment, the tension-controlled auto- trawl system provided the best footrope contact overall and the lowest standard deviation (Fig 3). High heave and moderate to strong crabbing were largely responsible for the differences observed in the bottom-tending performance of the footrope among treatments (Figs. 5-7). The tension treat- ment was most effective in counteracting the ef- fects of environmental conditions, whereas the locked-winches treatment was the least stable of the three treatments given the changing environ- mental conditions. We found both autotrawl systems have a po- tential to improve bottom trawl survey biomass estimates by increasing the dynamic stability of the trawl, thereby reducing the variance in its catchability. The equal-tension system proved bet- ter than the symmetry system in improving the overall stability of the footrope bottom-tending performance given the low bottom current ve- locities observed. Symmetry autotrawl systems, designed to react to crosscurrent conditions at towing depth, could prove to be more effective under stronger current conditions, but during our experiment, current velocities may have been too low (<35 cm/s) for us to be able to detect a differ- ence between the symmetry and tension towing modes. Bridles Flatfish are stimulated to herd by close proximity to or actual contact with the lower bridle (Main and Sangster, 1981a), whereas semipelagic species such as Atlantic cod (Gadus morhua) are probably stimulated to herd by the sight of the otter doors and mud clouds created by the doors and lower bri- dles (Main and Sangster, 1981b). For this reason, the length of the lower bridle in contact with the bottom and the frequency of this contact affect the herding efficiency of the trawl (Somerton, 2003). If the among- and within-tow variability of the bridle bottom-tending distance could be reduced by using an autotrawl system, then standardizing survey 16 12 AL dS ot A. S 0 10 40 60 80 100 Heave (cm/s) Figure 8 Comparison of regression lines illustrating the relationship between mean bridle distance off-bottom and heave for locked winches (L), symmetry (Si, and tension (T) treatments. Treat- ments in which the heave had a statistically significant effect on the mean are identified with an asterisk (*). 2 b. 1 '^'- DS OT 0 A A .^ 8 6 D a A '^ D o o s- - □ a □ T 4 A O O 2 - 0 E u Q 20 40 60 Heave (cm) 80 100 1,5 2 Crabbing (cm) Figure 9 Comparison of regression lines illustrating the relationship between standard deviation (SD) in bridle distance off-bottom and environmental variables for locked winches (L), symme- try (S), and tension (T) treatments. Treatments in which the environmental variable had a statistically significant effect on the SD are identified with an asterisk (*). 44 Fishery Bulletin 104(1) 5 o a A AL as OT 1 4 a 3 - 6 A O A A D i ^^— L' 2 O D □ 0 o O D D - Q,,----^ 0 n M o O a T S 1 0 o ° ° D 0 10 20 30 Crosscurrent (cm/s) 40 Figure 10 Comparison of regression lines illustrating the relation- ship between bridle symmetry and crosscurrent for locked winches (L). symmetry (S), and tension (T) treatments. Treatments in which crosscurrent had a statistically sig- nificant effect on bridle symmetry are identified with an asterisk (*). procedures to include towing with an autotrawl system would likely be beneficial. Our results showed that both autotrawl systems re- duced the mean distance of the bridles off-bottom and the standard deviation of this distance and increased the symmetry between bridles (Fig. 4) over the locked- winches treatment. Our experiment also demonstrated that both autotrawl systems increased the stability of lower bridle bottom contact in changing environmental conditions, but we were unable to discern which of the systems was better. The data indicate that bridle bot- tom-tending performance was the least affected by en- vironmental conditions during the symmetry treatment; however the standard deviation in the bridle distance off-bottom was almost always lower during the tension treatment, the exception being under extreme heave conditions (Fig. 9). The locked-winches treatment had the highest standard deviation and was affected the most by heave, crabbing, and crosscurrent velocity. Autotrawl systems counteract the effects of warp length differential created by trawl crabbing. During our locked-winches treatment crabbing likely caused unequal tension in the warps similar to that seen while towing straight behind the boat with unequal warp lengths. Weinberg and Somerton (2006) reported that warp tension changes significantly with offset in warp lengths. To compensate for the crabbing angle and to assure that the trawl is pulled square to the direction of the tow during crabbing, one warp should be shorter than the other. Tension-controlled autotrawl systems ad- just the length of the warps to equalize the tension on them. Symmetry-controlled autotrawl systems change the length of the warps to minimize crosscurrent and also increase the symmetry of the trawl in relation to the tow direction. In summary, autotrawl systems proved to be effective in decreasing some of the adverse effects of environ- mental factors on some aspects of the 83-112 bottom trawl performance and, as a result, have the potential to reduce variance in among-haul catchability of the survey trawl. For this trawl, footrope and bridle dis- tances off-bottom were significantly different among the three towing modes, albeit the differences in actual dis- tances off-bottom were small. Trawls deploying heavier groundgear, thus enabling more constant contact with the bottom, may not be as affected. Further investiga- tions are needed to assess the effects of the two types of autotrawl systems on other types of survey trawl gear, such as high-opening bottom trawls, trawls using differ- ent footropes, shrimp trawls, and midwater trawls. The effect of using autotrawls in areas with stronger current and different depths also needs to be investigated for the different types of trawl gear. For bottom trawl surveys, we are concerned with two potential shortcomings in the symmetry-controlled autotrawl system tested. First, videos of trawling in rough terrain have revealed footrope distortion occur- ring when the footrope or doors snag on the bottom (Weinberg, unpublished data). Typically, the side op- posite the snag is pulled forward while the snagged side remains stationary or is pulled ahead at a slower pace. This uneven pull on the net causes asymmetry in the headrope shape, by skewing it from the general tow direction (Weinberg and Somerton, 2006). Distortion of the headrope introduces error in the current direc- tion and velocity values obtained by the current sensor mounted on the headrope from which warp length is determined, and thus impacts trawl performance. Our second concern involves the overall warp adjustment period required for the trawl to equilibrate to current Kotwicki et al,: Effect of autotrawl systems on the performance of a survey trawl 45 direction and velocity based on the current sensor in- put. Because many surveys standardize to a 15-min tow duration, a proportionally large percentage of tow time may be involved with adjusting warp lengths in pursuit of optimal symmetry and therefore vary trawl catchability during the tow. Under commercial trawling conditions, the manufacturer of the symmetry winch control system recommends using a 2-min minimum signal collection and analysis period between warp ad- justments. After each adjustment, new sensor signals would be received and evaluated, followed by another warp adjustment, if necessary. We are uncertain as to how many warp adjustments may be necessary to orient the trawl in relation to crosscurrent flow, but based on our experiences and the frequency with which warp lengths changed during our experiment, it is con- ceivable that several adjustment periods spanning the majority of the survey tow may be necessary, leaving minimal time for the net to actually fish symmetrically. Furthermore, each 2-min warp adjustment would be based on a maximum of only five current flow readings because the current sensor refreshes data at 24-s inter- vals. Should signal loss occur, then fewer data points would be available. We recommend taking a cautious approach before switching a trawl survey from locked winches to auto- trawl. A change in survey method may require an ex- tensive calibration experiment between the two trawl- ing methods in order to maintain the continuity of a survey time series. Furthermore, autotrawl systems, like other mechanical devices, require service, appro- priate inspections and periodic testing to ensure that they are functioning correctly within manufacturer specifications. Autotrawl calibration parameters are dependent upon accurate measurements of the diameter and length of each winch drum, warp diameter, and construction (e.g., compacted vs. traditional or wire core vs. fiber core), layers of warp, and the number of windings per layer on each drum. Hydraulic pumps and lines, electric motors, valves, solenoids, computer- ized control panels, and geometric counters must all be inspected to assure proper operation. Of course, should surveys operate with a symmetry-style autotrawl sys- tem, then all of the above procedures would hold true in addition to the need for accurate calibration of the current sensor and the proper mounting of the sensor to the headrope. Acknowledgments The success of this project is due to the efforts of many. We extend our gratitude to Tim Cosgrove, Brad Lougheed, and the crew of the FV Vesteraalen. Our knowledge of winch control systems was greatly enhanced by the guid- ance of Doug Dixon and Ed Ramberg, as was our ability to communicate at-sea with a wide range of instru- ments thanks to Scott Furnish, David Roetcisoender, Jon French, and Dennis Benjamin. We are also grateful to our reviewers including Captain John Gruver, Mark Wilkins, Gary Walters, and Peter Munro. Literature cited Main, J., and G. I. Sangster. 1981a. A study of the sand clouds produced by trawl boards and their possible effect on fish capture. Scott. Fish. Res. Rep. 20:1-20. 1981b. A study of fish capture process in a bottom trawl by the direct observations from a towed underwater vehicle. Scott. Fish. Res. Rep. 23:1-23. Munro, P. T., and D. A. Somerton. 2002. Estimating net efficiency of a survey trawl for flatfishes. Fish. Res. 55:267-279. Neter, J., M. H. Kutner, C. J. Nachtsheim, and W. Wasserman. 1996. Applied linear regression models, third ed., 720 p. Irwin, Chicago, IL. Rose, C. S., and E. P. Nunnallee. 1998. A study of changes in groundfish trawl catch- ing efficiency due to differences in operating width, and measures to reduce width variation. Fish. Res. 36:139-147. Somerton. D. A. 2003. Bridle efficiency of a survey trawl for flatfish: measuring the length of the bridles in contact with the bottom. Fish. Res. 60:273-279. Somerton D. A., and P. T. Munro. 2001. Bridle efficiency of a survey trawl for flatfish. Fish. Bull. 99:641-652. Somerton, D.A.. and R. S. Otto. 1999. Net efficiency of a survey trawl for snow crab, Chionoecetes opilio, and Tanner crab, C. bairdi. Fish. Bull. 97:617-625. Somerton, D.A., and K. L. Weinberg. 2001. The affect of speed through the water on footrope contact of a survey trawl. Fish. Res. 53:1724. Stauffer, G. 2004. NOAA protocols for groundfish bottom trawl sur- veys of the nation's fishery resources. NOAA Tech. Memo. NMFS-F/SPO-65, 205 p. Alaska Fish. Sci. Center, 7600 Sand Point Way NE, Seattle, WA 98115. Weinberg, K. L., R. S. Otto, and D. A. Somerton. 2004. Capture probability of a survey trawl for red king crab iParalithodes camtschaticus). Fish. Bull. 102:740-749. Weinberg, K. L., and D. A. Somerton. 2006. Variation in trawl geometry due to unequal warp length. Fish. Bull. 104:21-34. Weinberg, K. L., D. A. Somerton. and P. T Munro. 2002. The effect of trawl speed on the footrope capture efficiency of a survey trawl. Fish. Res. 58:303-313. 46 Abstract — The California market squid iLoligo opalescens) has been harvested since the 1860s and it has become the largest fishery in California in terms of tonnage and dollars since 1993. The fishery began in Monterey Bay and then shifted to southern California, where effort has increased steadily since 1983. The California Department of Fish and Game (CDFGl collects information on landings of squid, including ton- nage, location, and date of capture. We compared landings data gathered by CDFG with sea surface tempera- ture (SST), upwelling index (UI). the southern oscillation index (SOI), and their respective anomalies. We found that the squid fishery in Monterey Bay expends twice the effort of that in southern California. Squid land- ings decreased substantially follow- ing large El Nino events in 1982-83 and 1997-98, but not following the smaller El Niiio events of 1987 and 1992. Spectral analysis revealed autocorrelation at annual and 4.5- year intervals (similar to the time period between El Nino cycles). But this analysis did not reveal any fortnightly or monthly spawning peaks, thus squid spawning did not correlate with tides. A paralarvae density index (PDI) for February correlated well with catch per unit of effort (CPUE) for the following November recruitment of adults to the spawning grounds. This stock- recruitment analysis was significant for 2000-03 iCPUE = 8.42 + 0.4lPDI. adjusted coefficient of determina- tion, r2 = 0.978, P=0.0074). Surveys of squid paralarvae explained 97. 89^ of the variance for catches of adult squid nine months later. The regres- sion of CPUE on PDI could be used to manage the fishery. Catch limits for the fishery could be set on the basis of paralarvae abundance surveyed nine months earlier. The fishery for California market squid {Loiigo opalescens) (Cephalopoda: Myopsida), from 1981 through 2003 Louis D. Zeidberg Monterey Bay Aquarium Research Institute 7700 Sandholdt Rd. Moss Landing, California 95039-9644 E-mail address: zelo g mbari org William M. Hamner Dept. ol Ecology and Evolutionary Biology Univ, California, Los Angeles 621 Charles E. Young Drive South Box 951606 Los Angeles, California 90095-1606 Nikolay P. Nezlin Southern California Coastal Water Research Proiect 7171 Fenwick Lane Westminster, California 92683-5218 Annette Henry California Department of Fish and Game Southwest Fisheries Science Center Marine Region, La Jolla Field Office 8604 La Jolla Shores Drive La Jolla, California 92037 Manuscript submitted 7 May 2004 to the Scientific Editor's Office. Manuscript approved for publication 20 June 2005 by the Scientific Editor Fish. Bull. 104:46-.59 12006). The recent discovery of falsification in Chinese fisheries reporting has led to the realization that the majority of the world's fisheries surpassed sustain- ability in 1988 (Watson and Pauly, 2001). The food chain has been fished down by removal of apex predators like swordfish and snapper beyond sustainability, and fisheries have subsequently shifted to prey items like sardine and mackerel (Pauly et al., 1998). We have reached the point where cephalopods are regularly the largest biomass of all commercial spe- cies harvested. Since 1970, groundfish landings of flounders, cods, and had- docks have either decreased or main- tained their levels while landings in cephalopod fisheries have increased (Caddy and Rodhouse, 1998). Some of the larger cephalopod landings may be due to increased demand, but lower levels of predation and competition from finfish, and the shorter lifespan of squid may have allowed cephalopods to increase in abundance worldwide. Loiigo opalescens is a small squid (130 mm mantle length) that occupies the middle trophic level in Califor- nia waters, and it may be the state's most important forage species. Mar- ket squid are principal forage items for a minimum of 19 species of fishes, 13 species of birds, and six species of mammals (Morejohn et al., 1978). The effective management of this fishery is of paramount importance not only to the fishermen involved but also to the millions of fishes, birds, and mam- mals that compete for this resource. Because cephalopods eat mostly zoo- plankton (Loukashkin, 1976), if we also deplete the squid population, it is not clear how oceanic food chains will respond. If the subannual popu- lation of L. opalescens fails to recruit a large biomass in a given year, the long-lived predators of this species in Zeidberg et al , The fishery for Loligo opalescens from 1981 through 2003 47 the California Current may encounter severe metabolic stress. Since the decline of the anchovy fishery, market squid probably constitutes the largest biomass of any single marketable species in the coastal environment of Cali- fornia (Rogers-Bennett. 2000). In the 1999-2000 season, fishermen landed 105,005 metric tons of California mar- ket squid [Loligo opalescens) with an exvessel (whole- sale) revenue of $.36 million (California Department of Fish and Game [CDFGJ landing receipts). These squid deposit egg capsules on sandy substrates at depths of 15-50 m in Monterey Bay (Zeidberg et al., 2004) and 20-90 m in the Southern California Bight. The majority of squid landings occur around the California Channel Islands, from Pt. Dume to Santa Monica Bay, and in southern Monterey Bay. The fishery comprises chiefly light-boats with high wattage illumination to attract and aggregate spawning squid to the surface, and seine vessels that net the squid (Vojkovich, 1998). Management to date has followed methods that are not dependent upon an estimate of population abun- dance because no estimate of squid biomass exists. In addition to limiting the catch and the number of ves- sels, management of the fishery has included weekend closures north of Point Conception since 1983, and these closures have recently extended to all of California coastal waters. This regulation is designed to allow a 48-hour period each week for undisturbed spawn- ing. For Monterey Bay, the weekend closure resulted in highest landings on Mondays and decreasing daily landings through Friday (Leos, 1998). Since 2000, light boat and seine vessel operators have been required to complete logbooks for CDFG, such that CPUE can be estimated from data on the cumulative effort required to land squid. Because of their short lifespan, many squid popula- tions have been more effectively correlated with local oceanographic conditions than have pelagic fish spe- cies with life spans of 4-8 years. Squid landings from all regions of the world fluctuate in conjunction with the temperatures of the previous season. Mclnnis and Broenkow (1978) found positive temperature anomalies preceded good Loligo opalescens landings by 18 months, and poor squid catches followed periods of anomalous low temperatures in Monterey Bay. Robin and Denis (1999) found similar results. Warmer waters (mild win- ters) were followed by increased cohort success for Lo- ligo forbesi in the English Channel, but this effect was not constant throughout the year. Conversely, Roberts and Sauer (1994) found Loligo vulgaris reynaudii land- ings in South Africa to increase with upwelling that coincided with La Nina (cold water) conditions in the equatorial Pacific. Rocha et al. (1999) also found an increase in squid paralarvae of many species during upwelling conditions on the Galacian-coast. Modern instruments for monitoring coastal ocean con- ditions, including weather buoys and satellites, provide a vast amount of information on the physical environ- ment of fish and squid populations. The correlation between cold, upwelled nutrient-rich water at the sea surface resulting from Eckman transport and phyto- plankton blooms a few days later is well established (Nezlin and Li, 2003). Mesoscale eddies generated by coastal processes and islands also serve to concentrate phytoplankton (Falkowski et al., 1991; Aristegui et al., 1997; DiGiacomo and Holt, 2001). The subsequent effect upon zooplankton grazers rapidly follows the cycles of upwelling and relaxation (Wing et al., 1995; Graham and Largier, 1997; Hernandez-Trujillo, 1999). Waluda et al. (1999) found that the CPUE for the II- lex argentinus fishery was not related to monthly local sea surface temperature (SST), but CPUE was inversely related to SST on the hatching grounds for the previous July, when hatchlings were in their exponential growth phase (Yang et al., 1986; Grist and des Clers, 1998). The largest catches followed cold water. Waluda et al. (2001) found a large CPUE when the Brazilian Current dominated and frontal waters diminished in the location where squid hatching occurs. Agnew et al. (2000, 2002) found that CPUE for Loligo gahi was inversely corre- lated with SST for hatching areas six months earlier. Sakurai et al. (2000) found that Todarodes pacificiis CPUE was highest following periods when there were large regions of hatchling-favorable habitat (17-23°C waters). They found a positive correlation between the density of paralarvae and the catch per unit of effort (CPUE) of adults in the same year (r~ = 0.91) and also in the CPUE of the following year (r2 = 0.77). The CDFG has an extensive database of landings data from 1981 to the present for market squid. Be- cause there is no record of effort prior to 2000 and be- cause the market is driven by demand, it is difficult to use landings and vessel-day data to calculate a CPUE and therefore estimate biomass. Fishermen report that even if squid are available, they may not be harvested when processors are not accepting squid (Brockman^). However, there is no other database as large and wide- spread temporally and spatially as fishery data. Even though there are no data recorded when boats attempt to catch squid and fail, we can still use landings and vessel-days to create a CPUE. This CPUE therefore is not a methodically collected estimate of biomass, but is still a robust enough estimate of abundance to draw preliminary conclusions as we wait for logbook data to accumulate. It is important to determine the effects of the envi- ronment on the California market squid fishery so that we can predict future landings from present conditions. Our investigation uses California market squid landings for 1981-2003 to examine correlations of landings and CPUE with physical oceanography. We compare land- ings data (time, location, vessel-days, and landings [in pounds]) to sea surface temperature (SST), upwelling index (UI), the Southern Oscillation index (SOI), the in- dex of sea surface temperature in the eastern equatorial tropcial Pacific NIN03, and their respective anomalies. We also compare CPUE to a paralarvae density index Brockman, D. 2002. Personal commun. Davie.s Locker Sportfishing, 400 Main St. Newport Beach, CA 92611. 48 Fishery Bulletin 104(1) (PDI) based upon distributions determined in the Southern California Bight (Zeidberg and Hamner, 2002). Materials and methods 35° - 0 -125° The CDFG database for commercial Cali- fornia market squid landings from 1981 to present includes weight, date, location (based on CDFG 10x10 nm blocks), and gear type. Accounting for general physical oceanographic properties (Harms and Winant, 1998; Bray et al., 1999: Brink et al., 2000; Hickey et al., 2003) and following our previous stud- ies (Nezlin et al., 2005), we organized the landings data into six areas to look at subtle differences between them: MB = northern coastal (because the majority of the land- ings in this area occur in southern Monterey Bay), CC=central coastal, SB = Santa Barbara Channel, SCB = Southern California Bight, SM= Santa Monica, and SD = San Diego. Also we grouped the fishery into two larger regions April (APR, equal to MB above) and October (OCT, a combination of the other five areas) based upon the month of greatest recruit- ment (Fig. 1). For the purpose of our study, recruitment is the aggregation of reproductive adults on the spawning grounds. When CDFG reports squid data, they make a distinction at Point Conception, thus our MB and CC areas are grouped as the "north" and our SB, SCB, SM, and SD are named "south." For this fishery we defined CPUE as the recorded tons landed in a day, divided by the number of seine vessels that landed these squid. Those days in which there were no landings were assigned a value of zero. This CPUE is impor- tant because, although not truly a quantifying effort, it does provide a means for estimating the abundance of squid by providing some basis for the amount of time taken to make a landing. Lampara, brail, and light boat data were not included because of increased variability in landings and effort and the fact that these vessels have dwindled from ten to zero percent since 1981. The landings and boat data for each area were summed for each block by day. For example, assume that on a particular day fishermen caught 10,000 metric tons with four boats in the area of SM, 18,000 tons from three boats in SCB, and 12,000 tons from three boats in SB. We would calculate a CPUE of 2500 tons/vessel-day in SM, 6000 tons/VD in SCB, and 4000 tons/VD in SB, respectively. Thus for every date for which there was a landing we were able to calculate CPUE value for each area. Until 2002, there had never been a landing in Mon- terey in January, when one vessel captured 75 tons. Data such as this produce misleadingly high CPUEs; therefore all months with less than seven vessel-days for the en- tire 22-year period were removed from the analysis. April recruitment (APR) October recruitment (OCT) km --F= -120 = Figure 1 The California coast with the fishery areas for the California market squid (Loligo opalescens) identified. Areas were classified according to physical oceanographic features: Northern Coast (MB), Central Coast (CC), Santa Barbara Channel (SB), Southern California Bight (SCB), Santa Monica Bay (SM), and San Diego (SD). Regions were also classified by fishery recruitment month: April recruiting (APR: same as MB area) and October recruiting (OCT: CC, SB, SCB, SM, and SD areas combined). Block 526 (indicated with a slender arrow) is where the majority of the MB-APR landings occur. Shaded area indicates the location of the paired-net surveys used to generate the paralarvae density index (PDI). Physical oceanography data were gathered from the In- ternet for sea surface temperature (SST),'- upwelling in- dex (UI),'^ southern oscillation index (SOI),^ and NIN03.5 Upwelling index (UI) is an Ekman offshore water trans- port (mVs per 100 m of the coastline) estimated from fields of atmospheric pressure (Bakun, 1973). Southern Oscillation index (SOI) is the difference between the standardized measurements of the sea level atmospheric - NOAA (National Oceanic and Atmospheric Administration). National Data Buoy Center. 2000. Website: http://facs. scripps.edu/surf/buoys.html [Accessed on 24 April 2003.1 ^ NOAA Pacific Fisheries Environmental Laboratory. 2003. Website: http://www.pfeg.noaa.gov/products/las.html [Accessed on 20 March 2003.1 * Australian Government Bureau of Meteorology. 2005. Website: http://www.bom.gov.au/climate/current/soihtml. shtml [Accessed on 29 March 2003.1 ^ IRI (International Research Institute for Climate Pre- diction). 2005. Website: http://ingrid.ldgo.columbia.edu/ SOURCES/.Indices/.nino/.EXTENDED/.NIN03/ [Accessed on 20 April 2004.1 Zeidberg et a\ The fishery for Loligo opalescens from 1981 through 2003 49 Table 1 Totals of vessel-days, landi ngs (metric tons), and CPUE (tons/vessel-day) for two ti me per iods of the Califor nia market squid {Loligo opalescens) fishery (1981- 2003). The majority of the fishery occurred in Monterey Bay, 1981-89, and in southern Cali- fornia. 1990 -2003. The six smal areas represent physical oceanographic features and the larger regions (APR and OCT) are | grouped by the month of greatest recruitment of t pawning adults to the fishery (see bolder border in Fig. 1). MB and APR are synonymous terms. MB = northern coastal area, pi edominantly southern Monterey B ay; CC =central coast SB = Santa Barbara; SCB = Southern California Bight; SM = Santa Monica; and SD = San Diego VD=ve.ssel days. Region Vessel-days Landing Percent landings CPUE Years and area (VD) Percent VD (tons) (tons) (tons/VD) 1981-89 APR MB 4918 87.4 27242 66.5 5.5 OCT CC 38 0.7 1040 2.5 27.4 SB 186 3.3 3811 9.3 20.5 SCB 169 3.0 3475 8.5 20.6 SM 122 2.2 1811 4.4 14.8 SD 194 3.4 3597 8.8 18.5 Subtotal 709 12.6 13735 33.5 19.4 Total 5627 40977 7.3 1990-2003 APR MB 6508 22.5 92323 13.6 14.2 OCT CC 1283 4.4 33964 5.0 26.5 SB 3822 13.2 104172 15.3 27.3 SCB 8037 27.7 212986 31.3 26.5 SM 4408 15.2 113569 16.7 25.8 SD 4918 17.0 124147 18.2 25.2 Subtotal 22468 77.5 588838 86.4 26.2 Total 28976 681161 23.5 pressure in Tahiti and Darwin. NIN03 is determined by averaging the SST anomalies over the eastern tropical Pacific (5°S-5°N; 150°W-90°W). The buoys used were the Monterey buoy (46042, 36°N 122°W) for the MB region, the east Santa Barbara buoy (46053, 34.24°N 119.85°W) for the SB region, and the Santa Monica buoy (46025, 33°N 119°W) for the remaining regions. We performed a spectral analysis of the entire time series to look for significant periodicities in the daily data for the entire 22-year data set. CPUE values were natural log-transformed and smoothed with a Parzen window (Ravier and Fromentin, 2001). We used a time series analysis method of cross correlation to deter- mine the lag period in months between CPUE and the physical features of SST, SOI, NIN03, and UI and their anomalies from averaged seasonal cycles. Using this lag period we calculated linear regression of the CPUE from SST. Sea surface temperature (SST) time series was ob- tained from infrared satellite measurements with ad- vanced very high resolution radiometers (AVHRRs) on National Oceanic and Atmospheric Administration (NOAA) meteorological satellites. The data were pro- cessed at the University of Miami's Rosenstiel School of Marine and Atmospheric Science (RSMAS) and the NOAA National Oceanographic Data Center (NODC) within the scope of Pathfinder Project (version 4.1, available from the Jet Propulsion Laboratory Physical Oceanography Distributed Active Archive Center [JPL PO DAAC]).!' We performed a stock-recruitment analysis from a paralarvae density index (PDI). Paralarvae were col- lected with paired nets (505-f 0 .0 J:Qii 0- i-.l..,-:^^ 0~-^^ ! •— --*- 1 B --f— 2 O 3 4 5 O ■ 1 6 O 1 7 O 8 9 10 o r 11 12 -•-APR o OCT ■ O -9- O 10 iC - a o.. o- ° o O • •o •» o ° — *l- — I 1 1 1 1 1 1 1 1 , ^ 1 2 3 4 5 6 7 8 9 10 11 12 Month Figure 3 Fishery data for Loligo opalescent:, summed by month for 1981-2003. (A) Landings (metric tons). (B) Effort in vessel-days (VD). (C) Catch per unit of effort (tons/VD). Monterey Bay (April [APR] — black circles) and southern California (October [OCT] — unfilled circles). Scale of y-axis changes between A, B, and C. Largest landings occurred one month later in May in the APR region and in November in the OCT region when SST was 11.7"C and 16.1°C, respectively. The most significant cross correlations of time lag analysis for CPUE to SST are listed in Table 3. In all cases of biological significance, CPUE lagged SST by 4, 5, or 10 months. MB CPUEs were highest (in May) when SST was low four months earlier (Jan), and hence gave a negative correlation. In all other regions, the four or five month correlation was positive, with CPUE high (Nov) when SSTs were high four months earlier (Jul). For the southern California areas there was a negative correlation with SSTs 10 months earlier (Janl. Therefore, cold winters and warm summers correlate with larger landings. Recruitment of spawning adults to the fishery occurs during the productive seasons in both APR, MB, and OCT. Productivity in APR and MB co-occurs with the spring-summer upwelling season, and in OCT productivity correlates with winter storms that lead to a deeper mixed layer. There were signifi- cant cross correlations with SOI, NIN03, and UI, but not with the anomalies of SOI and UI. Interestingly correlations for NIN03 were greater than those for SOI (Fig. 4), indicating that the CPUE of Loligo opalescens is more closely related to the "oceanic teleconnection" than to "atmospheric teleconnection," 52 Fishery Bulletin 104(1) Table 2 Periods of greatest spectral variance in the daily CPUE data of the market squid tLoligo opnlescens) fishery for 1981-2003. Significance: P<0.01. Numbers in bold are repeated in more than one area. Harmonics of factors of 2: 2, 4, ... 4096 (blank spaces I are omitted because they are inherent in spectral analysis and not relevant to this spe- cies. MB=northern coastal area, predominantly southern Monterey Bay; CC = central coast; SB = Santa Barbara; SCB = Southern California Bight; SM = Santa Monica; and SD = San Diego. Top ten periods of spectral variance MB CC Rank Days Years Month Days Years Month 1 372.4 1 12.2 2 356.2 1 11.7 372.4 1 12.2 3 1638.4 4.5 53.7 4 7 0 0.2 2730.7 7.5 89.5 5 315.1 0.9 10.3 455.1 1.2 14.9 6 2730.7 7.5 89.5 481.9 1.3 15.8 7 341.3 0.9 11.2 356.2 1 11.7 8 455.1 1.2 14.9 390.1 1.1 12.8 9 3.5 0 0.1 10 1365.3 3.7 44.8 SB SCB Rank Days Years Month Days Years Month 1 372.4 1 12.2 372.4 1 12.2 2 356.2 1 11.7 1638.4 4.5 53.7 3 390.1 1.1 12.8 356.2 1 11.7 4 1365.3 3.7 44.8 1365.3 3.7 44.8 5 1638.4 4.5 53.7 390.1 1.1 12.8 6 7 8 910.2 2.5 29.8 264.3 0.7 8.7 9 819.2 2.2 26.9 10 182 0.5 6 682.7 1.9 22.4 Rank SM SD Days Years Month Days Years Month 1 372.4 1 12.2 356.2 1 U.7 2 1638.4 4.5 53.7 3 182 0.5 6 372.4 1 12.2 4 182 0.5 6 5 1638.4 4.5 53.7 6 356.2 1 11.7 682.7 1.9 22.4 7 390.1 1.1 12.8 341.3 0.9 11.2 8 315.1 0.9 10.3 1365.3 3.7 44.8 9 204.8 0.6 6.7 390.1 1.1 12.8 10 7 0 0.2 178.1 0.5 5.8 Table 3 Results of the time series analysis. Significant correla- tion coefficients occurred when CPUE lagged behind sea | surface temperature (SST) from buoys and advanced very high resolution radiometers (AVHRRs) by 4 -10 months for all areas, exc ept CC. Negative correlation coefficients deni- onstrate that high CPUE corresponds with low water tem- | peratures in the lags zed month from column two; positive | values may indicate a direct relationship. MB=northern coastal area predominantly southern Monterey Bay; CC = central coast ; SB = Santa Barbara; SCB = Southern Califor- nia Bight; SM=Santa Monica; and SD = San Diego. CPUE Lagg( ?d SST Correlation region (months) source coefficient MB 4 MB buoy -0.481 5 AVHRR -0.358 SCB 4 SM buoy 0.206 10 SM buoy -0.344 9 AVHRR -0.340 SD 4 SM buoy 0.220 10 SM buoy -0.398 10 AVHRR -0.387 SM 5 SM buoy 0.176 10 SM buoy -0.340 10 AVHRR -0.334 SB 4 ESB buoy 0.387 4 SM buoy 0.415 9 AVHRR -0.372 Assuming a 6-9 month lifespan for L. opalescens, we used linear regression to compare SST from buoy data for the previous 6-10 months. We performed com- parisons up to 10 months because squid eggs take 30 days to hatch at 12°C, which is a typical time period for eggs to hatch in winter in Southern California and spring-summer in Monterey Bay. The only significant re- gression occurred in the SM region with SSTs 10-months earlier (r- = 0.46, P=0. 00.33, Fig. 5). We compared satel- lite-derived (AVHRR) estimates of SST for 1985-2002 from areas with high densities of paralarvae and juve- niles (within 8 km of shore) with CPUE by using cross correlation time series analysis. Although there were significant correlations, linear regression yielded no sig- nificant predictions for landings or CPUE from SST. Stock recruitment analysis We compared a paralarvae density index (PDI) with CPUE for the SCB and SM regions (Fig. 6, shaded area of Fig. 1). Collections of paralarvae were made in Febru- ary (Fig. 6). After the initial surveys of 1999, methods developed in Zeidberg and Hamner (2002) resulted in 34-50 stations for oblique bongo tows to collect paralar- vae in SCB and SM regions. Paralarvae/1000 m-' from Zeidberg et al : The fishery for Loligo opalescens from 1981 through 2003 53 APR CPUE/SOI 10 12 14 16 -16 -14 -i: -10 Time lag (months) 10 12 14 16 Figure 4 Time-series analysis: cross-correlation between CPUE and the global climatic indices: Southern Oscilla- tion index (SOI for atmospheric pressure differences between Tahiti and Darwin, A and B) and NIN03 (SST anomaly in eastern equatorial Pacific, C and D) for regions OCT (A, C) and APR/MB (B, D). The correlation between CPUE and N1N03 is greater than the correlation between CPUE and SOI in both regions. CPUE lags NIN03 by 9-11 months in APR and 4 months in OCT. Thus, the effects of an El Nifio event cause declines in CPUE for Loligo opalescens in Southern California 4 months later and in Monterey Bay 9-11 months later (long arrows). High correlation coefficients at a 10-month lag in Southern California (OCT) may be due to a second generation of market squid responding to changes in SST (shorter arrow). all stations were averaged to create the February PDI (Fig. 6 lower right), and this PDI was then compared to the November recruitment of spawning adults (CPUE) to the fishery for the same year. Linear regression was not significant for 1999-2003 {r'- = 0.522. P=0.1683). How- ever, if 1999 was treated as an outlier because it lacked nearshore sampling sites where 76% of the paralarvae were captured subsequently, the regression explains 97.8% of the variance, and the F-value of the ANOVA ratio test for this regression is significant, P=0.007 (Fig. 7J. From 1992 through 2002, SCB (36.2%) and SM (16.2%) represented nearly half of the landings for the state, and therefore this technique (regression of CPUE on PDI) could apply throughout the state. Discussion We report landings, effort, and catch per unit of effort for Loligo opalescens in California for 1981-2003. It is important to reiterate that CPUE is an approximation of abundance in the fishery and fails to estimate biomass of squid in California waters. Vessels that attempt to cap- 40 ■ "-« CPUE= 131.19 -7.24 (SST); r2 = 0.46 f 35 - • o W^ "-^ r 30 ■ ^*^^ --., 1990 o CD Z. 25 ■ o . ^^^ ~ - .t: ^^ ^^^w_ " ^ — '"^^ § 20 - ""^ ^"V,^ CD Q- 15 - 2002 •^"v ^V^ x: • ^ ^^V^ 5 10 - 1986 X ^\,^ • 13 5 14 14,5 15 15.5 16 16 5 17 Sea surface temperature (SST) 10 months prior Figure 5 Linear regression and 95'* confidence intervals for CPUE in November, the highest recruitment month, from SST of the previous January in the SM region. CPUE=131.19 - 7.24 xSST; r2 = 0.46, P=0.0033. 54 Fishery Bulletin 104(1) 2000 c^ •■ .•o;Ooo« ^^ 'f*\ V- 2001 \ 2002 "'°o ■•»-\ >.*' oas® ~. \ "^■r^.w 0 0- 1 1 -10 10-100 i 100- 1000 ■ •••••«> C,e* ', f*\ •^• i-o "^^ ^1000-10000 ^ Paralarvae/1 000 m3 V 2003 ^•" ■j^-S ^°° Paralarvae density index (PDI) 1 80 o o ° 60- V g Strong > 40 La Nina i5 year Weak El Nirio year 2000 2001 rear Figure 6 Density profiles (exponential bubble ploti for Loligo opalesceus paralarvae in the Southern California Bight, February 1999-2003 (shaded area of Fig. 1). Size of circles corresponds to number of paralarvae/1000 m-' seawater sampled. Data for 1999-2001 reprinted with permission from Springer-Verlag, originally in Zeidberg and Hamner (2002) Mar. Biol. 141(1):111-122. Data from all tows were averaged to obtain a paralarvae density index (PDI) for each year, lower right. 1999 La Niiia (cold) and 2002 El Nifio (warm) events are labeled above bars in the index. Note the lack of any sample sites within 8 km of shore or with >100 paralarvae/1000 m-= in 1999. ture squid and fail cannot be tracked with this method, and squid that are not harvested commercially are not accounted for in this report. Loligo opalescens reproduces by aggregating from small, foraging groups of hundreds of individuals to groups of millions of individuals. It is possible, therefore, that a large decrease in biomass can be masked by a larger percentage of the population aggregating in seemingly similar-size spawning masses. Such species are vulnerable to highly mobile fisheries (Oostenbrugge et al., 2002). Trends in the fishery The fishery for market squid (Loligo opalescens) has increased in all parts of the study area since 1983 because of an increase in demand for calamari inter- nationally and because of the collapse of other fisheries both within and outside California waters. The major- ity of fishing activity has shifted from Monterey Bay to the Southern California Bight since the 1980s. Fishing activity in the bight experienced a second increase in Zeidberg et al. The fishery for Loligo opalescem from 1981 through 2003 55 40- i- 35- Q_ ro o y rS 30- CD -S 25 • ■5 g If 20- X^ if '^- •/^ CPUE = 8.423 + 0.407 (PDI); ^^ r- = 0.978, P = 0.007 i^ 10. o "• 5 . 0 10 20 30 40 50 60 70 80 Paralarvae density index (PDI) February Figure 7 Stock-recruitment model: linear regression of catch per unit of effort I CPUE) for spawning adults in November on the February paralarvae density index I PDI) in the SM and SCB areas for 2000-2003. the 1990s, reflecting fishery participants from Alaska, Washington, and Oregon. The most economically harm- ful trend has been the substantial decrease in landings during the second year of strong El Niiio events, and the slight decrease in landings after weak ones. The initial impetus of performing the spectral analy- sis was to determine if the squid were migrating to the spawning grounds in relation to a lunar or tidal signal. It is important to note that the spectral analysis with CPUE and landings data (not shown) did not show that squid recruit to spawning sites in a fortnightly cycle. There was no 14-day period in any area. Spectral analy- sis demonstrated periodicities for CPUE of Loligo opal- escens on scales ranging from days to years. The most common periods for all areas were annual. Varying from 315 to 390 days, annual cycles made up more than half of the top ten signals in the analysis. The 4.5-year cycle corresponds well with the El Nirio events of 1982-83, 1987. 1992, and 1997-98 (Hayward et al., 1999). In each of these cases the CPUE anomalies were negative (Zeidberg, 2003). The longest period was 7.5 years in the MB and CC areas. There were evident leaps in the mean CPUE based on mean CPUE ±5 months in MB at mid-1988 and the end of 1995, when out-of-state fish- ermen began to harvest squid in California (Zeidberg, 2003). Although these leaps may correspond to changes in the biomass of the squid, they are more likely due to enhancements in the capacity of the fishery to capture squid as acoustic and communication technology has improved. The 3.7-year period is probably a statistical harmonic of the 7.5-year period. Paralarvae density index (PDI) can predict CPUE Zeidberg and Hamner (2002) have sampled the SCB and SM areas for Loligo opalescens paralarvae since 1999 and we used that data to create a paralarvae density index (PDI). CPUE appears to be a better indicator of stock abundance than landings data for squid (Sakurai et al., 2000). Adults recruiting to the fishery in November, measured in CPUE, can be pre- dicted by linear regression from the PDI of February. A regression of the CPUE data from the PDI data for 1999-2003 is not significant, but if 1999 is treated as an outlier the remaining four points (2000-03) create a regression that explains 97.8% of the variance. Our 1999 sampling of paralarvae may not be representative of the fishery because it was the first sampling year and the sampling sites were located farther offshore than those sampled in 2000-03. In 1999 there were no sites within 7.4 km of shore, where 76% of the paralar- vae were captured in the following four years of sam- pling. Despite these caveats, this method could provide the first opportunity to manage California's market squid fishery according to scientifically gathered bio- logical indicators and with very few of the inherent assumptions needed for many other types of forecast- ing (Mangel et al., 2002). As the years of logbook data accumulate, estimates of CPUE will be more closely related to the actual biomass of the species. By the end of February, we can have a prediction for the CPUE for the following year's adult recruitment. Paralarvae may be the best stage of the life cycle for a fishery prediction because juveniles can escape trawls, fewer assumptions need to be made than with estimates from spawning females (Macewicz et al., 2004), and there is sufficient time (6-9 months) to develop predictions. These predictions could help managers set catch limits and aid fishermen in deciding how to invest in gear for the following season. In addition to our paralarvae sampling, CalCOFI has sampled the waters of California for zooplankton in a manner similar to ours since 1949. Paralarval distributions for Loligo opalescens have been described from these data (Okutani and McGowan, 1969). The greatest difference between the two sampling efforts is the number of stations that are in close proximity to land. The majority of the paralarvae (76%) captured by Zeidberg and Hamner (2002) were at stations less than 8 km from shore, but there is only one CalCOFI station at this proximity to land. After reviewing their surveys and models of larval dispersal (Franks, 1992; Botsford et al., 2001; Siegel, 2003), we predict that a PDI calculated from CalCOFI samples will be substan- tially lower than ours, but given the long time period of the CalCOFI sampling program, any significant cor- relations could be more powerful statistically than ours. Furthermore, fishermen could be employed to perform bongo tows for paralarvae in proximity to shore to com- plement CalCOFI data. If the CalCOFI bongo net data were sorted for Loligo opalescens paralarvae, and fisher- men collected paralarvae nearshore, Monterey Bay and southern California CPUE could be predicted months in advance. Separate management of the two regions would be necessary given the time lag of recruitment (APR and OCT). 56 Fishery Bulletin 104(1) Comparison of fishery data with physical data We found a correlation between CPUE of the largest recruitment month with SST buoy data from 10 months earlier in the SM area only. There may be physical features specific to this area that increase the cor- relation between spawning recruitment and SST. For example, SM is a small area, it is close to the buoy, most of the area is sandy bottom, and it contains the Redondo Canyon. Thus if further attempts to match physical oceanography to the biology of a pelagic species were to occur, the Santa Monica Bay could be the most ideal location. But this correlation between CPUE of the largest recruitment month with SST in the SM area may be a seasonal effect because the regression is significant for SST only and not an SST anomaly. Furthermore, we caution that the significance of the correlation between CPUE and SST in SM may be a type-I error because it was the only significant test of the 30 tests run with an alpha level of 0.05. The size of the recruitment event was not strongly related to small deviations from aver- age monthly SST; thus the timing of squid recruitment to spawning grounds in APR and OCT may be tied to annual fluctuations of prey availability and correlations with temperature may be coincidental. The 10-month lag corresponds to the egg-laying date of 9-month-old squid. The lack of a greater number of correlations may be due to the small spatial resolution of the buoy data and the enormous variability of SST data due to meso- scale oceanographic features in the large fishery areas. In some areas the nearest buoy was quite distant from the fishery zone. To address the spatial distance of spawning grounds from buoys, we compared SSTs derived from satel- lite AVHRR images with CPUE. AVHRR data were collected from just the shelves and slopes of the six fishery areas because these are the most important areas for the growth of hatchlings and juveniles. Cross- correlation time series analyses were significant at 5-10 month lags (Table 3), but this significance did not translate into any predictive capabilities with linear regression. Similarly, cross-correlations of CPUE with SOI and NIN03 were significant at a 10-month lag in Monterey Bay and a 4-month lag in the Southern California Bight. Thus the Monterey fishery (10% of landings) is offset by six months (roughly one short cohort) from the SCB fishery. The correlation coefficients for NIN03 were greater than those of SOI, corroborating the idea that the direct influence of the coastal waves ("oceanic teleconnection") is the main source of the changes in the hydrographic and ecological features of the Califor- nia Current system (Huyer and Smith, 1985; Rienecker and Mooers, 1986; Lynn et al., 1995; Chavez, 1996; Ramp et al., 1997) rather than the ENSO (El Niiio Southern Oscillation)-related changes of atmospheric circulation ("atmospheric teleconnection") (Simpson, 1983; Simpson, 1984a, 1984b; Mysak, 1986; Breaker and Lewis, 1988; Breaker et al., 2001; Schwing et al., 2002). Loliginid life cycles and future management of squid fisheries The correlation between SST and CPUE in the following season may have resulted from the unique development pattern of teuthids. The use of CPUE as an index of abundance of the population (Sakurai et al., 2000), in combination with studies of squid growth in relation to SST (Jackson and Domeier, 2003), could explain large fluctuations in landings data from year to year. In terms of bottom-up forcing, individual squid health and the resulting population size result from a combination of prey availability and metabolic rates. Squids grow exponentially in the first two months of life and then logarithmically until senescence. In rearing tanks and given a constant food supply, loliginids also grow faster in warmer temperatures (Yang et al., 1986; Forsythe et al., 2001) as their metabolic rates increase (O'Dor, 1982). Grist and des Clers (1998) predicted that annual fluctuations in SST that cause differential growth in squids can lead to younger cohorts hatched in warm temperatures and surpassing in size older cohorts born in colder seasons. Thus in October, a large 6-month-old squid that hatched in April and developed in warm water may spawn with a smaller 9-month-old squid that hatched in the cold waters of January. However, in the California system and possibly in other upwelling systems the situation is more complex than in rearing tanks. For example, Jackson and Do- meier (2003) demonstrated that due to the influences of El Nino and La Nifia cycles and upwelling. the mean mantle length of Loligo opalescens is shortest when larvae are hatched in the warmest temperatures and longest when hatched in cold waters. Mantle length is also positively correlated with the upwelling index. In the ocean, squid do not have a constant food supply. The high productivity and cold temperatures caused by upwelling and La Nina combined to create a period of rich food resources and lower metabolic rates for squid, probably enhancing the recovery of the fishery in 1999. During the El Nifio event the squids were small and less abundant because they had a high metabolic rate due to increased temperatures and were exposed to lower levels of available prey due to decreased ocean produc- tivity. Seasonal maxima of phytoplankton in Monterey Bay occur in summer; but in the southern part of the Southern California Bight productivity peaks in win- ter (Nezlin et al., 2002). These differences may be an indicator of why the fishery operates in Monterey Bay from April to November, coinciding with the upwelling season, and in the Southern California Bight from No- vember to May, coinciding with less stratification of the water column and more mixing due to winter storms and colder air temperatures. Lowry and Carretta (1999) corroborated the tempera- ture-induced plasticity of mantle length (ML) from beaks of squid in California sea lion (Zalophus californianus) scats and spewings. MLs of squid prey were half the size during El Nino years on San Clemente and Santa Barbara islands. However, at San Nicholas Island dur- ing El Nino events, there were both small- and regular- Zeidberg et al : The fishery for Loligo opolescens from 1981 through 2003 57 size squid prey; this finding may indicate that the squid stock moved offshore to find productive waters. Alter- natively, San Nicholas sea lions may have been feeding on squids from Baja California. Zeidberg and Hamner (2002) suggested the possibility of a northern shift of the squid population in El Niiio years, as has been found for most zooplankton (Colebrook, 1977). However, the growth plasticity and fluctuating re- productive success for Loligo opalescens should not be underestimated. The possibility remains that the huge fluctuations in squid landings during strong ENSO events is due to the entire biomass of market squid waning and waxing rather than to population migra- tions away from traditionally fished spawning grounds. Triennial groundfish surveys demonstrate that market squid experienced a coast-wide population decrease, not a poleward migration during the 1997-98 El Nino.*^ With the exception of El Nino years, the fishery in- creased its landings each year until 2000. However, it remains unknown if the capacity of the fishery is close to reaching the total biomass of squid in California. The California sardine {Sardinops sagax) fishery col- lapsed in the 1960s, and a twenty-year moratorium was required before there was recovery to a fraction of prior spawning biomass (Wolf, 1992). Whether over-fishing or large scale, multidecadal climatic regime shifts caused this collapse is matter of debate (Chavez et al., 2003), but without an effective management plan, squid will continue to be fished because of market demand. Mar- kets are driven by economic forces and traditionally do not control themselves in a biologically sustainable manner. A full recovery of the squid fishery occurred from 1998 to 2000 and thus spanned four generations of squid given a 6-9 month lifecyle; for the California sar- dine (with a 6-8 year lifecycle), a proportionally similar recovery period would be 24-32 years (Parrish^). In 1998-99 the fishery for Loligo opalescens decreased to low levels during the El Nino event, then recovered to record levels in the following years. This was most likely due to the plasticity of squid development in rela- tion to water temperature and upwelling and the short (4-6 month) life span of squid. One should not assume that the ability of this species to recover from environ- mental stress like El Nino applies also to the recent anthropogenic stresses associated with increasing fish- ery capacity. It remains to be seen if the large decline in southern California landings in the last five years (119,780-24,449 tons/year) is due to the small El Nino of 2002-03, the climate-regime shift in 1998, overfish- ing, or some other factor such as increased water strati- fication due to global warming. Although the short-lived squid may not be able to recover from overexploitation " CDFG (California Department of Fish and Game). 2005. Fi- nal market squid fishery management plan. Website: http:// www.dfg.ca.gov/mrd/msfmp/index.html [Accessed on 6 June 2005.] ' Parrish, R. 2005. Personal commun. Fisheries and Marine Ecosystems Program, Pacific Fisheries Environ- mental Laboratory, 1352 Lighthouse Ave. Pacific Grove, CA 93950-2097. in short order, the huge number of long-lived birds, fish, and marine mammals (Morejohn et al., 1978; Lowry and Carretta. 1999) that depend on squid as a key forage species may not be able to recover rapidly from lack of management foresight. The recent establishment of the marine reserve system in the Channel Islands elimi- nates 139f of key squid fishing grounds. This ecosystem- based management approach may assist in protecting not only the squid but also their predators. Acknowledgments This research was funded by a California Fish and Game award (no. FG7334MR) and a Coastal Environmental Quality Initiative Program grant (no. 783828-KH-19900) and was supported in part by the David and Lucile Pack- ard Foundation through the Monteray Bay Aquarium Research Institute. Bruce Robison, Rich Ambrose, Peggy Fong, and Dick Zimmer provided thoughtful reviews. Andrea Steinberger assisted with data management. Literature cited Agnew, D. J., J. R. Beddington, and S. L. Hill. 2002. The potential use of environmental information to manage squid stocks. Can J. Fish. Aquat. Sci. 59:1851-1857. Agnew, D. J., S. Hill, and J. R. Beddington. 2000. Predicting the recruitment strength of an annual squid stock: Loligo gahi around the Falkland Islands. Can J. Fish. Aquat. Sci. 57:2479-2487. Aristegui, J., P. Tett, A. Hernandez-Guerra, G. Basterretxea, M. F. Montero, K. Wild. P. Sangra, S. Hernandez-Leon, M. Canton, J. A. Garcia-Braun, M. Pacheco, and E. D. Barton. 1997. The influence of island-generated eddies on chloro- phyll distribution: a stufy of mesoscale variation around Gran Canaria. Deep Sea Res. 44:7-96. Bakun, A. 1973. Coastal upwelling indices, west coast of North America, 1946-1971, 103 p. NOAA Tech. Rep. NMFS SSRF-671. Botsford L. W., A. Hastings, and S. D. Gaines. 2001. Dependence of sustainability on the configuration of marine reserves and larval dispersal distance. Ecol. Lett. 4(2):144-150. Bray N. A., A. Keyes, and W. M. L. Morawitz. 1999. The California Current system in the Southern California Bight and Santa Barbara Channel. J. Geo- phys Res 104 C4:7695-7714. Breaker, L. C, and P. A. W. Lewis. 1988. A 40-50 day oscillation in sea-surface temperature along the central California coast. Estuar. Coast. Shelf Sci. 26:395-408. Breaker, L. C, P. C. Liu, and C. Torrence. 2001. Intraseasonal oscillations in sea surface tempera- ture, wind stress, and sea level off the central California coast. Cont. Shelf Res. 21(6-7):727-750. Brink, K. H., R. C. Beardsley, J. Paduan, R. Limeburner, M. Caruso, and J. G. Sires. 2000. A view of the 1993-1994 California Current based on surface drifters, floats, and remotely sensed data. J. Geophys. Res. Oceans. 105(C4):8575-8604. 58 Fishery Bulletin 104(1) Caddy, J. F., and P. G. Rodhouse 1998. Cephalopod and groundfish landings: evidence for ecological change in global fisheries? Rev. Fish Biol. Fish. 8:431-444. Chavez. F. P. 1996. Forcing and biological impact of onset of the 1992 El Nino in central California. Geophys. Res. Let. 23(3):265-268. Chavez, F. P., J. Ryan, S. E. Lluch-Cota, and M. Niquen-C. 2003. From anchovies to sardines and back: multidecadal change in the pacific ocean. Science 299:217-221. Colebrook, J. M. 1977. Annual fluctuations in biomass of taxonomic groups of zooplankton in the California Current 1955- 59. Fish. Bull. 75:357-368. DiGiacomo, P. M., and B. Holt. 2001. Satellite observations of small coastal ocean eddies in the Southern California Bight. J. Geophys. Res. 106 (C10):22, 521-22, 543. Falkowski, P. G.. D. Ziemann, Z. Kolber, P. K. Bienfang. 1991. Role of eddy pumping in enhanced primary pro- duction on the ocean. Nature 352:55-58. Forsythe, J. W., L. S. Walsh, P. E. Turk, and P. G. Lee. 2001. Impact of temperature on juvenile growth and age at first egg-laying of the Pacific reef squid Sepio- teuthis lessoniana reared in captivity. Mar. Biol. 138:103-112. Franks, P. J. S. 1992. Sink or swim: accumulation of biomass at fronts. Mar. Ecol. Prog. Ser. 82:1-12. Graham, W. M., and J. L. Largier. 1997. Upwelling shadows as nearshore retention sites: the example of northern Monterey Bay. Cont. Shelf Res. 17:509-532. Grist, E. P. M., and S. des Clers. 1998. How seasonal temperature variations may influ- ence the structure of annual squid populations. IMA J. of Math. Appl. Med. Biol. 15:187-209. Harms, S., and C. D. Winant. 1998. Characteristic patterns of the circulation in the Santa Barbara Channel. J. Geophys. Res. V103, C2:3041-3065. Hayward T. L., T. R. Baumgartner, D. M. Checkley, R. Durazo, G. Gaxiola-Castro, K. D. Hyrenbach, A. W. Mantyla, M. M. Mulhn, T. Murphree, F. B. Schwing, P. E. Smith, M. J. Tegner. 1999. The state of the California Current in 1998-1999: Transition to cool-water conditions. Calif Coop Oceanic Fish Invest Rep 40:29-62. Hernandez-Trujillo, S. 1999. Variability of community structure of Copepod related to El Ninol982-83 and 1987-88 along the west coast of Baja California Peninsula, Mexico. Fish. Oceanog. 8:284-295. Hickey B. M., E. L. Dobbins, and S. E. Allen. 2003. Local and remote forcing of currents and tem- perature in the central Southern California Bight. J. Geophys. Res. Oceans 108(C3):3081-3084. Huyer, A., and R. L. Smith. 1985. The signature of El Nifio off Oregon in 1982-83. J. Geophys. Res. 90:7133-7142. Jackson, G. D., and M. L. Domeier. 2003. The effects of an extraordinary El Nino/La Niiia event on the size and growth of the squid Loligo opalescens off Southern California. Mar. Biol. 142:925-935. Leos, R. R. 1998. The biological characteristics of the Monterey Bay squid catch and the effect of a two-day-per-week fishing closure. CalCOFI Reports 39:204-211. Loukashkin, A. S. 1976. On biology of market squid, Loligo opalescens, a contribution toward the knowledge of its food habits and feeding behavior. CalCOFI Reports 18:109-111. Lowry, M. S., and J. V. Carretta. 1999. Market squid {Loligo opalescens) in the diet of Cali- fornia sea lions {Zalophus californianus) in southern Cal- ifornia (1981-1995). CalCOFI Reports 40:196-207. Lynn, R. J., F. B. Schwing, T. L. Hayward. 1995. The effect of the 1991-1993 ENSO on the California Current system. CalCOFI Reports 36:57-71. Macewicz, B. J., J. R. Hunter, N. C. H. Lo, E. L LaCasella. 2004. Fecundity, egg deposition, and mortality of market squid [Loligo opalescens). Fish. Bull. 102:306-327. Mangel, M., B. Marinovic, C. Pomeroy, and D. CroU. 2002. Reqiuem for Ricker: unpacking MSY. Bull. Mar. Sci. 70:763-781. Morejohn, G. V., J. T. Harvey, and L. T. Krasnow. 1978. The importance of Loligo opalescens in the food web of Marine vertebrates in Monterey Bay, Calif Fish Game Fish Bull. 169:67-98. Mclnnis, R. R., and W. M. Broenkow. 1978. Correlations between squid catches and oceanog- raphy conditions in Monterey Bay, Calif. Fish Game FishBull. 169:161-170. Mysak, L. A. 1986. El Nino, interannual variability and fisheries in the northeast Pacific Ocean. Can. J. Fish. Aquat. Sci. 43:464-497. Nezlin, N. P., W. M. Hamner, and L. D. Zeidberg. 2005. Influence of El Nino 1997-1998 on the Pelagic Ecosystem off California: remote sensed analysis. In California and the World Ocean '02 (O. T. Magoon, H. Converse, B. Baird, B. Jines, and M. Miller-Henson, eds.), p. 1233-1252. Nezlin, N. P., and B. L. Li. 2003. Time series analysis of remote-sensed chlorophyll and environmental factors in the Santa Monica-San Pedro Basin off Southern California. J. Mar. Syst. 39:185-202. O'Dor, R. K. 1982. Respiratory metabolism and swimming perfor- mance of the squid Loligo opalescens. Can J. Fish. Aquat. Sci. 39:580-587. Okutani T., and J. A. McGowan. 1969. Systematics, distribution and abundance of the epiplanktonic squid (Cephalopoda, Decapoda) larvae of the California Current April, 1954-March, 1957. Bull. Scripps Inst. Oceanog. 14:1-90. Oostenbrugge, J. A. E., J. J. Poos, W. L. T. van Densen, and M. A. M. Machiels. 2002. In search of a better unit of effort in the coastal liftnet fishery with lights for small pelagics in Indonesia. Fish. Res. 59:43-56. Pauly, D., V. Christensen, J. Dalsgaard, R. Froese, and F. Torres Jr. 1998. Fishing down marine food webs. Science 279: 860-863. Ramp, S. R., J. L. McClean, C. A. Collins, A. J. Semtner, and K. A. S. Hays. 1997. Observations and modeling of the 1991-1992 EI Nifio signal off central California. J. Geophys. Res. 102(C3):5553-5582. Zeldberg et a\ The fishery for Loligo opolescens from 198f through 2003 59 Ravier, C, and J. M. Fromentin. 2001. Long-term fluctuations in the eastern Atlantic and Mediterranean bluefin tuna population. ICES J. Mar. Sci. 58:1299-1317. Rienecker, M. M., and C. N. K. Mooers. 1986. The 1982-1983 El Niiio signal off" northern California. J Geophys. Res. 91:6597-6608. Roberts, M. J., and W. H. H. Sauer. 1994. Environment: the key to understanding the South African chokka squid iLoligo vulgaris reynaudii) life cycle and fishery? Antarctic Science 6:249-258. Robin. J. P., and V. Denis. 1999. Squid stock fluctuations and water temperature: temporal analysis of English Channel Loliginidae. J. Appl. Ecol. 36:101-110. Rogers-Bennett, L. 2000. Review of some California fisheries for 1999: market squid, dungeness crab, sea urchin, prawn, abalone, groundfish, swordfish and shark, ocean salmon, nearshore finfish. Pacific sardine. Pacific her- ring. Pacific mackerel, reduction, white seabass, and recreational. CalCOFI Rep. 41:8-26. Rocha, F., A. Guerra, R. Prego, U. Piatkowski. 1999. Cephalopod paralarvae and upwelling condi- tions off Galician waters ( NW Spainl. J. Plank. Res. 21:21-33. Sakurai, Y., H. Kiyofuji, S. Saitoh, T. Goto, and Y. Hiyama. 2000. Changes in inferred spawning areas of Toda- rodes pacificus (Cephalopoda: Ommastrephidae) due to changing environmental conditions. ICES J. Mar. Sci. 57:24-30. Schwing, F. B., T. Murphree, L. deWitt, and P. M. Green. 2002. The evolution of oceanic and atmospheric anom- alies in the northeast Pacific during the El Nifio and La Nina events of 1995-2001. Prog. Oceanog. 54(l-4):459-491. Siegel, D. A., B. P. Kinlan, B. Gaylord, and S. D. Gaines. 2003. Lagrangian descriptions of marine larval dis- persion. Mar. Ecol. Prog. Ser. 260:83-96. Simpson, J. J. 1983. Large-scale thermal anomalies in the California Current during the 1982-1983 El Nino. Geophys. Res. Let. 10:937-940. 1984a. El Nino-induced onshore transport in the Cali- fornia current during 1982-1983. Geophys. Res. Let. ll(3):241-242. 1984b. A simple model of the 1982-1983 Californian "El Nino". Geophys. Res. Let. ll(3):243-246. Vojkovich, M. 1998. The California fishery for market squid (Loligo opalescens). CalCOFI Rep. 39:55-61. Waluda, C. M., P. N. Trathan, and P. G. Rodhouse. 1999. Influence of the oceanographic vriability on recruit- ment in the Illex argentinus i Cephalopoda: Ommas- trephidae) fishery in the South Atlantic. Mar. Ecol. Prog. Ser. 183:159-167. Waluda, C. M., P.G. Rodhouse, G. P. Podesta, P. N. Trathan, and G. J. Pierce. 2001. Surface oceanography of the inferred hatching grounds of Illex argentinus (Cephalopoda: Ommastrephi- dae) and influences on recruitment variability. Mar. Biol. 139:671-679. Watson, R., and D. Pauly. 2001. Systematic distortions in world fisheries catch trends. Nature 414:534-536. Wing, S. R., J. L. Largier, and L. W. Botsford. 1995. Settlement and transport of benthic invertebrates in an intermittent upwelling region. Limnol. Oceanog. 40(2):316-329. Wolf, P 1992. Recovery of the Pacific sardine and the California sardine fishery. CalCOFI Reports 33:76-86. Yang. W. T, R. F. Hixon, P. E. Turk, M. E. Krejci, W. H. Hulet, and R. T. Hanlon. 1986. Growth, behavior, and sexual maturation of the market squid, Loligo opalescens, cultured through the life cycle. Fish. Bull. 84:771-798. Zar, J. H. 1984. Biostatistical analysis, 2"'^ ed., 718 p. Prentice Hall, Englewood Cliffs, NJ. Zeidberg, L. D., and W. M. Hamner. 2002. Distribution of squid paralarvae, Loligo opal- escens (Cephalopoda: Myopsida), in the Southern Cali- fornia Bight in the three years following the 1997 El Niiio. Mar. Biol. 141:111-122. Zeidberg, L. D. 2003. The early life history and fishery of the Califor- nia market squid, Loligo opalescens. Ph.D. diss., 136 p. Univ. California Los Angeles, Los Angeles, CA. Zeidberg, L. D., W. Hamner, K. Moorehead, and E. Kristof 2004. Egg masses of Loligo opalescens (Cephalop- oda: Myopsida) in Monterey Bay, California follow- ing the El Nino event of 1997-1998. Bull. Mar. Sci. 74:129-141. 60 Abstract — The variability in the supply of pink shrimp (Farfante- penaeus duorarum) postlarvae and the transport mechanisms of plank- tonic stages were investigated with field data and simulations of trans- port. Postlarvae entering the nursery grounds of Florida Bay were collected for three consecutive years at chan- nels that connect the Bay with the Gulf of Mexico, and in channels of the Middle Florida Keys that connect the southeastern margin of the Bay with the Atlantic Ocean. The influ.x of postlarvae in the Middle Florida Keys was low in magnitude and varied sea- sonally and among years. In contrast, the greater postlarval influx occurred at the northwestern border of the Bay, where there was a strong seasonal pattern with peaks in influx from July through September each year. Planktonic stages need to travel up to 150 km eastward between spawning grounds (northeast of Dry Tortugast and nursery grounds (western Florida Bay) in about 30 days, the estimated time of planktonic development for this species. A Lagrangian trajectory model was developed to estimate the drift of planktonic stages across the SW Florida shelf. The model simu- lated the maximal distance traveled by planktonic stages under various assumptions of behavior. Simulation results indicated that larvae traveling with the instantaneous current and exhibiting a diel behavior travel up to 65 km and 75% of the larvae travel only 30 km. However, the eastward distance traveled increased substan- tially when a larval response to tides was added to the behavioral variable (distance increased to 200 km and 85% of larvae traveled 150 km). The question is, when during larval devel- opment, and where on the shallow SW Florida shelf, does the tidal response become incorporated into the behavior of pink shrimp. Variability in supply and cross-shelf transport of pink shrimp iFarfantepenaeus duorarum) postlarvae into western Florida Bay Maria M. Criales' John D. Wang^ Joan A. Browder^ Michael B. Robblee'* Thomas L. Jackson^ Clinton Hittle" ' Rosenstiel School of Marine and Atmospheric Science, MBF University ol Miami 4600 Rickenbacker Causeway Miami, Florida 33149 E-mail address (for M M Cnales) mcrialesa'Tsmas miami edu ^ Rosenstiel School of Marine and Atmospheric Science, AMP University of Miami 4600 Rickenbacker Causeway Miami, Florida 33149 3 NOAA Fisheries, Southeast Fisheries Science Center 75 Virginia Beach Drive Miami, Florida 33149 ■* United States Geological Survey Center for Water and Restoration Studies 3110 SW9'^ Avenue Ft Lauderdale, Florida 33315 Manuscript submitted 16 September 2003 to the Scientific Editor's Office. Manuscript approved for publication 30 June 2005 by the Scientific Editor. Fish. Bull. 104:60-74 (2006). Patterns of recruitment of coastal spe- cies are highly variable, mainly because of the complex interaction of biotic anii abiotic factors across the different life history stages. These factors include but are not limited to reproductive dynamics, larval dispersal and behav- ior, physiological tolerances, and the hydrometeorological regime in which their life stages develop (e.g., Shanks, 1995; Cowen, 2002). The commercially valuable tropical penaeid shrimps that use different habitats during their life cycle (offshore spawning grounds and estuarine nursery habitats) have to cope with a large suite of physical processes and stimuli (Rothlisberg et al., 1995, 1996). The pink shrimp iFarfantepenaeus duorarum) of Dry Tortugas is one of the most economi- cally and ecologically important spe- cies in southwest Florida. The pink shrimp supports an important year- round fishery of about 4000 metric tons in an area of 10,000 km- between Dry Tortugas and Key West (Iversen et al., 1960; Klima et al., 1986). The Tortugas fishery is directly depen- dent on young shrimp that migrate from inshore nursery areas onto the offshore fishing grounds (Sheridan, 1996; Browder et al., 2002). Recruit- ment shows no relationship to spawn- ing size; therefore harvest fluctuations are apparently due to environmental conditions rather than fishing opera- tions (Nance and Patella, 1989). To effectively manage this species, it is necessary to have accurate informa- tion on the processes linking nursery and spawning ground populations. Pink shrimp population dynamics are affected by physical processes and environmental conditions occurring in the southwestern (SW) Florida Shelf of the Gulf of Mexico, the At- lantic coastal zone, and the Florida Bay. Early research on gonad develop- ment (Cummings, 1961), distribution of larval stages (Jones et al., 1970), and analysis of length frequency distributions of fishery stock data (Iversen et al., 1960; Roberts, 1986) indicated that the center of spawn- Criales et al Variability in supply and cross shelf transport of Farfanlepenaeus duorarum postlarvae into Florida Bay 61 ing is northeast of the Dry Tortugas. If gravid females spawn northeast of the Dry Tortugas, larvae need to travel up to 150 km to reach the main nursery ground in western Florida Bay. Females spawn on the continen- tal shelf at about 30 m of depth, where larvae develop, passing through several changes in feeding habitats, behavior, and physical stages (nauplii, zoeae, myses) (Dobkin, 1961; Ewald, 1965; Jones et al, 1970). Postlar- vae undergo between three and eight additional plank- tonic stages before settlement. Larvae develop rapidly, needing only about 30 days to become postlarvae ready to settle to the bottom (Ewald, 1965; Dobkin, 1961). Planktonic stages (larvae and postlarvae) approach the coast and postlarvae enter the nursery grounds of Florida Bay at about 9-10 mm total length (Tabb et al., 1962; Allen et al., 1980; Criales et al., 2000). Larval development and ocean hydrodynamics must be tightly linked to successfully bring these planktonic stages to their coastal nursery grounds. Mechanisms of transport used by planktonic stages of penaeid shrimps are highly variable, depending on the species, different environmental conditions, oceanic physical processes, and complexity of larval behaviors (e.g., Dall, 1990; Rothlisberg et al., 1995, 1996; Wenner et al., 2005). Physical oceanographic processes signifi- cantly affect the transport of planktonic stages from spawning to nursery grounds (Yeung and Lee, 2002; Criales et al., 2003). Two main immigration routes have been hypothesized for pink shrimp postlarvae entering Florida Bay: 1) postlarvae may drift south-southeast downstream with the Florida Current and enter Florida Bay through the tidal channels of the Lower and Middle Florida Keys (Rehrer et al., 1967; Munro et al., 1968), and 2) postlarvae may move northeast across the SW Florida shelf and enter the Bay at its northwestern boundary (Jones et al., 1970; Criales and Lee, 1995). The most widely recognized pathway for postlarvae to reach Florida Bay up-to-now has been by transport up the Atlantic side through the tidal channels of the Middle Florida Keys (Munro et al., 1968; Criales and McGowan, 1994; Criales et al., 2003). The favorable Ekman transport generated by the southeastern winds along the west-east oriented coast, and coastal counter- current flow generated by cyclonic eddies provide favor- able onshore transport mechanisms along the Florida Keys coast (Criales and Lee, 1995; Lee and Williams, 1999). In contrast, larval transport across the broad, shallow SW Florida Shelf has not been well studied and questions exist about the feasibility of this pathway. Subtidal frequency flows are weak in the SW Florida shelf and mainly in the alongshore (north-south) direc- tion as a direct response to wind events (Koczy et al., 1960; Weisberg et al., 1996; Lee et al., 2001). Tidal currents are strong mainly in the cross-shelf direction (Wang, 1998; Smith, 2000). Freshwater discharges from the Everglades affect a broad area of the SW Florida shelf (Lee et al., 2001; Jurado, 2003). Isopleths less than 32 are typically confined to the region between Cape Sable and Cape Romano, and from 32 to 36 extend from near Cape Romano to the vicinity of Dry Tortugas in a highly variable annual pattern (Lee et al., 2001; Johns and Szymanski'). For tropical penaeid shrimps that undergo larval development offshore, but whose nursery grounds are inshore, migratory behavior is a key factor for their advection to nursery grounds (Dall et al., 1990; Shanks, 1995). The simplest migratory behavior is vertical movement, and three types of vertical migrations are known to mediate horizontal transport of larvae: on- togenic, diel, and tidal (for reviews see Sponaugle et al., 2002). For some Australian penaeid species (ba- nana prawn [Fenneropenaeus merguiensis], grooved tiger prawn [Penaeus seinisulcatus], and eastern king prawn [Melicertus plebejus]) it has been shown that early planktonic stages (protozoeae and myses) perform diel vertical migration cued by light and that later in development (as postlarvae) the migration is cued by tides (Rothlisberg, 1982; Rothlisberg et al., 1983, 1995) and there is no cross-shelf displacement of larvae dur- ing the 15 days of diel behavior. Previous studies of pink shrimp in South Florida have clearly indicated ontogenic behavior for pink shrimp; postlarvae have a higher degree of mobility than earlier protozoeae and myses (Temple and Fischer, 1965; Eldred et al., 1965; Jones et al., 1970; Criales and Lee, 1995). On the other hand, diel behavior is not so well determined. Although protozoeae, myses and postlarvae were more abundant at the surface during the night than during the day, day and night differences have not been statistically significant in any of these previous studies. The effect of diel, tidal, or ontogenic combinations of behavior on cross-shelf transport from South Florida spawning grounds to western Florida Bay nursery grounds has not previously been explored. The postlarvae of many penaeid shrimps, including pink shrimp, are known to synchronize vertical migration with the tides at the entrance to estuarine nursery grounds (for reviews see Garcia and Le Reste 1981, Dall et al., 1990). This pro- cess is known as selective tidal stream transport (STST) (Forward and Tankersley, 2001). Penaeid postlarvae as- cend in the water column during the flood and sit on the bottom during the ebb to maximize up-estuary move- ment (e.g., Rothlisberg et al., 1995). This behavior has been shown for pink shrimp postlarvae inside Florida Bay (Tabb et al., 1962; Roessler and Rehrer, 1971), but not along the border of the bay with the Gulf of Mexico. When during the life cycle and where on the shelf this tidal behavior begins and what the environmental cues are — these questions remain unanswered. The purpose of our research was 1) to determine pat- terns of supply of pink shrimp postlarvae into Florida Bay through two distinct regions, 2) to define the most important transport route for planktonic stages from the Dry Tortugas into Florida Bay, 3) to examine al- ternative behavioral responses of larvae and postlar- vae, and 4) to propose a recruitment mechanism for ' Johns, E., and D. Szymanski. 2003. Mixing it up in Florida Bay. Florida Bay News, summer 2003:1-3. 62 Fishery Bulletin 104(1) 26°N (s 25°N 24-N 83 W 82'W 81 "W 80°W Figure 1 Map of the study area (with bathymetryl showing the channel net sampling stations at the northwestern and southeastern borders of the nursery grounds of Florida Bay, ADCP moorings, and CMAN and COMP stations. The Tortugas fishing grounds (area enclosed by dashed line) includes the center of spawning for Farfantepenaeus duorarum (filled ellipse). Northwestern stations: MG = Middle Ground and SK=Sandy Key; Florida Key stations: WH=Whale Harbor and PH = Panhandle; A and B = onshore and offshore ADCP moorings respectively; l = Long Key CMAN station; 2 = NW Florida Bay COMP station. Small map at the left corner indicates major currents in the Gulf of Mexico and off the coast of Florida. CC = Caribbean Current; LC = Loop Current; FC = Florida Current; GS = Gulf Stream. pink shrimp across the SW Florida shelf that combines the effect of hydrodynamics with larval behavior. Four modes of behavior-related transport were simulated under hydrodynamic conditions occurring on the SW Florida shelf in order that the resulting distances trav- eled under each condition might be contrasted. Material and methods Pink shrimp postlarvae were collected in two regions of Florida Bay to evaluate postulated hypotheses of eastern and western gateways and pathways of larval transport into the bay. The two study sites consisted of two large channels connecting northwestern Florida Bay with the SW Florida shelf of the Gulf of Mexico (Sandy Key Chan- nel [SK], and Middle Ground [MG]); and of more confined tidal channels in the Middle Florida Keys that connect the Bay with the Atlantic Ocean (Whale Harbor [WH], and Panhandle Key [PH]) (Fig. 1). Adjacent mud banks that are occasionally exposed at low tide and over-topped at higher tide stages define these channels. The averaged depth, tidal flow, and cross sectional area of the four channels are summarized in Table 1 to show the differ- ent levels of water flow through these pathways. Chan- nels depths are similar, but western channels (Middle Ground and Sandy Key) have higher tidal flows and larger cross sectional areas than the Florida Keys chan- nels (Table 1). Tidal fluctuations are primarily semidiur- nal at all stations and have weaker diurnal constituents (Smith, 1998, 2000). Acoustic Doppler velocity meters (ADVMs) that measure continuous velocity and depth (tide and stage) and associated CTD instruments that measure conductivity and temperature were installed at each channel in January 2002. A boat-mounted acoustic Doppler current profiler (ADCP) was used to calculate total discharge across the cross-section of the channel during monthly sampling (Hittle et al., 2001). Cnales et al : Variability in supply and cross shelf transport of Farfanlepenaeus duororum postlarvae into Florida Bay 63 Postlarvae were collected monthly in each channel during two nights around the new moon from January 2000 to December 2002. At the PH station sampling began in June 2000 after the original site {Captain Key channel) was abandoned because it had insufficient wa- ter flow for effective sampling. Two moored subsurface channel nets (net 1 and net 2) of 0.75-m- opening, 1-mm mesh size, and SOO-nm mesh in the codend were sus- pended with floats at 0.5 m depth. Nets were deployed each night before dusk and removed shortly after dawn each day. General Oceanic flowmeters (2030R16, Low Speed Rotor, Miami, FL) were mounted in the mouth of the nets and the volume of water filtered through the nets was calculated for each net. Farfanlepenaeus duorarum postlarvae were sorted, identified, and pre- served in 90% ethanol. The raw catch in each sample was standardized to numbers of postlarvae per 1000 m'^ of water filtered. The average number of postlarvae over the two sampling nights was used as the mean month- ly concentration for each net. The average of monthly postlarval concentration for each region (northwestern Florida Bay vs. Florida Keys) was compared by using a nonparametric two-way analysis of variance (ANOVA) (Anderson, 2001). Three 12-hour experiments were conducted in sum- mer 2002 in the SK channel to document the behavioral response of pink shrimp postlarvae to ebb and flood tides. Consecutive pairs of night (i.e. dark) flood and dark-ebb plankton samples were taken hourly from 19:00 to 07:00 h from 9 to 10 July (new moon), 23 to 24 July (full moon), and 8 to 9 August (new moon). In addition, plankton samples were taken for 10 consecu- tive hours daily on 8 August to verify the response of postlarvae to light. The nonparametric Kruskal-Wallis test and analysis of variance (ANOVA) were used to determine differences in concentration of postlarvae between dark-ebb and dark-flood periods. To evaluate the possible effect of environmental vari- ables on larval supply to Florida Bay and the pattern of seasonality in postlarval concentrations, available time series of winds and sea surface temperature (SST) on the coastal shelf were examined in relation to our time series of monthly measured larval concentrations. In particular we were looking for a pattern that might help to determine the reason for the marked summer peak in the concentration of postlarvae at the western border of Florida Bay. Time series of hourly winds and sea surface temperature (SST) for the 3-year sampling period were obtained from the Coastal Marine Auto- mated Network (CMAN) station at Long Key, and from the Coastal Ocean Monitoring and Prediction System (COMP) station in NW Florida Bay (Fig. 1). Tempera- ture and wind time series from the Long Key CMAN station and the NW Florida Bay COMP station were highly correlated with each other (7-2 = 0.9; P<0.01). The longer Long Key time series were used for coastal SST and wind analysis. Wind speed and direction over the Keys, as measured at CMAN sites, are highly coherent (Peng et al., 1999) and useful for explaining currents on the SW shelf (Lee and Williams 1999). Monthly av- Table 1 Mean cross section depth, peak tidal flow, and cross sec- tion area of the four channels sampled for pink shrimp iFarfantepenaeus duorarum) postlarvae at the north- western border of Florida Bay, Middle Ground (MG) and Sandy Key (SKl, and at the southeastern edge in the Middle Florida Keys, Whale Harbor (WH), and Panhan- dle Key (PH). Area was calculated at zero (m) of mean sea level. Station MG SK WH PH epth Peak tidal flow C ross area (m) (m^/sec) (m-) 3.0 1420 2723 3.1 570 1345 3.3 280 407 2.0 30 74 erages of wind vectors and SST were calculated from hourly CMAN time series data (years 2000 to 2002) to examine the effect of winds and SST on the monthly postlarval collections. Time series of current data from two established sta- tions with moored ADCPs and temperature and salinity sensors were used to drive our transport model. Ini- tially, these data were examined to determine whether prevailing currents alone could explain the transport of larvae from the Tortugas spawning grounds to Florida Bay nursery grounds. These stations also were a source of salinity data used in one set of simulations. These stations were located on the inner SW shelf of the Gulf of Mexico, about 30 km from our MG station (Lee et al., 2001) (Fig. 1). This array monitored coastal currents as part of the Florida Bay Circulation and Exchange Study (Lee et al., 2001). The ADCP moorings were located at depths of 6.4 m (mooring A=onshore) and 11.6 m (moor- ing B=offshore) and recorded data every 30 minutes for a 3-year period (A=21 September 1997 to 15 October 2000; B=22 September 1997 to 17 October 2000). The two ADCP moorirr^s were about 30 km apart. Lee et al. (2001) reported insignificant differences between currents in the vertical for cross-shelf transport in the shallow SW Florida shelf. Wind and current vectors were resolved into cross-shelf (u=east [+] and west [-]) and alongshore ((;=north [-I-] and south [-] constituents). The east-west and north-south displacement of current and wind constituents (half-hour current data and the hourly wind data) was the product of each current and wind constituent by the respective time interval. Cor- relation analysis was conducted on currents and wind time series. A harmonic analysis was conducted on the three-year ADCP raw data to define tidal constituents and current magnitude. Period (Pi) and tidal excursion (Ti) were calculated for each constituent from amplitudes (Ai) and frequencies of the constituents as follows: Ti = AiPi/n. 64 Fishery Bulletin 104(1) A simple Lagrangian trajectory model was devel- oped to estimate the drift of planktonic stages. The model used the observed currents at ADCP moorings A and B to calculate trajectories. Because of the lack of information on spatial current variations and the difficulty of extrapolating vertical current profiles from one depth and bottom relief to other conditions, a simple two-dimensional (horizontal) simulation model was used. For the computations we selected the high- est bin from the ADCP that had good data throughout the tidal cycle at each station. This bin typically was 1 m below the mean water level at that location and was the highest bin not affected by instrument sidelobe interference. Because instantaneous bottom currents were closely aligned in direction with surface currents and because magnitudes were only 25% lower, the tidal currents were barotropic and we used, therefore, the surface currents to estimate the largest possible distance traveled. The larval transport was calculated with the equation dxjdt = w,, where .r, = position vector of the larvae; t = time; and «, = the local current velocity. An Euler (forward in time) integration rule was used to numerically solve this equation. Because only two ADCPs (A and B) provided the current data for this large region, no attempt was made to extrapolate a current field from them. Simulations were run by using current meters A and B independently and by assum- ing that the current field was spatially homogeneous. The trajectories therefore were two-dimensional in the horizontal plane and the result was identical to a progressive vector diagram. Comparison between the two sets of trajectories (A and B) provided an estimate of the possible variability. The model was used to ex- plore the potential transport of planktonic stages under various assumptions of behavior controlled by environ- mental cues: 1) a behavioral response to salinity and light, 2) a diel behavior, 3) a diel and tidal behavior throughout the planktonic phase, and 4) an ontogenic change that began with diel behavior and added a tidal behavior at the 15''^ day. All four hypotheses of larval behavior in relation to transport were simulated in order that their effects on distance traveled could be compared and contrasted. In all simulations, we assumed that larvae traveled only at night. We also assumed that the source of the pink shrimp larvae was located immediately northeast of the Dry Tortugas about 150 km from western Florida Bay (Cummings, 1961; Jones at al., 1970; Roberts, 1986). The program simulated distances traveled by particles for a period of 30 days (e.g., days 1-30, 31-60, 61-90, etc.), a pe- riod that corresponds to the estimated developmental period for pink shrimp from the time of hatching to the postlarval stage when larvae are ready for settlement (Dobkin, 1961; Ewald, 1965). Results Patterns of postlarval supply, SST, and winds The monthly influx of postlarvae through the Middle Florida Keys channels (WH and PH) exhibited a highly variable temporal pattern from year to year. Postlarvae were observed every month through the three years (Fig. 2A). Peaks of postlarvae through the Middle Keys channels occurred in May, July, August, and October 2000; in January, July, and October 2001; and in Janu- ary, March, June, and September 2002. In contrast, the monthly influx of postlarvae through the northwestern stations (SK and MG) showed a strong seasonal pattern with one distinct high peak centered in summer from July through September for each year of the 3-year period (Fig. 2B). The number of postlarvae entering through northwestern Florida Bay was much higher than through the Keys stations. The mean concentration of postlarvae per station over the 3-year period indicated that concentrations of postlarvae entering northwestern Florida Bay through SK and MG channels were about eight times greater than through the Florida Keys chan- nels of WH and PH (Fig. 3). Results from a two-way ANOVA indicated that there was a significant effect of site (northwestern stations vs. Florida Key stations) and month on the supply of postlarvae entering Florida Bay (Table 2). Winds showed a seasonal pattern; the spring and summer were dominated by weak southeasterly winds and the fall and winter, by strong northerly winds (Fig. 2, C-D). The monthly average alongshore showed a weak northward constituent in spring-summer of each year (Fig. 2C). The monthly average cross-shelf winds were consistently negative (toward the west) and showed no seasonality (Fig. 2D). The average tem- perature over the three-year time series was 26.1°C, and winter temperatures in 2000-01 were lower than in 2001-02 (Fig. 2E). In summer, during the period of peak postlarval immigration through the northwestern stations (MG and SK), the alongshore wind constituent was mainly northward, the cross-shelf wind was west- ward, and the SST was above average during the three- year period (Fig. 2, A-E). Postlarval concentrations at the northwestern stations were correlated with SST and alongshore winds (Fig. 2, B-E; Table 3). Postlarval con- centrations at the Florida Keys stations (WH and PH) were not correlated with either winds or SST. Subtidal and tidal currents at the SW Florida shelf Advective displacement derived from two ADCP velocity records indicated that the net current is primarily in the alongshore direction (Fig. 4, A and B). The alongshore flow from onshore mooring A was northward, had a total mean velocity of 0.0062 m/sec, and a total water dis- placement of 589.3 x 10' m over the three years (Fig. 4A). The cross-shelf flow was westward, had a mean veloc- ity of -0.0005 m/sec, and a total water displacement of -289.7 X 10^ m. The alongshore flow from offshore moor- Cnales et al Variability in supply and cross-shelf transport of Farfantepenaeus duorarum postlarvae into Florida Bay 65 Florida Key stations J FMAI^ JJ/JJ AS0NDJFI\/1ANi1J JASONDJ Fti^AIVlJ JASOND 2000 2001 2002 Figure 2 Monthly time series of pink shrimp iFarfantepenaeus duorarum ) postlarval concentrations, winds, and sea-surface temperature: (A) mean concentration of pink shrimp postlarvae at the Florida Key stations of Whale Harbor (WHi and Panhandle (PH); (B) mean concentration of pink shrimp postlarvae at the northwestern Florida Bay stations of Middle Ground (MG) and Sandy Key (SK); concentrations of two nets are depicted as 1 and 2; (C) alongshore winds d'); (D) cross-shelf winds iu); (E) sea-surface temperature (SST). Winds and SST data are from Long Key CMAN station. ing B was southward, had a mean velocity of -0.0038 m/sec and a total water displacement of -364.7x10' m (Fig. 4B). The cross-shelf flow was eastward, had a mean speed of 0.0015 m/sec, and a total water displacement of 145.5 X 10'^ m and a resultant southeastward current. The alongshore winds measured at Long Key were southward and had a total water displacement of 16.1 x 10^ m, and the cross-shelf winds were westward and had a total 66 Fishery Bulletin 104(1) SK MG Northwestern stations WH PH Florida Key stations Figure 3 Mean concentration of pink shrimp (Farfantepenaeus duorarum ) postlarvae over a three-year period at each sampling station: northwestern Florida Bay stations — Middle Ground (MG) and Sandy Key (SK); Florida Key stations— Whale Harbor (WH) and Panhandle (PH). Table 2 Results of a nonparametric two-way ANOVA of the effect of site and month on pink shrimp {Farfantepenaeus duo- rarum) postlarvae entering Florida Bay. P<0.05 is indi- cated with an asterisk. SS = sum of squares; MS=mean of squares. Factor df SS MS Month Site Month X site Error 80 1 1 461 756.7 65.9 1191.5 772.2 9.5 65.9 1191.5 1.7 5.6 0.007* 39.3 0.002* 711.3 0.007* Table 3 Correlation coefficients of pink shrimp {Farfantepenaeus duorarum) postlarval concentrations with sea surface temperature (SST), cross-shelf wind (t/), and alongshore wind (V) at the four sampling stations of MG= Middle Ground, SK= Sandy Key, WH= Whale Harbor and PH = Panhandle. Environmental data are from Long Key CMAN station. Significant correlations (P<0.05 ) are indi- cated with an asterisk. SST U V MG 0.82* -0.025 0.67* SK 0.73* 0.11 0.57* WH 0.15 0.16 0.36 PH -0.17 0.14 -0.09 water displacement of -86.3 x 10^ m (Fig. 4C ). The subtidal currents in this region are very weak as also observed by other investigators (Koczy et al., 1960; Rehrer et al., 1967). Correlation coefficients between winds and currents in the alongshore direction iv) were significant for both onshore and offshore currents time series (Table 4). Correlation coefficients between winds and currents in the cross-shelf direction were higher in the onshore than in the offshore data series. This analysis indicated that prevailing currents, overall, were not favorable to passive transport of larvae from the Tortugas to western Florida Bay. A harmonic analysis of the 3-year ADCP data showed that semidiurnal tidal constituents (M.2, S.,, K.,, and N„, see Table 5 for explanation of abbreviations) are dominant on the SW Florida shelf. M., was the strongest tidal constituent and its east-west constituent explained 95% of the total current variance. The east-west amplitude of the M., constituent was 0.32 m/s for both moor- ings, and the north-south was 0.07 m/s for moor- ing A and 0.04 m/s for mooring B. The east-west amplitude (0.32 m/s) was much stronger than the long-term averaged subtidal cross-shelf constitu- ent at both stations (0.001 m/s). The east-west tidal excursion of the M, constituent was similar for both moorings, but a few meters larger on the offshore station (Table 5). This result indicates that it is reason- able to consider that there are similar tidal excursions on the SW shelf up to 50 km. Ebb versus flood catches Concentration of postlarvae collected hourly over a com- plete dark portion of a tidal cycle showed clearly that dark-ebb catches were negligible (<10%) by comparison with dark-flood catches (>90'7f ) (Fig. 5, A-C). Hourly con- centrations of postlarvae collected on the dark flood were 92.7% from 9 to 10 July, 90.2% from 23 to 24 July, and 86.9% from 8 to 9 August 2002. A nonparametric Krus- kal-Wallis test showed significant differences in catches of postlarvae between dark-ebb and dark-flood peri- ods (Kruskal-Wallis; H=15.5, n = 36, P<0.001; (ANOVA: 7^=18.6; P<0.001). Samples taken during the day on 8 August confirmed the hypothesis that postlarvae are present mostly at night in the water column (Table 6). A total of 18 postlarvae (31.5 postlarvae/1000 m^) were captured during 10 consecutive daylight hours, against 2657 (3702.5 postlarvae/1000 m^) captured during 10 consecutive hours of darkness. Although concentrations of postlarvae tend to increase with the tidal current, these differences were not significant (Kruskal-Wallis: //=10.3, /! = 36, P=0.5). From 9 to 10 July, no postlar- vae were captured during the hours of highest current speed (2:00 to 3:00 h), which coincided with high rain and winds (Fig. 5A). It is not clear whether these fac- tors caused a change of behavior in the postlarvae or a malfunction of the net. From 23 to 24 July, concentra- tion of postlarvae on the dark flood followed the cur- Cnales et al Variability in supply and cross-shelf transport of Farfontepenaeus duoroium postlarvae into Florida Bay 67 rent speed; highest catches occurred at the maximum speed and lowest catches at the minimum speed (Fig. 5B). From 8 to 9 August, the highest peak of postlarvae occurred at the end of the dark-flood period when the current had decreased in speed (Fig. SCI. Transport simulations Results of channel net sampling showed that the greatest postlarval influx occurred at the northwestern border of the bay, where there was a strong seasonal pattern with maxima from July through September each year. Postlarval concentrations at the northwestern stations were cor- related with the alongshore winds and with surface temperature but did not correlate with the cross-shelf winds. However, postlarvae need to travel up to 150 km across the shelf to reach their nursery grounds. Cross-shelf transport mechanisms were explored by using a Lagrang- ian trajectory model that simulated the maximum distance traveled by planktonic stages moving at night for a 30-day period. Four scenarios of transport were modeled under dif- ferent assumptions of behavior: 600x10-" 400x10' 200x103 . A / N y Mooring A •''' __^ *-i~'--^' " W ^ -100x10' -200x10-' -300x10' -400x10' • -500x10' 20x10' Mooring B "-V r -20x10' ■ -40x10' ■ -60x10' ■ -80x10^ -100x10' Long Key CMAN Scenario 1 With the assumption that planktonic stages (larvae and postlarvae) of pink shrimp respond to light and salinity changes (Hughes 1969a, 1969b). the first simulation postulates that larvae and postlar- vae move horizontally with a vertical migratory behavior cued by changes in salinity. Larvae and postlarvae swim to the surface at night when an increase in salinity is detected and remain near the bottom when the salinity decreases. Onshore mooring The maximum distance traveled in the cross-shelf direction was 98 km eastward and 70% of larvae traveled up to 30 km (Fig. 6A). The average eastward distance in all simulations was 22 km. The maximum displacement occurred in fall-winter. Along- shore distances were as much as 25 km northward (70% of larvae) and 15 km southward (30%) (Fig. 6, A and B). Offshore mooring The maximum distances traveled in the cross-shelf direction was 100 km (84%^ of larvae traveled 40 km). Alongshore distances were as much as 15 km northward (70%) and 20 km southward (30%) (Fig. 6, C and D). 1011121 23456789 101 1121 23456789 101 1121 2345678910 1997 1998 1999 2000 Figure 4 Time series of current and wind displacement. October 1997-October 2000. Current data are from two ADCPs located at the SW Florida shelf, and wind data are from Long Key CMAN station; lA) onshore ADCP mooring A, (B) offshore ADCP mooring B; see ADCP locations in Figure 1. (C) Wind displace- ment. Alongshore constituents are represented by interrupted lines (N=north l-^], S=south [-]) and cross-shelf by continuous lines (E = east [+], W=west [-]). Scenario 2 The second simulation assumes that plank- tonic stages travel at night using the instantaneous current. Onshore mooring Distances traveled in the cross- shelf direction reached a maximum of 68 km and 75% of larvae traveled only 30 km eastward (Fig 6A). The average eastward distance in all simulations was 15 km. The maximum displacement occurred in spring-summer (May to September 1998) during the first year and in fall-winter (October 1999 to February 2000) during the second and third year. Distances recorded in the along- shore direction were as far as 40 km northward (48% of larvae) and 30 km southward (52%) (Fig. 6, A and B). 68 Fishery Bulletin 104(1) Offshore mooring Maximum distance traveled in the cross-shelf direction reached 60 km eastward and 80% of larvae traveled less than 30 km. Distances re- corded in the alongshore direction were as far as 40 km northward (51%) and 40 km southward (49%) (Fig. 6, C and D). Scenario 3 Under the assumption that pink shrimp larvae and postlarvae migrate vertically in a tidal cycle, the third simulation postulates that larvae and post- larvae swim in the water column near the surface at night during the flood tide and remain near the bottom during the ebb tide. In this simulation it is assumed that planktonic stages move by using the eastward cur- Table 4 Correlation coefficients of wind and current components. U and V are the east-west and north-south components respectively. Nearsurface currents are from onshore (A) and offshore (B) ADCP moorings over the SW Florida shelf Wind data are from Long Key CMAN station. Signif- icant correlations (P<0.05l are indicated with an asterisk. Mooring A Mooring B onshore offshore [/-current V-current f7-current V-current {/-wind V-wind -0.25* 0.26* -0.22* 0.55* -0.10 0.12 -0.06 0.60* rent (flood tide) during the postulated 30 days of larval development. Onshore mooring The maximum distance traveled in the cross-shelf direction was 200 km eastward and 86% of the larvae exceeded 150 km. The average eastward distance in all simulations was 132 km. The maxi- mum larval displacement occurred from December 1999 through March 2000. Distance traveled in the along- shore direction was 45 km northward (70%) and 5 km southward (30%) (Fig. 6, A and B). Offshore mooring The maximum larval displace- ment in the cross-shelf direction was 200 km eastward and 85% of the larvae reached 150 km. The maximum eastward displacement occurred in fall and in winter. Distance traveled in the alongshore direction was as much as 40 km northward (82%) and 5 km southward (18%). The maximum distance traveled was recorded in March-April and June 2000 (Fig. 6, C and D). Scenario 4 In this simulation, it is assumed that there is a change in behavior for pink shrimp larvae — an assumption similar to the one taken in simulations for some Australian penaeid species (Rothlisberg, 1982; Rothlisberg et al., 1995, 1996). Early larval stages (pro- tozoea and myses) migrate vertically in a diel cycle and there is no cross-shelf displacement during the first 15 days of development. Later in development, postlarvae migrate by using tidally induced behavior superim- posed on the diel behavior for the remaining 15 days of planktonic development, and the eastward current (flood tide). Onshore mooring The maximum displacement of larvae in the cross-shelf direction was 100 km eastward Table 5 Harmonic analysis results conducted on three years of ADCP data September 1997 -October 2000). Moorings were located at the SW Florida shelf: A (onshore) and B (offshore) moorings. U and V are the east-west and north-south tidal consistuents. respectively. Explained variance of constituent U was 95<:{ and of V was 50% (for A), and ge-Zr and 29% (for B), respectively. M2 = semidiurnal lunar: S2= semidiurnal solar; N2 =semidiurna larger lunar elliptic; Kl-diurnal lunisolar Ml = diurnal smaller | lunar elliptic; 01 = diurnal principal lunar ,K2= semidiurnal lunisolar. Tidal Period Tidal constituent U Tidal constituent V Amplitude Tidal excursion Amplitude Tidal excursion constituent (h) (m/sec) (m) (m/sec) (ml A M2 12.42 0.321 4561.7 0.070 993.5 N2 12.66 0.056 810.8 0.013 191.5 S2 12.00 0.100 1368.2 0.017 239.3 Kl 23.93 0.049 1332.9 0.015 408.7 Ml 24.84 0.005 139.5 0.001 31.3 01 25.82 0.039 1139.1 0.010 289.9 K2 11.97 0.025 345.6 0.010 135.8 B M2 12.42 0.323 4593.0 0.038 539.4 N2 12.66 0.057 828.3 0.007 105.9 S2 12.00 0.105 1439.7 0.009 129.3 Kl 23.93 0.052 1434.4 0.008 213.9 Ml 24.84 0.003 93.9 0.002 54.1 01 25.82 0.043 1278.1 0.002 68.0 K2 11.97 0.023 319.5 0.005 63.1 Cnales et al.: Variability m supply and cross-shelf transport of Foi fantepenaeus duorarum postlarvae into Florida Bay 69 -40 and 86% of larvae reached 80 km. The aver- age eastward distance traveled in Simula- 60 tions was 66 km. The maximum distances traveled in the alongshore direction was 25 40 ■ km northward (70% of larvae) and 10 km southward (30%). The maximum distance 20 traveled was recorded in March-April and June 2000 (Fig. 6, A and B). Offshore mooring The maximum displace- ment of larvae in the cross-shelf direction -20 was 90 km eastward and 85% of larvae reached 80 km. Distance traveled in the alongshore direction reached 20 km north- ward (86% of larvae) and 5 km southward (14%) (Fig. 6, C and D). Discussion The monthly influx of pink shrimp postlar- vae entering Florida Bay through its north- western border showed a strong seasonal pattern of annual high peaks in summer over the 3-year period. Postlarval concentra- tions were correlated with alongshore winds and sea surface temperature. Alongshore winds were seasonal, with a weak northward constituent in spring-summer of each year changing to a strong southward in fall and winter. This seasonal pattern agrees with the general circulation described for the SW Florida shelf, highly dependent on synoptic- scale winds, coupled with a strong seasonal- ity and strong tidal currents (e.g., Weisberg et al.. 1996; Wang, 1998; Smith, 2000; Lee et al., 2001). Although tidal currents seem to be the main vehicle of eastward transport for planktonic stages, alongshore winds may be fundamental for moving larvae northward along the SW Florida shelf by avoiding a drift with the Florida Current or with the cyclonic circulation of the gyres that form southwest of the Dry Tortugas (Lee et al., 1994; Fratantoni et al., 1998). Under these circumstances larvae may reach Florida Bay by the shortest route in summer using tidal currents and winds across the shallow SW Florida shelf and entering Florida Bay by its northwestern border. The summer sea- sonality of postlarval immigration may also be amplified by the seasonality of spawning because higher temperatures induce higher spawning activity (Cripe, 1994) and conse- quently more recruits to enter the Bay during favorable onshore conditions. Another alternative explanation for the summer lar- val immigration is that larvae may take advantage of a distinct annual tidal cycle produced in summer every year as a result of the interaction of the periods of the diel vertical migration and the tidal constituent (S.,, A Flood i 0 ° Storm ^ ll ^ ^ • new Ebb ^^3/ ^ 0 0 0 0 July 9 -10, 2002 - 1 - -1 19h 20h 21ti 22ti 23h 24ti n Figure 5 Hourly concentration of pink shrimp (Farfantepenaeus duorarum) postlarvae at Sandy Key station (SK) during a complete dark (night) tidal cycle conducted during (A) new moon, 9-10 July 2002, (B) full moon, 23-24 July 2002, and (C) new moon, 8-9 August 2002. Right y-axis indicates concentration of postlarvae/m''; lefty-axis is tidal current speed (cm/sec) measured by acoustic Doppler velocity meter (filled circle and solid line) and by flowmeter (empty circles and interrupted line). Positive values in they-axes indicate transport into Florida Bay (flood), and negative values transport out of the Bay (ebb). Horizontal bars on the bottom indicate hours of darkness versus light. K,) with the annual cycle of the length of the night (Criales et al., 2005). Larvae moving vertically in the water column with a diel behavior can be transported up to 70 km onshore in summer because the eastward current of the tidal constituents matches the diel cycle over extended intervals in the shorter summer nights 70 Fishery Bulletin 104(1) Table 6 Data from ebb-flood experiment conducted at D = dark. VWF = vohime of water filtered. Sandy Key (SK) station during 20 consecutive hours. E = ebb, F=flood, L = light. Date Collection time Tide stage Light vs. dark VWF (m3) Number of postlarvae Concentrations (postlarval m'l 08/08/2002 11:00 E L 395.8 7 17.7 08/08/2002 12:00 E L 535.3 2 3.7 08/08/2002 13:00 F L 1360.1 0 0.0 08/08/2002 14:00 F L 1262.3 0 0.0 08/08/2002 15:00 F L 971.3 8 8.2 08/08/2002 16:00 F L 531.2 1 1.9 08/08/2002 17:00 F L 0.0 0 0.0 08/08/2002 18:00 F L 0.0 0 0.0 08/08/2002 19:00 E L 450.6 0 0.0 08/08/2002 20:00 E L 575.6 0 0.0 08/08/2002 21:00 E D 530.9 5 9.4 08/08/2002 22:00 E D 477.6 14 29.3 08/08/2002 23:00 E D 60.7 7 115.4 08/09/2002 0:00 E D 263.3 87 330.4 08/09/2002 1:00 F D 648.6 141 217.4 08/09/2002 2:00 F D 944.3 247 261.6 08/09/2002 3:00 F D 1164.1 491 421.8 08/09/2002 4:00 F D 977.5 529 541.2 08/09/2002 5:00 F D 737.8 1006 1363.6 08/09/2002 6:00 F D 315.1 130 412.5 (Criales et al., 2005). Therefore, pink shrimp and other fish and invertebrate species that use tidal currents for transport with a daily vertical behavior may take advantage of this annual tidal cycle to improve their chances of reaching coastal nursery habitats. The monthly influx of postlarvae through the Mid- dle Florida Keys channels exhibited a highly variable seasonal pattern from year to year and from station to station. This section of the Keys coastal waters is frequented by coastal cyclonic eddies that originate at the Dry Tortugas and move downstream at 5-17 km/day along the edge of the shelf at intervals of 1 to 3 months (Fratantoni et al., 1998; Yeung et al., 2001). For planktonic stages, these eddies may serve as a delivery mechanism from offshore spawning grounds to the southeastern border of Florida Bay, allowing an onshore transport by the coastal countercurrent flow generated by the cyclonic circulation (Lee et al., 2001; Yeung et al., 2001; Criales et al., 2003). Episodic, me- soscale events associated with boundary current fronts and eddies may cause high variability in transport (Lee et al., 1994; Limouzy-Paris et al., 1997; Yeung and Lee, 2002). This variability was reflected in the influx of pink shrimp postlarvae through Middle Keys channels (Criales et al., 2003). The influx of postlarvae at WH and PH channels in the present study was also highly variable and there was no correlation with winds or sea surface temperature. Therefore, we hypothesized that the high variability in postlarval influx detected at the Florida Keys channels reflects the temporal and spatial variability associated with the passage of coastal eddies. Simulations of transport based on current observa- tions indicated that passive larvae could not consistent- ly be advected the estimated 150 km eastward across the shelf between spawning and nursery grounds in 30 days. This is true primarily because of the weak eastern current and the reversing nature of tides. In contrast, planktonic stages (larvae and postlarvae) mov- ing at night with the eastward current (flood tide) can consistently travel 100 to 200 km in 30 days. Hypotheti- cally, 85% of the larvae can be transported far enough from the known spawning grounds of Dry Tortugas to the nursery grounds in western Florida Bay in 30 days. Previous works conducted inside Florida Bay (Tabb et al., 1962; Roessler and Rehrer, 1971) and our own data obtained along the western border of Florida Bay dem- onstrated the ability of pink shrimp postlarvae to re- spond to the dark flood tide and to distinguish between day and night. Over 90% of postlarvae were caught in the dark flood period and only a few postlarvae were caught during daylight hours. This behavior needs to be investigated for early larval stages to define the exact age at which larvae begin reacting to change in tides and to environmental cues that trigger the vertical movement in relation to tidal stage. For other penaeid Cnales et al Variability in supply and cross shelf transport of Farfontepenaeus duorarum postlarvae into Florida Bay A Onshore ADCP A, Cross-shelf (+East. -West) ■ ■♦QCioooS -•♦••OOOOOOO- -r-T — I — ^ — r— 1 — t I 1 I I I I I I I I — I I I I I — n — rn — i i i — i — i i i i i — i i i 9 10 11 i; 1 : 3 J 5 6 7 8 9 10 11 12 I 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 C Offshore ADCP B, Cross-shelf (+East, -West) ▼ T * »▼▼▼- ^^0^0 TTT ▼ AA ▼ T ^aaa^SOa ••• « gd^^'^ "^^aaA .-88ooooflfil 2ooooa* '.^^^^m^ oo- I I I I I I I I I I I I I I I I I I I 1 I I — n — n — I I I I I I I I I I I I 9 10 11 12 1 234 56789 10 11 12 t 234 56789 10 11 12 1 234 56789 10 1997 1998 1999 2000 1997 1998 1999 2000 B Onshore ADCP A, Alongshore (-i-North, -South) ▼oQ ' SjjJi '^ «o%^ 8 ^5 'a o ooo • •• • 8 •^ 8 .• • I I I I I I I I I I I I 1 I I I I I I I I I I I I I I I I I I I I 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 D Offshore ADCP B, Alongshore (+North, -South) T^ T^ O ?▼» _ -TAA^A, ■ 8. o o o 1997 1998 1999 2000 9 10 11 12 1 234 56789 10 11 12 1 234 56789 10 11 12 1 234 56789 10 1997 1998 1999 2000 • scenario 1 o scenario 2 ▼ scenario 3 | A scenario 4 | Figure 6 Outputs of transport simulations of pink shrimp iFarfantepenaeus duorarum) planktonic stages (larvae and postlarvae) developed from in situ current observations and under different scenarios of larval behavior. Each point represents 30 days of travel from September 1997 to October 2000; (A and Bi outputs from onshore ADCP A currents, cross-shelf and alongshore constituents, respectively; (C and D) outputs from offshore ADCP B currents, cross-shelf and alongshore constituents, respectively. Scenario 1 (filled circles) = planktonic stages migrating with the instantaneous current and a diel behavior (at night) cued by changes in salinity; scenario 2 (empty circles) = planktonic stages migrating with the instantaneous current and a diel behavior (at night); scenario 3 (filled triangles) = planktonic stages migrating with a diel (at night) and a tidal (eastward current) behavior; scenario 4 (empty triangles) = planktonic stages migrating by an induced ontogenic change of behavior (a diel behavior during the first 15 days and a tidally induced behavior (eastward current) superimposed on the diel behavior (at night) for the remaining 15 days). species (Fenneropenaeus merguiensis, Penaeus plebejus. and Penaeus semisulcatus) in the Gulf of Carpentaria, Australia, an ontogenic change in behavior has been (iocumented for planktonic stages (Rothlisberg, 1982; Rothlisberg et al., 1995. 1996; Condie et al., 1999). Results of these studies showed that during the first two weeks of planktonic development, larvae migrate vertically in a diel cycle; later in development, the verti- cal migration is tidally induced. Under those conditions there was little or no systematic cross-shelf displace- ment of larvae during the first two weeks; afterward, cross-shelf displacement occurred rapidly (Rothlisberg et al., 1995; 1996). Simulation of transport for pink shrimp larvae using this ontogenic change of behavior indicated that planktonic stages can be transported up to 100 km eastward and 85% of larvae reach 80 km. If this behavior applies to pink shrimp, the location of the spawning grounds needs to be reconsidered in favor of areas closer to Florida Bay. Alternatively, the behavior of early planktonic stages may have been underesti- mated. Results of this research indicated once again the extreme importance of defining larval behavior and including behavior in dispersal models. The cue(s) that penaeid postlarvae use to migrate in the water column during the flood tide and to return to the bottom on the ebb tide are not completely un- derstood and could be species specific. Hughes (1969a, 1972) demonstrated in laboratory experiments that 72 Fishery Bulletin 104(1) pink shrimp postlarvae react to changes in salinity by changing swimming direction. Postlarvae were more active in the water column with increases in salinity, and this finding implies a shoreward displacement with the flood tide. Similar responses to salinity changes have been found for postlarvae oi Farfantepenaeus cali- forniensis, Farfantepenaeus brevirostris, Litopenaeus stylirostris, and Litopenaeus vannamei from the Mexi- can Pacific (Mair, 1980). The importance of a salinity cue for transport has been questioned for other penaeid species that inhabit hypersaline estuaries (southeast African, western Australia) in which penaeid postlar- vae would need to move against a salinity gradient (Penn, 1975; Forbes and Benfield, 1986; Rothlisberg et al., 1995). Our simulations of transport guided by salinity changes indicated that planktonic stages could travel distances in the range of only 30 km in 30 days. This result may indicate that salinity is not the only environmental factor controlling long cross-shelf migra- tions of pink shrimp. However, Hughes (1969a, 1969b) in early experiments suggested that a salinity cue could apply to postlarvae near the nursery grounds. Changes in water pressure have been proposed as the only environmental factor that triggers the vertical migration of postlarval shrimps in the Gulf of Car- pentaria, Australia (Penn, 1975; Forbes and Benfield, 1986; Rothlisberg et al., 1995). Laboratory experiments and numerical models have shown that tiger shrimps iPenaeus semisulcatus and Penaeus esculentus) larvae switch behavior when the change in water pressure with tides becomes a significant fraction of the total pressure (Rothlisberg et al., 1996; Condie et al., 1999; Vance and Pendrey''^). This behavior only occurred in larvae above a certain size. However, it still remains to be determined whether, in a natural ecological context, the rates of relative changes of pressure are consistent with the tidal cycle periodicity and are detectable at absolute amounts in order to permit a behavioral re- sponse. By means of simulations of transport, we have identi- fied a potential STST mechanism for planktonic pink shrimp to migrate the estimated 150 km in 30 days from spawning to nursery grounds over the Florida shelf Organisms inhabiting coastal ecosystems domi- nated by tides have the potential to control their cross- shelf movement through STST (Shanks, 1995). The extent of the transport depends on the speed of the tidal current and the time that organisms spend in the water column. Success in reaching the nursery grounds depends upon the stage in larval or postlarval develop- ^ Vance D. J., and R. C. Pendrey. 2001. Vertical migration behaviour of postlarval penaeid prawns: a laboratory study of the effect of tide, water depth and day/night. In The definition of effective spawning stocks of commercial tiger prawns in the Northern prawn fishery and king prawns in the eastern king prawn fishery-behaviour of postlarval prawns, p. 28-52. Fisheries Research and Development Corporation (FRDC ) Final Report ( Project 97/108 ). 68 p. CSIRO Marine Research Laboratories, P.O. Box 120, Cleveland, Qld. 4163, Australia. ment when the tidal behavior is added. A dependence on tidal currents for the entire larval transport period was postulated for Melicertus latisultacus in Western Australia (Penn, 1975). The shrimp Lucifer faxoni is the only species for which a STST mechanism across the shelf has been shown (Woodmansee, 1966). Strong tidal currents and several coastal sources of fresh wa- ter define the spawning grounds of the wide and shal- low SW Florida shelf (Lee et al., 2001; Jurado, 2003). Under these conditions, parts of the SW Florida shelf may behave as an estuary in which planktonic pink shrimp may easily recognize tides by means of endog- enous behavior or environmental variables. From this study we determined that the greatest influx of postlarval pink shrimp occurred at the north- western border of Florida Bay in summer. Postlarvae entering Florida Bay through the channels of the Mid- dle Florida Keys occurred at a much lower magnitude and there were only sporadic peaks and no appar- ent seasonality in influx. The transport mechanism of planktonic stages of pink shrimp across the SW Florida shelf seems to depend heavily on semidiurnal tides and larval behavior, and much less on seasonal winds. The response of postlarvae to the tidal currents was clearly observed at the western margin of Florida Bay. Such behavior needs to be explored in early stages to define the age at which larvae begin to respond to tides, the location on the Florida shelf at which this response occurs, and the specific environmental cues linked to such behavior. With this information, more realistic simulations of transport can be made with a complete hydrodynamic model that would incorporate spatial and vertical variations in currents. Depend- ing on the resulting transport, the location of spawn- ing grounds may be better defined, leading to better protection of this valuable fishery resource. Informa- tion on recruitment variability and key environmen- tal factors affecting larval transport are essential to accurately interpret stock assessments, maintain the ecological integrity of both spawning grounds and nursery grounds, and to effectively manage the pink shrimp fishery. Acknowledgments We are especially grateful to Thomas Lee and Elizabeth Williams (RSMAS, University of Miami), Elizabeth Johns, Ryan Smith, Shailer Cummings, and Nelson Melo (NOAA/AOML, Miami), and Ned Smith (Harbor Branch Oceanographic Institute) for providing ADCP data and valuable comments; to William Richards (NMFSC/ NOAA Miami) and Robert Cowen (RSMAS, University of Miami) for their constructive comments and support; to Hernando Cardenas (NMFSC) for his assistance sorting plankton samples; to Andre Daniels (USGS, Miami), for his valuable support of fieldwork; and to the personnel involved in collecting and managing data from C-MAN and COMP stations. This research was funded by the NOAA South Florida Ecosystem Resto- Criales et al Variability in supply and cross shelf transport of Forfantepenaeus duoraium postlarvae into Florida Bay 73 ration Prediction and Modelling (SFERPM) program through a cooperative agreement between Southeast Fisheries Science Center, USGS and CIMAS, RSMAS. University of Miami, and the DOI Critical Ecosystems Studies Initiative of the Everglades Restoration Pro- gram. The sampling that supported this research was conducted pursuant to National Park Service Permit no. EVER-2003-SCI-0109 and Florida Fish and Wildlife Conservation Commission Special Activity License no. 03SR-502. Literature cited Allen, D. M., J. H. Hudson, and T. J. Costello. 1980. Postlarval shrimp iPenaeus) in the Florida Keys: species, size, and seasonal abundance. Bull. Mar. Sci. 30:21-33. Anderson, M. J. 2001. A new method for non-parametric multivariate analysis of variance. Aust. Ecol. 26: 32-46. Browder, J. A., Z. Zein-Eldin, M. M. Criales, M. B. Robblee, S. Wong, T. L. Jackson, and D. Johnson. 2002. Dynamics of pink shrimp (Forfantepenaeus duor- ariim) recruitment potential in relation to salinity and temperature in Florida Bay. Estuaries 25:1355-71. Condie S. A., N. R. Loneragan, and D. J. Die. 1999. Modelling the recruitment of tiger prawns Penaeus esculentus and P. semisulcatus to nursery grounds in the Gulf of Carpentaria, northern Australia: implications for assessing stock-recruitment relationships. Mar. Ecol. Prog. Ser. 178:55-68. Cowen, R. B. 2002. Larval dispersal and retention and consequences for population connectivity. In Coral reef fishes: dynamics and diversity in a complex ecosystem (P. Sale, ed.), p. 149-170. Academic Press, London. Criales, M. M., M. J. Bello, and C. Yeung. 2000. Diversity and recruitment of penaeoid shrimps at Bear Cut, Biscayne Bay, Florida. Bull. Mar. Sci. 67:773-788. Criales, M. M., and T. N. Lee. 1995. Larval distribution and transport of penaeoid shrimps during the presence of the Tortugas Gyre in May-Junel991. Fish. Bull. 93:471-482. Criales, M. M., and M. F. McGowan. 1994. Horizontal and vertical distribution of penaeidean and caridean larvae and micronektonic shrimps in the Florida Keys. Bull. Mar. Sci. 54:843-856. Criales, M. M., J. Wang, J. A. Browder, and M. B. Robblee. 2005. Tidal and seasonal effect of transport of pink shrimp postlarvae. Mar. Ecol. Prog. Ser. 286:231-238. Criales, M. M., C. Yeung. D. Jones. T. L. Jackson, and W. J. Richards. 2003. Variation of oceanographic processes affecting the size of pink shrimp (Farfantepenaeus duorarum i postlarvae and their supply to Florida Bay. Estuar. Coast. Shelf Sci. 57 (3): 457-468. Cripe, G. M. 1994. Induction of maturation and spawning of pink shrimp, Penaeus duorarum, by changing water tempera- ture, and survival and growth of young. Aquaculture 128:255-260. Cummings, D. C. 1961. Maturation and spawning of pink shrimp, Penaeus duorarum Burkenroad. Trans. Am. Fish. Soc. 90: 462-468. Dall, W., B. J. Hill, P. C. Rothiisberg, and D. J. Staples. 1990. The biology of Penaeidae. Adv. Mar. Biol. 27, 489 p. Dobkin, S. 1961. Early development stages of pink shrimp, Penaeus duorarum. from Florida waters. Fish. Bull. 61: 321-349. Eldred, J. W., G. T. Martin, and E. A. Joyce Jr. 1965. Seasonal distribution of penaeid larvae and postlar- vae of the Tampa Bay area, Florida. Fl. Board Conserv. Mar. Res. Lab., Tech. Ser. 44, 47 p. Ewald, J. J. 1965. The laboratory rearing of pink shrimp, Penaeus duorarum Burkenroad. Bull. Mar. Sci. 15(2):436- 449. Forbes, A. T.. and M. C. Benfield. 1986. Tidal behaviour and post-larval penaeid prawns (Crustacea: Decapoda: Penaeidae) in a southeast African estuary. J. Exp. Mar. Biol. Ecol. 102:23-34. Forward, R. B., and R. A. Tankersley. 2001. Selective tidal-stream transport of marine ani- mals. Oceanog. Mar. Biol. 39:305-353. Fratantoni, P S., T. N., Lee, G. P. Podesta, and F Muller-Karger. 1998. The influence of Loop Current perturbations on the formation and evolution of Tortugas eddies in the southern Straits of Florida. J. Geophys. Res. 103:24759-79. Garcia, S., and L. Le Reste. 1981. Life cycles, dynamics, exploitation and management of coastal penaeid shrimp stocks. FAO Tech. Paper 203, 215 p. FAO, Rome. Hittle, C, E. Patino, and M. Zucker. 2001. Freshwater flow from estuarine creeks into north- eastern Florida Bay. USGS Water-Resources Invest. Rep. 01-4164, 32 p. Hughes, D. A. 1969a. Responses to salinity change as a tidal trans- port mechanism of pink shrimp, Penaeus duorarum Burkenroad. Biol. Bull. 136:45-53. 1969b. Evidence for the endogenous control of swim- ming in pink shrimp, Penaeus duorarum. Biol. Bull. 136:398-404. 1972. On the endogenous control of tide-associated displacements of pink shrimp, Penaeus duorarum Burkenroad. Biol. Bull. 142:271-280. Iversen, F. S., A. E. Jones, and C. P. Idyll. 1960. Size distribution of pink shrimp, Penaeus duor- arum, and fleet concentrations on the Tortugas fishery grounds. U.S. Fish. Wild. Ser. Spec. Sci. Rep. 356: 1-62. Jones, A. C., D. E. Dimitriou, J. J. Ewald, and J. H. Tweedy. 1970. Distribution of early developmental stages of pink shrimp, Penaeus duorarum. in Florida water. Bull Mar. Sci. 20:634-661. Jurado, J. L. 2003. The dynamics of diatom blooms and silicon cycling in coastal waters of the southwestern Florida shelf and northwestern Florida Bay. Ph.D. diss, 268 p. Univ. Miami, Miami, FL. Klima E. F., G. A. Matthews, and F. J. Patella. 1986. Synopsis of the Tortugas pink shrimp fishery, 1960- 1983, and the impact of the Tortugas Sanctuary. North Am. J. Fish. Manag. 6:301-310. 74 Fishery Bulletin 104(1) Koczy, F. F., M. O.Rinkel, and S. J. Niskin. 1960. The current patterns on the Tortugas shrimp grounds. Proc. Gulf Carib. Fish. Inst. 12"' Annual Sess., Nov. 1959:112-125. Lee, T. N. M. E. Clarke. E. Williams, A. F. Szmant, and T. Berger. 1994. Evolution of the Tortugas Gyre and its influence on recruitment in the Florida Keys. Bull. Mar. Sci. 54:621-646. Lee, T. N, E. Johns, D. Wilson, E. Williams, and N. P. Smith. 2001. Transport processes linking south Florida coastal ecosystems. In Linkages between ecosystems in the South Florida hydroscape (J. Porter and K. Porter, eds.), p 309-342. CRC Press, Boca Raton, FL. Lee, T. N., and E. Williams. 1999. Mean distribution and seasonal variability of coastal currents and temperature in the Florida Keys with implications for larval recruitment. Bull Mar. Sci. 64:35-56. Limouzy-Paris, C, H. C. Graber, D. L. Jones, A. W. Ropke. and W. J. Richards. 1997. Translocation of larval coral reef fishes via sub- mesoscale spin-off eddies from the Florida Current. Bull. Mar. Sci. 60: 966-83. Mair, J. McD. 1980. Salinity and water-type preferences of four species of postlarval shrimps iPenaeus) from west Mexico. J. Exp. Mar. Biol. Ecol. 45:69-82. Munro, J. L., A. C. Jones, and D. Dimitriou. 1968. Abundance and distribution of the larvae of the pink shrimp iPenaeus duorarum] on the Tortugas Shelf of Florida. August 1962-October 1964. Fish. Bull. 67:165-181. Nance, E. F, and F. J. Patella. 1989. Review of the Tortugas shrimp fishery from May 1987 to January 1989. NOAA Tech. Memo., NMFS- SEFC-238, 11 p. Peng, G., C. N. K. Mooers, and H. Graber. 1999. Coastal winds in South Florida. J. Appl. Meteor. 38:1740-1757. Penn, J. W. 1975. The influence of tidal cycles on the distribu- tion pathway of Penaeus latisulcatus Kishinouye in Shark Bay, Western Australia. J. Mar. Freshw. Res. 26:93-102. Rehrer, R., A. C. Jones, and M. A. Roessler. 1967. Bottom water drift on the Tortugas Grounds. Bull. Mar. Sci. 17:562-575. Roberts, T. W. 1986. Abundance and distribution of pink shrimp in and around the Tortugas Sanctuary, 1981-1983. North Am. J. Fish. Manag. 6:311-327. Roessler M. A., and G. Rehrer. 1971. Relation of catches of postlarval pink shrimp in Everglades National Park, Florida, to the commercial catches on the Tortugas Grounds. Bull. Mar. Sci. 21 (4):790-805. Rothlisberg, P. C. 1982. Vertical migration and its effect on dispersal of penaeid shrimp larvae in the Gulf of Carpentaria, Australia. Fish. Bull. 80:541-554. Rothlisberg, P. C, J. A. Church, and C. Fandry. 1995. A mechanism for near-shore density and estua- rine recruitment of post-larval Penaeus plebejus Hess (Decapoda, Penaeidae). Estuar. Coast. Shelf Sci. 40: 115-138. Rothlisberg, P. C. J. A. Church, and A. M. G. Forbes. 1983. Modelling advection of vertically migrating shrimp larvae. J. Mar. Res. 41:511-538. Rothlisberg, P. C, P. D. Craig, and J. R. Andrewartha. 1996. Modelling penaeid prawn larval advection in Alba- tross Bay, Australia: defining the effective spawning population. Mar. Freshw. Res. 47:157-168. Shanks, A. L. 1995. Mechanisms of cross-shelf dispersal of larval inver- tebrates and fish. In Ecology of marine invertebrate larvae (L. R. McEdward, ed.), p. 323-367. CRC Press, Inc., Boca Raton, FL. Sheridan. P. 1996. Forecasting the fishery for pink shrimp Penaeus duorarum, on the Tortugas Grounds, Florida. Fish. Bull. 94:743-755. Smith, N. P 1998. Tidal and long-term exchanges through channels in the Middle and Upper Florida Keys. Bull. Mar. Sci. 62:199-211. 2000. Transport across the western boundary of Florida Bay. Bull. Mar. Sci. 66:291-304. Sponaugle, S., R. K. Cowen, A Shanks, S. G. Morgan, J. M. Leis, J. Pineda, G. W. Boehlert. M. J. Kingsford, K.C. Linderman, C. Grimes, and J. L. Munro. 2002. Predicting self-recruitment in marine populations biophysical correlations and mechanisms. Bull. Mar. Sci. 70(1) suppl.:341-375. Tabb, D. C, D. L Dubrow, and A. E. Jones. 1962. Studies on the biology of pink shrimp Penaeus duorarum Burkenroad, in Everglades National Park, Florida. Fla. Board Conser. Tech. Ser. 37:1-32. Temple, R., and C. C. Fischer. 1965. Vertical distribution of the planktonic stages of penaeid shrimp. Inst. Mar. Sci. Publ. 10:59-67. Wang, J. D. 1998. Subtidal flow patterns in western Florida Bay. Es- tuarine Coast. Shelf Sci. 46:901-915. Weisberg, R. H., B. D. Black, and J. Yang. 1996. Seasonal modulation of the West Florida continental shelf circulation. J. Geophys. Res. 23:2247-2250. Wenner, E. L., D. M. Knott, C. A. Barans, S. Wilde, J. O. Blanton, and J. Amft. 2005. Key factors influencing transport of white shrimp {Litopenaeus setiferus) postlarvae into the Ossabaw Sound system, Georgia, USA. Fish. Ocean. 14(3)175-194. Woodmansee, R. A. 1966. Daily vertical migration of Lucifer. Planktonic numbers in relation to solar and tidal cycles. Ecology 47:847-850. Yeung, C. D. L. Jones, M. M. Criales, T. L. Jackson, and W. J. Richards. 2001. Influence of coastal eddies and counter-currents on the influx of spiny lobster, Panulirus argus, postlarvae into Florida Bay: influence of eddy transport. Mar. Freshw. Res. 52:1217-32. Yeung. C, and T. L. Lee. 2002 Larval transport and retention of the spiny lobster, Panulirus argus, in the coastal zone of the Florida Keys, USA. Fish. Oceanog. 11:86-309. 75 Abstract — The population biology and status of the painted sweeplips iDiagramma pictum) and spangled emperor (Lethrinus ncbiilosus) in the southern Arabian Gulf were estab- lished by using a combination of size-frequency, biological, and size- at-age data. Transverse sections of sagittal otoliths were characterized by alternating translucent and opaque bands that were validated as annuli. Comparisons of growth characteris- tics showed that there were no sig- nificant differences (P>0.05) between sexes. There were well defined peaks in the reproductive cycle, spawning occurred from April to May for both species, and the mean size at which females attained sexual maturity was 31.8 cm fork length (Lp) for D. pictum and 27.6 cm (Lp) for L. nebulosus. The mean sizes at first capture (21.1 cm Lp for D. pictum and 26.4 cm Lp for L. nebulosus) were smaller than the sizes for both at first sexual maturity and those at which yield per recruit would be maximized. The range of fishing-induced mortality rates for D. pictum (0.37-0.62/yr) was sub- stantially greater than the target (F„p, = 0.07/yr) and limit lF,,^„ = 0.09/ yr) estimates. The range of fishing- induced mortality rates for L. nebu- losus (0.15/yr to 0.57/yr) was also in excess of biological reference points (F„p^ = 0.10/yr and F,,^„ = 0.13/yr). In addition to growth overfishing, the stocks were considered to be recruit- ment overfished because the biomass per recruit was less than 20'^* of the unexploited levels for both species. The results of the study are important to fisheries management authorities in the region because they indicate that both a reduction in fishing effort and mesh-size regulations are required for the demersal trap fishery. Biology and assessment of the painted sweetlips iDiagramma pictum (Thunberg, 1792)) and the spangled emperor (Lethrinus nebulosus (Forsskal, 1775)) in the southern Arabian Gulf Edwin M. Grandcourt Thabit Z. Al Abdessalaam Ahmed T. Al Shamsi Franklin Francis Marine Environment Research Centre Environmental Research and Wildlife Development Agency Corniche Road P O Box 45553 Khalidlya Abu Dhabi, United Arab Emirates E-mail address (for E M Grandcourt) egrandcourtS'erwdagovae Manuscript submitted 29 December 2003 to the Scientific Editors Office. Manuscript approved for publication 11 July 2005 by the Scientific Editor. Fish. Bull. 104:75-88 (2006). The painted sweetlips iDiagramma pictum (Thunberg, 1792)). is a member of the family Haemuli(iae and is widely distributed throughout the Indo-West Pacific, from the Red Sea and East Africa to Japan and New Caledonia (Randall et al., 1997). Adults are found in shallow coastal waters and coral reefs down to a depth of 80 m, and juveniles are often found in weedy areas (Smith and McKay, 1986). The diet of this species consists of benthic invertebrates and fishes (Sommer et al., 1996). It is a rela- tively large tropical species attaining 100 cm fork length and 6 kg in total weight (Torres, 1991); consequently it is exploited throughout its range with a variety of gears, including hand- lines, traps, and nets (Fischer and Bianchi, 1984). Diagramma. pictum has a gonochoristic reproductive mode and spawning occurs annually with one clear seasonal peak (Breder and Rosen, 1966). Fishes of the family Lethrinidae are abundant in the coastal tropical and subtropical Indo-Pacific (Young and Martin, 1982). The spangled em- peror (Lethrinus nebulosus (Forsskal, 1775)), is distributed throughout the Indo-West Pacific from the Red Sea and East Africa to southern Japan and Samoa. It is found in a variety of habitats including coral reefs, sea grass beds and mangroves from near shore to a depth of 75 m (Randall, 1995). Adults are either solitary or are found in small schools, and ju- veniles form large schools in shal- low, sheltered sandy areas. The diet of this species is mainly composed of moUusks, crustaceans, polychaete worms, and echinoderms (Fischer and Bianchi, 1984). As with other representatives of the family Lethrinidae, L. nebulosus is a protogynous hermaphrodite, and sexual transformation from female to male occurs over a wide range of sizes (Young and Martin, 1982; Ebisawa, 1990). Lethrinids are considered to have long spawning seasons, running from spring to at least early fall, with spring and fall peaks. Spawning oc- curs after dark for most species in aggregations along the inner or outer edge of the fringing reef (Johannes, 1981). Lethrinus nebulosus is a large tropical species reaching 80.0 cm total length and 8.4 kg total weight (Ran- dall, 1995) and is exploited through- out its range with a variety of gears (Fischer and Bianchi, 1984). Both species form an important part of fisheries landings in the southern Arabian Gulf, where they are mainly caught with dome-shaped wire traps that have a hexagonal mesh of ap- proximately 3.5 cm in diameter. Traps are either set individually or in strings from traditional wooden dhows, sets are made to a maximum depth of 40 m, and vessels fish an 76 Fishery Bulletin 104(1) i *> \\ W':- -^^: r '__■; "■ V ^ - ,^ Figure 1 Study site (stippled area) showing the location of the Emirate of Abu Dhabi, off the coast of which data were collected for the painted sweetlips iDiagramtna pictum) and spangled emperor iLethrinus nehulosus) from commercial catches. average of 210 traps each. Collection of catch-and-effort data for the fisheries of the Emirate of Abu Dhabi in the United Arab Emirates was initiated in 2001. However, many species, including D. pictum and L. nebulosus, are recorded to the family level, and therefore the use of statistical catch-at-age methods can not be used for conducting assessments at the species level. Landings of haemulids (predominantly D. pictum) and lethrinids (predominantly L. nebulosus), totaled 719 and 2911 metric tons, respectively, in the Emirate of Abu Dhabi during 2003 (Grandcourt et al.M. Despite the limited time scale for which catch and effort data are available, there has been an overall increase in fishing effort and catches since 2001. Many of the fish populations in the Arabian Gulf have been heavily exploited (Samuel et al., 1987), and fishing effort may have already been above optimum levels for most demersal species (Siddeek et al., 1999). The expan- sion of the fishing fleet of the United Arab Emirates and the lack of appropriate data on most stocks underscore the need to assess the fisheries resources of the region. The goal of this study was to evaluate the status of £1. pictum and L. nebulosus and to provide biological refer- Grandcourt, E. M., F. Francis, A. Al Shamsi, K. Al Ali, and S. Al Ali. 2004. Annual fisheries statistics for Abu Dhabi Emirate 2003, 87 p. Environmental Research and Wildlife Development Agency, P.O. Box 45553, Government of Abu Dhabi, United Arab Emirates. ence points and other pertinent information required for management. Specific objectives included establishing key demographic parameters by using validated age estimates, identifying reproductive characteristics and conducting yield-per-recruit analyses for the selected study species. Materials and methods Study site and sampling protocol Size-frequency data were collected from commercial catches made off the coast of the Emirate of Abu Dhabi in the United Arab Emirates (Fig. 1) between September 2000 and March 2003. Fish were selected at random from landings and fork length (Lp) was recorded to the nearest cm by using a measuring board. Monthly target sample sizes were 500 individuals per species. Biological data was collected from specimens pur- chased from commercial catches between June 2002 and May 2003. Samples were obtained from 30 individuals of each species from a representative size range during the last week of each month. Standard length (Lg), fork length (Lp), and total length (L.j.) were recorded to the nearest mm by using a measuring board. Whole wet weight was measured with an electronic balance and recorded to the nearest g. The sex of a fish was deter- mined by macroscopic examination of the gonad, which Grandcourt et al : Biology and assessment of Diagramma pictum and Lethrinus nebulosus in the southern Arabian Gulf 77 was removed and subsequently weighed to 0.1 g with an electronic balance. Sagittal otoliths were extracted, cleaned in water, dried, and stored in seed envelopes. One of each pair of sagittae was weighed to 0.1 mg, burnt on a hotplate until it changed to a dark brown color, and embedded in epoxy resin. Transverse sections through the nucleus (of approximately 200-300 jim thickness) were obtained by using a twin blade saw. Sections were mounted on glass slides and examined with a low-power microscope and transmitted light. Age and growth The number of opaque bands in transverse sections was recorded in addition to the optical characteristics of the outer margin (opaque or translucent). The proportion of samples with opaque or translucent margins was calculated for each month and used to infer the timing and periodicity of increment formation. The age at which the first opaque band formed was calculated as the time between the mean birth date and the time of formation of opaque bands. Subsequently, the absolute age was calculated as the age at formation of the first band plus the number of opaque bands outside the first band and the time between the formation of the last band and cap- ture. In order to establish the relationship of the timing of opaque zone formation with trends in sea water tem- perature, time series data were converted by using the scaling process given in Newman and Dunk (2003). Growth was investigated by fitting the von Berta- lanffy growth function (von Bertalanffy, 1938) to size- at-age data using standard nonlinear optimization methods. The model was fitted to pooled data and each sex separately. The von Bertalanffy growth function is defined as follows: L, = L., il -e -k U-l„) «'), where L, = length at time t\ L^= the asymptotic length; k = the instantaneous growth coefficient; and t^ = the hypothetical time at which length is equal to 0. Growth curves were compared between sexes for each species by using the analysis of residual sums of squares method of Chen et al. (1992). The growth performance index (P (Gayanilo and Pau- ly, 1997) was calculated in order to provide a basis for the comparison of growth characteristics in terms of length: *' = - 2/3 logio (a), where = logjg ik) + 0.67 logm iWj and W, = aLj. The constant, a, was derived from length-weight rela- tionships and k and L^ were obtained from the von Bertalanffy growth function. Parameters of the length-weight relationship were ob- tained by fitting the power function W = oLp* to length and weight data where W is the total wet weight, a is a constant determined empirically, Lp is the fork length, and b is close to 3.0 for species with isometric growth. Ratios of total length (Lj) to fork length (Lp) were also calculated for each species. Reproduction The mean size at first sexual maturity was estimated for females by fitting the logistic function to the proportion of mature fish in 2-cm (Lp) size categories. The mean size at first maturity was taken as that at which 50% of individuals were mature. Gonadosomatic indices, calculated by expressing gonad weight as a proportion of total body weight, were plotted against the sample period by month to establish the timing and seasonality of spawning. The mean birth date was estimated from patterns in reproductive indices. Population sex ratios were examined by using x" good- ness-of-fit tests. Independent tests were conducted to determine whether sex ratios differed significantly from unity for whole samples and for size and age categories within samples. The probability level was set at 0.05 (a=0.05) and Yates's correction factor was used on ac- count of there being only one degree of freedom for each comparison. Juvenile retention was calculated as the proportion of fish in aggregated size-frequency samples below the mean size at first sexual maturity. Mortality and selectivity Size-at-age data were used to construct age-length keys following the method of Ricker (1975) and these were used to convert aggregated length-frequency data into age-frequency distributions. The number of fish above the age at which fish were fully recruited to the fishery was calculated as a proportion of the total number of fish. The annual instantaneous rate of total mortality (Z) was subsequently determined with the age-based catch curve method (Beverton and Holt, 1957). The natural logarithm of the number of fish in each age class was plotted against the corresponding age, and Z <±95% CI) was estimated from the descending slope of the best fitting line by using least-squares linear regression. Initial ascending points representing fish that were not fully recruited to the fishery were excluded from the analyses. The annual instantaneous rate of total mortality was also estimated with the length-converted catch method of Pauly (1983). Pooled length-frequency samples were converted into relative age-frequency distributions by using parameters of the von Bertalanffy growth func- tion. The natural logarithm of the number of fish in each relative age group divided by the change in rela- tive age was plotted against the relative age, and Z (±95% CI) was estimated from the descending slope of the best fitting line with least-squares linear regres- sion. The estimates of Z from age-based and length- 78 Fishery Bulletin 104(1) converted catch curves were compared by using a modified Ntest (Sokal and Rohlf, 1995 . Backwards extrapolation of the length-con- verted catch curves was used to estimate the probability of capture data. Selectivity curves were generated by fitting a logistic function to the plot of the probability of cap- ture against size, from which values of the parameters L^g, L^g, and the size at which fish were fully recruited to the fishery (Ljqq) were obtained. Estimates of the annual instantaneous rate of natural mortality (M) were obtained for each species with the empirical equation derived by Hoenig (1983). Maximum age estimates of 31 years for D. pictiim and 21 years for L. nebulosus from the literature (Loubens, 1980; Edwards and Shaher. 1991) were used because the maximum ages and sizes obtained in our study were considerably lower than other reported values. The annual instantaneous rate of fish- ing-induced mortality (F) was calculated by subtracting the natural mortality rate (M) from the total mortality rate (Z) derived from age-based catch curves (F=Z-M). The calcu- lation was also made for the upper and lower 95% confidence intervals for estimates of Z in order to derive a range of fishing mortal- ity rate estimates. The exploitation rate (E) was calculated as the proportion of the fish- ing mortality in relation to total mortality {E=FIZ). Assessment of the fishery Relative yield and biomass-per-recruit analyses were used to assess the fishery. Growth (k and L^), mortal- ity (M), and selectivity (Z-f^) parameters were used as model inputs, and knife-edge selection was assumed. The Beverton and Holt (1966) yield-per-recruit (YPR) model modified by Pauly and Soriano (1986) was used to estimate the sizes at maximum yield per recruit ^L^^^) and to predict the effects of increasing the mean size at first capture (Z-5q) to the mean size at first sexual maturity (L^^^) and that at which yield per recruit would be maximized (^,„ax*- Estimates of exploitation rates representing 1) a marginal increase of relative yield per recruit which is 0.1 of its value at the origin (£gi) and 2) maximum yield iE^^^^) were also derived from the model. The exploitation rates corresponding to i^^pt and ^iin,,, (^upt and£|,|^|,) were calculated and used to estimate the relative biomass per recruit for each ,^J^ V ^^^E> ■■^/>---l Figure 2 Photomicrographs of transverse sections through the sagittal oto- liths of (A) Diagramma pictum. 56.0 cm Lj, and (B) Lethrinus nebu- losus, 45.7 cm Lp. Dots show opaque zones (scale bar=l niml. and i'niax frorn relative biomass- species for Ljq, L^ per-recruit curves. Precautionary target iF^^) and limit ^^limit' biological reference points were calculated as 0,5 and 2/3 M, respectively, and used to infer resource status by direct comparison with the fishing mortality rates established for the study species. Results Age and growth Alternating translucent and opaque growth increments were observed in transverse sections of the sagittal otoliths of D. pictum and L. nebulosus when viewed with transmitted light under low-power magnification (Fig. 2). For both species, one growth increment consist- ing of an opaque and translucent zone was formed on an annual basis. Opaque bands formed in the summer months between May and September in association with increasing sea water temperatures (Fig. 3); conversely translucent zones were deposited in the autumn and winter (October to February) in association with decreas- ing sea water temperatures. The maximum age estimates determined from counts of opaque bands were 13 and 14 years for D. pictum and L. tiebulosus, respectively. Size-at-age relationships were asymptotic in form and there was considerable individual variability in growth (Fig. 4) (parameters of the von Bertalanffy growth function are given in Table 1). A comparison of the growth characteristics between sexes revealed that there were no significant differ- ences in parameter estimates for both species (P=0.125, df=319 for D. pictum and P=0.878, df =324 for L. nebu- Grandcourt et a\ Biology and assessment of Diogiamma pictum and Lethnnus nebulosus in the southern Arabian Gulf 79 losus). Values of the growth performance index

— >^ 0 -0 1 - < 1 1 1 1 1 1 I 1 1 1 1 01 23456789 10 11 12 Month Figure 3 The proportion of otoliths with opaque outer margins for (A) Diagramma pictum (n = 348l and (B) Lethnnus. nebulosus (n = 343) and monthly sea temperatures off the Emirate of Abu Dhabi. Note that the values have been converted to a standardized scale to enable comparison of the trends. Modal age groups in age-frequency distributions derived from age-length keys and size-frequency data were 3 years for D. pictum and 5 years for L. nebulosus (Fig. 6). The proportion offish above the age at which fish were fully recruited was 13.8% and 45.7% for D. pictum and L. nebulosus, respectively. There were no significant differences between the total mortality rate estimates derived from age-based and length-converted catch curves for D. pictum {t=0.81, P=0.43, 15 df) and L. nebulosus (^=0.03, P=0.98, 11 df) (Fig. 7). Fishing mortality rates were in excess of the natural mortality rates, accounting for 79% and 64% of the total mortality for D. pictum and L. nebulosus, respectively (Table 4). The selectivity range derived from plots of the prob- ability of capture at size was 25.0 cm for D. pictum (18.0-43.0 cm) and 34.0 cm for L. nebulosus (13.0-47.0 cm) (Fig. 8). Values of the sizes where the probability of capture was 50% (Ljq), 75% (L-g), and 100% (Lj,,,,) are given in Table 5. For both species, fish were recruited Table 1 Parameters of the von Bertalanffy growth function, coef- ficients of determination (/■-), and sample sizes in) for Dia- gramma pictum and Lethrinus nebulosus in the southern Arabian Gulf D pictum L. nebulosus Males Fe- males All Males Fe- males All ;; 0.29 0.23 0.24 0.10 0.11 0.11 L^cm(Lp) 60.6 63.8 63.0 69.9 65.2 66.2 ^olyr) -1.2 -1.5 -1.4 -3.3 -2.9 -3.0 r2 0.86 0.83 0.84 0.91 0.88 0.79 n 81 244 325 86 244 330 80 Fishery Bulletin 104(1) Figure 4 The von Bertalanffy growth function fitted to size-at- age relationships for (A) Diagramma pictuni and (B) Lethriiius nebulosus. 4.0 1 A 1 3.0 ■ 1 2,0 ■ ' • Females — »— Males 1.0 ■ ' , r T £ L J L. ■ I v: [ 1 0,0 1 c u ro E o S 7.0-1 CO S 6 0 ■ ^^~ a ^~' "*■♦ •^ J r-'-'-N.. m"-""'^ 2 B 3 1 5 6 7 8 9 10 11 12 5,0 - 4,0 ■ 3 0 ■ I ^ \ 2,0 ■ - / \ .,,,... Females \ — •— Males 1.0 ' ■ \ . M f f f f M 1 2 3 4 5 6 7 8 9 10 11 12 Month Figure 5 Mean monthly gonadosomatic indices (±SE) for (A) Diagramm pictum (n = 359) and (B) Lethrinus nebulosus (/?=360l. Table 2 Results of chi-square goodness -of-fit tests on sex ratios within age categories forZ) agra mma pictum and Lethrinus nebulosus (* significant at it- 0.05). Age category (yr) No of males No. of females Chi square total) P D. pictum 0-2 29 130 64.2 <0.01* 3-5 38 91 21.8 <0.01* 6-13 16 29 3.8 >0.05 L, nebulosus 0-2 34 93 27.4 <0.01* 3-6 29 88 29.8 <0.01* 7-14 27 69 18.4 <0.01* Grandcourt et al Biology and assessment of Diagiamma pictum and Lethnnus nebulosus in the southern Arabian Gulf 81 2000 1500 1000 500 n = 5006 0 1 2 3 4 5 6 7 8 9 10 11 12 13 2000 ■ 1500 1000 500 n = 13.129 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Age group (yr) Figure 6 Age-frequency distributions for I A) Diagramma pictum and (Bi Lethrinus nebulosus derived from age-length keys and size-frequency data. to the fishery at a mean size iLc^g) which was smaller than the mean size at which first sexual maturity was attained (L,„j,(). Assessment of the fishery The size at which yield per recruit would be maximized *-^max' w^s ■^4.4 cm (Lp) for D. pictum and 36.9 cm (Lp) for L. nebulosus. For both species, these values were considerably greater than the mean size at first capture and the mean size at first sexual maturity (Fig. 9). The exploitation rate for D. pictum (0.79/yrl was greater than that which would maximize yield per re- cruit (0.57/yr) at the existing mean size at first capture (Table 6). Furthermore, the same yield per recruit could be achieved at a much lower exploitation rate and at an increased relative biomass per recruit (Fig. 10). The yield-per-recruit function also indicated that an increase in the size at first capture to that which would Table 3 Results of chi -square goodness-of-fit tests on sex ratios within size categories for Diagramma pictum and Lethn- 1 nus nebulosus (*=sign ficantat a = 0.05). Size Chi- category No. 0 f No. of square (cm Lp) males females (total) P D. pictum 20-34 26 128 67.6 <0.01* 35-49 41 93 20.2 <0.01* 50-64 27 43 3.7 >0.05 L. nebulosus 15-29 40 121 40.8 <0.01* 30-44 41 89 17.7 <0.01* 45-54 26 73 22.3 <0.01* Table 4 Mortality and exploitation rates (±95 CI) for Diagramma pictum and Lethrinus nebulosus. D. pictum L. nebulosus Total mortality rate/yr(Z) (Age-based catch curve) 0.63 (0.50-0.75) 0.56 (0.35-0.77) (Length- converted catch curve) 0.69(0.58-0.79) 0.56(0.53-0.60) Natural mortality rate/yr (Mt 0.13 0.20 Fishing mortality rate/yr (F) 0.50(0.37-0.62) 0.36(0.15-0.57) Exploitation rate/yr (£) 0.79 0.64 Table 5 Probability of capture ( selectivity) for Diagramma pictum and Lethrinus nebulosus. Probability of capture 50% (L50) D. pictum L. nebulosus cm (Lp) cm (Lp) 21.1 26.4 30.7 35.1 40.3 43.8 82 Fishery Bulletin 104(1) • Length converted * Age based 7 6 5 - t 4 ■ 3 2 1 B »oo' 10 12 14 16 • Length converted A Age based 5 10 15 20 Age (yr)- Relative age (yr) 25 Figure 7 Age-based catch curves (InF against A^e) and length- converted catch curves (In F/dt against Relative age) for Diagramma pictum (A) and Lethrinus nebulosus (B). Dashed and solid lines show the regression equation {y = b.x+a) fitted to data for age-based and length-con- verted catch curves, respectively. Only solid points are included in the regressions. Table 6 Exploitation rates iEg^ and E^^^^) at the existing size at first capture (L^p), the size at first capture at sexual maturity iL^^,) and the size at first capture at maximum yield per recruit (L^^^) for Diagramma pictum and Leth- rinus nebulosus. D. pictum L. nebulosus '•ate L50 ^mat i^ax ^50 ^mat -^max £u,(/yr) 0.50 0.56 0.75 E„3, (/yr) 0.57 0.65 0.81 0.51 0.56 0.75 0.63 0.65 0.86 10 20 30 50 60 70 B 30 40 50 Fork length (cm) 60 70 Figure 8 Selectivity curves for (A) Diagramma pictum and (B) Lethrinus nebulosus, showing the mean size at first capture at probabilities of 0.5 (^5,,), 0.75 (L-5). and the size at which fish are fully recruited to the fishery 0 u.us ■ 0.04 ■ 0.03 ■ ^^ ^ ^"^^^ 0.02 ■ ^ v 0.01 ■ / L50 / - Lmat X ^^ toi ^^^^^:ssw 0 • , ^^^ 0.6 0.4 0.2 0 0.2 0,4 0,6 0.8 1 Exploitation rate (/yr) Figure 10 Relative yield and respective biomass-per-recruit curves (descending lines) for (A) Diagramma pictum and (B) Lethrinus nebulosus showing the existing exploitation rate (£) and the exploitation rate at which the slope of the yield-per-recruit curve is 0.1 of its value at the origin iEqj). Curves show the effect of increasing the existing mean size at first capture (LjqI to the mean size at first sexual maturity (imat* ^^^ ^^^ ^'^^ which would maximize yield per recruit ih^^^). Estimates of precautionary target and limit exploita- tion rates (E ^ and £|,j„,t) for D. pictum were 0.34 and 0.40, respectively. The exploitation rate for L. nebulosus (0.64/yr) was marginally greater than that which would maximize yield per recruit (0.63/yr) at the existing mean size at first capture (Table 6). The yield-per-recruit function indicated that an increase in the size at first capture to that which would maximize yield per recruit would be associated with an increase in yield at the current level of exploitation. An increase in the size at first capture to that at which sexual maturity occurs iL^^^) was also predicted to be associated with an increase in yield, although to a lesser degree (Fig. 10). The relative biomass per recruit for L. nebulosus at the current exploitation rate was less than 20% of that Table 7 Relative biomass per recruit at precautionary exploita- tion rates (E ^ and E^^^^^) at the existing size at first cap- ture (L5Q ), the size at first capture at sexual maturity (Lj^^^^j), and the size at first capture at maximum yield per recruit '•^max' '*"" Diagramma pictum and Lethrinus nebulosus. Relative biomass per recruit Exploitation rate D. pictum L. nebulosus ^50 mat max ^50 mat max £i,m,t') in terms of growth in length with other estimates obtained for the same or a similar species (Gayanilo and Pauly, 1997). Values of

can vary among three eras of exploitation; a "prehistoric" period, during which little data are available; a "modern" period, when presumably there are some data on abundance or mortality rates; and a "future" period, when fishing mortality rates are con- trolled (input). The absence of data during the "prehis- toric era" generally precludes the estimation of annual deviations in recruitment (f) or fishing mortality rate (b) during that period. The average weight or fecundity of the plus group is expressed as a function of the average age of the plus- group. Initially, it is assumed that the age composition of the plus-group is in equilibrium consistent with Equa- tion 1, in which case the average age of the plus-group at the beginning of the first year is approximately -M,, a^i = A + 1-e -M^ (7) Subsequently, the age of the plus-group is updated as AN A-l.y -'^v''.4-l-'W.4-l / — .1 V AT - i" J' X~ ^1 A '.4.y+l N (8) A,y+\ Reference points Equations 1-4 describe the relative dynamics of a popu- lation apart from its absolute abundance. As such they are suitable for developing management plans where the fishing mortality rate is controlled directly (e.g., by reducing effort) and the biomass reference points are expressed on a relative scale. When the virgin spawn- ing biomass itself is used as the reference point, the estimated value of s^ is a direct measure of the status of the stock. For example, if the management goal is to maintain spawning biomass at or above 50% of the virgin level, then estimates of s below 0.5 may trigger some action to reduce fishing pressure. A related reference point is the equilibrium spawning potential ratio (Goodyear, 1993), defined as the expected lifetime fecundity per recruit at a given F (i/>^) divided by the expected lifetime fecundity in the absence of fishing (ly'fl): Wo a=0 (9) -J^Fi^+M, As shown in Appendix 2, the corresponding equilibrium level of relative spawning biomass (denoted by a tilde) may be computed as 1-1- log. P log, a ap -1 a-1 Ricker Beverton and Holt (10) Beverton and Holt 0-8 /T^"^ ^^^ 0.6 // ."""^ 0,4 \fx -♦-2 -*-10 0,2 i/ -•-40 oK 0 0.2 0-4 0,6 0.8 1 Ricker 4 -1 y^^*"^^ -•-2 3 - / X. -*-10 -•-40 2 - 1 - /x^^^^"^^:!::: ^ .] k::-— - 0 0.2 0.4 0.6 0.8 1 S Figure 1 Examples of scaled Beverton-Holt and Ricker spawner-recruit relationships for various values of | a (maximum lifetime fecundity). The labels s and r refer to relative spawning biomass and relative recruitment, respectively. Note that s is independent of the vulnerability vector I'. Accordingly, MSST definitions based on s will have the desirable property of being insensitive to changes in fishery behavior. Other management plans employ reference points such as ^,„„,. or recruit statistic Fqj, which are based on the yield per ^V-ly- 1-e -(R' +M„) -I /;•,+*', (11) Fv+M„ where w^ is some measure related to the average weight of the catch. Inasmuch as there are no terms involving the absolute abundance of the stock, the calculation of such statistics poses no special problems for the relative framework presented in the present study. Prescriptions based on the maximum sustainable yield (MSY) are slightly more complicated because equilibrium yield is the product of equilibrium recruitment R and equilib- rium yield per recruit: A 1 .p-'/^'n+^a' -Y.F",*I^, Y = RpJ^w,,Fv„ ^ .^ e -0 . (12) Fv+M„ 92 Fishery Bulletin 104(1) However, the fishing mortality rate that maximizes Equation 12 also maximizes Equation 12 divided by the virgir. recruitment /fg (a constant). Thus, i^^v/sv ™3y be obtained from max F e ■-" (13) where s Jp has been substituted for RplR^y The values of p and s corresponding to F„,^,^ , Fq ,, or F,„,,. may be calculated by means of Equations 9 and 10, respectively. Note however, that p is no longer the target value specified by management, but a derivative of the targeted values of F. This means that MSST defi- nitions based on s, ,, , s,, ,, and s,., will vary somewhat III (l\ '.I 1 I'lsv -^ with the behavior of the fishery. In some cases this could lead to risk prone situations where the percep- tion of stock status changes simply because the fishery targets different age groups (i.e., the definition of MSST changes rather than the abundance of the resource). In the case of MSY. a more stable alternative is to define the MSST in terms of a "spawn at least once" policy where mature animals are regarded as fully vulnerable to the fishery and immature animals area regarded as invulnerable. Parameter estimation The equations above include numerous "unknowns" rep- resenting the processes of reproduction, mortality, and growth. In the case of "data-poor" stocks like goliath grouper, there are insufficient data to estimate all of these unknown parameters with an acceptable level of precision. However, it is often possible to increase the precision of the estimates through the use of Bayesian prior probability densities constructed to reflect expert opinion (e.g., Wolfson et al., 1996; Punt and Walker, 1998) or based on meta-analyses involving similar spe- cies (e.g., Liermann and Hilborn, 1997; Maunder and Deriso, 2003). The Bayesian approach to estimation seeks to develop a "posterior" probability density for the parameters 0 that is conditioned on the data D, P{0 I D). By applica- tion of Bayes rule it is easy to show that P{0\D)cl0)-log,P(0)|. (16) In the present model, a prior needs to be specified for the parameters reflecting recruitment (o and £ ), mortality (M, (p. d^,, i',, ), fecundity (£„', and growth in weight ((f„). It is assumed in the present study that the parameters are statistically independent with respect to prior knowledge, such that the joint prior is merely the product of the marginal priors for each parameter. The exceptions are the process error functions for the annual recruitment and fishing mortality rate devia- tions, f^ and (3^. These are assumed to be autocorrelated lognormal variates with negative- log density functions of the form -logP(f) = 2a fr + I*' -pr£.y - log cr^. (17) where p, = the correlation coefficient; and o'\ = the variance of log^.j;^. For stability reasons, it is assumed that £g = 0. It is possible, at least in principle, to conduct an as- sessment based on prior specifications alone. However, it may be difficult to develop sufficiently informative pri- ors for some of the parameters, particularly for the fish- ing mortality rates. The preferred approach, of course, is to condition the estimates on data. With the present model it is assumed that catch data are either unavail- able or unreliable, otherwise a standard age-structured production model (cf. Restrepo and Legault, 1998) would be more appropriate. However, time series of catch-per- unit-of-effort data or fishery-independent surveys are often available even when total catches are not. To the extent that changes in these data (r) are proportional to changes in the abundance of the population as a whole (N), they may be modeled as C,..=Q,Y.^,,,Na.ye -(f,r„+M„l(, y,^^ (18) y, ^, - NonnahO.a^., ), where / t. = and ndexes the particular survey time series; = the proportionality coefficient scaling the time series to the relative abundance of the population; the fraction of the year elapsed at the time of the survey; the standard deviation of the fluctuations in logp c, owing to observation errors or changes in the distribution of the stock; and the relative vulnerability of each age class to the fishery and the /"' survey, respectively. The corresponding negative logarithm of the sampling density is Porch et al : A catch-free assessment model with application to Epinephelus ita/ara 93 log P(c\0} = 'Iog,,(c,,,)-log,, II 2a- + logCT,., (19) Anecdotal observations may be treated in similar fashion. For example the perceptions of constituents on the abundance of the resource relative to virgin levels {/}) can be modeled as 6 (22) Z = 200.6(l-e-''^26,a.0.49lj 1 - A /\ M B ^^-^ ^^^ 0.8 - / \ °^ " ^^ ^ 0.6 - / \ °^ " ^ 0 4 - / \ ° '' ■ >• 0.2 ■ / >^ °^ " ■S 0- J ^ ^ 1 1 -1 U 1 . , . o 0 0.1 0.2 0.3 0 0 1 0.2 0.3 0 4 0 5 ? M 0, Relativ C I ^-"-■"""^^ 1 - ) 0,8 - \. 0.8 - 0.6 - \^^ 0.6 - 0.4 - N,^ 0 4 0.2 - ^\^ 0 2 ■ ^ _— — 1 0 2 4 6 8 10 0 20 40 60 80 100 02 1-03 (%) Figure 3 Priors for the mortality rate parameters: (A) lognormal prior for natural mortality rate, (B) truncated normal prior for proportionality factor 0j, (C) truncated normal prior for multiplier $2 (D) gamma prior for percent reduction in F associated with the 1990 harvest ban. The upper and lower boundaries for each parameter are as given on the horizontal axes. where w = weight in kg; and / = length in cm expressed as a von Bertalanffy function of age (see Bullock et al., 1992). Natural mortality The maximum observed age of 37 years (Sadovy and Eklund, 1999) suggests a value for M of about 0.11/yr according to the method of Hoenig (1983). Legault and Eklund"* suggested a plausible range of 0.037 yr to 0.19/yr (midpoint 0.11) based on an analy- sis of the fraction surviving to various maximum ages. To reflect this uncertainty, a lognormal prior with a median of 0.11 and CV of 0.4 was used (Fig. 3A). Fishing mortality rate and relative vulnerability A large fraction of the recreational landings of goliath grouper appear to come from the Ten Thousand Islands area in Southwest Florida, where most of the animals caught have been between the ages of one and five years. How- ever, large animals were often targeted by commercial and recreational fishermen in other areas. Accordingly, we assumed the vulnerability of goliath grouper gener- ally increased with age according to the sigmoid-shaped logistic curve (23) 1-i-e -(n-a„, l/rf Estimates for the parameters ar,o and d were obtained by fitting the curve (weighted by cumulative mortality at age ) to the relative frequency of ages in two different data sets. The first data set included mostly juveniles animals between the ages of 0 and 5, obtained during creel censuses of recreational catches in the Ten Thousand Islands area of the Ever- glades National Park (see Porch et 3 Legault, C. M., and A.-M. Eklund. 1998. Generation times for Nassau grouper and jewfish with comments on M/K ratios. Sustainable Fisher- ies Division Contribution SFD-97/98- lOA, 5 p. Southeast Fisheries Science Center, 75 Virginia Beach Drive, Mi- ami, Florida 33149. Porch et al : A catch-free assessment model with application to Epinephelus ita/ara 95 al.^). The second data set included mostly adult animals obtained opportunistically from recreational and com- mercial catches in the eastern Gulf of Mexico (Bullock et al., 1992). The SEDAR stock assessment review panel based their advice on models that used the former selec- tion curve (Kingsley'); however the effect of using the latter curve was examined as a sensitivity analysis. The two curves are contrasted in Fig. 4A. The fishing mortality rate on the most vulnerable age class was modeled as follows: 0/,, 1900 2, 03, 6y are parameters to be estimated. In the present study, effort was assumed to track the U.S. Census'' for the number of people living in South Florida coastal counties between 1900 and 1980. From 1980 to 1989 this assumption was no longer required owing to the availability of several time series of relative abundance (see below). Instead, interannual variations in fishing mortality were modeled according to Equation 5 with median (^.,Fjc,-c|, log-scale variance a7,.=0.15 and correlation coefficient p^, = 0.5, which essentially amounts to a mild constraint on year-to-year changes in F. The nonzero correlation coefficient is intended to reflect the momentum in effective fishing effort from one year to the next that arises from a combination of market demands and the tendency of many fishermen to target only the species they are most adept at catching. Even so, the relatively large variance term admits substantial inter- annual variations if the data warrant them. Moreover, runs with p^,= 0.0 (no year-to-year momentum) did not produce substantially different results. The effect of the harvest moratorium was modeled as a percentage ^g of the average fishing mortality rate in the 1980-89 period. Relatively uninformative priors were used for (p^ and (p^ (Fig. 3, B and C). A somewhat * Porch, C. E., A-M. Eklund and G. P. Scott. 2003. An assess- ment of rebuilding times for goliath grouper. Sustainable Fisheries Division Contribution SFD-2003-0018. Southeast Fisheries Science Center, 75 Virginia Beach Drive, Miami, Florida 33149. 26 p. 5 Kingsley, M. C. S., ed. 2004. The Goliath Grouper in southern Florida: assessment review and advisory report. Report prepared for the South Atlantic Fishery Management Council, the Gulf of Mexico Fishery Management Council, and the National Marine Fisheries Service, 17 p. South Atlantic fishery Management Council, 1 Southpark Circle, Charleston SC"29406. ^ Population of Florida Counties by Decennial Census: 1900 to 1990, 4 p. 1995. Compiled and edited by Richard L. Forstall. Population Division, U.S. Bureau of the Census. Washington, DC 20233 more informative prior with bounds between 0.01 and 0.5 was used for ip., based on the opinions of members of the SEDAR panel (Fig. 3D). Survey information Porch and Eklund (2004) have developed relative indices of abundance from two visual surveys: the personal observations of a professional spearfisher (DeMaria") and a volunteer fish-monitor- ing program administered by the Reef Education and Environmental Foundation (REEF 2000). In addition. Cass-Calay and Schmidt*^ have standardized catch rate data collected in the Ten Thousand Islands area by the Everglades National Park (ENP). The two visual surveys are assumed to reflect the abundance of mature fish ages 6 and older (based on diver reports of size). The ENP catch rate index, on the other hand, is assumed to reflect the relative abundance of juveniles with relative vulnerabilities given by the dome-shaped gamma func- tion (normalized to a maximum of 1): ■'ENP. a ° ^,l-°/Oin.y. ^100% (25) where Ojoo', - ^he most vulnerable age; and CV = the coefficient of variation. Estimates for ajQ,,,; (3.47) and CV (0.34) were obtained by fitting the mortality-weighted gamma curve to the frequency of ages 0 -7 in the Ten Thousand Islands data mentioned earlier (for more detail see Porch et al.^). The resulting curve is shown in Figure 4B. Anecdotal Impressions of stock status Johannes et al. (2000) pointed out that local fishermen often disagree with the conclusions drawn by scientists in data-poor situations and suggest that many times additional data will prove the fishermen correct. As mentioned ear- lier, expert judgements about the relative abundance of a stock can be treated as data or represented by a "prior." We collected information on the value of s at the time moratoriums began (1990) by interviewing fishermen and divers who had been active in southern Florida since the early 1960s or before. Specifically, interviewees were asked to state their perception of the percent reduction in goliath grouper populations from the time they began diving to the time the mora- torium on catch was imposed (1990). The average per- cent reduction reported for large goliath (approximately age 6 and older) was 86% (standard deviation of about 13%, Table 1). This information was modeled as data in accordance with Equation 20. ■ DeMaria, D. 2004. Personal, obsery. P.O. Box 420975, Summerland Key, FL 33042. ** Cass-Calay, S. L., and T. W. Schmidt. In review. Stan- dardized catch rates of juvenile goliath grouper, Epinephelus itajara, from the Everglades National Park Creel Survey, 1973-1999. 96 Fishery Bulletin 104(1) Results The model was able to fit the ENP index of juvenile goli- ath grouper very well but could not reconcile the conflict- ing trends indicated by the DeMaria and REEF indices Figure 4 Selection curves used to represent the vulnerability of goliath grouper [Epinephelus itajarat to (A) the overall fishery and (B) Everglades National Park anglers. The logistic curves shown in (Al were fitted to either age- composition data derived from the Everglades National Park (ENPl creel census or opportunistic samples from offshore fishing trips (Bullock et al., 1992). DeMaria 1970 Year Figure 5 Base model fitted to the four indices of abundance for goliath grouper iEpinepheliis itajara) in southern Florida. for adult goliath grouper (Fig. 5). The estimated trends in spawning biomass were rather uncertain (Fig. 6A), but nevertheless indicated a rapid decline to about 5% of virgin levels by the time the harvest ban was imposed in 1990, followed by a significant increase. The estimates of fishing mortality were also somewhat uncertain, but gen- erally indicated a gradual increase in fishing mortality to moderate levels during the 1970s followed by a rapid increase during the 1980s (Fig. 6B). The harvest mora- torium was estimated to have been about 83% effective in reducing fishing mortality, nevertheless losses owing to human activities (e.g., illegal harvest and release mor- tality of animals caught at depth) were still estimated to be substantial (F =0.05/yr). If, in accordance with the Gulf of Mexico Management Council's generic Sus- tainable Fisheries Act amendment, the limit reference point is taken to be the equilibrium spawning biomass corresponding to a spawning potential ratio of 50% , then the model indicates that current fishing mortality rates are near F^^c, and that there is less than a 50% chance the stock will recover within 15 years (Fig. 7). Sensitivity runs were conducted to examine the im- plications of 1) dropping one or more of the indices, 2) increasing the assumed minimum age represented in the REEF and DeMaria indices from 6 to 10, 3) assuming that the historical period began in 1950 rather than 1900 and using the anecdotal informa- tion as a tuning index and (4) using the alternate fishery selection curve that was fitted to the data from Bullock et al. (1992), where adult animals were much more vulnerable to the fishery than were juveniles. Of these, the results were most sensitive to removal of the DeMaria index — the projected trends being much more optimistic (Fig. 8). This is because the DeMaria index indicates that the adult population increased rapidly during the first few years of the harvest ban, but then suffered a set back in 1999 and has since leveled off. In contrast, the REEF index in- dicates that the population continued to increase during that time. Thus, when the DeMaria index is removed, the model allows for a faster postmoratorium increase in the adult population by esti- mating a low fishing mortal- ity rate of about 0.01/yr (i.e., a harvest ban that is 97% effective). The fishing mor- tality rate estimates for the 1980s are also lower without the DeMaria index inasmuch as the DeMaria index indi- cates a more precipitous de- cline during that time than the ENP index (the REEF index does not begin until 1994). 2000 Anecdotal 1950 1960 1970 1980 1990 2000 Porch et al.: A catch-free assessment model with application to Epinephelus ita/aro 97 1950 1960 1970 1980 1990 2000 2010 2020 Year 1950 1960 1970 1980 1990 Year 2000 2010 2020 Figure 6 Base model predictions of (A) spawning biomass of goli- ath grouper (£. itajara) in southern Florida in relation to the equilibrium level associated with a spawning potential ratio of 50%, s/Sjq,.,, and (B) the apical fishing mortality rate on goliath grouper (£. itajara), F^ .^i. The lines with small dashes represent approximate 80% confidence limits and the dashed horizontal lines represent the levels associated with an SPR of 50% . The sensitivity run with the alternate selection curve also produced more optimistic results (Fig. 8). Inasmuch as the model now attributes most of the fishing mortal- ity to age classes well beyond the age at first maturity (see Fig. 41, the spawning stock biomass is estimated to have been reduced to a lesser extent (to about 10% of virgin levels by 1990 as compared to 5%). Thus, other things being equal, recovery requires less time. The lev- el of F, increased with the alternate selection curve because fewer age classes are affected by fishing. Discussion All of the model formulations examined depicted the same qualitative patterns: escalating fishing mortality rates and rapidly declining spawning biomass, particu- larly during the 1980s, followed by a sharp decrease in fishing mortality and strong recovery in spawning biomass after the 1990 harvest ban. These trends are remarkably consistent with the anecdotal observations shown in Table 1 and Figure 5 as well as with the expert testimony given during the SEDAR stock assessment 1 00 1 ~ 0 80 ■ o CO « 0 60 ■ >. ■i 0 40 • J3 o D- 0 20 ■ ^-^^^^ y^ 1995 2000 2005 2010 2015 2020 Year Figure 7 Predicted probability that the stock of goliath grouper (E. itajara) in southern Florida will have recovered to levels exceeding the equilibrium spawning biomass associated with a spawning potential ratio of 50%. review. The estimated rapid increase in fishing mortality during the 1980s appears to reflect a real increase in effort that occurred due to elevated demand and selling prices (Sadovy and Eklund, 1999), as well as the wide- spread use of the LORAN-C navigational system (which made it easier for fishermen to relocate productive off- shore shipwrecks). Thus, it seems safe to conclude that the population was overfished at the time the harvest ban was imposed and is currently undergoing a sub- stantial recovery. Less clear is the extent to which the population has recovered since the harvest ban. Using the base model, we estimated that the harvest ban has reduced fishing pressure by more than 50%, but probably less than 90% (Fig. 9). Thus, there is a strong chance that the current fishing mortality rate, although greatly reduced as compared to the 1980s, remains greater than i^jo'; 'i-^-' above 0.05/yr). This in turn translates into less than a 40% chance that the population will recover to levels above SgQ,- within the next 15 years. Several fishermen have testified that the harvest ban is probably less than 90% effective because goliath grouper are still taken illegally in places and because animals caught and released in deeper water often do not survive^; therefore this result does not ap- pear unrealistic. More optimistic results, implying a 70% to 80% chance of recovery within 15 years, were obtained when the DeMaria index was excluded or when selection was oriented more towards older animals. There does not appear to be a strong a priori case for excluding the DeMaria index in favor of the REEF and ENP indices. Although the coverage is rather limited, the trends of the DeMaria index are consistent with those of the ENP index (with a suitable time lag) and with anecdotal accounts of the trends in other areas." The issue of selection is more vexing. It can be argued that the age- composition data from the ENP creel census adequately reflects the composition of the juvenile catch inasmuch 98 Fishery Bulletin 104(1) (D 20 1 5 - 1 0 0 5 00 I 0 -, 0 0) > o o 0) 0 6 ^ 0 4 io 0.2 Exclude DeMarIa index Selection favors older fish Year 20 -J 1 5 ■ ^l^ 1 0 ■ ? }H 0 5 • J ■ 1950 1970 1990 2010 Year Year Figure 8 Trends in relative spawning biomass (s/Sjqsj), apical fishing mortality rate (F). and the probability of recovery is>s^Q,.^) for two sensitivity runs — one exclud- ing the DeMaria index (left) and the other with the selection curve favoring older fish (right). as it comes from the center of juvenile abundance; how- ever most aciults were caught outside this area of abun- dance. Thus, the relative contribution of juveniles and adults to the overall catch is unclear and the directional bias in the fitted logistic selection curve is uncertain. 1 - ^ 08 - » / \ . ' / \ } probabili o posterior 1 / > > 04 - ra 0) '^ 02 - _^ '^ 0 - 1 1 1 1 1 1 0 20 40 60 80 100 Percent reduction in F(1 - ^3) Figure 9 Posterior and prior distributions for the effectiveness of the 1990 harvest ban in reducing the fishing mortality rate F on southern Florida goliath grouper (£. itajara) populations (in relation to the fishing mortality rate levels observed during the 1980s). The only other age composition information that has come to light comes from the study by Bullock et al (1992). which was not designed to provide a random sample of the catch and is probably biased towards larger animals caught on offshore wrecks. In principle, one could reflect this uncertainty more formally either by developing a prior for the selectivity parameters or else by weighting the results from the two selection models. The scientists on the SEDAR stock assessment review panel based their advice on the selection curve derived from the ENP data," which is equivalent to placing negligible weight on the curve derived from the Bullock et al. (1992) data; however they recognized the selection curve as an important source of uncertainty that is difficult to address without adequate data. It is important to emphasize that the Bayesian ap- proach adopted in the present study allows one to ex- plicitly model the uncertainty about parameters such as M, for which no data may exist, but a prior distribution covering the plausible range of values may be specified. There is, of course, the potential for introducing bias when one or more of the priors are based on expert opinion or otherwise subjective information. However, the same sorts of bias can be introduced by conducting sensitivity analyses where the unknown parameters are fixed to various values selected by the analysts. Fur- thermore, if unbiased data are in short supply, analyses Porch et al.: A catch-free assessment model with application to Epinephelus ita/ora 99 based on completely uninformative priors will be useless for generating advice because the range of plausible outcomes is too large. Accordingly, we view the use of subjective priors primarily as a vehicle for provid- ing more realistic limits on uncertainty and prefer to express the model outcomes in terms of probability statements. For example, the point estimate from the base model indicated that the population would never recover to Sgg,, because the fishing mortality rate under the harvest ban was still slightly above F^g.,. However, consideration of the uncertainty led to the conclusion that the chance of recovering to Sggr; within 15 years was nearly 40%. Some sources of uncertainty have not been adequately accounted for in the above assessment. For example, the relationship between fecundity and age is unknown. We used weight-at-age as a proxy for the relative fe- cundity-at-age in our analysis, but it is often the case that fecundity increases with age faster than weight. If this is true for goliath grouper, then our projections would be too optimistic. It should also be remembered that the results apply strictly to the goliath grouper population in southern Florida. It is believed that the center of abundance for the population in U.S. waters is off southern Florida, particularly in the Ten Thou- sand Islands area, but goliath grouper are known to have occurred throughout the coastal waters of Gulf of Mexico and along the east coast of Florida, and on up through the Carolinas. Inasmuch as goliath grouper are not highly migratory, it is possible it may take some additional time for the species to fully occupy its historical range, thus delaying the overall recovery of the U.S. population. The primary advantage of the catch-free assessment model proposed in the present study is that it does not require knowledge of the total number of removals. In this light it is worth noting that 623 of the 905 stocks included in the 2000 annual report to Congress on the Status of Fisheries were listed as having unknown sta- tus, often because catch data were either unavailable or deemed unreliable. Thus we expect the proposed method will become increasingly useful as fishery scientists are asked more and more to develop FMPs for poorly moni- tored fisheries. The fact that the model estimates the population's relative abundance, rather than its absolute abundance, is of little consequence when, as is often the case, adjustments to the target fishing mortality rate or catch quota are made in relation to the levels in previous years (Caddy, 2004). Moreover, certain bi- ases tend to cancel out when dimensionless quantities like relative abundance are used. If, for example, only a consistent fraction of the population were sampled, then the absolute estimates of abundance would be biased but the relative estimates would not (Prager et al., 2003). The greatest drawback of the catch-free method is probably its inability to provide direct estimates of the equilibrium catch levels associated with particular reference points (e.g., MSY). This situation could per- haps be ameliorated by obtaining estimates of absolute abundance from a comprehensive short-term survey covering the entire range of the animal, in which case the relative outputs from the model (including relative catch) could be appropriately scaled. Alternatively, a long-term monitoring program at select sites located throughout the known range of the animal could be es- tablished to detect changes in relative abundance under various closely monitored trial levels of catch. Acknowledgments The present paper benefitted greatly from the scrutiny given to a related document'' by members of the SEDAR stock assessment review panel (R. Allen, L. Barbieri, J. Brodziak, M. Cufone, D. DeMaria, M. Kingsley, D. Murie, M. Murphy, J. Neer, J. Rooker, R. Taylor, E. Toomer, and J. Wheeler). S. Turner, L. Brooks, and two anonymous reviewers also gave helpful comments on the manuscript. S. Cass-Calay and T. Schmidt secured the goliath grouper length-composition data from the Everglades National Park creel survey; J. Brusher and J. Schull provided the age-length data from their sampling program in the Ten Thousand Islands area. Literature cited Annala, J. 1993. Fishery assessment processes in New Zealand's ITQ system. In Proceedings of the international symposium on management strategies for exploited fish populations. 21-24 October 1992, Anchorage Alaska (G. Kruse, D. M. Eggers, R. J. Marasco, C. Pautzke, and T. J. Quinn II, eds. ), p. 791-806. Alaska Sea Grant College Program, Univ. Alaska, Fairbanks. AL. Bard, Y. 1974. Nonlinear parameter estimation. Academic Press, San Diego, CA, 341 p. Bullock, L. H., M. D. Murphy, M. F. Godcharles, and M. E. Mitchell. 1992. Age, growth, and reproduction of jewfish Epineph- elus itajara in the eastern Gulf of Mexico. Fish. Bull. 90:243-249. Caddy, J. F 1998. A short review of precautionary reference points and some proposals for their use in data poor situations. FAQ Fish. Tech. Pap. 379, 30 p. FAO, Rome. 2004. Current usage of fisheries indicators and reference points, and their potential application to management of fisheries for marine invertebrates. Can. J. Fish. Aquat. Sci. 61:1307-1324. Cadrin, S. X., J. A. Boutillier, and J. S. Idoine. 2004. A hierarchical approach to determining reference points for Pandalid shrimp. Can. J. Fish. Aquat. Sci. 61:1373-1391. FAO (Food and Agricultural Organization). 1995. Precautionary approach to fisheries. FAO Fish. Tech. Pap. 350, 210 p. FAO, Rome. Gelman, A., J. B. Carlin, H. 8. Stern, and D. B. Rubin. 1995. Bayesian data analysis, 526 p. Chapman and Hall, London. Goodyear, C. P. 1993. Spawning stock biomass per recruit in fisheries 100 Fishery Bulletin 104(1) management: foundation and current use. In Risk evaluation and biological reference points for fish- eries management (Smith, S. J., J. J. Hunt, and D. Rivard, ed.), p. 67-81. Can Spec. Pub). Fish. Aquat. Sci 120. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mor- tality rates. Fish. Bull. 82:898-903. IWC (International Whaling Commission). 1994. Report of the Scientific Committee, Annex D. Report of the sub-committee on management procedures. Rep. Int. Whal. Comm. 44:74-92. Johannes, R. E., M. M. R. Freeman, and R. J. Hamilton. 2000. Ignore fishers" knowledge and miss the boat. Fish Fish. 1:257-271. Liermann, M., and R. Hilborn. 1997. Depensation in fish stocks: a hierarchic Bayesian meta-analysis. Can. J. Fish. Aquat. Sci. 54: 1976- 1984. Maunder, M. N., and R. B. Deriso. 2003. Estimation of recruitment in catch-at-age models. Can. J. Fish. Aquat. Sci. 60:1204-1216. Myers, R. A., K. G. Bowen, and N. J. Barrowman. 1999. Maximum reproductive rate offish at low popula- tion sizes. Can. J. Fish. Aquat. Sci. 56: 2404-2419. Porch, C. E., and A.-M. Eklund. 2004. Standardized visual counts of goliath grouper off South Florida and their possible use as indices of abundance. Gulf Mex. Sci. 22:155-163. Prager, M. H., C. E. Porch, K. W. Shertzer, and J. F. Caddy. 2003. Targets and limits for management of fisheries: A simple probability-based approach. N. Am. J. Fish. Manag. 23:349-361. Punt, A. E.. and T. I. Walker. 1998. Stock assessment and risk analysis for the school shark (Galeorhinus galeus) off southern Australia. Mar. Freshw. Res, 49:719-731. Restrepo, V. R., and C. M. Legault. 1998. A stochastic implementation of an age-structured production model. In Fishery stock assessment models (F. Funk, T. J. Quinn II, J. Heifetz, J. N. lanelli. J. E. Powers. J. F. Schweigert, P. J. Sullivan, and C.-I. Zhang, eds.), p. 435-450. Alaska Sea Grant College Program Report No. AK-SG-98-01, Univ. Alaska. Fair- banks, AL. Restrepo, V. R., G. G. Thompson, P. M. Mace, W. L. Gabriel, L. L. Low. A. D. MacCall, R. D. Methot, J. E. Powers, B. L. Taylor, P. R. Wade, and J. F. Witzig. 1998. Technical guidance on the use of precautionary approaches to implementing National Standard 1 of the Magnuson-Stevens Fishery Conservation and Manage- ment Act. NOAA Technical Memo. NMFS-F/SPO-31, 54 p. National Technical information Center, 5825 Port Royal Road, Springfield, VA 22161. Rose, K. A., J. H. Cowan, K. O. Winemiller, R. A. Myers, and R. Hilborn. 2001. Compensatory density dependence in fish popu- lations: importance, controversy, understanding and prognosis. Fish Fish. 2:293-327. Sadovy, Y., and A-M. Eklund 1999. Synopsis of biological data on the Nassau grouper, Epinephelus striotiis iBloch, 1792), and the jewfish, E. itajara (Lichtenstein, 1822). NOAA Tech. Report NMFS 146, 65 p. Wolfson, L. J., J. B. Kadane, and M. J. Small. 1996. Bayesian environmental policy decisions: two case studies. Ecol. Appl. 6:1056-1066. Appendix 1: reparameterized spawner-recruit relationships The number of young fish recruiting to a population (R) is often relateci to the aggregate fecundity of the spawn- ing stock (S) by using one of two functional forms: fl = abS Ricker Beverton and Holt (A.l) b + S The parameter a is the slope of the curve at the origin and the parameter b controls the degree of density dependence. Notice that the domain of both functions extends from zero to infinity, whereas in practice there must be some limitation on S and R even in the absence of fishing owing to environmental constraints (call them jSq and i?0' respectively). This being so, we obtain 12.- ,*s„ l-i-So/6. Ricker Beverton and Holt (A.2) The ratio S„/Ry represents the maximum expected life- time fecundity of each recruit and a represents the sur- vival of recruits in the absence of density dependence. Accordingly, the product o = qSq/Rq may be interpreted as the maximum possible number of recruits produced by each spawner over its lifetime (Myers et al., 1999). The dimensionless character of a makes it useful for interspecies comparisons, or for borrowing values from species with similar life history strategies. Solving for b in terms of a one obtains log,, a I Sf, Ricker Sy/(l-a). Beverton and Holt Substituting Equation A. 3 into Equation A.l gives (A.3) R- aSa -s/s„ aS, 0 Ricker Beverton and Holt (A.4) l-Ka-l)S/S„ Porch et al. A catch free assessment model with application to Epinephelus ita/aia 101 and, since a = a/J^/Sy, R- ^^^A«l-S«o Rn aS/Sn 'l + (a-l)S/S„ Ricker Beverton and Holt (A. 5) Dividing through by i?y and defining s as S/Sq gives Equation 4. Appendix 2: formula for equilibrium spawning biomass The spawning potential ratio (p) is defined as the number of spawners produced by each recruit at equilibrium with a given fishing mortality rate F divided by the number of spawners per recruit under virgin conditions (F-O). This may be written Vo Sq / /?o ^f ^ ^0 (A.6) where the tilde signifies equilibrium values. At equilibrium we also obtain from Equation 4 sa^-' as (l + s(a-l)) Ricker Beverton and Holt (A.7) Dividing both sides of Equation A.7 by r, substituting p for Equation A.6, and solving for s gives Equation 10. 102 Abstract — Fish assemblages were investigated in tidal-creek and sea- grass habitats in the Suwannee River estuary, Florida. A total of 91.571 fish representing 43 families were collected in monthly seine samples from January 1997 to December 1999. Tidal creeks supported greater densities of fish (3.89 fish/m-; 83% of total) than did seagrass habitats (0.93 fish/m- 1. We identified three distinct fish assemblages in each habitat: winter-spring, summer, and fall. Pinfish iLagodon rhomboides), pigfish iOrthopristis chrysoptera I, and syngnathids characterized seagrass assemblages, whereas spot fLeiosto- rnus xanthurus), bay anchovy (Anchoa mitchiUi), silversides (Menidia spp.), mojarras iEucinostornus spp.), and fundulids characterized tidal-creek habitats. Important recreational and commercial species such as striped mullet (Mugil cephahis) and red drum (Sciaeiiops ocellatus) were found primarily in tidal creeks and were among the top 13 taxa in the fish assemblages found in the tidal-creek habitats. Tidal-creek and seagrass habitats in the Suwannee River estu- ary were found to support diverse fish assemblages. Seasonal patterns in occurrence, which were found to be associated with recruitment of early- life-history stages, were observed for many of the fish species. Fish assemblages found in tidal-creek and seagrass habitats in the Suwannee River estuary Troy D. Tuckey Florida Fish and Wildlife Conservation Commission Flonda Wildlife Research Institute Apalachicola Field Laboratory East Point, Florida 32328 Present address: School of Marine Science Virginia Institute of Marine Science College of William and Mary PO. Box 1346, Route 1208 Create Road Gloucester Point, Virginia, 23062 E-mail address, tuckeytg'vims edu Mark Dehaven Department of Agnculture and Consumer Services Division of Aquaculture 11350 SW 153^'^ Ct Cedar Key Florida 32625 Manuscript submitted 20 July 2004 to the Scientific Editor's Office. Manuscript approved for publication 19 July 2005 by the Scientific Editor. Fish. Bull. 104:102-117 12006). The Suwannee River estuary, located on the gulf coast of Florida, is rela- tively pristine and supports commer- cial and recreational fisheries. It is an unusual estuary, with an orientation along the open coastal shoreline, and its habitats include oyster bars, mud- flats, seagrasses, tidal creeks, and an extensive salt marsh (Comp and Seaman, 1985). In other estuaries of the United States, fish assemblages, species abundance, and habitat asso- ciations within estuaries have been studied extensively. Particular atten- tion has focused on estuaries as nurs- ery habitats for young-of-the-year (YOY) fishes that use seagrasses, tidal-creeks, and marshes during their early-life stages (Shenker and Dean, 1979; Bozeman and Dean, 1980; Liv- ingston, 1984; Cowan and Birdsong, 1985; Gilmore, 1988; McGovern and Wenner, 1990; Baltz et al., 1993; Peterson and Turner, 1994; Rooker et al., 1998). In addition, comparisons between different habitats within estu- arine systems have been conducted to evaluate the value of each habitat as a nursery (Weinstein and Brooks, 1983; Sogard and Able, 1991; Rozas and Minello, 1998; Paperno et al., 2001). Aside from basic species-composition studies of marsh fishes (Kilby, 1955; Nordlie, 2000) and fishes that inhabit shallow waters near Cedar Key (Reid, 1954), only one recent study (Tsou and Matheson, 2002) has investigated the distribution patterns of fishes in the Suwannee River estuary. Tsou and Matheson (2002) found that the nekton community structure for the Suwannee River estuary had a strong seasonal pattern that was consistent among years and followed patterns for water temperature and river dis- charge. They found assemblages that were associated with warm and cold seasons, and wet and dry seasons, but they did not examine habitat spe- cific assemblages (Tsou and Matheson, 2002). For proper management of fish- ery resources, it is beneficial to have detailed, current information concern- ing the status of all life-history stages and associated habitats of species that reside in the area, as well as informa- tion concerning species interactions and associated food webs. Although human development in the Suwannee River estuary is not a current threat, the potential withdrawal of freshwater from the Suwannee River for human consumption is a possibility and could impact fish assemblages found in the estuary (Tsou and Matheson, 2002). This article describes habitat-spe- cific assemblages by examining the fish fauna collected in seagrass habi- Tuckey and Dehaven: Fish assemblages found in tidal creek and seagrass habitats in the Suwannee River estuary 103 tats with those collected in tidal-creek habitats in the Suwannee River estuary. We performed com- parisons of monthly collections offish found along tidal-creek shorelines and those found in seagrass habitats to define fish assemblages and incor- porated abiotic parameters as potential factors influencing the assemblages. We also compared length distributions of species in each habitat to examine the success of YOY recruitment and the subsequent influence of YOY recruitment on fish assemblages in order to understand the nursery function of each habitat. Methods Study location This study took place in the Suwannee River estu- ary, which lies along the gulf coast of Florida, extending from just north of the Suwannee River to Cedar Key (Fig. 1). The Suwannee River emp- ties directly into the Gulf of Mexico forming an unusual open estuary that stretches 13 kilometers north of the river mouth, southeastward to the islands of Cedar Key, and extends approximately 8 kilometers offshore (Leadon'). The Suwannee River estuary is shallow (water depth <2.2 m below mean sea level), and has semidiurnal tides with a tidal range of 0.7 m. The shoreline is relatively undeveloped; the city of Cedar Key (pop. 898) along the southeastern edge of the estuary and the small town of Suwannee approximately 4.8 kilometers inland along the Suwannee River are the only populated areas. The remainder of the coastline, consisting of the Lower Suwannee and Cedar Keys National Wildlife Refuges and the Cedar Key State Preserve, is owned by the public. Study design Randomly selected sites were sampled monthly within tidal-creek and seagrass habitats from January 1997 through December 1999. Juvenile and small adult fish from each site were collected using a 21.3-mxl.8-m nylon seine with 3.2-mm mesh and a center bag measur- ing 1.8-mx 1.8-mx 1.8-m. Sampling methods depended on the habitat sampled, and all seines were deployed during daylight hours. Collections in tidal creeks consisted of six hauls per month in 1997 and increased to nine hauls per month in 1998 and 1999. Tidal creeks consisted of soft mud, deep channels, oyster bars, and steep banks. Shoreline vegetation included saltmarsh cord grass {Spartina al- ri tt.jt Suwannee River Suwannee , ■- River estuary -29 20'N ^i> jfe .^ y. ' '•••. • ^4 '( "k ?:t ^ Gulf •? ' >^ c^f; of • * * ^C ^ Mexico • " '"**m.j. / ■29 10'N • • .'^!fb^^ ^^^^ North • • •■ 44i^. Key J?_^, ,• i? .• ♦ • •• *^ Land * >*••• •♦ A Tidal-creek seine hauls \f • Seagrass seme hauls 0 15 3 6 9 12 83-W Leadon, C. J. 1979. Unpubl. manuscr. Environmental effects of river flows and levels in the Suwannee River sub- basin below Wilcox and the Suwannee River estuary, Florida, 59 p. Suwannee River Water Management District Interim Report, 922.5 County Road 49, Live Oak, FL 32060. Figure 1 udy area showing location of tidal-creek and seagrass seine hauls in e Suwannee River estuary, located on the gulf coast of Florida. terniflora) and needle rush (Juncus roemerianus) near the creek mouths and changed to a variety of freshwa- ter marsh grasses and terrestrial vegetation upstream. The seine was set from a boat in a semicircular pattern along the shoreline, retrieved onshore, and sampled an average area of 68 m- per haul. Shoreline areas inun- dated with vegetation were not sampled if the water depth was greater than 0.5 m in order to reduce the interference of vegetation during sample collections. Sampling in tidal-creeks was limited to the shoreline because the water was too deep in the channels to deploy the seine. In addition, despite the importance of oyster bar habitat, oyster bars located inside tidal creeks were not sampled because they interfered with the proper deployment of the seine. Seagrass habitats generally surrounded the major is- lands near Cedar Key, including North Key and nearby islands (Fig. 1). In addition, vegetated patches extended from North Key northwestward to the mouth of the Su- wannee River and were present in shallow areas ap- proximately three kilometers west of the Suwannee River (Fig. 1). Dominant seagrass species in the Suwannee River estuary included turtle grass (Thalassia testudi- 104 Fishery Bulletin 104(1) num.), manatee grass {Syringodium filliforme), and shoal gprass (Halodule ivrightii). Percent coverage was estimat- ed visually, or if water clarity was insufficient to visually inspect the bottom, bottom samples were collected at 3-m intervals during deployment of the seine. For ana- lytical purposes, areas sampled had to contain at least ten percent seagrass to he considered seagrass habitat. Samples were classified as "vegetated" or "unvegetated" after sampling; and monthly collections varied from one to seven hauls per month depending on seagrass cov- erage. Seagrass habitats were sampled by pulling the seine into the current or wind, whichever was strongest. We kept the opening of the net at a constant width by maintaining tension on a 15.5-m line that was attached between each end pole of the seine while the seine was hauled for a distance of 9.1 m. The distance covered by the seine was measured by a weighted line from the starting point. The net was retrieved by bringing the end poles together and pulling the net at an angle around a vertical pole that closed the wings of the net and forced the catch into the bag. A typical seine haul over seagrass habitat covered approximately 140 m-. In the field, all fish were identified to the lowest pos- sible taxon. counted, and released. Up to 40 individuals of species of special interest (important to the commer- cial or recreational fishery) and 10 individuals of all other species were measured to the nearest millimeter standard length (SL). For quality-control purposes, three specimens of each species collected were returned to the laboratory so that species identification could be confirmed. At each site, Secchi depth and water depth were measured, and water temperature (°C), salinity, dissolved oxygen level (mg/L), and pH were measured by using a Hydrolab Surveyors'- water-quality instru- ment (Hach Environmental, Loveland, CO). Data analysis Multivariate analyses were used to compare fish com- munity structures along tidal-creek shorelines to those found in seagrass habitats (Field et al., 1982). Average monthly abundance estimates (number of fish divided by the number of hauls) were calculated separately for each fish species in each habitat type. Average monthly abundance estimates were then converted to percent composition to correct for bias introduced by the two different net-deployment methods and for the different levels of effort in each habitat. Fishes that were not identified to species were eliminated (<0.1% of total fish collected) except where species complexes, such as silversides (Menidia spp.), mojarras (Eucinostomus spp. <50 mm SL), menhaden {Brevoortia spp.), and minnows {Notropis spp.) were substituted. Species complexes were used when meristic characters for juveniles were insufficient to distinguish between two or more pos- sible species (Eucinostomus spp. and Notropis spp.) or where there was possible hybridization (Menidia spp. and Brevoortia spp.). Fish-community comparisons based on percent spe- cies composition by habitat and month were conducted by using algorithms in PRIMER (Plymouth Routines in Multivariate Ecological Research, vers. 5, Plymouth Marine Laboratory, UK) for the study of community structure (Clarke and Warwick, 1994). To identify fish assemblages, hierarchical agglomerative cluster analysis was performed with the Bray-Curtis similar- ity matrix calculated on fourth-root transformed per- centage data. The fourth-root transformation reduced the dominance of abundant species and increased the influence of less abundant species in the community analysis. The cluster analysis was run on one matrix consisting of all transformed fish abundance esti- mates collected in tidal-creek and seagrass habitats combined. To identify species that were responsible for the pat- terns observed in the cluster analysis, similarity and dissimilarity percentage breakdowns were conducted by using the SIMPER procedure in PRIMER (Clarke, 1993). Average similarities between assemblages were analyzed to determine the contribution of each species to the overall similarity. This procedure reduced the number of species required to explain the patterns observed in the cluster diagram and allowed for a sim- plified interpretation of the species assemblages. The higher the similarity value, the more alike samples were within assemblages. Alternatively, dissimilarity values were examined to identify species that were characteristic of a particular assemblage. Species that have high average dissimilarity values and low stan- dard deviations are those that contribute consistently to samples within their group, with the result that they can be used to distinguish between groups. The relationship between environmental variables and fish community structure was examined by us- ing the BIO-ENV procedure. A Spearman rank cor- relation test was used to compare ranked values from the aforementioned biota similarity matrix to ranked values from an environmental similarity ma- trix, which was created from environmental variables measured in this study. Comparisons were based on normalized Euclidean distance. The environmental variables used to create the environmental similar- ity matrix included pH, water temperature, salinity, water depth, and Secchi depth. Dissolved oxygen was strongly correlated with water temperature and was therefore not included in the analysis because it would produce results similar to those produced by water temperature. Abiotic variables were standardized by subtracting each mean and dividing by the standard error to remove any bias associated with the different measurement scales. The influence of recruitment of YOY fishes in defining fish assemblages was examined. By relating increases in abundance of species that were identified to be im- portant contributors through the SIMPER procedure to decreases in their average length in both habitats, we were able to identify the timing of juvenile recruitment. Length-frequency distributions showed that the seine continued to catch larger individuals and therefore the decrease in average length was not due to a decrease Tuckey and Dehaven: Fish assemblages found in tidal-creek and seagrass habitats in the Suwannee River estuary 105 -1 — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — 1— Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov 1997 1998 1999 Figure 2 Monthly mean values (± standard error) for water temperature, dissolved oxygen, and salinity from January 1997 to December 1999. Circles represent values measured in seagrass habitats; triangles represent values measured in tidal-creek habitats. in the number of larger fish. The mean length of each species in each sample was calculated by habitat, and differences in mean length were tested by using fish assemblages as the main factor in a one-way ANOVA. Multiple comparisons tests (Tukey's test) were then conducted to examine significant ANOVA results and these tests determined which assemblages contained fishes with significantly different lengths. Results Environmental conditions The combination of water temperature, salinity, and water depth had the highest correlation (p,^=0.659) with the fish assemblages of any possible combination of measured abiotic variables. Seasonal patterns were observed for water temperature and dissolved oxygen, whereas salinity fluctuated in seagrass habitats and generally increased during 1999 in tidal-creek habitats (Fig. 2). Water temperatures ranged from 7.5° to 32.5°C (mean=23.0°C, SE = 0.99) in the seagrass habitats and ranged from 10.3° to 33.3°C (mean=23.1°C, SE = 0.91) in the tidal-creek habitats. Minimum values of dissolved oxygen coincided with the highest water temperatures in each habitat and ranged from 2.9 to 12.7 mg/L in the seagrass habitats and from 2.9 to 13.6 mg/L in the tidal- creek habitats. Salinity, however, was lower during all seasons in the tidal-creek habitats than in the seagrass habitats (Fig. 2). Mean salinity in the seagrass habitats was 27.1%f (SE = 0.91) and ranged from nearly fresh (1.3^?^) to marine (34.8%f ) depending on river discharge, whereas in the tidal creeks, mean salinity was 9.5%c (SE = 0.81) and ranged from 0.0%t. to 29. 0%^. An unusu- ally high period of rain during February and March of 1998 decreased salinity values in the tidal-creek and seagrass habitats. 106 Fishery Bulletin 104(1) Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov Jan Mar May Jul Sep Nov 1997 1998 1999 Figure 3 Cumulative number of species collected in both tidal-creek and seagrass habitats (open circles) and number of species collected in tidal-creeks (trianglesi and seagrass habitats (filled circles) by year and month from January 1997 to December 1999. Fish fauna At the conclusion of the three-year study, 111 fish species and 4 additional species complexes had been collected. During the first year of the study, 61 species were col- lected in samples taken in tidal-creek habitats, and 48 species were collected in seagrass habitats (Fig. 3). Thirteen new species were added to the species list from tidal creek samples and 20 new species were added to the list from samples taken in seagrass habitats during 1998. During the final year of sampling only six addi- tional species were collected in tidal creeks and 12 new species were collected in seagrass habitats. Overall, tidal creeks contained greater relative (uncorrected for gear efficiency) densities offish (3.89 fish/m'-) compared with seagrass habitats (0.93 fish per m-). In seagrass habi- tats, 15,395 individuals were collected in 118 samples that covered approximately 16,520 m- of seagrass habi- tat (Table 1). In tidal-creek habitats, a total of 76,176 individuals were collected from 288 samples that covered approximately 19,584 m- of tidal-creek shoreline habitat (Table 2). Thirty five species were restricted to seagrass habitats, and another 35 species were collected only in tidal creeks. The remaining 45 species were collected in both habitats at least once during the study. Overall, twelve families were restricted to seagrass habitats: phycid hakes (Phycidae), toadfishes (Batrachiodidae), batfishes (Ogcocephalidae), flyingfishes (Exocoetidae), cardinalfishes (Apogonidae), barracudas (Sphyraenidae), wrasses (Labridae), combtooth blennies (Blenniidae), mackerels (Scombridae), triggerfishes (Balistidae), box- fishes (Ostraciidae), and porcupinefishes (Diodontidae; Table 1). Seven families were restricted to tidal-creek habitats: minnows (Cyprinidae), sunfishes (Centrarchi- dae), killifishes (Cyprinodontidae), gars (Lepisosteidae), eagle rays (Myliobatidae), pikes (Esocidae), and livebear- ers (Poeciliidae; Table 2). Fish assemblages A clear separation of fish assemblages was identified and indicated by two main branches that corresponded to fishes found in seagrass habitats and those found in tidal-creek habitats (Fig. 4). There were two months iden- tified from seagrass samples (January 1997 and March 1998) that had a species composition that was more closely linked to samples taken from tidal creeks. Seagrass habitats Seasonal fish assemblages in seagrass habitats were evi- dent during all three years of the study, which included winter-spring, summer, and fall assemblages (Fig. 4). The winter-spring assemblage consisted principally of pinfish {Lagodon rhomboides). pigfish (Orthopris- tis chrysoptera). dusky pipefish (Syngnathus floridae), southern puffer (Sphoeroides nepheliis), and gulf pipe- fish [Syngnathus scovelU). which together accounted for more than 95% of the cumulative percent similarity. The summer assemblage had a higher average similarity value (43.67) than did the winter-spring assemblage (42.56) and consisted of more than 21 species. Eleven of these species — silver perch (Bairdiella chrysoura), S. floridae, bay anchovy {Anchoa mitchilli), L. rhom- boides, Eucinostomus spp., spotted seatrout (Cynoscion nebulosus), S. scovelli, planehead filefish (Monacanthiis hispidus), striped anchovy {Anchoa hepsetus). inshore liz- ardfish {Synodus foetens), and O. chrysoptera — accounted for more than 75% of the cumulative similarity of the summer assemblage. The fall assemblage had the high- Tuckey and Dehaven Fish assemblages found in tidal creek and seagiass habitats in the Suwannee Rivei estuary 107 Table 1 Taxonomic list of individuals collected in seagrass habitats for each species by month and total number collected, a 11 years combined. Family Species Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total Dasyatidae Dasyatis sabina 0 0 0 0 0 0 1 0 1 3 1 1 7 Dasyatis say 0 0 0 0 0 2 0 0 0 0 0 0 2 Ophichthidae Myrophis punctatus 0 0 0 1 0 0 0 0 0 0 0 0 1 Clupeidae Harengula jaguana 0 0 0 0 0 0 6 230 0 162 0 0 398 Bi-evoortia spp. 0 0 1 0 0 0 0 0 0 0 0 0 1 Opisthonema ogliniini 0 C) 0 0 0 4 1 4 35 9 0 0 53 Sardinella aurita 0 u 0 0 0 0 0 16 0 0 0 0 16 Engraulidae Anchoa hepsetus 0 0 0 0 1 540 2 1 235 13 3 0 795 Anchoa mitchilli 0 0 4 3 0 38 1567 3838 632 3504 53 0 9639 Ariidae Arius felis 0 0 0 0 0 1 1 14 45 1 0 0 62 Bagre marinus 0 0 0 0 0 0 0 0 1 0 0 0 1 Synodontidae Synod us foetens 0 0 0 0 1 2 3 8 3 1 1 1 20 Gadidae Urophycis floridana 0 0 0 0 0 0 0 0 0 0 0 6 6 Batrachiodidae Opsanus beta 0 0 0 0 0 0 0 0 1 0 0 0 1 Ogcocephalidae Ogcocephalus radiatus 0 0 0 1 0 0 0 0 0 0 0 0 1 Exocoetidae Hyporhamphus meeki 0 0 0 0 1 7 2 0 0 0 0 0 10 Belonidae Strongylura marina 0 0 0 0 0 1 3 1 0 0 0 0 5 Atherinidae Membras niartinica 0 10 0 0 1 10 49 21 13 15 3 0 122 Menidia spp. 0 0 0 0 0 0 0 0 0 0 1 2 3 Syngnathidae Hippocampus erectus 0 0 0 0 0 0 0 0 0 1 0 0 1 Hippocampus zosterac 0 1 0 0 0 0 1 1 2 0 1 2 8 Syngnathus floridae 2 7 1 9 2 34 33 53 48 26 52 38 305 Syngnathus louisianae 0 0 0 1 0 3 4 2 1 4 3 0 18 Syngnathus scoveHi 0 3 6 4 0 11 32 4 9 10 13 13 105 Serranidae Mycteroperca microlepis 0 0 0 0 0 0 0 0 0 1 0 0 1 Centropristis striata 0 0 0 0 0 1 2 7 12 3 74 12 111 Serraniculus pumilio 0 0 0 0 0 0 0 0 0 0 10 0 10 Serranus subligarius 0 0 0 0 0 0 0 0 2 0 0 0 2 Apogonidae Astrapogon alutus 0 0 0 0 0 0 0 0 0 1 0 0 1 Carangidae Caranx hippos 0 0 0 0 0 0 0 0 0 2 0 0 2 Chloroscombrus chrysurus 0 0 0 0 0 0 2 12 5 0 0 0 19 Oligoplites saurus 0 0 0 0 0 7 2 2 4 2 0 0 17 Selene vomer 0 0 0 0 0 0 0 0 0 2 0 0 2 Lutjanidae Lutjanus griseus 0 0 0 0 0 0 0 1 0 1 0 0 2 Lutjanus synagris 0 0 0 0 0 0 3 0 0 1 2 0 6 Gerreidae Eucinostomus gula 0 0 0 0 0 0 0 0 6 6 4 1 17 Eucinostomus harengulus 0 0 0 0 0 1 0 2 1 0 0 0 4 Eucinostomus spp. 0 0 0 0 0 34 172 162 73 13 54 36 544 Haemulidae Haemulon plumieri 0 0 0 0 0 0 1 6 11 4 3 7 32 Orthopristis chrysoptera 1 0 6 26 738 30 38 0 8 4 0 1 852 Sparidae Diplodus holbrooki 0 0 0 0 7 17 0 4 1 7 0 0 36 Lagodon rhomboides 17 65 63 87 21 54 24 27 40 36 32 87 553 Sciaenidae Bairdiella chrysoura 0 0 0 3 24 159 273 59 184 44 2 1 749 Cynoscion arenarius 0 0 0 0 0 0 1 24 0 0 0 0 25 Cynoscion nebulosus 0 0 0 0 0 4 70 37 10 6 1 0 128 Leiostom u s xa n th u ru s 2 34 1 0 0 0 0 0 2 0 0 0 39 Menticirrhus a m erica nus 0 0 0 0 0 2 13 119 2 0 1 0 137 continued 108 Fishery Bulletin 104(1) Table 1 (continued) Family Species Jan Feb Mar Apr Maj Jun Jul Aug Sep Oct Nov Dec Total Sciaenidae Menticirrhus saxatilis 0 0 1 0 0 7 0 5 0 0 0 2 15 ifont.i Sciaenops ocellatus 0 0 0 0 0 0 0 0 0 0 9 0 9 Ephipipidae Chaetodipterus faber 0 0 0 0 0 0 2 12 2 0 0 0 16 Mugilidae Mugil cephalus 0 5 1 0 0 0 0 0 0 0 0 0 6 Mugil curema 0 0 0 1 0 0 0 0 0 0 0 0 1 Sphyraenidae Sphyraena borealif: 0 0 0 0 4 0 0 0 0 0 0 0 4 Labridae Halichoeres bivittatus 0 0 0 0 0 0 0 3 0 0 0 0 3 Lachnolaimus maximus 0 0 0 0 0 0 0 0 2 1 0 0 3 Blennhdae Cliasmodes saburrae 0 0 0 0 0 0 9 0 5 0 2 0 16 Parablennius mannoreus 0 0 0 0 0 0 0 1 0 0 1 0 2 Hypsoblennius hentzi 0 0 0 0 0 0 0 0 1 0 0 2 3 Hypleurochilus geminatus 0 0 0 0 0 0 2 0 0 0 0 0 2 Gobiidae Gobionellus boleo.'ioma 0 0 0 0 0 0 9 0 0 0 0 0 9 Gobiosoma bosc 0 0 0 0 0 0 29 0 0 0 1 0 30 Gobiosoma longipala 0 0 0 0 0 0 0 0 0 1 0 0 1 Gobiosoma i-obiiatuni 0 0 0 2 0 0 7 2 0 1 0 1 13 Microgobius gulnsiiti 0 0 0 0 0 0 3 2 0 0 2 0 7 Microgohiu s thala usin us 0 0 0 0 0 0 0 77 0 0 0 0 77 Scombridae Scomberomorus maculatu ■; 0 0 0 0 0 0 0 0 1 0 0 0 1 Triglidae Prionotus scitulus 0 2 0 0 0 2 5 4 2 1 0 1 17 Prionotus tribulus 0 0 0 0 0 0 0 2 0 0 1 0 3 Bothidae Paralichthys albigutta 0 0 2 2 0 0 2 0 2 0 0 0 8 Etropus crossotus 0 3 1 0 0 0 2 1 1 0 0 1 9 Etropus microstomus 0 0 1 0 0 0 0 0 0 0 0 0 1 Cynoglossidae Symphiiriis plagiusa 0 0 1 0 0 1 1 9 0 0 0 1 13 Soleidae Achirus lineatus 0 0 0 0 0 0 1 7 0 0 0 0 8 Trinectes maculatus 0 0 0 0 0 0 0 0 1 0 0 0 1 Balistidae Aluterus schoepfi 0 0 0 0 0 1 1 2 0 1 0 0 5 Monacanthus citiatus 0 0 0 0 0 0 0 2 10 8 2 19 41 Monacanthus hispidus 0 0 0 0 2 10 1 36 3 9 7 25 93 Ostraciidae Lactophrys quadricornis 0 0 2 0 0 2 0 6 6 3 3 1 23 Tetraodontidae Sphoeroides nephelus 0 0 2 16 4 2 2 2 1 4 2 2 37 Diodontidae Chilomycterus schoepfi 0 1 2 0 4 2 0 11 5 10 5 9 49 Column total 22 131 95 156 810 989 2382 4839 1429 3921 349 272 15,395 Number of hauls 6 10 9 4 5 15 10 11 11 11 10 16 118 est average similarity level at 48.90, and nine species — S. floridae, M. hispidus, black seabass (Centropristis striata), L. rhomhoides, Eucinostomus spp., fringed file- fish {Monacanthus ciUatus), S. scovelli, striped burrfish (Chilomycterus schoepfi), and S. nephelus — accounted for more than 91% of the cumulative similarity. Lagodon rhomhoides and S. floridae were characteristic of all three assemblages, and Eucinostomus spp., S. scovelli, and M. hispidus were important contributors to the summer and fall assemblages. Other abundant species from seagrass assemblages included southern kingfish {Menticirrhus americanus), rough silverside {Membras niartinica), and gobiids, particularly green goby {Micro- gobius thalassinus) and naked goby (Gobiosoma bosc). A significant drop in salinity during March 1998 corresponded to an alteration in the fish community collected in seagrass habitats. The species present in the seagrass habitats during March 1998 were more consistent with species collected in tidal creeks and included A. mitchilli, Brevoortia spp., and spot (Leiosto- mus xanthurus), none of which were collected in March of 1997 or 1999 in seagrass habitats. Samples collected during March 1997 and 1999 contained individuals more characteristic of seagrass habitats, such as C. schoepfi, L. rhomhoides, O. chrysoptera, and iS. nephelus in 1997 and scrawled cowfish (Acanthostracion quadri- cornis), O. chrysoptera, L. rhomhoides, S. floridae, and S. scovelli in 1999. Tuckey and Dehaven Fish assemblages found In tidal-creek and seagrass habitats in the Suwannee River estuary 109 Table 2 Taxonomic list of individuals collected in tidal-creeks for each species by month and total number collected all years combined. Family Species Jan Feb Mar Apr Ma> Jun Jul Aug Sep Oct Nov Dec Total Dasyatidae Dasyatis sabina 0 0 6 1 1 0 0 0 1 1 3 0 13 Myliobatidae Rhinoptera bonasus 0 0 0 0 3 0 0 0 0 0 0 0 3 Lepisosteidae Lepisosteus osseus 0 0 0 0 0 2 2 1 0 0 0 0 5 Lepisosteus platyrhincus 0 0 1 1 0 1 1 0 2 2 0 0 8 Ophichthidae Myrophis punctatus 0 0 0 0 1 0 0 0 0 0 0 0 1 Clupeidae Harengula jaguana 0 0 0 0 2 1 2 9 10 2 0 0 26 Brevoortia spp. 15 139 93 400 394 7 5 0 0 0 0 1 1054 Sardinella aurita 0 0 0 0 0 0 22 0 0 0 0 0 22 Engraulidae Anchoa hepsetus 3 0 0 1 175 482 71 33 7 0 3 0 775 Anchoa mitchilli 86 497 825 108 612 329810,329 6970 5776 7436 1637 2107 39,681 Cyprinidae Notemigoniis crysoleucas 0 2 1 0 0 0 0 0 0 0 0 0 3 Notropis spp. 0 0 0 0 4 9 0 0 0 0 0 0 13 Ariidae Arius felis 0 0 0 0 0 0 0 2 3 0 0 0 5 Esocidae Esox niger 0 0 0 1 1 0 0 0 0 0 0 0 2 Synodontidae Synod us foe tens 0 0 0 3 7 0 0 0 1 0 3 1 15 Belonidae Strongylura marina 0 0 0 2 4 1 8 1 1 1 0 0 18 Strongylura notata 2 0 0 0 0 0 3 2 8 0 1 0 16 Strongylura timucu 0 0 0 0 11 1 8 3 11 0 0 0 34 Cyprinodontidae Adinia xenica 7 7 0 4 1 0 12 23 0 153 14 25 246 Cyprinodon variegatus 4 0 0 1 0 0 0 0 0 0 2 3 10 Lucania goodei 0 0 11 0 0 0 0 0 0 0 0 0 11 Lucania parva 0 2 2 0 0 1 2 15 4 0 0 6 32 Fundulidae Fundulus confluentus 2 0 0 0 0 0 0 0 0 1 0 3 6 Fundulus grandis 33 39 6 28 9 30 21 127 9 39 170 218 729 Fundulus majalis 29 12 24 7 0 18 0 42 3 10 13 65 223 Fundulus seminolis 7 2 0 1 0 14 0 14 5 6 0 1 50 Poeciliidae Gambusia holbrooki 6 0 59 2 4 1 0 1 12 4 1 7 97 Heterandria formosa 0 1 0 0 0 0 0 0 0 1 0 0 2 Poecilia latipinna 1 0 0 0 2 2 0 17 0 4 4 12 42 Atherinidae Membras martinica 0 0 4 0 9 566 2286 28 114 7 1 0 3015 Menidia spp. 383 429 201 265 145 695 982 571 814 602 749 358 6194 Syngnathidae Syngnathus floridae 0 1 0 0 0 0 0 0 1 0 0 0 2 Syngnathus louisianae 0 0 0 0 0 0 0 0 1 0 1 0 2 Syngnathus scovelli 2 1 2 0 1 0 1 7 5 1 2 11 33 Serranidae Diplectrum hivittatum 0 0 0 0 0 0 0 0 0 0 1 0 1 Centrarchidae Enneacanthus gloriosus 0 0 2 0 0 0 0 0 0 0 0 0 2 Elassoma zonatum 0 0 0 0 4 0 0 0 0 0 0 0 4 Lepomis gulosus 0 1 1 0 0 0 0 0 0 0 0 0 2 Lepomis niacrochirus 0 0 0 0 0 0 0 0 1 0 0 0 1 Lepomis marginatus 0 0 0 1 0 0 0 0 0 0 0 0 1 Lepomis microlophus 0 0 0 0 1 0 0 0 0 0 0 0 1 Lepomis punctatus 1 0 0 1 1 1 2 2 3 2 0 0 13 Micropterus salmoides 0 0 3 1 14 4 0 5 8 2 0 0 37 Carangidae Chloroscombrus chrysurus 0 0 0 0 0 0 1 3 15 2 0 0 21 Otigoplites saurus 0 0 0 0 0 8 53 25 47 17 6 0 156 Trachinotus falcatus 0 0 0 0 0 0 0 0 0 3 0 0 3 continued 110 Fishery Bulletin 104(1) Table 2 (continued) Family Species Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total Lutjanidae Lutjanus griseus 0 0 0 0 0 2 2 9 8 6 9 1 30 Gerreidae Eucinostoinus giila 1 0 0 0 0 0 0 4 7 25 49 3 89 Eucmostomus harengulu^ 0 25 5 2 6 3 41 150 88 100 65 45 530 Eucinostoinus spp. 17 38 6 1 1 87 1035 1082 503 862 691 358 4681 Haemulidae Orth opristis ch rysoptera 0 0 0 0 0 0 0 2 0 0 0 0 2 Sparidae Archosargus probatocephatus 7 1 1 0 1 1 0 1 0 0 1 0 13 Lagodon rhomboides 87 231 146 142 67 114 111 85 38 29 53 10 1113 Sciaenidae Bairdiella chrysoura 0 0 7 7 200 154 58 246 117 4 3 0 796 Cynoscion arenarius 0 0 0 5 236 102 90 91 46 76 5 0 651 Cynoscion nebulosus 3 1 1 0 5 11 21 53 62 23 36 2 218 Leiostomus xanthurus 9526 2044 834 771 274 42 125 24 6 5 2 31 13,684 Menticirrhus americanus 0 0 0 0 3 12 50 36 12 10 5 0 128 Micropogonias undulatiis 4 4 0 5 2 0 0 0 1 0 6 2 24 Pogonias cromis 0 0 0 0 2 2 0 3 0 1 0 1 9 Sciaenops ocellatus 26 28 20 15 3 7 4 1 1 48 105 65 323 Ephipipidae Chaetodipterus faber 0 0 0 0 0 1 3 1 3 0 0 0 8 Mugilidae Mugil cephalus 175 159 59 25 2 39 22 7 18 1 6 6 519 Mugil curema 1 0 0 0 0 25 2 0 0 0 0 1 29 Mugil gyrans 0 0 0 0 0 0 0 0 2 0 0 4 6 Gobiidae Bathygobius soporator 0 0 0 4 0 0 2 1 18 5 6 3 39 Gobionellus boleosoma 0 0 3 6 9 0 1 6 15 0 4 0 44 Gobiosonia bosc 8 3 14 7 1 40 21 17 52 20 18 67 268 Gobiosonia robustum 0 6 3 1 0 8 7 10 6 0 4 15 60 Microgobius gulosus 5 23 7 5 4 7 0 25 5 8 3 14 106 Microgobius thalassinus 0 0 0 0 0 0 1 1 1 0 1 0 4 Triglidae Prionotus scitulus 0 0 0 0 0 1 0 0 0 0 0 0 1 Prionotus tribulus 0 0 0 1 0 0 0 0 1 11 4 0 17 Bothidae Paralichthys albigutta 5 1 0 0 0 2 0 0 1 0 0 0 9 Paralichthys lethostigma 0 0 1 0 0 1 0 0 0 0 0 0 2 Etropus crossotus 0 0 0 0 0 1 0 0 0 0 0 0 1 Cynoglossidae Symphurus plagiusa 6 0 0 0 0 2 2 1 13 5 1 5 35 Soleidae Achirus lineatus 0 0 0 0 0 2 4 9 10 8 2 0 35 Tnnectes maculatus 1 8 10 0 5 13 1 13 2 4 2 1 60 Tetraodontidae Sphoeroides nephelus 0 0 0 1 6 0 0 0 0 4 0 1 12 Column total 10,453 3705 2358 1826 2233 5821 1 5.414 9779 7898 9551 3685 3453 76,176 Number of hauls 24 24 24 24 24 24 24 24 24 24 24 24 288 Recruitment of YOY fishes had an influence on defining fish assemblages in seagrass habitats. The winter-spring assemblage was dominated by YOY L. rhomboides and O. chrysoptera (Table 3), which had significantly shorter standard lengths than did the other assemblages (Table 4). The summer assemblage showed an increase in the number of species and an in- crease in their abundance, but there were no significant differences in length for YOY for any species between assemblages. Tidal-creek habitats Three fish assemblages (winter-spring, summer, and fall) were identified from samples taken in tidal-creek habitats and reflected similar seasonal patterns com- pared with fish assemblages identified from seagrass habitats (Fig. 4). The winter-spring assemblage had an average similarity level of 51.76 and consisted of L. xanthurus, Menidla spp., A. initchilli, L. rhomboides, M. cephalus, and red drum (Sciaenops ocellatus). These six Tuckey and Dehaven Fish assemblages found in tidal creek and seagrass habitats in the Suwannee River estuary 111 Bray-Curtis similarity r- sg 12 Seagrass (fall) !g_i-' so 11 y ig.l : ig.11 sg.?. sg_6 sg_9 ?g 8 Seagrass !g_' (summer) sg 10 !g_9 59.6. sg_6. y S0_9. sg 10 sg-5. sg_5. SSJ. Seagrass sg_.'. so 3 (winter and spring) sg_4 sg_'. sg.i sg_i. sa.5. ^ =9.3. S9_1. so 3 r tc 6 tc 7 tc 6 tc 9 tc 6 tc To Tidal-creek 'c_11. (summer) tc 8 tc 9 tc 7 tc 9 tc e tc_10. > tc 12 tc a Tidal-creek Ic 12 (tall) Is 11 tc 11 r tc 1 tc 2 1c 1 lc.2. tc 3 Tidal-creek 1c_l_ (winter and spnng) tc 5 tc 4 tc 5 tc S Ic 3 Ic 4 V- tc 4 tc.S. Figure 4 Results of cluster analysis for fish assemblages found in the Suwannee River estuary. SG = seagrass habitats, TC = tidal-creek habitats, l = January, 2 = February, etc., and the year is denoted by the last two digits of the year. For example, SG_1_97 corresponds to samples collected during January of 1997 in seagrass habitats. Note: Clusters are free to rotate at the point at which the lines branch. species accounted for more than 68% of the cumulative similarity within the winter-spring assemblage. The summer assemblage had an average similarity level of 62.29 and was characterized by ten species that accounted for 70.65% of the cumulative similarity; A. mitchilli, Menidia spp., Eucinostomus spp., sand seatrout (Cynoscion arenarius), B. chrysoura, C. nebulosus, L. rhomboides, E. harengulus, M. martinica, and leather- jacket (Oligoplites saurus). The fall assemblage had an average similarity level of 60.36 and was characterized hy Eucinostomus spp., Menidia spp., gulf killifish (Fiin- dulus grandis), A. mitchilli, E. harengulus, diamond kil- lifish (Adinia xenica), clown goby (Microgobius gulosus). and M. cephalus, which accounted for more than 67% of the cumulative similarity. Species tolerant of low salin- ity, such as A. xenica, marsh killifish {Fundulus con- fluentus), mosquitofish (Gambusia holbrooki), and least killifish {Heterandria formosa) were commonly collected in tidal creeks. There were 35 species collected, includ- ing groups such as cyprinids, cyprinodontids, poeciliids, lepisosteids, and centrarchids, which were restricted entirely to tidal creeks (Table 2). Seasonal recruitment of juvenile fishes to tidal-creek habitats was evident and remained consistent through- out the study resulting in three clearly defined assem- blages. Young-of-the-year L. xanthurus recruited to 112 Fishery Bulletin 104(1) Table 3 Results of SIMPER procedure showing the average percent dissimilarity id'^i) for important species between seagrass assem- blages, di is the average contribution of the ith species to the disssimilarity between groups; bi/SD(di) is the ratio between the average contribution of the /'*' species and the standard deviation of 6i; Cum di'7c is the cumulative contribution to the total dis- similarity. Species are listed in decreasing contribution to average dissimilarity. Species Average abundance di (V/SD(cV) di^-i Cumulative di% Eucinosto??ius spp. winter-spring summer 4.8 1.38 6.31 6.31 0.24 9.9 Lagodon rho/nboides 10.75 1.83 4.5 1.69 5.91 12.22 Bairdiella chrysoura 1.01 12.76 4.38 1.25 5.75 17.97 Orth opristis ch rysoptera 26.58 0.94 3.66 1.17 4.81 22.78 Monacanthus hispidus 0.09 1.89 3.25 1.22 4.27 27.05 Syngnathus floridae 1.2 3.88 3.24 1.1 4.25 31.3 Centropristis striata 0.01 1.82 2.68 0.99 3.52 34.82 Anchoa hepsetus 0.03 11.95 2.67 0.8 3.51 38.33 Syngnathus scovelli 0.8 1.13 2.48 1.36 3.25 41.58 Sphoeroides nephelus 1.24 0.19 2.27 1.2 2.99 44.57 Anchoa mitchilli 0.33 4.46 2.27 0.95 2.98 47.55 Chilomycterus schoepfi 0.18 0.46 2.12 1.07 2.78 50.33 Anchoa niilchilli winter-spring fall 10.87 2.32 13.31 13.31 0.33 478.1 Lagodon rhomboides 10.75 4.57 6.43 1.36 7.87 21.18 Orth opristis ch rysoptera 26.58 1.33 4.73 1.05 5.79 26.97 Bairdiella chrysoura 1.01 8.26 3.39 1.58 4.15 31.12 Leiostomus xanthurus 0.67 0.14 3.26 0.61 3.99 35.11 Syngnathus scovelli 0.8 0.71 2.83 1.19 3.46 38.57 Syngnathus floridae 1.2 3.26 2.81 1.27 3.44 42.01 Eucinostomus spp. 0.24 5.38 2.68 1.24 3.28 45.29 Harengula jaguana 0 9.74 2.57 1.24 3.15 48.44 Sphoeroides nephelus 1.24 0.12 2.55 0.91 3.12 51.56 Anchoa mitchilli summer fall 6.13 1.98 9.30 9.30 4.46 478.1 Eucinostomus spp. 9.9 5.38 3.11 1.34 4.72 14.02 Bairdiella chrysoura 12.76 8.26 2.70 1.21 4.10 18.12 Anchoa hepsetus 11.95 0.48 2.36 0.98 3.59 21.71 Syngnathus floridae 3.88 3.26 2.35 1.05 3.56 25.27 Monacanthus hispidus 1.89 0.38 2.34 1.14 3.55 28.82 Harengula jaguana 2.79 9.74 2.15 1.25 3.27 32.09 Centropristis striata 1.82 0.64 2.06 1.01 3.13 35.22 Syngnathus scovelli 1.13 0.71 1.82 1.35 2.77 37.99 Lagodon rhomboides 1.83 4.57 1.75 1.08 2.65 40.64 Chilomycterus schoepfi 0.46 0.52 1.72 1.13 2.61 43.25 Cynoscion nebulosus 1.54 1.95 1.66 1.31 2.53 45.78 Monacanthus ciliatus 0.45 0.45 1.62 0.83 2.47 48.25 Orthopristis chrysoptera 0.94 1.33 1.62 1.04 2.46 50.71 Tuckey and Dehaven: Fish assemblages found in tidal creek and seagrass habitats in the Suwannee River estuary 113 Table 4 Results of ANOVA comparing PRIMER. * <0.05, **<0.01, **■ standard length <0.001. of each species between habitats and seasonal assemblages identified through Species Factor df F P Tukey HSD Bairdiella chrysoura habitat season 1 2 0.02 0.38 0.8750 0.6819 Cynoscion nehulosus habitat 1 2.74 0.1008 season 2 8.03 *** winter > summer Lagodon rliomboides habitat season 1 2 6.61 80.12 * *** tidal-creek > seagrass summer > fall > winter Leiostomus xanthurus habitat 1 6.84 * season 2 29.87 *** summer > fall, winter Mugil cephalus habitat season 1 2 1.23 2.54 0.2713 0.0862 Orthopristis chrysoptera habitat 1 2.96 0.0996 season 2 4.6 * summer > winter Sciaenops ocellatus habitat 1 1.98 0.1633 season 2 6.28 ** summer, winter > fall tidal creeks and dominated samples collected during January and February (Tables 2 and 5). Recruitment of YOYL. rhomboides and C. arenarius also contributed to the winter-spring species assemblage (Tables 4 and 5). The summer assemblage was influenced by recruit- ment of YOY F. grandis and C. nebulosus, which had significantly shorter standard lengths than they had in the winter-spring assemblage. The recruitment of S. ocellatus helped to characterize the fall assemblage. Emigration of larger individuals could also account for a decrease in mean length; however length-frequency plots showed that larger individuals remained vulner- able to the gear and that the reduction in mean length was due to recruitment of YOY fishes. Discussion Fishes collected in seagrass habitats in this study were similar to those found in other studies of seagrass hab- itats; resident species were present year-round and there were seasonal pulses of juveniles that used the seagrass habitats as a nursery (Reid, 1954; Livingston, 1982; Weinstein and Brooks, 1983). The assemblages we identified were the result of the staggered influx of YOY fishes of different species to seagrass habitats through- out the year. For example, YOY L. rhomboides and O. chrysoptera recruited during winter and spring, whereas other abundant species such as YOY B. chrysoura and Eucinostomus spp. entered the nursery during summer and fall. We found an increase in species abundance and species richness during summer and fall similar to that found by Reid (1954), who conducted his study near Cedar Key. The same pattern was evident in other estuarine systems (Cowan and Birdsong, 1985; Rooker et al., 1998), demonstrating that recruitment of many juvenile fish species to seagrass habitats during summer and fall allows the juveniles to use the protection pro- vided by the growing seagrasses (Stoner, 1983) and to use the food resources found within them (Carr and Adams, 1973). Early-life-history stages of species with commercial or recreational importance were found in each habi- tat, but seagrass habitats contained a greater variety of juveniles from offshore reef species than did tidal creeks. Along the southeastern United States, juveniles of many economically important species use a variety of habitats in estuaries as nurseries, including man- groves, oyster reefs, marshes, tidal creeks, and seagrass habitats (Coleman et al., 1999, Coleman, et al., 2000). In our study. YOY reef fish taxa, such as serranids, lutjanids, and haemulids, were more abundant in sea- grass habitats than they were in tidal-creek habitats, except for gray snapper [Lutjanus griseus). Juveniles of several reef species (C. striata, Mycteroperca microlepis, Serraniculus pumilio, Serranus subligaris, and Lachno- laimus maximus) were found only in seagrass habitats. However, a complicating factor in our study was the elimination of oyster habitats from our sampling design. Oyster reefs are known to harbor juvenile C. striata and M. microlepis (Coleman et al., 2000) and they may have been under-estimated in our study because we did not sample these habitats. Other economically im- portant species, such as C. nebulosus, also recruited to the seagrass habitats and are known to reside in them much of their life (Reid, 1954; McMichael and Peters, 1989; Mason and Zengel, 1996). These economically im- portant species use seagrass habitats in the Suwannee River estuary as a nursery and eventually enter local fisheries. Consequently, the maintenance of healthy 114 Fishery Bulletin 104(1) Table 5 Results of SIMPER procedure showing the average percent dissimilarity {d9r) for important species between tidal-creek assem- blages, di is the average contribution of the ;■'' species to the disssimilarity between groups; di/SDidi) is the ratio between the average contribution of the ("' species and the standard deviation of di; Cum di9c is the cumulative contribution to the total dis- similarity. Species are listed in decreasing contribution to average dissimilarity. Species Average abundance 6i di/SD(di) di% Cumulative bi'/c Leiostomus xanthurus winter-spi-ing summer 4.17 2.07 7.03 7.03 177.1 8.43 Anchoa mitchilli 4.93 243.61 3.83 2.4 6.46 13.49 Brevoortia spp. 10.87 0.74 2.65 1.42 4.46 17.95 Eucinostomus spp. 0.24 22.28 2.62 1.74 4.42 22.37 Cynoscion arenarius 1.62 3.02 1.94 1.77 3.26 25.63 Mugil cephalus 4.52 0.88 1.91 1.57 3.21 28.84 Membra^ martinica 0.04 21.61 1.9 1.10 3.19 32.03 Lagodon rhombnides 6.46 2.92 1.83 1.46 3.08 35.11 Leiostomus xanthurus winter-Spring fall 4.33 2.32 7.85 7.85 177.1 0.96 Eucinostomus spp. 0.24 24.63 4.03 2.63 7.30 15.15 Brevoortia spp. 10.87 0.03 2.58 1.41 4.68 19.83 Eucinsotomus herengulus 0.07 2.86 2.51 2.73 4.56 24.39 Fundulus grandis 1.26 12.02 2.19 1.72 3.97 28.36 Fundulus majalis 0.77 3.47 1.66 1.40 3.02 31.38 Adinia xenica 0.23 1.82 1.62 1.55 2.94 34.32 Cynoscion nebulosus 0.07 1.16 1.60 1.68 2.90 37.22 Fundulus g/-andis summer fall 2.59 2.34 4.84 4.84 1.16 12.02 Anchoa mitchilli 243.61 14.92 2.52 1.69 4.70 9.54 Sciaenops oceUatus 0.53 5.36 2.09 1.39 3.91 13.45 Eucinostomus spp. 22.28 24.63 1.89 1.27 3.53 16.98 Fundulus majalis 0.26 3.47 1.88 1.45 3.51 20.49 Bairdiella chrysoura 4.06 1.53 1.81 1.71 3.38 23.87 Adinia xenica 0.97 1.82 1.76 1.9 3.28 27.15 Cynoscion arenarius 3.02 0.87 1.7 1.78 3.18 30.33 seagrasses in this area is important to the preservation of these resources. Tidal-creek habitats in the Suwannee River estuary provided resources for many species that had restricted distributions related to salinity tolerances and includ- ed taxa that were also found in the nearby seagrass habitats. We found recreationally important freshwater taxa, such as Micropterus salmoides and Lepomis punc- tatus, in tidal-creek habitats. Other groups restricted to tidal-creeks were those tolerant of low salinity and in- cluded the fundulids, poeciliids, and cyprinodontids. In addition, some economically valuable species were more abundant in tidal-creeks than in seagrass habitats, in- cluding M. cephalus. C. arenarius, S. ocellatus, and L. griseus. Tidal creeks also supported a greater density of fishes than did seagrass habitats — a density that could have resulted from habitat preferences, differen- tial mortality between habitat types, or gear avoidance. Because the seine was set along the shoreline in tidal creeks, fishes were trapped between the seine and the shoreline, perhaps making them more vulnerable to the gear, whereas in seagrass habitats, the seine was pulled along the bottom with the end open prior to retrieval and fishes could have used the opening to escape. The results of our study showed that there was a more consistent assemblage of fishes in tidal creeks, whereas Tuckey and Dehaven Fish assemblages found in tidal-creek and seagrass habitats in the Suwannee River estuary 115 fish assemblages found in seagrass habitats had greater variability in the species present and in the abundance of those species. The more consistent assemblage of fishes found in tidal creeks could be explained by the persistence of vegetation throughout the year in tidal creeks, which may have contributed to reduced preda- tion and may have provided direct or indirect sources of food. Vegetation coverage in seagrass habitats was seasonal and Strawn (1961) found that above-ground seagrass biomass declined during winter and increased during summer and fall. The increased complexity re- sulting from blade density and seagrass species hetero- geneity offered by the growing seagrasses is known to affect fish abundance and composition (Stoner, 1983). The fish community structure in our study reflected this seasonal change; fewer fish species were present during winter and spring than in summer and fall. As seagrass biomass increased, fish species composition and total numbers also increased, resulting in greater variability within seagrass fish assemblages. We found that the combination of water tempera- ture, salinity, and water depth, more than any other combination of abiotic variables, helped to explain the fish community structure found in the Suwannee River estuary. Although water temperatures between the two habitats were similar, tidal creeks typically had soft mud sediments instead of sand and mud, marsh-grass species instead of submerged aquatic vegetation, deeper average depths, and lower salinity values. Water tem- perature has been shown to correlate with timing of recruitment for YOY fishes, which is ultimately re- lated to adult spawning patterns (Subrahmanyam and Coultas, 1980; Nelson, 1998; Paperno, 2002). Because water temperatures were similar in each habitat, dif- ferences in fish-community structures were more likely related to salinity tolerances, factors that correlate with salinity and water depth. Water depth in the Su- wannee River estuary varies seasonally; lowest water levels occur during winter (Strawn, 1961). The result is a confounding effect of water temperature and water depth that probably act in concert to limit distribution of fishes. A strong indicator that salinity may be the major abiotic factor that determines fish distributions, and ultimately species assemblages, was the low-salin- ity event during March 1998 that changed the seagrass fish assemblage to one more closely resembling a tidal- creek assemblage. If vegetation type were the primary factor controlling species assemblages in these habitats, tidal-creek species would remain in tidal-creeks and not invade seagrass habitats when salinity values changed to more favorable conditions. Therefore, varying salini- ties allowed different groups of fishes to use habitats according to their salinity tolerance (Wagner, 1999). Nordlie (2003) examined 20 studies of estuarine salt marsh fish communities in eastern North America and characterized communities based on the life history patterns exhibited by the species. General life history categories were originally established by McHugh (1967) and included permanent residents, marine nursery, ma- rine transients, diadromous, and freshwater transients. The 45 species that had overlapping distributions among habitats in our study were consistent with the classifi- cations for marine nursery or marine transient species. Marine transient species do not require estuarine habi- tats for development, but venture into estuaries during periods of low rainfall, whereas marine nursery species require estuarine conditions for development. The two exceptions in our study (Gobionellus bolesoma and M. gulosus) were considered primary residents of saltmarsh communities, but were frequently found in estuaries. We collected 80 fish species in tidal creeks in the Suwannee River estuary — more species than have been found in most other studies of tidal creeks — and this number could be related to the long-term duration of sampling. For example, Peterson and Turner (1994) observed 29 fish species inhabiting Louisiana marshes in a one-year study, whereas we found 51 additional species in our tidal-creek habitats. Similarly, Hettler (1989) found 35 species in a one-year study of saltmarsh fishes in North Carolina, and Weinstein (1979) recorded 61 species from his one-year study of the Cape Fear River, North Carolina. Furthermore, Cain and Dean (1976) found 51 species in a one-year examination of fishes in an intertidal creek in South Carolina. The first year of our study resulted in the collection of 61 species from tidal-creek habitats. It is likely that three years of sampling in our study increased our chances of collecting rare species, which resulted in the higher level of species richness. Another reason for the high species diversity and abundance of fishes that we found in tidal creeks could be attributed to our sampling along the tidal-creek edge, which is known for its structural complexity (Montague and Wiegert, 1990) and importance as a foraging and refuge area (Baltz et al., 1993; Kneib and Wagner, 1994; Peterson and Turner, 1994). For example, Baltz et al. (1993) collected fishes in Louisiana marsh edges to look at the importance of the marsh-edge microhabitat and found that the 15 most abundant fishes were concentrat- ed near the marsh edge and consisted mostly of early- life-history stages. They hypothesized that the fishes aggregated near the marsh edge to take advantage of the protection provided by the vegetation and the avail- able food resources. Our sampling targeted the tidal- creek edge, and the gear we used selected for juveniles and small-adult species, which could explain the higher diversity than that seen in other studies. Another pos- sibility is that our randomly chosen sampling sites cov- ered a greater variety of microhabitats along tidal-creek shorelines than did the sampling of Weinstein (1979), Hettler (1989), and Peterson and Turner (1994), which could also explain the higher species richness. Despite differences in sampling methods, the collection of 80 fish species in tidal creeks appears to be unusual. The withdrawal of fresh water from the Suwannee River would likely change the salinity regime in the Suwannee River estuary, which may in turn reduce spe- cies diversity in the region by reducing habitat availabil- ity to groups tolerant of low salinity. Furthermore, the high abundance of juvenile fishes that use low-salinity 116 Fishery Bulletin 104(1) tidal creeks as a nursery would be altered and the re- sponses could vary on a species-specific basis (Tsou and Matheson, 2002). A decline in the amount of freshwater inflow into the tidal-creeks could lead to an overall shift towards a more saline environment and result in the expansion of seagrass habitats. However, Strawn (1961) showed that the distribution of seagrasses at Cedar Key was affected by water depth, water clarity, and the inter- action of temperature and tides during winter months, making the prospect of seagrass expansion unlikely. Although a decrease in the amount of fresh water may result in an increase in water clarity through a reduction in dissolved nutrient input and reduced primary produc- tivity, as has been seen in Apalachicola Bay, Florida (Liv- ingston, 2003), the extreme low tides, cold temperatures, wave action, and sediment geochemistry in the Suwannee River estuary may negate the effects of increased light penetration (Koch, 2001). Therefore, a decrease in fresh water may result in an increase in high-salinity bare substrate that has been shown to be less suitable as a fish nursery than either seagrass or tidal-creek habitats (Sogard and Able, 1991; Rozas and Minello, 1998). Tidal-creek and seagrass habitats in the Suwannee River estuary contained diverse fish communities that reflected seasonal changes associated with recruitment of YOY fishes. Many of these species are the targets of commercial and recreational activities, which support local economies. Although much of the land surround- ing the Suwannee River estuary has been preserved, measures must be taken to ensure that the supply of fresh water from the Suwannee River is also preserved to maintain the integrity of the aquatic environment and the associated estuarine fish community. Acknowledgments We appreciate the effort of our coworkers at the Florida Marine Research Institute's Cedar Key Field Laboratory for assisting with the collection of data for this study and to Fred Vose and Cynthia Cooksey for the initial concept of this study. This paper benefitted from reviews by D. Nemeth, D. Adams, B. Winner, F. Vose, J. Whittington, K. Tisdel, R. McMichael, J. Leiby, J. Quinn. T. Tsou, and two anonymous reviewers. This project was supported in part by funding from the Department of the Interior, U. S. Fish and Wildlife Service, Federal Aid for Sport Fish Restoration Project number F-43, and Florida rec- reational fishing license revenues. Literature cited Baltz, D. M., C. Rakocinski, and J. W. Fleeger. 1993. Microhabitat use by marsh-edge fishes in a Loui- siana estuary. Environ. Biol. Fish. 36:109-126. Bozeman, E. L. Jr., and J. M. Dean. 1980. The abundance of estuarine larval and juvenile fish in a South Carolina intertidal creek. Estuaries 3:89-97. Cain, R. L., and J. M. Dean. 1976. Annual occurrence, abundance and diversity of fish in a South Carolina intertidal creek. Mar. Biol. 36:369-379. Carr, W. E. S., and C. A. Adams. 1973. Food habits of juvenile marine fishes occupying seagrass beds in the estuarine zone near Crystal River, Florida. Trans. Am. Fish. Soc. 102:511-540. Clarke, K. R. 1993. Non-parametric multivariate analyses of changes in community structure. Austr. J. Ecol. 18:117-143. Clarke, K. R., and R. M. Warwick. 1994. Change in marine communities: an approach to statistic analysis and interpretation, 144 p. National Environmental Research Council, UK. Coleman, F. C, C. C. Koenig, A. Eklund, and C. B. Grimes. 1999. Management and conservation of temperate reef fishes in the grouper-snapper complex of the southeastern United States. Am. Fish. Soc. Symp. 23:233-242. Coleman, F. C, C. C. Koenig, G. R. Huntsman, J. A. Musick, A. M. Eklund, J. C. McGovern, R. W. Chapman, G. R. Sedberry, and C. B. Grimes. 2000. Long-lived reef fishes: the grouper-snapper com- plex. Fisheries 25(3):14-21. Comp, G. S., and W. Seaman Jr. 1985. Estuarine habitat and fishery resources of Florida. /;; Florida aquatic habitat and fishery resources, IW. Seaman jr., ed.), p. 337-436. Florida Chapter of the American Fisheries Society, Eustis, FL. Cowan, J. H. Jr., and R. S. Birdsong. 1985. Seasonal occurrence of larval and juvenile fishes in a Virginia Atlantic coast estuary with emphasis on drums (Family Sciaenidae). Estuaries 8:48-59. Field, J. G., K. R. Clarke, and R. M. Warwick, 1982. A practical strategy for analyzing multispecies distribution patterns. Mar. Ecol. Prog. Ser. 8:37-52. Gilmore, R. G. Jr. 1988. Subtropical seagrass fish communities: population dynamics, species guilds and microhabitat associations in the Indian River Lagoon, Florida. Ph.D. diss., 199 p. Florida Inst. Technology, Melbourne, FL. Hettler, W. F, Jr. 1989. Nekton use of regularly-flooded saltmarsh cord- grass habitat in North Carolina, USA. Mar. Ecol. Prog. Ser. 56:111-118. Kilby. J. D. 1955. The fishes of two gulf coastal marsh areas of Florida. Tulane Stud. Zool. 2:175-247. Koch, E. W. 2001. Beyond light: Physical, geological, and geochemical parameters as possible submersed aquatic vegetation habitat requirements. Estuaries 24:1-17. Kneib, R. T, and S. L, Wagner. 1994. Nekton use of vegetated marsh habitats at differ- ent stages of tidal inundation. Mar. Ecol. Prog. Ser. 106:227-238. Livingston, R. J. 1982. Trophic organization of fishes in a coastal sea- grass system. Mar. Ecol. Prog. Ser. 7:1-12. 1984. The relationship of physical factors and biologi- cal response in coastal seagrass meadows. Estuaries 7:377-390. Tuckey and Dehaven: Fish assemblages found in tidal creek and seagrass habitats in the Suwannee River estuary 117 2003. Trophic organization in coastal systems, 388 p. CRC Press, Boca Raton, FL. Mason, W. T., Jr., and S. A. Zengel. 1996. Foods of juvenile spotted seatrout in seagrasses at Seahorse Key, Florida. Gulf Mex. Sci. 2:89-104. McGovern. J. C, and C. A. Wenner. 1990. Seasonal recruitment of larval and juvenile fishes into impounded and non-impounded marshes. Wetlands 10:203-221. McHugh, J. L. 1967. Estuarine nekton. In Estuaries (Lauff, G. H., ed.), p. 581-620. AAAS, Publ 83, Washington, D.C. McMichael. R. H.. Jr., and K. M. Peters. 1989. Early life history of spotted seatrout, Cynoficion neb- ulosus (Pisces: Sciaenidae), in Tampa Bay, Florida. Es- tuaries 12:98-110. Montague, C. L., and R. G. Wiegert. 1990. Salt marshes. In Ecosystems of Florida (R. L. Myers and J. E. Ewel, eds.), p. 481-516. Univ. Central Florida Press, Orlando, FL. Nelson, G. A. 1998. Abundance, growth, and mortality of young-of- the-year pinfish, Lagodon rhomboides, in three estu- aries along the gulf coast of Florida. Fish. Bull. 96:315-328. Nordlie, F. G. 2000. Patterns of reproduction and development of selected resident teleosts of Florida salt marshes. Hy- drobiologia 434:165-82. 2003. Fish communities of estuarine salt marshes of eastern North American, and comparisons with temper- ate estuaries of other continents. Reviews Fish. Biol. Fisheries 13:281-325. Paperno, R. 2002. Age-0 spot {Leiostomus xanthuriis) from two estuaries along central Florida's east coast: compari- sons of the timing of recruitment, seasonal changes in abundance, and rates of growth and mortality. Fl. Sci. 65:85-99. Paperno, R., K. J. Mille, and E. Kadison. 2001. Patterns in species composition offish and selected invertebrate assemblages in estuarine subregions near Ponce De Leon Inlet, Florida. Estuar. Coast. Shelf Sci. 52:117-130. Peterson, G. W., and R. E. Turner. 1994. The value of salt marsh edge vs. interior as a habitat for fish and decapod crustaceans in a Louisiana tidal marsh. Estuaries 17:235-262. Reid, G. K. Jr. 1954. An ecological study of the Gulf of Mexico fishes, in the vicinity of Cedar Key, Florida. Bull. Mar. Sci. Gulf Car. 4:3-94. Rooker, J. R., S. A. Holt, M. A. Soto, and G. J. Holt. 1998. Postsettlement patterns of habitat use by sciae- nid fishes in subtropical seagrass meadows. Estuaries 21:318-327. Rozas L. P., and T. J. Minello. 1998. Nekton use of salt marsh, seagrass, and nonveg- etated habitats in a South Texas (USA) estuary. Bull. Mar. Sci. 63:481-501. Shenker, J. M., and J. M. Dean. 1979. The utilization of an intertidal salt marsh creek by larval and juvenile fishes: abundance diversity and temporal variation. Estuaries 2:154-163. Sogard, S. M., and K. W. Able. 1991. A comparison of eelgrass, sea lettuce macroalgae, and marsh creeks as habitats for epibenthic fishes and decapods. Estuar. Coast. Shelf Sci. 33:501-519. Stoner, A. W. 1983. Distribution of fishes in seagrass meadows: role of macrophyte biomass and species composition. Fish. Bull. 81:837-846. Strawn, K. 1961. Factors influencing the zonation of submerged monocotyledons at Cedar Key, Florida. J. Wild. Manag. 25:178-189. Subrahmanyam, C. B., and C. L. Coultas. 1980. Studies on the animal communities in two north Florida salt marshes Part III. Seasonal fluctuations of fish and macroinvertebrates. Bull. Mar. Sci. 30:790-818. Tsou, T., and R. E. Matheson Jr. 2002. Seasonal changes in the nekton community of the Suwannee River estuary and the potential impacts of freshwater withdrawal. Estuaries 25:1372-1381. Wagner, C. M. 1999. Expression of the estuarine species minimum in littoral fish assemblages of the lower Chesapeake Bay tributaries. Estuaries 22:304-312. Weinstein, M. P. 1979. Shallow marsh habitats as primary nurseries for fishes and shellfish. Cape Fear River, North Caro- lina. Fish. Bull. 77:339-357. Weinstein, M. P., and H. A. Brooks. 1983. Comparative ecology of nekton residing in a tidal creek and adjacent seagrass meadow: community compo- sition and structure. Mar. Ecol. Prog. Ser. 12:15-27. 118 Abstract — Examination of 203 adult bluefish iPomatomus saltatrix) from Long Island, New York, in 2002 and 2003 and 66 from the Outer Banks, North Carolina, in 2003 revealed the presence of dracunculoid nem- atodes (PhiloinetJ-a saltatrix) in the ovaries of female fish. Percent prevalence reached 88% in July and then decreased after the peak of the spawning season. Bluefish contained up to 100 parasites per fish. Infec- tion was associated with a range of disorders, including hemorrhage, inflammation, edema, prenecrotic and necrotic changes, and follicular atresia, that may prevent proper de- velopment of oocytes and probably affect bluefish fecundity. Historical occurrences, life cycle, and geographi- cal distribution of this nematode remain largely unknown, but may play important roles in recruitment processes of bluefish. Prevalence, intensity, and effect of a nematode iPhilometra saltatrix) in the ovaries of bluefish iPomatomus saltatrix) Lora M. Clarke Marine Sciences Research Center Stony Brook University Stony Brook, New York 11794-5000 E-mail address Lora ClarkefQ*msrc sunysb edu Alistair D. M. Dove Department of Microbiology and Immunology Cornell College of Veterinary Medicine c/o Marine Sciences Research Center Stony Brook University, Stony Brook, New York 11794-5000 David O. Conover Marine Sciences Research Center Stony Brook University Stony Brook, New York 11794-5000 Manuscript submitted 14 July 2004 to the Scientific Editor's Office. Manuscript approved for publication 25 July 2005 by the Scientific Editor. Fish. Bull. 104:118-124 (2006). Factors influencing recruitment vari- ability in marine fishes are often com- plex and poorly understood. Slight variations in mortality rates, growth rates, and stage durations in the early life stages can result in tenfold or greater fluctuations in abundance (Houde, 1987). Recruitment variation appears to be driven by a combina- tion of factors, such as environmental and oceanographic processes (Munch and Conover, 2000), diet (Friedland at al., 1988; Marks and Conover, 1993; Juanes and Conover, 1995), growth and development (McBride and Conover, 1991; Hare and Cowen, 1997) and habitat use (Able et al., 2003). The importance of parasitism and disease has, however, seldom been considered. In the Northwest Atlantic, the bluefish [Pomatomus saltatrix) is dis- tributed from Florida to the Gulf of Maine and is both commercially and recreationally important. This highly migratory species has at least two distinct spawning seasons. The first occurs in the spring, from March to May, south of Cape Hatteras, North Carolina (NO (Kendall and Walford, 1979; Collins and Stender, 1987) and the second occurs off the coast of New York (NY) from late June to August (Norcross et al., 1974; Sherman et al., 1984), peaking in July (Chiarella and Conover, 1990). Ichthyoplankton surveys have indicated that a third spawning event occurs south of Cape Hatteras, NC, in the autumn, but juveniles spawned during this time frame have rarely been captured (Col- lins and Stender, 1987). During the collection of bluefish ovaries for another study, the nema- tode Philometra saltatrix Ramach- andran, 1973 was detected in the ovaries of adult bluefish. Previous studies of Philometra spp. in other host species have indicated that their presence can have a negative effect on fecundity (Oliva et al., 1992; Hesp et al., 2002), implying that a more complete understanding of parasites may be important to understanding reproductive success. Although fac- tors such as female size and condi- tion are often considered in deter- mining reproductive success, the role of parasitism is rarely investigated (Marshall et al., 1998; Marteinsdottir and Begg, 2002). The potential effect of this nematode on the reproductive Clarke et al Prevalence of the nematode Philometra saltatrix in the ovaries of Pomatomus saltan i 119 Long Island, NY 41N '05 -^t, iO -"6 ts' -"5 50 "5" 25" Outer Banks, NC '^^'T^^/ Cape Hatteras -Tros -Tew -Te' 15' -75' sc -sas' Figure 1 Map of the sampling area, Long Island, New York, and Outer Banks, North Carolina, where adult bluefish iPomatomus xaltatrix) were collected from commercial gill-netters, trawlers, and seafood markets in 2002-03 for examination of infestation bv the nematode Philometra saltatrix. potential and early life history success of bluefish is unknown. No information other than location of oc- currence (Ramachandran, 1973) and a brief abstract describing the presence of philometrids in the heart of juvenile bluefish is available (Cheung et al.M. The purpose of this study is to investigate the prevalence, intensity, and effect of Philometra saltatrix in the ova- ries of bluefish. Materials and methods Adult bluefish were collected from commercial gill-net- ters, trawlers, and seafood markets on Long Island, NY, and the Outer Banks, NC (Fig. 1). In NY, fish were caught off the southern coast of Long Island from Shin- necock Inlet to Montauk Point, approximately 1-15 km offshore. In NC, fish were caught approximately 1-40 km off the coast, and the majority of fish were caught 30-40 km east of Oregon Inlet. Cheung, P. J., R. F. Nigrelli, and G. D. Ruggieri. 1984. Philo- metra saltatrix infecting the heart of the 0-class bluefish. Pomatomus saltatrix (L.), from the New York coast. In S. F. Snieszko commemoration fish disease workshop, p. 27. Joint Workshop of Fish Health Section, AFS, and Midwest Disease Group, Little Rock, AR. Sampling dates were determined by the availability of fish through the local fishermen. In 2002, bluefish were sampled from mid-July through early October off the southern coast of Long Island, NY (80 females, 108 males). In 2003, fish were collected in NC in April (43 females, 21 males) and in NY from the end of June through September (123 females, 42 males). Fork length (FL), fish weight, gonad weight, preva- lence of both live and dead worms, total worm weight, and gonadosomatic indices (GSI) were recorded for each fish. Worms were often intertwined making it difficult to count the number of worms in each ovary; therefore, total worm weight per ovary was used as a proxy for in- tensity. Representative samples were fixed in a solution of 95% glacial acid and 5% formalin for identification. Initially, examinations of both male and female fish were conducted, but after preliminary evidence showed that nematodes were not present in the gonads of male fish, future examinations were restricted to female fish. Haphazardly selected ovaries were preserved in 10% formalin and processed according to standard histo- logical methods (Luna, 1968) to investigate pathologies associated with the parasite. Transverse sections were cut from the same region in the center of each ovary. These were examined under a light microscope and images were captured with a Spot Insight digital CCD and processed with ImagePro Plus software (Media Cybernetics, Silver Spring, MD). 120 Fishery Bulletin 104(1) Jul 02 Aug 02 Sep 02 Ocl 02 Apr 03 Jun 03 Jul 03 Aug 03 Sep 03 Month Figure 2 Monthly prevalence of live Philometru t^altatrix in the ovaries of bluefish (Pomatomus saltatrix). Sample sizes are noted on the top of the bars. [ZZ12002 IB 2003 T "' " I — ' — \ JullH Jul 24 Jul 30 Aug 5 Aug n Aug 1 7 Aug 23 Aug 29 Sep 4 Date Figure 3 Daily prevalence of live Philnmetra saltatrix in the ovaries of bluefish tPomatoniuti saltatrix) in 2002 and 2003, Results Description and location of worm Philoinetra saltatrix was identified in the gonads of female fish ranging in size from 363 to 815 mm (FL) in both NC and NY samples. The majority of worms found in the ovaries were gravid females. Gravid female worms were visible macroscopically and most often visible even before the initiation of ovary dissection. Female worms reached a maximum of 150 mm in length and approxi- mately 300 iim in width. Dead worms were present in all months sampled and were encapsulated by several melanised layers of fibrotic tissue. Approximately 20 juvenile male and female worms were identified in the ovary of a female fish in July 2003. These immature worms reached a maximum length of 2 mm. It should be noted, however, that not all fish were examined microscopically; therefore, it is possible that other male and juvenile worms were present in other hosts but not detected. Prevalence and intensity Bluefish in the western North Atlantic represent a single genetic stock (Graves et al., 1992), allowing us to combine data from New York and North Carolina to examine seasonal trends. Pooled by month of capture, the prevalence of infection varied seasonally during the two sampling years. In NC, prevalence of live nematodes was 7% in April, the spring spawning season (Fig. 2). Prevalence of live worms in NY in June was 4.8% and reached a maximum of 79% (2002) and 42% (2003) in July during the summer spawning season (Fig. 2). It is important to note that sampling in 2002 did not begin until mid-July but was conducted for the entire month in 2003. If only the second half of July is examined for both years, the prevalence is 79% and 83% for 2002 and 2003, respectively. This peak in prevalence was followed by a slow decrease in live worms until late August and early September after which time live worms were no longer detected. Examination of prevalence by sampling day showed greater evidence of interannual variation. In the middle of July 2002, 100% of bluefish examined were infected with live Philometra saltatrix, whereas in 2003 the highest percentage of fish sampled that were infected was 91% (Fig. 3). Furthermore, in 2002 prevalence peaked in the middle of July, whereas in 2003 preva- lence was highest at the end of the month. Additionally, live worms were detected until the end of August in 2002 but were detected two weeks later in 2003. Intensity of parasite infection, as indicated by total live and dead worm weight per fish, reached a maxi- mum 0.145 g (mean of 0.081 g ±0.1367) and was high- est during the month of July in both years of the study (Fig. 4). Patterns of intensity were similar to those of prevalence; intensity was highest during the summer spawning season and then later decreased. One bluefish caught just south of Cape Hatteras, NC, in the begin- ning of May 2003 had the highest intensity observed with over 100 nematodes present (Fig. 5, A and B). This fish was not included in the data analysis because it was the only fish examined during that month, but this extreme instance of infection sets the upper bound on the level of observed intensity. Binary logistic regression was used to test whether the presence or absence of the nematode was related to fork length (P=0.096), fish weight (P=0.292), or GSI (P=0.783), but no significant relationships were found. Clarke et al Prevalence of the nematode Philometra saltatnx in the ovaries of Pomatomus saltatnx Jul 02 Aug 02 Sep 02 Oct 02 Apr 03 Jun 03 Jul 03 Aug 03 Sep 03 H/lonth Figure 4 Monthly intensity of live and dead Philometra saltatrix in the ovaries of bluefish (Pomatomus saltatrix). However, a significant positive correlation was found between intensity and fish weight (R= 0.169, P=0.000), intensity and fork length (i?=0.221, P=0.003), and in- tensity and GSI (i? = 0.262, P=0.000). The Bonferroni method was applied to account for the multiple com- parisons. Adjusted P-values are given. Observations in adult males and YOY Although not the target of this study, we also detected the presence of Philometra saltatrix in the pericardial cavity of one adult male bluefish collected in NC and three adult males in NY, Fish size ranged from 411 to 429 mm FL. Worms were not detected in the gonads of male fish. Additionally, we detected worms in the pericardial cavity of three young of the year (YOY) bluefish caught in the Long Island Sound. Fish ranged in size from 140 to 185 FL. Histopathology Worms found in the ovary were surrounded by oocytes at various stages of development. The guts of these worms were most often filled with host erythrocytes. A diversity of pathological responses was noted, and significant variability was found among host individuals (Fig. 5). In many cases, infection was associated with interstitial hemorrhage in the connective tissues of the ovary. Occasionally, hemorrhage into the lumen of the ovary was also observed, but it was not clear whether this was caused by feeding activities of the worm or from tissue damage by other means. In many cases, moderate lymphocytic infiltration was observed in ovarian connec- tive tissues. Some specimens showed marked edema in the perifollicular spaces. Prenecrotic changes, including pyknosis and cellular swelling, were observed in connec- tive tissue cells, and necrosis of both connective tissue and oocytes (atresia) was observed in several instances. Granulomatous inflammation and fibrosis occurred in association with dead worms and, occasionally, atretic follicles. Fibrotic capsules surrounding dead worms were polylaminate and melanised and appeared to represent female worms that had expelled larvae and had subse- quently died. Encapsulated worms that still contained larvae were also observed occasionally. Discussion Bluefish along the east coast of the United States appear to be heavily infected with the ovarian nematode Philo- metra saltatrix. Although many studies provide descrip- tions of various philometrid species, very few provide information about the prevalence, intensity, or effect of these nematodes. Ramachandran (1973) described the presence of several female and male worms in a single ovary but provided no information on pathological changes associated with infection. Intensity and prevalence of Philometra saltatrix cycle seasonally and appear to be synchronous with the blue- fish spawning cycle. Although live worms were detected from April to October, the levels of infection were higher in July than in any other month sampled and this tim- ing is coincident with the peak of the summer spawning season off NY (Chiarella and Conover, 1990). The life cycle of Philometra saltatrix is unknown and, therefore, it is not clear whether initial infection coincides with the spawning season or if nematodes are acquired at an earlier time and reside in some other host tissue site before migrating to the ovaries during the spawning season, perhaps stimulated by hormonal cues from the host. Prevalence and intensity were much lower during the spring spawning season south of Cape Hatteras, NC. It is difficult to determine whether this was due to geographical differences or the limited sampling that month; it is possible that we missed the peak spawning period in 2003. The rapid decline in both prevalence and intensity after the summer spawning period indi- cates that Philometra saltatrix are released or migrate out of the host fish synchronously with the release of fish eggs. We speculate that the reproduction cycle of the nematode is closely matched to that of the host bluefish. Other studies have reported that infection occurs af- ter first host maturity and that female nematodes have only been observed in the gonads of females (Oliva et al., 1992; Hesp et al., 2002). Although not the focus of this study, the pericardial cavity was also found to be infected by Philometra saltatrix in adult males in NC and NY and in YOY bluefish in NY waters, in contradic- tion to these other findings. It is not clear if these YOY infections represent separate infections or if they are the same infections observed in adult females. Review- ing the reported lifecycle of other spirurid nematodes 122 Fishery Bulletin 104(1) N cl^^l*^^. *5> 100 |jm E '- v>^l^li\- Figure 5 Observations associated with Philometra saltatrix infection in bluefish (Pomatomus saltatrix) ovaries. (A) Gross, bisected ovary from infected bluefish, showing heavy infection with Philometra (B) Low power view of transverse section (approximately 1.5 cm in diameter) of heavily-infected bluefish ovary. M=muscular capsule, F=ovarian follicles. W=nematodes in lumen of ovary, displacing oocytes. (C) Histological section of healthy ovarian tissue showing dense packing of healthy oocytes in various developmental stages, little connective tissue, and no cellular infiltrates. (D) Histological section of infected bluefish ovary, showing relatively fewer oocytes (Ol and the presence oi Philometra (P). In addition, there is severe interstitial hemorrhage (H) and necrosis (N) of ovarian tissue, with heavy cellular infiltration and atresia (A) of ovarian follicles. and Philometra spp. (Williams and Jones, 1994), we speculate that nematode larvae are released at spawn- ing time, pass through the copepod intermediate stage, and a second perhaps paratenic intermediate host (or perhaps pass through only one intermediate host), and infect YOY bluefish. It is important to note that most YOY bluefish larger than approximately 40 mm TL are piscivorous (Marks and Conover 1993); thus, it is possible that the infected YOY bluefish observed in this study were infected through a second intermedi- ate host. The existence of a second intermediate fish host is supported by the observed positive relationship between host body size and parasite intensity. After ini- tial infection, these worms reside in non-ovarian sites such as the pericardial cavity (as we observed), where they may remain quiescent until maturation of the host. At that time, worms may migrate through the tissues to the ovary for spawning and the completion of their life-cycle. This proposed life history would explain the ontogenetic and seasonal patterns of worm distribution that we observed and explain the rapid appearance of well-developed worms in the ovary at the onset of the spawning season. The prevalence and intensity of infections we ob- served in bluefish were significantly higher than those reported in most other fish species. For example, inten- sity of infection in Glaucosoma hebraicum by Philometra lateolabracis ranged from 1 to 7 nematodes (mean of 2 nematodes) (Hesp et al., 2002), whereas prevalence of Philometra margolisi in the gonads of red grouper, Epinephelus morio, ranged from 14 to 28% depending on locality (Moravec et al., 1995) and prevalence of Philo- metra lateolabracis in Parupeneus indicus was 5.3% (intensity of 1-2 nematodes) (Moravec et al., 1988). Clarke et a\. Prevalence of the nematode Philometra saltan ix in the ovaries of Pomatomus saltalnx 123 Although most studies indicate the occurrence of these nematodes to be lower than in bluefish, in a later study Moravec et al. (1997) reported prevalence of Philometra margolisi in the gonads of Epinephelus morio to be as high as 88% with an intensity of up to 84 nematodes. Additionally, although most studies report the absence of male worms (Hesp et al., 2002), we found up to 20 males in the ovary of one female fish. The high per- centage of adult female bluefish that are infected may indicate that Philometra saltatrix are significantly af- fecting the reproductive success of bluefish. Despite some studies showing deleterious effects of other parasites on fish ovaries (see for example, Adler- stein and Dorn, 1998), studies on the effects of philome- trids on fish ovaries remain scarce. Indeed, in a recent review of histopathological assessments of gonadal tis- sue in wild fishes, Blazer (2002) made no mention of Philometra spp., despite the fact that it is probably the most common adult helminth observed in fish gonads. Annigeri (1962) described damage to ovaries of Oto- lithiis argenteiis by an unidentified philometrid, and Ramachandran (1975) mentioned necrosis in the ovaries of Mugil cephalus caused by Philometra cephalus. The infection of testes of pink snapper Pagriis auratus by Philometra lateolabracis resulted in partial or extensive atrophy of those gonads (Hine and Anderson 1981). Oliva et al. (1992) alluded briefly to lower fecundity in Paralabrax humeralis as a result of infection with Philometra but provided no data in support of this as- sertion. Hesp et al. (2002) studied the dynamics and effects of P. lateolabracis in the gonads of the Austra- lian predatory marine fish Glaucosoma hebraicum, but did not observe significant pathological abnormalities associated with infection. Histopathological changes associated with philome- trid infection in bluefish were significant. The hem- orrhaging, edema, inflammation, atresia, and necro- sis observed most likely reduce oocyte number and quality, leading to lower fecundity. The presence of erythrocytes in the guts of nematodes indicates that the parasites were feeding on the blood of the host; this diversion of nutrients to the parasite may exacer- bate the impacts of the worm on ovarian tissue. Fur- thermore, intense infections can reduce the effective volume of the ovary and thus lead to lower fecundity (Fig. 5, A and B). The high prevalence, intensity, and pathological dam- age observed could be an important factor in recruit- ment variation in bluefish. Although modest interannual differences in prevalence and timing of infection were observed in this study (Fig. 4), Philometra saltatrix was not observed by researchers conducting studies of the GSI of bluefish from NY waters in the 1980s (Chiarella and Conover, 1990; Conover, personal observ. ). Hence, it is possible that the abundance of Philometra may fluctuate greatly over long time scales. The reason for the current prevalence of the worm and its effect on bluefish recruitment is unknown. Future studies of this host-parasite system should seek to account for temporal variations of this kind. Acknowledgments We thank R. Knotoff and numerous other commercial fishermen from Shinnecock Inlet, NY, for helping us to obtain samples. B. Burns and J. Pearce of the NC Division of Marine Fisheries helped collect North Caro- lina bluefish. We thank R. Clarke, C. Knakal, and S. Abrams for laboratory and field assistance. We also thank two anonymous reviewers for helpful comments on the manuscript. Funding was provided by the Rut- gers/NOAA Bluefish-Striped Bass Research Program and the New York State Department of Environmental Conservation. Literature cited Able K. W., P. Rowe, M. Burlas, and D. Byrne. 2003. Use of ocean and estuarine habitats by young-of- year bluefish iPomatonuis saltatrix) in the New York Bight. Fish. Bull. 101:201-214. Adlerstein, S. A., and M. W. Dorn. 1998. The effect of Kudoa paniformis infection on the reproductive effort of female Pacific hake. Can. J. Zool. 76:2285-2289. Annigeri, G. G. 1962. A viviparous nematode, Philometra sp., in the ovaries of Otolithua argenteus (Cuvier). J. Mar. Biol. Assoc. (India) 3:263-265. Blazer, V. S. 2002. Histopathological assessment of gonadal tissue in wild fishes. Fish Physiol. Biochem. 26: 85-101. Chiarella, L., and D. O. Conover. 1990. Spawning season and first year growth of adult bluefish (Pomatomus saltatrix) from the New York Bight. Trans. Am. Fish. Soc. 119:455-462. Collins, M. R., and B. W. Slender. 1987. Larval king mackerel (Scomberomorus caralla), Spanish mackerel, (S. maculatus) and bluefish (Poma- tomus saltatrix) off the southeast coast of the United States, 1973-1980. Bull. Mar. Sci. 41:822-834. Friedland, K. D., G. C. Garman, A. J. Bejda, A. L. Studholme, and B. 011a. 1988. Interannual variation in diet and condition in juvenile bluefish during estuarine residency. Trans. Am. Fish. Soc. 117:474-479. Graves, J. E., J. R. McDowell, A. M. Beardsley, and D. R. Scoles 1992. Stock structure of the bluefish Pomatomus sal- tatrix along the mid-Atlantic coast. Fish Bull. 90 (4):703-710. Hare, J., and R. K. Cowen. 1997. Size, growth, development, and survival of the planktonic larvae of Pomatomus saltatrix. Ecology 78:2415-2431. Hesp, S. A., R. P. Hobbs, and I. C. Potter. 2002. Infection of the gonads of Glaucosoma hebraicum by the nematode Philometra lateolabracis: occurrence and host response. J. Fish Biol. 60:663-673. Hine, P. M., and C. D. Anderson. 1981. Diseases of the gonads and kidneys of New Zealand snapper, Chrysophrys auratus Forster (F. Sparidae). In Wildlife diseases of the Pacific Basin and other coun- tries (M. E. Fowler, ed.), p. 166-170. Academic Press, London. 124 Fishery Bulletin 104(1) Houde, E. D. 1987. Early life dynamics and recruitment variability. Am. Fish. Soc. Symp. 2:17-29. Juanes, F., and D. O. Conover. 1995. Size-structured piscivory; advection and linkage between predator and prey recruitment in young-of-the- year bluefish. Mar. Ecoi. Prog. Ser. 128:287-304. Kendall. A. W. Jr., and L. A. Walford. 1979. Sources and distribution of bluefish. Poinatoinus saltatrix, larvae and juveniles off the east coast of the United States. Fish. Bull. 77:213-227. Luna, L. G. 1968. Manual of histological staining of the Armed Forces Institute of Pathology, 258 p. McGraw-Hill Book Com- pany, New York, NY. Marks, R., and D. O. Conover. 1993. Ontogenetic shift in the diet of young-of-the-year bluefish iPomatomus saltatrix) during the oceanic phase of the early life history. Fish. Bull. 91:97-106. Marshall, C. T., O. S. Kjesbu, N. A. Yaragina, P. Solemdal, O. Ulltang. 1998. Is spawner biomass a sensitive measure of the reproductive and recruitment potential of Northeast Arctic cod? Can. J. Fish. Aquat. Sci. 55:1766-1783. Marteinsdottir, G., and G. A. Begg. 2002. Essential relationships incorporating the influ- ence of age, size and condition on variables required for estimation of reproductive potential in Atlantic cod Gadus morhua. Mar. Ecol. Prog. Ser. 235:235-256 McBride, R. S., and D. O. Conover. 1991. Recruitment of young-of-the-year bluefish {Poina- toinus saltatrix) to the New York Bight: variation in abundance and growth of spring and summer-spawned cohorts. Mar. Ecol. Prog. Ser. 78:205-216. Moravec, F., P. Orecchia, and L. Paggi. 1988. Three interesting nematodes from the fish Paru- peneus indices (Mullidae, PerciformesI of the Indian Ocean, including a new species, Ascarophis pariipenei sp. N. (Habronematoidea). Folia Parasitol. 35:47-57. Moravec, F., V. M. Vidal-Martinez, and L. Aquirremacedo. 1995. Philometra margolisi n.sp. (Nematoda, Philome- tridae) from the gonads of the red grouper, Epinephelus morio (Pisces, Serranidae), in Mexico. Can. J. Fish. Aquat. Sci. 52(suppl. 1) 161-165. Moravec, F., V. M. Vidal-Martinez, J. Vargas-Vazquez, C. Vivas- Rodriguez, D. Gonzalez-Solis, E. Mendoza-Franco, R. Sima- Alvarez, and J. Guemez-Ricalde. 1997. Helminth parasites o{ Epinephelus morio (Pisces: Serranidae) of the Yucatan Peninsula, southeastern Mexico. Folia Parasitol. 44:255-266. Munch, S. B., and D. O. Conover. 2000. Recruitment dynamics of bluefish, Pomatomus sal- tatrix, on the continental shelf from Cape Fear to Cape Cod, 1973-1995. ICES J. Mar. Sci. 57:393-402. Norcross, J. J., S. L. Richards, W. H. Massraann, and E. B. Joseph. 1974. Development of young bluefish iPomatomus sal- tatrix) and distribution of eggs and young in Virginian coastal waters. Trans. Am. Fish. Soc. 103:477-497. Oliva, M. E., A. S. Borquez, and A. N. Olivares. 1992. Sexual status of Paralahrax humeralis (Serra- nidae) and infection by Philometra sp. (Nematoda: Dracunculoidea). J. Fish Biol. 40:979-980. Ramachandran, P. 1973. Philometra saltatrix sp. n., infecting the gonads of the common bluefish Pomatomus saltatrix (L.) off the New England coast of the United States. Zool. Anz. 191:325-328. Ramachandran, P. 1975. Philometra cephalus sp. n. infecting the gonads of the striped mullet, Mugil cephalus L. from the Arabian Coast of Kerala, India, with a note on its pathology. Zool. Anz. 194:140-144. Sherman, K., W. Smith, W. Morse, M. Berman. J. Green, and L. Ejsymont. 1984. Spawning strategies of fishes in relation to cir- culation, phytoplankton production, and pulses in zoo- plankton off the northeastern United States. Mar. Ecol. Prog. Ser. 18:1-19. Williams, H., and A. Jones. 1994. Parasitic worms offish, p. 139-146. Taylor and Francis Ltd., Bristol. PA. 125 Abstract — Size-related differences in power production and swim speed duration may contribute to the observed deficit of nursing calves in relation to lactating females killed in sets by tuna purse-seiners in the eastern tropical Pacific Ocean (ETP). Power production and swim-speed duration were estimated for north- eastern spotted dolphins iStenella attenuata), the species (neonate through adult) most often captured by the fishery. Power required by neonates to swim unassisted was 3.6 times that required of an adult to swim the same speed. Estimated unassisted burst speed for neonates is only about 3 m/s compared to about 6 m/s for adults. Estimated long-term sustainable speed is about 1 m/s for neonates compared to about 2.5 m/s for adults. Weight-specific power requirements decrease as dolphin calves increase in size, but power estimates for 2-year-old spotted dol- phin calves are still about 40% higher than power estimates for adults, to maintain the same speed. These esti- mated differences between calves and adults are conservative because the calculations do not include accommo- dation for reduced aerobic capacity in dolphin calves compared to adults. Discrepancies in power production are probably ameliorated under normal circumstances by calves drafting next to their mothers, and by employing burst-coast or leap-burst-coast swim- ming, but the relatively high speeds associated with evasion behaviors during and after tuna sets likely diminish use of these energy-saving strategies by calves. Duration of unassisted swimming activity for spotted dolphin iStenella attenuata) calves: implications for mother-calf separation during tuna purse-seine sets Elizabeth F. Edwards Southwest Fisheries Science Center National Marine Fisheries Service 8604 La Jolla Shores Drive La Jolla, California 92037 E mail address Elizabeth Edwardsigi noaa.gov Manuscript submitted 15 February 2005 to the Scientific Editor's Office. Manuscript approved for publication 25 July 2005 by the Scientific Editor. Fish. Bull. 104:125-135 (2006). Dolphin calves draft in echelon posi- tion next to their mothers for the first few weeks after birth and continue to return to drafting in echelon posi- tion frequently throughout at least their first year (EdwardsM. "Echelon position' is the physical positioning of the calf within a few centimeters of the mother, near her mid-section, with fin motions reduced or absent (e.g., Norris and Prescott, 1961). This position takes advantage of the moth- er's flow field, reducing or effectively eliminating the energy cost to the calf of moving forward through the water (Weihs, 2004). Because draft- ing appears to be ubiquitous among dolphin calves, particularly during the neonate stage, it appears likely that drafting is an essential factor in maintaining physical association between calves and their mothers, especially when the calves are small. In the eastern tropical Pacific Ocean (ETP) a situation occurs in which it becomes important to con- sider the consequences for calves of losing their drafting association with their mother. In this area, a tuna purse-seine fishery targets schools of large yellowfin tuna that associate closely with schools of dolphins, pri- marily the spotted dolphin (SteneUa attenuata) (NRC, 1992). The associ- ated schools of tunas and dolphins are located by a helicopter sent out from the purse-seine vessel and are subsequently captured through the actions of several high-powered speed- boats, which are released from the purse-seiner to overtake and herd the associated animals into the closing arc of the purse-seine (NRC, 1992). Examination of the dolphins found dead in the net has revealed that 75% to 95% of the lactating females killed in the sets are not killed with an accompanying calf (Archer et al., 2004). This observed calf deficit is a potentially important factor in the lack of recovery of ETP dolphin popu- lations despite over a decade of very low fishery mortality (Wade et al.^), but little is known about why or how the separation of mothers and calves occurs (Archer et al., 2004). The chase, encirclement, and re- lease procedure of purse-seine fish- ing operations tends to be relatively prolonged — chases averaging about 30 minutes, capture and confine- ment about 90 minutes, and postre- lease swimming at least 90 minutes (Myrick and Perkins, 1995; Chivers 1 Edwards, E. 2002. Behavioral contri- butions to separation and subsequent mortality of dolphin calves chased by tuna purse-seiners in the eastern tropi- cal Pacific Ocean. National Oceano- graphic and Atmospheric Administra- tion Administrative Report LJ-02-28, 33 p. Southwest Fisheries Science Center, 8604 La Jolla Shores Drive, La Jolla, CA 92037. - Wade, P., S. Reilly, and T. Gerrodette. 2002. Assessment of the population dynamics of the northeastern offshore spotted and eastern spinner dolphin populations through 2002. National Oceanographic and Atmospheric Ad- ministration Administrative Report LJ-02-13, 58 p. Southwest Fisheries Science Center, 8604 La Jolla Shores Drive, La Jolla, CA 92037. 126 Fishery Bulletin 104(1) and Scott^). The set procedure also tends to induce rela- tively high, sustained swimming speeds in the dolphins. Swimming speeds of two spotted dolphins carrying ve- locity tags during and after tuna purse-seine sets in the ETP averaged 1.7-2.8 m/s during chase and 2.6-3.1 m/s after release, compared to undisturbed speeds of 1.2-1.9 m/s (Chivers and Scott^). Compared to adults, dolphin calves have smaller, less-coordinated muscles and lower aerobic capacity, especially at birth but per- sisting through the first year or more (Dolar et al., 1999; Dearoff et al., 2000; Noren et al., 2001; Noren et al., 2002; Eitner et al, 2003; Noren et al., 2004). Therefore, it appears possible that the observed calf deficit in the kill may result at least in part from energetics-related separation of calves from mothers during the chase or after release, as well as the inability of calves to main- tain speed with adults while swimming alone. The present study examines the possibility of ener- getics-related separation, by estimating the length of time (duration) during which spotted dolphin calves of different sizes ranging from neonate through two years of age can swim unassisted at various velocities. These velocity-durations are then compared to adult swim- ming capacities, in order to determine whether calves swimming without assistance may experience energy- based difficulty keeping up with adult dolphins during evasion from tuna purse-seine sets. Materials and methods Duration limits for unassisted swimming by spotted dolphin calves were calculated by combining estimates of mass-specific energy cost of steady, submerged, unas- sisted swimming by neonate through adult spotted dolphins, with reported swimming speed durations of adult dolphins of various species (Table 1). Swimming duration for adults was used because this information was not available for dolphin calves. The data were com- bined based on the assumption that the duration a given unit of dolphin muscle can sustain a given mass-spe- cific energy cost is the same for both adults and calves within any species of dolphin. It is much more likely that swimming capacity of calves within any dolphin species is significantly lower than that of adults and that dif- ferences are particularly pronounced at birth {Dolar et al., 1999; Dearoff et al., 2000; Eitner et al., 2003; Noren et al., 2001, 2002, 2004), but the exact differences are unknown. The swimming duration estimates presented in this study therefore represent maximum durations. It is also possible that species differ in power produc- tion capacities, based on observations of blood chemistry differences between species (Ridgway and Johnston, 1965), but it is not unreasonable to assume that relative differences between species are maintained throughout the size ranges of individuals within a species, so that younger dolphins are probably less adept than adults in any particular species. Thus, although it is not yet possible to quantify differences between calf and adult swimming-duration capacities in Stenella attenuata in nature, they are likely greater than estimated in the present study, so that problems apparent from this study are likely greater during actual tuna purse-seine sets. The energy cost of a steady, submerged, unassisted swimming rate was estimated in this study for veloci- ties ranging from 1.5 to 6.0 m/s in order to encompass speeds from slow undisturbed swimming to the maxi- mum likely to occur during attempted evasion from tuna purse-seine sets. Average velocities observed dur- ing tracking of dolphins before and after experimental sets varied from about 1.5 m/s m to about 3 m/s (Chiv- ers and Scott') but short term speeds attained during evasion maneuvers were likely to have exceeded these averages (Table 1). Total body energy costs were esti- mated first and then converted to mass-specific costs by dividing total estimated energy costs by the estimated total muscle weight of each size of modeled dolphin. Muscle-mass-specific measures are more appropriate than simply dividing by total body weight for compari- sons between sizes because muscle fraction of body weight increases with body weight in Stenella attenuata in the ETP, from 35-40% in neonates to 55-60% in adults (Edwards, 1993). Data sources Total body energy costs were estimated for eight mod- eled Stenella attenuata ranging in size (age) from new- born through adult. Sizes were selected to emphasize changes during the early months (Table 2). Size-at-age to two years old was estimated from Hohn and Ham- mond (1985). Size of an adult reproductive female was estimated from Perrin and Reilly (1984). Total body wet weight, wetted surface area of body, fins, and flukes, maximum body diameter, and fraction of total body weight composed of muscle were estimated for each size of modeled dolphin by using regression equations devel- oped from morphological measurements of ETP dolphins (Edwards 1993, Edwards"*). The morphological measurements were taken from 35 spotted dolphins ranging in size from 0.71 to 2.1 m total length (tip of rostrum to fluke notch). This size range encompasses all life-history stages (near-term fetus through mature adult) of the spotted dolphins found in the eastern tropical Pacific Ocean. Spotted dolphins are about 0.8 m total length at birth (Hohn and Hammond, 1985). Specimens included 22 females and 13 males (Edwards, 1993). Male specimens included ■' Chivers, S., and M. Scott. 2002. Tagging and tracking of Stenella specie.s during the 2001 Chase and Encirclement Stress Studies cruise. National Oceanographic and Atmo- spheric Administration Administrative Report LJ-02-33. 23 p. Southwest Fisheries Science Center, 8604 La JoUa Shores Drive, La Jolla, CA. 92037. ^ Edwards, E. 2002. Energetics consequences of chase by tuna purse-seiners for spotted dolphins (Stenella attenuata) in the eastern tropical Pacific Ocean. National Oceano- graphic and Atmospheric Administration Administrative Report LJ-02-29. 32 p. Southwest Fisheries Science Center, 8604 La Jolla Shores Drive, La Jolla, CA 92037. Edwards: Duration of unassisted swimming by Stenella attenualo 127 three fetuses, six immature, and four mature individu- als. Female specimens included three fetuses, six im- mature, one mature resting, one mature lactating, and 11 mature pregnant individuals. All specimens were killed during tuna fishing operations in the eastern tropical Pacific Ocean. Two of the specimens were col- lected in February 1980, one in July 1983, nine in July 1985, two in August 1985, seventeen in December 1985, and four were collected without a date noted (Edwards, 1993). All specimens were processed according to the same procedures prior to dissection. Immediately after the sets, the dolphin specimens were brought on board the vessel and frozen whole in the brine wells of the vessel. The specimens were transported frozen to port and then transported frozen to the Southwest Fisheries Science Center. Specimens were kept frozen until thawed in fresh water (about 27°C) just prior to dissection. Not all measurements were made on all specimens; therefore sample sizes differ between the regression equations presented below. Energetics model The energetics model used to esti- mate total body cost of swimming was taken from Edwards (1992, based on Magnuson, 1978), except that 1) new data were used to estimate dolphin body param- eters and 2) the estimate of fin plus induced drag was replaced by the multiplier 3 (see below). The model used standard hydrodynamic equations and methods (Hoerner, 1965; Hertel, 1969; Webb, 1975) to estimate hydrodynamic drag on a fully submerged streamlined body of revolution moving steadily in turbulent flow. Body surface area was increased to specifically include the surface area of fins and flukes (Fish''), and drag estimates were increased to account for body and fin movements. Because energy to move forward (thrust energy) must exactly balance the drag experienced by a steadily swimming animal, estimating total drag energy is equivalent to estimating thrust energy, i.e., the energy cost to swim (Fish and Rohr, 1999). studies by Fish (1998), Webb (1975), and Yates (1983). P was estimated as a function of total hydrodynamic drag (Di, in dynes) and velocity (V, in m/s) as P,„=D,V/lOl where the factor 10" converts (D,V) to watts. Total drag was estimated as a function of drag due to body, fins, and movements of body parts as D,=0.5pV'S,.C,, = 1.5pV%,C„ where p = density of seawater (1.025 g/cm'^); iS,,, = wetted surface area; Cj = coefficient of total drag; and 3 = drag augmentation factor. S,^ includes surface area of body plus fins and flukes, where estimated planar area of fins was increased by 69c to account for the curvature of the fins, based on measurements of individual slices from fins and flukes from one small and one large dolphin, 1.32 m and 1.93 m in length, respectively (Edwards, unpubl. data). Drag augmentation factor generally varies between 3 and 5 (e.g., Lighthill, 1971; Fish, 1993) and was assumed equal to the value of 3 in the present study, based on studies of gliding vs. actively swimming dolphins (Skrovan et al., 1999). The factor 3 accounts for the increase over gliding drag caused by body movements during active swim- ming. Use of a squared relationship between V and JD, is supported by the observed relationship between total drag and velocity in free swimming Tursiops truncatus (Skrovan et al. 1999, Eq. 6). Use of a cubic relation- ship between V and P, is supported by observations of swimming kinematics of Tursiops truncatus swimming between 1 and 6 m/s (Fish, 1993). Sj,, (in cm^) was estimated as S, =0.299L2°5, Model formulation Total power (P,, in watts) required to overcome drag during steady, submerged swimming (Hertel, 1969) by a modeled dolphin of a given total length (L, rostrum to fluke notch) was estimated as where P^ = mechanical power (in watts) required to overcome hydrodynamic drag; £„, = muscle efficiency; and E = "propeller efficiency" (efficiency of propul- sion by flukes). based on measurements from 19 Stenella attenuata ranging in size from 0.71 to 2.01 m, where L is total length (in cm). Cj was estimated from the formula for drag of sub- merged streamlined bodies of revolution moving at constant velocity (Hoerner, 1965; Hertel, 1969; Webb, 1975) as Cj=Cf\\ + [l.b(d I Lf'^) + l[(d I L\^]^, Cf = the coefficient of friction drag; and d = 0.12 £„, was assumed to be 0.2 from studies of muscle efficien- cies in terrestrial animals (e.g., Goldspink 1988), man (Alexander, 1983; quoting Dickinson, 1929) and dolphins (Fish, 1993, 1996). E^ was assumed to be 0.85 based on 5 Fish, F. 2002. Personal comniun. Liquid Life Laboratory, West Chester Univ. Pennsvlvania, West Chester, PA. where d = maximum body diameter (in cm) based on measurements from 24 Stenella attenuata ranging in size from 0.71 to 2.01 m. C, was estimated from the formula for submerged stream- lined bodies of revolution moving at constant velocity in turbulent flow (e.g., Webb, 1975) as 128 Fishery Bulletin 104(1) Table 1 Reported duration of various swimming speeds for several species of dolphins. Velocity (m/s) Duration Publication Species Burst (seconds) 11.2 leap Rohretal. (2002) Tursiops 11.1 2 Lang (1975) Stenella 10.6 2 Lang (1975) S ten el! a 10.3 2 Lang (1975) Stenella 9.7 leap Rohretal. (2002) Tursiops 8.8 <2 Rohret al. (2002) Delphinus 8.8 leap Rohretal. (2002) Tursiops 8.3 7.5 Lang (1975) Tursiops 8.2 <3 Rohretal. (2002) Tursiops 8.2 <120 Auetal. (1988) Stenella 8.0 7.6 Lang and Norris 1 1966) Tursiops 8.0 <3 Rohretal. (2002) Delphinus 7.8 2 Lang (1975) Lagenor 7.7 2-3 Lang and Prior (1966) Stenella 7.5 leap Lang (1975) Tursiops 7.4 leap Lang (1975) Tursiops 7.3 leap Au and Weihs (1980) Stenella 6.8 10 Lang and Norris ( 1966) Tursiops 6.7 1.4 Rohretal. (2002) Delphinus 6.7 <3 Rohretal. (2002) Delphinus 6.3 10 Lang and Norris ( 1966 ) Tursiops 6.4 240 Au and Weihs (19801 Stenella 6.2 <2 Rohretal. (2002) Tursiops 5.9 50 Lang and Norris (1966) Tursiops 6.0 2 Rohretal. (2002) Delphinus 5.6 <10 Rohretal. (2002) Tursiops 5.3 6.6 Maximum (minutes) Lang and Norris (1966) Tursiops 4.7 11 Au and Perry man (1982) Stenella 4.4 21 Au and Perryman (1982) Stenella 4.2 <8.5 Prolonged (hours-days) Rohret al.( 20021 Delphinus 3.5 1 hour Au and Perryman ( 1982 ) Stenella 3.2 9.3 hour Leatherwood and L. (1979) Stenella 3.1 extended period Lang and Norris ( 1966 ) Tursiops 3.1 90min Chivers and Scott'' Stenella 3.0 1.5 hours Au and Perryman ( 1982 ) Stenella 2.8 15 min Chivers and Scott-' Stenella 2.6 106 mm Chivers and Scott"* Stenella 2.1 days Chivers and Scott^ Stenella 1.9 hours Chivers and Scott'' Stenella 1.8 hours Chivers and Scott-' Stenella 1.7 32 min Chivers and Scott^ Stenella 1.6 4 days Hui(1987) Delphinus 1.6 hours-days Chivers and Scott-' Stenella 1.5 hours Chivers and Scott-' Stenella 1.2 hours Chivers and Scott-' Stenella 1.2 <50 hours Perrinetal. (1979) Stenella Edwards: Duration of unassisted swimming by Stenella attenuafa 129 No. of Length Weight Swimming animals (m) ikgl location Swimming condition 3 tank maximum observed leap velocity 1 1.86 52.7 lagoon accelerating, 25 m course 1 1.86 52.7 lagoon accelerating, 25 m course 1 1.86 52.7 lagoon accelerating, 25 m course 3 tank average maximum leap velocity 1 wild maximum observed speed 3 tank average leap velocity 1 1.91 89 lagoon 61 m course in 300 m lagoon 6 2.6 197 small tank maximum speed, 8 m course 300 wild evading helicopter 1 1.91 89 lagoon 61 m course; maximum speed observed 1 1.83 105 tank 5 or 8 m course 1 2.09 91 tank accelerating, long narrow tank 1 captive 70 m circular pool; maximum speed 1 1.91 89 ocean swimming in speedboat waves 1 1.91 89 ocean swimming in speedboat waves not specified wild during chase by speedboats 1 1.91 89 ocean swimming in speedboat waves "A school" wild average maximum observed speed 1 1.83 105 captive mean high swim speed 1 1.91 89 lagoon 61 m course; maximum observed speed not specified wild after escaping purse seine 6 2.6 197 tank mean high swim speed 1 1.91 89 ocean swimming in speedboat waves 41 wild average maximum speed evading aircraft 4 wild maximum speed after release 1 1.91 89 ocean swimming in speedboat waves school no. 5 wild maximum observed speed and duration school no. 5 wild swimming with ocean swells "A school" wild average maximum velocity evading aircraft school wild average speed evading vessel 1 2.05 wild average speed estimated from radiotag 1 1.91 89 ocean swimming in speedboat waves D230 wild postrelease velocity school no. 2 wild average speed evading vessel D230 wild velocity during chase by seiner D19 wild postrelease velocity 12 wild average speed, 1992, 1993 D230 wild average nonchase velocity, night D230 wild average nonchase velocity, day D19 wild velocity during chase by senior 2 1.76 95 tank small, shallow round tank 2 wild average speed, 2001 D19 wild average nonchase velocity, day D19 wild average nonchase velocity, night 26 wild minimum distance traveled, radiotag 130 Fishery Bulletin 104(1) Table 2 Dolphin mode parameters See text for formulas and rationale. Wetted Maximum Dolphin Total length Body weight Muscle weight surface area diameter Fineness no. Age (cml (kg) (kg) (cm2) (cm) ratio 1 new born 85 6.40 2.62 2700 13.5 6.30 2 1 week 87 6.85 2.81 2832 13.8 6.29 3 1 month 90 7.58 3.18 3036 14.3 6.28 4 3 months 98 9.76 4.29 3615 15.7 6.24 5 6 months 110 13.76 6.33 4581 17.7 6.20 6 1 year 129 22.08 11.04 6351 21.0 6.14 7 2 years 154 37.37 20.18 9131 25.4 6.07 8 adult 190 69.73 41.84 14045 31.7 .=1.99 C/ =o.o72i^-"^ where R is Reynold's number, estimated here as R = LV/v, where v = kinematic viscosity ( = 0.01 Stokes). The calculations above generated estimates of the power required for a whole dolphin of a given size to swim at a given velocity (P,, in watts). Power required per kilogram of wet weight muscle (P„, ), for a given velocity was estimated as where M„, = total muscle mass (wet weight in kg), esti- mated from total body mass (M,, wet weight in kg) as M„ -2.97M' based on In-ln regression of measurements from a sample of 26 Stenella attenuata from the ETP ranging in size from 0.71 to 2.06 m. Total muscle was used rather than some portion of measured musculature because the com- plex and interconnected muscle and connective tissue of dolphins makes it difficult to isolate any particular por- tion as uniquely responsible for locomotion (Pabst, 1990). M-i was estimated from total length (L, in cm) as M, =0.0000119L-^' based on In-ln regression of measurements from a sample of 23 Stenella attenuata from the ETP ranging in size from 0.71 to 2.01 m. Results Model corroboration Model estimates of the cost of swimming compared reasonably well with the cost of swimming in published reports for other species of dolphins swimming 1-6 m/s, in cases where the published reports can be appropri- ately compared with the present model. Although a number of studies present a variety of estimates of drag, thrust power, and metabolic power at various swimming speeds for a variety of dolphins (reviewed by Fish and Rohr, 1999), comparisons of present results with many of these earlier studies would be inappropriate because either the estimates were derived from completely dif- ferent models from the one used in the present study or from generally similar models but where different assumptions were made about model parameters, such as propulsive efficiency, metabolic efficiency, drag for- mulation, and body structure (Edwards'). Only appro- priately comparable published results are discussed in the present study. All weight-specific measurements in the following paragraph refer to total body mass. At estimated opti- mum velocities ranging from about 1.2 m/s in neonate spotted dolphins to about 1.7 m/s in adults (Edwards^), model estimates are approximately 3 W/kg for all sizes of spotted dolphin, compared to measurements (de- rived under various methods) of about 2.5-5.5 W/kg for adults of various species of dolphins, either resting or swimming about 2 m/s (Hui, 1987; Worthy et al., 1987; Williams et al., 1992; Fish, 1993; Yadzi et al., 1999). Observed average total metabolic rates calculat- ed from oxygen consumption by two Tursiops (average weight 162 kg), swimming 2 and 3 m/s, were approxi- mately 2.5 and 3.7 W/kg (Yadzi et al., 1999), compared to model estimates of approximately 2.9 and 5.9 W/kg for an adult spotted dolphin (about 70 kg) swimming at the same speeds. Model estimates of thrust power output for an adult spotted dolphin (about 70 kg) also compared well with thrust power estimated as a func- tion of velocity from videos of five Tursiops swimming between 1 and 6 m/s (Fish, 1993). Given an average adult Tursiops weight of 230 kg, average estimated thrust power for Tursiops swimming 1, 3, 5, and 6 m/s was 1, 3, 14, and 23 W/kg, compared to spotted dolphin model estimates of 1, 3, 13, and 21 W/kg, respectively. These comparisons refer to mechanical power output Edwards: Duration of unassisted swimming by Stenello oltenuata 131 only, because Fish's (1993) analysis does not include metabolic efficiency. Assum- ing the same metabolic efficiency of 0.2 for the Tursiops in Fish's study as used in the present model, an estimate of total power by his Tursiops swimming 6 m/s is 115 W/kg (i.e., 23x(l/0.2)), compared to the model estimate of about 125 W/kg for the adult spotted dolphin. Mass-specific cost of swimming 300 250 - 200 E 5' 150 100 50 Model estimates of mass-specific power requirements (watts per kilogram muscle, W/kg,„) for unassisted swimming by spot- ted dolphins in the ETP increase quickly with velocity, regardless of dolphin size, but increase much more quickly for smaller dolphins (Fig.l). Model results indicated that a neonate (85 cm) spotted dolphin must produce 3.6 times more power per kilogram of muscle than an adult swim- ming at the same speed. The factors are 3.5, 3.3, 2.8, 2.4, 1.8, and 1.4 times adult power for 87-, 90-, 98-, 110-, 129-. and 154-cm spotted dolphins. The factors do not vary with velocity because the relative differences between body dimensions in dolphins of different sizes remain constant regardless of swimming speed. For example, at speeds typical of ETP dolphins at- tempting to evade speedboats during an impending tuna-purse-seine set or during escape from the net after release (about 3 m/s) (Chivers and Scott^), the model predicts that an 85-cm neonate swimming un- assisted by its mother would require a muscle power of about 105 W/kg„, versus 30 W/kg„, for a 190-cm adult. The difference is still relatively marked even for two-year-old calves (154 cm), which would require an estimated 42 W/kg,,^ to maintain a steady submerged unassisted swim speed of 3 m/s, i.e., 1.4 times adult power requirements. Velocity duration limits Velocity duration limits for adults, drawn from litera- ture sources, appear to range from a few seconds for burst speeds above about 5 m/s, to several minutes for maximum speeds of about 4 m/s, to hours at prolonged speeds below about 3.5 m/s (Table 1). Estimated mass-specific muscle power required of adult spotted dolphins to achieve these speeds ranges from 208 W/kg,,, for burst speeds of 6 m/s greater, to 67 W/kg,,, for maximum sustainable speeds around 4 m/s, and 4-10 W/kg„, for prolonged duration speeds between 1 and 2 m/s (Fig.l). If one assumes that these power-duration limits apply also to dolphin calf muscle, neonate dolphins (85 cm) for example, could generate a burst duration power of about 200 W/kg,„ for a few seconds at unassisted swim- ming speeds of about 3.0 m/s, maximum duration power Estimated duration limits -Seconds -Minutes -Hours 00 m/s Figure 1 Estimated swimming duration limits for eastern tropical Pacific Ocean spotted dolphins iStenella attenuata) of various sizes (total length, tip of rostrum to fluke notch), swimming at various speeds. of about 70 W/kg,,, for some minutes at about 2.2 m/s, and prolonged duration power of about 5-10 W/kg„, for hours at about 1.0 m/s (Fig.l). For two-year-old spotted dolphins (154 cm), estimated swimming duration limits were about 5 m/s for burst effort at 200 W/kg„,, about 3.5 m/s for maximum effort at 70 W/kg„,, and about 2 m/s for prolonged effort at 5-10 W/kg,„. Intermediate ages produced intermediate results. Discussion Model results demonstrated clearly that swimming capa- bility of dolphin calves is much lower than that of adults. The especially pronounced difference between neonates and adults appears a likely basis for development of the ubiquitous drafting behavior observed in both captive and wild neonates. Given the large difference in swim- ming capacities, it is difficult to imagine how small dolphin calves could maintain any long-term association with the mother without the hydrodynamic benefits of drafting, even under normal circumstances. During the relatively fast-paced evasion swimming associated with tuna purse-seine sets in the ETP, it appears that spotted dolphin calves swimming unas- sisted are likely to have serious difficulty maintaining the same speed as adults — difficulties being especially pronounced for younger dolphins. It is also likely that the differences in swimming capacity presented in this study are underestimates because they do not include effects such as positive bouyancy (Cockroft and Ross, 1990; Mann and Smuts, 1999) and soft, flexible fins and flukes that are present for several hours after birth (McBride and Kritzler, 1951; Tavolga and Essapian, 132 Fishery Bulletin 104(1) 1957; Wells, 1991). Other differences persist for weeks or months, including undeveloped musculature (Ei- tner et al., 2003) and uncoordinated swimming and respiratory movements (Tavolga and Essapian, 1957; Taylor and Saayman, 1972; Reiss, 1984; Cockroft and Ross, 1990; Peddemors, 1990; Herzing, 1997; Mann and Smuts. 1999). Aerobic capacity is also likely to be reduced throughout the first few months, and not likely to reach adult levels until two to three years of age (Dolar et al., 1999; Dearoff et al., 2000; Noren et al., 2001, 2002, 2004). The results presented in this study also do not include the added costs of increased drag due to swimming near the surface, and repeatedly piercing it. which will occur much more often during evasion of tuna sets than during normal swimming. These other effects increase the likelihood that dolphin calves, particularly the younger individuals, will have problems coping with the swimming conditions induced during tuna purse-seine sets. Difficulties with unassisted swimming for dolphin calves may be ameliorated to some extent by employing drafting (Weihs, 2004), burst and coast (Au and Weihs, 1980), or leap-burst-coast swimming behaviors (or a combination of these behaviors) (Weihs, 2002). Theo- retically, these strategies could significantly reduce the cost to calves of moving through the water. However, it is not clear that any of these energy-saving strategies will be consistently attainable by spotted dolphin calves during herd movements associated with evading or es- caping sets by tuna purse-seiners in the ETP. Drafting can only be sustained through respiratory leaps if mother and calf leave and reenter the water with equal speed and efficiency (Weihs, 2004). Because evasion of tuna purse-seine sets involves sustained high-speed swimming characterized by repeated full- body respiratory leaps from the water (Au and Wiehs. 1980), calves are likely to tire and lose coordination more quickly than adults, and the smallest calves will be the first to experience problems. These factors may have contributed to the observed failure of a neonate dolphin calf in the ETP to successfully maintain a drafting relationship with its assumed mother during a respiratory leap while attempting to evade a vessel (Weihs, 2004). Once the drafting relationship is disrupted, the calf appears likely to fall behind because of its physical limitations, unless its mother alters her speed so that the calf can reestablish the drafting relationship. How- ever, review of dolphin mother-calf behavior indicates that the mother is not likely to voluntarily leave other adults during attempted evasion of tuna purse-seine sets in the ETP (EdwardsM. Thus, the faster or longer the chase or postrelease escape period (or combination of all three factors), and the younger the calf, the more likely it appears that the calf will become separated from its mother and be left behind during periods of fast swimming by the adults. Although fast swimming for a few minutes may not pose a great problem for many calves, longer periods appear increasingly likely to lead to significant calf loss in purse-seine operations. The duration of the high-speed period is at least as important as the speed maintained during the period, given the power-duration relationship. If fast swim- ming is concluded quickly, it is more likely that calves could achieve the required power during the short time required. As fast-swimming persists, power capacity decreases rapidly (Fig. 1), so that more and more calves are likely to be lost because they have exceeded their ability to keep up with the adults in their school. Even under normal circumstances, burst and coast, or burst-leap-coast, swimming patterns are not likely to provide sustained benefits to free-swimming calves until they approach adult size, coordination, and swimming capacity, because the calves will still tire more quickly than their associated adults. In the case of tuna purse- seine set evasion, sustained use of these swimming patterns may not be employed even by adults because these behaviors become less efficient as swim speed increases (Weihs, 2002). Video studies of swimming behaviors of adult Tursiops have shown that burst and coast periods decrease with increasing drag and cease altogether if the drag is large (Skrovan et al., 1999). In a study of dolphin gaits, an adult Tursiops experiencing increased drag due to an instrument pack did not use burst and coast propulsion during horizontal swimming speeds and depths at which four other adult Tursiops (without instrument packs) did employ short periods of burst and coast swimming. Even animals swim- ming without instruments at 1.5-3.7 m/s incorporated only short periods of burst and coast, and glide periods rarely exceeded 2 seconds. At the speeds characteristic of tuna purse-seine sets, dolphins tend to swim steadily or employ burst-leap-coast swimming behaviors (Weihs, 2002, Au and Weihs, 1980). With their smaller power production capacities, calves will need to shift into leap-burst-coast mode at slower speeds than will adults (Weihs, 2002), and again will tire more quickly. A potential shortcoming of the results presented in the present study is that the model produces absolute values from single calculations, without any associated statistics or sensitivity analyses. It is more likely that a range of values for morphological, physiological and behavioral characteristics occurs within real spotted dolphin populations in the ETP, so that a range of re- sponses is much more likely than a single response for a given size, swimming speed, and duration of chase. Lack of statistical and sensitivity analyses can be a problem where responses tend to be subtle and there- fore difficult to discern. However, the model results presented in this study with respect to estimated size- related differences in power production capacities of spotted dolphins in the ETP are far from subtle, and model results are reasonably similar to energy mea- sures derived from observations on real dolphins indi- cating that these model results are not unrealistic. In general, results presented here imply that older calves may be able to swim unassisted as fast and for as long as their associated adults during or after most sets (or during and after most sets) by tuna purse-seine vessels in the ETP. The markedly smaller energy production Edwards Duration of unassisted swimming by Stenella atlenuota 133 capacities of younger calves, particularly neonates and infants, compared to adults appears to be a reasonable factor contributing to mother-calf separation during or after sets by tuna vessels in the ETP. Management and research implications From these observations, it appears that energetic limi- tations may contribute significantly to the observed calf deficit in the retrieved purse-seine by facilitating mother-calf separation during purse-seine set activ- ity. Separation risk appears to increase strongly with decreasing calf size and with increased speed and dura- tion of evasion-related swimming. If separation is pro- longed, subsequent mortality of milk-dependent calves appears likely due to predation or starvation because adoption by a surrogate mother is unlikely in the ETP (EdwardsM. Examination of existing aerial photographs of spotted dolphins attempting to evade helicopters and research vessels, as well as directed future experiments in which aerial observations are conducted specifically to identify and monitor mother-calf pairs over time during various chase scenarios, would be very helpful in evalu- ating the extent to which calf size effects predicted here occur during relatively high-speed evasion behaviors by ETP spotted dolphin schools. Estimation of the rate at which dolphins of various ages are likely to encounter tuna purse-seine sets will be another important factor in relating the results of this analysis to potential calf loss during purse-seine sets. Assuming that power limitations are a significant factor in calf loss during evasion of tuna purse-seine sets, management strategies that could be implemented to minimize calf loss include 1) avoiding setting on schools that contain calves, particularly young calves, 2) minimizing the speed or duration of chase, and 3) minimizing the length of time dolphins are retained in the net so that mothers and calves separated during the set may reunite more quickly, if possible. Minimiz- ing the length of time in the net is already a desired fishery goal, related to minimizing cost and the poten- tial for dolphin mortality in the net. Minimizing chase duration is also already a desired fishery goal, because shorter chases are both less expensive and tend to be more successful. Minimizing chase speed is not a re- alistic goal because it would likely lead to the escape of the dolphins and the targeted tuna. Thus, the only potential improvement that might be made under cur- rent conditions (i.e., while setting on dolphin calves is permitted) would be to concentrate effort on identifying and avoiding dolphin schools with calves, presumably by scrutinizing the school from the vessel's helicopter prior to initiating a set. This may or may not be possible, depending on the ability of observers in helicopters to spot calves from the air. Given the model results presented here, it appears that further research and perhaps management actions should be implemented to better understand and reduce the risk of separation of mothers and calves during sets by tuna purse-seiners in the ETP. Acknowledgments This research was conducted under Section 304 (a)(1): Stress studies, of the 1997 International Dolphin Con- servation Program Act amendments to the U.S. Marine Mammal Protection Act, and subsequent Congressional directives. Literature cited Alexander. R. 1983. Animal mechanics. Blackwell Scientific Publica- tions, London, England. Archer, F., T. Gerrodette, S. Chivers, and A. Jackson. 2004. Annual estimates of the unobserved incidental kill of pantropical spotted dolphin (Stenella attenu- ata attenuata) calves in the eastern tropical Pacific Ocean. Fish. Bull. 102:233-244. Au, D., and W. Ferryman. 1982. Movement and speed of dolphin schools responding to an approaching ship. Fish. Bull. 80:371-379. Au, D., M. Scott, and W. Ferryman. 1988. Leap-swim behavior of "porpoising" dolphins. Cetus 8:7-10. Au, D., and D. Weihs. 1980. At high speeds dolphins save energy by leaping. Na- ture 284:548-550. Cockroft, v., and G. Ross. 1990. Observations in the early development of a captive bottlenose dolphin calf In The bottlenose dolphin (S. Leatherwood and R. Reeves, eds.), p. 461-478. Aca- demic Press, San Diego, CA. Dearoff, J., W. McLellan, R. Dillaman, D. Frierson, and D. Fabst. 2000. Precocial development of axial locomotor muscle in bottlenose dolphins (Tursiops truncatusl J. Morph. 244:203-215. Dickinson, S. 1929. The efficiency of bicycle pedaling, as affected by speed and load. J. Phys. Lond. 76:242-255. Dolar, M., F. Suarez, P. Fonganis, and G. Kooyman. 1999. Myoglobin in pelagic small cetaceans. J. E.xp. Biol. 202:227-236. Edwards, E. 1992. Energetics of associated tunas and dolphins in the eastern tropical Pacific Ocean: a basis for the bond. Fish. Bull. 90:678-690. 1993. Allometry of energetics parameters in spotted dolphin iStenella attenuata) from the eastern tropical Pacific Ocean. Fish. Bull. 91:428-439. Eitner, S., J. Dearoff, W. McLellan, and D. Fabst. 2003. Postural role of lateral axial muscles in develop- ing bottlenose dolphins {Tursiops truneatus). Proc. R. Soc. Lond. B 271:909-918. Fish, F 1993. Power output and propulsive efficiency of swim- ming bottlenose dolphins (Tursiops truneatus). J. Exp. Biol. 185:179-183. 1996. Transitions from drag-based to lift-based propulsion in mammalian swimming. Am. Zool. 36:628-641. 1998. Comparative kinematics and hydrodynamics of odontocete cetaceans: morphological and ecological correlates with swimming performance. J. Exp. Biol. 201:2867-2877. 134 Fishery Bulletin 104(1) Fish, F., and J. Rohr. 1999. Review of dolphin hydrodynamics and swimming performance. Tech. Rapt. 1801, 136 p. SPAWAR Sys- tems Center, SSC San Diego, San Diego, CA. Goldspinlt, G. 1988. Muscle energetics. In Mechanics and energetics of animal locomotion (R. Ale.xander and G. Goldspink, eds.), p. 57-81. Chapman and Hall, London, UK. Hertel, H. 1969. Hydrodynamics of swimming and wave-riding dolphins. In The biology of marine mammals (H. Ander- son, ed.), p. 31-63. Academic Press, New York, NY. Herzing, D. 1997. The life history of free-ranging Atlantic spot- ted dolphins iStenella frontalis): age classes, color phases, and female reproduction. Mar. Mamm. Sci. 13(4)1.576-595. Hoerner, S. 1965. Fluid-dynamic drag, 460 p. Hoerner, 148 Busteed Drive, Midland Park, NJ. Hohn, A., and P. Hammond. 1985. Early postnatal growth of the spotted dolphin, Stenella attenuata, in the offshore eastern tropical Pacific. Fish. Bull. 83:553-566. Hui, C. 1987. Power and speed of swimming dolphins. J. Mamm. 68:126-132. Lang, T. 1975. Speed, power, and drag measurements of dolphins and porpoises. /;; Swimming and flying in nature (T. Wu and C Brokaw, eds.), p. 553-572. Plenum Press, New York, NY. Lang, T., and K. Norris. 1966. Swimming speed of a Pacific bottlenose por- poi.se. Science 151(3710):588-590. Lang, T., and K. Pryor. 1966. Hydrodynamic performance of porpoises (Stenella attenuatta). Science 152:531-533. Leatherwood, S., and D. Ljungblad. 1979. Nightime swimming and diving behavior of a radio- tagged spotted dolphin, Stenella attenuata. Cetology 34:1-6. Lighthill, Sir J. 1971. Large-amplitude elongated-body theory of fish locomotion. Proc. R. Soc. Lond. B. 179:125-138. Magnuson, J. 1978. Locomotion by scombroid fishes; hydromechanics, morphology and behavior. /;; Fish physiology, vol. 7 (W. Hoar and J. Randall, eds.), p. 240-313. Academic Press, New York, NY. Mann, J., and B. Smuts. 1999. Behavioral development in wild bottlenose dolphin newborns (Tursiops sp.}. Behavior 136:529-566. McBride, A., and H. Kritzler. 1951. Observations on pregnancy, parturition, and post- natal behavior in the bottlenose dolphin. J. Mamm. 32.251-266. Myrick, A., and P. Perkins. 1995. Adrenocortical color darkness as indicators of con- tinuous premortem stress in chased and purse-seined male dolphins. Pathophys. 2:191-204. NRC (National Research Council). 1992. Dolphins and the tuna industry, 176 p. Committee on Reducing Porpoise Mortality from Tuna Fishing, National Research Council. National Academy Press, Washington, D.C. Noren, S., V. Cuccurulo, and T. Williams. 2004. The development of diving bradycardia in bottle- nose dolphins (Turs/op.s- truncatua). J. Comp. Physiol. B 174:139-147. Noren, S., G. Lacave, R. Wells, and T. Williams. 2002. The development of blood oxygen stores in bot- tlenose dolphins (Tursiops truncatus): implications for diving capacity. J. Zool. Lond. 258:105-113. Noren, S., T. Williams, D. Pabst, W. McLellan, and J. Dearoff 2001. Development of diving in marine endotherms: preparing the skeletal muscles of dolphins, penguins, and seals for activity during submergence. J. Comp. Physiol. 171:127-134. Norris, K., and J. Prescott. 1961. Observations on Pacific cetaceans of Califor- nian and Mexican waters. Univ. Calif. Publ. in Zool. 63:291-402. Pabst, A. 1990. Axial Muscles and Connective Tissues of the Bottle- nose Dolphin. In The bottlenose dolphin (S. Leather- wood and R. Reeves, eds. I. p. 51-67. Academic Press, San Diego, CA. Peddemors, V. 1990. Respiratory development in a captive-born bottle- nose dolphin. S. Afr. J. Zool. 5:178-184. Perrin, W., W. Evans, and D. Holts. 1979. Movements of pelagic dolphins iStenella spp.) in the eastern tropical Pacific as indicated by results of tagging operations, 1969-1976. NOAA Tech. Rept. NMFS SSRF-737, 14 p. Perrin, W., and S. Reilly. 1984. Reproductive parameters of dolphins and small whales of the family Delphinidae. Rep. Int. Whal. Comm. (special issue 6):97-133. Reiss, D. 1984. Observations on development of echolocation in young bottlenose dolphins. In Animal sonar ( P. Natchti- gall and P. Moore, eds.), p. 121-127. Plenum Press, New York, NY. Ridgway, S., and D. Johnston. 1965. Blood oxygen and ecology of porpoises of three genera. Science 151:456-458. Rohr, J., F. Fish, and J. Gilpatrick. 2002. Maximum swim speeds of captive and free- ranging delphinids: critical analysis of extraordinary performance. Mar. Mamm. Sci. 18(1):1-19 Skrovan, R., T. Williams. P. Berry, P. Moore, and R. Davis. 1999. The diving physiology of bottlenose dolphins (Tur- siops truncatus] II. Biomechanics and changes in bouy- ancy at depth. J. Exp. Biol. 202:2749-2761. Taylor, C, and G. Saayman. 1972. Social organization and behavior of dolphins and baboons. Ann. Cape Prov. Mus. Nat. Hist. 9(2): 11-49. Tavolga, M., and F. Essapian. 1957. The behavior of the bottle-nosed dolphin (Tursiops truncatus): mating, pregnancy, parturition, and mother- infant behavior. Zoologica 42:11-31. Webb, R 1975. Hydrodynamics and energetics of fish pro- pulsion. Bull. Fish. Res. Board Can. 190:1-158. Weihs, D. 2002. Dynamics of dolphin porpoising revisited. Integr. Comp. Biol. 42:1071-1078. 2004. The hydrodynamics of dolphin drafting. J.Biol. 3(8):l-23. Edwards Duration of unassisted swimming by Stenella atlenuota 135 Wells, R. 1991. Bringing up baby. Nat. Hist. August:56-62. Williams, T., W. Friedl, M. Fong, R. Yamada, P. Sedivy, and J. Haun. 1992. Travel at low energetic cost by swimming and wave-riding bottle-nosed dolphins. Nature 355 (6363): 821-823. Worthy, G., S. Innes, B. Braune, and R. Stewart. 1987. Rapid acclimation of cetaceans to an open-system respirometer. //; Approaches in marine mammal ener- getics (A. Huntley, D. Costa, G. Worthy and M. Castellini, eds.l, p. 115-126. Soc. Mar. Mamm. Spec. Publ. 1. Soc. Mar. Mamm., P.O. Box 368. Lawrence, KS 66044. Yadzi, P., A. Kilian, and B. CuHk. 1999. Energy expenditure of swimming bottlenose dolphins. Mar. Biol. 134:601-607. Yates, G. 1983. Hydrodynamics of body and caudal fin pro- pulsion. In Fish biomechanics (P. Webb and D. Weihs, eds.), p.177-213. Praeger Publishers, New York, NY. 136 Abstract — We assayed allelic varia- tion at 19 nuclear-encoded microsat- ellites among 1622 Gulf red snapper iLutjanus campechanus) sampled from the 1995 and 1997 cohorts at each of three offshore localities in the northern Gulf of Mexico (Gulf). Localities represented western, cen- tral, and eastern subregions within the northern Gulf. Number of alleles per microsatellite per sample ranged from four to 23, and gene diversity ranged from 0.170 to 0.917. Tests of conformity to Hardy-Weinberg equi- librium expectations and of genotypic equilibrium between pairs of micro- satellites were generally nonsignifi- cant following Bonferroni correction. Significant genie or genotypic het- erogeneity (or both) among samples was detected at four microsatellites and over all microsatellites. Levels of divergence among samples were low (Fj,.;. sO.OOl). Pairwise exact tests revealed that six of seven "significant" comparisons involved temporal rather than spatial heterogeneity. Contem- poraneous or variance effective size (N^y) was estimated from the tempo- ral variance in allele frequencies by using a maximum-likelihood method. Estimates of N^,y ranged between 1098 and >75,000 and differed significantly among localities; the N^.y estimate for the sample from the northcentral Gulf was >60 times as large as the esti- mates for the other two localities. The differences in variance effective size could reflect differences in number of individuals successfully repro- ducing, differences in patterns and intensity of immigration, or both, and are consistent with the hypothesis, supported by life-history data, that different "demographic stocks" of red snapper are found in the northern Gulf. Estimates of N^,y for red snap- per in the northern Gulf were at least three orders of magnitude lower than current estimates of census size (A^). The ratio of effective to census size {N^,/N) is far below that expected in an ideal population and may reflect high variance in individual reproductive success, high temporal and spatial variance in productivity among subre- gions or a combination of the two. Manuscript submitted 17 July 2004 to the Scientific Editor's Office. Manuscript approved for publication 25 July 2005 by the Scientific Editor. Fish. Bull. 104:136-148 (2006). Population structure and variance effective size of red snapper (Lutjanus campechanus) in the northern Gulf of Mexico* Eric Saillant John R. Gold Center for B)osystematics and Biodiversity Texas A&M University, TAMU-2258 College Station, Texas 77843-2258 E-mail address (for E Saillant) esaillantiStamu edu Red snapper (Lutjanus campechanus) is a highly exploited marine fish found primarily on the continental shelf of the Gulf of Mexico (Hoese and Moore, 1977). Red snapper abundance in the northern Gulf of Mexico (hereafter. Gulf) has decreased by almost 90% in the past two decades (Goodyear and Phares') owing to overexploita- tion by commercial and recreational fishermen, high juvenile mortality due to the shrimp-trawl fishery, and habi- tat change (Christman-; Gallaway et al., 1999). An important question for management and conservation of red snapper resources regards delineation of geographic stock structure. Should separate stocks exist, management of the fishery, including assessment and allocation, could be subdivided to avoid subregional over-exploitation and to maintain potentially adaptive genetic variation (Carvalho and Hauser, 1995; Hauser and Ward, 1998). A second important question for management is whether sufficient genetic resources exist to ensure long-term integrity of red snapper stocks. Preliminary esti- mates of the number of red snapper adults in the northern Gulf range from 7.8 to 11.7 million (Cowan-'; Porch""), which may indicate a priori that suf- ficient genetic resources are available. However, recent studies in other, com- mercially exploited marine fishes have shown that genetic effective size iN^.) can be three-five orders of magnitude smaller than estimates of census size or N (Hauser et al., 2002; Turner et al., 2002). Briefly, A^,, is defined as the number of individuals in an "ideal" population that would experience the same magnitude of genetic drift as the actual population (Hartl and Clark, 1989). N^. is an important biological parameter, in part because it reflects the relative effects of genetic drift and selection on nonneutral loci, and in part because it can indicate long-term risk of extinction from genetic factors (Turner et al., 2002). As long-term sustainability requires maintenance of sufficient genetic resources (Allendorf and Waples, 1996), populations (or stocks) with small N, potentially may * Contribution 133 from the Center for Biosystematics and Biodiversity, Texas A&M University, College Station, Texas 77843-2258. 1 Goodyear, C. P., and P. Phares. 1990. Status of red snapper stocks of the Gulf of Mexico: report for 1990, 72 p. Na- tional Marine Fisheries Service, South- east Fisheries Centre, Miami Laboratory, CRD 89/90-05, 75 Virginia Beach Drive, Miami, FL 33149-1099. - Christman, M. C. 1997. Peer review of red snapper (Lutjanus caii^pechanus) research and management in the Gulf of Mexico: statistics review, 51 p. Of- fice of Science and Technology, National Oceanic and Atmospheric Administra- tion, National Marine Fisheries Service, 1315 East-West Highway 9th Floor F/CS, Silver Spring, MD 20910. ■^ Cowan, J. H. 2004. Personal commun. Department of Oceanography and Coastal Sciences, Coastal Fisheries Institute, Louisiana State University, Baton Rouge, LA 70803. ■• Porch, C. E. 2004. Personal commun. National Marine Fisheries Center, Southeast Fisheries Science Center, 75 Virginia Beach Drive, Miami, FL 33149. Saillant and Gold: Population structure and variance effective size of Lut/anus campechanus in tfie nortfiern Gulf of Mexico 137 Sampl in the centra suffer reduced capacity to respond to changing or novel environmental pressures (Frankham, 1995; Higgins and Lynch, 2001). At present, red snapper resources in the northern Gulf are managed under a single- stock hypothesis (GMFMC-^ «). This hypoth- esis is supported by a number of prior genetic studies that employed allozymes (Johnson"), mitochondrial (mt)DNA (Gold et al., 1997; Gar- ber et al., 2004), and microsatellites (Gold et al., 2001b). In each study, genetic homogene- ity was observed across sampling localities, leading to the inference that gene flow was sufficient to maintain statistically identical allele distributions across the sampling area. All of these studies, however, either involved individuals of mixed cohorts or were based on relatively small sample sizes. Alternatively, tag-and-release and ultrasonic tracking (Fable, 1980; Szedlmayer and Shipp, 1994; Szedlmay- er, 1997) have indicated that adult red snapper are sedentary and exhibit high site fidelity (but see Patter- son et al., 2001). In addition, Pruett et al. (2005) used nested-clade analysis of red snapper mtDNA sequences and found evidence of different temporal episodes of both range expansion and restricted gene flow due to isolation by distance. They suggested that the spatial distribution of red snapper in the northern Gulf had a complex history that likely reflected glacial advance and retreat, habitat availability and suitability, and that the latter (i.e., physical conditions, and habitat availability and suitability) could partially restrict gene flow among present-day red snapper. Our objectives were to more rigorously assess genetic stock structure in the northern Gulf by employing a large sample size of individuals from discrete cohorts. We report allelic variation at 19 nuclear-encoded mic- rosatellites sampled from each of two cohorts at three different localities in the northern Gulf. Genetic homo- geneity among localities was tested and contemporane- ous or variance effective size (A/',,) at each locality was estimated from the temporal variance in allele frequen- cies (Waples, 1989) by using a maximum likelihood method (Wang, 2001). N t ^ "T^ ^ AL \ ^'^ ") MS \ TX \ U r— -^— — A ' V/~A J^ ^^Jjpv^ \ \ ^ % Alabama FL V. r 1 Louisiana Texas t '^ 33°N 27°N 96 W 84°W = GMFMC I Gulf of Mexico Fishery Management Coun- cil). 1989. Amendment number 1 to the Reef Fish Fishery Management Plan. 356 p. Gulf of Mexico Fishery Manage- ment Council, 2203 N. Lois Avenue, Suite 1100, Tampa, FL 33609. •^ GMFMC (Gulf of Mexico Fishery Management Coun- cil). 1991. Amendment number 3 to the Reef Fish Fishery Management Plan, 17 p. Gulf of Mexico Fishery Manage- ment Council, 2203 N. Lois Avenue, Suite 1100, Tampa, FL 33609. ' Johnson, A. G. 1987. An investigation of the biochemical and morphometric characteristics of red snapper tLutjanus canipecltanus) from the southern United States. Unpubl. manuscr., 26 p. National Marine Fisheries Service, Panama City Laboratory. 3.500 Delwood Beach Road, Panama Citv, FL 32407-7499. Figure 1 ing localities for adult red snapper iLiitJaniis campechanus) northern Gulf of Mexico: northwestern Gulf (Texas), north- 1 Gulf (Louisiana), and northeastern Gulf (Alabama). Materials and methods Adult red snapper were sampled between 1999 and 2001 by angling 40-50 km offshore of Port Aransas (Texas), Port Fourchon (Louisiana), and Dauphin Island (Alabama). These localities represent western, cen- tral, and eastern subregions, respectively, within the northern Gulf (Fig. 1) but hereafter for convenience are referred to as Texas, Louisiana, and Alabama. Individual fish were aged by otolith-increment analysis (following Wilson and Nieland, 2001) and individuals belonging to the 1995 and 1997 cohorts were selected for genetic analysis. Sample sizes for the 1995 and 1997 cohorts at each locality were 203 and 211 (Texas), 286 and 272 (Louisiana), and 376 and 274 (Alabama). Tissue samples (heart and muscle) were removed from each fish and stored as described in Gold et al. (2001b). The genotype of all fish was determined at 19 microsatel- lites by using PCR primers and methods described in Gold et al. (2001b). Summary statistics, including number of alleles, allel- ic richness (a measure of number of alleles independent of sample size), and unbiased gene diversity (expected heterozygosity) were computed for each microsatellite in each sample, with F-stat, version 2.9.3 (Goudet, 1995). Homogeneity of allelic richness and gene diversity among samples was tested with Friedman rank tests. Departure of genotypic proportions from Hardy-Wein- berg equilibrium expectations was measured within samples as Weir and Cockerham's (1984) f; probability of significance (Pjjw' was assessed with a Markov-chain method (Guo and Thompson, 1992), as implemented in Genepop (Raymond and Rousset, 1995) and by using 5000 dememorizations, 500 batches, and 5000 itera- tions per batch. Genotypic disequilibrium between pairs of microsatellites within samples was tested by exact tests, as implemented in Genepop and by employing the same Markov-chain parameters as above. Sequential Bonferroni correction (Rice, 1989) was applied for all multiple tests performed simultaneously. 138 Fishery Bulletin 104(1) Homogeneity of allele and genotype distributions among samples was examined with exact tests; signifi- cance of probability values was assessed by a Markov- chain method, as implemented in Genepop and using the same Markov-chain parameters as above. The degree of differentiation between pairs of samples was estimated as Weir and Cockerham's (1984) 9. as implemented in F-Stat. Sequential Bonferroni correction (Rice, 1989) was applied for all multiple tests performed simultane- ously. Spatial (geographic) differences among samples was assessed from multilocus data by estimating the likelihood that any given individual could be assigned to the sample locality from which it was drawn. The Bayesian method of Rannala and Mountain (1997), as implemented in Geneclass vers. 2.0 (Piry et al., 2005), was used to "assign" sampled individuals to a locality; the probability that an individual belonged to a given locality was calculated by using the resampling algo- rithm in Paetkau et al. (2004) and was based on 1000 simulated individuals. A locality was excluded as a potential origin of a given individual if the probability of the individual belonging to that locality fell below a threshold level of 0.05. Temporal changes in allele frequencies between the two cohorts were used to estimate variance effective size (Afj.y.) at each locality. This "temporal" method (Wa- ples, 1989) estimates effective size from the temporal variance in allele frequencies over the time interval between sampling, thus providing a contemporaneous estimate of A^,.. The pseudo-maximum-likelihood method described in Wang (2001) was used to obtain estimates and 95% confidence intervals of N^y by using the pro- gram MLNE available at http://www.zoo.cam.ac.uk/ ioz/software.htm#MLNE. The 95% confidence intervals were obtained as the range of support associated with a drop of two logarithm units of the likelihood func- tion, as inferred from the likelihood distribution (Wang, 2001). We used the analytical method developed by Jorde and Ryman (1995, 1996) to account for effects of overlapping generations on temporal-method estimates of N,. In a population with overlapping generations, the magnitude of temporal allele-frequency change is depen- dent in part on age-specific survivorship (/,) and birth rate (&,). Survivorship was calculated by assuming an equal probability (S) of surviving from one year class to the next and equal probability of survival of males and females. The value of -S (0.56 for Texas and 0.604 for Louisiana and Alabama) was estimated by using age- structure data of red snapper to calculate age-specific survivorship r high Texas 1098 Louisiana >75.000 Alabama 1235 6.52 3275 777 2706 >75,000 2515 gin. In addition, for four individuals from the Alabama sample (0.6%), all three localities were excluded as the potential origin. The pseudo-maximum-likelihood (temporal-method) estimates of variance effective size (N^y), corrected for overlapping generations, and their 95% confidence in- tervals for all three localities are shown in Table 4. 140 Fishery Bulletin 104(1) Estimates for the samples from Texas (A/py=1098) and Alabama (N^,y=1235) were essentially the same, falling well v/ithin the 95% confidence intervals of one another. An exact, maximum-likelihood estimate could not be generated for the sample from Louisiana because the value of N^.y with highest likelihood was >75,000 and the likelihood of higher values of N^,y could not be com- puted. This estimate is more than an order of magni- tude greater than the N ^• estimates for the other two localities and is significantly higher than those based on 95% confidence intervals. Discussion Genetic population structure Results obtained from pairwise exact tests indicated that the majority of genetic differentiation detected among the twelve spatial-temporal samples of red snap- per was due to allele and genotype distributions in the 1995 cohort sampled from the northwestern Gulf (Texas) and in the 1997 cohort sampled from the north- eastern Gulf (Alabama). In addition, exact tests among cohorts sampled in 1997 were nonsignificant and F^y. values among localities (both cohorts) averaged less than 0.001. These results indicate that the genetic differences observed in the present study are temporal (between cohorts within localities I and not spatial (among locali- ties). A "hint" of spatial differentiation was suggested by assignment tests. A total of 53% offish were reclassified (assigned) to their original locality, a proportion that differed significantly from that expected if genotypes were distributed randomly among localities. However, for 98% of the fish, none of the three localities could be unequivocally excluded as the locality of origin. The above results are in general agreement with oth- er, genetics-based studies of red snapper in the northern Gulf in that little to no significant geographic hetero- geneity in genetic markers, ranging from allozymes to mtDNA to microsatellites, has been detected (John- son^; Gold et al., 1997; Garber et al., 2004; Gold et al., 2001b). The one exception was a study by Bortone and Chapman* where significant heterogeneity in both temporal and spatial restriction-fragment patterns of the mitochondrially-encoded 16S ribosomal (r)RNA gene was reported. Bortone and Chapman'* suggested that the observed genetic heterogeneity likely stemmed from nonrandom sampling where individuals related by descent had remained in close spatial proximity to one another. In general, the "consensus" inference has been that gene flow among present-day red snapper in the northern Gulf is sufficient to offset divergence by genetic drift of the (presumed) selectively neutral genetic markers assayed. Such gene flow could involve movement of adults (Patterson et al., 2001), hydrody- namic transport of pelagic eggs and larvae (Goodyear-'), or both. The foregoing notwithstanding, there are a number of caveats (discussed in Pruett et al., 2005) to the infer- ence that significant gene flow occurs among present- day red snapper in the northern Gulf Briefly, tag-and- recapture and ultrasonic-tracking studies (Fable, 1980; Szedlmayer and Shipp. 1994; Szedlmayer, 1997) have indicated that adult red snapper are largely sedentary and nonmigratory. Significant movement of adults in the northeastern Gulf was reported by Patterson et al. (2001), but movement per se was mostly unidirec- tional (west to east) and the average distance covered in roughly a year was only -30 kilometers. Movement of (pelagic) red snapper eggs and larvae likely occurs, but neither egg nor larval type nor length of larval life are effective predictors of gene flow in marine fishes (Shulman and Bermingham, 1995) and larval exchange rates of marine species generally appear overestimated (Cowen et al., 2000). In addition, regardless of the life- history stage at which gene flow might occur in red snapper, movement across the continental shelf should be more-or-less linear and would be expected to fol- low a pattern of isolation by distance where fish from proximal localities are more similar genetically than fish from more distal ones. However, the correlations between genetic and geographic distance expected from isolation by distance have not been found (Gold et al., 1997; 2001b; this article). Finally, salient differences in geologic structure, habitat structure, and ecological conditions (Rezak et al., 1985; Gallaway et al., 1998), significant differences in salinity due to freshwater outflow from river systems in the northcentral Gulf (Morey et al., 2003), and the present-day occurrence during the summer months of a major hypoxic zone that extends out along the continental shelf from the Mississippi Delta westward (Rabalais et al.'°; Ferber, 2001) potentially could serve as barriers to movement and gene flow. Despite these caveats, the bulk of the genetics data has indicated essentially no difference among present- day red snapper sampled across the northern Gulf This is consistent with the unit stock hypothesis and with the inference that observed genetic homogeneity is due to substantial gene flow. However, it is important to ' Bortone, S. A., and R. W. Chapman. 1995. Identification of stock structure and recruitment patterns for the red snap- per, Lutjanus campechanus. in the Gulf of Mexico. Final report for Marfin Program Grant Number NA17FF0379-03, 39 p. Southeast Regional Office, National Marine Fisher- ies Service, 263 13"^ Avenue South, Saint Petersburg, FL 33701. •' Goodyear, C. P. 1995. Red snapper stocks in U.S. waters of the Gulf of Me.xico, 171 p. National Marine Fisheries Service. SE Fisheries Center, Miami Laboratory, CRD 95/96- 05, 75 Virginia Beach Drive, Miami, FL 33149-1099. i"Rabalais, N. N., R. E. Turner, D. Justic, Q. Dortch, and W. J. Wiseman. 1999. Characterization of hypoxia: topic I report for the integrated assessment on hypoxia in the Gulf of Mexico. NOAA Coastal Ocean Program Decision Analy- sis Series No. 15, 203 p. NOAA Coastal Ocean Program, 1305 East-West Hwy,. N/SC12 SSMC4, Silver Spring, MD 20910. Saillant and Gold Population structure and variance effective size of Lutianus campechonus in ttie norttiern Gulf of Mexico 141 note the following. First, it is possible that gene flow among present-day red snapper in the northern Gulf is limited but there has been insufficient time for semi- isolated lineages to completely sort into monophyletic assemblages. Pruett et al. (2005), on the basis of re- sults of nested-clade analysis of mtDNA haplotypes ob- tained from representative samples of the same cohorts (and localities) studied in the present study, hypoth- esized that semi-isolated assemblages of red snapper in the northern Gulf may exist over the short term, yet over the long term comprise a larger metapopu- lation tied together by periodic gene flow. Similarity in allele frequencies of genetic markers (such as used here and in previous studies of red snapper) presumed to be neutral to natural selection in theory could be maintained in such a metapopulation during periods when gene flow was limited or even absent. Second, all the genetic markers studied to date are presumed to be selectively neutral and to be affected primarily by the interaction(s) between gene flow and genetic drift. Genes affecting life-history and other traits that are in- fluenced by natural selection need not necessarily follow the same pattern(s), and geographic differences in adap- tively useful alleles at such genes can be maintained even in the face of substantial gene flow (Conover et al., 2005). It is thus not implausible that red snapper across the northern Gulf could differ in allele frequency at adaptively useful genes yet be homogeneous at selec- tively neutral ones. Contemporaneous effective size (N^^y) and present-day demographic dynamics Estimates of contemporaneous or variance effective size {N^y) for the Texas and Alabama localities (-1100) were essentially the same, but were at least an order of magnitude less than the N^,y estimate (>75,000) for the Louisiana locality. These estimates reflect differences in effective population size under the assumption that no immigration into a locality has occurred during the study interval, an assumption at odds with the general absence of allele-frequency heterogeneity among locali- ties as well as the low estimates of F>,.j. between pairs of samples. Short-term immigration (within the time interval of the study) could increase the variance in allele frequencies, thus resulting in an overestimate of A^^;. (Wang and Whitlock, 2003); whereas longer-term immigration at a (more-or-less) constant rate from a source population would have the opposite effect. The observed differences among localities could thus reflect differences in effective sizes, differences in patterns and intensity of immigration, or both. Temporal variation in allele frequencies also could occur if only a fraction of potential spawners at a locality actually contributed to recruitment and if such "temporal" subpopulations dif- fered in allele frequencies between years. Regardless, the differences in 7V,^• may indicate that different demo- graphic dynamics currently exist among localities. Wang and Whitlock (2003) recently extended previous maximum-likelihood methods to allow simultaneous es- timation of N^.y and m (rate of migration), provided data from multiple loci were available and all sources of im- migrants into a focal population were known. Because of the latter, we were able to generate estimates of N^.y and m only for the sample from the Louisiana locality (focal population), using the samples from the Texas and Alabama localities as source populations. Surpris- ingly, the estimate of N^,y for the Louisiana sample (4887, 95% confidence intervals of 1543 and 31,254) was at least -15 times smaller than the estimate based on no migration; m was estimated to be 0.0097 (95% confi- dence intervals of <0.001 and 0.0355). Clearly, more ex- tensive sampling across the northern Gulf is warranted to obtain estimates of N^.y and ;?; at other localities and to place this finding into perspective. Effective size (/Vp)/census size(AV) ratios Estimates of N^,y for all three sample localities were two or more orders of magnitude less than the current, pre- liminary estimates of adult census size (7.8-11.7 million) across the northern Gulf (Cowan-^; Porch^). Given that empirically derived N^JN ratios from a variety of verte- brates are 0.10-0.11 on average (Frankham, 1995), this result is somewhat surprising in that red snapper have a long reproductive life-span and overlapping generations (Wilson and Nieland, 2001), life-history features that are expected to increase N^./N by limiting variance in lifetime reproductive success among individuals (Jorde and Ryman, 1995; Waite and Parker, 1996). The issue is of importance in that census sizes of many commer- cially exploited marine fish populations are generally orders of magnitude larger than sizes where genetic resources might be lost (Franklin, 1980; Schultz and Lynch, 1997). However, species or populations with exceedingly small N^JN ratios potentially could be in danger of losing genetic resources, resulting in reduced adaptation and population productivity (Hauser et al., 2002). In addition, low N^JN ratios may explain in part why there often is a poor relationship between spawn- ing stock size and recruitment (Hauser et al., 2002). To date, N^,/N ratios smaller than 10" ' have been found for four other exploited marine fish species (Hauser et al., 2002; Turner et al., 2002; Hutchinson et al., 2003; Gomez-Uchida and Banks^M. Factors that theoretically can lower genetic effec- tive size with respect to census size include fluctuating adult number and year-class strength (Hedgecock, 1994; Vucetich et al., 1997), and variance in reproductive suc- cess. The latter can arise from biased sex ratio, high variance in male or female reproductive success, vari- ance in productivity among habitats, or any combina- tion of these factors (Nunney, 1996, 1999; Whitlock and Barton, 1997). Virtually any of these factors could lower N./N ratios in red snapper. Biased sex-ratio, however. " Gomez-Uchida. D., and M. A. Bank.s. 2003. Oregon State Univ., Hatfield Marine Science Center, Department of Fish- eries and Wildlife, 20.30 SE Marine Sci. Dr. Newport, OR 97365. 142 Fishery Bulletin 104(1) seems unlikely, because the ratio of males and females across three years of red snapper catch data was 0.97 and did not differ significantly from unity (Nieland^-). Variation in population number and in year-class strength, alternatively, seems likely, given the annual differences in commercial and recreational landings and the annual differences in abundance of age-0 and age- 1 red snapper, respectively (Schirripa and Legault^'*). Variance in reproductive success is far more difficult to assess but can include mating systems (Nunney, 1993) that lead to differences in reproductive success between males and females, and a "sweepstakes" process (Hedge- cock, 1994) where size-dependent fecundity, combined with random but family-specific early mortality (Hauser et al., 2002), leads to a large variance in the number of (surviving) offspring per parent. The latter could be ef- fected in red snapper by nonrandom removal of related subadults or juveniles either by localized overfishing or by shrimp trawling. Finally, variance in productiv- ity among habitats across the northern Gulf can be inferred from subregional differences in red snapper growth rates (Fischer et al., 2004) and from subregional ecological differences (Gallaway et al., 1998) that dis- tinguish the northeastern Gulf from the northwestern Gulf. Future ecological and behavioral studies to gen- erate estimates of variance in individual reproductive success or variation in productivity among localities are clearly warranted. Demographic stocks The differences in variance effective size (A''^,i) among the geographic samples of red snapper indicate present-day differences in demographic dynamics that may include the number of individuals that produce surviving off- spring, and hence by inference, census size. The factors, ecological or otherwise, promoting these demographic differences are difficult to assess but likely relate in some way to variation in food availability, habitat qual- ity, or mortality (or a combination of all three factors). Accordingly, one might expect one or more of these factors to differ among the sample localities, given the differences in variance effective size among localities. In addition, one might expect other demographic param- eters to differ as well. Our study was part of a larger, multidisciplinary proj- ect that involved studies of age-and-growth and repro- duction of red snapper at the three localities. The age- and-growth studies of Fischer et al. (2004) documented that fork length, total weight, and age-frequency dis- tributions differed significantly among localities. Red snapper sampled at the Texas locality were significantly '- Nieland, D. 2005. Personal commun. Coastal Fisheries Institute, Louisiana State University, Baton Rouge, LA 70803-7503. " Schirripa, M. J., and C. M. Legault. 1999. Status of the red snapper in U.S. waters of the Gulf of Mexico. Report SFD-99/00-75, 86 p. Southeast Fisheries Science Center, 75 Virginia Beach Drive, Miami, FL 33149-1099. smaller at age and reached smaller maximum size than did red snapper sampled at the Louisiana and Alabama localities; fish sampled at the latter two localities did not differ in size-at-age or maximum size. There also was a significantly higher proportion of smaller, young- er fish at the Texas locality than at the other two. The studies of reproductive capacity (Woods et al., 2003) involved only fish sampled from the Louisiana and Alabama localities but revealed that females sampled from the Alabama locality reached sexual maturity at a smaller size and younger age than did females sampled from the Louisiana locality. The differences in growth rate are likely a function in part of the more produc- tive, nutrient-rich waters found at the Louisiana and Alabama localities and caused by the plume produced from the Mississippi River (Fischer et al., 2004), a hy- pothesis reinforced by the observation (Grimes, 2001) that between 70-80'7f of fishery landings in the north- ern Gulf of Mexico come from waters surrounding the Mississippi River delta. The differences in female age and size at maturity, alternatively, are thought to indi- cate a stressed population and to reflect a compensatory response to growth overfishing or declining population size, or a response to both (Trippel, 1995; Woods et al., 2003). Collectively, the life-history differences and the differences in genetic-based estimates of effective size strongly suggest that red snapper at the three localities represent three different, demographic stocks. A critical issue is whether the demographic differ- ences in life history observed among red snapper in the northern Gulf are genetic or phenotypic (environmen- tally induced) in origin. Most discussions of stock struc- ture in commercially exploited marine fishes involve an explicit genetics component (Gold et al., 2001a), and typically, the absence of genetic heterogeneity within a fishery leads to management planning for a single unit stock. However, life-history traits can change rapidly in response to environmental pressures (e.g., size-selective fishing), and it has been hypothesized that the pool of genotypes that code for life-history traits is a highly dynamic property of populations, and moreover, that lo- cal adaptation(s) differentiating populations can evolve even in the presence of extensive gene flow (Conover et al., 2005). Thus, demographically different stocks could differ genetically, but not necessarily in selectively neu- tral markers that respond primarily to the interaction between gene flow and genetic drift. The issue also is of importance to management planning because phenotypi- cally plastic responses due to environmental differences generally can be reversed fairly quickly, whereas genetic responses are typically much slower (Hutchings, 2004; Conover et al., 2005). The geographic differences in red snapper in growth rates and shifts in timing of female maturity in all likelihood are due to a mix of genetic and environmen- tal factors, as are most life-history traits in a variety of animal species, including fishes (Mousseau and Roff, 1987; Conover and Munch, 2002). A significant genetic component to growth rate is well documented in a va- riety of fishes under aquaculture (Dunham et al., 2001) Saillant and Gold: Population structure and variance effective size of Lut/anus compechanus in tfie nortfiern Gulf of Mexico 143 and genotypes for smaller size and younger age at ma- turity clearly exist (Gjerde, 1984; Tipping, 1991; Trip- pel, 1995). These considerations indicate that a genetic component to growth rate and age at maturity may exist in red snapper, and if so, stock-structure consid- erations solely on the basis of homogeneity in selectively neutral genetic markers may not be warranted. Acknowledgments We thank A. Fisher for providing individual age data, D. Nieland for providing life-history data, T. Turner for extensive assistance with statistical analysis, the Texas A&M supercomputing facility staff for help with parallel processing of MLNE, and S.C. Bradfield, C. Burridge, and L. Richardson for technical assistance. We also thank W. Patterson and T. Turner for helpful com- ments on early drafts of the manuscript, and J. Cowan and C. Porch for providing estimates of red snapper census size. Work was supported by the Gulf and South Atlantic Fisheries Development Foundation (Grant 70- 04-20000/11824), by the Marfin Program of the U.S. Department of Commerce (Grant NA87-FF-0426I, and by the Texas Agricultural Experiment Station (Project H-6703). This article is number 46 in the series "Genet- ics Studies in Marine Fishes." Literature cited Allendorf, F. W., and R. S. Waples. 1996. Conservation and genetics of salmonid fishes. In Conservation genetics: case histories from nature (J. C. Avise, and J. L. Hamrick. eds.l p. 238-280. Chapman and Hall. New York, NY. Carvalho, G. R.. and L. Hauser. 1995. Molecular genetics and the stock concept in fisheries. In Molecular genetics in fisheries (G. R. Carvalho, and T. J. Pitcher eds.l p. 55-80. Chapman and Hall, London. Conover, D. O., S. A. Arnott, M. R. Walsh, and S. B. Munch. 2005. Darwinian fishery science: lessons from the Atlan- tic silverside. Can. J. Fish. Aquat. Sci. 62:730-737. Conover, D. O., and S. B. Munch. 2002. Sustaining fisheries yields over evolutionary time scales. Science 297:94-96. Cowen, R. K., K. M. M. Lwiza, S. Sponaugle, C. B. Paris, and D. B. Olson. 2000. Connectivity of marine populations: open or closed? Science 287:857-859. Dunham, R. A., K. Majumdar, E. Hallerman, D. Bartley, G. Mair, G. Hulata, Z. Liu, N. Pongthana. J. Bakos, D. Penman, M. Gupta, P. Rothlisberg, and G. Hoerstgen-Schwark. 2001. Review of the status of aquaculture genetics. In Aquaculture in the third millennium (R. P. Subasinghe, P. Bueno, M. J. Phillips, C. Hough, S. E. McGladdery, and J. R. Arthur, eds.) p. 137-166. FAG, Rome. Fable, W. A. 1980. Tagging studies of red snapper iLiitjanus campecha- nus) and vermilion snapper [Rhomboplites auroru- bens) off the south Texas coast. Contrib. Mar. Sci. 23:115-121. Ferber, D. 2001. Keeping the Stygian waters at bay. Science 291: 968-973. Fischer, A. J., M. S. Baker Jr., and C. A. Wilson. 2004. Red snapper, Lutjanus campechanus, demographic structure in the northern Gulf of Mexico based on spatial patterns in growth rates and morphometries. Fish. Bull. 102:593-603. Frankham, R. 1995. Effective population-size adult-population size ratios in wildlife: a review. Genet. Res. 66:95-107. Franklin, I. R. 1980. Evolutionary change in small populations. Chap- ter 8 m Conservation biology: an evolutionary-eco- logical perspective (M. Soule, and B. Wilcox, eds.) p. 135-149. Sinauer, Sunderland, MA. Gallaway, B. J., J. G. Cole, R. Meyer, and P. Roscigno. 1999. Delineation of essential habitat for juvenile red snapper in the northwestern Gulf of Mexico. Trans. Am. Fish. Soc. 128:713-726. Gallaway, B. J., M. Longnecker, J. G. Cole, and R. M. Meyer. 1998. Estimates of shrimp trawl bycatch of red snapper, Lutjanus campechanus, in the Gulf of Mexico. In Fish- ery stock assessment models (F. Funk F, T. J. Quinn II, J. Heifetz, J. N. lanelli, J. E. Powers, J. F. Schweigert, P. J. Sullivan, and C.-I. Zhang, eds.) p. 817-839. Alaska Sea Grant College Program, No. AK-SG-98-01. Fair- banks, AK. Garber, A. F, M. D. Tringali. and K. C. Stuck. 2004. Population structure and variation in red snapper {Lutjanus campechanus) from the Gulf of Mexico and Atlantic Coast of Florida as determined from mitochon- drial DNA control region sequence. Mar. Biotechnol. 6:175-185. Gjerde, B. 1984. Response to individual selection for age at sexual maturity in Atlantic salmon. Aquaculture 38:229- 240. Gold, J. R., C. P Burridge. and T. F. Turner. 2001a. A modified stepping-stone model of population structure in red drum, Sciaenops ocellatus (Sciae- nidae) from the northern Gulf of Mexico. Genetica 111:305-317. Gold, J. R., E. Pak, and L. R. Richardson. 2001b. Microsatellite variation among red snapper [Lut- janus campechanus) from the Gulf of Mexico. Mar. Biotechnol. 3:293-304. Gold, J. R., F. Sun, and L. R. Richardson. 1997. Population structure of red snapper from the Gulf of Mexico as inferred from the analysis of mitochondrial DNA. Trans. Am. Fish. Soc. 126:386-396. Goudet, J. 1995. F-STAT (Version 1.2): A computer program to cal- culate F-Statistics. J. Hered. 86:485-486. Grimes, C. B. 2001. Fishery production and the Mississippi River discharge. Fisheries 26:17-26. Guo, S., and E. Thompson. 1992. Performing the exact test of Hardy-Weinberg propor- tion from multiple alleles. Biometrics 48: 361-372 Hartl, D. L., and A. G. Clark. 1989. Principles of population genetics, 2"'' edition, 669 p. Sinauer Assoc. Inc.. Sunderland, MA. Hauser, L., and R. D. Ward. 1998. Population identification in pelagic fish: the limits of molecular markers. In Advances in molecular ecol- 144 Fishery Bulletin 104(1) ogy (G. R. Carvalho, ed.), p. 2191-2242. lOS Press, Amsterdam. Hause-, L., G. J. Adcock, P. J. Smith. J. H. B. Ramirez, G. R. Carvalho. 2002. Loss of microsatellite diversity and low effective population size in an overexploited population of New Zealand snapper (Pagrus auratus). Proc. Natl. Acad. Sci. USA. 99:11742-11747. Hedgecoek, D. 1994. Does variance in reproductive success limit effective population sizes of marine organisms? In Genetics and evolution of aquatic organisms (A. R. Beaumont, ed.) p. 122-134. Chapman and Hall, New York, NY. Higgins, K., and M. Lynch. 2001. Metapopulation extinction caused by mutation accumulation. Proc. Natl. Acad. Sci. USA 98: 2928- 293.3. Hoese, H. D., and R. H. Moore. 1977. Fishes of the Gulf of Mexico, 327 p. Texas A&M Univ. Press, College Station, TX. Hutchings. J. A. 2004. Life history consequences of over-exploitation to population recovery in northwest Atlantic cod, Gadus morhua. Can. J. Fish. Aquat. Sci. 62:824-832. Hutchinson, W. F., C. van Gosterhout, S. L Rogers, and G. R. Carvalho. 2003. Temporal analysis of archived samples indi- cates marked genetic changes in declining North Sea cod [Gadus morhua). Proc. R. Soc. London, B 270: 2125-2132. Jorde, P. E., and N. Ryman. 199.5. Temporal allele frequency change and estima- tion of effective size in populations with overlapping generations. Genetics 139:1077-1090. 1996. Demographic genetics of brown trout iSalmo trutta) and estimates of effective population size from temporal change of allele frequencies. Genetics 143:1369-1381. Morey, S. L., W. L. Schroeder, J. J. O'Brien, and J. Zavala- Hidalgo. 2003. The annual cycle of riverine influence in the east- ern Gulf of Mexico. Geophys. Res. Letters 30:OCE 5.1-5.4. Mousseau, T. A., and D. A. Roff 1987. Natural selection and the heritability of fitness components. Heredity 59:181-198. Nunney, L. 1993. The influence of mating system and overlapping generations on effective population size. Evolution 47:1329-1341. 1996. Influence of variation in female fecundity on effec- tive population size. Biol. J. Linn. Soc. 59: 411-425. 1999. The effective size of a hierarchically structured population. Evolution 53:1-10. Paetkau, D., R. Slade, M. Burden, and A. Estoup. 2004. Direct, real-time estimation of migration rate using assignment methods: a simulation-based exploration of accuracy and power. Mol. Ecol. 13:55-65. Patterson, W. F., J. C. Watterson, R. L. Shipp, and J. H. Cowan Jr. 2001. Movement of tagged red snapper in the northern Gulf of Mexico. Trans. Am. Fish. Soc. 130:533-545. Piry, S., A. Alapetite, J. M. Cornuet, D. Paetkau, L. Baudouin, and A. Estoup. 2005. GeneClass2: a software for genetic assignment and first generation migrants detection. J. Hered. 95:536-539. Pruett, C. L., E. Saillant, and J. R. Gold. 2005. Historical population demography of red snap- per iLutjanus campechanus} from the northern Gulf of Mexico based on analysis of sequences of mitochondrial DNA Mar. Biol. 147:593-602. Rannala, B., and J. L. Mountain. 1997. Detecting immigration by using multilocus genotypes. Proc. Nat. Acad. Sci. USA 94:9197-9201. Raymond, M., and F. Rousset. 1995. Genep(ip (version 1.2): a population genetics soft- ware for exact test and Ecumenicism. J. Hered. 86: 248-249. Rezak, R., T. J. Bright, and D. W. McGrail. 1985. Reefs and banks of the northwestern Gulf of Mexico, 259 p. John Wiley and Sons, New York, NY. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-225. Schultz, S. T., and M. Lynch. 1997. Mutation and extinction: the role of variable muta- tional effects, synergistic epistasis, beneficial mutations, and degree of outcrossing. Evolution 51:1363-1371. Shulman, M. J., and E. Bermingham. 1995. Early life histories, ocean currents, and the pop- ulation genetics of Caribbean reef fishes. Evolution 49:897-910. Szedlmayer, S. T. 1997. LTltrasonic telemetry of red snapper, Lutjanus campechanus, at artificial reef sites in the northeast Gulf of Mexico. Copeia 1997:846-850. Szedlmayer, S. T., and R. L. Shipp. 1994. Movement and growth of red snapper, Lutjanus campechanus, from an artificial reef area in the north- eastern Gulf of Mexico. Bull. Mar. Sci. 55:887-896. Tipping, J. M. 1991. Heritability of age at maturity. North Am. J. Fish. Manag. 11:105-108. Trippel, E. A. 1995. Age at maturity as a stress indicator of fish- eries. BioScience 45:759-771. Turner, T. F., J. P. Wares, and J. R. Gold. 2002. Genetic effective size is three orders of magnitude smaller than adult census size in an abundant, estua- rine-dependent marine fish tSciaenops ocellatus). Ge- netics 162:1329-1339. Vucetich, J. A.. T. A. Waite. and L. Nunney. 1997. Fluctuating population size and the ratio of effective to census population size. Evolution 51: 2017-2021. Waite, T. A., and P. G. Parker. 1996. Dimensionless life histories and effective popula- tion size. Conserv. Biol. 10:1456-1462. Wang, J. 2001. A pseudo-likelihood method for estimating effective population size from temporally spaced samples. Genet. Res. 78:243-257. Wang, J., and M. C. Whitlock. 2003. Estimating effective population size and migration rates from genetic samples over space and time. Ge- netics 163:429-446. Waples, R. S. 1989. A generalized approach for estimating effec- tive population size from temporal changes in allele frequency. Genetics 121:379-391. Weir, B. S., and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of popula- tion structure. Evolution 38:1358-1370. Saillant and Gold: Population structure and variance effective size of Lutianus campechanus in the northern Gulf of Mexico 145 Whitlock, M. C, and N. H. Barton. 1997. The effective size of a subdivided population. Ge- netics 146:427-441. Wilson. C. A., and D. L. Nieland. 2001. Age and growth of red snapper, Lutjanus campechanus, from the northern Gulf of Mexico off Louisiana. Fish. Bull. 99:65.3-664. Woods, M. K., A. J. Fischer, J. H. Cowan, Jr., and D.L. Nieland. 2003. Size and age of female red snapper Lutjanux campechanus in the northern Gulf of Mexico. Proc. Gulf Caribb. Fish. Inst. 54:526-537. Appendix Table 1 Summary statistics at 19 nuclear-encoded microsatellite loci for the 1995 cohort of red snapper (Lutjanus campechanus} sampled at three localities in the northern Gulf of Mexico, n is sample size, no. of A is the number of alleles, A^ is allelic richness, H^ is gene diversity (expected heterozygosity), Pfj^y is probability of conforming to expected Hardy- Weinberg genotypic proportions, and Fjg is an inbreeding coefficient measured as Weir and Cockerham's ( 1984 ) /; Boldface indicates significant departures from HW equilibrium following (sequential) Bonferroni correction. Locus Texas Louisiana Alabama Locus Texas Louisiana Alabama Lca20 Pr.s240 n 199 286 373 n 140 276 372 No. of A 5 5 6 No. of A 20 21 23 Ar 3.39 3.37 3.00 An 15.84 14.94 14.96 He 0.170 0.215 0.172 He 0.917 0.901 0.885 ^HW 0.088 0.816 0.007 ^HW 0.010 0.129 0.096 F,s 0.053 -0.009 0.097 F,s 0.065 0.079 -0.012 Lca22 Prs248 n 198 281 376 n 195 285 372 No. of A 14 17 14 No. of A 21 21 23 ^H 8.45 9.62 8.92 An 13.40 12.61 12.88 He 0.686 0.741 . 0.712 He 0.889 0.851 0.874 P 0.176 0.317 0.013 ^HW 0.468 0.393 0.291 F,s 0.013 0.002 0.055 P,s -0.010 0.006 0.047 Lca43 Prs257 n 202 275 340 n 165 273 269 No. of A 10 11 9 No. of A 16 16 16 An 6.41 5.98 6.19 An 12.95 12.57 12.50 H, 0.535 0.553 0.530 He 0.903 0.909 0.904 ^HW 0.585 0.763 0.669 ^HW 0.3.50 0.140 0.392 F,, -0.028 -0.006 0.017 Fis 0.021 0.013 0.005 Lea 64 Prs260 ;? 197 286 377 n 189 283 376 No. of A 12 14 13 No. of A 4 5 7 An 7.19 7.54 6.97 A„ 3.45 3.39 3.50 H, 0.777 0.778 0.764 He 0.361 0.390 0.339 P 0.239 0.684 0.028 Phw 0.507 0.285 0.185 F,s 0.027 -0.025 0.014 Fis -0.011 -0.005 -0.045 continued 146 Fishery Bulletin 104(1) Appendix Table 1 (continued) Locus Texas Louisiana Alabama Locus Texas Louisiana Alabama Lco91 Prs275 n 201 285 375 /) 199 286 374 No. of A 6 7 7 No. of A 9 10 9 Ar 4.49 4.16 4.43 Ar 5.46 4.91 5.25 He 0.608 0.559 0.580 He 0.635 0.595 0.590 ^HW 0.002 0.957 0.162 ^HW 0.000 0.604 0.183 F,s 0.002 -0.066 -0.021 F,s 0.105 -0.014 0.011 Prs 55 Prs303 n 184 277 377 n 200 285 374 No. of A 8 7 9 No. of A 7 13 11 Ar 3.30 3.82 3.60 Ar 5.02 5.77 5.29 He 0.158 0.228 0.209 He 0.365 0.416 0.375 ^HW 0.353 0.326 0.102 ^HW 0.785 0.559 0.007 F,s -0.030 0.001 0.073 F,s -0.029 -0.012 -0.026 Lea 107 Prs282 ;; 189 286 375 n 202 285 377 No. of A 11 12 11 No. of A 14 14 14 Ar 8.67 8.59 7.90 Ar 8.57 8.62 8.06 He 0.809 0.806 0.796 He 0.664 0.669 0.623 ^HW 0.871 0.451 0.776 "hw 0.311 0.039 0.066 F,s 0.013 0.006 -0.031 F,s -0.006 0.072 -0.035 Prsl37 Prs328 n 201 286 376 n 200 286 377 No. of A 13 13 17 No. of A 6 8 6 Ar 7.83 7.92 8.33 Ar 3.70 4.06 3.53 He 0.706 0.700 0.711 He 0.555 0.557 0.557 ^HW 0.103 0.331 0.000 ^HW 0.008 0.002 0.034 F,s 0.049 0.071 0.125 F„ 0.072 -0.086 0.020 Prs221 Prs333 n 197 282 376 n 202 283 371 No. of A 16 20 19 No. of A 8 6 8 Ar 9.78 10.26 9.73 Ar 4.22 3.98 4.70 He 0.791 0.802 0.792 He 0.288 0.294 0.371 'l]\V 0.043 0.037 0.102 ■P//IV 0.008 0.638 0.002 F,s 0.018 0.050 0.053 F,s 0.055 0.013 0.128 Prs229 n 201 285 376 No. of A 8 7 8 Ar 5.96 5.44 5.39 He 0.470 0.508 0.486 P ' HV, 0.689 0.392 0.782 F„ 0.038 0.088 0.005 Saillant and Gold; Population structure and variance effective size of Lut/anus campechanus in ttie northern Gulf of Mexico 147 Appendix Table 2 Summary statistics at 19 nuclear-encoded microsatellite loci tor tfie 1997 cohort of red snapper iLutjanus campechanus) sampled at three localities in the northern Gulf of Mexico, ii is sample size. No. of A is number of alleles, A,j is allelic richness, Hg is gene diversity lexpected heterozygosity). P,,u is probability of conforming to expected Hardy-Weinberg genotypic proportions, and F,j; is an inbreeding coefficient measured as Weir and Cockerhani's 1 1984 1 f. Boldface indicates significant departures from HW equilibrium following (sequential) Bonferroni correction. Locus Texas Louisiana Alabama Locus Texas Louisiana Alabama Lca20 Prs240 n 211 272 269 n 188 234 240 No. of A 6 6 6 No. of A 20 20 22 A« 3.72 3.35 3.78 ^H 14.22 15.37 14.34 He 0.238 0.184 0.206 He 0.897 0.898 0.885 ^HW 0.121 0.561 0.009 ^HW 0.001 0.012 0.317 F,s 0.042 0.043 0.082 F,s 0.092 0.043 -0.021 Lca22 Prs248 n 208 244 266 n 211 271 272 No. of A 16 14 15 No. of A 21 22 21 An 9.88 9.36 9.45 Afl 12.58 12.91 12.67 liE 0.769 0.757 0.771 He 0.872 0.867 0.882 ^HW 0.000 0.892 0.086 Phw 0.176 0.616 0.334 F,s 0.106 0.004 -0.009 F,s 0.001 0.030 0.017 Lca43 Prs257 n 210 272 272 n 206 266 246 No. of A 8 12 11 No. of A 17 17 18 Ar 6.22 6.48 6.19 A« 13.24 12.86 13.51 He 0.587 0.536 0.528 He 0.908 0.898 0.915 ^HW 0.325 0.669 0.981 ^HW 0.282 0.113 0.464 F,s 0.010 -0.049 -0.003 Pis 0.011 0.008 0.005 Lca64 Prs260 n 211 271 271 n 211 272 272 No. of A 11 13 11 No. of A 6 6 6 Ar 7.33 6.93 6.90 A« 3.70 3.41 3.88 He 0.784 0.765 0.769 He 0.367 0.344 0.429 ^HW 0.086 0.749 0.495 Phw 0.275 0.780 0.311 Fjs 0.027 0.020 0.012 Fis -0.019 0.026 -0.002 Lca9l Prs275 n 202 268 262 n 211 272 273 No. of A 7 8 8 No. of A 7 9 8 Ah 4.22 4.38 4.43 A« 5.00 5.07 4.61 He 0.560 0.575 0.570 He 0.608 0.612 0.579 ^HW 0.895 0.927 0.005 ^HW 0.711 0.334 0.441 ^,.S -0.070 0.039 0.030 F,s 0.034 0.015 0.031 Lea 107 Prs282 n 211 264 269 n 211 272 273 No. of A 10 11 11 No. of A 13 12 12 Afl 7.90 8.07 8.15 Aa 8.46 8.40 7.89 He 0.799 0.798 0.775 He 0.636 0.639 0.614 ^H\V 0.249 0.669 0.346 ^HW 0.886 0.556 0.141 F,s -0.104 -0.015 -0.045 F,s -0.051 0.028 0.022 continued 148 Fishery Bulletin 104(1) Appendix Table 2 (continued) Locus Texas Louisiana Alabama Locus Texas Louisiana Alabama Prs55 Prs303 n n 211 272 270 No. of A 7 6 6 No. of A 10 9 12 Aft 4.34 3.62 3.52 Ar 5.33 5.17 6.05 He 0.266 0.210 0.221 He 0.375 0.400 0.400 ^HW 0.100 0.525 0.199 'HW 0.527 0.344 0.781 F,s -0.017 -0.052 0.051 Pis -0.010 -0.011 -0.055 Prsl37 Pre328 n 211 272 271 11 211 272 273 No. of A 13 13 12 No. of A 6 6 5 -^ft 7.88 7.43 8.00 Aft 3.54 3.45 3.71 He 0.721 0.694 0.715 He 0.542 0.545 0.568 ^HW 0.127 0.001 0.051 ^HW 0.323 0.108 0.191 F,s 0.008 0.105 0.019 F,s 0.090 -0.018 0.007 Prs221 Prs333 n 211 271 270 ;i 211 272 272 No. of A 19 18 17 No. of A 6 7 6 Aft 10.06 9.54 10.32 Aft 3.84 4.52 4.20 He 0.800 0.792 0.802 He 0.342 0.320 0.323 Phw 0.108 0.006 0.797 ^HW 0.571 0.819 0.412 F,s 0.016 0.100 -0.025 F,s 0.029 -0.022 0.032 Prs229 /! 211 271 269 No. of A 7 9 9 Aft 5.05 5.08 5.51 »£ 0.495 0.464 0.527 ^HW 0.001 0.000 0.020 ^/.S 0.081 0.181 0.118 149 The relationship between smolt and postsmolt growth for Atlantic salmon iSalmo salar) in the Gulf of St. Lawrence Kevin D. Friedland UMass/NOAA Cooperative Marine Education and Research Program Blalsdell House University of Massachusetts Amherst, Massachusetts 01003 Present address: National Marine Fisheries Service 28 Tarzwell Dr Narragansett, Rhode Island 02882 E-mail address: kevinfriedlandiginoaagov Lora M. Clarke Department of Natural Resources Conservation Holdsworth Hall University of Massachusetts Amherst, Massachusetts 01003 Jean-Denis Dutil Ministere des Peche et des Oceans Institut Maurice-Lamontagne 850 route de la Mer, Mont-Joli Quebec, G5H 3Z4, Canada Matti Salminen Finnish Game and Fisheries Research Institute Viikinkaari 4 P.O. Box 2, FIN-00791 Helsinki, Finland The interaction of ocean climate and growth conditions during the postsmolt phase is emerging as the primary hypothesis to explain pat- terns of adult recruitment for indi- vidual stocks and stock complexes of Atlantic salmon (Sabno salar). Friedland et al. (1993) first reported that contrast in sea surface tempera- ture (SST) conditions during spring appeared to be related to recruitment of the European stock complex. This hypothesis was further supported by the relationship between cohort specific patterns of recruitment for two index stocks and regional scale SST (Friedland et al., 1998). One of the index stocks, the North Esk of Scotland, was shown to have a pat- tern of postsmolt growth that was positively correlated with survival. indicating that growth during the postsmolt year controls survival and recruitment (Friedland et al., 2000). A similar scenario is emerging for the North American stock complex where contrast in ocean conditions during spring in the postsmolt migra- tion corridors was associated with the recruitment pattern of the stock complex (Friedland et al., 2003a, 2003b). The accumulation of addi- tional data on the postsmolt growth response of both stock complexes will contribute to a better understanding of the recruitment process in Atlantic salmon. Anadromous salmonids produce co- horts of juvenile smolts that migrate over a short period of time at a nearly uniform size; however, the variability that occurs in migration timing and the size spectra of smolts is of poten- tial interest in the study of recruit- ment (Hoar, 1976; Wedemeyer et al., 1980). Enhancement and restoration programs provide data on the effect of smolt size on return rate where hatchery practices both intentionally and unintentionally produce fish of varying size and quality. There are many case studies that show positive correlations between smolt size and return rate in salmonids, but non- significant relationships have also been observed, which underscore the fact that the contrast in size that can be achieved in hatcheries is often outside ecologically relevant limits (Farmer, 1994; Salminen et al., 1994). The functional relation- ship between smolt size and return rate is probably more accurately de- scribed as nonlinear and as having some optimality within the range of hatchery releases (Bilton et al., 1982; Henderson and Cass, 1991). Researchers have also considered the effect of smolt size at ocean entry and the effect of smolt size on the en- suing growth patterns of postsmolts. Ward and Slaney (1988) described a positive relationship between return rate and smolt size for rainbow trout iSaliJio gairdneri). suggesting that the recruitment rate of a year class was mediated by the mortality that occurred when the fish migrated to sea. When new data were added to that relationship, the original con- clusions were no longer supported, indicating that factors other than smolt size at ocean entry contribute to recruitment (Ward, 2000). Al- though size at ocean entry may con- tribute to mortality risk, postsmolt growth, and the biotic and abiotic factors affecting postsmolt growth, often play a dominant role (Salmin- en et al., 1995). However, what has remained obscure is whether size at ocean entry influences postsmolt growth. Skilbrei (1989) and Nicieza and Branfia (1993) reported negative relationships between smolt size and Manuscript submitted 4 February 2004 to the Scientific Editor's Office. Manuscript approved for publication 10 April 2005 by the Scientific Editor. Fish. Bull. 104:149-155 (2006). 150 Fishery Bulletin 104(1) marine growth in Atlantic salmon, whereas Lundquist et al. (1988) and Salminen (1997) reported positive relationships. Einum et al. (2002) reported an inverse relationship between pre- and postsmolt growth of At- lantic salmon and suggested that phenotypic charac- teristics favoring growth in one environment may not necessarily favor growth in another. The body of mixed results sheds little light on whether smolt size confers a growth advantage to postsmolts. It also leaves in doubt the importance of freshwater experience on postsmolt survival. Common to this body of work is the depen- dence on samples coming from river returns and fishery catches that may not be representative of the full range of growth signatures in postsmolt populations because they do not include data from mortalities and often do not include data from some return age groups. In this study we report on an analysis of scale growth indices from a collection of postsmolts from the Gulf of St. Lawrence. We measured freshwater growth signa- tures representative of smolt size and freshwater growth prior to migration and postsmolt growth was partitioned by season. These are samples of a life stage in an in- termediate phase of marine life that is infrequently sampled and may be free of some of the assumed biases associated with samples from spawning fish. Figure 1 Postsmolt scale of Atlantic salmon iSalmo aalar) in the Gulf of St. Lawrence show- ing focus, freshwater growth zone, and postsmolt growth zone. Material and methods We collected data on scale circuli spacing that was rep- resentative of the freshwater and postsmolt growth for juvenile salmon captured in the Gulf of St. Lawrence. These postsmolt salmon were collected in 1982-84 and were originally reported in Dutil and Coutu (1988). They were captured in experimental gill nets along the northwest shore of the Gulf during the months of August to October. Freshwater and postsmolt growth descriptors were in- terpreted from circuli spacing patterns deposited on the scale. Scales were cleaned and mounted between glass slides and the spacings of scale circuli were measured with a Bioscan Optimas (Media Cybernatics, Inc., Sil- ver Spring, MD) image processing system. Freshwater zone length (FZL) was taken as the total distance from the center of the scale focus to the transition between freshwater and postsmolt growth (Fig. 1). This tran- sition zone is defined by the appearance of the first marine intercirculus spacing, i.e., as a wider spacing (according to the reader's judgment) than the progres- sion of spacings in the freshwater zone. The FZL ends at the last freshwater circulus. FZL was interpreted as a proxy for smolt length at migration. The mean of the last five circuli spacings during the freshwater phase was computed for each sample (CSLF, circuli spacing during last freshwater period). This circuli spacing index was interpreted as an indication of freshwater finishing growth (the final phases of freshwater growth prior to migration). Two circuli spacing indices were extracted from the postsmolt zone, the mean spacing between intercirculi 2 through 6 — an interval which was interpreted as the growth index for the early ma- rine period (CSFM, circuli spacing during first marine period) — and the mean intercirculi spacing of the next five pairs occurring later in the postsmolt growth — an interval interpreted as an index of summer growth (CSSM, circuli spacing during summer marine period). The total length of the postsmolt growth zone was also extracted (MZL, marine zone length), which was the distance from the freshwater-marine transition, starting at the last freshwater circulus to the outer edge of the scale. All measurements were made on a single scale from each specimen along the 360° axis of the scale. We tested the functional relationship between fresh- water and marine growth, early and late marine growth, and seasonal progression of the growth zones using linear regression. We tested the significance of the slope parameters for each relationship as an indica- tion of variable dependency. We also considered whether these data might contrib- ute to our understanding of size selective mortality by constructing a time series plot of FZL and MZL against date of capture. An anticipated positive trend in MZL would reflect growth during the marine phase, but any trend in FZL would reflect size-specific removals at- tributable to size at ocean entry. Results Measurements were taken from the scales of 587 posts- molts collected during 1982-84. FZL averaged 0.67 mm (SD = 0.163) for all samples. CSLF averaged 0.022 mm (SD = 0.0049), reflecting the slower growth and more narrowly spaced circuli of the freshwater zone. The two NOTE Fnedland el al Relationship between smoll and postsmolt growth in Salmo solar circuli spacing indices from the postsmolt growth zone were larger in magnitude than the index from the freshwater zone. CSFM averaged 0.061 mm (SD = 0.0109) and reflected a threefold increase in circuli spacing between the freshwater and marine zones. CSSM averaged 0.064 mm (SD = 0.0100), reflecting the increased growth occurring within the duration of marine residency'. MZL increased over time from approximately 0.6 mm in early August to 1.2 mm by mid-October. The data we examined to test the relation- ship between smolt size and freshwater finishing growth and postsmolt growth indicate an absence of any linkage between the two growth environ- ments. We found no significant relationships be- tween smolt sizes as represented by FZL and ei- ther spring (CSFM) or summer (CSSM) postsmolt growth as represented by the circuli spacing indi- ces. The scatter between CSFM and FZL suggests there were no linear trends for any of the study years (Fig. 2A). None of the three regressions had slopes significantly different from zero (Table 1). There was also an absence of any relationship between CSSM and FZL as evidenced by a similar pattern of scatter for the coordinates (Fig. 2B) and by an absence of significant slopes. The relationship was further tested by consid- ering the growth that occurs just prior to smolt migration as an indication of smolt condition at entry into the marine environment. As was the case with the bivariate relationships with FZL, CSLF was not a significant predictor of either CSFM or CSSM (Fig. 3). None of the slopes were significantly different from zero (Table 1). Growth in the early and later part of the post- smolt season were correlated. There were signifi- cant positive linear relationships between CSSM and CSFM for all three years (Fig. 4), indicating that the postsmolts began their marine residence with rapid growth and continued to grow at a rapid rate through summer. The slopes of all three regressions were sig- nificantly different from zero (Table 1). A time series plot of MZL, plotted by date of capture, shows an increasing trend reflecting the growth that postsmolts experience during early marine residence (Fig. 5). For the same fish, we also plotted FZL by date of capture and found a negative trend over time, indi- cating that FZL decreased during the period (Table 1), which would not be consistent with selective mortality of smolts of smaller initial size. Discussion Our main findings indicate that marine growth of post- smolt Atlantic salmon sampled from August to October in the Gulf of St. Lawrence was independent of fresh- water growth history. Neither smolt size nor the last freshwater growth period of smolts prior to migration was related to postsmolt growth patterns. Furthermore, 0,12 0 10' 0 08- 0,06 ■ £■ 0,04. E 9- 0 12- O 0,10' 0.08- 0.06 0 04 1982 1983 1984 ■1982 1983 •1984 o -o 0 0 „-,=•*- V^'-.-Jf _i_.'-- — qiij ^y^T* , ■• sf-®' - B * » ° ooo.\s«' ce o ■* O 0,2 — I — 0,6 — I — 1,0 0,4 0,6 0.8 1,0 1,2 Freshwater zone length (mm) 1,4 Figure 2 Relationships between circuli spacing during the early marine period (A) and circuli spacing during the summer marine period (B) and freshwater zone length for Atlantic salmon tSalmo salar) in the Gulf of St. Lawrence. Table 1 Linear associations between scale growth indices and date of capture for Gulf of St. Lawrence Atlantic salmon iSalnw salar) postmolts. FZL = freshwater zone length; CSLF=mean of the last five freshwater intercirculi spac- ings; CSFM=mean spacing of intercirculi spacings 2 through 6; CSSM = mean spacing of intercirculi spacings 7 through 11; DOC = date of capture. Linear associations Probability Hg slope = 0 Predictor iX) Dependent (7) 1982 1983 1984 FZL FZL CSLF CSLF CSFM DOC DOC Sample size CSFM CSSM CSFM CSSM CSSM FZL MZL 0.53 0.50 0.27 0.33 <0.01 <0.01 <0.01 374 0.87 0.61 0.70 0.57 <0.01 158 0.57 0.22 0.22 0.27 <0.01 55 152 Fishery Bulletin 104(1) there was no evidence that size at the smolt stage of postsmolts captured later in the season was larger than those captured earlier in the season, indicating that there was no size-selective attrition on smaller smolts during August to October. The strength of these data is that the samples were derived from collections of posts- molts, not river returns or fishery catches, and thus it is less biased and more representative of natural mortality factors. It will take considerably more data to fully test this hypothesis, but the highest priority should be given to similar sample collections, especially those collected earlier in the year (Holm et al., 2000). These data will be pivotal in interpreting the relative impact of size at ocean entry versus postsmolt growth for determing postsmolt survival. Understanding survival during the postsmolt period is of great importance in the face of declining stock abundance and possible stock extirpations over por- tions of the range of Atlantic salmon (Anderson et al., 2000). With evidence showing the success of conserva- tion efforts to increase freshwater populations (Swans- burg et al., 2002), the range of potential life stages 0.12- 0,10' 0.08- 0.06 ■ 0.04 0.12 O 0.10 0.08- 0.06 0.04 1982 1983 1984 1982 1983 1984 B „«oe o**.>-?' ^^^■■^ if-- 0.01 0.05 0.02 0.03 0.04 Circuli spacing, late freshwater period (mm) Figure 3 Relationships between circuli spacing during the early marine period (A) and circuli spacing during the summer marine period (B) and circuli spacing during the late freshwater period for Atlantic salmon (Salmo salar) in the Gulf of St. Lawrence. and habitats causing recruitment failure is reduced. Evidence at the population level showing coherence be- tween early marine growth and survival at sea is still a long way from providing sufficient evidence to fully describe survival mechanisms or climate linkages. The hypothesized climate forcing appears to be related to a mismatch between ocean entry time for smolts and the thermal regime they find during their transition to marine life (Friedland et al., 2003b). The concept is supported by the growth response of postsmolts over a range of temperatures that they would likely encounter during that period of time. Atlantic salmon postsmolts follow a nonlinear growth response to temperature and optimum growth occurs at 13°C (Handeland et al., 2003). Variations in migration timing could result in fish encountering ocean conditions that are too warm or cool; these variations could explain the different associations of North American and European stocks to ocean temperatures. It would also be interesting to know if the thermal growth optimum is robust to vary- ing feeding rations, which may be an issue for regional stock groups in the North Atlantic. The transition period from freshwater to ma- rine life is characterized by a number of dy- namic changes in growth and predation. The highest mortality rates for postsmolts occur during the first weeks at sea (Eriksson, 1994). As a consequence, even a minimal variation in these rates will have a large impact on adult recruitment. Size-selective mortality on juve- niles has been demonstrated for other salmo- nids such as sockeye salmon {Ojjcorhynchiis nerka) and has been shown to be seasonally concentrated (West and Larkin, 1987). Size- specific predation often occurs but predators, such as adult bird species, are not likely to grow at corresponding rates to those of the fish (Dieperink et al., 2002). Thus, size at ocean entry and the ability to grow out of the size range vulnerable to specific predators must affect the dynamics of predation rate. The pe- riod of predation vulnerability is mediated by growth rate; therefore the argument returns to what confers faster growth on postsmolts in some years? One dynamic that challenges all stocks is the transition from predominantly invertebrate to piscivorous prey. Some authors suggest it is the success in making this transition, or the time it takes to make the transition, that dictates the amount of time the postsmolts take to grow out of the predation-vulnerable size range (Du- til and Coutu, 1988; Andreassen et al., 2001; Salminen et al., 2001). Depending on prevail- ing foraging conditions, this transition may be promoted by increased smolt size, as was the case in the northern Baltic Sea in 1990-93 (Salminen et al., 2001). One potential expla- nation for the contradictory views on the rela- tionship between pre- and postsmolt growth in NOTE Fnedland et al : Relationship between smolt and poslsmolt growth in Salmo solar 153 Atlantic salmon may arise from geographical dif- ferences or year-to-year changes (or both) in prey regimes during the critical diet-transition period. In lake-run brown trout, Niva and Jokela (2000) found a positive correlation between hatchery growth rate and postsmolt growth in a lake with small fish as the main prey. In contrast, hatchery growth did not predict growth in a lake where the fish had to prey on bottom-dwelling invertebrates. This finding may indicate that assessment of in- dividual performance may be highly specific to environments in migratory salmonids. Size of postsmolts at ocean entry in salmonids also manifests itself in predictable patterns of life history variation. In sea trout (Salmo triitta) there is a close correspondence between size of juvenile stage and adult size, which is believed to have an impact on recruitment through size variation at critical life stages (Elliott, 1985). Hutchings and Jones (1998) found that smolt age and size had a small effect on the growth-rate threshold for maturity in Atlantic salmon. Smolt-size variation in sea trout has been associated with latitude, indicating that for this species recruitment strate- gies are likely adapted to local physical conditions, predators, and prey regimes (L' Abee-Lund et al., 1989). Wild smolts are often smaller than hatch- ery smolts for the same or allied strains, yet they consistently survive at higher rates (Poole et al., 2003). Obviously there are other cofactors that can explain why the wild fish overcome any advantage conferred on hatchery fish by their larger size at migration. The wild versus hatchery comparison is a relatively weak piece of evidence to use to dismiss the idea that smolt size is an important survival factor. However, a recent study of size at ocean entry in a Quebec stock of wild fish may provide more useful evidence (Dodson^ Caron-). The investigators found that the size spectra of smolts were not different from the size spectra of back-calculated smolt sizes for returning adults, suggesting some factor other than size at ocean entry was controlling survival. It should be noted that postsmolts collected in the present study were of unknown origin, but the scale pattern for the freshwater zone clearly indicates that they were of wild origin. The growth response of postsmolts may be gov- erned by a simple response to temperature optima occurring in coastal marine waters, but it may al- so be more complex than that. The surface waters in the Gulf of St. Lawrence circulate in wide-scale Dodson, J. 2004. Personal comniun. Departement de Biologie Pavilion Alexandre-Vachon, local .30.58-B Universite Laval, Quebec, GIK 7P4. Canada. Caron. F. 2004. Personal commun. Direction de la Recherche Faune et Pares Quebec, 67.5 est, Boul. Rene-Levesque Boite 92, lie etage, Quebec, GIR .5V7, Canada. 012- 0-10- 0.08- 0.06- 0,04- 1982 1983 1984 -1982 1983 1984 0.02-1 1 1 1 1 1 1 1 1 1 r 0.02 0.04 0,06 0.08 010 0 12 Circuli spacing, early marine period (mm) Figure 4 Relationships between circuli spacing during the summer marine period and circuli spacing during the early marine period for Atlantic salmon (Salinn salar) in the Gulf of St. Lawrence. e 0 9 03- Freshwater Marine •»•••• . . • ' « z: Aug 6 Aug 20 Sep 3 Sep 17 Date of capture — r- Oct • Figure 5 Freshwater zone length and marine zone length, by date of capture during the 1982 field season for Atlantic salmon t Salmo salar) in the Gulf of St. Lawrence. 154 Fishery Bulletin 104(1) gyres over the extremely cold waters of the cold inter- mediate layer (CIL), thus creating a thermally complex habitat. Strong winds periodically force cold waters from the CIL to the surface (Ouellet, 1997). Postsmolts appear to be opportunistic feeders in the ocean, and despite their predominant use of surface waters, they have been known to use benthic habitats and water column features as well (Levings, 1994; Sturlaugsson, 1994). These behaviors would distribute fish in a man- ner independent of surface temperatures because forage behaviors are not necessarily controlled by temperature preferences. Growth in postsmolts is also affected by other physical parameters, such as photoperiod (Fors- berg, 1995), and possibly by variation in sea surface salinity as shown for chum salmon (Oncorhynchus keta) (Morita et al., 2001). We can predict migration timing and initial migration trajectories for postsmolts. but we are hard pressed to predict the physical nature of their habitats over time and the physiological effect of these habitats on postsmolt growth. Acknowledgments We thank T. Sadusky for help with the initial phases of this study and two anonymous reviewers for productive comments. Literature cited Anderson, J. M., F. G. Whoriskey. and A. Goode. 2000. Atlantic salmon on the brink. Endangered Spe- cies Update 17:15-21. Andreassen, P. M. R.. M. B. Martinussen, N. A. Hvidsten, and S. O. Stefansson. 2001. Feeding and pre-selection of wild Atlantic salmon post-smolts. J. Fish Biol. 58:1667-1679. Bilton, H. T., D. F. Alderdice, and J. T. Schnute. 1982. Influence of time and size at release of juvenile coho salmon [Oncorhynchus kisutch) on returns at maturity. Can. J. Fish. Aquat. Sci 39:426-447. Dieperink, C., B. D. Bak, L. F. Pedersen, M. I. Pedersen, and S. Pedersen. 2002. Predation on Atlantic salmon and sea trout during their first days at sea as postsmolts. J. Fish Biol. 61:848-852. Dutil, J.-D., and J.-M. Coutu. 1988. Early marine life of Atlantic salmon, Salmo salar, postsmolts in the northern Gulf of St. Lawrence. Fish. Bull. 86:197-212. Einum, S., E. B. Thorstad, and T. F. Nssje. 2002. Growth rate correlations across life stages in female Atlantic salmon. J. Fish Biol. 60:780-784. Elliott, J. M. 1985. Growth, size, biomass, and production for dif- ferent life stages of migratory trout Salmo trutta in a Lake District stream, 1966-83. J. Animal Ecol. 54:985-1001. Eriksson, T. 1994. Mortality risks of Baltic salmon during downstream migration and early sea-phase: effects of body size and season. Nord. J. Freshw. Res. 69:100. Farmer, G. J. 1994. Some factors which influence the survival of hatch- ery Atlantic salmon [Salmo salar) smolts utilized for enhancement purposes. Aquaculture 121:223-233. Forsberg, O. I. 1995. Empirical investigation on growth of post-smolt Atlantic salmon iSalmo salar L.) in land-based farms. Evidence of photoperiodic influence. Aquaculture 133:235-248. Friedland, K. D., L. P. Hansen, and D. A. Dunkley. 1998. Marine temperatures experienced by postsmolts and the survival of Atlantic salmon (Salmo salar L.) in the North Sea area. Fish. Oceanogr. 7:22-34. Friedland, K. D., L. P. Hansen, D. A. Dunkley, and J. C. MacLean. 2000. Linkage between ocean climate, post-smolt growth, and survival of Atlantic salmon (Salmo salar L. I in the North Sea area. ICES J. Mar. Sci. 57:419-429. Friedland, K. D., D. G. Reddin, D. G., and M. Castonguay. 2003a. Ocean thermal conditions in the post-smolt nurs- ery of North American Atlantic salmon. ICES J. Mar. Sci. 60:343-355. Friedland, K. D., D. G. Reddin, and J. F. Kocik. 1993. Marine survival of North American and Euro- pean Atlantic salmon: effects of growth and environ- ment. ICES J. Mar. Sci. 50:481-492. Friedland, K. D., D. G. Reddin, J. R. McMenemy, and K. F. Drinkwater. 2003b. Multideeadal trends in North American Atlantic salmon (Salmo salar) stocks and climate trends rel- evant to juvenile survival. Can. J. Fish. Aquat. Sci. 60:563-583. Handeland, S. O., B. T. Bjbrnsson, A. M. Arnesen, and S. O. Stefansson. 2003. Seawater adaptation and growth of post-smolt Atlantic salmon (Salmo salar) of wild and farmed strains. Aquaculture 220:367-384. Henderson, M. A., and A. J. Cass. 1991. Effect of smolt size on smolt-to-adult survival for Chilko Lake sockeye salmon (Oncorhynchus nerka ). Can. J. Fish. Aquat. Sci. 48:988-994. Hoar. W. S. 1976. Smolt transformation: evolution, behaviour, and physiology. J. Fish. Res. Board Can. 33:1233-1252. Holm, M., J. C. Hoist, and L. P. Hansen. 2000. Spatial and temporal distribution of post-smolts of Atlantic salmon (Salmo salar L.) in the Norwegian Sea and adjacent areas. ICES J. Mar. Sci. 57:955-964. Hutchings, J. A., and M. E. B. Jones. 1998. Life history variation and growth rate thresholds for maturity in Atlantic salmon, Salmo salar. Can. J. Fish. Aquat. Sci. 55(suppl. l):22-47. Levings, C. D. 1994. Feeding behavior of juvenile salmon and significance of habitat during estuary and early sea phase. Nord. J. Freshwater Res. 69:7-16. L Abee-Lund, J. H., B. Jopnsson, A. J. Jensen, L. M. Saettem, T. G. Heggberget, B. . Johnson, T. F. Ntesje. 1989. Latitudinal variation in life-history characteris- tics of sea-run migrant brown trout Salmo trutta. J. Animal Ecol. 58:525-542. Lundquist, H., W. C. Clarke, and H. Johansson. 1988. The influence of sexual maturation on survival to adulthood of river stocked Baltic salmon, Salmo salar, smolts. Holarct. Ecol. 11:60-69. NOTE Fnedland et a\ Relationship between smolt and postsniolt growth in Salmo salai 155 Merita, S. H., K. Morita, and H. Sakano. 2001. Growth of chum salmon [Oncorhynchiis keta) correlated with sea-surface salinity in the North Pacific. ICES J. Mar. Sci. 58:1335-1339. Nicieza, A. G, and F. Braniia. 1993. Relationships among smolt size, marine growth, and sea age at maturity of Atlantic salmon iSahno salar) in northern Spain. Can. J. Fish. Aquat. Sci. 50:1632-1640. Niva, T., and J. Jokela. 2000. Phenotypic correlation of juvenile growth rate between different consecutive foraging environments in a salmonid fish: a field experiment. Evol. Ecol. 14:111-126. Ouellet, P. 1997. Characteristics and vertical distribution of Atlantic cod {Gadus morhua) eggs in the northern Gulf of St. Lawrence, and the possible effect of cold water tem- perature on recruitment. Can. J. Fish. Aquat. Sci. 54:198-210. Poole, W. R., D. T. Nolan, T. Wevers, M. Dillane, D. Cotter, and O. Tully. 2003. An ecophysiological comparison of wild and hatch- ery-raised Atlantic salmon (Salmo salar L.) smolt from the Burrishoole system, western Ireland. Aquaculture 222:301-314. Salminen, M. 1997. Relationships between smolt size, postmolt growth and sea age at maturity in Atlantic salmon ranched in the Baltic Sea. J. Appl. Ichthyol. 13:121-130. Salminen, M., E. Erkamo, and J. Salmi. 2001. Diet of post-smolt and one-sea-winter Atlantic salmon in the Bothnian Sea, Northern Baltic. J. Fish Biol. 58:16-35. Salminen, M., S. Kuikka, and E. Erkamo. 1994. Divergence in the feeding migration of Baltic salmon i Sal mo salar L.I; the significance of smolt size. Nord. J. Freshw. Res. 69:32-42. 1995. Annual variability in survival of sea-ranched Baltic salmon, Salmo salar L.: Significance of smolt size and marine conditions. Fish. Manag. Ecol. 2:171-184. Skilbrei, O. T. 1989. Relationships between smolt lengths and growth and maturation in the sea of individually tagged Atlantic salmon iSalmo salar). Aquaculture 83:95-108. Sturlaugsson, J. 1994. Food of ranched Atlantic salmon iSalmo salar L. I postsmolts in coastal waters. West Iceland. Nord. J. Freshw. Res. 69:43-57. Swansburg, E, G. Chaput, D. Moore, D. Cassie, and N. El-Jabi. 2002. Size variability of juvenile Atlantic salmon: links to environmental conditions. J. Fish Biol. 61:661-683. Ward, B. R. 2000. Declivity in steelhead ( Oncorhyncbus mykiss ) recruit- ment at the Keogh River over the past decade. Can. J. Fish. Aquat. Sci. 45:298-306. Ward, B. R., and P. A. Slaney. 1988. Life history and smolt-to-adult survival of Keogh River steelhead trout (So/mo gairdneri) and the rela- tionship to smolt size. Can. J. Fish. Aquat. Sci. 45: 1110-1122. Wedemeyer G.A., R. L. Saunders, and W. C. Clarke. 1980. Environmental factors affecting smoltification and early marine survival of anadromous salmon. Mar. Fish. Rev. 42:1-14. West, C. J., and P. A. Larkin. 1987. Evidence for size-selective mortality of juvenile sock- eye salmon ( Oncorhynchus nerka I in Babine Lake, British Columbia. Can. J. Fish. Aquat. Sci. 44:712-721. 156 Fishery Bulletin 104(1) Erratum Erratum: Fishery Bulletin 102(4): p. 568. Calambokidis, John, Gretchen H. Steiger, David K. Ellifrit, Barry L. Troutman, and C. Edward Bowlby Distribution and abundance of humpback whales (Megaptera nouaeangliae) and other marine mammals off the northern Washington coast Correction: Table 3 should read as follows be bo a. O too c to >. en w M C JS so CO ro g a> c 03 s "3 s ^ = c cc .5 2 be c ra -: E 2 c o to Z be 2 be fM r- -^T CO O CO C^) lO "* 50 CO (>] CO CD T-H 00 ^ CO 0) 1 Ji a. CO 3 > X " ^ c S O § CQ CO O^ CO CD CT; Tj- OJ O C30 CO I> CO Ol CO OJ CD CO CO CJl 00 00 X ^ G^ CO CO CD Tt ^ CD CO Tji I> Tf ^ ^ CO -H ^ a o "ffl s . . at flJ 1- a; m S T3 .g ^ CO m 2: (X o o CO ^H ic O^ CO -cf^ C^ br r o, c br c Cfl a '-- 157 Fishery Bulletin Guidelines for authors Content of manuscripts Contributions published in Fishery Bulletin describe original research in marine fishery science, fishery engi- neering and economics, as well as the areas of marine environmental and ecological sciences (including model- ing). Although all contributions are subject to peer review, responsibility for the contents of papers rests upon the authors and not upon the editor or publisher. Submission of an article implies that the article is original and is not being considered for publication elsewhere. Manuscripts must be written in English. Authors whose native lan- guage is not English are strongly advised to have their manuscripts checked by English-speaking colleagues prior to submission. Articles may range from relatively short contributions (10-15 typed and double-spaced pages, tables and figures not included) to extensive contributions (20-30 typed pages). Notes are reports of 5 to 10 pages without an abstract and describe methods or results not supported by a large body of data. Manuscript preparation Title page should include authors' full names and mailing addresses and the senior author's telephone, fax number, and e-mail add/-ess, and a list of key words to describe the contents of the manuscript. Abstract should be limited to 150 words (one-half typed page), state the main scope of the research, and emphasize the author's conclusions and relevant findings. Do not review the methods of the study or list the contents of the paper. Because abstracts are cir- culated by abstracting agencies, it is important that they represent the research clearly and concisely. Text must be typed in 12 point Times New Roman font throughout. A brief introduction should convey the broad significance of the paper; the remainder of the paper should be divided into the following sections; Materials and methods, Results, Discussion (or Conclusions), and Acknowl- edgments. 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, abbre- viations, and technical terms should be written out in full the first time they are used. Include FAO common names for species in the list of keywords and in the introduction. Regional common names may be used throughout the rest of the text if they are different from FAO common names which can be found at http;//www.fishbase.org/search. html. Follow the U.S. Government Printing Office Style Manual ( 1984 ed. ) and the CBE Style Manual ( 6th ed. ) for editorial style, and the most current issue of the Ameri- can Fisheries Society's Common and Scientific Names of Fishes from the United States and Canada for fish nomen- clature. Dates should be written as follows: 11 November 2000. Measurements should be expressed in metric units, e.g., 58 metric tons (t); if other units of measurement are used, please make this fact explicit to the reader. Write out the numbers zero through nine unless they form part of measurement units (e.g., nine fish but 9 mm). Text footnotes should be inserted in 9-point font at the bottom of the page that displays the first citation of the footnote. Footnotes should be formatted in the same manner as citations. Footnote all personal communica- tions, unpublished data, and unpublished manuscripts with full address of the communicator or author, or, as in the case of unpublished data, where the data are on file. Authors are advised to avoid references to nonstandard (gray) literature (such as internal, project, processed, or administrative reports, ICES Council Minutes, IWC Minutes or Working Papers, any "research" or "working" documents, laboratory or contract reports. Management Council reports, and manuscripts in review) wherever possible. If these references are used, present them as footnotes and list whether they are available from NTIS (National Technical Information Service) or from some other public depository. Cite all software and special equipment or chemical solutions used in the study, not in a footnote but within parentheses in the text (e.g., SAS, vers. 6.03, SAS Inst., Inc., Cary, NO. Literature cited comprises published works and those accepted for publication in peer-reviewed litera- ture (in press). Follow the name and year system for citation format. If there is a sequence of citations in the text, list chronologically; (Smith, 1932; Green, 1947; Smith and Jones, 1985). Abbreviations of serials should conform to abbreviations given in the Serial Sources for the BIOSIS Previews Database. Authors are respon- sible for the accuracy and completeness of all citations. Literature citation format; Author (last name, followed by first-name initials). Year. Title of report or manu- script. Abbreviated title of the series to which it belongs. Always include number of pages. If there is a sequence of citations by the same first author, list the works alphabetically according to the last name of following authors (e.g.. Smith G. P., L. C. Brown, 1982; Smith, G. P., and T. P. Stuart, 1982). If the authorship is identi- cal, list works chronologically. Tables and figures — general format • Zeros should precede all decimal points for values less than one. • Sample size, n, should be italicized. • Capitalize the first letter of the first word in all labels within figures. • Do not use overly large font sizes in maps and for units of measurements along axes in figures. • Do not use bold fonts or bold lines in figures. • Do not place outline rules around graphs. • Do not use horizontal lines in graphs to indicate mea- surement units on axes. 158 Fishery Bulletin 104(1) • Use a comma in numbers of five digits or more (e.g. 13,000 but 3000). • Maps should have a North arrow (or compass sign) and degrees latitude-longitude (e.g., 170 E) Tables are often overused in scientific papers; it is seldom necessary or even desirable to present all the data associated with a study. Tables should not be excessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All unusual symbols must be explained in the table legend. Other incidental comments may be footnoted with italic footnote markers. Use asterisks to indicate probability in statistical data. Do not type table legends on a separate page; place them on the same page as the table data. Do not submit tables in photo mode. Figures include line illustrations, photographs (or slides), and computer-generated graphs and must be cited in numerical order in the text. Graphics will aid in the comprehension of the text, but they should be limited to presenting patterns rather than raw data. Figures are costly to print and should be limited to six. Figures must be labeled with author's name and number of figure. Avoid placing labels vertically (except on y-axis). Figure legends should explain all symbols and abbreviations and should be double-spaced on a separate page at the end of the manuscript. Please note that we do not print graphs in color. FAILURE TO FOLLOW THESE GUIDELINES WILL DELAY PUBLICATION OF A MANUSCRIPT Copyright law does not apply to Fishery Bulletin. which falls within the public domain. However, if an author reproduces any part of an article from Fishery Bul- letin in his or her work, reference to source is considered correct form (e.g.. Source: Fish. Bull 97:105). Reprints 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. Submission The Scientific Editorial Office encourages authors to submit their manuscripts as a single PDF (preferred), Word (zipped), or WordPerfect (zipped) document by e- mail to Fishery.Bulletin@noaa.gov. Please use the sub- ject heading "Fishery Bulletin manuscript submission." Do not send encrypted files. For further details on elec- tronic submission, please contact the Scientific Editorial Office directly (see address below). Or you may send your manuscript on compact disc in one of the above formats along with four printed copies (one original plus three copies [stapled]) to the Scientific Editor, at the address shown below. Send photocopies only of figures with initial submission of manuscript; do not send original figures. Original figures and electronic copies of figures will be requested later when the manuscript has been accepted for publication. Dr. Adam Moles Scientific Editor, Fishery Bulletin 11305 Glacier Hwy Juneau, AK 99801-8626 Once the manuscript has been accepted for publication, you will be asked to submit a final software copy of your manuscript. When requested, the text and tables should be submitted in Word or Word Rich Text Format. Figures should be sent as PDF files, Windows metafiles, tiff files, or as EPS files. Send a copy of figures in original software if conversion yields a degraded version. [^^Sftsu/ mMsam9Xftst2/ mps U.S. Department of Commerce Volume 104 Number 2 April 2006 Fishery Bulletin U.S. Department of Commerce Carios M. Gutierrez Secretary National Oceanic and Atmospiieric Administration Vice Admiral Conrad C. Lautenbacher Jr., USN (ret) Under Secretary for Oceans anci Atmosphere National Marine Fisheries Service William T. Hogarth Assistant Administrator for Fislneries Scientific Editor Adam Moles, Ph.D. Technical Editor Elizabeth Calvert National Marine Fisheries Service, NOAA 11305 Glacier Highway Juneau, Alaska 99801-8626 ^OFCo, ^ATBS 0» " Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C15700 Seattle, Washington 98115-0070 The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Puijlications Office. National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115-0070. Periodicals postage is paid at Seattle, WA. POSTMASTER: Send address changes for subscriptions to Fish- ery Bulletin, Superintendent of Docu- ments, 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 Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $55.00 domestic and $68.75 foreign. Cost per single issue: $28.00 domestic and $35.00 foreign. See back for order form. Editorial Committee Harlyn O. Halvorson, Ph.D. Ronald W. Hardy, Ph.D. Richard D. fVlethot, Ph.D. Theodore W. Pietsch, Ph.D. Joseph E. Powers, Ph.D. Harald Rosenthal, Ph.D. Fredric M. Serchuk, Ph.D. George Walters, Ph.D. University of Massachusetts, Boston University of Idaho, Hagerman National Marine Fisheries Service University of Washington, Seattle National Marine Fisheries Service Universitat Kiel, Germany National Marine Fisheries Service National Marine Fisheries Service Fishery Bulletin web site: www.fishbulI.noaa.gov The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. It began as the Bulletin of the United States Fish Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963. each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office. Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions. State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 104 Number 2 April 2006 Fishery Bulletin MBLWHOI Library MAY 2 2006 WOODS HOLE -:3achuoetts 02542 Contents Articles 159—166 Garci'a-Diaz, Mercedes, Jose A. Gonzalez, Marfa J. Lorente, and Victor M. Tuset Spawning season, maturity sizes, and fecundity in blacktail comber (Serranus atricauda) (Serranidae) from tfie eastern-central Atlantic 167-181 Tissot, Brian N., Mary M. Yoklavich, Milton S. Love, Keri York, and Mark Amend Benthic invertebrates that form habitat on deep banks off southern California, with special reference to deep sea coral The conclusions and opinions ex- pressed in Fishery Bulletin are solely those of the authors and do not represent the official position of the National Marine Fisher-ies Ser- vice ( NOAAl or any other agency or institution. The National Marine Fisheries Service (NMFS) does not approve, recommend, or endorse any proprie- tai*y product or proprietai"y 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, recom- mends, or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. 182—196 Li, Zhuozhuo, Andrew K. Gray, Milton S. Love, Akira Goto, Takashi Asahida, and Anthony J. Gharrett A key to selected rockfishes (Sebastes spp ) based on mitochondrial DNA restriction fragment analysis 197—203 Torres-Orozco, Ernesto, Arturo Muhlia-Melo, Armando Trasviiia, and Sofia Ortega-Garcfa Variation in yellowfin tuna iThunnus a/bacares) catches related to El Niho-Southern Oscillation events at the entrance to the Gulf of California 204-214 Pepin, Pierre Estimating the encounter rate of Atlantic capelin iMallotus villosus) with fish eggs, based on stomach content analysis 215—225 Davis, Michelle L., Jim Berkson, and Marcella Kelly A production modeling approach to the assessment of the horseshoe crab (Limulus polyphemus) population in Delaware Bay Fishery Bulletin 104(2) 226-237 Hoff, Gerald R. Biodiversity as an index of regime shift in the eastern Bering Sea 238-246 Pietsch, Theodore W., and James W. Orr Tnglops dorothy, a new species of sculpin (Teleostei: Scorpaeniformes: Cottidae) from the southern Sea of Okhotsk 247—255 Chen, Yong, Sally Sherman, Car! Wilson, John Sowles, and Minora Kanaiwa A comparison of two fishery-independent survey programs used to define the population structure of American lobster (Homarus amencanus) in the Gulf of Maine 256—277 Walsh, Harvey J., Katrin E. Marancik, and Jonathan A. Hare Juvenile fish assemblages collected on unconsolidated sediments of the southeast United States continental shelf 278-292 Goldman, Kenneth J., and John A. Musick Growth and maturity of salmon Sharks (Lamna ditropis) in the eastern and western North Pacific, and comments on back-calculation methods Notes 293-298 Hurton, Lenka, and Jim Berkson Potential causes of mortality for horseshoe crabs (Limulus polyphemus) during the biomedical bleeding process 299-302 Kirshenbaum, Sheril, Scott Feindel, and Yong Chen A study of tagging methods for the sea cucumber Cucumana frondosa in the waters off Maine 303-305 Kimura, Daniel K., and Martin W. Dorn Parameterizing probabilities for estimating age-composition distributions for mixture models 306-310 May-Ku, Marco A., Uriel Ordonez-Lopez, and Omar Defeo Morphometric differentiation in small luveniles of the pink spotted shrimp (Farfantepeneaus brasiliensis) and the southern pink shrimp (F. notialis) in the Yucatan Peninsula, Mexico 311-320 Cope, Jason M. Exploring intraspecific life history patterns in sharks 321 Guidelines for authors 159 Abstract — Blacktail comber iSer- ranus atricauda Gunther) is a com- mercially important species in the Canary Islands fisheries. A total of 425 individuals were collected and histological techniques were used to investigate the reproductive data. Results indicated that the spawning season occurred throughout the year, peaking between March and July. Individuals reached 50'^'v maturity at 19.3 cm TL and 95'~, at 33.1 cm TL. Batch fecundity estimates ranged from 21,774 to 369,578 oocytes per spawning event for specimens from 22.2 to 39.8 cm TL. Blacktail comber was estimated to spawn 42 times/year and 26. 5^^^ of individuals spawned, on average, every 3.8 days. Estimates of potential annual fecundity for the species ranged from 0.91 to 15.5 mil- lion oocytes, and an average of 5.1 ±4.1 million. Spawning season, maturity sizes, and fecundity in blacktail comber iSerranus atricauda) (Serranidae) from the eastern-central Atlantic Mercedes Garcia-Diaz Jose A. Gonzalez Institute Canarlo de Clencias Marinas Departamento de Blologia Pesquera PO Box 56. E-35200 Telde (Las Palmas), Spam Maria J. Lorente Departamento de Biologia Animal (U Morfologia Microscopica) Universidad de Valencia Moliner 50 Buriassot (Valencia), Spain Victor M. Tuset Institute Canarie de Ciencias Marinas Departamento de Bielegia Pesquera PO Bex 56, E-35200 Telde (Las Palmas), Spam E-mail address (fer V M Tuset, contact author); vlcterta@iccm.rcanaria.es Present address (fer V M Tuset) Institute Canane de Ciencias Mannas PO Box 56, E-35200 Telde Las Palmas, Spain Manuscript submitted 20 November 2003 to the Scientific Editor's Office. Manuscript approved for publication 23 June 2005 by the Scientific Editor. Fish. Bull. 104:159-166 (2006). Fish of the genus Serranus are syn- chronous hermaphrodites; male and female tissues are simultaneously functional (Smith, 1965). The most common style of reproduction is serial monogamy, in which individuals are solitary during the day before pair- ing up and spawning in the late afternoon (Fischer, 1986). During spawning one fish in each pair func- tions as a male and the other as a female and cross-fertilization occurs. This special characteristic, and the possibility of self-fertilization, have encouraged detailed studies of the gonad structure of the genus Serranus (Atz, 1965; Fishelson, 1970; Reinboth, 1970; Febvre et al., 1975; Zanuy, 1977; Brusle, 1983; Abd-el-Aziz and Rama- dan, 1990; Garci'a-Diaz et al., 1997, 2002), although knowledge of other reproductive features is scarce. Models of dynamic population used in the management of fishery re- sources and in biological studies of fish require knowledge of the repro- ductive system (Koslow et al., 1995). This includes gonad morphology (ex- ternal and cellular description of the ovary and testis), reproductive pat- tern (hermaphoditism or gonocorism), reproductive behavior, reproductive cycle, spawning season duration, size at maturity, sex ratio, size at sexual transition, and fecundity. Of all these reproductive features, fecundity is the most difficult biological parameter to obtain, although it is of interest to fishery scientists both as a critical pa- rameter for stock assessments based on egg production methods and as a basic aspect of fish biology and popu- lation dynamics. To calculate fecun- dity, fish are divided into determinate and indeterminate spawners (Hunter and Macewicz, 1985a; Hunter et al., 1992). With indeterminate spawners, multiple or serial batches are noted, and spawning may take place many times during a protracted spawning 160 Fishery Bulletin 104(2) •40* JO -20' ^. . season. Therefore fecundity is very dif- ficult to determine before the onset of the spawning season because 1) there is no way to differentiate between the oocytes that are going to be shed dur- ing the next season from those which will remain for future seasons, and 2) the number of reserve oocytes that will mature during the next spawning season cannot be predetermined (Hunter and Goldberg, 1980; Hunter et al., 1985). To estimate indeterminate annual fecun- dity, the mean number of eggs per batch and spawning frequency throughout the spawning season must be calculated. Ow- ing to the complexity of obtaining these data, fecundity studies are normally lim- ited to determinate spawners or pelagic species of substantial economical inter- est (Hunter and Goldberg, 1980; Hunter and Macewicz, 1985a; Karlou-Riga and Economidis, 1997; Murua et al., 1998). The blacktail comber (Serranus atricau- da Gunther, 1874), is a littoral (3-150 m) benthic species, ranging throughout the eastern central Atlantic (from the Bay of Biscay to Mauritania, the Azores, Madeira, and the Canary Islands) and in the west- ern Mediterranean. It is, therefore, a spe- cies of wide distribution and commercial interest in many regions (Bauchot, 1987; Smith, 1990). In the Canary Islands, it is an economically important species for small-scale inshore fisheries (Perez-Barroso et al., 1993). Garcia-Di'az et al. (1999, 2002) determined that this species is a functional simultaneous hermaphro- dite. The ovary is classified as asynchronous, i.e., various stages of oocyte development occur simultaneously (pri- mary growth stage, yolk vesicle formation, vitellogenesis, oocyte maturation, and mature egg). The histological structure of the gonad and of the sperm indicates that this species is an externally fertilizing teleost. The objective of this article is to increase our under- standing of the reproductive biology of the blacktail comber in the Canary Islands by estimating its spawn- ing season duration, size at maturity, and fecundity. Materials and methods 0' 15" Europe Atlantic Ocean Canary Islands 0 500 km Location o [Serranus Atlantic). Figure 1 f sampling areas (shown with dark shading) for blacktail comber atricauda) specimens from the Canary Islands (eastern-central buffered formaldehyde. After 24-48 h, they were dehy- drated, embedded in paraffin wax, sectioned in longi- tudinal or cross-sections (4 to 5 .urn thick), and stained with Harris "hematoxylin-Puttis" eosin. Oocyte stages and spermatogenic cells were classified according to Garcia-Diaz et al. (2002). Postovulatory follicles and atresia were also characterized according to the crite- ria used for Engraulis mordax by Hunter and Goldberg (1980) and Hunter and Macewicz (1985b). Maturity stages (MS) were determined from histological obser- vations (the development of the ovary and testis and also the presence and absence of different types of the oocytes and spermatocytes) according to Garcia-Diaz et al. (1997, 2002) (Table 1). The developmental stages of the oocytes were categorized according to Selman et al. (1993): primary growth (stage I); cortical alveoli formation (stage II); vitellogenesis (stage III); oocyte maturation (stage IV); and mature egg (stage V). Sampling A total of 425 individuals of S. atricauda. ranging from 15.7 to 43.2 cm total length were sampled monthly from commercial catches off the Canary Islands between Sep- tember 1992 and November 1994 (Fig. 1). Total length (TL), to the nearest centimeter, was measured for each specimen, together with the total and gutted weight (TW and GW, respectively), gonad weight (GNW), and liver weight (LW), with an accu- racy of 0.1 g. Gonads were removed and fixed in 10% Seasonality of gonad development Monthly changes in the five following variables were analyzed to determine the spawning season of blacktail comber (Garcia-Diaz et al., 1997): 1 Percent frequency of the maturity stages; this indi- cates population changes in gonad development. 2 Gonadosomatic index (GSI=(gonad weight /gutted weight)xlQO): this shows differences in development of the gonads with respect to gutted body weight; Garcia-Diaz et al ; Spawning season, maturity sizes, and fecundity of Seiranus atncauda 161 Table 1 Description of gonad stages according to Garci'a-Diaz et al. (1997, 2002). Maturity stage Histological appearance I. Immature II. Developing virgin, or recovering-spent III. Developing, maturing IV. Ripe V. Spent Ovary contains oogoniaand oocytes from primary growth (stage I). Testis formed mainly by sperma- togonia and primary spermatocytes. Ovary and testis joined by connective tissue. Ovary begins to acquire ovarian lamellae. Contains stage-I oocytes and yolk-vesicle-formation- stage oocytes (stage II) in advanced phases. Testis arranged in tubules with spermatogonia and primary and secondary spermatocytes. Ovary with oocytes at all previous stages and oocytes in vitellogenic stage (stage III). Seminiferous tubules contain all spermatogenic cells. Ovary with oocytes at all previous stages and with maturation oocytes (stage IV), mature and hydrated eggs (stage V). Atresic oocytes and postovulatory follicles appear. Testis completely mature, tubules filled with spermatozoa (which accumulate in deferent duct next to gonadal wall) to be expelled. Oocytes undergoing regression and reabsorption. Numerous atresic oocytes. Testis in regression; cells appear fused, form semicontinuous mass. 3 Hepatosomatic index (HSl={Uver weight I gutted weight )y.lQO): this estimates the relative size of the liver to body weight: 4 Condition factor (Kn={total weight/ TL '•)x 1000): this is as an overall measurement of robustness of the fish, b being the exponential of the regression TW = aTL^, which is 3.25 according to Tuset et al. (2004); 5 Oocyte diameter (DO): this gives information on the cell development of each individual. To obtain this value, often fish randomly chosen from the monthly samples, the diameters of the first 50 oocytes encoun- tered were measured with an ocular micrometer. Measurements were taken only of oocytes sectioned through the nucleus. Length at sexual maturity Total length of all individuals was used to estimate the size at first maturity and size at mass maturity. These are defined as the sizes (TL) at which 50% (TLj^y,^) and 95% (TLggtj) of all fish sampled are at the relevant maturity stage (developing MS III, ripe MS IV, or spent MS V) (Garcia-Diaz et al., 1997). The proportions were estimated at length classes of 1 cm, and the data were fitted to the logistic curve (Pope et al., 1983): p = WO/(l + exp{a+bTL)), where p = percentage of mature individuals as a func- tion of size class (TL); and a and b = specific parameters that can change during the life cycle. A logarithmic transformation was applied to the equa- tion to calculate the parameters a and b by means of linear regression. Reproductive potential The pattern of annual fecundity (indeterminate or deter- minate) was assessed by oocyte size-frequency distribu- tion (Hunter and Macewicz, 1985a). Ten fish ranging between 18.1 and 31.5 cm TL in ripe stage were selected for analysis of oocyte size-frequency distribution — five fish in March and another five in July — to represent early and late gonad development (White et al., 2003). Because the mean size (^-test, P>0.05) was not signifi- cantly different between both samples, one frequency distribution was obtained for each month. The diameters of the first 500 oocytes encountered were measured with an ocular micrometer for each fish. Measurements were taken only of oocytes sectioned through the nucleus. Gonads of individuals in MS III (developing) and MS IV (ripe) were collected to estimate fecundity. One lobule was fixed in 10% buffered formaldehyde (24 h) for histological study, and the other was weighed and preserved in Gilson's fluid to estimate fecundity. Gonads with postovulatory follicles were omitted from batch fecundity estimates (Hunter et al., 1985). The oocyte size-frequency method was applied to de- termine the batch fecundity (Hunter et al., 1985). This entails counting the number of oocytes in the most de- veloped modal group of oocytes and plotting the size-fre- quency. In each lobule 200 oocytes were measured with the Nikon digital counter SC-112 and data processor DP- 201 connected to the micrometer stage of a profile projec- tor (V-12A, Nikon, Melville, NY). The routine NORMSEP (normal distribution separator) of FAO-ICLARM Stock Assessment Tools (FISAT II, vers. 1.0.0, FAO, Rome, Italy) program was used to separate the most developed modal group of oocytes based on Hasselblad's maximum likelihood method (Hasselblad, 1966). Batch fecundity was estimated gravimetrically by using the oocyte size-frequency method (Hunter et 162 Fishery Bullelm 104(2) al., 1985; Yoneda et al., 2001). Samples were filtered through a 100-mm mesh sieve (approximately the mini- mum diameter of vitellogenic oocytes, Garcia-Diaz et al., 2002) and were carefully washed under running distilled water to eliminate tissue remains and fixer. We used a sieve made from a piece of nylon plankton net inserted between two sections of PVC pipe, 15 cm in diameter and 10 cm in depth (Lowerre-Barbieri and Barbieri, 1993). A compact mass of oocytes was dried for 24 hours on filter paper and weighed precisely to 0.001 g. Afterwards, a subsample of 0.2 g was selected and placed in the count dish, covered with 3-4 drops of glycerin, and the number of oocytes in the sample was tallied with a hand counter. Batch fecundity was considered to be the total number of oocytes within the most developed modal group of oocytes (Hunter et al., 1985). Spawning frequency (the number of spawnings per year by an individual) was estimated by dividing the duration of the spawning season by the average number of days between spawning for all individuals (Hunter and Macewicz, 1985a). The duration of the spawning season was the number of days between the first (7"^ February) and last (16''^ July) occurrence of hydrated oocyte or postovulatory follicles. The average number of days between spawning was the inverse of the percent frequency of hydrated individuals, multiplied by 100 (Collins et al., 1998). The potential annual fecundity estimates (PAFEs) were obtained by multiplying batch fecundity by spawn- ing frequency, and relative fecundity was defined as PAFE divided by individual weight (Hunter et al., 1992). The relationships between PAFE and TL, PAFE and TW, and PAFE and gonad weight (GW) were calcu- lated by the following compound equation; where X = TL, TW, or GW; and a and h = specific parameters. A logarithmic transformation was applied to the equa- tion to calculate parameters a and b by using linear regression (Zar, 1996). Results Spawning season and maturity sizes The specimens revealed the presence of maturing and ripe individuals between December and October, exclud- ing January because only one individual was sampled then. This finding may indicate that the population spawns throughout this period, peaking between Febru- ary (37.5'7f ) and June (89.8%) (Fig. 2). The highest values for GSI (Fig. 3A) occurred between March (2.06) and July (3.37) and a peak of maximum activity occurred in June (3.93), decreasing over the following months. The HSI (Fig. 3B) and Kn (Fig. 3C) presented irregular values during the annual cycle. Consequently, no correlation was observed between either ovarian and liver growth with ovarian and fish growth. Oocyte diameter (DO) (Fig. 3D) values varied similarly to those from the GSI; highest values occurred from March (261.37 .iim) to July (275.07 ^, o § 50 g' 40 it 30 ■ 7 ■ i -/jL'^o%= 19,3 cmi /"i i 20 / 1 10 ^^ \ \ 1 5 9 13 17 21 25 29 33 37 41 Total length (cm) Figure 4 Sexual maturity curve and size-s at maturity for blacktail comber iSerraus atricauda) from the Canary Islands, 164 Fishery Bulletin 104(2) gonad (qualitative methods), somatic indices (e.g., GSI, HIS, and Kn), or evolution of oocyte diameter (quan- titative methods) (or all four methods) are commonly applied. Many authors have disagreed about their ap- plications and biological interpretation (De Vlaming et al., 1982; Wootton, 1990; Shapiro et al., 1993; Sadovy, 1996; Karlou-Riga and Economidis, 1997). Our results demonstrate that the GSI indicates only the spawning peak and not the presence of batch or multiple spawn- ing within the species; and the HSI and Kn do not exhibit a clear trend throughout the year. Normally, variations of these indices (HSI and Kn) imply energy storage for reproduction (Hoar et al., 1983; N'Da and Deniel, 1993). However, S. atricauda does not require such storage because feeding activity is not altered during the reproductive period (Morato et al., 2000). Oocyte diameter presents a trend similar to that of GSI, although its biological interpretation is different. The oocyte may begin the vitellogenesis phase and there □ March ■July Oocyte diameter (microns) Figure 5 Oocyte size-frequency distributions l25-iim intervals) for blacktail comber tSerraus atricauda) ovaries from the Canary Islands. may be no significant increase in gonad weight. Conse- quently. GSI values do not necessarily change signifi- cantly when the spawning period has begun. Thus, the quantitative indices (GSI, HSI, Kn) show the general trend of the population, whereas the oocyte diameter provides information about the population as a whole, as well as at the individual level. The analysis of qualitative and quantitative data in S. atricauda seems to indicate that spawning takes place throughout the year, although the general popu- lation spawns between March and July. In the Canary Islands, other members of the genus spawn over long periods: eight months in S. cabrilla and nine months in S. scriba (Garcia-Diaz et al., 1997; Garcia-Diaz 2003). These three species have the longest spawning season of all fish studied in the Canaries, although they are also the only species whose reproductive patterns have been analyzed with histological procedures. Nevertheless, environmental or biological factors (or both) must be related to this extended spawning period. To estimate spawning frequency, the hy- drated oocyte method is less time consum- ing but the POP method is better because the spatial and temporal distribution of the postovulatory follicle structures is not con- tinuous (Hunter and Goldberg, 1980). Re- productive potential measured as potential annual fecundity has not been addressed in any species of the genus Serranus with the POP method to date. It is true that this fecundity must be considered as a rough estimation because the spawning frequency was a preliminary estimation because of the scarce number of individuals in sam- pling months. Siau and Bouian (1994) cal- culated the total fecundity in S. cabrilla and S. scriba, considering all the vitel- logenesis oocytes. Their method has been widely rejected because it is valid only for 16 14 - 12 10 8- 6 4 2 0 20 22 24 26 28 30 32 34 36 38 40 42 Total length (cm) 100 200 300 400 500 600 700 800 900 1000 Total weight (g) 100 200 300 400 500 600 700 800 900 Gutted weight (g) Figure 6 Relationships between potential annual fecundity and total length (A), total weight (B), and gutted weight (C) for black- tail comber (^Serranus atricauda) from the Canary Islands. Garcia-Diaz et al Spawning season, maturity sizes, and fecundity of Serronus atncauda 165 determinate spawners, and both the above-mentioned serranids are indeterminate spawners (Garci'a-Diaz et al., 1997; Garcia-Diaz, 2003). The fact that the method was used for these two indeterminate spawners could explain why the values of maximum fecundity obtained by these authors for S. cabrllla (441,502 oocytes) and S. scriba (54,649 oocytes) were lower than that for S. atricauda (15,5 millions oocytes). Consequently, the present study enables us to provide new reproductive data for blacktail comber, where potential annual fe- cundity is essential for future egg production models and estimates of spawning stock biomass. Acknowledgments We wish to thank Jose Ignacio Santana for the collection of samples, Antonio Valencia for his help in the building of the mechanism to separate the vitellogenic oocytes, Gordon Hamilton for proofreading the submission copy of the manuscript, and the reviewers for their comments and suggestions. Literature cited Abd-el-Aziz, S. H., and A. A. Ramadan. 1990. Sexuality and hermaphroditism in fishes. I. Syn- chronous functional hermaphroditism in the serranid fish Serranus scriba L. Folia Morph. 38:86-103. Atz, J. W. 1965. Hermaphroditic fish. Science 150:789-797. Bauchot, M. L. 1987. Serranidae. /;; Fiches FAO d'identification des especes pour les besoins de la peche (revision 1), Medi- terranee et Mer Noire, zones de peche 37, II (vertebres) (W. Fischer, M. L. Bauchot, and M. Schneider, eds.), p. 1301-1319. FAO-CEE, Rome. Brusle, S. 1983. Contribution to the sexuality of a hermaphro- ditic teleost, Serranus hepatus L. J. Fish Biol. 22: 283-292. Collins, L. A., A. G. Johnson, C. C. Koenig, and M. S. Baker Jr. 1998. Reproductive patterns, sex ratio, and fecundity in gag, Mycteroperca microlepis (Serranidae), a protogynous grouper from the northeastern Gulf of Mexico. Fish. Bull. 96:415-427. De Vlaming, V., G. Grossman, and F. Chapman. 1982. On the use of the gonadosomatic index. Comp. Biochem. Physiol. 73A (11:31-39. Febvre, M., M. Michele, and M. Lafaurie. 1975. Comparaison de la sequence ovogenetique chez des Teleosteens (sic) ovipares gonochoriques et hermaph- rodites (Mullus, Serranus, Boops). Pubbl. Staz. Zool. Napoli 39:140-152. Fischer, E. A. 1986. Mating systems of simultaneously hermaphroditic serranid fishes. In Proceedings of the second Indo- Pacific fish conference (T. Uyeno, R. Arai, T. Taniuchi, and K. Matsuura, eds.), p. 776-784. Ichthol. Soc. Japan, Tokyo, Japan. Fishelson, L. 1970. Protogynous sex reversal in the fish Anthias squa- mipmnis regulated by the presence or absence of a male fish. Nature 227:90-91. Garcia-Diaz, M. M. 2003. Applicacion de tecnicas histologicas para la deter- minacion de aspectos reproductores en tres teleosteos marinos (Serranidae, Serranus), 228 p. Servicio de Publicaciones de la Universidad de La Laguna, Ten- erife. Garcia-Diaz, M. M., V. M. Tuset, J. A. Gonzalez, and J. Socorro. 1997. Sex and reproductive aspects in Serranus cabrilla (Osteichthyes: Serranidae): macroscopic and histological approaches. Mar. Biol. 127:379-386. Garcia-Diaz, M. M., M. J. Lorente, J. A. Gonzalez, and V. M. Tuset. 1999. Comparative ultrastructure of spermatozoa of three marine teleosts of the genus Serranus: S. atricauda, S. cabrilla and S. scriba. J. Submicrosc. Cytol. Pathol. 31:503-508. 2001. Application of histology in reproductive biology of fishery studies. In Perspective in comparative endocrinology: unity and diversity (H. J. T. H. Goos, R. K. Rastogi, H. Vaudry. and R. Pierantoni, eds.), p. 1095-1100. Monduzzi Editore, Bologna, Italy. 2002. Morphology of the ovotestis of Serranus atricauda (Teleostei, Serranidae). Aquat. Sci. 64:87-96. Hasselblad, V. 1966. Estimation of parameters for a mixture of normal distribution. Technometrics 8: 431-444. Hoar, W. S., D. J. Randall, and E. M. Donaldson (eds.). 1983. Fish physiology, vol. IX (part B), 477 p. Academic Press, New York, NY. Hunter, J. R., and R. Goldberg. 1980. Spawning incidence and batch fecundity in Northern anchovy Engraulis mordax. Fish. Bull. 77:641-652. Hunter, J. R., and B. J. Macewicz. 1985a. Measurement of spawning frequency in multiple spawning fishes. In An egg production method for esti- mating spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mordax (R. Lasker, ed.), p. 79-94. NOAA Tech. Rep. NMFS 36. 1985b. Rates of atresia in the ovary of captive and wild anchovy, Engraulis mordax. Fish. Bull. 83: 119-136. Hunter, J. R., N. C. R. Lo, and R. J. H. Arnold. 1985. Batch fecundity in multiple spawning. /;; An egg production method for estimating spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mordax (R. Lasker, ed.), p. 66-77. NOAA Tech. Rep. NMFS 36. Hunter, J. R., B. J. Macewicz, N. C. H. Lo. and C. A. Kimbrell. 1992. Fecundity, spawning, and maturity of female Dover sole Microstomas pacificus, with an evaluation of assump- tions and precision. Fish. Bull. 90:101-128. Lowerre-Barbieri, S. K., and L. R. Barbieri. 1993. A new method of oocyte separation and preservation for fish reproduction studies. Fish. Bull. 91:165-170. Karlou-Riga, C, and P. S. Economidis. 1997. Spawning frequency and batch fecundity of horse mackerel, Trackurus trachurus (L.), in the Saranikos Gulf (Greece). J. Appl. Ichthyol. 13:97-104. Koslow, J. A., J. Bell, P. Virtue, and D. C. Smith. 1995. Fecundity and its variability in orange roughy: effects of population density, condition, egg size, and senescence. J. Fish Biol. 47:1063-1080. Morato, T., R. S. Santos, and J. P. Andrade. 2000. Feeding habits, seasonal and ontogenic diet shift of blacktail comber, Serranus atricauda (Pisces: Serra- 166 Fishery Bulletin 104(2) nidae). from the Azores, north-eastern Atlantic. Fish. Res. 49:51-59. Murua, H., L. Motos, and P. Lucio. 1998. Reproductive modality and batch fecundity of the European hake [Merluccius meiiucciiin L.) in the Bay of Biscay. CalCOFI Rep. 39:196-203. N'Da, K., and C. Deniel. 1993. Sexual cycle and seasonal changes in the ovary of the red mullet, MiiUus surmuletus, from the southern coast of Brittany. J. Fish Biol. 43:229-244. Perez-Barroso, F. I., S. Jimenez, I. J. Lozano, J. A. Gonzalez, and J. I. Santana. 1993. Estudio preliminar de la pesqueria artesanal de Serranus atru rnida Gunther, 1974 y de Serranus cabrilla (Linnaeus. 1758) (Osteichthyes, Serranidae) en el archipielago canario. Publ. Esp. Inst. Esp. Oceanogr. 11:365-373. Pope, J. A., A. R. Marggets, J. M. Hamley. and E. F. .\kyuz. 1983. Manual of methods for fish stock assessment. Pt. 3. Selectivity of fishing gear. FAO Fish.Tech. Pap. 41 (l):l-65. Reinboth, R. 1970. Interse.xuality in fishes. Men. Soc. Endocrinol. 18:515-543. Selman, K., R. A. Wallace, A. Sarka, and X. Qi. 1993. Stages of oocyte development in the zebrafish, Brachydanio rerio. J. Morphol. 218:203-224. Sadovy. Y. J. 1996. Reproduction of reef fishery species. In Reef fish- eries, vol. 2 (N. V. C Polunin, and C. M. Roberts, eds.), p. 15-59. Chapman and Hall, London. Shapiro, D. Y., Y. Sadovy, and A. McGehee. 1993. Size, composition, and spatial structure of the annual spawning aggregation of the red hind, Epi- nephelus guttatus (Pisces: Serranidae). Copeia 1993: 367-374. Siau, Y., and A. Bouain. 1994. Variations in spawning of two species of coastal hermaphrodite fishes, genus Serranus, related to their bathymetric distribution. Oebalia 10:1-20. Smith, C. L. 1965. The patterns of sexuality and classification of serranid fishes. Am. Mus. Novit. 2207:1-20. 1990. Serranidae. In Check-list of the fishes of the eastern tropical Atlantic, vol. 2 (J. C. Quero, J. C. Hureau, C. Karrer, A. Post, and L. Saldanha, eds. I, p. 695-706. UNESCO-SEI-JNICT, Lisbon, Portugal. Tuset, V. M., J. A. Gonzalez, I. J. Lozano, and M. M. Garcia-Diaz. 2004. Age and growth of the blacktail comber, Serranus atricauda (Serranidae), off the Canary Islands (central- eastern Atlantic). Bull. Mar. Sci. 74:53-68. White, G. G., T.A. Munroe, and H. M. Austin. 2003. Reproductive seasonality, fecundity, and spawn- ing frequency of tautog iTautoga onitis) in the lower Chesapeake Bay and coastal waters of Virginia. Fish. Bull. 101:424-442 Wootton, R. J. 1990. Ecology of teleost fishes, 404 p. Chapman and May, London. Yoneda, M., M. Tokymura, H. Fujita, N. Takeshita, K. Takeshita. M. Matusyama and, S. Matsuura. 2001. Reproductive cycle, fecundity, and seasonal dis- tributions of the anglerfish Lophius Utulon in the East China and Yellow seas. Fish. Bull. 99:56-370. Zanuy, S. 1977. Induccion a la puesta y estudio de la ovogenesis en un teleosteo marine: Paracentropristis cabrilla L. Inv. Pesq. 41:337-384. Zar, J. H. 1996. Biostatistical analysis, 3^'^ ed., 662 p. Prentice- Hall International, LTpper Saddle River, New Jersey. 167 Abstract — There is increasing inter- est in the potential impacts that fish- ing activities have on megafaunal benthic invertebrates occurring in continental shelf and slope ecosys- tems. We examined how the structure, size, and high-density aggregations of invertebrates provided structural relief for fishes in continental shelf and slope ecosystems off southern California. We made 112 dives in a submersible at 32-320 m water depth, surveying a variety of habi- tats from high-relief rock to flat sand and mud. Using quantitative video transect methods, we made 12,360 observations of 15 structure-form- ing invertebrate taxa and 521,898 individuals. We estimated size and incidence of epizoic animals on 9105 sponges, black corals, and gorgonians. Size variation among structure-form- ing invertebrates was significant and 909c of the individuals were <0.5 m high. Less than 1% of the observa- tions of organisms actually shelter- ing in or located on invertebrates involved fishes. From the analysis of spatial associations between fishes and large invertebrates, six of 108 fish species were found more often adjacent to invertebrate colonies than the number of fish predicted by the fish-density data from transects. This finding indicates that there may be spatial associations that do not neces- sarily include physical contact with the sponges and corals. However, the median distances between these six fish species and the invertebrates were not particularly small (1.0-5.5 m). Thus, it is likely that these fishes and invertebrates are present together in the same habitats but that there is not necessarily a functional relationship between these groups of organisms. Regardless of their associations with fishes, these invertebrates provide structure and diversity for continental shelf ecosystems off southern Califor- nia and certainly deserve the atten- tion of scientists undertaking future conservation efforts. Benthic invertebrates that form habitat on deep banks off southern California, with special reference to deep sea coral Brian N. Tissot Program in Environmental Science and Regional Planning 14204 NE Salmon Creek Ave. Washington State University Vancouver, Washinton 98686-9600 E-mail address tissot'a Vancouver wsu edu Mary M. Yoklavich Fisheries Ecology Division Southwest Fisheries Science Center National Marine Fisheries Service, NOAA Santa Cruz Laboratory 110 Shaffer Road Santa Cruz, California 95060 Milton S. Love Marine Science Institute University of California Santa Barbara, California 93106 Keri York Program in Environmental Science and Regional Planning 14204 NE Salmon Creek Ave. Washington State University Vancouver, Washington 98686-9600 Mark Amend Southwest Fisheries Science Center National Marine Fisheries Service, NOAA Santa Cruz Laboratory 110 Shaffer Road Santa Cruz, California 95060 Manuscript submitted 8 December 2004 to the Scientific Editor's Office. Manuscript approved for publication 21 July 2005 by the Scientific Editor. Fish. Bull. 104:167-181 (2006). Science and conservation communi- ties are increasingly interested in the potential impacts that fishing activi- ties have on megafaunal benthic inver- tebrates found in continental shelf and slope ecosystems (Dayton et al., 2002; NRC, 2002; Malakoff, 2004; Roberts and Hirshfield, 2004; Rogers, 2004). Megafaunal invertebrates (>5 cm in height) contribute significantly to bio- diversity, play important functional ecological roles, and can be indica- tors of long-term environmental condi- tions (e.g., Riedl, 1971; Palumbi, 1986; Brusca and Brusca, 1990). Moreover, because large invertebrates, such as sponges and corals, enhance the diversity and structural component of fish habitat and are vulnerable to impacts by at least some fisheries, they also may signify habitat areas of particular concern (HAPC) and as such would be protected under the Magnuson-Stevens Fishery Conser- vation and Management Act (Freese, 2001; Etnoyer and Morgan^). ' Etnoyer, P., and L. Morgan. 2003. Oc- currences of habitat-forming deep sea corals in the Northeast Pacific Ocean. Technical Report, NOAA Office of Habitat Conservation, 31 p. Marine Biologv Conservation Institute, 15806 NE 47'h Ct., Redmond, WA 98052. 168 Fishery Bulletin 104(2) Deep-sea corals, such as gorgonians (sea fans), antipa- ththarians (black corals), scleractinians (stony corals), and hydrocorals, are of particular interesc because they are often long-lived and slow-growing (Andrews et al., 2002; Heifetz, 2002), poorly studied (Etnoyer and Mor- gan'), and in certain situations vulnerable to human activities (e.g., mobile fishing gear) (Watling and Norse 1998; Freese et al., 1999; Krieger, 2001; Dayton et al., 2002: Fossa et al., 2002; NRC, 2002;). Other megafau- nal invertebrates, such as crinoids, basket stars, and sponges also may enhance the structural components offish habitat (Puniwai, 2002) and may be disturbed or destroyed by some fishing activities 'Freese, 2001; Krieger, 2001). The potential for invertebrates to add functional structure to benthic communities has centeicd large- ly around their size and complex morphology. A size threshold of 1 m has often been used as an indicator of structure-forming species because marked changes in benthic community structure have been observed in areas where rocky substrata exceed 1 m (Lissner and Benech, 1993). The complex structure of deep-sea corals also has been discussed as an important factor that contributes to microhabitat diversity (Krieger and Wing, 2002; Etnoyer and Morgan'). In this article, be- sides forming complex structure and large size, we also believe that megafaunal invertebrates form structure if they aggregate in high numbers, especially in areas of low relief For example, aggregations of sea urchins and sea pens may provide significant structural relief for fishes in mud- and sand-dominated habitats (Bro- deur, 2001). An important question is the extent to which struc- ture-forming invertebrates are ecologically important to fishes, especially those of economical value. Most studies have focused on "associations" between structure-form- ing invertebrates and fishes as a measure of ecological importance at several spatial scales. Fishes have been considered to be associated with invertebrates if they are found in the same trawl sample (Heifetz, 2002), if fishing is higher in areas with corals than without corals (Husebo et al., 2002), if they are found together within similar habitats observed from a submersible (Hixon et al.-), or if they are observed "among or within 1 m" from corals (Krieger and Wing, 2002). In this article we investigated association at three different levels: 1) fishes that are physically touching large in- vertebrates; 2) fishes that are found statistically more frequently near large invertebrates in relation to their overall abundance patterns; and 3) fishes that are found as nearest neighbors to large invertebrates. The goal of this study was to describe patterns in the density, distribution, and size of structure-forming megafaunal invertebrates on deep rocky banks and outcrops off southern California. Given the recent inter- est in these organisms as potentially important habitat for groundfishes, and thus targets for protection from fishing activities, these organisms deserve a critical examination of their potential to contribute structure to continental shelf and slope ecosystems and an ex- amination of their associations with fishes and other marine organisms. Accordingly, our specific objectives were the following: • Identify structure-forming invertebrates based on cri- teria of size, morphological complexity, and density; • Quantify the density and size distributions of these invertebrates according to depth and substratum types; • Quantify associations between large, structure-form- ing invertebrates and other organisms, particularly fishes; and • Assess the health of these organisms in terms of obvious physical damage. Materials and methods Underwater surveys were conducted off southern Califor- nia by using nonextractive video-transect methods and direct observations from an occupied research submers- ible (Delta) from 8 October to 6 November 2002. These surveys were conducted as part of a larger investigation into the abundance, size, and distribution of cowcod (Sebastes levis) and associated benthic fishes and habi- tats inside and around the newly established Cowcod Conservation Areas (CCAs) off southern California (Fig. 1). The CCAs, which encompass 14,750 km- and are closed to groundfish harvest in water depth >37 m, were established in 2001 to assist in rebuilding the depleted cowcod population off southern California. Digital, georeferenced maps of seafloor substratum types, interpreted from side-scan sonar, multibeam ba- thymetry, seismic reflection, and other past geophysical surveys, were used to identify and select sites of rocky habitats (Greene et al.'). We attempted to restrict the substratum types to mixed sediment and rock and to 30-330 m depth (i.e., likely cowcod habitat). The Delta submersible was tracked by using an ORE Trackpoint II plus (ORE Offshore, West Wareham, MA) USBL system and WINPROG (vers. 3.1, FUGRO. San Diego, CA) software. We linked the tracking system to our ArcView"^ GIS (vers. 3.2, ESRI Corp., Redlands, CA) seafloor mapping project and tracked the submersible real-time in relationship to depth and seafloor habitat maps. 2 Hixon, M. A., B. N. Tissot, and W. G. Pearcy. 1991. Fish assemblages of rocky banks of the Pacific northwest, Heceta, Coquille, and Daisy Banks. OCS Study MMS 91-0052. 410 p. U.S. D.I. Minerals Management Service 770 Paseo Camarillo. 2"'' Floor, Camarillo, CA 93010. 3 Greene, H. G., J. J. Bizzarro, D. M. Erdey. H. Lopez, L. Murai, S Watt, and J. Tilden. 2003. Essential fish habi- tat characterization and mapping of California continental margin. Moss Landing Marine Laboratories Technical Pub- lication Series No. 2003-01, 29 p., 2 CDs. Moss Landing Marine Laboratories, 8272 Moss Landing Rd., Moss Landing, CA 9.5039. Tissot et al Invertebrates on deep banks off souttnern California 169 120"30'W 120=W 119 'SOW 119°W 118'30'W 118°W 117°30'W 117-W 34°30'N 34 N 33°30'N 33°N 32°30'N Point Conception Saci Miguel Is. Santa Cruz Is Anacapa Is ^ "Harrison'iS^^i^ '- " Reef A \, Hidden Reef /?T n V ^ ' 'ft J—i TJ»~^v wcv"' (lx "^ S Potato ©imk. • S^N1«pIas.ls7 Kidney BanH<- ~ ^ , — Santa-Bjirbir jTs^ .^i ^■^ Ost)Orn^ank 25 ;-fV ..,., <> V.\ :■■ % , ^^pheri-y Bank 50 I L. 100 km 120°30W 120°W 119"30W 119''W 118'30'W 118°W 117°30'W 34°30'N 34°N 33 30N 33°N 32'>30N 117°W Figure 1 Locations of survey sites and dives (black dots) conducted inside and outside the cowcod conservation areas (delineated by outlined boxes) off southern California in 2002 to ascertain patterns in the density, distribution. and size of habitat-forming invertebrates. We documented dives continuously during daytime hours with an externally mounted high-8 video camera positioned above the middle viewing-porthole on the starboard side of the submersible. The observer ver- bally annotated all tapes with observations on fishes, invertebrates, and physical habitats. Two parallel lasers were installed 20 cm apart on either side of the external video camera for estimating fish and invertebrate sizes and delineating a 2 m-wide belt transect for counting fishes and invertebrates. We used personal dive sonar from inside the submersible to verify the width of swath for the belt transects. Digital still and video cameras were used inside the submersible to help document fishes, invertebrates, and habitats. We defined "habitat" using a combination of nine dif- ferent categories of substratum and standard geologi- cal definitions (see Stein et al., 1992; Yoklavich et al., 2000). In order of increasing particle size or relief, these substrata were the following: mud (code M), sand (S). gravel (G), pebble (P), cobble (C), boulder (B), continu- ous flat rock (F), rock ridge (R), and pinnacles (T). A two-character code was assigned each time a distinct change in substratum type was noted along the tran- sect, thus delineating habitat patches of uniform type. The first character in the code represented the substra- tum that accounted for at least 50% of the patch, and the second character represented the substratum ac- counting for at least 20% of the patch (e.g., "BC" repre- sented a patch with at least 50% cover by boulders and at least 20% cover by cobble). Each habitat patch also was assigned a code based on the degree of its three- dimensional structure as defined by the vertical relief of the physical substrata from the seafloor. Habitats were coded as l=low (<1 m), 2=moderate (1-5 m), or 3=high relief (>5 m). Patches less than 10 seconds in duration were not recorded. The area of each habitat patch was determined by calculating the distance between the beginning and end of habitat patches with ArcGISC© and multiplying by the width of the transect (2 m). A total of 58 different types of habitat patches were observed across all dives. These data were analyzed by a cluster analysis (Euclidean distance, group average method) by using the abundances of the 20 most com- mon invertebrate species, and the resulting dendrogram was used to pool the number of codes into the 17 most distinct habitat types exhibiting a similarity of >50%. Direct counts of megafaunal invertebrates were made from videotapes within each habitat patch; patch areas varied from 12-1472 m-. Densities of invertebrates were estimated by dividing the total number of individuals 170 Fishery Bulletin 104(2) Table 1 Number of submersible dives and habitat patches sur- veyed for structure-forming invertebrates on rocky out- crops inside and outside the cowcod conservation areas in | the southern CaUfornia borderland. No. of Study site No . of dives habitat patches Inside the cowcod conservation areas 43-fathom 4 142 Cherry Bank 8 266 Hidden Reef 5 156 Kidney Bank 21 474 Osborn Bank 5 199 Potato Bank 4 113 San Nicolas Island 23 785 Santa Barbara Island 9 203 Tanner and Cortes banks 21 587 Outside cowcod conservation areas Harrison's Reef 9 198 Dume Canyon 3 66 Total 112 3189 in each group, identified to the lowest taxonomic level, by the area of their associated habitat patches. For the larger invertebrates we recorded the geographic position and estimated maximum height of each solitary sponge, gorgonian. and black coral. We also noted the color of black corals. We made observations on the occurrence of animals (i.e., epizoids) found directly on sponges, gor- gonians, and black corals, and also noted any damaged or dead individuals. Voucher specimens were collected to assist in taxonomic identification. To quantify fish-invertebrate associations we used ArcGIS - to estimate the distance between each sponge, gorgonian, and black coral invertebrate and the nearest fish. These data were compared to the total number of fishes counted in habitats that contained sponges, gorgonians, and black corals by using a chi-square test to look for significant differences in the frequency of fish observed near corals in relation to overall fish abundance. Results We completed 112 dives and surveyed 3189 habitat patches (Table 1), covering 26.1 hectares at 32-320 m depths (median depth of 110 m). The distributions of number of patches and of surface area of habitats were similar except for sand (SS), sand-gravel (SG). and mud (MM) habitats, all of which had greater surface areas in relation to number of patches (Fig. 2, A and B). Overall, cobble-sand (CS), sand, mud, and sand-gravel TT RR RM BB BC BS CB MB SB CS SC SP MG SG MM MS SS CD 2 - 26 1 hectares -l?l m ra kl FU I a ■4 TT RR RM BB BC BS CB MB SB CS SC SP MG SG MM MS SS ^, 2 n n TT RR RM BB BC BS CB MB SB CS SC SP MG SG MM MS SS Habitat type Figure 2 Characteristics of habitat patches surveyed on southern California rocky banks. (A) Number of patches in each substratum type. (B) Total area of each substratum type. (C) Mean relief (±1 standard error). See text for description of method and habitat codes. See page 169 for definitions for the substrate abbreviations along the .V axis. constituted the largest habitat areas, but cobble-sand and sand were the most frequent habitat types. Verti- cal physical structure varied among habitat types; the highest structure was found in high-relief rock areas (TT to BS) and lower structure was found in low-relief mixed rock (CB to SP) and mixed sediment areas (MG to SS; Fig. 2C). The frequency of patches of each sub- stratum type varied by depth (Fig. 3). The incidence of most high-relief rock categories (TT, RR, BB, and BS) decreased with depth, and the occurrence of mud-domi- nated habitat patches (MB, MG, MS, MM) increased with depth. Overall, 12,360 observations were made on 521,898 individuals from 15 taxa of megafaunal, structure-form- ing invertebrates; during these observations, estimates of size and incidence of epizoic animals on 9105 sponges, black corals, and gorgonians were made (Table 2). The most common structure-forming invertebrates (98% of total) included the crinoid Florometra serratissima (40%), the brittle star iOphiacantha spp. [33%]), bra- chiopods (order Terebratulida [11%]), the white sea ur- chin {Lytechinus anamesus [9%]), the fragile sea urchin Tissot et al : Invertebrates on deep banks off southern California 171 (Allocentrotus fragilis [4%]), and sea pens (suborder Subselliflorae; [2%]; Fig. 4). The density of common structure-forming invertebrates was variable across habitat types; some species were found over a wide range of habitats. Crinoids and basket stars were found on all 17 habitat types but were most dense on either high-relief rock or low- relief mixed rock (Fig. 5). In contrast, brittle stars and brachiopods were dense in low-re- lief mixed rock but rare or absent in low-re- lief mixed sediment. White sea urchins were most dense in habitats with sand, whereas fragile sea urchins were most dense in habi- tats with mud. White-plumed anemones were most dense in mud-gravel habitats, and sea pens were most dense in low-relief mixed- sediments (Fig. 5). Deep sea corals and sponges were the larg- est structure-forming invertebrates but were relatively uncommon (27^ of total) (Table 2). Gorgonians were difficult to distinguish and were categorized into one group (order Gor- gonacea). The black coral is a new species that recently has been described and named the Christmas tree coral (Antipathes den- drochristos) (Opresko, 2005). Sponges were categorized into five groups based on their structure and shape: flat, barrel, shelf, vase, and foliose sponges (Fig. 6). Gorgonians and black corals were most dense on low-relief mixed rock areas (Fig. 7). However, gorgonians were found in only four habitat types at 144-163 m depth, whereas black corals were found on 12 habitat types at 100-225 m depth, including pinnacle, boulder, and sand areas. These differences may be due to the un- equal number of observations (i.e., 27 gorgonian vs. 135 black coral colonies). The five morphological groups of sponges displayed broad distributions across habitat types but were espe- cially dense on high-relief rock and low-relief mixed rock (Fig. 7). Flat, barrel, vase, and foliose sponges were found in all habitats; shelf sponges were found in all habitats except MM, MS, and SS. Foliose sponges were found at significantly deeper depths (mean=191 m; SE = 53; n=1259) than were other sponge groups (pooled mean=152 m; SE = 0.6; n=7545), which were not significantly dif- ferent from each other (Kruskal-Wallis //=594; df=4; P<0.01). Generally sponge size increased with increasing depth, although the correlation was low (/■=0.07; P<0.001; n = 6551). Although sponges were found throughout the study area, gorgonians and black corals were restricted in their distribution to a small number of sites (Fig. 8). Structure-forming invertebrates displayed wide varia- tion in size; maximum height ranged from 4 cm for brachiopods to 2.5 m for black corals (Table 2). There was no significant correlation between size of the in- vertebrate taxa and structural relief of the substratum types (r=0.28, P=0.30, /z = 15). 20 10 < 90 m. n =770 patcties IJ nH^^n r-1 n 20 - 91-125 m. n = 1087 patcties 10 ■ n , , p in rrn ^ n n n n 20 - 10 - 126-175 m. n=751 patcties n CL - n 20 - 10 >175m, n =570 patches _a n^rin LX TT RR RM BB BC BS CB IV1B SB CS SC SP IVIG SG IVIIV1 fVIS SS Habitat code Figure 3 The frequency of habitat patches of each substratum type, stratified by depth. See page 169 for definitions for the substrate abbrevia- tions along the .r axis. Gorgonians and black corals had different size distri- butions (Fig. 9). Black corals ranged from 10-250 cm in height (mean=33.6; SE=1.9; n=195) and most indi- viduals were in the 10-50 cm size category. Individuals were found in three color forms: gray-to-white (SO'/f ), rusty-brown-to-red (47%), and gold (3%). Gorgonians ranged in size from 10 to 40 cm (mean=21.7; SE = 1.2; n=27) and were found in multiple morphological forms from elongate to fan-like (Fig. 6). Sponges displayed similar size distributions to those of gorgonian corals, but had different mean and max- imum sizes (Fig. 9). Mean sizes of flat, barrel, and foliose sponges were not significantly different from each other (pooled mean=19.8 cm; SE = 0.1; ;!=7373) but were significantly smaller than vase and shelf sponges (pooled mean = 20.9 cm; SE = 0.2; ?! = 1289), which were not different from each other (one-way ANOVA; F=4.52; df=4,8657; P=0.001). The maximum observed height was 50 cm for shelf sponges, 60 cm for foliose sponges, and 100 cm for barrel, flat, and vase sponges (Fig. 9). Most (98.2% by number) sponges, gorgonians, and black corals by number did not have any other organ- isms living on them (Table 3). Overall, crinoids (1.4%) were most commonly associated with these large in- vertebrates, followed by sponges (0.1%), and nine other 172 Fishery Bulletin 104(2) Table 2 Significant criteria of structure-forming invertebrates on rocky outcrops off southern California. "X" indicates taxa that exhibit large size, complex morphological shape, and high-density aggregations. SE is standard error and n is the total number of obser- vations per taxon. The sample sizes for density and depth calculations were reduced because of incomplete observations on all individuals. Organisms are ordered from largest to smallest group. Taxa n Criteria Density (no. /hectare) Maximum height (cm) Depth (m) Mean physical structural relief Size Complex morphology High density Mean SE Mean SE Black coral 13.5 tAntipathes dendrochristos ) X X 10 4 250 183 24 1.7 Flat sponge 4240 X 312 53 100 149 43 1.9 Barrel sponge 2138 X 163 28 100 151 46 2.0 Vase sponge 1167 X 90 16 100 163 44 1.9 Sea pen (Subselliflorae) 9726 X X 819 153 100 157 61 1.4 Basket star iGorgonocephalus eucriemis) 733 X X 58 12 90 162 44 1.7 White-plumed anemone iMetridium farcuneni 57 X 5 2 80 161 46 1.9 Foliose sponge 1259 X 96 25 60 191 53 2.0 Shelf sponge 139 X 10 3 50 158 47 2.1 Gorgonian 1 Gorgonacea I 27 X 3 3 40 159 4 1.2 Crinoid [Florometra serratissima 174,231 ) X X 22,045 8466 35 146 46 1.6 Brittle star (Ophiuridae 207,667 X 17,786 4301 15 148 47 1.8 Fragile sea urchin (Allocentrntus fragilif:) 18,363 X 2031 668 8 185 45 1.8 White sea urchin iLytechinu.s anamesui^i 45,092 X 4631 1388 4 85 20 1.2 Brachiopod 1 Order Terebratulida ) 56,924 X 3623 1705 4 100 17 1.8 Table 3 Associations of organisms with large structure-form ng invertebrates on rocky banks off southern California. Associated organisms {% of total observations) Egg Basket Brittle Sea Taxa u None C inoid s Sponges Fishes cases Crabs stars stars stars Anemone Algae Salps Foliose sponge 1262 99.0 0.6 0.3 0 0 0 0 0 0 0 0 0 Flat sponge 4043 99.5 0.5 <0.1 0 0 0 <0.1 0 0 0 0 0 Barrel sponge 2068 98.0 1.9 0 0 0 0 0 0.1 0.1 0 0 0 Shelf sponge 134 99.3 0.7 0 0 0 0 0 0 0 0 0 0 Vase sponge 1155 96.1 3.6 0 0.1 0.1 0 0.1 0 0 0 0 0 Gorgonian 27 100 0 0 0 0 0 0 0 0 0 0 0 Black coral 194 84.7 6.8 3.1 1.3 0 1.8 0.4 0.4 0 0.4 0.4 0.4 Total 8883 98.2 1.4 0.1 <0.1 <0.1 <0.1 <0.1 <0,1 <0.1 <0.1 <0.1 <0.1 Tissot et al : Invertebrates on deep banks off soutfiern California 173 Figure 4 Some major structure-forming invertebrates on southern California rocky banks: (A) crinoids {Flo- rometra fierratissima), (B) basket stars (Gorgonocephalus eucnemis), (C) brittle stars (Ophiacantha spp.), (D) brachiopods (Terebratulida i, lEl white sea urchins [Lytechinus anamesus), (F) white-plumed anemones (Metridium farcimen), (G) fragile sea urchins (AUocentrotus fragilis). and IH) sea pens iSubselliflorae). 174 Fishery Bulletin 104(2) taxa fall <0.1'7(). Black corals had the largest incidence of associated animals (15.3% of individuals), followed by vase (3.9%), barrel (2%), foliose (1%), shelf (0.7%), and flat (0.5%) sponges. Fish were most commonly observed on black corals (1.3%) but were also observed on vase sponges, including an attached egg case. No organisms were observed living on gorgonians. The frequency of fish species near sponges, gorgoni- ans, and black corals was significantly different from the frequency of those same species found elsewhere along transects (chi-square, all P<0.01; Table 4). Of the 108 species adjacent to these large invertebrates, six species were found at significantly higher frequencies than predicted by their density along transects: cowcod, bank rockfish iSebastes rufus). swordspine rockfish iSebastes ensifer). shortbelly rockfish iSebastes jordani), pinkrose rockfish iSebastes simulator), and members of the rockfish subgenus Sebastomus (Table 4). 600 Cnnoids nn^nnnrir^nn ■^^-^L. The distribution of mean nearest-neighbor distances between fishes and large invertebrates varied from 0.1 to 9.9 m (Fig. 10). Overall median distances varied from 0.9 m (shelf sponges) to 1.8 m (black corals). However, there were no statistical differences between the me- dian distances for each group (Kruskal-Wallis, H=WA; df=6; P=0.11). For the six fishes that were found more frequently near large invertebrates than on transects, the overall median distances to the invertebrates were 5.5 m (cowcod). 1.0 m (bank rockfish), 1.3 m (swordspine rockfish), 1.5 m (shortbelly rockfish), 1.7 m (pinkrose rockfish), and 1.4 m (Sebastomus). The overall incidence of damaged and dead sponges, gorgonians, and black corals was low (0.3% of total num- ber observed). Black corals were more commonly damaged (1.7%) or dead (1.1%), followed by vase sponges (0.6% and 0.1%, respectively), barrel sponges (0.5% and 0%), and foliose sponges (0% and 0.1%). No dead or damaged flat or shelf sponges or gorgonians were observed. Damage in black corals included portions of the colony that appeared discolored; dead black cor- als lacked polyps and were discolored. Among sponges, the most common damage was individu- als that had broken from the substratum and were lying on their side or broken colonies. Basket stars I r^ r^ .J. r^ n r;-! 0 fi ri r^ Pn .=,,—, Pn Fl 900 - Brittle stars ^r^^OUO ' I ' ' t ' ' I ' '1 n JL ^n^^ 500 E o Bractiiopods ^^^^ ^1 1^ >, 200 - Wtiite sea urchins T T n=45,092 c Q 0 - -n- 1 r- rn n fl , ^ 2 n White plumed anemone i rJ^ ^ , ^ , 200 - Fragile sea urchins 1 .n. o ^. -Oa 50 Sea pens rjl ^ -^ -, p O rp [7] n=9,726 X n_ TT RR RIVl BB BC BS CB MB SB CS SC SP MG SG MM MS SS Habitat code Figure 5 Mean density of structure-forming invertebrates in habitat patches of each substratum type. Vertical bars are + one standard error. See page 169 for definitions for the substrate abbreviations along the X axis. Discussion Several groups of invertebrates were distin- guished by their large size, such as black corals, sponges, crinoids, basket stars, anemones, and sea pens. Organism size is an important aspect of structural habitat because it contributes to vertical relief and increases the availability of microhabitats. For example, yelloweye iSebastes rube?Ti?nus) rockfish may use the large gorgo- nian coral Primnoa as a vantage point to prey upon small fishes (Krieger and Wing, 2002). Size variation among structure-forming inver- tebrates was significant. Individual black corals, sea pens, and sponges greater than 1 m in height represented only 0.1% of all organisms, and 90% of the individuals were <0.5 m high. Similarly, the complex structures of crinoids, gorgonians, black corals, and basket stars may increase the availability of microhabitats and create a high surface area for settlement or retention of other organisms. Fish egg cases have been observed attached to both gorgonians (Etnoyer and Morgan') and vase sponges (pres- ent study. Table 3). High-density aggregations have not been used as a criterion for defining structure-forming invertebrates, but there are several examples that illustrate the potential importance of these aggregations. High density "forests" of crinoids provide refuge and substrata for a wide variety of small fishes and invertebrates in rocky areas (Lissner and Benech, 1993; Puniwai, 2002). Tissot et al Invertebrates on deep banks off southern California 175 Figure 6 Structure-forming invertebrates; (A) gorgonians (Gorgonacea), (B) black coral iAntipathes dendrochristos), (Cl flat, (D) barrel, (E) shelf, (F) vase, and (G) foliose sponges on southern California rocky banks. 176 Fishery Bulletin 104(2) Similarly, high-density aggregations of brittle stars and brachiopods in boulder-cobble areas and fields of sea pens and sea urchins in sand and mud habi- tats also may provide space and structure for other organisms (e.g., Brodeur, 2001). With the exception of black corals and sea pens, the largest structure-forming invertebrates, such as sponges, Metridium spp., and crinoids, were most common on high- to moderate-relief rocky habi- tats. These long-lived organisms are likely to be favored in stable habitats that are more insulated from sediment transport and high particle loads than low-relief, mud-dominated areas (Lissner and Benech, 1993). Large invertebrates add structure and micro-scale complexity to these rocky habitats that already contain high-to-moderate amounts of relief Sponges, with their broad distributions, may also provide structure for flatfishes in low-relief mud habitats (Ryer et al., 2004) Black corals and gorgonians, in contrast, are more commonly found in current-swept areas near drop-offs and under ledges (Grigg, 1974; Parrish, 2004). In our study, these invertebrates were found in low-relief mixed cobble-boulder-sand habitats at 100-225 m depths, providing significant vertical structure for potential use by a wide variety of organisms. Aggregations of sea pens and sea urchins may provide important structure in low-relief sand and mud habitats where there is little physical habi- tat complexity. In addition, these organisms may provide refuge for small planktonic and benthic invertebrates, which in turn may be preyed upon by fishes. They also may alter water current fiow, thereby retaining nutrients and entraining plank- ton near the sediment. Urchins rapidly respond to patches of drift kelp (Harrold and Reed, 1985), which provide organic material to deep sea habitats (Harrold et al., 1998). One of the central issues currently relevant to structure-forming invertebrates is the degree to which these species contribute to the spawning, breeding, feeding, or growth-to-maturity of eco- nomically important fishes. Although there are several studies that report fish-invertebrate associations within common habitats (Hixon et al.^), or make anecdotal or general observations on fish-invertebrate associations (e.g., Krieger and Wing, 2002), few studies have sys- temically quantified these relationships. In our study. for 9105 observations on the larger invertebrates found on southern California rocky banks only 1.8% of indi- viduals had other organisms lying on or attached to them. Moreover, the vast majority of these organisms were other invertebrates, including crinoids, sponges, crabs, basket stars, brittle stars, seastars, anemones and salps. Less than 1% of the observations of organ- isms actually sheltering in or located on invertebrates involved fishes (a total of five individuals and one egg case), and most were observed on large black corals (Table 3). This result implies that fishes are not strong- 0 10 0 05 ■ 0 00 0,5 0,0 10 0,6 - 0,0 4 0 10 Gorgonians n=27 Blacl< corals 1 A ^^ -ia Flat sponges n nll^nHn^ n^ Barrel sponges on A r;i r^ r^-i n ^ ,J^ r^ Stielf sponges ^^^n 1 Habitat Code n=139 i htJx M- Vase sponges 1 A I, m Oii r^ ,^ ^ r-p Foliose sponges 1 M n r;n ri rp I I r-1 i XU- Tl X -r -I- T TT RR RM BB EC BS CB MB SB CS SC SP I^G SG MM MS SS Habitat code Figure 7 Mean density of gorgonians, black corals, and sponges in habitat patches of each substratum type. Vertical bars are ± one standard error. See page 169 for definitions for the substrate abbreviations along the .v axis. ly associated with structure-forming invertebrates in the areas we surveyed off southern California. However, we should note that our observations were limited to daylight hours and that the viewing angle from the submersible generally precluded seeing inside some of the sponges (especially vase and barrel types). Moreover, our analyses focused on associations between fishes and individual solitary invertebrates, most of which were <0.5 m in height. We did not examine as- sociations between all structure-forming invertebrates, nor did we examine associations between invertebrates and assemblages of fishes at the level of discrete habitat patches (100-1000 m scale) (e.g., Tissot et al.^) ■• Tissot, B. N., M. A. Hixon, D. L Stein. Unpubl. manu- script. Habitat-base submersible assessment of groundfish assemblages at Heceta Bank, Oregon from 1988-1990. Tissot et al , Invertebrates on deep banks off southern California 177 120 W 119 30'W 34°30'N 34 N 33°30N 33°N 32°30N Point Conception San Miguel Is. Santa Cruz Is ^^o ^... Santa Rosa Is. ^^nacapa Is Harrison's Reef Dume Canyon Reef 1 1 7 30 W 117 No. of Colonies • 1- 2 • 3- 4 • 5- 7 • 8- 23 # 24 -57 Hidden Osborn Bank-c^/ . \\V.. v. ^< (Cv'^'l Potato Bank. Saa^icdtes Is. - — .V '-^ .TV\ f — ■^^^ X/ -^y^t 0 25 50 100 km \^ Cherry Bank ■A r-'.Tannel/Cortes[6ank \fc:"' -V. • »\'> :).4 rtO'N 33 30 N ^^^ ¥A" 32'30N 120 30' W 120 W 1 1 9°30'W 117°30W 119 'SOW 119"W 118 30W 33'30'N 32°30N B Point Conception ^""'^ u San H/liguel Is. Santa Cruz Is. ^ .^ * - --■ _ Anacapals. "~- •~», , ' -— ' --^. •n:^~-'^ - "^ ^^ vr _ V tSrrison's Reef Dumfe Canyon ^- -" ' <;nntfl RAfin l«. - No. of Colonies • 1 • 2 X-^-v , Santa Rosa Is. % %. Hidden] Reef "*v _rv v^-- V \ Kidney BaTm' < I ..^..^ Santa Barljara lb ., , 5,. ^J \ ''< 'sa^liit'olasls. -—X X W'CH^ryBank tx ^^- V. '^ 0 12 5 25 50 km 1 1 I 1 III ■ "" ^■S5i^;U:-/Tar)nel/CortesBank\ '.-;> ;:i: r-g>i|th^ t\ ■^ r--N 34 30'N 34 N 33"N 120-30W 120 W 119"30'W 11 9° W 117°30'W Figure 8 Distribution and number of (A) black coral and (B) gorgonian colonies inside and outside the cowcod conservation areas (delineated as outlined boxes) of southern California in 2002. Dive sites are indicated by black dots. 178 Fishery Bulletin 104(2) From the analysis of spatial associations between fishes and large, individual structure-forming inver- tebrates, six of 108 species were found more often ad- jacent to colonies than predicted by their abundance along transects. This result indicates that there may be spatial associations that do not necessarily include physical contact with the sponges and corals. However, the median distances between these six fish species and large invertebrates (1.0-5.5 m) were not particularly small. Thus, it is likely that these fishes and inverte- brates are present in the same types of habitats and that there is not necessarily a functional relationship between these two groups of organisms. Parrish (2004) reached similar conclusions on stud- ies of black coral in Hawaii. Although fish densities were higher in areas that included corals, when bot- tom relief and depth were accounted for these densi- ties were not higher than those for surrounding areas without corals. Thus, there was no clear evidence that corals served to aggregate fish. Rather, fishes and cor- als co-occurred in areas with similar physical relief and unique flow regime (Parrish, 2004). Auster (2005) also reached similar conclusions by finding no significant difference in the density of a common rockfish species iSebastes fasiatus) between areas of rock and boulders with dense coral cover and similar areas having dense epifaunal cover (i.e., without coral). Auster concluded that although dense coral and dense epifaunal habitats were functionally equivalent, the epifaunal habitat was more widespread in his study area, making that habitat perhaps more important to the greater rockfish popu- lation. Finally, Syms and Jones (2001) demonstrated that removal of high densities of soft corals caused no significant changes in the associated fish communities and that the heterogeneity of habitat generated by soft corals was indistinguishable from equivalent habitat formed by rock alone. Thus, fish-invertebrate associa- tions, by themselves, do not necessarily demonstrate the functional importance of invertebrates as habitat to benthic fish populations. One possible conclusion from our study is that ob- served fish-invertebrate associations, like those reported for many cold-water corals, can be overstated. In the absence of quantitative information, observers may re- member the few positive associations between fishes and structure-forming invertebrates but ignore (or forget) the more numerous observations of large invertebrates with no associated fishes. Indeed, the general impression of the authors after making submersible observations was that there were higher numbers of fishes associated with large invertebrates when in reality only five fishes were observed lying directly on a large invertebrate in the video transects. A more likely conclusion, however, is that the continental shelf communities of southern California are unique and that large black corals do not have the high number of commensals as seen, for example, on Primnoa in Alaska. An additional consider- ation is the relatively low number and size of individual sponges, gorgonians, and black corals observed in this study. Primnoa in Alaska can form massive stands 3 m 20 10 0 200 100 0 3000 1500 0 CO ■g > 1000 "D g 500 o 0 d z 100 50 0 400 0 800 400 0 Gorgonians X. 15 20 25 30 40 Black corals 50 100 150 200 250 n=4,043 Flat sponges r^ n i 10 20 30 40 50 60 70 80 90 100 n=2,068 1— Barrel sponges n n t 10 20 30 40 50 60 70 80 90 100 "=134 Shelf sponges 10 20 30 40 50 60 70 80 90 100 n=1,155 Vase sponges rn n „ i 10 20 30 40 50 60 70 80 90 100 n=i.262 Foliose sponges 10 20 30 40 50 60 70 80 90 100 Height (cm) Figure 9 Size distributions of large, structure-forming inver- tebrates. Arrows indicate ma.ximum sizes observed in each group. tall and 7 m wide in some areas (Krieger and Wing, 2002). Moreover, the majority of fishes were observed on the largest individuals in their study (>15 m^ in volume) Most of the corals in our study were <0.5 m in height. Given the importance of this issue, we argue for more rigorous quantitative studies on fish-invertebrate associations that would include densities of fishes and sizes of both fishes and invertebrates. Regardless of their associations with fishes, the struc- ture-forming invertebrates described in this study are very likely to be ecologically important on continental shelf ecosystems and are certainly significant in their own right. Observation on the health of the larger in- vertebrates indicates few damaged (0.1%) or dead (0.2%) individuals and a low incidence of fishing gear in the areas surveyed (Tissot et al., unpubl, data). These ob- servations are consistent with the absence of a signifi- cant commercial bottom trawling fishery in our survey area, which has been associated with negative impacts on large invertebrates in other locations (Watling and Norse, 1998; Freese et al. 1999: Krieger. 2001). Thus, this study affords a unique view of what appears to be Tissot et dl Invertebrates on deep banks off southern California 179 6 3 -a a -a ^ c a ■- Si ^ a to «} -c . a -2 >. o -a v. 0) *^ a ^ c _o CO CO c a; HI /■ ^ QJ •^ -C ".I CO CO ^ c S c ^ M CO bo *^ , c/l M g S o. o o a r: u T. crt y= 0) > T3 c !B q; an cl; c o 0- CO m o o H e-s H C bfl — J= a; CO "S tin E-> CO Q. c H CO -2 2^ — j5 0) to =C be o '- C [2 CO o c ■-« V ~ o CO CO •-^ Vi CO ^ CO X C ■^ r; ■^ en CJ CO 0) CO r* CO Si -2 c 1 5 3 t>5 0^ ^ s •n CC J:! p 8 3 ^ o o ;~i 1 £ o CO ■J: "5, o ;- _Cj CC P a a CO tC a a CO cc ^ ■^ 8 ■a CD cc Q. a. CO Li o O "j; S-H n C>", Q o ■-A a; cc £! >, n a ■j; y= o: s o. T3 Ql ^,) ^ O) ^^■• S a -a o ■j; to CC en 01 en CO C = OJ 'iJ T3 O CJ •*^ S CO o C^ >. CI Cl CI n to o n M ■A CO c c cj CO bB CO C3 rr o r! 1 c CO -o CO "to -a CO c c c > 3 CD CO o CO S 1 CO CO OJ CO D w a. tl. C/2 C/^ pa E :u m ^ u c/: O Di 180 Fishery Bulletin 104(2) -u 15 t-H ^ o i=! V V V V V a Xj ;- o u ^ en s ? m s ^ 4J o o o o o c CO i-H e2 V V V c -D .2 'S o en bo a a "3 o V, .2 o ts c a; o o o ^ o o H C bo o c CO ^ CO ^ ^ y: o V V Ci H ca o a « Cm ;<-- 1 s ^ o ^ "n ^ T-K ^ T-H I-l u CO o V 3 C H M Q. 'S c 0 jj LI ^^ u to o c; V c 2 o c CO .—1 Tj- V ^ V ^ 2 £ ."I 1- "5 -tj c^ o o CM X ^ « 1 OJ b£) QJ C "3 o O ? CI. ■ x; oj '11 CO s f >— t !— ! r-H o ^ 2 1 V V V S c o O ^ s CQ tc bo •„ c c^ IN H 01 o c <" t. a) CC n! bo ^ S ta Q. c CO CM .—1 '^. rX o 3 ■C C O Q f^ en 2 ■3 o o "^ c; 2? CO ■5 ^ 2 e ^1 3 a. -21 ij CO tlD o 5 o a; OJ ^ ■a .2 « t. cc >^ O -a o C- 0 M e OJ c s-i CJ c t, -o ? -o -^ ';:: a -o ro CO "iJ 01 ^ u "-^ bo a. CO ^ X ra f « t ^ M tt; c O g^ H Q O 3 J CQ 0 30 0 20 0 15 0 20 0 20 -1 I-^ r-n Gorgonians median=1.4m — 1 — — Black corals median=l 7 m n l— 1 Flat sponges median=1.4m — 1 r-l Barrel sponges median=i 3m — Shelf sponges median=0.9m (— 1 Vase sponges median=1.4 m — Foliose sponges median=i 3 m 01 23456789 10 Mean distance to nearest fish (m) Figure 10 Mean nearest neighbor distances of fish species to gorgonians, black corals, and sponges. relatively undisturbed megafaunal invertebrate commu- nities in the southern California Borderlands, and sup- ports the continued protection of these animals within the Cowcod Conservation Areas. Acknowledgments We thank the dedicated crews of the RV Velero and Delta for support in the field. D. Schroeder, M. Nishimoto, L. Snook, T. Laidig, T.Anderson, R. Starr, B. Lea, and C. Wahle assisted with the survey work. K. Greenwood and S. Green assisted with habitat and invertebrate analyses. Funding was provided by NOAA Fisheries, SWFSC Fisheries Ecology Division, Offices of Habitat Conservation and Protected Resources; the National Undersea Research Program; NOAA Marine Protected Area Science Center; the David and Lucile Packard Foundation; Washington State University Vancouver; and the Monterey Bay Sanctuary Foundation. Tissot et al Invertebrates on deep banks off southern California 181 Literature cited Andrews, A. H., E. E. Cordes, M. M. Mahoney, K. Munk, K. H.Coal, G. M. Calliet, and J. Heifetz. 2002. Age, growth, and radiometricage validation of a deep-sea, habitat-forming gorgonia iPrimnoa reife- daeformin) from the Gulf of Alaska. Hydrobiologia 471:101-110. Auster, P. J. 2005. Are deep water corals important habitat for fishes? Cold-water corals and ecosystems (A, Freiwald and J. M. Roberts (eds.l, 1244 p. Springer, New York, NY. Brodeur, R. D. 2001. Habitat-specific distribution of Pacific ocean perch iSebastes ahitiis) in Pribilof Canyon, Bering Sea. Cont. Shelf Res. 21:207-224. Brusca, R. C, and G. J. Brusca. 1990. Invertebrates, 922 p. Sinauer Associates, Inc., Sunderland, MA. Dayton, P. K., S. Thrush, and F. C, Coleman. 2002. Ecological effects of fishing in marine ecosystems of the United States, 52 p. Pew Oceans Commission, Arlington, VA. Fossa, J. H., P. B. Mortensen. and D. M. Furevik. 2002. The deep-sea coral Lophclia pertufa in Norwegian waters: distribution and fishery impacts. Hydrobiologia 471:1-12. Freese, J. L. 2001. Trawl-induced damage to sponges observed from a research submersible. Mar. Fish. Rev. 63(3):7-13. Freese, J. L., P. J. Auster, J. Heifetz, and B. L. Wing. 1999. Effects of trawling on seafloor habitat and associ- ated invertebrate taxa in the Gulf of Alaska. Marine Ecol. Prog. Ser. 182:119-126. Grigg, R. W. 1974. Distribution and abundance of precious corals in Hawaii. In Proc. second international coral reef sym- posium (A. M. Cameron et al., eds.), p. 235-340. Great Barrier Reef Committee, Brisbane, Queensland, Aus- tralia. Harrold, C, and D. Reed. 1985. Food availability, sea urchin grazing, and kelp forest community structure. Ecology 66:1160-1169. Harrold, C, K. Light, and S. Lisin. 1998. Organic enrichment of submarine-canyon and continental-shelf benthic communities by macroalgal drift imported from nearshore kelp forests. Limnol. Oceangr. 43:669-678. Heifetz, J. 2002. Coral in Alaska: distribution, abundance, and species associations. Hydrobiologia 471:19-28. Husebo, A. L. Nottestad, J. H. Fossa, D. M. Furevik, and S. B. Jorgensen. 2002. Distribution and abundance of fish in deep-sea coral habitats. Hydrobiologia 471: 91-99. Krieger, K. J. 2001. Coral \Pniunoa) impacted by fishing gear in the Gulf of Alaska. In Proceedings of the first international symposium on deep-sea corals (Willison et al., eds.), 106-116. Ecology Action Centre, Halifax, Canada. Krieger. K. J., and B. L. Wing. 2002. Megafaunal associations with deepwater corals iPrimnoa sp.) in the Gulf of Alaska. Hydrobiologia 471:83-90. Lissner, A., and S. Benech. 1993. Benthic hard-substrate community ecology of the Santa Maria basis and western Santa Barbara channel. In Taxonomic atlas of the benthic fauna of the Santa Maria Basin and western Santa Barbara Channel (J. A. Blake and A. L. Lissner, eds.), p. 33-45. Santa Barbara Museum of Natural History, Santa Barbara, CA. Malakoff, D. 2004. U.S. asked to act immediately to protect deep-sea coi-als. Science 304(5667):31. NRC (National Research Council). 2002. Effects of trawling and dredging on seafloor habitat. Committee on Ecosystem Effects of Fishing: Phase 1-effects of bottom trawling on seafloor habitats, 136 p. National Research Council, National Academy Press, Washington, D.C. Opresko, D. M. 2005. A new species of antipatharian coral (Cnidaria: Anthozoa: Antipatharia) from the southern California Bight. Zootaxa 852:1-10. Palumbi, S. R. 1986. How body plans limit acclimation: responses of a demosponge to wave force. Ecology 67( 1):208-214. Parrish, F. A. 2004. Foraging landscape of the Hawaiian monk seal. Ph.D. diss., 146 p. Univ. Hawaii, Manoa. Puniwai, N. P.F. 2002. Spatial and temporal distribution of the crinoid Florometra serratissma on the Oregon continental shelf. M.S. thesis, 34 p. Washington State Univ., Vancouver, WA. Roberts, S., and M. Hirshfield. 2004. Deep-sea corals: out of sight, but no longer out of mind. Front. Ecol. Environ. 2(3):123-130. Rogers. A. 2004. The biology, ecology and vulnerability of deep-sea coral reefs. lUCN publication:l-13. Riedl, R. 1971. Water movement: animals. Marine ecology: a com- prehensive, integrated treatise on life in oceans and coastal waters (O. Kinne, ed.), 1(2):1123-1156. Wiley- Inter Science, London. Ryer, C. H. A. W. Stoner and R. H. Titgen. 2004. Behavioral mechanisms underlying the refuge value of benthic habitat structure for two flatfishes with differing anti-predator strategies. Mar. Ecol. Prog. Ser. 268:231-243. Stein, D. L., B. N. Tissot, M. A. Hixon and W. Barss. 1992. Fish-habitat associations on a deep reef at the edge of the Oregon continental shelf. Fish. Bull. 90:540-551. Syms, C, and G. P. Jones. 2001. Soft corals exert no direct effects on coral reef assemblages. Oecologia 127:560-571. Watling, L., and E. A. Norse. 1998. Disturbance of the seabed by mobile fishing gear: a comparison to forest clearcutting. Cons. Biol. 12(6):1180-1197. Yoklavich, M. M., H. G. Greene, G. M. Cailliet, D. E. Sullivan, R. N. Lea, and M. S. Love. 2000. Habitat associations of deep-water rockfishes in a submarine canyon: an example of a natural refuge. Fish. Bull 98:625-641. 182 Abstract — Larval and juvenile rock- fishes iSebastes spp.) are difficult to identify using morphological charac- ters. We developed a key based on sizes of restriction endonuclease frag- ments of the NADH dehydrogenase-3 and -4 (ND3/ND4) and 12S and 16S ribosomal RNA (12S/16SI mitochon- drial regions. The key makes use of variation in the ND3/ND4 region. Restriction endonuclease Dde I varia- tion can corroborate identifications, as can 12S/16S variation. The key, based on 71 species, includes most North American taxa, several Asian species, and Sehastolnbus alascanus and Heli- colenus hilgendnrfi that are closely related to rockfishes. Fifty-eight of 71 rockfish species in our database can be distinguished unequivocally, using one to five restriction enzymes; identities of the remaining species are narrowed to small groups: 11 S. polyspinis. S. crameri, and S. ciliatus or variabilis (the two species could not be distinguished and were consid- ered as a single species) ; 2) S. chlo- rostictus, S. eos, and S. rosenblatti: 3) S. entomelas and S. mystinus; 4) S. emphaeus, S. variegatus, and S. wilsoni; and 5) S. carnatus and S. chrysomelas. A key to selected rockfishes iSebastes spp.) based on mitochondrial DNA restriction fragment analysis Zhuozhuo Li' Andrew K. Gray' Milton S, Love' Akira Goto^ Takashi Asahida^ Anthony J. Gharrett' ' Fisheries Division, School of Fisheries and Ocean Sciences University of Alaska Fairbanks 11120 Glacier Highway Juneau, Alaska 99801 Email address (for A J Gharrett, contact author) ffaigia'uafedu ^Marine Science Institute University of California Santa Barbara, California 93106-6150 3 Graduate School of Fisheries Sciences Hokkaido University 3-1-1 Minato-cho, Hakodate Hokkaido 041-8611, Japan '' School of Fisheries Sciences Kitasato University Sanriku, Ofunato Iwate 022-0101, Japan Manuscript submitted 11 October 2004 to the Scientific Editor's Office. Manuscript accepted for publication 1 August 2005 by the Scientific Editor. Fish. Bull. 104:182-196 (2006). More than 100 species of rockfishes (genus Sebastes) are found worldwide (Kendall, 2000), and the majority are distributed along the Pacific coast of North America and in the northwest Pacific from the western Bering Sea south to Japan and Korea (Love et al., 2002). Sixty-five rockfish species are found along the California coast (Moser, 1996). Within the genus, there is a high degree of similarity of the morphological characters among many species. These similarities are in part due to recent divergence but also may have resulted from convergence of con- geners occupying similar habitats. Historically, species identification of Sebastes has been based on morphol- ogy; however, this approach is often insufficient, especially for identify- ing sympatric species, where there is considerable similarity in size and physical features and in pigmenta- tion, as well as overlap in meristic characters. The difficulty in identify- ing Sebastes to species level exists for all developmental stages, but larvae are especially difficult to identify be- cause they lack diagnostic characters (Kendall, 1991), Currently, only 15 species at the larval stage can be sep- arated by physical characters such as body shape, pigmentation patterns, and head spine development (Love et al., 2002). Rockfishes are important ecologi- cally, and most species are economi- cally valuable. Sebastes larvae form a large portion of ichthyoplankton collections and rank third or fourth in abundance among all fish larvae taken during California Cooperative Fisheries Investigations (CalCOFI) surveys that have covered the entire length of the California and Baja Cal- ifornia coast and now cover waters off southern California. The ability to identify Sebastes accurately and efficiently at all developmental stages will greatly improve both our ability to learn about their life histories and our management and conservation ef- Li et a\. A key to selected Sebastes spp based on mitochondrial DNA restriction fragment analysis 183 forts. An increased understanding of life history varia- tion can improve the systematic descriptions of Sebastes species, which has mostly been based on morphology. Several methods have been developed to obtain spe- cies-specific information that supplements the use of morphological and meristic characters for species identification. Otolith microstructure and other hard structures have been used to distinguish the late- stage larvae and young-of-the-year pelagic juveniles of some Sebastes species (Laidig and Adams'; Laidig and Ralston, 1994; Laidig et al., 1996). Biochemical genetic methods have been used to identify adults (Barrett et al.. 1966; Seeb, 1986), and may also be used to identify juveniles of some species (Seeb and Kendall, 1991). LeClair and Buckley (2001) used allozyme variation at 37 loci to positively identify 155 individuals of juvenile Sebastes diploproa and Sebastes melatwps; but they were unable to identify another 29 individuals, partly because of a limited database. Despite some success with their use, allozymes often have low resolution, and there are complexes of species that cannot be dis- tinguished by using allozymes alone (Seeb and Kendall, 1991). DNA sequence-based methods have some advantages over the use of allozymes for the identification of juve- nile fish species. They often have greater resolution, due in part to the fact that expression of some metabolic enzymes changes during development, whereas DNA sequences generally do not. DNA is less susceptible to degradation than the enzymes used in allozyme stud- ies. Also, DNA sequence-based methods require a small amount of tissue sample, which is especially suited for work with samples from early life stages (e.g.. Gray et al., 2006). Because it has a relatively high rate of substitution (Moritz et al., 1987), mitochondrial DNA (mtDNA) can be useful for distinguishing closely re- lated species. Mitochondrial DNA is usually inherited maternally in vertebrates and does not recombine with paternal mtDNA if it is leaked into the zygote (Gyllen- sten et al., 1991; but see Rokas et al., 2003). Thus, the mtDNA sequence represents a matrilineal phylogeny and is often useful in delineating phylogenetic rela- tionships of closely related species. Sequences of the mitochondrial cytochrome b gene were used to identify the pelagic young of Sebastes constellatus and Sebastes ensifer (Rocha-Olivares et al., 2000). Multiplex PCR of haplotype-specific regions of mtDNA has also been used to identify early stages of Sebastes species (Rocha- Olivares, 1998). The objective of this study was to devise a key for species identification using mtDNA restriction fragment data from both the ND3/ND4 and 12S/16S regions. In a previous study, data for species-specific mtDNA restriction site variation in the ND3/ND4 region were ' Laidig, T. E., and P. B. Adams (eds. ). 199L Methods used to identify pelagic juvenile rockfish (genus Sebastes) occurring along the coast of central California. NOAA-TM-NMFS- SWFSC-166. 180 p. NMFS Southwest Fisheries Science Center, 110 Shaffer Rd., Santa Cruz, CA 95060. presented for 15 Sebastes species (Gharrett et al., 2000). We have included data for 56 additional species using the ND3/ND4 and 12S/16S regions of the mtDNA as the target region of PCR-amplification. We have included specimens of Helicolenus hilgendorfi and Sebastolobus alascanus for contrast. Materials and methods Adult specimens of 71 species of rockfishes were collected from the Gulf of Alaska, the coastal waters of Califor- nia and Baja California, and from the coast of Japan. Samples of H. hilgendorfi and Sebastolobus alascanus, species from two sister genera, were collected from Japa- nese coastal waters and the Gulf of Alaska, respectively, to provide outgroup comparisons. A sample of heart tissue from each specimen was preserved in either 95% ethanol or a solution that was 80% 0.25M ethylenediami- netetraacetic (EDTA) acid at pH 8 and saturated with NaCl (Seutin et al., 1991) and 20% dimethyl sulfoxide (DMSO). At least five individuals of each species were analyzed, except for a few species for which less than five specimens were available. We did not distinguish between S. ciliatus and S. variabilis, which had not yet been described (Orr and Blackburn, 2004), when we collected samples. Total genomic DNA was isolated by using a Purgene DNA^M isolation kit (Gentra Systems, Inc., Minneapolis, MN). Two target regions of mtDNA were amplified by using the polymerase chain reaction. The ND3/ND4 region begins in the glycyl tRNA gene and spans the NADH-dehydrogenase subunit-3, arginyl tRNA, NADH- dehydrogenase subunit-4L, and NADH-dehydrogenase subunit-4 genes, ending in the histidyl tRNA gene. The 12S/16S region starts near the phenylalanyl tRNA end of the 12S rRNA gene, and runs through the valyl tRNA gene to near the leucyl tRNA end of the 16S rRNA gene. Primers for target regions have been used to amplify these regions in northern Pacific rockfish (Gharrett et al., 2001). The lengths for the amplified ND3/ND4 and 12S/16S regions are about 2385 and 2430 base pairs, respectively, based on the aggregate restriction maps. Subsamples of the PCR products of each individual were digested with one or more of the restriction endo- nucleases BstN I, BstU I, Cfo I, Dde I, Hind II, Hinf I, Mbo I, Msp I, Rsa I, and Sty I. Fragments were sepa- rated electrophoretically in 1.5% agarose gels (one part agarose [Sigma-Aldrich, St. Louis, MO] and two parts SynergeF'"^' [Diversified Biotech Inc., Boston MA]) in 0.5xTBE buffer (TBE is 90mM Tris-boric acid, and 2 mm EDTA, pH 7.5). A 100 base-pair ladder provided molecular weight markers to estimate restriction frag- ment sizes. Gels were stained with ethidium bromide and photographed on an ultraviolet light transillumi- nator. Restriction fragments that could not be accu- rately measured from agarose gels were separated on 8% polyacrylamide gels and stained with SYBR Green I Nucleic Acid Stain'^^' (Molecular Probes, Eugene, OR) using a 25-bp ladder for a molecular weight standard. 184 Fishery Bulletin 104(2) A restriction site map was cre- ated for each endonuclease by using all observed restriction fragment patterns and neces- sary double digests. Results Baseline data 2000' 1000- _ 500- Q. S 300- 200- 100- 50- Interspecific variation was observed for all enzymes in both mitochondrial regions, except for Hind II in the 12S/16S region. Intraspecific variation was observed for sev- eral enzymes. A total of 215 restriction sites were detected in the ND3/ND4 and 12S/16S regions (Appendix 1). Of the 215 sites, 97 were unique to Sebastes species, 21 were unique to Sebastolobus alas- carius, seven were unique to H. hilgendorfi, and one was shared by Sebastolobus alas- caniis and H. hilgendorfi. The ND3/ND4 region had 141 sites, and the 12S/16S region had 74 sites. In the ND3/ND4 region, 36 sites were common in all haplotypes. whereas 46 occurred in only a single haplotype. A total of 132 composite haplotypes resulted from site differences in the two mtDNA regions (Table 1). The individuals of Sebastes species had 127 haplotypes; Sebastolobus alascanus had four haplotypes; and H. hilgendorfi had a single haplotype. Thirty-four of the 71 species displayed intraspecific variation and were represented by more than one composite haplotype; the remaining 37 species had a single composite hap- lotype. Two pairs and three triplets of Sebastes species had identical composite haplotypes and could not be sepa- rated at the species level with this restriction site in- formation (see identification key). These groups were 1) S. carnatus and S. chrysomelas; 2) S. chlorostictus, S. eos, and S. rosenblatti; 3) S. ciliatus or variabilis, S. crameri, and S. polyspinis; 4) S. emphaeus, S. var- iegatus, and S. wilsoni\ and 5) S. entomelas and S. mystinus. Several pairs of haplotypes between variable Sebastes species were separated by a single restriction site. These included 1) S. hopkinsi and S. ovalis — Dde I; 2) S. zacentrus and the S. emphaeus-S. variegatus-S. wilsoni complex — Rsa I; and 3) S. reedi and the S. cilia- tus or variabilis-S. crameri-S. polyspinis complex — Mbo I. The most variable species was S. mystinus, with five haplotypes. Five Sebastes species, S. dalli, S. hubbsi, S. polyspinis, S. trivittatus, and S. zacentrus, as well as Sebastolobus alascanus, had four haplotypes each. T I I I I I — I — I — I — I — I — I — I — I — I — I — I — r — 1 — I — I — I — I — I — I — I — I — I — 1 — I — I — I — I — r OPABIpMqmRVgcCJfiDeHNhEGKnaFrd I ikb Mbo I haplotype Figure 1 A mock gel showing expected fragment patterns for rockfish (Sebastes spp.) hap- lotypes as a result of digestion of the mitochondrial ND3/ND4 PCR product by restriction endonuclease Mbo I. The mobilities are the logarithm of the fragment sizes and separation was assumed to have taken place in l.S'S agarose gels (one part agarose and two parts SynergeF™). The haplotypes correspond to the fragment sizes in Appendix 1 and to the haplotypes in Table 1 and the identification key. A haplotype may occur in more than one species. Most of the fragments that occurred in the shaded region would probably not be well resolved in an agarose gel. The remaining 58 Sebastes species have species-specific markers that allow unambiguous identification. Use of the key The key we developed for Sebastes species was based exclusively on variation in the ND3/ND4 region because it required only a single PCR amplification and variation in the 12S/16S region contributed no additional resolu- tion between species to that provided by the ND3/ND4 region. The key is not a dichotomous key, but it is applied in the same way that taxonomic keys are applied. The first step is to digest PCR-amplified DNA from the ND3/ND4 region with restriction endonuclease Mbo I and to estimate the sizes of the fragments produced by separating the fragments on an agarose-SynergeF*^' gel. The best results will be achieved by using molecular weight markers, digitally photographing the gels, and estimating the fragment sizes with appropriate software. Alternatively, visual recognition of the fragment pat- tern can be accomplished by constructing a graphic key (e.g., Fig. 1 for Mbo I fragments) from fragment sizes predicted by the restriction site maps in Appendix 1. Note that in the figure, the logarithm of the size of the fragment (in base pairs) is used to estimate the mobility of a fragment (Sambrook et al., 1989). When the Mbo I haplotype has been identified, proceed to the next step. For example if the Mbo I digest results in haplotypes D or e, proceed to step g. in the key, which specifies digestion of another subsample of the ND3/ND4 PCR Li et al, A key to selected Sebastes spp based on mitochondrial DNA restriction fragment analysis 185 Table 1 Composite haplotypes for Seba mtDNA regions. The haplotype ites spp., Hclicniciuis hilgcndnrfi. and Seba. codes refer to haplotypes in Appendix 1. ■tolobu s ala scaiuis in the ND3/ND4 and 12S/16S Species ND3/ND4 12S/16S Bs(N I BslVl Cfol Dde\ Htnill Hin(\ Mho I Msp Rsttl Sly I SsfNI BrfUI Cfol Dde\ HindU HmU Mho I Msp I Rsa I S(vl aleutianus A F C B N A E K D D C A A A D A B C B A A aleiitianiis B F C B N A F K D D c A A A D A B c B A A alutiia A B C A J A A B B C c A A A D A B c B A A alutus B B C A J A A B B C c A A A D A A c B A A atrovircnti O B D b C G .1 C B I A A A D A B A A A A aiii-iciilatiia F B D V A G k C B A A A A D A B A A A A aurora B C D m A A b L A e A A A D A A C B A A bahcocki A F C D G A A G B B C A A A C A B c B B A babcocki B F c D F A A H B B C A A A C A B c B B A borealis F c D D A A F B B C B A A A A B c A C A breiispinis G c D C A A K D B D A A A D A A c A C A capensis A F c D G B A E B D 0 A A A C A A c A D A capensis B F c D F B A E B D 0 A A A C A A A A D A capensis C F c D G B A E B D 0 A A A D A A C A D A carnatiis A F a A I C G C a B A A A A D A B A A A A carnatus B F a A I C A C a B A A A A D A B A A A A carnatus C F a a I C G C a B A A A A D A B A A A A caurinus A F B C L A G c C B A A A A D A B A A A A caurinus B F B C L A G J C B A A A A D A B A A A A chlorosticus F B D G B A E B C C A A A C A B C A D A chrysomelas A F a A I C G C a B A A A A D A B A A A A chrysometas B F a A I C G c a B A A A A D A B A A C A ciliatus variabili s A D A M A A F B F C A A A D A B C A A A constellatus A F C D G F A E E D M G A A C A A C E D A constellatus B F C D G F A E E D M A A A C A A C A D A era;?!?;'! A D A M A A F B F C A A A D A B C A A A da«/ A F C D bb b G P C B g A A A D A H A A A A rfa//; B F C D bb b G 1 C B g A A A D A H A A A A da//; C F C D bb b G P c a b A A A D A H A A A A da/// D F C D bb b G P c B g A A A D A B A A A A diploproa g B C aa A A C M C C A A A C A B G F A A elongatus A W C I h A A c M P C A A A D A B I A A A elongatus B W C I h A A A M P c A A A D A B I A A A emphaeus A F C D E A A D B D c A A A B A B C A C A emphaeus B F c D E A A D B B c A A A B A B C A c A ensifer A F c D F B Q E B C c A A A C A B A A c A ensifer B F c D F B A E B C c A A A C A B C A c A entomelas F c J 1 A A E a B c A A A D A A A F B A eos F B D G B A E B C c A A A C A B C A D A exsul F C D D B A i B D M A A A C A B C A B A flavidus E C D H A C A B A c A A A A A B C B D A giUi A b C D X A P m A R N A E D D A B C B A A g//// B b C D X A P K A R N A E D D A B C B A A ^oode/ Y C D P A O a B D D A A A D A I C A A A helvoinaculatus A F C D K B A A B C C A A A C A B B A D A conti nued 186 Fishery Bulletin 104(2) Table 1 (continued) Species ND3/ND4 12S/16S Bs^N I Bs/UI Cfo I Ddel Hinill HinSl Mho I Msp I Rsa I S(v I BrfN I BstV 1 Cfo I DdelHmill Hmtl Mho I Msp 1 Rsal Sh\ helvnmactilatus B F C D K B A A B E C A A A C A B C A D A hopkinsi A Z c A n A g K B C e A A A c A B C A B A hopkinsi B Z c A n A g K A C e A A A c A B c A B A hopkinsi C Z c A u A g g B C e A A A c A A c A B A hiihbsi A L D R A I 0 A C G A C A D A A c A A A huhhui B M D S A I 0 A C G A C A D A A c A A A hiibhsi C M D T A I P A c G A c A D A B c A A A huhhsi D M D S A F 0 A c G A c A D A A c A A A inermis A F D X A A D e J I A A A A A B c A C A inermis B F H D X A A D e J I A A A A A B c A C A inermis C F I D cc A I D e J I A A A A A B c A c A jordani A a c D c E A c A p C F A A H A B c A D A jordani B a c D c E A d A p c F A A H A B c A D A jordani C a c D q E A c A p c F A A H A B c A D A joyneri A 0 H D a A A D e L A A A A A A B c A C A joyneri B O H D a A I D e L A A A A A A B c A C A /(?;!/(gnosus F C D F B A n B D M A A A C A B c A B A /et'i's B A cr to d A A q d B C A A A D A B c A D A mafrfona/c/( P b A e A b c g D C B A A G A G c A A D »!a/;g^(?/- F B D L C D c C B A A A A D A B A A B A melanops A D C D F A B E A A C A A A D A B c B D A melanops B D C D F A B E B A C A A A D A B c B D A me/anostomou.s A 0 C D g A A K A 0 C A A A D A B c B C A me/a?ioston!0!/.s B 0 c D g A A K A B C A A A D A B c B C A miniatus A F c K r C A e B A C A A A C A A c A A A miniatus B F c K r c N e B A C A A A C A A c A A A mystinus A F c J A A E a B C A A A D A A A B B A mystinus B B c J A C E D B c A A A D A A A A B A mystinus C F c J A A E a B e A A A D A A A B B A mystinus D B c J A A E D B C A A A D A A A B B A mystinus E B c J A C E D B C A A A D A A A B B A nebulosus A F G A Z A G C C B A A A A D A B A A A A nehulosus B F G A Z B G C C B A A A A D A B A A A A n!groc(/ic c c a> 5 0-001 ^ 0 4 8 12 16 20 24 Time of day Figure 3 Least squares estimates of the mean number of fish eggs per stomach in relation to time of day. Error bars represent 1 standard error. swimming speed of capelin would be very low (si cm/s) or that fish eggs are not strongly selected as prey dur- ing feeding. If on the other hand we consider modal capelin (0.13 m) swimming at 0.5-1 body length/s, the reactive distance would range from 0.024-0.034 m at the time of peak foraging activity, assuming 100% probability of attack within the reactive area. To de- termine the effect of variations in the density of eggs encountered by capelin, I also calculated the variability 210 Fishery Bulletin 104(2) 10 01 0.01 0.001 8 12 16 Time of day (h) 20 24 Figure 4 Estimated effective volume swept (A'), with ±1 standard error (solid symbols and thick error bars) based on mean number of eggs per stomach and mean density of eggs from plankton collections, in relation to time of day (offset by +1 hour from midpoint of interval). Confidence intervals are based on the maxi- mum likelihood estimates of mean numbers of eggs per capelin stomach. Box plots show the 25"^, 50"', and 75"' percentiles of the estimated effective volume swept (K) based on the mean number of eggs per stomach and each estimate of egg density based on plankton collections; whiskers show the 10"' and 90''' percentiles of the distribution and open symbols show the outlying points. in estimated K using each observation of egg density from the plankton samples. This calculation generally resulted in less variability in the estimate of effective volume swept than the uncertainty in eggs per stomach (Fig. 4, box plots) because of the skewed distribution of egg densities based on plankton tows. The latter reflects the skewed distribution of egg densities based on plankton tows. Discussion Capelin prey composition observed in this study is con- sistent with that observed in previous studies, which show that copepods, euphausiids, and amphipods repre- sent the principal prey items; the overall results of this study may therefore have general applicability. Fish eggs represent a relatively minor part of the diet of capelin. Neither Huse and Toresen (1996) nor Assthorsson and Gislason (1997) reported any occurrence of pelagic fish eggs in the stomachs of capelin from the Barents Sea and waters north of Iceland. O'Driscoll et al. (2001) reported an occurrence of less than 1% in capelin sam- pled during May-June in inshore waters and during August-September in offshore areas off Newfoundland. In the present study, an occurrence of 4*7^ to 8%, which contributed an average of 1% of prey by numbers, was found for collections in coastal Newfoundland waters during July. There is little basis for explaining the differences between studies because only this analysis provides an estimate of the density of eggs in the water column where capelin were sampled. Fish eggs were approximately 1% of the prey found in the stomachs of capelin but they represented slightly more than 0.1% of the organisms sampled by the plankton nets used in this study. Huse and Toresen (1996) collected plankton data with different gear than that used in the present study, but they did not report the occurrence of pelagic fish eggs in their samples, possibly because pelagic fish eggs were in very low abundance or were absent during the collection periods of their study. In general, fish eggs were more likely to be found in capelin stomachs as the length of individual fish in- creased. These fish tend to feed more heavily on larger prey types (e.g., euphausiids and amphipods) but the occurrence of fish eggs in individual fish tended to in- crease as the relative stomach fullness decreased. When considered with the general pattern in feeding periodic- ity, fish eggs were more likely to be eaten at the onset of feeding activity shortly after dusk, but as individual fish filled their guts with larger or more numerous prey, feeding on fish eggs tended to decrease. Fish eggs may represent an attractive prey for capelin capable of feed- ing on larger prey types (e.g., stage V of C. finmarchi- cus, euphausiids, and amphipods) because their weight is comparable to that of large copepods (Darbyson et al., 2003). However, as larger prey types are encountered or as hunger decreases, capelin may become more selective feeders and focus their foraging on energetically more profitable prey. The diurnal pattern of occurrence of fish eggs in capelin stomachs may also reflect differences in the visibility of eggs as a result of changing light levels, as well as predator hunger. Alternatively, the greater number of eggs in stomachs at dusk may reflect increased spatial overlap between predator and prey as capelin move toward the surface in the evening to feed (O'Driscoll et al.^). The effective volume swept by capelin estimated in the present study indicates that the probability that a capelin will attempt to eat a pelagic fish egg is low, because it is highly unlikely that fish will swim at a very slow speed based on an assumed reactive dis- tance. The predation rates based on average densi- ties and encounter rates range from 0.04 to 0.86 egg/h 'O'Driscoll R. L., J. T. Anderson, and F. K. Mow- bray. 2001b. Abundance and distribution of capelin from an acoustic survey in conjunction with the 1999 pelagic juvenile survey. Capelin in SA2 + Div. 3KL during 1999. Canadian Science Advisory Secretariat Research Document 2001/161. [Available from http://www.dfo-mpo.gc.ca/csas/ or Canadian Science Advisory Secretariat, Department of Fisheries and Oceans, 200 Kent Street, stn. 12032, Ottawa, Ontario, Canada KIA 0E6.] Pepin: The encounter rate of Mallotus villosus witli fish eggs 211 based on an evacuation rate of 5Q'7< of eggs per hour. A low predation rate on fish eggs is not entirely unex- pected. The eggs of cunner and cod from our samples had an average diameter of 0.86 and 1.3 mm, whereas the modal length of capelin from our collections was approximately 130 mm, making the eggs -1% of the predator's body length. Based on Paradis et al.'s (1996) analysis, maximum predation rates by fish feeding on ichthyoplankton occur when the prey are 10% of the predator's length, and the minimum predation rates were observed when prey were -0.5-1% of the predator's length. Paradis et al.'s (1996) summary equations indi- cate that an average predation rate of 0.007-0.03 eggs/ h would not be inconsistent with laboratory estimates of predation rates for a range of evacuation times from 1 to 4 hours and would be considerably lower than the predation rates estimated in the present study. How- ever, estimates based on a simple foraging model (e.g., Paradis et al., 1999) for a modal capelin swimming at 1 body length/s and with a reactive distance of 1 body length yield a mean of 12 eggs/h, and halving these parameter values yields a predation rate of 1.6 eggs/h. These alternative methods of estimating predation rates indicate that capelin are somewhat more likely to feed on fish eggs in the field than laboratory studies would indicate but less likely than would be anticipated from a simple foraging model. The results from this study do not provide an exhaustive estimate of encounter rates between capelin and pelagic fish eggs, but the range of effective encounter rates derived from the different approaches highlights the lack of understanding of this key parameter essential for any model of multispecies interactions. Such a broad range of expected predation (and encounter) rates indicates that in order to develop predictive skills in estimating the potential impact of planktivores on fish eggs, a better understanding of the factors that prompt a predator to feed on a fish egg is needed. Most laboratory studies dealing with predation by fish on fish eggs and larvae have used single prey and few have provided a measure of the probability of attack (Bailey and Houde, 1989; Paradis et al., 1996). Feeding patterns can be affected by the complexity of the available prey community (Kean-Howie et al., 1988; Gotceitas and Brown, 1993; Pepin and Shears, 1995), but methods to relate prey consumption with prey avail- ability in the field are still limited. Uncertainty in encounter and predation rates are due partly to errors in the estimation of prey availability and stomach content and partly due to our general lack of knowledge concerning the evacuation rate of fish eggs from predator stomachs. For this study I relied on Hunter and Kimbrell's (1980) observations of evacuation rates for northern anchovy from experiments conducted at 15°C. Water temperatures in Trinity Bay in July 2000 showed considerable variation with depth, rang- ing from ~13°C at the surface to -1°C at 100 m, with the overall average temperature over the water column being ~3-4°C. Given that most metabolic processes increase by 1.3-2 times over ten degrees Celsius (e.g., Brett and Groves, 1979), evacuation rates in the pres- ent study could have been half of those measured by Hunter and Kimbrell (1980) and thus would imply that our observations of stomach contents could represent 70% of the eggs consumed in the last hour rather than the 50% assumed in the analysis. The result would be that estimates of volumes swept by capelin would have to be reduced by -30%, further reducing their potential impact on ichthyoplankton populations. The general un- certainty in our knowledge of evacuation rates for fish eggs from the stomachs of capelin, and from most other planktivorous fish, therefore limits the inferences that can be derived about their impact on ichthyoplankton survival. A second source of uncertainty is due to the inher- ent sampling variability associated with the study of elements that form a small fraction of the prey of a predator. This study was based on more than 1000 specimens from a small region in order to provide the greatest accuracy possible within the confines of the study area. Stomach sampling is often restricted to fewer specimens sampled over a much broader geo- graphic range. Without a large number of observations from a local environmental setting (i.e., site), estimates of mean numbers of rare prey types would likely be unable to reflect the effect of site-specific differences in environmental conditions on predator feeding and would thereby increase the uncertainty around the estimated effective encounter rates. The situation is well illustrated by the doubling in the frequency of occurrence of fish eggs in the specimens that were ran- domly selected for complete gut analysis versus all the specimens collected for analysis. Further uncertainty comes from the variability in estimates of egg density encountered by capelin — a variability that tended to result in slightly tighter confidence intervals in esti- mates of effective volume swept than the uncertainty due to variability in the number of prey per predator stomach. Variations in egg densities from plankton samples did result in lower estimates of effective vol- ume swept, partly as a result of the rare high densities (typical of highly skewed distribution) that characterize the variability in plankton catches (Power and Moser, 1999). Although this study is not intended to provide a general estimate of the effective encounter rate of capelin feeding on fish eggs, the approach provides a useful example of the sampling requirements for the study of feeding on prey that are a minor part of the diet of a predator but which may be greatly affected by the predator's impact on the prey population. Greater efforts must be directed at understanding patterns of predation in the field, in relation to prey availability, if the role of planktivorous fish in ichthyo- plankton dynamics is to be understood. The assump- tions in the estimation approach could affect the effec- tive volume swept: estimates of the effective volume swept are based on average prey densities integrated over the water column, and the simplification of the encounter model (Eq. 1) incorporates possible variations in the probability of attack (e.g., due to diurnal varia- tions in visibility of eggs) into the value of K. Pepin et 212 Fishery Bulletin 104(2) al. (2005) noted that in the study region, pelagic fish eggs are generally more abundant near the surface and decrease exponentially in abundance with increasing depth. As a result, if feeding by capelin occurs predomi- nantly in surface waters, the effective density of prey encountered would be greater than the average density over the water column, which would in turn decrease the estimate of the effective volume swept based on our estimates of encounter rates. However, the pat- terns of vertical distribution of both eggs (Pepin et al., 2005) and capelin (O'Driscoll et al., 2001) are highly variable among sites and without accurate estimates of these elements from the current study, inferences about the impact on estimates of the effective volume swept would be speculative. In contrast, I implicitly incorporated the probability of attack into the estimate of the effective volume search, essentially giving it a value of 100%. However, Wieland and Koster (1996) found that the visibility of eggs increased with devel- opmental stage and hence may alter the probability of attack by a predator. Lower visibility of early stage eggs could lead to an increase in the estimate of the effective volume swept, but because we could not stage most eggs sampled from the stomachs, the net effect on the estimated volume swept is unclear. Planktivorous fish represent important forage for piscivores in several marine ecosystems but they can also play a significant role as predators of early life stages. In the Baltic Sea, spratt iSprattus sprattus L.) feed extensively on cod eggs (Koster and MoUmann, 2000) and they may play an important role in regulat- ing the underlying form of the stock-recruitment rela- tionship (Koster et al., 2001). In upwelling systems, anchovies and sardines may also have a significant impact on egg and larval mortality rates (Smith et al., 1989). It has also been suggested that herring (C harengus L.) and mackerel (Scomber scombrus L.) predation on ichtyplankton may play an important role in fish population dynamics in the Gulf of St. Lawrence (Swain and Sinclair, 2000) and on Georges Bank (Garrison et al., 2002). However it is unclear whether capelin play a significant role in the fish community dynamics in the ecosystems they inhabit, despite evidence that they are likely to have a signifi- cant impact on zooplankton abundance (Vesin et al., 1981; Akenhead et al., 1982; Skjoldal and Rey, 1989; Hassel et al., 1991). Clupeids, scombrids and engrau- lids can alternate between filter- and particulate-feed- ing modes, which may allow them to more effectively consume relatively small prey (~1 mm), such as fish eggs. There is no information on the feeding modes of capelin, but another osmerid, rainbow smelt (Osmerus mordax, Mitchill 1814), is known to be a primarily particulate feeder (Mills et al., 1995). Particulate feed- ing is generally directed toward larger prey, whereas filter feeding is used when prey are smaller or more abundant. Gill raker structure does not provide signifi- cant insight into the possible feeding modes of these various planktivorous fish. The estimate of reactive distance from the present study (0.024-0.034 m) is of the same order as the gape width of capelin (3.6% for TL; Pepin, unpubl. data), making filter feeding a distinctly possible mode of feeding. For known filter feeders, Uotani (1985) reported a maximum gill raker length of 6 mm in Engraiilis japonicus (100 mm TL), Gibson (1988) reported an average gill raker length of 4 mm in Clupea harengus (150 mm TL), and Molina et al. (1996) reported gill raker lengths ranging from 5 to 9 mm in Scomber japoiucus (150-250 mm SL). On the other hand, Lecomte and Dodson (2004) reported gill raker lengths of 3.3-3.9 mm in Osmerus mordctx (150 mm TL). In the case of capelin, I found that gill rak- ers ranged from 3.0-4.5 mm in length (130-170 mm TL) (Pepin, unpubl. data). Inter-raker gaps among all these species ranges from 0.19 mm (£. japonicus) to 0.4 mm (S. Japonicus) and from 0.36 to 0.45 mm in capelin (Pepin, unpubl. data). Thus all these spe- cies, which feed extensively on zooplankton but which use different ingestion modes, are equipped with gill rakers designed to retain small particles. To better understand the role of capelin in Arctic ecosystems, a greater knowledge of feeding will be required. If filter feeding is a dominant feeding mode, a simple filtra- tion or volume-swept model may provide an accurate measure of the potential impact of capelin on fish eggs, whereas greater knowledge of behavior and selectivity will be required if particulate feeding is the dominant feeding mode because selection for or against fish eggs could be influenced by availability of alternate prey (e.g., Kean-Howie et al., 19881. In most of the areas where planktivores can have a significant impact on the mortality of fish eggs, the ecosystems can be characterized as being "wasp-waist- ed" (Cury et al., 2000), i.e, as having a relatively high diversity of plankton species and top predators that are all influenced by the abundance and productivity of relatively few and dominant forage fish species that act to transfer energy from secondary producers to up- per trophic levels in a manner similar to that found in many Arctic regions. It is clear that capelin are important prey for a number of top predators in the waters of the Barents Sea (Gjosaster, 1998), Iceland (Vilhjalmsson, 2002), and on the Newfoundland and Labrador coast (Akenhead et al., 1982). and they may also have an important impact on the plankton commu- nity (Skjoldal and Rey, 1989; Hassel et al., 1991). What remains uncertain is whether they play an important regulatory role in the dynamics of the early life stages of fish. Observations from this study revealed that the highest occurrence of fish eggs in the stomachs of cap- elin came from capelin near one of the few remaining major spawning aggregations of the northern cod stock (Rose, 2003). The dynamics of such circular prey-preda- tor interactions need to be studied further in order to understand their potential importance to population regulation (Walters and Kitchell, 2001) but to do so will require consideration of the spatial and temporal patterns of predator-prey interactions if predictive re- lationships are to be derived (Pepin et al., 2002, 2003; Baumann et al., 2003). Pepin The encounter rate of Mallotus villosus with fish eggs 213 Acknowledgments I would like to thank Tim Shears, Gary Maillet, Hugh Maclean, and Greg Redmond for their assistance in the field, Leah Wilson for performing the laboratory analysis of stomach contents, and the officers and crew of the CCGS Templeman for their diligence. Comments by J. Carscadden, J. Dower, G. Huse, F. Mowbray, and two anonymous referees helped to improve the manuscript. Literature cited Akenhead S. A., J. Carscadden , H. Lear, G. R. Lilly, and R. Wells. 1982. Cod-capelin interactions off Northeast New- foundland and Labrador. Can. Spec. Pub. Fish. Aquat. Sci. 56:141-148. Arrhenius F., and S. Hansson. 1994. In situ food consumption of young-of-the-year Baltic Sea herring Clupea harengus: a test of predic- tions from a bioenergetic model. Mar. Ecol. Prog. Ser. 11:145-149. Assthorsson O. S., and A. Gislason. 1997. On the food of capelin in the subsarctic waters north of Iceland. Sarsia 82:81-86. 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. Baumann H., P. Pepin, F. J. M. Davidson, F. Mowbray, D. Schnack, and J. F. Dower. 2003. Reconstruction of environmental histories to investigate patterns of larval radiated iUlvaria sub- bifurcata) growth and selective survival in a large bay of Newfoundland. ICES J. Mar. Sci. 60:243-258. Brett J. R., and T. D. D. Groves. 1979. Physiological energetics. In Fish physiology, vol. 8 (W. S. Hoar, D. J. Randall, and J. R. Brett, eds.), p. 279-352. Academic Press, New York, NY. Cowan J. H., K. A. Rose. E. D. Houde, S. B. Wang, and J. Young. 1999. Modeling effects of increased larval mortality on bay anchovy population dynamics in the mesohaline Chesapeake Bay: evidence for compensatory reserve. Mar. Ecol. Prog. Ser. 185:134-146 Cury P., A. Bakun, R. J. Crawford, A. Jarre, R. A. Quinones, L. J. Shannon, and H. M. Verheye. 2000. Small pelagics in upwelling systems: patterns of interraction and structural changes in "wasp-waist" ecosystems. ICES J. Mar. Sci 57:603-618. Darbyson, E., D. P. Swain, D. Chabot, and M. Castonguay. 2003. Diel variation in feeding rate and prey consumption of Atlantic herring iClupea harengun L.) and Atlantic mackerel (Scomber scomrbus L.) in the southern Gulf of St. Lawrence. J. Fish Biol. 63:1235-1257. Folkvord, A. 1993. Prey recognition in stomachs of cannibalistic juve- nile cod iGcidus morhua L.). Sarsia 78: 97-100. Garrison, L. P., W. Mieheals, J. S. Link, and M. J. Fogarty. 2002. Spatial distribution and overlap between ichthyo- plankton and pelagic fish and squids on the southern flank of Georges Bank. Fish. Oceanogr. 11:267-285. Gerritsen, J., and J. R. Strickler. 1977. Encounter probabilities and community structure in zooplankton: a mathematical model. J, Fish. Res. Board Can. 34:73-82. Gibson, R. N. 1988. Development, morphometry and particle retention capability of the gill rakes in herring, Clupea harengusi L. J. Fish. Biol. 32:949-962. GjosEBter, H. 1998. The population biology and exploitation of cap- elin (Mallotus villoaus^ in the Barents Sea. Sarsia 83: 453-496. Gotceitas, V., and J. A. Brown. 1993. Risk of predation to fish larvae in the presence of alternative prey: effects of prey size and number. Mar. Ecol. Prog. Ser. 98:215-222. Hassel, A., H. R. Skjoldal, H. Gjosseter, H. Leong, and L. Omli. 1991. Impact of grazing from capelin (Mallotus villosus) on zooplankton: a case study in the northern Barents Sea in August 1985. Polar Res. 10:371-388. Hunter, J. R., and C. A. Kimbrell. 1980. Egg cannibalism in the northern anchovy, Eit- graulis mordcix. Fish. Bull. 78:811-816. Huse, G., and R. Toresen. 1996. A comparative study of the feeding habits of herring (Clupea liarengus, clupeidae, L.) and capelin (Mallotus villosus. Osmeridae, MiiUer) in the Barents Sea. Sarsia 81:143-153. 2000. Juvenile herring prey on Barents Sea capelin larvae. Sarsia 85:385-391. Kean-Howie, J. C, S. Pearre, and L. M. Dickie. 1988. Experimental predation by sticklebacks on larval mackerel and protection of fish larvae by zoo- plankton alternative prey. J. Exp. Mar. Biol. Ecol. 124:239-259. Koster, F. W., H. H. Hinrichsen, M. A. St. John, D. Schnack, B. R. MacKenzie, J. Tomkiewicz, and M. Plikshs. 2001. Developing Baltic cod recruitment models. 2. Incorporation of environmental variability and species interaction. Can. J. Fish. Aquat. Sci. 58:1534-1556. Koster, F. W., and C. Mollmann. 2000. Trophodynamic control by clupeid predators on recruitment success in Baltic cod? ICES J. Mar. Sci 57: 310-323. Lecomte, F., and J. J. Dodson. 2004. Role of early life-history constraints and resource polymorphism in the segregation of sympatric populations of an estuarine fish. Evol. Ecol. Res. 6:631-658. Mills, E. L., R. O'Gorman, R. F. Roseman, C. Adams, and R. W. Owens. 1995. Planktivory by alewife (Alosa pseudoharengus) and ranbow smelt (Osmerus mordax) on microcrustacan zooplankton and dreissenid (Bivalvia: Dreissenidae) veligers in southern Lake Ontario. Can. J. Fish. Aquat. Sci. 52:925-935. Molina, R. E., Manrique, F. A. and H. E. Velasco. 1996. Filtering apparatus and feeding of the Pacific mack- erel (Scomber japonicus) in the Gulf of California. Rep. Cal. Coop. Ocean. Fish. Invest. 37:251-156. Munk, P. 2002. Larval sand lance (Ammodytes sp.) in the diet of small juvenile wolffish (Anarhiclias spp.): preda- tory interactions in frontal water masses off western Greenland. Can. J. Fish. Aquat. Sci. 59:1759-1767. O'Driscoll, R. L., M. J. D. Parsons, and G. A. Rose. 2001. Feeding of capelin (Mallotus villosus) in Newfound- land waters. Sarsia 86:165-176. Paradis, A. R., and P. Pepin. 2001. Modelling changes in the frequency distributions of 214 Fishery Bulletin 104(2) fish larvae using field estimates of predator abundance and size distributions. Fish. Oceanogr. 10:217-234. Paradis, A. R., P. Pepin, and J. A. Brown. 1996. Vulnerability to predation offish eggs and larvae: a review of the influence of the relative size of prey and predator. Can. J. Fish. Aquat. Sci. 53:1226-1235. Paradis, A. R., M. Pepin, and P. Pepin. 1999. An individual-based model of the size-selective vulnerability offish larvae to four major predator types: disentangling the effect of size-dependent encounter rates and susceptibility to predation. Can. J. Fish. Aquat. Sci. 56:1562-1575. Pepin, P., and T. H. Shears. 1995. The influence of body size and alternate prey abun- dance on the risk to predation of fish larvae. Mar. Ecol. Prog. Ser. 128:279-285. Pepin, P., J. F. Dower, J. A. Helbig, and W. C. Leggett. 2002. Regional differences in mortality rates of a larval fish, the radiated shanny iUlvaria subbifurcata): estimat- ing the relative roles of dispersal and predation. Can. J. Fish. Aquat. Sci. 59:105-114. Pepin, P., J. F. Dower, and F. J. M. Davidson. 2003. A spatially-explicit study of prey-predator interac- tions in larval fish: assessing the influence of food and predator abundance on growth and survival. Fish. Oceanogr. 12:19-33. Pepin, P., P. V. R. Snelgrove, and K. P. Carter. 2005. Accuracy and precision of the continuous underway fish egg sampler iCUFES) and bongo nets: a comparison of three species of temperate fish. Fish. Oceanogr. 14:432-447. Power, J. H., and E. B. Moser. 1999. Linear model analysis of net catch data using the negative binomial distribution. Can. J. Fish. Aquat. Sci. 56:191-200. Rose, G. A. 2003. Monitoring coastal northern cod: towards an opti- mal survey of Smith Sound, Newfoundland. ICES J. Mar. Sci 60:453-462. Skjoldal, H. R., and F. Rey. 1989. Pelagic production and variability of the Barents Sea ecosystem. In Biomass yields and geography of large marine ecosystems (K. Sherman and L. M. Alex- ander, eds.l, p. 241-286. Selected Symposia Series 111, American Association for the Advancement of Science. Westview Press Inc., Boulder, CO. Smith, P. E., H. Santander, and J. Alheit. 1989. Comparison of the mortality rates of Pacific sar- dine, Sardinops saga.x, and Peruvian anchovy, Engraulis ringens. eggs off Peru. Fish. Bull. 87:497-508. Swain, D. P., and A. F. Sinclair. 2000. Pelagic fishes and the cod recruitment dilemma in the northwest Atlantic. Can. J. Fish. Aquat. Sci. 57:1321-1325. Temming, A.. B. Bohle. D. W. Skagen, and F. R. Knudsen. 2002. Gastric evacuation in mackerel: the effect of meal size, prey type and temperature. J. Fish Biol. 61:50-70. Uotani, I. 1985. The relation between the development of feeding organs and feedings modes of the anchovy. Bull. Jap. Soc. Sci. Fish. 51:197-204. Utne-Palm, A. C. 2000. Prey visibility, activity, size and catchabilitiy's (evasiveness) influence on Gobisculus flavescens prey choice. Sarsia 85:157-165. Vesin, J. P., W. C. Leggett. and K. W. Able. 1981. Feeding ecology of capelin [Mallotus villosus) in the estuary and western Gulf of St. Lawrence and its multispecies implications. Can. J. Fish. Aquat. Sci. 38:257-267. Vilhjalmsson, H. 2002. Capelin {Mallotus villosus) in the Iceland-East Greenland-Jan Mayen ecosystem. ICES J. Mar. Sci 59: 870-883. Walters, C, and J. F. Kitchell. 2001. Cultivation/depensation effects on juvenile sur- vival and recruitment: implications for the theory of fishing. Can. J. Fish. Aquat. Sci. 58:39-50. Wieland, K., and F. W. Koster. 1996. Size and visibility of Baltic cod eggs with refer- ence to size-selective and stage-dependent predation mortality. J. Appl. Ichthyol. 12:83-89. Werner, F. E., J. A. Quinlan, R. G. Lough, and D. R. Lynch. 2001. Spatially-explicit individual based modeling of marine populations: a review of the advances in the 1990s. Sarsia 86:411-421. 215 Abstract — Horseshoe crab {Liiuuhis polyphemus) is harvested commer- cially, used by the biomedical indus- try, and provides food for migrating shorebirds, particularly in Delaware Bay. Recently, decreasing crab popula- tion trends in this region have raised concerns that the stock may be insuf- ficient to fulfill the needs of these diverse user groups. To assess the Del- aware Bay horseshoe crab population, we used surplus production models (programmed in ASPIC), which incor- porated data from fishery-independent surveys, fishery-dependent catch- per-unit-of-effort data, and regional harvest. Results showed a depleted population (B._,|,„3/B;^,gY=0.03-0.71 ) and high relative fishing mortality <^2002'^MSV=0-9-9-5). Future harvest strategies for a 15-year period were evaluated by using population projec- tions with ASPICP software. Under 2003 harvest levels (1356 t), popula- tion recovery to B,,,,.,- would take at least four years, and four of the seven models predicted that the population would not reach Bi/sy within the 15- year period. Production models for horseshoe crab assessment provided management benchmarks for a spe- cies with limited data and no prior stock assessment. A production modeling approach to the assessment of the horseshoe crab iUmulus polyphemus) population in Delaware Bay Michelle L. Davis Department of Fisheries and Wildlife Sciences Virginia Polytechnic Institute and State University 210 Cheatham Hall Blacksburg, Virginia 24061-0321 Email address midavistSvtedu Jim Berkson National Marine Fisheries Service RTR Unit at Virginia Tech 100 Cheatham Hall Blacksburg, Virginia 24061-0321 Marcella Kelly Department of Fisheries and Wildlife Sciences Virginia Polytechnic Institute and State University 210 Cheatham Hall Blacksburg, Virginia 24061-0321 Manuscript submitted 25 January 2005 to the Scientific Editor's Office. Manuscript approved for publication 10 August 2005 by the Scientific Editor. Fish. Bull. 104:215-225 (2006). The horseshoe crab {Limulus poly- phemus) has become a source of con- troversy on the Atlantic coast of the United States (Berkson and Shuster, 1999; Walls et al., 2002). This species is commercially harvested for use as bait, is used by the biomedical indus- try, and is an important source of food for a large number of species, includ- ing migrating shorebirds. However, population trends in the Delaware Bay region in recent years have indicated a possible decline in horseshoe crab abundance, raising concerns that the population may be unable to fulfill the present and future needs of these diverse user groups. Horseshoe crabs are harvested commercially for use as bait in the American eel (Anguilla rostrata), whelk, and conch (family Melongeni- dae) pot fisheries (ASMFC). Histori- cally, horseshoe crabs were considered "trash fish" of little commercial value and were used primarily as fertilizer or animal feed. When the bait fish- ery began, there were few restrictions on harvest and no harvest-reporting requirements. A maximum reported coastwide harvest of about 2 million crabs (3100 metric tons [t]) occurred in 1998 (ASMFC2). Commercial har- vest has decreased in recent years owing to the adoption of state-by- state quotas in 2000 (ASMFC'') and the increased use of bait-saving de- vices for the eel and conch fisheries, both of which have reduced the de- mand for crabs. Horseshoe crabs are also used by the biomedical industry. The blood of horseshoe crabs contains Limulus Amebocyte Lysate (LAL), a substance used to detect the presence of endo- toxin contamination in injectable and implantable drugs and devices (No- vitsky, 1984). The Food and Drug Ad- ministration estimated that 260,000 horseshoe crabs were bled for LAL in 1997. After bleeding, the animals were released at the capture site, and 1 ASMFC (Atlantic States Marine Fisher- ies Commission). 1998. Interstate fish- ery management plan for horseshoe crab, 57 p. ASMFC, 1444 Eye Street, NW, Sixth Floor, Washington, DC 20005. - ASMFC. 2004. Horseshoe crab 2004 stock assessment report, 87 p. ASMFC, 1444 Eye Street, NW, Sixth Floor, Wash- ington, DC 20005. ■' ASMFC. 2000. Addendum I to the fish- ery management plan for horseshoe crab, 9 p. ASMFC, 1444 Eye Street, NW, Sixth Floor, Washington, DC 20005. 216 Fishery Bulletin 104(2) the mortality from the bleeding process was estimated to be 7.5% (Walls and Berkson, 2003). Currently, there is no substitute for LAL that offers comparable speed and sensitivity. Horseshoe crabs also play an important role in ma- rine and terrestrial food webs (Botton and Shuster, 2003). Shorebirds migrating from South America to Arctic breeding grounds stop in the Delaware Bay to re- build depleted energy reserves (Botton et al., 1994). The time and place of their stop-over coincides with that of annual horseshoe crab breeding, when the crabs arrive en masse to spawn on sandy beaches during high tides of May and June (Botton and Harrington, 2003). An adult female horseshoe crab lays approximately 88,000 eggs per year (Shuster, 1982), and a single red knot (Calidris canutus) can consume an estimated 18,000 crab eggs daily (USFWS-^). Horseshoe crabs account for substantial economic value in the Delaware Bay region. Regional economic contribution of the eel and conch fisheries is approxi- mately $2.2 to $2.8 million annually. The regional eco- nomic value of the horseshoe crab biomedical industry is $26.7 to $34.9 million annually. Ecotourism related to migrating shorebirds has become increasingly im- portant to the economy of the Delaware Bay region. An estimated 6,000 to 10,000 recreational bird watchers visit the Delaware Bay in spring and contribute $6.8 to $10.3 million to the regional economy. Recently, there has been concern that the horseshoe crab population can fulfill the current needs of these user groups. These concerns led the Atlantic States Marine Fisheries Commission (ASMFC) to develop a fishery management plan for horseshoe crabs. Unfortu- nately, very few abundance data are available for this species. Many state and federal trawl surveys record horseshoe crabs caught during sampling, but gear and sampling methods are not designed for horseshoe crabs and catches are not common. Only recently have sta- tistically robust horseshoe crab-specific surveys been initiated: a Delaware Bay spawning survey (Smith et al., 2002) and an offshore trawl survey (Hata and Berk- son, 2003). Because of the limited data available, previous stock assessments of the Delaware Bay population have been restricted to trend analyses to determine whether a single survey identifies a significant change in the pop- ulation or whether there is a consensus among data sets. However, many Delaware Bay surveys have high variability and low power to detect population change and would therefore only be able to identify dramatic changes in population size. Also, trend analyses do not provide estimates of stock status (Caddy, 1998) such as relative biomass (B/Syt,.y) and relative fishing mortality iF/fusY^- 'I'^^ ultimate goal for horseshoe crab assess- ment employs a stage-based catch-survey methodology (Collie and Sissenwine, 1983; HSC-SAS^), incorporating data from harvest and surveys. It will be a number of years before this modeling approach can be implement- ed, however, since stage-class data from commercial harvest are not currently being collected. The surplus production modeling approach (Prager, 1994) used in our study is an appropriate bridge be- tween these two methods. Production models allow for the incorporation of harvest data and multiple surveys, improving predictive power over that of a single survey. This technique does not include a stage-structure in the model; instead it focuses on the dynamics of the popula- tion as a whole. Similar methods have been successfully applied to horseshoe crab data from Rhode Island (Gib- son and Olszewski'') and production models have been widely used for assessments of other species (Booth and Punt, 1998; Cadrin and Hatfield, 2002; Vaughan and Prager, 2002). Surplus production models assume a low population growth rate at small population sizes and as the population nears the carrying capacity (Quinn and Deriso, 1999). In the logistic growth form of the model used in the present study, maximum growth rate (and maximum surplus production) occurs at one-half of the carrying capacity. At this point, the maximum surplus population growth can be harvested while still main- taining a stable population size. Surplus production models provide estimates of maximum sustainable yield (MSY; the largest harvest that can continuously be removed from a stock), population biomass, and fishing mortality, as well as allow for the estimation of effects of future management. We fitted a regional-scale production model to Dela- ware Bay horseshoe crab data in order to quantify the current status of the Delaware Bay population and to estimate impacts of future management actions. The results from this production model will allow the ASMFC and member states to manage the Delaware Bay population of horseshoe crabs more effectively with the goal of providing a sustainable resource for com- mercial harvest, the biomedical industry, and migrating shorebirds. Methods Production model We used an age-aggregated production model with the Prager (1994) form of the Graham-Schaefer surplus-pro- duction model (i.e., logistic population growth). ^ USFWS( U.S. Fish and Wildlife Service). 2003. Delaware Bay shorebird-horseshoe crab assessment report and peer review. Migratory Bird Publication R9-03/02, 107 p. USFWS, 4401 N. Fairfax Dr., MBSP 4107, Arlington, VA 22203. ^ HSC-SAS (Horseshoe crab stock assessment subcom- mittee). 2000. A conceptual framework for the assess- ment of horseshoe crab stocks in the mid-Atlantic region, 19 p. ASMFC, 1444 Eye Street, NW, Sixth Floor, Wash- ington, DC 20005. ^ Gibson, M., and S. Olszewski. 2001. Stock status of horse- shoe crabs in Rhode Island in 2000 with recommendations for management, 13 p. RI Division of Fish and Wildlife, 4808 Tower Hill Rd., Wakefield, RI 02879. Davis et al Population assessment of Limulus polyphemus 217 3.00 200 1.00 0 00 rn -100 -2 00 MD coastal bays DE 16-tt trawl, <160-mm crabs DE 16-ft YOY DE 30-ft trawl NJ ocean NMFS spring NMFSfall DB spawning 1 1 1 1 r 1991 1993 1995 1997 Year 1999 2001 2003 Figure 1 Fishery-independent survey indices for the Delaware Bay (DBi region from 1991 through 2003, standardized by survey for comparison. MD = Maryland; DE = Delaware: NJ = New Jersey; DB = Delaware Bay. dB, rB' . ' r-F,)B, d, K where /• = the stock's intrinsic growth rate; K = the carrying capacity, both of which are assumed to be constant (Prager, 19941; and F, and B, = Fishing mortality and biomass, respec- tively, at time t. In addition, the harvest was not assumed to equal sur- plus production (i.e., the model was a dynamic or non- equilibrium model; Quinn and Deriso, 1999). This model assumes B^jy = 0.5K, where By^jy is the spawning biomass that would produce MSY. This form is often used because of its theoretical simplicity and because it is central among possible production model shapes. This production model was conditioned on catch, mean- ing that landings data were assumed to be more precise than abundance indices. By assuming that abundance indices are correlated measures of population abun- dance, the model is able to incorporate multiple indices by interpreting differences among indices as sampling error. To fit the production model, we used the ASPIC software (vers. 5.02) of Prager (1994), a program that has been used extensively in stock assessments (Cadrin and Hatfield, 2002; MacCall, 2002). We included data from fishery-independent and fishery-dependent sources in model runs. Abundance indices In the Delaware Bay region, there are a number of fish- ery-independent surveys that collect data on horseshoe crabs (Fig. 1). These include National Marine Fisheries Service (NMFS) trawl (spring, 1968-2003, and fall, 1963-2002), New Jersey (NJ) ocean trawl (1989-2002), Maryland (MD) coastal bays trawl (1988-2002), Dela- ware (DE) 16-ft trawl (juvenile and young-of-the-year; 1992-2002), DE 30-ft trawl (1990-2002), and Delaware Bay spawning survey (1999-2003). Detailed descriptions of these surveys can be found in ASMFC.^ We selected 1991-2003 as the modeling timeframe because both harvest data and abundance index data were available for this period. In spite of the large number of surveys and long time series for some of these surveys, many had high vari- ability and low power to detect a decline. Additionally, many of these surveys were negatively correlated with each other for the years investigated (Table 1). Be- cause an underlying assumption for the model is that each survey is representative of the population being evaluated, total disagreement (i.e., negative correla- tion) among any pair of surveys cannot be reconciled by the model, resulting in model errors. We therefore used three subsets of fishery-independent surveys in which all pairs were positively correlated for popula- tion modeling (Table 2), incorporating six of the eight fishery-independent surveys into production model runs. Fishery-dependent data were also available from 1991 to 2002 from the Delaware hand and dredge fisheries. Abundance indices based on catch-per-unit-of-effort (CPUE) for these fisheries were calculated from the annual number of trips and landings for each fishery (Fig. 2). Within the models, abundance indices were weighted by the inverse of the coefficient of variation (CV) from regressions, which gave more weight to surveys with less variability. For comparison, we also conducted model 218 Fishery Bulletin 104(2) runs where surveys were weighted equally within the models. Table 2 lists the three fishery-independent mod- els (referred to as FI) and the four fishery-dependent models (referred to as FD) used in production model applications and the CV and weighting of each survey within the models. a. o 30 25 20- 1 5- 1 0 05 0.0 -0.5- -1 0 -1 5 Hand harvest CPUE Dredge harvest CPUE 1991 1993 1995 1997 1999 2001 Year Figure 2 Fishery-dependent abundance indices based on catch-per-unit- of-effort data for the Delaware hand and dredge horseshoe crab tLimiiliiti polyphemus) fisheries. 1991-2002. standardized by index for comparison. Harvest We obtained horseshoe crab harvest data for Virginia, Maryland, Delaware, Pennsylvania, and New Jersey from 1995 to 2003 (NMFS'; Michels*). We combined harvests among states for regional-scale production model runs. We estimated the regional harvest from 1991 to 1994 from available Delaware landings data (Fig. 3). Reporting of harvest was not manda- tory during this time period, and harvest data NMFS (National Marine Fisheries Service). 2004. Fisheries Statistics and Economics Divi- sion, Silver Spring, MD. Landings data (by state and year) obtained in August 2004 from http:// www.st.nmfs.gov/stl/commercial/index.html. Web page was last modified 24 September 2003. ' Michels, S. Personal commun. 2004. DE Div. of Fish and Wildlife, 89 Kings Highway, Dover, DE 19901. Table 1 Correlation matrix for abundance indices derived from 10 surveys, with number of pairwise theses for 1991-2003. Negatively correlated indices are shown in bold font. Indices used in pi with an asterisk (*). comparisons (i.e., years) in paren- oduction model runs are identified 4 5 1 NMFS fall 2 NMFS spring 3 DE 30-ft trawl DE 16-ft trawl, <160-mm crabs DE 16-ft trawl. YOY 6 N.J ocean 7 MD coast 8 9 10 Spawning DE DE survey dredge hand 1 NMFS sail 1.000 (12) 2 NMFS spring* -0.409 (12) 1.000 (13) 3 DE 30-ft* trawl 0.706 (12) -0.370 (12) 1.000 (12) 4 DE 16-ft trawl, < 160- mm' crabs -0.261 (11) 0.182 (11) 0.215 (11) 1.000 (11) 5 DE 16-ft trawl, YOY* -0.153 (10) 0.214 (10) 0.341 (10) 0.617 (10) 1.000 (10) 6 NJ ocean* -0.102 (12) -0.249 (12) 0.479 il2) 0.252 (11) -0.132 (10) 1.000 (12) 7 MD coastal bays* -0.18X (12) 0.086 (12) -0.038 (12) 0.639 (11) 0.336 (10) -0.116 (12) 1.000 (12) 8 Spawning survey -0.786 (4) 0.211 (5) -0.069 (4) -0.657 (4) -0.781 (3) 0.392 (4) -0.087 (4) 1.000 (5) 9 DE dredge harvest* 0.210 (12) -0.329 (12) 0.450 (12) 0.379 (11) 0.031 (10) 0.506 (12) 0.135 (12) 0.509 1.000 (4) (12) 10 DE hand harvest* -0.536 il2i 0.002 (12) -0.214 (12) 0.639 (11) 0.084 (10) 0.485 Il2i 0.123 il2i 0.167 0.564 1.000 111 112) (12) Davis et al Population assessment of Limulus polyphemus 219 from other states were unavailable or unreliable. We therefore expanded the Delaware harvest to represent the Delaware Bay region by making the assumption that regional landings were equal to ten times the landings in Delaware, the approxi- mate relationship from 1997 through 2003 when reporting was mandatory. When applicable, we converted numbers to metric tons (t) (Prager and Goodyear, 2001) using the relationship of Gibson and Olszewski'' (1.8182 kg/horseshoe crab). Com- mercial harvest of horseshoe crabs has been below the regional quota of 2595 t since 2000 (Fig. 3). Assumptions associated with the production models There were a number of general assumptions asso- ciated with production models (Quinn and Deriso, 1999). We assumed that productivity (change in biomass over time) responded instantaneously to changes in population size. Changes in the biotic and abiotic environments were ignored, and /• (the intrinsic rate of population growth) and K (the carrying capacity) were assumed to be constant. Because produc- tion models combine all age classes, it was assumed that size or age structure of the population would not have major effects on population dynamics. Specific assumptions about population values were also required by the model. All starting values of 4000 n Max = 3678 t 3500 ■ ^V y/^*~~A. S 3000 ■ / ^'''^^ v to Quota = 2595 1 / \ 5 2500 ■ . _ / \ " to / \ ■c 2000 - / \ ro f5 'L o 1500 ■ / ^v^,^^,^^ O) X ^tr ^^ S. 1000 ■ J 500 . 1991 1993 1995 1997 1999 2001 2003 Year Figure 3 Delaware Bay regional hor.seshoe crab iLimulus polyphemus) harvest (in t). Regional harvest, 1991-1994 (open diamonds). was estimated to be ten times the Delaware harvest. The current regional quota of 2595 t is shown by the horizontal line. MSY and K were based on the maximum harvest in the Delaware Bay from 1991 through 2003. The initial guess for MSY was 1850 t (half of the largest catch), and the initial guess for K was 37,000 t (ten times the largest catch). For fishery-independent model runs, we had the model freely estimate initial biomass in relation to carrying capacity (B^IK), with our start- Table 2 Description of production model runs, includ ng the data sources. years , and coefficient of variation (CV) from regression analyses. The weighting of surveys within the model s is the inverse of the CV. The starting estimate of B,/K fo r each model is also shown FI refers to models that include fishery- independent ab undance indices, and FD identifies models where fisherv- dependent indices from CPUE data were used. Relative Initial Model Data sources Years CV weighting Bi/K" value FI-1 NMFS spring DE 16-ft trawl, <160-mm crabs DE 16-ft trawl YOY MD coastal bays 1991-2003 1992-2002 1992-2001 1991-2002 0.636 0.517 0.928 0.515 0.241 0.296 0.165 0.297 0.5 FI-2 DE 30-ft trawl DE 16-ft trawl, <160-mm crabs DE 16-ft YOY 1991-2002 1992-2002 1992-2001 0.524 0.517 0.928 0.388 0.393 0.219 0.5 FI-3 DE 30-ft DE 16-ft trawl, <160-mm crabs NJ ocean trawl 1991-2002 1992-2002 1991-2002 0.524 0.517 0.345 0.283 0.287 0.430 0.5 FD-1 DE hand CPUE DE dredge CPUE 1991-2002 1991-2002 0.449 0.555 0.553 0.447 Fixed at 0.1 FD-2 DE hand CPUE DE dredge CPUE 1991-2002 1991-2002 0.449 0.555 0.553 0.447 Fixed at 0.2 FD-3 DE hand CPUE DE dredge CPUE 1991-2002 1991-2002 0.449 0.555 0.553 0.447 Fixed at 0.3 FD-4 DE hand CPUE DE dredge CPUE 1991-2002 1991-2002 0.449 0.555 0.553 0.447 Fixed at 0.4 220 Fishery Bulletin 104(2) ing estimate of this value equal to 0.5. In the absence of other information about starting biomass, B^IK = 0.5 (i.e., Bj=BygY' is an appropriate default value (Punt, 1990; Prager, 1994). In some cases B^/K was poorly estimated, and we employed a common solu- tion of fixing the value o{ B^IK (Vaughan and Prager, 2002). For model runs based on fishery-dependent indices, we fixed B^IK at 0.1, 0.2, 0.3, and 0.4 (Table 2). The CPUE value for the Delaware hand fishery was relatively low in 1991; therefore we assumed that the population biomass was below B^jgy. ie. B^IK was less than 0.5. lo- g- ♦ FI-2 ♦ CV s' 7 ■ J ^• O o U.~ 4- a a Equal High F gFI-3 LowB and High F 3- 2- 1 - FI-1 C>^ FI-2 P1.3 FI-1 °^»__^j»__£X Low B Ideal 0. M 0.25 0.50 0 75 100 125 ^2003/6^5^ Figure 4 Relative 2003 biomass (B2003''5msy> and relative 2002 fishing mortality (^2002^^;MSy' of Delaware Bay horseshoe crab iLimu- lus polyphemus) for each of the seven model runs presented. with abundance indices weighted equally or inversely to CV within the models. Quantities estimated by the model The production model estimated several benchmarks and status indicators useful in understanding horse- shoe crab biology and in improving management. These quantities included relative biomass (B/Sj^jgyl. relative fishing mortality (F/Fy^^y). population biomass (B), and maximum sustainable yield (MSY). Point estimates and 80% confidence intervals for each of these quantities were calculated for each model run. Population projections We used the Delaware Bay population estimates calculated by the surplus production model to project the population forward in time for a period of 15 years to evaluate potential harvest levels. We conducted projections using ASPICP (Prager, 1994), with annual landings specified for each year of the projections. We selected a 15-year time period because this is the longest projection period that can be programed in ASPICP, and confidence in projection model results decreases at longer time intervals. We evaluated trends in biomass over time for a range of harvest levels, including harvest at 2003 levels and proportional reductions of that harvest. These harvests were 0% of 2003 catch (no harvest), 25% (339 t annu- ally), 50%' (678 t), 75% (1017 t), and 100% (1356 t). We identified the number of years under each harvest scenario required for the population (and 80% confidence intervals) to rebuild to B^/^y. We also compared relative biomass (with 80%5 confi- dence intervals) in the final year of projections (2018) for each harvest level. I.b ■ B/Bmsy 1.4 - FI-1 1.2 ■ . - - ■ FI-3 10 08 06 — • — / 04 . X ^ ^ 02 - T77T — ~— -~— ___^ _ n n - ■ - - - -■-.»-,_ 1991 1993 1997 Year Figure 5 Production model estimates of relative biomass (B/B/^igy'i of horseshoe crabs (Limulus polyphemus) in the Delaware Bay region, 1991-2004. Results from fishery-independent model runs are shown, and the hori- zontal line represents B/B^,^y=l. 80% confidence intervals each model run are the same line pattern in gray. Results Results differed little between model sim- ulations for surveys weighted inversely to CV and for simulations with equally weighted surveys (Fig. 4). Production model runs showed that BIB^i^Y in the Delaware Bay region increased in the early 1990s and has declined steadily since 1995 (Figs. 5 and 6). Slight increases since 2001 were evident in some model runs. Relative biomass in 2003 was estimated to be low, with point estimates ranging from 0.03 to 0.20 for models with fishery- independent data and 0.20 to 0.71 for models with fishery-dependent data (Table 3). Eighty-percent confidence intervals for B^oos/B^gy ranged from 0.005 to 1.16. Davis et a\ Population assessment of Limulus polyphemus 221 Table 3 Status indicators and management benchmarks for horseshoe crabs in the Delaware Bay region, estimated from production model runs with fishery-independent iFIi or fishery-dependent (FDi indices weighted by the inverse of the CV. Point estimates and lower (L) and upper (U) 80"* confidence intervals (CIs) are shown for each model run. The objective function is a measure of how well the model was able to fit the data using a lognormal error structure — a lower value representing a better fit. FI-1 FI-2 FI-3 FD-1 FD-2 FD-3 FD-4 "imrx^MSY 0.232 0.022 0.030 0.201 0.427 0.588 0.710 80'* CKL) 0.109 0.005 0.008 0.127 0.238 0.319 0.388 80^f CKU) 0.479 0.154 0.041 0.347 0.762 1.015 1.157 ■'^2002''^.MSi' 2.156 9.501 6.560 1.795 1.269 1.034 0.911 BO'/rCKL) 1.442 6.005 3.999 1.258 0.770 0.588 0.513 80% CKU) 3.299 18.320 21.300 2.592 2.069 2.112 2.184 -S2003 ^^ FD-3 ._ pQ ^ 02- „•''* **------ 1991 1993 1995 1997 1999 2001 2003 Year Figure 6 Production model estimates of relative biomass iS/B,,,;..,.) of horseshoe crabs {Limulus polyphemus) in the Delaware Bay region, 1991-2003. Results from fishery-dependent model runs are shown, and the horizontal line represents B/Bn,gY=l- 80% confidence intervals for models FD-1 and FD-4 are shown in gray. to rebuild the population to B„sy varied substantially among models (Table 4). At 2003 harvest levels (i.e., 100%), projections showed population recovery in a mini- mum of four years, although four of the seven models did not reach B^^.y in the 15-year projection period. In the absence of harvest (i.e., 0%), recovery could occur in as few as 2 years, but two models did not reach B^^^y in the projection period. Estimates of B/Bj^g^y in the final year of projections (2018) also differed among model applications (Table 5). 222 Fishery Bulletin 104(2) Discussion According to surplus production model runs and corre- sponding projections, the horseshoe crab population in the Delaware Bay region has been depleted and current harvest levels may be too high to allow the population to rebuild to S^gy within 15 years. Biomass in this region has decreased steadily since 1995 and is currently well below B A/.sy This decline was evident in models that incorporated regional fishery-independent surveys and Delaware fishery-dependent indices. Figure 4 shows current relative biomass and relative fishing mortality for each of the model applications. All model runs indi- cated a depleted population with high fishing mortality, 10 -. 9 F/F„,5, FI-1 / 8 / 7 . - - - FI-3 / -. // s":: .-..-''^-^'^^^^■' 3- /""^ 2 . 1 . 0 ■ -^>: - -'--''''"^"^--'■- ~' — 1991 1993 1995 199?' 1999' ' 2001 ' 2003' 1 Year Figure 7 Production model estimates of relative fishing mortality (F/F,,,,-) of horseshoe crabs (Limulus polyphemus) in the Delaware Bay region. 1991-2003. Results from fishery-independent model runs are shown. and the horizontal line represents F/F„j,.y=l. SOVr confidence inter- vals for each model run are the same line pattern in gray. 45 40 3.5 3.0 25 2.0 15 10 05 0.0 F/F, 1993 1995 1997 1999 Year Figure 8 Production model estimates of relative fishing mortality ^F/F^|^y) of horseshoe crabs iLimulus polyphemus) in the Delaware Bay region, 1991-2002. Results from fishery-dependent model runs are shown, and the horizontal line represents FIFmsy-^- ^0% confidence inter- vals for models FD-1 and FD-4 are shown in gray. although the estimated extent differed among models (Fig. 4). Projections for some of the model runs predicted a relatively fast population recovery. In the absence of harvest, five of the seven model applications predicted rebuilding to 5^/.,y in five or fewer years. However, two of the model runs estimated such low population biomass that rebuilding the Delaware Bay population to B^,^y would take greater than 15 years, even with no harvest. A precautionary management strategy may therefore be appropriate for this population in the short term as more data are being collected. Population model results may have been affected by assumptions that we made during the modeling pro- cess. In order to extend the time period of the model back to 1991, we had to estimate regional harvest based on Delaware harvest data for the early years of the model. Because har- vest data from most states were not avail- able prior to 1995, we assumed that regional harvest was equal to ten times the Dela- ware harvest for the years 1991-94. This was the approximate relationship between Delaware and regional harvest from 1997 through 2003 when reporting was manda- tory. Actual regional harvest for 1991-94 was unknown because there were no har- vest-reporting requirements. We conducted sensitivity analyses which showed very little difference among runs with 1991-94 harvest equal to Delaware landings, 10 times Dela- ware landings, or 20 times the Delaware landings (the mean difference for Bo(fQ-^IB^,^y from these runs was 0,011). If the actual harvest was substantially greater than 20 times Delaware landings, differences in re- sults may occur. Results were also influenced by the data sources that were included. We based the inclusion of fishery-independent surveys on positive correlations among surveys, incor- porating the largest number of data sources possible into model runs. However, two fish- ery-independent surveys were not included: the NMFS fall trawl and the Delaware Bay spawning survey. The NMFS fall trawl was negatively correlated with seven of the nine other data sources (Table 1), and we there- fore assumed that it was not a reliable index of horseshoe crab abundance. The spawning survey was negatively correlated with five of the nine other data sources (Table 1), This survey also had a very short time series be- cause it was redesigned in 1999 to improve statistical power (Smith et al,, 2002). With so few data points, the production model would be unable to distinguish population trends from survey variability; therefore spawning survey data were excluded. However, as one of only two horseshoe crab-specific surveys currently in place, the Delaware Bay spawn- Davis et aL: Population assessment of Limulus polyphemus 223 ing survey will likely prove to be a valuable source of information in the future. Negative correlations were present among a relatively large number of Delaware Bay surveys (Table 1) which are assumed to be sampling the same population. This is a common problem in fisheries stock assess- ments (Richards, 1991; Schnute and Hilborn, 1993). The observed differences among these surveys could be attributed to a number of factors. Catches of horseshoe crabs are not common, leading to small sample sizes and high variability. Also, these surveys may differ in location, time of year, and gear selectivity. Future studies could employ an analysis of variance (ANOVA) to attempt to separate these factors from underlying horseshoe crab abundance trends. The data sources included in individual model runs also led to differing results. Models incorporating fishery-dependent data tended to present a slightly more optimis- tic view of the population and the fishery. predicting higher relative biomass and lower relative fishing mortality. Although harvest data often have the benefit of having been derived from large sample sizes (and have resulting low variance estimates), there is often a bias associated with fishery-dependent data. Fisheries do not sample randomly because they target areas of highest abundance, and thus biased indices are produced (Quinn and Deriso, 1999). In addition, the use of fishery-dependent abundance indices is often complicated by changes in gear, regulations, or sam- pling methods over time, any of which could affect catch rates. Fishery-independent surveys are usually more 2 °°1 Projected B/B^sy (model FI-1) 1 75- 1 50 1 25- CO 1 00 HI 0 75- 0 50 0 25- 0 00 2004 2006 2008 2010 2012 Year 2014 Figure 9 Example of projection results. Projected relative biomass iB/B^/^y) in shown for model FI-1, with each line representing a harvest level applied annually in the 15-year projections. The percentage refers to the percent of the 2003 Delaware Bay regional landings of 1356 t (i.e., 50% = 678 t). appropriate for assessments, assuming there is high consistency in sampling methods among years (Chen et al., 2003). Differences also existed among fishery-independent surveys. Models FI-2 and FI-3 predicted a much low- er population size than FI-1 or the fishery-dependent models. Projections with these runs predicted that the population was too depleted to recover in 15 years, even in the absence of harvest. The models differed only in the abundance indices included. In the trend analyses conducted for ASMFC,- the most significant population declines since 1997 were identified in the NJ ocean Table 4 Results of Delaware Bay population projections from production model runs from fishery-independent (FI) and fishery- dependent (FDl indices. Projections were conducted for 15 years, with a constant harvest (in t) applied annually. Harvest levels were based on 2003 harvest and are listed in the left column. The time period (and 80'7f confidence intervals) shown represent the number of years (starting in 2003) required for the population biomass to reach B,,^.).. "n/a" indicates that the biomass did not reach S.i/sy during the 15-year projection period. Harvest level in relation to that of 2003 Years to rebuild tc Bms\ FI-1 FI-2 FI-3 FD-1 FD-2 FD-3 FD-4 QVc (Ot) SOTc CI 5 years (2, 12) n/a (3, n/a) n/a ( n/a, n/a) 4 years (3,7) 3 years (1,6) 2 years (0,8) 2 years (0, 15) 25'7r(339t) 80% CI 5 (2, n/a) n/a (3, n/a) n/a ( n/a, n/a) 5 (3,8) 3 (1,8) 2 (0,9) 2 (0, n/a) 50'7f (678 t) 80% CI 6 (2, n/a) n/a (n/a, n/a) n/a (8, n/a) 6 (3, 11) 4 (1, 10) 3 (0, 14) 2 (0, n/a) 75% (1017 t) 80% CI 9 (3, n/ai n/a (n/a, n/a) n/a (5. n/a) 7 (4, 15) 4 (1, n/a) 3 (0, n/a) 3 (0, n/a) 100% (1356 t) 80% CI n/a (3, n/a) n/a (n/a. n/a) n/a ( n/a, n/a) n/a (5, n/a) 6 (2, n/a) 5 (0, n/a) 4 (O.n/a) 224 Fishery Bulletin 104(2) Table 5 Results of Delaware Bay population projections from production model runs from fishery-independent (FI) and fishery- dependent ( FD 1 indices. Projections were conducted for 15 years, with a constant harvest (in t) applied annually. Harvest levels were based on 2003 harvest and are listed in the left column. The relative biomass (B/B„gy) in the final year of projections (2018J and 80^r confidence intervals are shown for each model simulation. Harvest level in relation to that of 2003 ^2018''^MSy FM FI-2 FL3 FD-1 FD-2 FD-3 FD-4 0%(Oti 2.00 0.41 0.14 2.00 2.00 2.00 2.00 80'7f CKLl 1.49 o.oa 0.04 1.91 1.87 1.63 1.05 80'7f CKU) 2.00 1.97 0.28 2.00 2.00 2.00 2.00 25<7f (339t) 1.94 0.00 0.06 1.95 1.93 1.92 1.91 SO'/f CKL) 0.00 0.00 0.00 1.95 1.90 1.42 0.86 80<7, CKU) 1.96 1.94 0.42 1.95 1.93 1.93 1.93 50% (678 t) 1.87 0.00 0.00 1.90 1.86 1.84 1.82 80% CKL) 0.00 0.00 0.00 1.89 1.80 1.07 0.61 80% CKU) 1.93 0.00 1.76 1.90 1.86 1.86 1.86 75% (1017 t) 1.78 0.00 0.00 1.83 1.78 1.74 1.71 80% CI (L) 0.00 0.00 0.00 1.66 0.86 0..57 0.26 80% CKU) 1.85 0.00 1.74 1.85 1.78 1.78 1.77 100% (13561) 0.00 0.00 0.00 0.76 1.67 1.62 1.58 80% CKL) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 80% CKU) 1.72 0.00 0.00 1.77 1.69 1.69 1.68 trawl and the DE 16-ft 160-mm survey, both of which were included in model FI-3, and possibly explain the low population estimates. In other trend analyses conducted during the previous assessment, a population decline in the Delaware Bay was less evident. Using data from 1997 through 2003, we found that only four of eight fishery-independent surveys showed a significant decline, partially owing to high variability and low power. By incorporating a number of these surveys into a production model, we also found that the decreasing biomass in recent years becomes more apparent. These production model runs provide the added benefit of estimating stock status and management benchmarks, as well as the benefit of evaluating future management options. Although interpretation of absolute biomass or popula- tion size from a surplus production model can be some- what problematic, our estimates are roughly comparable to estimates from previous studies. Hata and Berkson (2003) calculated a mean 2001 population size of 4.4 million adults (95% confidence intervals of 2.1 million and 6.8 million) in the Delaware Bay from daytime trawl survey data. Botton and Ropes (1987) estimated 2.3 to 4.5 million adults in this region. In the present study, our estimates of the 2003 population size ranged from 0.6 million to 3.6 million crabs, and the mean of the seven model runs equaled 2.0 million crabs. Eighty percent confidence intervals ranged from 0.3 million to 6.6 million crabs for all model applications. Although within the range of results from the other studies, these wide confidence intervals provide little information for management. It will therefore be more appropriate to interpret relative biomass (B/B„j;j.) and relative fishing mortality (i^/.F^/.sj) for use in management decisions (Prager, 1994). It is important to understand the spatial scale and population represented by these models and analyses. This regional model is a compilation of a number of localized fishery-independent surveys, most of which encompassed a relatively small spatial area. However, the model results should not be interpreted at a more localized scale because landings data are combined for the region. Similarly, interpretation of results should not be expanded to represent other Atlantic horseshoe crab populations outside the Delaware Bay region be- cause neither survey nor harvest data in this model extend to other regions. Nevertheless, the Delaware Bay is believed to be the center of abundance and spawning activity for Atlantic horseshoe crabs; therefore popula- tion trends in this region may have significant implica- tions for adjacent populations. In analyses conducted for ASMFC,- trend analyses identified dramatic regional differences in horseshoe crab population trends. Although the Delaware Bay, eastern Long Island Sound, and New England popula- tions have experienced declines in recent years, the southeast population and the western Long Island popu- lation have remained stable or have increased. Future production models applied to these other regions will hopefully clarify these trends and allow managers to Davis et al Population assessment of Limulus polyphemus 225 determine regional harvest regulations. By identify- ing appropriate management for each region, we will improve our ability to rebuild the Atlantic horseshoe crab population and provide a sustainable resource for the diverse user groups. Acknowledgments This research was funded by the National Marine Fish- eries Service. We thank M. Prager and M. Gibson for providing modeling advice, D. Smith, S. Michels, and B. Andres for providing horseshoe crab data and guid- ance, and the members of the Horseshoe Crab Stock Assessment and Technical Committees of the ASMFC for their contribution to this study. This manuscript was improved by the comments and suggestions from M. Prager, M. Millard, S. Michels, B. Murphy, D. Hata, and S. Klopfer, and two anonymous reviewers. We greatly appreciate the time and effort of all involved. Literature cited Berkson, J., and C. N. Shuster. 1999. The horseshoe crab: the battle for a true multiple- use resource. Fisheries 24:6-10. Booth, A. J., and A. E. Punt. 1998. Evidence for rebuilding in the panga stock on the Agulhas Bank. South Africa. Fish. Res. 34:103-121. Botton, M. L., and B. A. Harrington. 2003. Synchronies in migration: shorebirds, horseshoe crabs, and Delaware Bay. In The American horseshoe crab (C. N. Shuster, R. B. Barlow, and H. J. Brock- man, eds.), p. 5-26. Harvard Univ. Press, Cambridge, MA. Botton, M. L., R. E. Loveland, and T. R. Jacobsen. 1994. Site selection by migratory shorebirds in Dela- ware Bay, and its relationship to beach characteristics and abundance of horseshoe crab {Limulus polyphemus) eggs. Auk 111:605-616. Botton, M. L., and J. W. Ropes. 1987. Populations of horseshoe crabs, Limulus polyphemus. on the northwestern Atlantic continental shelf. Fish. Bull. 85:805-812. Botton, M. L., and C. N. Shuster. 2003. Horseshoe crabs in a food web: who eats whom? In The American horseshoe crab (C. N. Shuster, R. B. Barlow, and H. J. Brockman, eds.), p. 33-153. Har- vard Univ. Press. Cambridge, MA. Caddy, J. 1998. A short review of precautionary reference points and some proposals for their use in data-poor situations. FAO Fish. Tech. Pap. 379, 29 p. FAO, Rome. Cadrin, S. X., and E. M. C. Hatfield. 2002. Relative biomass and production of longfin inshore squid, Loligo pealeii. Bull. Mar. Sci. 71:1115-1116. Chen, Y.. L. Chen, and K. I. Stergiou. 2003. Impacts of data quantity on fisheries stock assessment. Aquat. Sci. 65:92-98. Collie, J. S., and M. P. Sissenwine. 1983. Estimating population size from relative abundance data measured with error. Can. J. Fish. Aquat. Sci. 40:1871-1879. Hata, D., and J. Berkson. 2003. Abundance of horseshoe crabs iLimulus polyphe- mus) in the Delaware Bay area. Fish. Bull. 101:933-938. MacCallA. D. 2002. Use of known-biomass production models to de- termine productivity of west coast groundfish stocks. N. Am. J. Fish. Manag. 22:272-279. Novitsky, T. J. 1984. Discovery to commercialization: the blood of the horseshoe crab. Oceanus 27:13-18. Prager, M. H. 1994. A suite of e.xtensions to a nonequilibrium surplus- production model. Fish. Bull. 92:374-389. Prager, M. H., and C. P. Goodyear. 2001. Effects of mixed-metric data on production model estimation: simulation study of a blue marlin-like stock. Trans. Am. Fish. Soc. 130:927-939. Punt. A. E. 1990. Is B] = K an appropriate assumption when apply- ing an observation error production-model estimator to catch-effort data? S. Afr. J. Mar. Sci 9:249-259. Quinn, T. J., and R. B. Deriso. 1999. Quantitative fish dynamics, 542 p. Oxford Univ. Press, New York, NY. Richards, L. J. 1991. Use of contradictory data sources in stock assessments. Fish. Res. 11:225-238. Schnute, J. T, and R. Hilborn. 1993. Analysis of contradictory data sources in fish stock assessment. Can. J. Fish. Aquat. Sci. 50:1916-1923. Shuster, C. N. 1982. A pictorial review of the natural history and ecology of the horseshoe crab, Limulus polyphemus, with refer- ence to other Limulidae. In Physiology and biology of horseshoe crabs: studies on normal and environmentally stressed animals (J. Bonaventura, ed.), p. 1-52. Alan R. Liss, Inc. New York. Smith, D. R., P. S. Pooler, B. L. Swan, S. F Michels, W. R. Hall. P. J. Himchak. and M. J. Millard. 2002. Spatial and temporal distribution of horseshoe crab {Limulus polyphemus) spawning in Delaware Bay: implications for monitoring. Estuaries 25:115-125. Vaughan. D. S., and M. H. Prager. 2002. Severe decline in abundance of the red porgy {Pagrus pagrus) population off the southeastern United States. Fish. Bull. 100:351-375. Walls, E. A., and J. Berkson. 2003. Effects of blood extraction on horseshoe crabs {Limulus polyphemus). Fish. Bull. 101:457-459. Walls, E. A., J. Berkson. and S. A. Smith. 2002. The horseshoe crab, Limulus polyphemus: 200 million years of existence, 100 years of study. Rev. Fish. Sci. 10:39-73. 226 Abstract — Data collected from an annual groundfish survey of the eastern Bering Sea shelf from 1975 to 2002 were used to estimate bio- mass and biodiversity indexes for two fish guilds: flatfish and roundfish. Biomass estimates indicated that several species of flatfish (particu- larly rock sole, arrowtooth flounder, and flathead sole), several large sculpins {Myoxocephalus spp.), big- mouth [Hemitripterus bolini), and skates iBathyraja spp.) had increased. Declining species included several flatfish species and many smaller roundfish species of sculpins, eel- pouts (Lycodes spp.), and sablefish (Anoplopoma fimbria). Biodiversity indexes were calculated by using bio- mass estimates for both guilds from 1975 through 2002 within three physi- cal domains on the eastern Bering Sea shelf. Biodiversity trends were found to be generally declining within the roundfish guild and generally increasing within the flatfish guild and varied between inner, middle, and outer shelf domains. The trends in biodiversity indexes from this study correlated strongly with the regime shift reported for the late 1970s and 1980s. Biodiversity as an index of regime shift in the eastern Bering Sea Gerald R. Hoff Alaska Fisheries Science Center National Marine Fisheries Service 7600 Sand Point Way N E Seattle, Washington 98115 Email address; jerry.hofftSnoaagov Manuscript submitted 25 March 2004 to the Scientific Editor's Office. Manuscript approved for publication 13 August 2005 by the Scientific Editor. Fish. Bull. 104:226-237 (2006). Environmental and biological events in the North Pacific can indicate regime shifts or reorganizations of the eco- system at the environmental and bio- logical level (Hare and Mantua, 2000). Measurable climatic events were iden- tified in the mid-1970s, late 1980s (Anderson and Piatt, 1999; Anderson, 2000; Hare and Mantua, 2000; Zhang et al., 2000; Minobe, 2002) and the late 1990s (Hunt and Stabeno, 2002; Minobe, 2002). Regime shifts in the eastern Bering Sea (EBS) have been correlated with several climatic events, including the Pacific Decadal Oscilla- tion, the El Niiio Southern Oscilla- tion (Hollowed et al., 2001), sea ice coverage (Stabeno et al., 2001; Hunt et al., 2002; Hunt and Stabeno, 2002), and summer sea surface temperatures (Bond and Adams, 2002; Hunt et al., 2002; Minobe, 2002). The far reach- ing effects that climate change has on an ecosystem are not well defined, but many studies have shown strong correlations between climate change, fish recruitment, and plankton pro- duction in the North Pacific (Brodeur and Ware, 1992; Anderson et al., 1997; Anderson and Piatt, 1999; Brodeur et al., 1999; Clark et. al., 1999; Hare and Mantua, 2000; Zhang et al., 2000; Hollowed et al., 2001; Sugimoto et al., 2001; Conners et al., 2002; Hunt et al., 2002; Wilderbuer et al., 2002). During periods of climatic change, some fishes may not be well adapted to dramatic changes over a short time scale (1-10 years), whereas others may proliferate in a more hospitable environment. Key triggers of regime shifts and the extent to which species will respond remain unclear; however, evidence may indicate that species diversity is correlated with primary production in many systems, and the two may be interdependent (Rosenz- weig and Abramsky, 1993; Waide et al., 1999). A biodiversity index can be useful for monitoring the stability, health, and productivity of an ecosys- tem, as well as for aiding manage- ment, by tracking exogenous changes and their far reaching effects on spe- cies. In the present study, I used bio- diversity indexes, reflecting richness and evenness, as indicators of species composition changes and related these changes to regime shift events for the eastern Bering Sea (EBS) shelf. Methods Data were synthesized from a con- tinuous 24-year period of standard- ized groundfish surveys conducted by the Alaska Fisheries Science Center (AFSC) on the EBS shelf from 1979 through 2002; additional estimates from 1975 were included. Biodiversity indexes (species richness and even- ness) were calculated from biomass estimates for two fish guilds, flatfish and roundfish, in each of the inner, middle, and outer domains of the EBS shelf. Biodiversity indexes were plot- ted for each fish group and domains and examined for changes over the study period. The observed changes were correlated with regime shift and ocean climatic change events for the EBS. Data collection Gear, station location, sampling pro- cedures, and time of year of the AFSC Hoff Biodiversity as an index of regime shift in tlie eastern Bering Sea 227 Survey area in the stations from 20 outer domains. eastern Bering Sea continental shelf survey have been standard- ized since 1982. Prior to 1982, a 400-mesh eastern trawl (smaller than the current 83-112 trawl) was used (Bakkala, 1993). Although the biomass estimates produced by the two trawls are not directly comparable because of unknown catchability differences between the nets, survey data from 1975 and 1979-81 were included in the present analysis because of the importance in the timing of the 1970s regime shift. The biodi- versity indexes are based on the relative proportion of the species and therefore are an indicator of assemblage structure. Data used for this study were collected by the Resource Assess- ment and Conservation Engineer- ing (RACE) division of the AFSC, which surveys the EBS shelf each summer (May- August). The sur- vey area extends from the Alaska Peninsula north to Nunivak Is- land and St. Matthew Island, and west to the 200 m shelf break. Trawl hauls were conducted on a grid of 356 fixed stations (20 nmi by 20 nmi) (Fig. 1) during daylight hours. Hauls were towed for 30 min at 3 knots and ranged in depth from 15 m in Bristol Bay to nearly 200 m near the shelf edge. Most trawling was conducted with the AFSC 83-112 eastern trawl (1982-2002), which is a low-opening two-seam trawl with a 26.5-m headrope and 34.1-m cable footrope wrapped with rubber striping and chain hangings that contact the bottom while the trawl is towed (Rose and Walters, 1990). Height and width of the net were measured with an acoustic SCANMAR (Scanmar, Asgardstrand, Norway) or NETMIND (Northstar Technical Inc., St. John's, NF, Canada) net mensuration system, or estimated by us- ing a function that relates trawl widths to tow depths from measured hauls. Each haul was measured with GPS or LORAN to record latitude and longitude data at the start and end of the trawl in order to determine distance fished. Processing of the catch was done entirely in the field at the time of capture. The entire catch was sorted to species, enumerated, and weighed, or a weighed sub- sample was used for very large catches. Catch per unit of area (CPUE) was determined for each trawl station completed during each survey. Bio- mass estimates for each species were then calculated by expanding the average CPUE for each stratum and then summing over all strata of the total survey area to obtain a biomass estimate in metric tons for each species. Figure 1 eastern Bering Sea. Crosses represent annual trawl survey 200 m. Bathymetry lines delineate the inner, middle and EBS shelf domains and study area The EBS is composed of three well-defined regions des- ignated as the inner, middle, and outer domains from east to west, respectively, across the shelf (Fig. 1). The three domains are distinct regions characterized by depth, water temperature, current flow, summertime primary production, and species composition (Bakkala, 1993; Schumacher and Stabeno, 1998; Stabeno et al., 2001). Briefly, the inner domain is a relatively shal- low (<50 m) and well-mixed warm basin with strong influences from a large coastal area, several major river systems, and a current flow northward along the coast from the Aleutian Islands. The middle domain is a relatively stagnant area of deeper water (50-100 m) that has strong summer water-column stratification and low current flows. The bottom water mass is relatively cold, overlaid by a layer of warmer wind-mixed water. Although characterized as a cold region, the water mass in the middle domain varies from year to year. The outer domain is influenced by the EBS slope region and by upwelling and northward current flow from the Aleutian Islands. This domain is relatively warm when compared to the middle domain (see Hunt et al., 2002, for review). For the purposes of this study the inner domain designation consisted of trawl stations less than 50 m in depth; the middle shelf designation was for trawl stations between 51 and 100 m in depth; and outer domain designation was for trawl stations between 101 and 200 m in depth. 228 Fishery Bulletin 104(2) Table 1 Flatfish guild species or species groups used for biodiversity indexes. * indicates the major proportion of the biomass for the group. Superscripts indicate the primary domains where each species occurs during the summertime AFSC survey. I=inner domain, M=middle domain, 0=outer domain. Group name Scientific name Pacific halibut Hippoglossus stenolepis"^'° Flathead/Bering flounder *Hippoglossoides elassodon and H. robustus'^"^ Greenland turbot Reinhardtius hippoglossoides'^ Arrowtooth/Kamchatka flounders '■Atheresthes stomias and A. evermaiiti'^"^ Starry flounder Platichthys stellatus'^' Alaska plaice Pleuronectes quadrituberculatus'^' Northern/southern rock sole *Lepidopsetta polyxystra and L. bilineata'''^"^ Longhead dab Limanda proboscidea' Yellowfin sole Limanda aspera"^ Rex sole Glyptocephalus zachirus^ Oher flatfish Hsopsetta isolepis'^'. 'Limanda sakhalinensis^"^, Psettichthys melanostictus, Microstomus pacificuti Fish assemblages differed noticeably among the inner, middle, and outer domains in the EBS; many species (sculpins, poachers, and eelpouts) primarily inhabited a single domain and many flatfishes and skates inhab- ited all three domains (Kaimmer'; Kinder and Schum- acher, 1981; Smith and Bakkala, 1982: Tables 1 and 2). Because of the distinct domain environments and assemblages, the flatfish and roundfish species groups were analyzed by inner, middle, and outer domains separately and all domains were combined to detect the influence of climate change on each assemblage. Species groups The taxonomic level of species identification for the entire survey period (1975-2002) has varied because of an incomplete knowledge of species characters and because of survey time constraints at-sea. Historically, survey goals were to obtain fisheries data on commer- cially or potentially commercially important species, and limited effort was put forth for identification of other species. However, given more efficient technolo- gies, better identification field guides, and increased focus on noncommercial species, species identification has improved and is approaching the current extent of taxonomic knowledge. Efforts in recent years (1998-2002) have increased our knowledge of current species distributions in the Kaimmer, S. M., J. E. Reeves, D. R., Gunderson, D. R.. Smith, G. B., and R. A. Macintosh. 1976. Baseline information from the 1975 OCSEAP survey of the demersal fauna of the eastern Bering Sea. In Demersal fish and shellfish resources of the eastern Bering Sea in the baseline year 1975 (W. T Pereyra, J. R. Reeves, and R. G. Bakkala, eds.), p. 157-367. NWAFC Processed Report. Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E. Seattle WA 98115. EBS shelf region. To assess species-identification con- fidence over the study period, current species distribu- tions from the AFSC survey and primary literature not associated with AFSC survey data, as well as historical fish collections of selected taxonomic groups (University of Washington Fish Collection, Oregon State University Fish Collection, and Auke Bay Fish Collections) were examined. After this assessment, it was subjectively determined whether species in this study were identified correctly throughout the study period. Species that were determined to be distinguishable and correctly identified were examined as individual species, and species that were possibly misidentified were grouped at a higher taxonomic level for this study. The resulting list of spe- cies or species groups includes the greatest number of fish taxa (Tables 1 and 2). At least 15 flatfish species (within the order Pleu- ronectiformes) and more than 75 species of fish and elasmobranchs not of the order Pleuronectiformes were recorded during the AFSC summer surveys (1975-2002) on the EBS shelf (20-200 m). Although flatfish make up about 16% of the total number of fish species, they contribute approximately 50% of the biomass of all fish combined (2001 and 2002 AFSC EBS survey estimates). Walleye pollock (Theragra chalcogramma) and Pacific cod (Gadus macrocephalus) represent approximately 42% of the survey biomass, and 8% of the biomass that included all other roundfish. Because of the extremely large biomass of walleye pollock and Pacific cod, the biodiversity indexes used are uninformative when wall- eye pollock and Pacific cod are included owing to the "swamping" effect by their relatively large biomasses when compared to those of other species. To account for this, two species guilds were defined: flatfish and roundfish. The flatfish guild included all Pleuronecti- formes recorded from the EBS survey and comprised 11 species or species groups (Table 1). The roundfish Hoff Biodiversity as an index of regime shift in tfie eastern Bering Sea 229 Table 2 Roundfish guild species or species groups used for biodiversity indexes. * indicates the major proportion of the biomass for the group. Superscripts indicate the primary domains where each species occurs during the summertime AFSC survey. I=inner domain, M=middle domain, 0=outer domain. IMO=inner, middle, and outer domains. Group name Scientific name Lamprey Lampetra tridentata Skate *Bathyraja parniifera"''". *B. interrupta", B. aleutica. B. taranetzi. Raja bmociilata Shark *Somniosus pacificus^, Squalus acanthiaa. Lamna ditropis Other poachers *Aspidophoroides bartoni'^, Bathyagonus ainscanus. Leptagonus leptnrhynchus. Pallasina barbata, Percis japonicus Sawback poacher Sarritor frenatus-'''"^ Sturgeon poacher Agonus acipenserinus"^' Bering poacher Ocellus dodecaedron' Bering wolffish and wolfeel Auarhichas nrieittalis and Anarrhichthys ocellatus Searcher Bathymaster signatus" Other sculpins Artediellus pacificus, Leptocottus armatus, Enophrys spp., Psychrolutes paradoxus, Blepsias bilobus, Nautichthys pribilovious, Icelinus spp. Staghorn sculpin *Gymnocanthus pistilligei-', *G. galeatus'^. G. dctrisus Darkfin sculpin Malacocottus zonurus'-' Yellow Irish lord Hem ilepidotu s jorda n i' Butterfly sculpin Hemitepidotus papilio-^' Triglops sculpins *Triglops scepticus^. *T. pingeli'^'. T. macellus. T. forficatus Myoxocephalus sculpt ns *Myoxocephalusjaok', *M. polyacanthocephalus^"^, M. verrucosus'^ Spinyhead sculpin Dasycottus setiger^ Bigmouth sculpin Hemitripterus holinfi Icelus sculpins ^'Icelus spmiger" and */. spatula'-'^' Pacific sandfish Trichodon trichodon' White-spotted greenl ing Hexagrammos stelleri' Snailfish *Liparis gibbus'^'-', *Careproctus rastrinus^'^, C. cyclospilus Smooth lumpsucker *Aptocyclus ventricosus'^ , Eumicrotremus orbis Pricklebacks *Lumpenus maculata"'"^, *L. sagittae'^"^, L. fabricii, L. medius. Poroclinus rolhrocki Other eelpouts *Lycodes spp. and Gyinnelus spp. Marbled eelpout Lycodes raridens' Wattled eelpout Lycodes palearis^° Shortfin eelpout Lycodes brevipes^ Pacific tomcod Microgadus proximus Saffron cod Elegiiius gracilis' Arctic cod Boreogadus saida^ Pacific sand lance Ammodytes hexapterus Pacific herring Clupea pallasii"^' Eulachon Thaleichthys pacificus'-' Capelin Mallotus villosus' Rainbow smelt Osmerus mordax"^"^ Salmon Oncorhynchus spp. Sablefish Anoplopoma fimbria^ Atka mackerel Pleurogrammus monopterygius Prowfish Zaprora silenus'^ Rockfish *Sebastes alutus°, S. polyspinis. S. ciliafus 230 Fishery Bulletin 104(2) guild included 40 species or species groups and excluded walleye pollock and Pacific cod (Table 2). The two guilds devised were subjective; however, owing to their dis- tinctive biomass and life history trends, the members of the flatfish guild were examined separately from all other species. Biodiversity indexes Biodiversity measures were calculated by using Ludwig and Reynolds's (1988) recommendations for species rich- ness and evenness. Richness and evenness are consid- ered robust measures and allow one to use biomass proportions for biodiversity estimations. The richness index used was as follows: Richness index = e" where H' = ^ 'iHi) A piecewise linear model (Neter et al., 1996) was applied to the biodiversity indexes using S-Plus (vers. 6.1, Insightful Corporation, Seattle, WA) to determine distinct breaks or inflection points in the linear data trends. The piecewise model finds a "knot," or inflec- tion point, that breaks the data set into two periods by using the least sum of squared residuals and the high- est i?'- for the two-line model to determine the best fit lines to the data. Linear regression models were then applied to the data sets by using the recommended inflection point as the breaking point for the two lines. The years covered for each best fit line, as well as the accompanying linear statistics for the individual lines and the piecewise linear model for the best-fit model are presented in Table 3. The indexes calculated for survey year 1975 are included for reference (denoted as an X on the diversity index graphs) but were not included in the linear regression models because of the time gap from 1975 to 1979 and the lack of confidence in the standardization of the 1975 survey compared with later surveys (Bakkala, 1993). and /?, = the biomass of individuals belonging to the /th of S species in the sample; and n = the total summed biomass for the entire guild for a given year (Ludwig and Reynolds, 1988). The evenness index is the "modified Hill's ratio" (Ala- talo, 1981) which is Evermess = where A=-;^py, where Pi = N and ?!, is the biomass of the (th species and A^ is the total biomass of all iS (species) in the guild (Ludwig and Reynolds, 1988). Biomass estimates obtained from the surveys were used to calculate biodiversity values for both flatfish and roundfish guilds on an annual basis. Biodiversity values calculated were used as an index of the relative proportion of species in each guild and allowed a direct comparison from year to year of the most abundant species. Species richness indicated the effective number of species that are influencing the index (Hill, 1973; Ludwig and Reynolds, 1988). A higher index indicates a larger number of predominant species in the assemblage (i.e., higher diversity). The evenness index determines the distribution of the biomass proportions of the more abundant species among the species in the guild. As the evenness index approaches zero, the biomass proportion of a single species dominates the guild; as the evenness index approaches one, there is less of a single dominant species and the biomass proportions are shared more equally among many species in the guild (Hill, 1973; Ludwig and Reynolds, 1988). Results Flatfish biomass Biomass estimates for the flatfish guild showed an increase from the late 1970s until the early 1990s, and then an overall slight decline between 1999 and 2002 for all domains (Fig. 2). Species that represented a large por- tion that showned an increase in biomass estimates over the study period included northern and southern rock sole (Lepidopsetta polyxystra and L. bilineata, spp. undeter- mined), arrowtooth and Kamchatka flounder (Atheresthes stomias and A. evermanni, spp. undetermined), flathead sole and Bering flounder (Hippoglossoides elassodon and H. robustus, spp. undetermined), and Pacific halibut iHippoglossus stenolepis). Flatfish species with declining biomass estimates included the yellowfin sole (Limanda aspera) and Greenland turbot {Reinhardtius hippoglos- soides), and a decline in lesser abundant species such as longhead dab {Limanda proboscidea) and Alaska plaice (Pleuronectes qiiadrituberciilafus). The inner and middle domains were dominated by northern rock sole, yellowfin sole, Alaska plaice, and flathead sole and Bering flounder (spp. undetermined), whereas main portions of the outer shelf comprised arrowtooth and Kamchatka flounder (spp. undetermined), flathead sole and Bering flounder (spp. undetermined). Pacific halibut, and northern and southern rock sole (spp. undetermined). Roundfish biomass Roundfish biomass increased from the 1970s to the 1980s and then declined steadily from the early 1990s through 2002 for all domains combined (Fig. 2). Among roundfish species there has been a decline since the 1970s and 1980s in many middle domain species, such as yellow Irish lord (Hemilepidotus jordani), butterfly sculpin (H. papilio), Hoff Biodiversity as an index of regime shift in tlie eastern Bering Sea 231 Table 3 Statistics of the 1 near regression models of biodiversity indexes for flatfish and roundfish guilds for the eastern Ber ng Sea shelf domains. The years encompassing each time period and the piecewise model r- for the two-1 ne model are included. Standard r^of error of piecewise Domain Index Years n Intercept Slope P-value r2 estimate model Flatfish guild Inner Richness 1979-1993 15 -112.9280 0.0579 0.0000 0.8797 0.0993 0.8850 1994-2002 9 3.9168 -0.0008 0.9645 0.0303 0.1270 Middle Richness 1979-1988 10 -358.4240 0.1821 0.0001 0.8829 0.2130 0.8640 1989-2002 14 7.8482 -0.0020 0.9171 0.0940 0.2887 Outer Richness 1979-2002 24 -2.8281 0.0030 0.6875 0.7493 0.2533 Combined Richness 1979-1989 11 -327.9090 0.1666 0.0000 0.8740 0.2212 0.9350 1990-2002 13 6.1232 -0.0012 0.8866 0.1932 0.1150 Inner Evenness 1979-1990 12 -48.3712 0.0247 0.0000 0.9291 0.0258 0.9610 1991-2002 12 -0.0349 0.0005 0.6927 0.0163 0.0161 Middle Evenness 1979-1987 9 -61.6390 0.0314 0.0003 0.8663 0.0361 0.8660 1988-2002 15 3.7805 -0.0015 0.5159 3.3171 0.0375 Outer Evenness 1979-2002 24 1.8798 -0.0006 0.6184 0.0115 0.0392 Combined Evenness 1979-1988 10 -59.8421 0.0305 0.0000 0.9318 0.0265 0.9630 1989-2002 14 2.0678 -0.0007 0.4732 0.0437 0.0136 Roundfish guild Inner Richness 1979-2002 24 -5.0326 0.0040 0.8287 0.2175 0.6197 Middle Richness 1979-1986 8 865.6880 -0.4338 0.0119 0.6789 0.7895 0.9040 1987-2002 16 203.6660 -0.1006 0.0000 0.7196 0.3095 Outer Richness 1979-1986 8 908.1290 -0.4564 0.0011 0.8511 0.5051 0.8690 1987-2002 16 -65.9678 0.0340 0.0762 0.2075 0.3276 Combined Richness 1979-1986 8 1535.3000 -0.7713 0.0003 0.9072 0.6527 0.9410 1987-2002 16 81.4226 -0.0394 0.1204 0.1634 0.4390 Inner Evenness 1979-1990 12 -14.9002 0.0078 0.3003 0.1066 0.0853 0.5620 1991-2002 12 -32.0753 0.0164 0.0057 0.5504 0.0561 Middle Evenness 1979-1990 12 51. .5981 -0.0257 0.0001 0.7779 0.0518 0.9210 1991-2002 12 20.3337 -0.0099 0.0002 0.7694 0.0206 Outer Evenness 1979-1987 9 66.1545 -0.0331 0.0024 0.7550 0.0552 0.7150 1988-2002 15 -16.5892 0.0085 0.0110 0.4028 0.0482 Combined Evenness 1979-1989 11 46.7485 -0.0233 0.0003 0.7782 0.0434 0.8630 1990-2002 13 1.1549 -0.0003 0.8818 0.0021 0.0306 armorhead sculpin (Gymnocanthus galeatus), snailfishes {Liparis gibbus and Careproctus ?-astrinus). marbled eel- pout (Lycodes raridens), and wattled eelpout [Lycodes palearis), and in outer domain species such, as shortfin eelpout (Lycodes brevipes) and sablefish (Anoplopoma fimbria). There has been a increase in large sculpins, including the bigmouth sculpin iHemitripteriis bolini) and the great sculpin (Myoxocephahis polyacanthocephcdus), in the outer domain. The biomass of skates predominantly the Alaska skate (Bathyraja parmifera), increased consid- erably from 1979 through 1990, followed by a prolonged period of little change from 1990 through 2002. domain the major portion of the assemblage comprises 3-5 species. Figure 3 shows the change in biodiversity indexes from 1975 to 2002 for each domain separately and all domains combined. The piecewise model shows a period of inflection for evenness and richness indexes for flatfish in the inner and middle domains that occurred between 1987 and 1993, with the average year of inflec- tion being 1990 (Fig. 3, Table 3). The outer domain for flatfish showed little change in evenness and richness throughout the 24-year study period. Three species dominated the entire period, with somewhat evenly distributed biomass proportions. Flatfish biodiversity Eleven species or species groups were used for biodi- versity indexes for the flatfish guild (Table 1). All 11 species are rarely found together, and within each shelf Roundfish biodiversity Forty-one species or species groups were used in the roundfish biodiversity analysis (Table 2). The even- ness and species richness index for the roundfish group 232 Fishery Bulletin 104(2) Flatfish guild Inner domain Middle domain Flatfish guild B other flatfish □ longhead dab H rex sole Year starry flounder Pacific halibut Alaska plaice $113 g @ Greendland turbot [H flathead/Berring flounder n arrowtooth/Kamchatka flounder Year yellowfin sole northern/southern rock sole Figure 2 Biomass estimates for flatfish guild and roundfish guild (left panel), and relative proportion of each species (right panel) for inner, middle, outer, and all three domains combined, respectively. Hoff: Biodiversity as an index of regime shift in the eastern Bering Sea 233 Roundfish guild Inner domain JSOQXi 4WXD0 Middle domain ■IS rjis ■i^V^fe' s«r l^w 03 0.6 0.4 02 1994 1999 2004 ^puO'^pooooo '■'»"" — a-e-6 1974 1979 1984 1989 1994 1999 2004 1 0.8- 06 04- 0.? oJ> oOnoOoOnooo^oo 1974 1979 1984 1989 1994 1999 2004 Year Figure 3 Biodiversity indexes (richness left, evenness right) for the flatfish guild determined with two linear models (or a single model) separated at the inflection point derived from the piecewise linear model. The graphs represent the inner, middle, outer, and all domains combined, respec- tively. The X represents biodiversity index for survey year 1975 which was not included in linear models or the piecewise model because of a gap in the time series. is plotted for each shelf domain and for all domains combined. The inner domain richness index revealed a steady level of approximately 2-3 dominant species or species groups iMyoxocephalus sculpins, skates, and Pacific herring) over the 24-year period. The gradual increase in skates and slight decline in the plain sculpin caused an increase in the evenness index to somewhat even proportions of the three dominant species in the inner domain, from 7 to 2 dominant species over the 24-year period, and evenness declined by about one-half (Fig. 4). The increase in skates and decline of many inner domain species from 1975 through the early 1990s resulted in an assemblage dominated by two species groups, Myoxocephalus sculpins and skates (Fig. 2). The outer domain showed a dramatic decline in rich- ness and evenness from 1975 to the mid-1980s and Hoff Biodiversity as an index of regime shift in tfie eastern Bering Sea 235 Roundfish Inner domain '" °°o"oo°""°''»°"°°"°°' 1 o.s 06 0 4- 0? 1974 1979 1984 1989 1994 1999 2004 1974 1979 1984 1989 1994 1999 2004 10- Middle domain a- 8i X 7i ~^ ° 6' ° >. 5- ^ 4- "^ 0 3- 0 6 ct"^-^ -T-<-^:^ 0 2- 1- rrf>-^ 1-, 0.8- 0.6- 0.4. 0.2. 1974 1979 1984 1989 1994 1999 2004 1974 1979 1984 1989 1994 1999 2004 Outer domain 10 9-\ 8 7 6 5- 4 3 2 1 1974 1979 1984 1989 1994 1999 2004 1974 1979 1984 1989 1994 1999 2004 0.8- 0.6-1 X 0.4. 0.2. All domains combined 1- 0.8- 0.6- 0.4- 0 2- 0- o^~a -e-r, — n P o .-■ f7< o P ' ■ '■' 0 '■' u o 1974 1979 1984 1989 1994 1999 2004 19'" 1979 1984 1989 1994 1999 2004 Year Year Figure 4 Biomass estimates of roundfish guild (left panel), and relative proportion of each species (right panel) for the inner, middle, outer, and all domains combined, respectively. then a slight increase through 2002. A decrease in the number of dominant species, such as sablefish, wattled and shortfin eelpouts from 1979 to 1986, and a large increase of skate biomass from 20% in 1975 to approxi- mately 75% of the total biomass by 1986, explain the index decline during this same period as the skate biomass dominated. After 1986 the skate populations stabilized somewhat to a slight decline which accounts for the slight increase in evenness and richness as other species gained a larger percentage of the total biomass. All indexes for the roundfish guild, except for inner domain richness, indicated two distinct linear periods. Inflection points ranged from 1986 to 1990 and 1988 was an average inflection year for the roundfish guild. Discussion Periods of climatic change have been shown to instill long-term changes in the North Pacific ecosystem (Hare 236 Fishery Bulletin 104(2) and Mantua, 2000). Biodiversity indexes are a robust measure for large ecosystem monitoring and possible indicators of resulting regime shift phenomenon from climatic change based on an assemblage's long-term responses. Hare and Mantua (2000) summarized over 100 studies from the eastern North Pacific identifying two distinct periods of change or "regime shifts" in 1977 and 1989. Recently. Stabeno et al. (2001) and Hunt and Stabeno (2002) suggested the possibility of a third period of significant climatic change in the late 1990s for the northeastern Pacific. Although environmental changes may be readily identified, the ecosystem fauna may respond more slowly, or in ways not immediately obvious; however the changes may be persistent and far reaching. According to survey data, the last 20 years have proven to be a very hospitable environment iFor many flatfish species in the EBS, where populations such as northern rock sole and arrowtooth and Kam- chatka flounder have rapidly expanded since the late 1970s. Biodiversity indexes are significantly higher than they were 20 years ago for flatfish as a group and have remained high and unchanged. The biodiversity indexes for flatfish guild corroborate the timing of the strong regime shift reported in the late 1980s. An inflection around 1990 is the strongest evidence of a regime shift from the biodiversity indexes. It would be expected that a purely biological response to a climatic event may lag by a time period, perhaps accounting for a later year than previously reported. Survey data indicate that many roundfish species have not fared as well as flatfish species and biodiversity indexes are significantly lower now than 20 years ago. The roundfish guild has undergone a significant reorga- nization in which a large group of species have declined and a single species has become dominant, causing bio- diversity to be suppressed. The regime shift is reflected in the roundfish biodiversity around 1988. which agrees with the results of other reported studies. Although EBS productivity has increased through the 1970s and 1980s and remained high through the 1990s, the trend in the ecosystem is towards fewer or single dominant species. The decline of many nonexploited spe- cies in the EBS is difficult to explain with the current regime shift theories. Unfortunately, a lack of informa- tion on the life history of the large number of species, such as poachers, sculpins, eelpouts, and skates in the EBS hinders a complete understanding of recruitment success or failure. Acknowledgments I thank all the dedicated scientists, vessel skippers, and crew for their hard work over the years collecting data and G. Walters and M. Martin for assistance with data analysis. Also I thank the reviewers for their contribu- tions in improving this manuscript. The reviewers were D. Somerton, G. Walters, J. Orr, S. Kotwicki, G. Stauffer, A. Hollowed, and an unknown journal reviewer. Thank you all. Lierature cited Alatalo, R. V. 1981. Problems in the measurement of evenness in ecology. Oikos 37:199-204. Anderson, P. J. 2000. Pandalid shrimp as indicators of ecosystem regime .shift. J. Northw. At). Fish. Sci. vol. 27:1-10. Anderson, P. J., J. E. Blackburn, and B. A. Johnson. 1997. Declines of forage species in the Gulf of Alaska. 1972-1995, as an indicator of regime shift. In Pro- ceedings of the international symposium on the role of forage fishes in marine ecosystems, p. 531-543. Univ. Alaska Sea Grant College Program Report AK-SG-97-01, Anchorage, AK. Anderson, F J., and J. F. Piatt 1999. Community reorganization in the Gulf of Alaska following ocean climate regime shift. Mar. Ecol. Progr. Ser. 189:117-123. Bakkala, R. G. 1993. Structure and historical changes in the groundfish complex of the eastern Bering Sea. NCAA Tech. Rep. NMFS 114, 91 p. Bond. N. A., and J. M. Adams. 2002. Atmospheric forcing of the southeast Bering Sea shelf during 1995-99 in the context of a 40-year histori- cal record. Deep-Sea Res. 11 49:5869-5887. Brodeur. R. D., M. T. Wilson, G. E. Walters, and I. V. Melnikov. 1999. Forage fishes in the Bering Sea: Distribution, spe- cies associations, and biomass trends. In Dynamics of the Bering Sea: a summary of physical, chemical, and biological characteristics, and a synopsis of research on the Bering Sea (T. R. Loughlin and K. Ohtani, eds.), p. 509-356. Univ. Alaska Sea Grant AK-SG-99-03, Anchorage, AK. Brodeur, R. D., and D. Ware. 1992. Long-term variability in zooplankton biomass in the subarctic Pacific Ocean. Fish. Oceanogr. 1:32-38. Clark. W. G., S. Hare, A. M. Parma, P. J. Sullivan, and R. J. Trumble. 1999. Decadal changes in growth and recruitment of Pacific halibut iHippoglossus stenolepis). Can. J. Fish. Aquat. Sci. 56:242-252. Conners, M. E., A. B. Hollowed, and E. Brown. 2002. Retrospective analysis of Bering Sea bottom trawl surveys: regime shift and ecosystem reorgani- zation. Progr. Oceanogr. 55:209-222. Hare. S. R., and N. J. Mantua. 2000. Empirical evidence for North Pacific regime shifts in 1977 and 1989. Progr. Oceanogr. 47:103-146. Hill, M. O. 1973. Diversity and evenness. A unifying notation and its consequences. Ecology 54:427-432. Hollowed, A. B., S. R. Hare, and W. S. Wooster. 2001. Pacific basin climate variability and patterns of northeast Pacific marine fish production. Prog. Ocean- ogr. 49:257-282. Hunt, G. L. Jr.. and P. J. Stabeno. 2002. Climate change and the control of energy flow in the southeastern Bering Sea. Prog. Oceaogr. 55:5-22. Hunt, G. L., P. Stabeno, G. Walters, E. Sinclair, R. D. Brodeur, J. M. Napp, and N. A. Bond. 2002. Climate change and control of the southeast- ern Bering Sea pelagic ecosystem. Deep-Sea Res. H 49:5821-5853. Hoff: Biodiversity as an index of regime shift in the eastern Bering Sea 237 Kinder, T. H., and J. D. Schumacher. 1981. Hydrographic structure over the continental shelf of the southeastern Bering Sea. In The eastern Bering Sea shelf: oceanography and resources, vol. 1 (D. W. Hood and J. A. Calder, eds.), p. 31-52. U.S. Govern- ment Printing Office, Washington D.C. Ludwig, J. A., and J. F. Reynolds. 1988. Statistical ecology: a primer on methods and com- puting, 337 p. John Wiley and Sons, New York, NY. Minobe, S. 2002. Interannual to interdecadal changes in the Bering Sea and concurrent 1998/99 changes over the North Pacific. Progr. Oceanogr. 55:45-64. Neter, J., M. Kutner, C. J. Nachtsheim, and W. Wasserman. 1996. Applied linear regression models, 3'"'' ed., 720 p. McGraw-Hill Com, Chicago, IL. Rose, C. S., and G. E. Walters. 1990. Trawl width variation during bottom trawl surveys: causes and consequences. In Proceedings of the sym- posium on application of stock assessment techniques to gadids (L-L. Low ed.), p. 57-67. Int. North Pac. Fish. Comm. Bull. 50. Rosenzwg, M. L., and Z. Abramsky. 1993. How are diversity and productivity related? In Species diversity in ecological communities historical and geographical perspectives (R.E. Rickles and D. Schulter, eds.), p. 52-65. Univ. Chicago Press, Chicago, IL. Schumacher, J. D., and J. P. Stabeno. 1998. Continental shelf of the Bering Sea. /;; The sea. vol. XI, The global coastal ocean: regional studies and synthesis (A.R. Robinson and K.H. Brink, eds.), p. 789-822. John Wiley, Inc., New York, NY. Smith, G. B., and R. G. Bakkala. 1982. Demersal fish resources of the eastern Bering Sea: spring 1976. NCAA Tech. Rep. NMFS SSRF-754, 129 p. Stabeno, P. J.. N. A Bond, N. B. Kachel. S. A. Salo. and J. D. Schumacher. 2001. On the temporal variability of the physical envi- ronment over the south-eastern Bering Sea. Fish. Oceanogr. 10ll):81-98. Sugimoto T., S. Kimura, and K. Tadokora. 2001. Impact of El Nino events and climate regime shift on living resources in the western North Pacific. Progr. Oceanogr. 49:113-127. Waide, R. B, M. R. Willig, C. R Steiner, G. Mittelbach, L. Gough, S. I. Dodson. G. P. Juday, and R. Parmenter. 1999. The relationship between productivity and species richness. Annu. Rev. Ecol. Syst. 30:257-300. Wilderbuer, T. K., A. B. Hollowed, W. J. Ingraham Jr., P. O. Spen- cer, M. E. Conners, N. A Bond, and G. E. Walters. 2002. Flatfish recruitment response to decadal climactic variability and ocean condition in the eastern Bering Sea. Prog. Oceanogr. 55:235-247. Zhang, C. I., J. B. Lee, S. Kim, and J. H. Oh. 2000. Climactic regime shifts and their impacts on marine ecosystem and fisheries resources in Korean waters. Progr. Oceanogr. 47:171-190. 238 Abstract — A new species of the cottid genus Triglops Reinhardt is described on the basis of 21 specimens collected in Aniva Bay, southern Sakhalin Island, Russia, and off Kitami, on the northern coast of Hokkaido, Japan, at depths of 73-117 m. Of the ten spe- cies of Triglops now recognized, the new species, Triglops dorothy, is most similar to T. pingeli Reinhardt, well known from the North Atlantic and North Pacific oceans and through- out coastal waters of the Arctic. The new species differs from T. pingeli in a combination of morphometric and meristic characters that includes most importantly the number of dorso- lateral scales; the number of oblique, scaled dermal folds below the lateral line; and the number of gill rakers. Triglops dorothy, a new species of sculpin (Teleostei: Scorpaeniformes: Cottidae) from the southern Sea of Okhotsk Theodore W. Pietsch School of Aquatic and Fishery Sciences College of Ocean and Fishery Sciences University of Washington Campus Box 355020 Seattle, Washington 98f95-5020 E-mail address: twp'g'u. Washington edu James W. Orr Resource Assessment and Conservation Engineering Division Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 7600 Sand Point Way NE Seattle, Washington 98115-6349 Manuscript submitted 28 March 2005 to the Scientific Editor. Manuscript approved for publication 16 August by the Scientific Editor. Fish. Bull. 104:238-246 (2006). Sculpins of the teleost family Cottidae are nearly ubiquitous in cold-water benthic habitats of the Northern Hemi- sphere, comprising nearly 200 species in the North Pacific alone (Yabe and Nakabo, 1984; Sheiko and Federov, 2000; Mecklenburg et al., 2002; Love et al., 2005), where they are found in almost every benthic habitat from the intertidal to the upper continental slope. Many species are preyed upon by larger fishes and marine mammals (Browne et al., 2002), and are them- selves predators primarily of smaller fishes and crustaceans (e.g., Tokranov, 1998; Hoff, 2000; Tokranov and Orlov, 2001). Although many members of the family are also commonly found in bycatch of commercial fisheries (Stevenson, 2004; Orlov, 2005), the systematics and life histories of most species are poorly known (Nelson, 1994; Hoff, 2000; Hoff, 2006) and new species continue to be described (e.g., Yabe, 1995; Yabe and Maruyama, 2001; Yabe et al., 2001). A more com- plete understanding of the diversity of the family is necessary to understand the role of cottids in the dynamics of North Pacific ecosystems. The cottid genus Triglops Rein- hardt (1830), including Prionistius Bean (1884), Elanura Gilbert (1896), and Sternias Jordan and Evermann (1898) as junior synonyms, contains 19 nominal species and subspecies, of which nine are currently recognized as valid (Pietsch, 1993). Members of the genus are characterized most strikingly by having a small head, a narrow, elongate body and a slender caudal peduncle, a long anal fin con- taining 18-32 rays, pelvic fins with a single spine and three soft rays, branchiostegal membranes united on the ventral midline but lying free from the isthmus, and scales below the lateral line modified to form dis- crete rows of tiny serrated plates that lie in close-set, oblique dermal folds. The species are generally distributed in rather deep water (primarily be- tween 18 and 600 m, but specimens have been taken more or less on the surface and as deep as 930 m; An- driashev, 1949; Hart, 1973; Fedorov. 1986) throughout cold-water, conti- nental shelf or slope regions of the North Pacific, North Atlantic, and Arctic oceans. All appear to feed pri- marily on small planktonic and ben- thic invertebrates (Andriashev, 1949; Fedorov, 1986). Spawning takes place from late summer to winter; the eggs are demersal (Andriashev, 1949; Mu- sick and Able, 1969; Fedorov, 1986). Triglops was first proposed by Jo- hannes Reinhardt (1830), based on a single specimen from West Green- land, but the type species T pingeli Pietsch and Orr: Tnglops dorothy. a new species from the southern Sea of OI100 km) (Krouse, 1980; Campbell, 1983). Size-dependent inshore-offshore distributions have been reported in many lobster studies (Skud, 1969; Cooper et al., 1975; Briggs, 1985; Campbell and Pe- zzack, 1986; Cobb et al., 1989; Harding and Trites, 1989). Information collected in commercial fisheries and scientific surveys has revealed that the spatial distri- bution of the American lobster may be size-dependent and that lobster of large sizes are more likely found in deep waters. This pattern has been identified but not quantified (Wilson, 1998). The spatial distribution of lobsters varies with seasons and scales with the size of lobsters. The distribution and abundance of lobsters have been studied with a variety of techniques. Diver and subma- rine and ROV surveys are generally limited to nearshore areas or in temporal resolution. Tagging programs are often limited in area and time covered and in number of lobsters tagged and returned. A fisheries-dependent sea sampling program covers a limited number of fish- ing boats within limited areas and thus may not be able to provide an overall picture of lobster distribution and abundance within the lobster stock area in the Gulf of Maine. Until recently, the only comprehensive fishery- independent data were those obtained from the NMFS Gulf of Maine fall and spring trawl surveys. NMFS has conducted a randomized stratified survey since the mid 1960s, but this survey is limited to depths deeper than 50 meters. In fact, the majority of the sampling stations are located in waters deeper than 120 meters as a result of an increased proportion of untrawlable areas and problems with lobster gear in inshore waters. In the fall of 2000, the DMR began a coast-wide inshore trawl survey, which mainly covered waters shallower than 120 meters (Fig. 1). The spatial distribution of lobsters is largely restrict- ed to the nearshore areas. Although found throughout the Gulf of Maine, 80% of the landings are estimated to come within three miles from shore (ASMFC, 2000). Clearly the lobster fishery will follow the concentrations of lobsters, although seasonal changes in movements and trapability of lobsters may make less productive fishing areas desirable to limited portions of the fleet at certain times of the year. As the fishing effort for lobster has increased, the traditional inshore fishery Chen et al,: Population structure of Homarus americanus in the Gulf of Maine 249 has expanded to nearshore waters (from 4.8 km to 32.2 km from shore). There is also a deepwater fishery for lobster that occurs farther from shore. An optimal and effective fisheries management plan requires a high-quality stock assessment, which, in turn, is dependent on data collected from fisheries- dependent and fisheries-independent survey programs. One of the most important pieces of information used in a lobster stock assessment is the abundance index which is derived from the fisheries-independent bottom trawl surveys conducted by NMFS. Because trawling is difficult in coastal waters, few sampling stations in the NMFS survey are located inshore, specifically along coastal Maine from which the majority of the lobster landings are landed. Previous studies have also indi- cated that inshore habitats are critical to the lobster fishery. Thus, it is critical to develop an inshore survey program that can cover the waters off coastal Maine. The data collected from such a program can overcome the problem of a lack of inshore coverage in the NMFS survey. An inshore survey and the NMFS survey can complement each other for a more complete abundance index of the lobster stock in the Gulf of Maine. In order to identify the importance of the DMR in- shore survey, we compared the differences between the NMFS survey and DMR inshore survey in 1) the size composition of survey catches and their temporal trends; 2) temporal trends in abundance indices; and 3) average size and weight of lobsters. From our comparative study, we determined whether the lobster stock in the Gulf of Maine has a size-dependent inshore-offshore distribution and whether it is essential to include both sampling pro- grams in a stock assessment to adequately describe the population dynamics of lobsters in the Gulf of Maine. Methods and materials Like the NMFS surveys, the inshore trawl survey is conducted during the spring and fall of each year. It has a stratified random design modeled after the NMFS and Massachusetts Department of Marine Fisheries (MADMF) surveys. The design has four depth strata (9-37 m, 37-64 m, 64-100 m, >100 m [its outer boundary roughly delin- eated by the 12-mile or 22-km limit]) and five regions based on oceanographic, geologic, and biological features. The fourth stratum was added in the spring of 2003. It expands the coverage area to equal the area in federal waters covered by the Atlantic States Marine Fisheries Commission (ASMFC) and allows some overlaps between annual inshore trawl survey area and the NMFS survey area. It also slightly reduces the sampling pressure in the shallower strata, which has been of concern to fixed- gear fishermen in the past. To randomize the survey area (-13,720 km-), each depth stratum is divided into 1-nmi (i.e., 1.852 km) sampling grids. A target of 100 stations is selected for sampl-ng in each survey, resulting in a sampling density of about 1 station /137 km-. This density compares to the sampling density of the NMFS survey (1 station/892 km-) and Massachusetts' survey (1 station/ 65 km'-). The number of stations per stratum is allocated in proportion to the area of each stratum. 'When a station is encountered that cannot be towed, an alternate tow is selected nearby over similar depth. Trawl design considerations for the survey included effectiveness of the gear for sampling the complex bot- tom in the Gulf of Maine and approximate comparabil- ity with previous and ongoing surveys, such as that performed by the NMFS. Net tapers are cut to permit the shape of the net to be of maximum height, while allowing the net to remain tight on the bottom. The net is shackled from the footrope to the frame by using two 0.95-cm shackles on a banded wire that runs parallel with the footrope. Heavy rubber wing bobbins retard bottom wing lift. The top leg is constructed of 5.1-cm mesh overall and has a 1.3-cm mesh liner in the codend. Doors are 7.5 Bison doors. Attached to the 21.3-m-long, 1.59-wide footrope is a roller frame. The 3.05-m-wide bosom section is made up of 20.3-cm rubber disks on 15.2-cm centers, and there are eight evenly spaced toggles. The spacing is maintained by small 10.2-cm cookies strung between the disks. Chain sweeps were not used. The headrope is 17.4 meter in length. Environmental data, including temperature and salin- ity profiles, wind, sea state, and weather, were collected at each station. A standard trawl tow, 20 minutes in duration, was made at each station. Shorter tow times were accepted under certain circumstances, such as the presence of fixed gears or untrawlable bottom in the survey pathway. Tow speed was maintained at 2.1 to 2.3 knots (i.e., 3.9 km/h to 4.3 km/h) and tow direction was oriented toward the tidal current whenever possible. All sampling was conducted during the day. After each tow, the net was brought aboard and emptied onto a sort- ing table. All individuals were identified and sorted by species. All lobsters were immediately separated and processed while the rest of the catch was sorted. Total weights (by sex), carapace length (mm), shell condition, presence and stage of eggs, V-notch condition, and inju- ries to the lobsters caused by the trawl were recorded. All lobsters were measured and the data were recorded in electronic format for analysis and made available on compact disk (CD). The data were geo-referenced and incorporated into geographic information systems (GISj for analysis. A no. 36 Yankee bottom trawl has generally been used in NMFS bottom trawl surveys. The trawl net is towed at approximately 3.5 knots (i.e., 6.5 km/h) for 30 minutes at each station. The survey is based on a stratified random design. Strata are defined ac- cording to water depth, latitude, and historical fishing patterns. Within each stratum, stations are assigned randomly; the number of stations allotted to a stratum is in proportion to its area (approximately one station per 892 km'^). Specifications for NEFSC standard no. 36 Yankee bottom trawl are defined by the NMFS.' ^ NMFS (National Marine Fisheries Service). Website: http:// www.nefsc.noaa. go v/femad/ecosurvey/mainpage. [Accessed August 2004.] 250 Fishery Bulletin 104(2) The trawl used in the DMR survey is similar in de- sign, but smaller than that used in NMFS, although the size of the mesh liner in the codend is the same. The DMR survey trawl net is also towed for less time (20 minutes versus 30 minutes for NMFS). The data collected in the inshore survey were ana- lyzed according to the methods described below. The stratified mean number of lobsters per tow, X. can be calculated as The stratified mean number of lobsters of length j per tow, X,, can be estimated as where the mean number of lobsters of size j per tow, X I,, is estimated as where g = the number of stratum in the survey; A^ = the survey area in stratum k\ ^ A = the total area covered by the survey (A='^A^ ); and *=' Xj. = the average catch per tow in stratum k. The X, can be calculated as X IkX .iJ.k l.k - The above equations were used for both the original data (without any transformation) and the data trans- formed by using log(.r-i-l). For results derived from the log (.t+l) transformed data, a retransformation was done to retransform the log-based results to the results of the original scale by using the following equation: X, -tX,.n 'k /=1 where X/, , = catch in tow t, stratum k; and T,, = the total number of tow in stratum K. The variance for the stratified mean, S'-(X), can be esti- mated as S~{X) = ^J^A;S'{X,J, A~ i=i where S.,(^,,) = the variance for the mean number per tow in stratum k, which can be estimated as S2(X, ) = -'=! - l100 m). na=not | available. Year Season Mean carapace length (mm) Average depth iml I II III IV I II III IV 2000 fall 61.9 64.4 80.0 na 28.6 .57.9 89.2 na 2001 fall 63.6 63.0 78.0 na 28.2 53.7 88.8 na 2002 fall 63.6 69.9 84.4 na 29.8 57.1 93.0 na 2001 spring 66.1 63.7 79.9 na 29.5 59.2 89.0 na 2002 spring 61.5 61.1 78.6 na 29.7 56.4 89.7 na 2003 spring 68.6 64.0 74.9 99.5 29.3 .59.0 S7.7 125.4 E 100 ra 20 00 (IN) 01 (IN) 02 (IN) 01 (OFF) 02 (OFF) Year (survey program) Figure 3 The stratified mean carapace length (CL) for American lobsters tHomarus amei'icaniis) for each sampling season and year for the Maine Depart- ment of Marine Resources inshore (IN I and the National Marine Fisheries Service offshore (OFF) survey programs. 00 = 2000; 01 = 2001; 02 = 2002. either fall or spring survey) for the inshore survey (Fig. 4). For the NMFS survey, however, significant dif- ferences were found in size distributions between 2001 and 2002 in the fall surveys, but not between the two years in the spring surveys. The seasonal differences (i.e., spring vs. fall) in size distribution within a year were much larger than the between-year differences for the inshore fall or spring survey. For the NMFS survey, however, the seasonal differences in size distributions were large in 2001, but small in 2002 (Fig. 4). The large seasonal differences in a year compared to the between- year differences in a season in size distributions for the inshore survey indicate that seasonal factors are impor- tant in determining the size distribution of lobsters in the inshore waters. Such seasonal variability was not so clear for the offshore waters. Large differences in size distributions were observed between the DMR inshore and NMFS surveys; the NMFS surveys consisted of lobsters of much larger sizes than those sampled inshore (Fig. 4). This finding may indicate that the inshore survey program had a limited coverage of the lobsters of large sizes, whereas the NMFS survey program had a limited coverage of the lobsters of small sizes. The size composition of lob- sters from the NMFS survey had large variations. This result was probably due to the small number of lobsters caught in the NMFS survey. An increase in sample size could make the size composition curve smoother, and thus better defined. Compared with size composition of Amercan lobsters from the NMFS survey, size composi- tion for the inshore survey was better defined, probably as a result of the larger number of lobsters caught in the inshore survey. The stratified mean sizes of lobsters were similar among years for the same sampling season in the in- shore survey (Fig. 3). For 2001, the stratified mean size in the fall survey was almost the same as that in the spring survey. For 2002, however, the stratified mean size in the fall survey was about 7 mm larger than that in the spring survey (Fig. 3). The stratified mean size of the 2001 fall survey was about 16 mm smaller than that of the 2001 spring survey (Fig. 3). The stratified mean size of lobsters in the NMFS survey was much larger than that of the inshore survey. This reconfirms the re- sults derived from the comparisons of size distributions between the inshore and NMFS surveys (Fig. 4). The abundance index derived from the inshore survey program revealed a consistent temporal pattern with original data or retransformed data (Table 2). For the original data and retransformed data, the fall inshore survey abundance index was the highest in 2001, fol- lowed closely by 2002, and the abundance index in 2000 was the lowest. For the log-transformed data, however, the differences among the three years were small (Table 2). For the spring survey, the abundance index in 2002 was much higher than the abundance indices in 2001 and 2003, and the abundance indices in 2001 and 2003 were similar (Table 2). The delta mean abundance indi- ces of the NMFS survey program were higher in 2002 Chen et a\ Population structure of Homaius ameiicanus in the Gulf of Maine 253 ^ 0.00 50 100 150 200 50 100 150 200 0,06 0.04 0.02- 0.00 Inshore Spring 2001 Spring 2002 Spring 2003 vy- 50 100 ^^r — 150 200 Spring 2001 Spring 1 2002 50 100 150 200 CL (mm) Figure 4 The size lof carapace ICLIl distributions for American lobsters [Hoinariis ainericanus) for each sampling season and year for the Maine Department of Marine Resources inshore and the National Marine Fisheries Service offshore survey programs. Table 2 Estimates of the survey abund ance index obtai ned ft oni the original data and log I.V + llt ransformed data and the i etransformed | abundance index derived from log(.v+li transformed data. 'Lower B" and "uppe • B" are the lower and i pper bou idaries of the 95'''r confidence intervals. Index Statistics Fall Spring 2000 2001 2002 2001 2002 2003 Original (inshore) Mean 55.5 68.5 64.1 22.6 49.2 21.2 Lower B 32.7 45.1 51.7 16.4 34.7 15.8 Upper B 78.2 91.9 76.5 28.8 63.6 26.7 Log(.r+l) Mean 2.9 2.9 3.1 2.1 2.7 2.2 transformed l inshore I Lower B 2.6 2.5 2.8 1.9 2.4 1.8 Upper B 3.2 3.3 3.5 2.4 3.0 2.6 Retransformed (inshore) Mean 54,5 119.6 88.8 25.1 62.4 24.3 Lower B 39.5 81.5 62.4 19.6 45.9 16.1 Upper B 75.2 175.1 126.3 32.0 84.7 36.5 Delta mean (NMFS) Mean 1.52 2.67 1.63 2.53 than those in 2001 for both the fall and spring sur- veys. Thus, for fall, DMR inshore survey program had a different temporal pattern between 2001 and 2002, compared with the NMFS surveys, whereas for spring, the temporal pattern of 2001 and 2002 was same for the two sampling programs. Discussions Large differences were found in size compositions and mean sizes of lobsters between the DMR inshore and NMFS surveys. Such differences indicated that large lobsters are more likely to appear in the NMFS survey, 254 Fishery Bulletin 104(2) and small lobsters are more likely to appear in the inshore survey. This is true for both sampling seasons, fall and spring. Two possible hypotheses can be devel- oped to explain such patterns: one is that large differ- ences in size compositions exist in lobsters inhabiting the areas covered by the inshore and NMFS survey pro- grams, and the other is that the observed differences in size compositions between the DMR inshore and NMFS surveys result from differences in gear selectivity for lobsters of different sizes. The first hypothesis implies that lobsters have size-dependent inshore-offshore distri- bution, namely that large lobsters tend to inhabit deep waters and small lobsters inhabit shallow waters. The second hypothesis implies that although lobsters may have no size-dependent inshore-offshore distribution, the NMFS survey gears are selective for large lobsters and the inshore survey gears are selective for small lobsters in the population. Such a difference in gear selectivity between the two programs may result from differences between the two sampling programs not only in sam- pling gears but in towing speed and duration. In order to determine which hypothesis is more plau- sible, we evaluated size composition for a depth stratum covered by both the DMR inshore and NMFS surveys. In the 2003 spring inshore survey, the fourth depth stra- tum had an average depth of 125 m which overlapped depth ranges in the NMFS surveys. The average size of lobsters in the fourth depth stratum in the 2003 spring inshore survey was almost identical to the average size of lobsters in the NMFS survey (Table 1 and Fig. 3). This finding indicates that the inshore survey gears can, like the NMFS survey gears, catch large lobsters if they are present in the areas covered by the survey program. Likewise, the differences in the average size between the first three strata and the fourth stratum (Table 1) in the 2003 spring inshore survey are likely to result from a lack of large lobsters in the first three strata in the inshore surveys, rather than from gear selectivity. Thus, the differences in size composition of lobsters between the DMR and NMFS surveys are likely the result of the differences in the size composition of the lobster population between the areas covered by the inshore and NMFS surveys, rather than the result of sampling methods. Although our study could not exclude the impacts of possible differences in gear selectivity on the survey size compositions, the results of our study seem to support the first hypothesis that the lobsters have a size-dependent inshore-offshore distribution, where large lobsters are more likely to be found in deep waters and small lobsters are to be found in shallow inshore waters. Thus, the two sampling programs tend to cover different segments of the stock in the Gulf of Maine. In order to have an adequate representation of the lobster population, it is necessary to include data from both sampling programs to describe the lobster population dynamics in the Gulf of Maine. The temporal changes in the abundance index in the spring surveys were rather consistent between the DMR and NMFS surveys and for data on the original and log scales. However, for the fall survey, the delta mean of the NMFS survey was consistent only with the mean of log-transformed data, but not with the mean of the original data and the mean retransformed data from the mean of log-transformed data (Table 2). This find- ing raises an interesting question regarding the type of data transformation we should choose and the potential impacts of each type on detecting temporal changes in the population abundance. A different choice of data transformation methods may lead to different interpre- tations of temporal variations in stock abundance. If the size-dependent inshore-offshore distribution pattern is not taken into consideration in making stock assessments, the entire stock will not be managed. We may need a separate set of abundance indices and size compositions for the lobsters in the inshore and offshore waters to describe the population dynamics of lobsters of different size groups more accurately. The inclusion of only NMFS survey data or inshore survey data in the stock assessment may lead to errors in determining the status of the lobsters if the population dynamics of large lobsters is different from that of small lobsters. The blending of DMR and NMFS data into a single set of data may be a solution. However, whether such a set of data can describe the dynamics of the whole population depends upon whether the lobster population dynamics are consistent among different size classes and whether there is a large difference in gear selectivity and in catchability of the trawl gear. Large differences in the population dynamics of large and small lobsters and in the gear selectivity and catchability of DMR and NMFS trawls may make such blended data less desirable for describing the population dynamics with DMR-NMFS blended data. For example, a significant increase or decrease in the abundance of lobsters of prerecruiting sizes may not be well defined with the NMFS surveys or DMR-NMFS blended data. We recommend both sets of data from the DMR inshore and the NMFS surveys be used separately in the stock assessment to define the population dynamics of the lobsters in the Gulf of Maine more accurately. Many fish species in the world exhibit a size-depen- dent distribution pattern similar to that of the Ameri- can lobster in the Gulf of Maine. Their population abundance and size structure are often assessed by fisheries-independent survey programs. The informa- tion is then used in the assessment and management of these fish species. Because it is often difficult with large trawls to survey inshore areas where the majority of stocks are present, the fishery-independent survey may not provide adequate coverage of inshore areas that are productive and critical, in particular, to recruitment. Thus, it is highly likely that fish in inshore areas are not adequately represented in an offshore-focused sur- vey, and this misrepresentation would lead to errors in data on stock-size structure. Such an error may, in turn, result in large errors in a stock assessment if an age- or length-based stock assessment model is used. We suggest that two sampling programs with different spatial focuses may help identify the problems associ- ated with a sampling program that does encompass fish Chen et al,: Population structure of Homarus americanus in the Gulf of Maine 255 size-dependent inshore-offshore distribution (or a size- dependent distribution that is related to depth or other environmental variables). The two sampling programs can greatly improve the quality and quantity of data col- lected in a fishery-independent program, leading to the improved assessment and management of fisheries. In the case where only one survey program can be conduct- ed (e.g., due to budget limitation or time constraints, or both), we suggest that area-specific fishery landing data, together with depth or other environmental variables that influence fish distributions, be used in allocating sampling efforts in a random stratified survey. Acknowledgments We would like to thank R. Tetrault of T&R Fish Inc. for providing the vessels, Capt. Curtis Rice of the FV Robert Michael and Capt. Samuel Galli of the FV Tara Lynn for contributing to the project from the start. We also would like to thank V. Manfredi, J. Brown, K. Stepanek, and D. Grout for field assistance and data entry and prepara- tion. L. Jacobson and J. Idoine of the Northeast Fisheries Science Center provided the NMFS survey information, which we appreciate greatly. The financial support of this study was provided by the NMFS. Maine Depart- ment of Marine Resources, Northeast Consortium, and Maine Sea Grant Program. Literature cited ASMFC (Atlantic State Marine Fisheries Commission). 2000. The American lobster stock assessment, 532 p. ASMFC, Washington, D.C. Briggs, P. T. 1985. Movement of the American lobster off the south shore of Long Island, New York. New York Fish Game 32:20-25. Caddy, J. F. 1975. Spatial model for an exploited shellfish popula- tion, and its application to the Georges Bank scallop fishery. J. Fish. Res. Board Can. 32:1305-1328. Campbell, A. 1983. Growth of tagged lobsters off Port Maitland, Homa- rus americanus, in the Bay of Fundy. Can. J. Fish. Aquat. Sci. 40:1667-1675. Campbell, A., and D. S. Pezzack 1986. Relative egg production and abundance of berried lobsters, Homarus americanus, in the Bay of Fundy and off Southweatern Nova Scotia. Can. J. Fish. Aquat. Sci. 43:2190-2196. Chen, Y., G. Liggins, K, J. Graham, and S. J. Kennelly. 1997. Modelling length-dependent offshore distribution of redfish, Centroberyx affinis. Fish. Res. 29:39-54. Cobb, J. S., D. Wang, D. B. Campbell, and P. Rooney. 1989. Speed and direction of swimming by postlar- vae of the American lobster. Trans. Am. Fish. Soc. 8:82-86. Cooper, R. A., R. A. Clifford, and C. D. Newell. 1975. Seasonal abundance of the American lobster, Homa- rus americanus, in the Boothbay region of Maine. Trans. Am. Fish. Soc. 104:669-674 Cooper, R. A., and J. R. Uzmann. 1980. Ecology of juvenile and adult Homarus. In The biology and management of lobsters, vol. 2. (J. S. Cobb and R. Phillips, eds. I, p. 97-141. Academic Press, New York, NY. Harding, G. C, and R. W. Trites. 1989. Dispersal of Homarus americanus larvae in the Gulf of Maine from Brown Bank. Can. J. Fish. Aquat. Sci. 46:1077-1078. Hilborn, R., and C. Walters. 1992. Quantitative fisheries stock assessment: choice, dynamics, and uncertainty, 570 p. Chapman and Hall, New York, NY. Krouse, J. S. 1980. Summary of lobster, Homarus americanus, tagging studies in American waters (1898-1978). Can. Tech. Rep. Fish. Aquat. Sci. 932:135-151. Lawton, P., and K. L. Lavalli. 1995. Postlarval, juvenile, adolescent and adult ecol- ogy. In Biology of the lobster Homarus americanus (J. F. Factor, ed.), p. 47-88. Academic Press, New York, NY. Palma, A. T, R. S. Steneck, and J. Wilson. 1999. Settlement-driven, multiscale demographic patterns of large benthic decapods in the Gulf of Maine. J. Exp. Mar. Biol. Ecol. 241:107-136. Paloheimo, J. E., and L. M. Dickie. 1964. Abundance and fishing success. Rapp. P. V. Reun. Cons. Int. Explor. Mer 1964:152-163. Pezzack, D. S., and D. R. Duggan. 1986. Evidence of migration and homing of lobsters on the Scotian shelf Can. J. Fish. Aquat. Sci. 43: 2206-2211. Rowling, K. 1994. Tiger flathead, Neoplatycephalus richardsoni. In The South East fishery (R. D. J. Tilzey, ed.), p. 124-136. Bureau of Resource Sciences, Australian Gov. Print Serv., Canberra, Australia. Skud, B. E. 1969. The effect of fishing on size composition and sex ratio of offshore lobster stocks. Fish Dir Skr. Ser. HavUnders:15:295-309. Taylor, C. C. 1953. Nature of variability in trawl catches. Fish. Bull. 54:145-166. Wahle, R. A. 1992. Body-size dependent anti-predator mechanisms of the American lobster. Oikos 65:52-60. Wahle, R. A., and R. S. Steneck. 1992. Habitat restrictions in early benthic life: experi- ments on habitat selection and in situ predation with the American lobster. J. Exp. Mar. Biol. Ecol. 157:91-114. Wilson, C. J. 1998. Bathymetric and spatial patterns of settlement in American lobsters, H. americanus, in the Gulf of Maine: Insights into processes controlling abundance. M.S. thesis, 37 p. Univ. Maine, Orono, ME. 256 Abstract — Patterns were investigated in juvenile fish use of unconsolidated sediments on the southeast United States continental shelf off Georgia. Juvenile fish and environmental data were sampled at ten stations along a 110-km cross-shelf transect, including four stations surrounding Gray's Reef National Marine Sanctuary (Gray's Reef NMFS). Cross-shelf stations were sampled approximately quar- terly from spring 2000 to winter 2002. Additional stations were sampled on three transects inshore of Gray's Reef NMS and four transects offshore of the Sanctuary during three cruises to investigate along-shelf patterns in the juvenile fish assemblages. Sam- ples were collected in beam trawls, and 121 juvenile taxa, of which 33 were reef-associated species, were identified. Correspondence analy- sis on untransformed juvenile fish abundance indicated a cross-shelf gradient in assemblages, and the station groupings and assemblages varied seasonally. During the spring, fall, and winter, three cross-shelf regions were identified: inner-shelf, mid-shelf, and outer-shelf regions. In the summer, the shelf consisted of a single juvenile fi.sh assemblage. Water depth was the primary environmental variable correlated with cross-shelf assemblages. However, salinity, den- sity, and water column stratification also correlated with the distribution of assemblages during the spring, fall, and winter, and along with tempera- ture likely influenced the distribu- tion of juvenile fish. No along-shelf spatial patterns were found in the juvenile fish assemblages, but the along-shelf dimension sampled was small (-60 km). Our results revealed that a number of commercially and recreationally important species used unconsolidated sediments on the shelf off Georgia as juvenile habitat. We conclude that management efforts would be improved through a greater recognition of the importance of these habitats to fish production and the interconnectedness of multiple habi- tats in the southeast U.S. continental shelf ecosystem. Manuscript submitted 26 October 2004 to the Scientific Editor's Office. Manuscript approved for publication 25 August 2005 by the Scientific Editor. Fish. Bull. 104:256-277(20061. Juvenile fish assemblages collected on unconsolidated sediments of the southeast United States continental shelf Harvey J. Walsh Katrin E. Marancik Jonathan A. Hare NOAA, NOS. NCCOS Center for Coastal Fisheries and Habitat Research 101 Pivers Island Road Beaufort, North Carolina 28516 Present address (for H J, Walsh): Biology Department MS #50 Woods Hole Oceanographic Institution Woods Hole, Massachusetts 02543 E-mail address (tor H J Walsh) hwalshM'whoi edu Fisheries management has been rap- idly evolving for the southeast United States with the development of a formal stock assessment process (i.e., SEDARM, and the identification of essential fish habitat (EFH; SAFMC, 1998). Single-species management plans have been used as the basis for fisheries management along the southeast United States for decades, and EFH is just beginning to be incor- porated into management plans (e.g., Sar-gassum and Oculina bank habitat of particular concern: FR, 2000). Glob- ally, as fish stocks continue to decline (Botsford et al., 1997) and harvest shifts to new species (Pauly et al., 1998), interest in ecosystem manage- ment and implementation of MPAs in many areas, including the south- east U.S. continental shelf, is growing (SAFMC, 2001). Understanding how juvenile fish use habitat is important for both single-species fishery man- agement and ecosystem approaches to fishery management. A description of habitat needs at each life history stage for species that represent the "significant food web" are also a recog- nized part of fishery ecosystem plans (N0AA2; CFEPTAP, 2004). Adult population variability, in most species, is caused by variability in the survival of the early-life stages; egg, larval, and juvenile (Sissenwine, 1984; Rothschild, 1986). Survival of these early stages is influenced by both habitat quantity and quality (Gibson, 1994; Peterson, 2003). One approach to improving marine fish- eries management is to incorporate habitat effects on early-life stage sur- vival into stock assessments (Beck et al., 2001; Peterson, 2003). Deter- mining how juvenile fish use habitat also benefits marine protected area (MPA) design and implementation. MPAs have been proposed to comple- ment traditional fishery management practices by imposing site-specific rules to protect enclosed resources (Parrish, 1999; Dayton et al., 2000; Beck and Odaya, 2001). In some in- stances, MPAs provide benefits to outside areas through spillover effects (Roberts et al, 2001). MPAs may also provide a source of recruits to other nonprotected areas through plank- tonic transport (Cowen et al., 2000; Shanks et al., 2003). To be effective in protecting fishery resources, an MPA or a network of MPAs needs to encompass the habitats used by spe- ' SEDAR (Southeast Data Assessment and Review). 2002. Southeast Data Assessment and Review (SEDAR). South Atlantic Fishery Management Council, Charleston, SC. Website: http://www. sefsc.noaa.gov/sedar.jsp [Accessed on 19 October 2004.1 -NOAA (National Oceanic and Atmo- spheric Administration). 1999. Eco- system-based fishery management. A report to Congress by the Ecosystems Principles Advisory Panel. lAvailable from National Technical Information Ser- vice, U.S. Department of Commerce, 5285 Port Royal Road, Springfield, VA.] Walsh et al,: Juvenile fish assemblages on the southeast United States continental shelf 257 cies throughout their life cycle (Browman and Konstan- tinos, 2004). Habitat use by juvenile fish has been studied exten- sively along the southeast United States, particularly for commercial and recreational species and structured habitats. For example, there is information available regarding the habitat use of specific estuarine (e.g., Cynoscion nebulosus: Thayer et al., 1999; paralichthid flounders: Walsh et al., 1999) and reef-associated spe- cies (e.g., Mycteroperca microlepis: Ross and Moser, 1995; Lutjanus griseus: Chester and Thayer, 1990; Centropristis striata: Lehnert and Allen, 2002). The importance of structural habitat to juvenile fish is also demonstrated by the consistent use of seagrass (Thayer et al., 1999), oyster reef (Meyer and Townsend, 2000), pelagic Sargassum (Moser et al., 1998), mangrove (Thayer et al., 1987), marsh (Hettler. 1989), and rocky reefs (Lindeman et al., 2000). Unconsolidated sedi- ments in estuarine systems and from the surf-zone out into shallow coastal areas (<10 m) are also important juvenile habitat (estuarine: Burke et al., 1991; Walsh et al., 1999; surf-zone: Wenner and Sedberry, 1989; Ross and Lancaster, 2002). Despite this large body of research, little is know about the use of unconsolidated sediments on the inner-, mid-, and outer-shelves of the southeast U.S. continen- tal shelf. Unconsolidated sediment is not identified as EFH of the southeast U.S. (SAFMC, 1998), although it is identified in several fishery management plans (red drum, snapper-grouper, rock shrimp, royal red shrimp, coastal migratory pelagics, golden crab, spiny lobsters, and calico scallops). The only offshore habitats that currently have specific protection through their own fishery management plans are coral, coral reefs, live hard bottom, and pelagic Sargassum (Barnette-^; FR, 2000). Research in other shelf ecosystems has identi- fied unconsolidated sediments on the shelf as important juvenile habitat (Norcross et al., 1997; Steves et al., 2000; Sullivan et al., 2000; Johnson et al., 2001). Fur- ther, unconsolidated sediments on the southeast U.S. continental shelf cover a majority of the area (-60-80%; Parker et al., 1983), and likely serve as juvenile habi- tat for the more than 1200 fish species reported from the ecosystem (estimated from Kendall and Matarese, 1994). Thus, well-defined use of unconsolidated sedi- ment habitat by juvenile fish is needed for sediment habitat to be incorporated into future fishery resource management on the southeast U.S. continental shelf. Our purpose was to document juvenile fish habitat use and to examine the structure of juvenile fish as- semblages in unconsolidated sediments off the Georgia coast region of the southeast U.S. continental shelf Spe- cific objectives were 1) to provide a list of species that •* Barnette, M. C. 2001. A review of the fishing gear utilized within the Southeast Region and their potential impacts on essential fish habitat. NOAA Tech. Memo. NMFS-SEFSC- 449, 62 p. (Available from National Technical Information Service, U.S. Department of Commerce, 5285 Port Royal Road, Springfield, VA.] use unconsolidated sediments on the shelf as juvenile habitat; 2) to determine which reef-associated species use unconsolidated sediments as juvenile habitat; and 3) to describe the relationships between the juvenile as- semblages and environmental factors. The focus on reef fish was motivated largely by the emphasis to protect reef areas by using MPAs, as part of the management of the reef-associated snapper-grouper complex (SAFMC, 2001). If species of the reef-associated snapper-grouper complex use unconsolidated sediments during the juve- nile stage, then protection of these habitats should be incorporated into the larger MPA effort. Materials and methods Study site The continental shelf off the Georgia coast is the widest part (200 km) of the southeast shelf. The shelf in this region is gently sloping and comprises four depth zones (Menzel, 1993); inner-shelf (0-20 m); mid-shelf (20- 40 m), outer-shelf (40-70 m), and shelf-edge (70-200 m). Each zone has different physical dynamics (Atkinson et al., 1985), climate (Blanton et al., 2003), and larval fish assemblages (Marancik et al., 2005). Demersal habitat consists of unconsolidated sediments, primarily medium to coarse quartz and carbonate sands (Nelson et al., 1999) interspersed with rocky reefs (Parker et al., 1983). Reefs of the southeast U.S. shelf range from no relief patchy live-bottom communities, to high-relief ledges. On the shelf off Georgia, rocky-reefs cover about 30% of the bottom (Parker et al., 1983). Rocky-reef habitats have a high diversity of invertebrate (Wenner et al., 1983) and vertebrate faunas (Chester et al., 1984; Parker and Mays, 1998; Lindeman et al., 2000). Located within the continental shelf off Georgia is the 56-km- Gray's Reef National Marine Sanctuary (Gray's Reef NMS, Fig. 1). The sanctuary depth ranges from 18 to 22 m and the benthic macrohabitat is mainly (-75%) unconsolidated sand sediments interspersed with patchy live-bottom and moderate relief (<2 m) hard-bottom ledges (Parker et al., 1994). The adult reef fish community of the sanc- tuary is typical of other inner-shelf (<30 m) reefs of the South and North Carolina shelves (Chester et al., 1984; Parker et al., 1994). Collection of juvenile fish and environmental data Sampling of juvenile fish was conducted approximately quarterly from April 2000 through February 2002 (Table 1). Ten cross-shelf stations (stations 1-7), approxi- mately 18.5 km apart, were sampled during most cruises (Fig. 1). Stations were missed on some cruises owing to weather and equipment failure (Table 1). The cross-shelf transect was 110 km long and stations were sampled on the inner-, mid-, and outer-shelf (10 to 50 m water depth). To avoid sampling within Gray's Reef NMS, four stations were placed immediately adjacent to the four sides of the sanctuary (stations 2.1-2.4). Additional 258 Fishery Bulletin 104(2) 81 OW 32 ON - 31 ON 80'0'W -^-l- South Carolina =^^^ n Grays Reef NMS 1 Beam trawl stations | • Cross-Shell o Inshore n Otishore - - -Depth contour (m) I Savannah (4-? 25 50 100 Kilometers Georgia f) ''"^'V o O o o ' o o oj^' D J o o o O jj^^curiswick 20 m • 3 D •5, D a ^7 30m 50 m 200m 31 0 W 1 80 OW 80 OW 32"0'N - 31 "ON Figure 1 Maps of the study area showing cross-shelf beam trawl stations (black dots) at which juvenile fish and environmental sampling occurred quarterly from April 2000-February 2002 (see Table li. Additional stations were sampled inshore (circles) and offshore (squares) of Gray's Reef NMS isee Table li (A). Station groups used in correspondence analysis of the inshore (Bl and offshore (C) juvenile fish data sets are separated by solid lines and labeled. stations were sampled along three transects inshore of Gray's Reef NMS and along four offshore transects during three cruises (Fig. 1 and Table 1). The benthic macrohabitat of the cross-shelf stations was determined by using a remotely operated vehicle (ROV, Phantom S2). The ROV was deployed off the starboard side of the ship. Approximately 30 m of the RO'V tether was let out before the tether was attached to a 45-kg weight. The weight was lowered to 3 m above the bottom and the ship was allowed to drift. During April 2000, two 15-min drifts were made at eight of the ten cross-shelf stations. Because of inclement weather, no drifts were conducted at the inner- or outer-most stations. Video (HI-8 mm) was recorded of each drift. Concurrent fish and hydrographic measurements were taken to examine the relationship between juvenile fish assemblages and the environment. At each sta- tion, temperature, salinity, density, and water depth were measured from the waters surface to one meter above the bottom with a Seabird CTD (SBE19, Seabird Electronics, Inc., Bellevue, WA). Juvenile fish were col- lected at each station by using a 2-m beam trawl with a 6-mm mesh body and a 3-mm mesh tail bag. Two kilogram weights were added to each skid of a standard 2-m beam trawl (Kuipers, 1975) to ensure that the trawl stayed on the bottom. Three 5-min bottom tows were made at each station. Samples were sorted on deck and fish and invertebrates were preserved in 95% ethanol. Preparation of fish data All beam trawl samples were sorted and fish were iden- tified to the lowest possible taxonomic level by using previously published descriptions (e.g.. Able and Fahay, Walsh et al Juvenile fish assemblages on the southeast United States continental shelf 259 Table 1 Number o fbe> im trawl stations sampled off the coast of Georgia for j uvenile fish. Station groups correspond to the station loca- | tions in F igure 1. Year Month Dates Season Stat ion group C ross-shelf Inshore Offshore 2000 April June August October 24-27 19-22 15-17 03-07 spring summer summer fall 9 9 9 10 2001 Jan/Feb March 30-01 21-23 wnter winter 10 10 Apr/May 30-04 spring 10 11 June 04-09 summer 10 August 03-06 summer 10 September 07-09 fall 6 October 11-13 fall 10 2002 February 08-13 winter 10 10 10 1998; Munroe, 1998). Identification to species was not possible for all fish. Taxa were separated into larval, juvenile, and adult age classes according to life-history characteristics from the literature and standard length distributions (SL, mm) from the beam trawl collec- tions. Fish were classified as larvae if fin development was not complete (and defined as "larvae at settlement stage" when fin formation was complete and pigmenta- tion incomplete; fish were classified as juveniles when pigmentation was complete, and were classified as adults if they exceeded the reported minimum SL for sexual maturity. Because the mesh of the beam trawl is large enough to extrude larvae and less efficient at catch- ing large (>150 mm SL) fish (Kuipers, 1975), analyses in this study were conducted only on settlement and juvenile stages, which were combined and referred to as juveniles. Taxa were also classified as not reef-associ- ated, strongly reef-associated or weakly reef-associated based on habitat-use information (Chester et al., 1984; Humann, 1994; Parker et al., 1994; Parker and Mays, 1998; Lindeman et al., 2000). Taxa found predominantly on rocky-reefs (e.g., Cetitropiistis striata ) were classified as strongly reef-associated. Those that use rocky-reefs and other habitats (e.g., Diplectrum formosum) were classified as weakly reef-associated. Standard catch per unit of effort (CPUE. fish/5-min bottom tow) was calculated for each taxon. CPUE for each taxon at each station was the average of all rep- licate tows. Three juvenile stage data sets (i.e., cross- shelf, inshore, and offshore) were compiled by using data from the station groups (Table 1, Fig. 1): the cross-shelf data set consisted of data from stations one through seven (Fig. lA); the inshore data set consisted of data from stations predominately inside the 20-m isobath (Fig. IB), and the offshore data set consisted of date from stations >30 m in depth (Fig. IC). For statistical analyses, the three data sets were subdi- vided into two groups so that rare taxa would not great- ly influence the classification of assemblages. The first group consisted only of abundant taxa, and the second group included abundant and rare taxa. Cross-shelf and offshore data sets were subdivided into taxa that made up at least lQ'7c (abundant) and 1% (abundant and rare) of the collections at any one station. The inshore data set was subdivided at 5% and 1% levels. The data sets were further condensed by eliminating all taxa not identified to species level with the exception of Stenotomus sp. and Rypticus sp. which were probably single species. Several abundant taxa were excluded from the analysis because they were only identified to genus and have multiple spe- cies common in the area, and each species may differ in distribution: i.e., Etropus spp. (4 species), Prionotus spp. (14 species), Sphoeroides spp. (11 species), Microgobius spp. (3 species), and Bothus spp. (3 species). As a result, the numbers of taxa in the rare and abundant data sets used in analyses were as follows: cross-shelf 82 (rare) and 28 (abundant) (Table 2); inshore 28 and 13 (Table 3), and offshore 51 and 17 (Table 4). Preparation of environmental data Season and five environmental (or hydrographic) vari- ables were chosen in an attempt to explain variation in the juvenile fish data. Cruises were assigned to one of four seasons (Table 1) based on wind (Atkinson et al., 1985) and temperature regimes (Marancik et al., 2005). Environmental variables were calculated as in Marancik et al. (2005). Briefly, CTD measurements of temperature (°C) and salinity were used to derive density (a,, Kg/m'^). Bottom temperature, salinity, and density were used in the multivariate analyses with juvenile fish because benthic habitats were sampled and bottom values were 260 Fishery Bulletin 104(2) Table 2 Taxa collected by beam trawl during two years of sampling (April 2000-February 2002) at the cross- shelf stations (see Fig. lA) off the coast of Georgia, and used in correspond ence analysis. Mean and standard dev ation (SD) of catch per unit of effort | (CPUE) were ca culated for the entire data set. Taxa that constituted one o " ten percent of any one station were us ed in the analyses, and an asterisk next to a ta.xon indicates that it was in the W^i data set. Seasonal juvenile assemblages from corre- | spondence analy sis are shown (1= inner-shelf: I/M = inner-/mid-shelf M=mid- shelf; 0-outer-shelf; and S = entire shelf: S). Taxa in bold indicate reef-associated species. Family Taxon CPUE Seasonal juvenile assemblage Mean SD Spring Summer Fall Winter Elopidae Elopa saurus 0.010 0.097 S M Ophichthidae Ophichfhiis gomesii 0.003 0.056 S Ophichthtis ocellatus 0.021 0.155 I/M S M M Congridae Ariosoma balearicum* 0.16.3 0.479 I/M S M M Clupeidae Brevoortia tyrannus 0.019 0.134 M Etrumeus teres 0.006 0.080 M Sardinella niirita 0.003 0.056 S Engraulididae Anchoa hepsetus* 0.303 1.313 I/M S M Anchoa lamprotaena 0.013 0.237 M Argentinidae Argentina striata 0.006 0.080 0 Synodontidae Synodus foetens'^' 0.746 1.626 I/M,0 S M M Synodus poeyi 0.048 0.487 0 Trachinocephalus myops 0.103 0.388 0 s M,0 0 Gadidae Urophycis regia* 0.234 0.759 I/M M Ophidiidae Ophidian selenops* 2.18.5 4.844 I/M s M M Otophidium omostigmum* 0.364 1.190 0 s 0 M Batrachoididae Ponchtys plectrodnn 0.006 0.072 M Antennariidae Antennarius radiosus 0.006 0.080 M,0 Ogcocephalidae Halieutichthys aculeatus 0.019 0.137 0 s Ogcocephalus nasutus 0.024 0.150 I/M s I Exocoetidae Hem ira mph us bra silieusis 0.003 0.056 s Hirundichthys affinis 0.022 0.169 s M Syngnathidae Hippocampus erectus 0.052 0.233 I/M s M I,M,0 Syngnathus scovelli* 0.025 0.192 s I Syngn a th u s springeri 0.025 0.158 I/M s M Scorpaenidae Scorpaena dispar 0.019 0.161 s M Scorpaena plumieri* 0.080 0.332 I/M s M M Triglidae Bellator hrachycliir 0.006 0.074 s 0 Bellator inilitaris 0.006 0.073 0 s Prionotus carolinus* 1.047 2.970 I/M s M M Prionotus ophryas 0.003 0.056 s Prionotus scituhis 0.022 0.146 M Serranidae Centroprisfis ocyurus* 0.167 1.344 0 s 0 Centropristis striata 0.022 0.186 I/M s Diplectrum formosum* 1.556 3.659 0 s M,0 M Serraniculits pumilio 0.180 0.817 I/M s M M Serranus phoebe* 0.085 0.466 0 s 0 0 Priacanthidae Priacanthus arenatus 0.003 0.056 s Apogonidae Pristigenys alta 0.016 0.123 s M Apogon pseudomaculatiis 0.012 0.108 s Carangidae Caranx bartholoniaci 0.003 0.051 0 Chloroscombrus chrysurus 0.014 0.124 s M Decapterus macarellus 0.006 0.080 0 M Decapterus punctatus* 0.022 0.218 s M Trachurus lathanii 0.003 0.054 I/M Lutjanidae Lutjanus analis 0.003 0.056 s continued Walsh et a\ Juvenile fish assemblages on the southeast United States continental shelf 261 Table 2 (continued) Family Taxon CPUE Seasonal juveni e assemblage Mean SD Spring Summer Fall Winter Haemulidae Haemulon aurolineatum* 0.022 0.214 S 0 Sparidae Stenotomus sp.* 0.119 0.811 I/M S M M Lagodon rhomboides 0.016 0.149 M Sciaenidae Cynosion nothus 0.009 0.110 M Cynosion regalis 0.011 0.118 M Larimus fasciatus* 0.036 0.363 I Leiostomus xanthurus* 0.621 3.752 I Menticirrhus americanus 0.013 0.185 I Mugilidae Mugil cephalus* 0.051 0.469 O Mugil curema 0.010 0.099 M Labridae Halichoeres bivattatus 0.006 0.113 S Xyrichtys novacula 0.003 0.056 s Uranoscopidae Kathetostoma albigutta 0.003 0.056 M Dactyloscopidae Dactyloscopus moorei* 0.837 2.329 I/M s M M Blenniidae Hypleurochilus gemmatus 0.010 0.126 I/M s Parablennius marmoreus* 0.006 0.080 s Callionymidae Diplogrammus paiuiradiatus' 0.025 0.174 s Gobiidae loglossus calliuriis 0.003 0.047 O s Stromateidae Peprilus triacaritlius 0.038 0.218 I/M M Bothidae Bothus lunatuH 0.006 0.080 0 0 Bothus ocellatus 0.021 0.155 s M 0 Bothus robinsi* 0.497 1.643 0 s M 0 Paralichthyidae Ancylcopsetta quadroccllata 0.063 0.300 M Citharichthys macrops 0.096 0.475 I/M s M M Cyclopsetta fimbriata 0.031 0.184 0 s M Syacium papillosum* 0.042 0.282 I/M s 0 Soleidae Gymnachirus melas"^ 0.019 0.134 0 s Cynoglossidae Symphurus diomedeanus 0.010 0.126 0 Symphurus minor* 0.519 1.801 0 s 0 0 Symphurus parvus 0.016 0.125 0 s Symphurus plagiusa 0.006 0.113 s Symphurus urospilus* 0.324 0.806 I/M s M M Balistidae Aluterus schoepfi 0.006 0.076 s M Monocanthus hispidus* 0.504 1.424 I/M s M 0 Monocanthus setifer 0.003 0.047 0 Ostraciidae Lactophyrs quadricornis 0.005 0.068 s highly correlated with the average water column and surface variables (Marancik et al., 2005). Vertical strati- fication was estimated by using Simpson's stratification parameter (Simpson and James, 1985),

Syacium papillosum IS -- ■ T : : 1 80'0'W -3rO'N - 30°0'N Fall Inner- She It Assemblage (sla 1 1 - Lanmus fascialus Syngnathus scovelli Mid-Shelt Assmblaqe (sla 2-5) Ophidion selenops Diplectrum formosum ' Dactyloscopus moorei Synodus foetens Monocarithus hispidus Anchoa hepsetus Bothus robinsi Symphurus urospilus Pnonotus carolmus Serraniculus pumilio Anosoma balearicum Stenotomus sp. Scorpaena plumien Outer-Shelf Assemblage (sla 6-7) Diplectrum formosum ' Symphurus minor Otophidium omostigmum Haemulon aurolineatum Syacium papillosum 3ro'N - 30'0'N 80 O'W 31 O'W 80 O'W sro'N- 30^0'N Inner- Shelf Assemblage (sla 1-2) Winter Mid-Shelf Assemblage (sla 2-5) Ophidton selenops Pnonotus carolmus Dactyloscopus moorei Urophyas regia Synodus foetens Otophidium omostigmum , Symphurus urospilus ' Diplectrum fotmosum Serraniculus pumilio ■ Stenotomus sp Scorpaena plumien Anosoma baleancum Decapterus punctatus Outer-She It Assemblage (sta 4-7) Bothus robinsi t^ugil cephalus Symphurus minor Centropristis ocyurus Monocanthus hispidus Serranus phoebe •3rO'N ■30 O'N 81 O'W 80 0'\A/ Figure 4 Cross-shelf juvenile fish assemblages identified from correspondence analysis for spring (A), fall (B). and winter (C). Summer was not plotted because there was only one summer assemblage. Hatched circles enclose stations that grouped together each season. Ta.xa from the 10% cross-shelf data set, in order of abundance, for each assemblage are shown in the tables. Asterisks indicate species that were found in multiple assemblages. and was mainly influenced by depth; the shallower inner- and mid-shelf stations were separate from the outer-shelf stations. The second dimension was related to bottom salinity (Fig. 5C) and correlated with the separation of the inner-shelf station from the higher saline mid- and outer-shelf stations (Fig. 2). A Monte Carlo permutation test of the relationships between the assemblages and environmental data for the fall indi- cated that depth and bottom salinity were significantly related to the juvenile assemblages and cross-shelf sta- tion groups (Table 6). The inclusion of environmental variables in the analy- sis of winter juvenile assemblage showed a weak rela- tionship with the cross-shelf groups and juvenile assem- blages, as indicated by the CCA correlation (Table 5). The addition of the environmental data also changed the ordination of the stations (Fig. 5D); yet there was still a cross-shelf gradient. Along the first dimension, depth, bottom temperature, and stratification were re- lated to the juvenile assemblages and cross-shelf groups (Fig. 5D). The shallow inner- and mid-shelf groups were lower in bottom temperature and stratification, whereas bottom density was lowest on the inner- and outer-shelves and highest on the mid-shelf (Fig. 2). A Monte Carlo permutation test for the winter relation- ships indicated that depth and bottom density were significantly correlated with the fish assemblages and station groups (Table 6), although the bottom density did not appear to have any relation to the patterns described (Fig. 5D). 268 Fishery Bulletin 104(2) CM < O O B » " o^SAL STRAT ^'"1^ c 3 3 CD c\j < O O C ' Inner Outer I-* *' BDEN Mid ' -^ STEM DE> \ BSAL 0) CM < o o D / Outer BDEN / \ \ * /sSAt Dtp ' Mid \ STR/!T * \ * " ^-■%a\ * ' - 1 BTEHI Vtn V -.^ Inner ^ -^ 3 CD CCA1 Figure 5 Canonical correspondence analysis ordinations (portraying the first and second dimension scores) of the \Q% cross-shelf data set showing the correlations between environmental variables, species, and station groups each season; spring (A), summer (B), fall (C), and winter (Dl. The solid triangles mark the locations of taxa, and the polygons enclose the bound- ary of each station group with three or more stations (as in Fig. 3). If stations were not grouped, circles mark the loca- tions. The arrows depict the gradient of each environmental variable (temperature = btemp, salinity=bsal, density=bden, stratification = strat, and depth = dep). The dashed lines intersect at the origin of the plot. Analyses were conducted by using both juvenile abundance and environmental data. Seasonal patterns In juvenile fish assemblages Many of the species (>68'7r) were present on the shelf in more than one season, which resulted in a large overlap of the seasonal assemblages (ordination not shown), including 12 of the 23 reef-associated species (Table 2). The assemblage data could be explained in three dimen- sions (as determined by CA of the 10% cross-shelf data set [Table 5]) and portrayed seasonal and cross-shelf patterns. Along the first dimension, there was separa- tion between the winter inner-shelf assemblage and all other assemblages. The second and third dimensions portrayed a cross-shelf gradient in the spring, fall, and winter juvenile assemblage data. The summer assem- blage overlapped all the cross-shelf groups from the other seasons except the winter inner-shelf assemblage. High abundances of Larimus fasciatus in the fall and Leiostomus xanthuriis and Mugil cephalus in the winter (Table 2) caused these two seasons to be more dissimilar than were spring and summer, which were very similar. Many of the juveniles were found in the same cross-shelf assemblages during each season, whereas others shifted assemblages (Fig. 4). Relation between seasonal juvenile assemblages and environmental variables There appeared to be no relationship between the seasonal juvenile assemblages and environmental variables. The products of the CCA eigenvalues and species-environment correlations were extremely low (Table 5), and the CCA ordination of seasonal juvenile assemblages and environmental variables was dif- ferent from the CA ordination. Thus, over the entire year, environmental variables did not help explain among-season variation in the juvenile assemblage data. Along-shelf and cross-shelf patterns in the inshore juvenile fish data set The inshore data set was made up of estuarine, coastal, and open-shelf species (Table 3, Appendix). There was a cross-shelf gradient in the spring and winter (Fig. 6). The innermost station group (8-m) separated from the other station groups along the first dimension (Fig. 7A) in spring. Most taxa from the 8-m assemblage (5 of 7) were not collected in high enough abundance to be included in the cross-shelf data set (Tables 2 and 3). In winter there was also a gradient from shallow to deeper stations (Fig. 6C). The assemblage nearest the origin in the winter (8-m assemblage) included estuarine species that were abundant at all the inshore stations (Table 3, Appendix). The coastal species, however, were more abundant at the deeper (12-m, 15-m, and 18-m) station groups (Table 3). There was no pattern in the along-shelf transects, from north to south, during either season (Fig. 6, B and D). Walsh et al Juvenile fish assemblages on the southeast United States continental shelf 269 Along-shelf and cross-shelf patterns in the offshore juvenile fish data set The offshore juvenile fish data set included primarily (76'7f) coastal, open-shelf, and slope species (Table 4, Appendix) and patterns in the data set varied with season. During the fall, there were both cross-shelf (Fig. 7A) and along-shelf patterns (Fig. 7B). Along the first dimen- sion, there was a cross-shelf gra- dient (Fig. 7A) from the shallow stations (30-45 m depth; Fig. IC) to the 50-m station group. The juve- nile assemblage at the 50-m sta- tion group had high abundances of reef-associated species (Table 4). The stations on the northern most transect (Fig. IC) also separated from the others along the second dimension, and had higher abun- dances of reef-associated species. The cross-shelf pattern was present in the winter, with greater abun- dances of reef-associated species at the 50-m station group (Fig. 7C), but there was no along-shelf pat- tern (Fig. 7D). Discussion Cross-shelf Along-shelf CA 1 CA 1 One hundred and twenty-one taxa of juvenile fish were collected on unconsolidated sediments on the continental shelf off Georgia. The actual number of species was higher, but taxonomic problems limited species identification (e.g., identification of Etropus spp., Pri- onotus spp., Sphoeroides spp., and Microgobius spp.). With concurrent ichthyoplankton sampling of the cross-shelf stations (Fig. 1), we collected 161 taxa (Marancik et al., 2005); the larval community had more pelagic species, including several scombrids (e.g., Auxis rochei, Scomberomorus cavcilla, and S. macu- latus) and 17 myctophids. The large juvenile and adult fish communities of the shallow coastal zone (<10 m) and continental shelf (9-180 m) comprised about 150 species, including both pelagic and demersal species (Wenner et al.^ ^ '^ ' ; Wenner and Sedberry, 1989). A maximum of 164 species were reported from rocky-reefs in the region (Chester et al., 1984; Parker et al. 1994; Figure 6 Correspondence analysis ordinations (portraying the first and second dimen- sion scores! of the inshore juvenile fish community data showing cross-shelf and along-shelf station groups in spring (A and B) and winter (C and D). Solid lines enclose the boundary of each station group with three or more stations. Station groups comprising one or two stations are not enclosed by a solid line. Each station group is labeled and portrayed with a different symbol. The dashed lines intersect at the origin of the plot. Analyses were conducted by using juvenile fish abundance data only. Baron et al., 2004). Fewer species inhabit estuaries than shelf habitats of the southeast (<90), particularly ■« Wenner, C. A., C. A. Barans, B. W. Stender, and F. H. Berry. 1979a. Results of MARMAP otter trawl investi- gations in the South Atlantic Bight. II, spring 1974, 78 p. lAvailable from Marine Resources Division, South Carolina Department of Natural Resources, 217 Ft. Johnson Rd.. P.O. Box 12559, Charleston, SC 29412.] •' Wenner, C. A., C. A. Barans, B. W. Stender, and F. H. Berry. 1979b. Results of MARMAP otter trawl investi- gations in the South Atlantic Bight. Ill, summer 1974. 62 p. [Available from Marine Resources Division, South Carolina Department of Natural Resources, 217 Ft. Johnson Rd., P.O. Box 12559, Charleston, SC 29412.1 I* Wenner, C. A., C. A. Barans, B. W. Stender, and F. H. Berry. 1979c. Results of MARMAP otter trawl investi- gations in the South Atlantic Bight. IV, winter-early spring 1975, 59 p. [Available from Marine Resources Division, South Carolina Department of Natural Resources, 217 Ft. Johnson Rd., P.O. Box 12559, Charleston, SC 29412.1 'Wenner, C. A., C. A. Barans, B. W. Stender, and F. H. Berry. 1980. Results of MARMAP otter trawl investiga- tions in the South Atlantic Bight. V, summer 1975, 57 p. [Available from Marine Resources Division, South Carolina Department of Natural Resources, 217 Ft. Johnson Rd., P.O. Box 12559, Charleston, SC 29412.] 270 Fishery Bulletin 104(2) Cross-shelf if different habitat types are con- sidered (Ross and Epperly, 1985; Hettler, 1989; Nelson et al., 1991). Thus, unconsolidated sediments are as rich a habitat in terms of the number of species, as are pelagic, rocky reef, and estuarine habitats in the southeast U.S. continental shelf ecosystem. Cross-shelf regions were defined on the Georgia shelf from spatial patterns in juvenile fish distribu- tions. The number and extent of cross-shelf regions varied season- ally, but in general, three regions were identified: inner-shelf, mid- shelf, and outer-shelf regions. Ju- venile assemblages were associated with the regions, yet some individ- ual species were distributed across regions. Juveniles of year-round residents (e.g.. Ophidian selenops, Diplectrum formosum. and Pn- onotus carolinus) were usually the most abundant species, although transient juveniles (e.g., Leiostomus xanthiirus, Lagodon rhomboides, and Brevoortia tyrannus) were sea- sonally abundant. Cross-shelf gra- dients in species distribution have been found for other organisms and life stages of fish on the southeast U.S. shelf (macroinfauna: Atkinson et al., 1985; larval fish; Marancik et al, 2005; adult reef fish; Ches- ter et al., 1984; adult demersal fish: Wenner et al.'*'' " "). Cross- shelf gradients also are common in demersal juvenile fish distribu- tions in other continental shelf ecosystems (northwest U.S.: Norcross et al., 1997; Toole et al., 1997; Abookire and Norcross, 1998; Bailey et al., 2003; Johnson et al., 2003; northeast U.S.: Steves et al., 2000; Sullivan et al., 2000; southwest U.S.: Johnson et al., 2001). The cause of cross-shelf gradients in juvenile fish distribution is difficult to determine. The primary envi- ronmental variable correlated with cross-shelf juvenile fish assemblages is often depth (Table 6; Norcross et al., 1997; Steves et al., 2000; Sullivan et al., 2000; Johnson et al., 2001; Johnson et al., 2003), although temperature, salinity, and sediment grain size also are correlated with depth (Norcross et al., 1997; Steves et al., 2000; Sullivan et al., 2000). On the Georgia shelf, salinity, density, and stratification correlated with the distribution of juvenile assemblages during the spring, fall, and winter (Table 6), and along with temperature (Figs. 2 and 6) likely influenced the distribution of ju- venile fish, but the causative mechanisms remain unre- solved. On the northeast U.S. shelf, cold bottom water left from winter resides on the mid-shelf during summer Along-shelf CA 1 CA1 Figure 7 Correspondence analysis ordinations (portraying the first and second dimen- sion scores) of the offshore juvenile fish community data showing cross-shelf and along-shelf station groups in fall (A and Bl and winter (C and D). Solid lines enclose the boundary of each station group with three or more stations. Station groups comprising one or two stations are not enclosed by a solid line. Each station group is labeled and portrayed with a different symbol. The dashed lines intersect at the origin of the plot. Analyses were conducted by using juvenile fish abundance data only. and fall (cold pool; Ketchum and Corwin, 1964). Steves et al. (2000) and Sullivan et al. (2000) hypothesized that cross-shelf patterns in settlement and juvenile fish distributions were caused by cross-shelf tempera- ture gradients related to the presence of the cold pool on the mid-shelf Similarly, juvenile flatfish species in Alaska exhibited cross-shelf gradients in habitat use that were influenced by temperature-depth interactions (Norcross et al., 1997). Hippoglossoides elassodon were most abundant in the colder, deeper locations, and Hip- poglossus stenolepis were more abundant in the warmer, shallow locations (Norcross et al., 1997). On the Georgia shelf, juvenile assemblages were less distinct during summer, when environmental gradients were weakest, providing some support for the hypothesis that cross- shelf patterns in juvenile distribution were caused by environmental factors. Alternatively, in other studies it has been hypothesized that juvenile distribution re- sults from selection of specific habitat characteristics within large-scale environmental gradients (Stoner and Abookire, 2002). For example, laboratory and field Walsh et al Juvenile fish assemblages on the southeast United States continental shelf 271 studies have found that some fish species select spe- cific sediment characteristics or biogenic structures (Stoner and Abookire. 2002; Diaz et al., 2003; Stoner and Titgen, 2003). and spatial patterns in these habitat characteristics may cause spatial patterns in juvenile fish distribution. We did not stratify sampling by sedi- ment characteristics or biogenic structures and thus we could draw no conclusion regarding the role of habitat characteristics in influencing juvenile fish di.stribution on the Georgia shelf No along-shelf spatial patterns were found in the juvenile fish community on the continental shelf off the Georgia coast; however, the along-shelf dimension sampled (-60 km) was relatively small compared to the along-shelf dimension of the southeast U.S. continental shelf ecosystem (-1000 km). Consistent patterns in along-shelf gradients in oceanographic features (Atkin- son et al., 1985; Lee et al., 1991) and adult fish commu- nities (Chester et al., 1984; Wenner and Sedberry, 1989) exist on the southeast U.S. shelf, but over larger along- shelf dimensions than sampled in the present study (200-400 km). Consistent patterns of juvenile use of along-shelf habitat over smaller dimensions (-100 km) have been reported for other shelf ecosystems (north- west U.S.; Norcross et al., 1997; Abookire and Norcross, 1998; northeast U.S.: Steves et al., 2000; Sullivan et al., 2000), but the relation to along-shelf environmen- tal and habitat characteristics is unclear. Along-shelf patterns in juvenile fish distribution on unconsolidated sediments may occur on the southeast U.S. shelf, but at dimensions larger than 60-100 km. Seasonal patterns in settlement and postsettlement movement both defined and blurred classification of juvenile fish assemblages on the Georgia shelf Many of the resident shelf species were consistently collected in the same cross-shelf regions (i.e., inner and mid-shelf: Ophidian selenops, Prionotus carolinus. Dactyloscopus moorei, Stenotomus sp., and Serranicithis piimilio; outer- shelf; Symphu)-us minor, Serranus phoebe, Centropristis ocyurus, and Bothus lunatus), and the consistent collec- tion of these species helped form a definition of juvenile assemblages. Other resident species shifted juvenile assemblages (e.g., Diplectriim formosum, Monacanthus hispidus, Bothus robinsi, Otophidium omostigmum, and Bothus ocellatus), possibly because of seasonal changes in settlement patterns or ontogenetic postsettlement movements, with the result that some classifications were blurred. Four scieanid species, present during the spring, summer, and fall as larvae (Marancik et al., 2005), were collected during the fall as juveniles and contributed to defining the fall inner-shelf (Larimus fasciatus and Menticirrhus americanus) and mid-shelf {Cynosion nothus and C. reg^a/is) juvenile assemblages. These species may shift settlement or juvenile habitat (or both) between estuaries and the coastal ocean; this shift has been shown to occur off the coast of New Jersey (Able et al., 2003; Neuman and Able, 2003). Further, a number of species in the ecosystem spawn on the shelf yet use estuarine habitats as juvenile nurser- ies (Warlen and Burke, 1990; Able and Fahay, 1998). In our study, settlement stage L. xanthurus, Brevoortia tyrannus, and Lagodon rhomboides were collected on the shelf, and contributed to the definition of the inner- and mid-shelf assemblages. Some individuals of these estuarine-dependent species may not settle directly into estuarine habitats as has been demonstrated for Centropristis striata off the coast of New Jersey (Able and Fahay, 1998). Only a few reef-associated species collected during the study used unconsolidated sediments consistently. Centropristis ocyurus, Diplectrum formosum, and Steno- tomus sp. were the most common reef fishes collected on unconsolidated sediments, and made up >1% of the total catch (Appendix). Other less frequently occur- ring juveniles of important reef-associated species that were collected were Centropristis striata, Epinephelus niveatus, Pristigenys alta, Priacanthus arenatus, and Lutjanus analis. Reef-associated species were collected across the en- tire shelf. Centropristis ocyurus and Stenotomus sp. were commonly collected on the outer-, inner-, and mid- shelves, respectively (Table 2, Appendix). Centropristis striata, which uses habitats in estuaries and coastal regions as juveniles (Able and Fahay, 1998), was col- lected on the inner- and mid-shelf in the spring and summer, during periods of high settlement (Able and Hales, 1997). The presence of many reef-associated species at the deeper (50 m) stations (Table 4) may in- dicate greater use of offshore unconsolidated sediments by reef-associated species or that sampling was done in an area closer to rocky-reefs. Several species of the South Atlantic Fishery Management Council snapper- grouper complex have been reported to use a variety of nonreef habitats as newly settled juveniles, including unconsolidated sediments on the shelf (Lindeman et al., 2000). Many coral reef fish also use several types of nearshore nonreef habitat as juveniles (de la Moriniere et al., 2002; Nagelkerken and van der Velde, 2002). These nonreef habitats are assumed to be important nursery habitats; however, strong evidence of movement from juvenile to adult habitats has been documented only for a few reef-associated fishes (Gillanders et al., 2003). The fact that reef-associated species also use unconsolidated sediments indicates there is an inter- dependence between reef habitats and unconsolidated sediments; yet the function of each habitat remains unquantified. In addition to reef species, a number of commercially and recreationally important demersal species used unconsolidated sediments on the shelf off Georgia as juvenile habitat. Juvenile Cynoscion nothus, C. regalis and Menticirrhus americanus were collected on the in- ner- and mid-shelves in the fall, and settlement-size Leiostomus xanthurus were collected in the winter. However, small noncommercial demersal species were the most abundant juveniles collected during beam trawl sampling. Much of the southeast U.S. continental shelf is im- pacted by trawl fisheries that can adversely impact unconsolidated sediments (Barnette^). Current manage- 272 Fishery Bulletin 104(2) ment concerns focus primarily on bycatch reduction, im- pacts on stock assessments of commercially and recre- ationally important finfish, and destruction of coral, live hard-bottom, and SAV habitat (NOAA''; Barnette"'). The presence of reef-associated and sciaenid species in our samples indicate that the unconsolidated sediments of the shelf are potentially important habitats for early life stages, and that trawl fisheries on the shelf may impact the population dynamics of these species. Equally im- portant may be the ecological effects on the small non- commercial demersal species that dominate the catch on unconsolidated sediments and that are abundant in the shrimp trawl bycatch (SAFMC, 1996). Unconsolidated sediments serve as juvenile habitat for a number of species in the southeast U.S. shelf eco- system. Some species use unconsolidated sediments throughout their life history, but for a number of spe- cies, unconsolidated sediments serve as only one of a mosaic of habitats through the life history (see Brow- man and Konstantinos, 2004). Additionally, there is growing recognition of the importance of unconsoli- dated sediments in the trophic ecology of the southeast U.S. shelf Benthic primary production of the shelf has been shown to be 4-6 times greater than water-column production at specific locations (Nelson et al., 1999). Unconsolidated sediments are areas of reef fish feed- ing (Sedberry, 1985, 1990) and several important fish predators inhabit unconsolidated sediments (triglids, synodontids, Ross, 1976; Richards et al., 1979). Cur- rent ecosystem approaches to fisheries (see Sissenwine and Murawski, 2004) in the region largely ignore un- consolidated sediments (SAFMC, 1998, 2001), and al- though the importance of this habitat to juvenile fish production remains unquantified (sensu Beck et al., 2001), management efforts would be improved through a greater recognition of the potential importance of these habitats to fisheries production and the intercon- nectedness of multiple habitats in the southeast U.S. continental shelf ecosystem. Acknowledgments We would like to thank all who helped with sample collections, sorting, and analyses: G. Bohne, R. Bohne, C. Bonn, J. Burke, M. Burton, B. Began, M. Duncan, J. Govoni, M. Greene, E. Jugovich, S. Lem, J. Loefer, R. Mays, G. McFall, R. McNatt, A. Powell, R. Rogers, S. Shoffler, S. Varnam, and T. Zimanski. We would like to thank NCAA's Undersea Research Center at the University of North Carolina at Wilmington for use of the ROV and L. Horn for operating the ROV. We appre- ciate the hard work and dedication of the officers and " NOAA (National Oceanic and Atmospheric Admini- stration). 1998. Managing the nation's bycatch: priorities, programs, and actions for the National Marine Fi.sheries Service, 199 p. [Available from National Technical Infor- mation Service, U.S. Department of Commerce, 5285 Port Royal Road, Springfield, VA 22161.1 crew of the NOAA Ship Ferrel, NOAA Ship Jane Yarn, NCAA Ship Oregon II, and RV Cape Fear. We would also like to thank A. Powell, M. Sullivan, J. Merriner, and an anonymous reviewer for their comments on pre- vious drafts. Gray's Reef National Marine Sanctuary, the National Marine Sanctuary Office, and Center for Coastal Fisheries and Habitat Research provided fund- ing for the project. Literature cited Able, K. W., and M. P. Fahay. 1998. The first year in the life of estuarine fishes in the Middle Atlantic Bight, 342 p. Rutgers Univ. Press, New Brunswick, NJ. Able, K. W., and L. S. Hales Jr. 1997. Movements of juvenile black sea bass Centropristis striata (Linnaeus) in a southern New Jersey estuary. J. Exp. Mar. Biol. Ecol. 213:153-167. Able, K. W., P. Rowe, M. Burlas, and D. Byrne. 2003. Use of ocean and estuarine habitats by young-of- year bluefish (Pomatomus saltatrix) in the New Yoi'k Bight. Fish. Bull. 101:201-214. Abookire, A. A., and B. L. Norcross. 1998. Depth and substrate as determinants of distribu- tion of juvenile flathead sole iHippoglossoides elassodon) and rock sole iPleuronectes bilineatus), in Kachemak Bay, Alaska. J. Sea. Res. 39:113-123. Atkinson, L. P., D. W. Menzel, and K. A. Bush (eds). 1985. Oceanography of the southeastern U.S. continen- tal shelf, 156 p. Am. Geophysical Union, Washington, D. C. Bailey, K. M., E. S. Brown, and J. T. Duffy-Anderson. 2003. Aspects of distribution, transport and recruitment of Alaska plaice ^Pleuronectes quadrituberculatus) in the Gulf of Alaska and eastern Bering Sea: compari- son of marginal and central populations. J. Sea Res. 50:87-95. Baron, R. M., L. K. B. Jordan, and R. E. Spieler. 2004. Characterization of the marine fish assemblage associated with the nearshore hardbottom of Broward County, Florida, USA. Estuarin. Coast. Shelf Sci. 60:431-443. Beck, M. W., K. L. Heck, K. W. Able, D. L. Childers, D. B. Egg- leston, B. M. Gillanders, B. Halpern, C. G. Hays, K. Hoshino, T. J. Minello, R. J. Orth, P. F. Sheridan, and M. R. Weinstein. 2001. The identification, conservation, and manage- ment of estuarine and marine nurseries for fish and invertebrates. Bioscience 51:633-641. Beck, M. W., and M. Odaya. 2001. Ecoregional planning in marine environments: identifying priority sites for conservation in the northern Gulf of Mexico. Aquat. Conserv. Mar. Freshw. Ecosys. 11:235-242. Blanton, B. O., A. Aretxabaleta, F. E. Werner, and H. E. Seim. 2003. Monthly climatology of the continental shelf waters of the South Atlantic Bight. J. Geophys. Res. Oceans 108:C8. Botsford, L. W., J. C. Castilla, and C. H. Peterson. 1997. The management of fisheries and marine eco- systems. Science 277:509-515. Browman, H. I., and I. S. Konstantinos. 2004. Marine protected areas as a central element of Walsh et al Juvenile fish assemblages on the southeast United States continental shelf 273 ecosystem-based management: defining their location, size and number. Mar. Ecol. Prog. Ser. 274:271-272. Burke. J. S.. J. M. Miller, and D. E. Hoss. 1991. Immigration and settlement pattern of Parali- chthys dentatiis and P. lethostigma in an estuarine nursery ground. North Carolona, USA. Neth. J. Sea Res. 27:393-405 Chester, A. J., G. R. Huntsman, P. A. Tester, and C. S. Manooch, III. 1984. South Atlantic Bight reef fish communities as represented in hook-and-line catches. Bull. Mar. Sci. 34: 267-279. Chester, A. J., and G. W. Thayer. 1990. Distribution of spotted seatrout [Cynoscion ucbu- losus) and gray snapper (Lutjanus griseiis) juveniles in seagrass habitats of western Florida Bay. Bull. Mar. Sci. 46:345-357. CFEPTAP (Chesapeake Fisheries Ecosystem Plan Technical Advisory Panel). 2004. Fisheries ecosystem planning for Chesapeake Bay, 363 p. NOAA Chesapeake Bay Office, Annapo- lis," MD. Cowen, R. K., K. M. M. Lwiza, S. Sponaugle, C. B. Paris, and D. B. Olson. 2000. Connectivity of marine populations: Open or closed? Science 287:857-859. Dayton, P. K., E. Sala, M. J. Tegner, and S. Thrush. 2000. Marine reserves: Parks, baselines, and fishery enhancement. Bull. Mar. Sci. 66:617-634. de la Moriniere, E. C, B. J. A. Pollux. I. Nagelkerken, and G. van der Velde. 2002. Post-settlement life cycle migration patterns and habitat preference of coral reef fish that use seagrass and mangrove habitats as nurseries. Estuar. Coast. Shelf Sci. 55:309-321. Diaz. R. J.. G. R. Cutter, and K. W. Able. 2003. The importance of physical and biogenic struc- ture to juvenile fishes on the shallow inner continental shelf. Estuaries 26:12-20. FR (Federal Register). 2000. Fisheries of the Caribbean, Gulf of Mexico, and South Atlantic; essential fish habitat for species in the South Atlantic; amendment 4 to the fishery management plan for coral, coral reefs, and live/hard bottom habi- tats of the South Atlantic region (Coral FMP). Federal Register 50 CFR, part 622 and 640, vol. 65, no. 115, p. 37292-37296. Government Printing Office, Wash- ington, D.C. Gibson, R. N, 1994. Impact of habitat quality and quantity on the recruitment of juvenile flatfishes. Neth. J. Sea. Res. 32: 191-206. Gillanders, B. M., K. W. Able, J. A. Brown. D. B. Eggleston. and P. F Sheridan. 2003. Evidence of connectivity between juvenile and adult habitats for mobile marine fauna: an impor- tant component of nurseries. Mar. Ecol. Prog. Ser. 247:281-295. Greenacre, M. J. 1984. Theory and applications of Correspondence Anal- ysis, 364 p. Academic Press, London, and Orlando, FL. Hettler, W. F, Jr. 1989. Nekton use of regularly-flooded saltmarsh cord- grass habitat in North Carolina, USA. Mar. Ecol. Prog. Ser. 56:111-118. Humann. P. 1994. Reef fish identification: Florida, Caribbean, Baha- mas, 396 p. Paramount Miller Graphics, Inc., Jack- sonville. FL. Johnson, K. A., M. M. Yoklavich, and G. M. Cailliet. 2001. Recruitment of three species of juvenile rockfish (Seba>iteii spp.) on soft benthic habitat in Monterey Bay, California. CalCOFl Rep. 42:53-166. Johnson, S. W.. M. L. Murphy, and D. J. Csepp. 2003. Distribution, habitat, and behavior of rockfishes, Sehastes spp., in nearshore waters of southeastern Alaska: observations from a remotely operated vehicle. Environ. Biol. Fish. 66:259-270. Kendall, A. W. J., and A. C. Matarese. 1994. Status of early life history descriptions of marine teleosts. Fish. Bull. 92:725-736. Ketchum, B. H., and N. Corwin. 1964. The persistence of "winter" water on the conti- nental shelf south of Long Island, New York. Limnol. Oceanogr. 4:467-475. Kuipers, B. 1975. On the efficiency of a two-meter beam trawl for juvenile plaice iPleuronectes platessa). Neth. J. Sea Res, 9:69-85. Lee. T. N., Yoder, J. A., and L. P. Atkinson. 1991. Gulf Stream frontal eddy influence on productiv- ity of the southeast LI.S. continental shelf. J. Geophy. Res. 96: 22,191-22,205. Lehnert, R. L., and D. M. Allen. 2002. Nekton use of subtidal oyster shell habitat in a southeastern US estuary. Estuaries 25:1015-1024. Lindeman, K. C, R. Pugliese, G. T. Waugh, and J. S. Ault. 2000. Developmental patterns within a multispecies reef fishery: Management applications for essential fish habitats and protected areas. Bull. Mar. Sci. 66:929-956. Marancik, K. E., J. A. Hare, and L. M. Clough. 2005. Cross-shelf and seasonal variation in larval fish assemblages on the southeast United States continen- tal shelf off the coast of Georgia, USA. Fish. Bull. 103:108-129. Menzel, D. W. 1993. Ocean processes: U.S. southeast continental shelf, 112 p. Office of Scientific and Technical Information, Oak Ridge, TN. Meyer. D. L., and E. C. Townsend. 2000. Faunal utilization of created intertidal eastern oyster iCrassostrea virginica) reefs in the southeastern United States. Estuaries 23:34-45. Moser, M. L., P. J. Auster, and J. B. Bichy. 1998. Effects of mat morphology on large Sargassum- associated fishes: observations from a remotely operated vehicle (ROV) and free-floating video camcorders. En- viron. Biol. Fish. 51:391-398. Munroe. T. A. 1998. Systematics and ecology of tonguefishes of the genus Symphurus (Cynoglossidae : Pleuronectiformes) from the western Atlantic Ocean. Fish. Bull. 96:1-182. Nagelkerken, I., and G. van der Velde. 2002. Do non-estuarine mangroves harbour higher densities of juvenile fish than adjacent shallow-water and coral reef habitats in Curacao (Netherlands Antilles)? Mar. Ecol. Prog. Ser. 245:191-204. Nelson, D. M., E. A. Irlandi, L. R. Settle, M. E. Monaco, and L. C. Coston-Clements. 1991. Distribution and abundance of fishes and inverte- brates in southeast estuaries, 167 p. Estuarine Living Marine Resources (ELMR) Rep. no. 9. NOAA/NOS 274 Fishery Bulletin 104(2) Strategic Environmental Assessments Division, Silver Spring. MD. Nelson, J. R., J. E. Eckman, C. Y. Robertson. R. L. Marinelli. and R. A. Jahnke. 1999. Benthic microalgal biomass and irradiance at the sea floor on the continental shelf of the South Atlan- tic Bight: Spatial and temporal variability and storm effects. Cent. Shelf. Res. 19:477-505. Neuman, M. J., and K. W. Able. 2003. Inter-cohort differences in spatial and temporal settlement patterns of young-of-the-year windowpane iScophthalmus aquosus) in southern New Jersey. Es- tuar. Coast. Shelf Sci. 56:527-538. Norcross. B. L., F. J. Muter, and B. A. Holladay. 1997. Habitat models for juvenile pleuronectids around Kodiak Island, Alaska. Fish. Bull. 95:504-520. Parker. R. O.. Jr., D. R. Colby, and T. D. Willis. 1983. Estimated amount of reef habitat on a portion of the U.S. South Atlantic and Gulf of Mexico continental shelf. Bull. Mar. Sci. 33:935-940. Parker, R. O., Jr., A. J. Chester, and R. S. Nelson. 1994. A video transect method for estimating reef fish abundance, composition, and habitat utilization at Gray's Reef National Marine Sanctuary, Georgia. Fish. Bull. 92:787-799. Parker, R. O.. Jr., and R. W. Mays. 1998. Southeastern U.S. deepwater reef fish assem- blages, habitat characteristics, catches, and life history summaries. NOAA Tech. Rep. NMFS 138, 41 p. Parrish, R. 1999. Marine reserves for fisheries management: Why not? CalCOFI Rep. 40:77-86. Pauly, D., V. Christensen, J. Dalsgaard, R. Froese, and F. Torres Jr. 1998. Fishing down marine food webs. Science 279: 860-863. Peterson, M. S. 2003. A conceptual view of environment-habitat-produc- tion linkages in tidal river estuaries. Rev. Fish. Sci. 11:291-313. Richards, S. W., J. M. Mann, and J. A. Walker. 1979. Comparision of spawning seasons, age, growth rates, and food of two sympatric species of searobins, Prionotus carolinus and Prionotus evolans, from Long Island Sound. Estuaries 2:255-268. Roberts, C. M.. J. A. Bohnsack, F. Gell, J. P Hawkins, and R. Goodridge. 2001. Effectsof marine reserves on adjacent fisheries. Sci- ence 294:1920-1923. Ross, S. T. 1976. Patterns of resource partitioning in sea robins (Pisces: Triglidae). Copeia 1977:561-571 Ross, S. W., and S. P. Epperly. 1985. Utilization of shallow estuarine nursery areas by fishes in Pamlico Sound and adjacent tributaries. North Carolina. In Fish community ecology in estuaries and coastal lagoons: towards an ecosystem integration (A. Yanez-Arancibia, ed.) p. 207-232. Universidad Nacional Autonoma de Mexico Press, Mexico. Ross, S. W., and J. E. Lancaster. 2002. Movements and site fidelity of two juvenile fish species using surf zone nursery habitats along the southeastern North Carolina coast. Environ. Biol. Fish. 63:161-172. Ross, S. W., and M. L. Moser. 1995. Life history of juvenile gag, Mycteroperca microl- epis, in North Carolina estuaries. Bull. Mar. Sci. 56: 222-237. Rothschild, B. J. 1986. Dynamics of marine fish populations, 277 p. Har- vard Univ. Press, Cambridge. MA. SAFMC (South Atlantic Fishery Management Council). 1996. Final amendment 1 to the fishery management plan for the shrimp fishery of the South Atlantic region (rock shrimp), 215 p. South Atlantic Fishery Manage- ment Council, Charleston, SC. 1998. Habitat plan for the South Atlantic region: essen- tial fish habitat requirements for fishery management plans of the South Atlantic Fishery Mangement Council, 449 p. South Atlantic Fishery Management Council, Charleston, SC. SAFMC (South Atlantic Fishery Man- agement Council). 2001. Magnuson-Stevens Act/NEPA scoping document: using marine protected areas in the South Atlantic, 22 p. South Atlantic Fishery Management Council, Charleston, SC. SEAMAP-SA ( Southeast Area Monitoring and Assessment Pro- gram-South Atlantic). 2001. Distribution of bottom habitats on the continental shelf from North Carolina through the Florida Keys, 166 p. SEAMAP-SA Bottom Mapping Workgroup, Atlantic States Marine Fisheries Commision, Wash- ington, DC. Sedberry, G. R. 1985. Food and feeding of the tomatate, Haemulon auro- lineatum (Pisces, Haemulidae), in the South Atlantic Bight. Fish. Bull. 83:461-466. Sedberry, G. R. 1990. Feeding habits of whitebone porgy. Calami/.': leucos- teus (Teleiostei: Sparidae), associated with hard bottom reefs off the southeastern United States. Fish. Bull. 87:935-944. Shanks, A. L., B. A. Grantham, and M. H. Carr. 2003. Propagule dispersal distance and the size and spac- ing of marine reserves. Ecol. Appl. 13:S159-S169. Simpson, J. H., and I. D. James. 1985. Coastal and estuarine fronts. In Baroclinic pro- cesses on continental shelves (C. N. K. Mooers ed.), p. 63-93. Am. Geophys. Union. Washington, D.C. Sissenwine, M. P. 1984. Why do fish populations vary? In Exploita- tion of marine communities (R. M. May, ed.), p. 129-152. Springer-Verlag, New York, NY. Sissenwine, M.. and S. Murawski. 2004. Moving beyond 'intelligent tinkering': advancing an ecosystem approach to fisheries. Mar. Ecol. Prog. Ser. 274:291-295. Steves, B. P., R. K. Cowen, and M. H. Malchoff 2000. Settlement and nursery habitats for demersal fishes on the continental shelf of the New York Bight. Fish. Bull. 98:167-188. Stoner, A. W., and A. A. Abookire. 2002. Sediment preferences and size-specific distribu- tion of young-of-the-year Pacific halibut in an Alaska nursery. J. Fish Biol. 61:540-559. Stoner, A. W., and R. H. Titgen. 2003. Biological structures and bottom type influence habitat choices made by Alaska flatfishes. J. Exp. Mar. Biol. Ecol. 292:43-59. Sullivan, M. C, R. K. Cowen, K. W. Able, and M. P. Fahay. 2000. Spatial scaling of recruitment in four continental shelf fishes. Mar. Ecol. Prog. Ser. 207:141-154. Walsh et al Juvenile fish assemblages on the southeast United States continental shelf 275 ter Braak, C. F. J., and P. Smilauer. 2002. CANOCO reference manual and CanoDraw for Windows user's guide: software forcanonical community ordination (vers. 4.5), 500 p. Microcomputer Power. Ithaca, NY. Thayer, G. W., D. R. Colby, and W. F. Hettler Jr. 1987. Utilization of the red mangrove prop root habitat by fishes in south Florida. Mar. Ecol. Prog. Ser. 35: 25-38. Thayer, G. W., A. B. Powell, and D. E. Hoss. 1999. Composition of larval, juvenile, and small adult fishes relative to changes in environmental conditions in Florida Bay. Estuaries 22:518-533. Toole, C. L., D. F. Markle. and C. J. Donohoe. 1997. Settlement timing, distribution, and abundance of Dover sole iMicrostomus pacificus) on an outer con- tinental shelf nursery area. Can. J. Fish. Aquat. Sci. 54:531-542. Walsh, H. J., D. S. Peters, and D. P. Cyrus. 1999. Habitat utilization by small flatfishes in a North Carolina estuary. Estuaries 22:803-813. Warlen, S. M., and J. S. Burke. 1990. Immigration of larvae of fall/winter spawning marine fishes into a North Carolina estuary. Estuaries 13:453-461. Wenner, E. L., D. M. Knott, R. F. Van Dolah, and V. G. Burrell Jr. 1983. Invertebrate communities associated with hard bottom habitats in the South Atlantic Bight. Estuar. Coast. Shelf Sci. 17:143-158. Wenner, C. A., and G. R. Sedberry. 1989. Species composition, distribution, and relative abun- dance of fishes in the coastal habitat off the southeastern United States. NOAA Tech. Rep. NMFS 79, 49 p. Appendix Juvenile taxa collected in beam trawl collect ions from the continental shelf off the coast of Georgia. USA. Seasons 1 catch per unit of effort (CPUE, fish/5 mint was calculated from cross-shelf stations (see Fig. 1 and Table 1). Mean (SD) of the total CPUE and standard len gth (SL. mm) was calculated from all stations. Adult distributions (AD) and whether Bpecies are reef-associated (RA) was determined from the literature. E = estuarine, C = coastal (0-20 m) O = open-shelf ( 20-70 m), SI = slope (>70 m), W = | weakly reef-associated, and S = strongly reef associated. Family Species Cross-shelf stations All stations AD RA Mean CPUE SL(mm) Spring Summer Fall Winter Mean(SD) Mean(SD) Elopidae Elops saurus 0.02 0.01 0.007 (±0.084) 46.7 (±15.91 E,C Muraenidae Gymnothorax saxicola 0.005 (±0.069) 69.5 (±0.71 CO Ophichthidae Ophichthus gomesii 0.01 0.009 (±0.119) 103.3 (±20.1) E,C, 0 Ophichthus ocellatus 0.05 0.03 0.02 0.016 (±0.133) 160.9 (±59.21 E,C,0 Congridae Ariosoma balearicum 0.39 0.10 0.30 0.02 0.135 (±0.431) 126.5 (±44.2) E,C.O Clupeidae Brevoortia tyrannus 0.07 0.014 (±0.1161 24.5 (±1.8) E.C.O Etrumeus teres 0.02 0.01 0.005 (±0.069) 31.8 (±1.1) CO Sardinella aurita 0.01 0.005 (±0.069) 36.0 (±2.8) CO Engraulididae Anchoa hepsetus 0.09 0.19 1.14 0.355 (±2.6311 31.3 (±10.21 E.CO Anchoa lamprotaena 0.07 0.015 (±0.2151 44.5 (±5.8) C Argentinidae Argentina striata 0.02 0.005 (±0.069) 52.0 (±1.4) SI Synodontidae Synodus foetens 1.00 0.45 1.55 0.39 0.643 (±1.5701 65.8(±40.1i E,C,0 Synodus poeyi 0.26 0.035 (±0.420) 33.1 (±3.5) 0 Trachinocephalus myops 0.13 0.06 0.22 0.06 0.091 (±0.354) 92.3 (±47.61 0 Gadidae Urophycis regia 0.28 0.68 0.302 (±0.867) 69.7 (±21.3) CO Ophidiidae Ophidian grayi 0.02 0.002 (±0.049) 39.0 CO Ophidian holhrooki 0.70 0.121 (±1.226) 98.8 (±47.3) CO Ophidian marginatum 0.11 0.03 0.01 0.03 0.050 (±0.270) 94.8 (±43.0) CO Ophidian selenops 3.63 0.35 4.92 1.49 1.823 (±4.373) 56.1 (±19.71 0 Ophidian welshi 0.12 0.02 0.019 (±0.194) 107.9 (±27.11 CO Otaphidium omostigmum 0.60 0.09 0.69 0.38 0.392 (±1.1921 60.0 (±21.81 0 Batrachoididae Porichtys plectrodan 0.03 0.007 (±0.0781 56.3 (±13.71 CO Antennariidae Antennarius ocellatus 0.002 (±0.048) 47.0 O.Sl w Antennarius radiosus 0.02 0.010 (±0.102) 24.0 (±4.1) 0. SI w Ogcocephalidae Halieutichthys aculeatus 0.02 0.04 0.014 (±0.1181 14.4 (±1.5) CO Ogcocephalus nasiitus 0.02 0.05 0.01 0.025 (±0.1541 29.0 (±17.8) CO Ogcocephalus parvus 0.02 0.012 (±0.1081 75.6 (±34.1) 0 continued 276 Fishery Bulletin 104(2) Appendix (continued) Family Species Cross-shelf stations All stations AD RA Mean CPUE SL (mm) Spring Summer Fall Winter Mean(SD) Mean(SD) E.xocetidae Heiniramphus balao 0.02 0.003 ±0.052) 80.0 C,0 Hemiramphus brasilieusis 0.01 0.002 ±0.0491 97.0 CO Hirundichthys affinis 0.05 0.02 0.017 ±0.146) 35.4 ( ±13.8) CO Parexocoetiis brachypterus 0.02 0.002 ±0.0481 34.0 0 Syngnathidae Hippocampus erectus 0.12 0.04 0.01 0.04 0.055 ±0.236) 60,8 ( ±19.9) E.C Syngnathus scovelli 0.06 0.02 0.019 ±0.166) 56.3 ( ±11.8) E,C Syngnathus springeri 0.02 0.01 0.07 0.021 ±0.144) 64.3 ( ±6.9) C,0 Scorpaenidae Scorpaenci dispar 0.02 0.07 0.043 ±0.4581 39.4 ( ±21.61 C,0 W Scorpaena plumieri 0.07 0.13 0.08 0.02 0.078 ±0.335) 43.5 ( ±31.0) CO W Triglidae Prionotus spp. 0.73 0.17 15.31 0.52 2.700 ±9.684) 11.6 ( t3.3) CO Bellator brachychir 0.01 0.01 0.007 ±0.0791 27.0 ( tl3.2) 0 Bellator militaris 0.01 0.01 0.004 ±0.063) 17.5 (±12.0) 0 Prionotus carolinus 3.04 0.32 0.42 1.08 1.127 ±2.801) 53.5 ( ±27.9) E,C,0 Prionotus evolans 0.03 0.004 ±0.081) 90.0 ( ±62.2) E,C,0 Prionotus ophryas 0.01 0.002 ±0.0491 16.0 CO Prionotus scitulus 0.08 0.033 ±0.180) 133.3 ( ±19.4) CO Serranidae Centropristis ocyurus 0.73 0.04 0.08 0.313 ±1.592) 31.0 ( tl2.0) 0 w Cenlropristis striata 0.07 0.04 0.017 ±0.161) 33.1 ( ±14.8) E.C.O s Diplectrum forniosum 0.96 2.50 2.39 0.16 1.360 ±3.378) 28.2 ( ±23.8) CO w Epinephelus niveatus 0.005 (±0.069) 29.0 ( ±7.1) 81 s Serraniculus pumilio 0.03 0.21 0.34 0.12 0.155 ±0.734) 22.5 ( ±7.5) CO w Serranus phoebe 0.31 0.04 0.02 0.04 0.220 ±1.618) 21.6(18.1) C.O w Serranus subligarius 0.007 ±0.146) 30.7 ( t9.3) CO s Grammistidae Rypticus spp. 0.01 0.046 ±0.590) 41.8 ( ±12.9) CO s Rypticus bistrispinus 0.054 (±0.596) 42.5 ( ±5.2) CO s Priacanthidae Priacanlhusarenatus 0.01 0.002 ±0.049) 15.0 0 s Pristigenys alta 0.02 0.05 0.016 ±0.126) 17.7 (±8.61 0 s Apogonidae Apogon sp. Astrapogon spp. fEpigonus spp 0.01 0.03 0.01 0.011 0.002 ±0.125) ±0.049) 20.9 ( 9.2 ±11.0) 0 w Apogon maculatus 0.061 ±0.6801 17.6 (±3.5) 0 w Apogon pseudomaculatus 0.04 0.028 ±0.213) 18.6 ( ±14.1) 0 w Carangidae Carangidae 0.02 0.005 ±0.068) 18.5 ( ±6.4) Decapterus spp. 0.14 0.031 ±0.278) 24.8 ( ±5.7) 0 Caranx bartholomaei 0.02 0.002 ±0.044) 12.0 CO Caranx crysos 0.02 0.002 ±0.049) 101.0 C.O w Chloroscombrus chrysurus 0.01 0.06 0.010 ±0.107) 15.1 ( t3.6) CO w Decapterus macarellus 0.02 0.01 0.005 ±0.069) 23.0 ( ±4.2) 0 Decapterus punctatus 0.05 0.01 0.017 ±0.188) 27.61 t5.8) CO Trachurus lath ami 0.02 0.002 (±0.046) 16.0 C.O Lutjanidae Lutjanus analis 0.01 0.002 (±0.049) 18.0 C.O s Haemulidae Haemulnn aurolineatum 0.05 0.05 0.016 ±0.185) 45.3 1 ±26.8) CO w Sparidae Stenotomus sp. 0.18 0.16 0.14 0.08 0.135 ±0.823) 46.1 ( ±28.4) c,o w Lagodon rhomboides 0.06 0.012 ±0.128) 27.1 (±27.9) E.CO w Sciaenidae Cynoscion nothus 0.04 0.006 ±0.095) 9.8 ( tO.9) E.C Cynoscion regalis 0.06 0.008 ±0.102) 10.3 ( tO.S) E, C Larimus fasciatus 0.19 0.027 ±0.313) 14.7 (±4.7) E,C Leiostomus xanthurus 2.27 3.361 22.193) 16.2 ( tl.7) E,CO continued Walsh et al Juvenile fish assemblages on the southeast United States continental shelf 277 Appendix (continued) Family Species Cross-shelf stations All stations AD RA Mean CPUE SL(mm) Spring Summer Fall Winter Mean(SD) Mean (SD) Sciaenidae(cont.) Menticirrhiisamericantis Stellifer laceolatus 0.07 0.010 (±0.159) 0.005 (±0.097) 13.3 (±9.1) 75.0 (±0.0) E,C E,C Mullidae Upeneus parvus 0.002 (±0.049) 48.0 CO Mugilidae Mugil cephali/s 0.18 0.047 (±0.422) 21.7 (±2.3) E,C,0 Miigil ciirema 0.01 0.02 0.027 (±0.214) 21.8 (±1.5) E,C,0 Labridae Halichoeres hivittatus 0.02 0.010 (±0.152) 71.0(±5.0) 0 S Xyrichtys novacida 0.01 0.002 (±0.049) 41.5 0 Scaridae Scaridae 0.03 0.009 (±0.093) 41.3 (±31.8) Uranoscopidae Kathetostoma albigutta 0.01 0.002 (±0.049) 26.0 0 Dactyloscopidae Dactyloscopus moorei 0.80 0.17 2.09 0.80 0.689 (±2.073) 22.1 (±5.1) 0 Blenniidae ChasmodeslParablennius mannoreus 0.04 0.007 (±0.109) 19.0 (±2.6) E.C S Hypleurochilus geminatus 0.04 0.01 0.021 (±0.160) 43.6 (±17.2) E.C Parablennius mannoreus 0.02 0.009 (±0.119) 20.3 (±4.3) E,C S Callionymidae Diplogrammus pauciradiatus 0.07 0.025 (±0.170) 9.5 (±1.0) 0 Gobiidae Microgobius spp. 0.06 0.82 1.42 0.01 0.454 (±1.424) 16.7 (±1.9) Gobiosoma bosci 0.01 0.014 (±0.150) 19.5 (±1.5) E.C Gobiosoma ginsburgi 0.02 0.005 (±0.068) 22.5 (±3.5) E.C W loghssus calliurus 0.01 0.002 (±0.041) 14.7 CO Stromateidae Peprilus triacanthus 0.07 0.01 0.08 0.042 (±0.250) 25.3 (±11.7) CO Bothidae Bothus oceUatus 1 robinsi 0.07 0.08 0.07 0.041 (±0.254) 15.9 (±2.6) Bothus spp. 0.86 0.41 4.50 0.33 0.971 (±3.728) 18.4 (±3.0) 0 Bothus lunatus 0.02 0.01 0.005 (±0.069) 79.0 (±9.9) E,C,0 Bothus ocellatus 0.03 0.04 0.01 0.046 (±0.341) 33.4 (±23.5) CO Bothus robinsi 1.21 0.23 0.52 0.34 0.435 (±1.464) 37.3 (±35.7) CO Paralichthyidae Citharichthys spp. 0.19 0.03 0.41 0.132 (±0.671) 13.7 (±6.0) Etropus spp. 8.89 1.31 2.96 15.97 10.563(20.9.36) 18.3 (±8.21 Ancylopsetta quadrocellata 0.22 0.114 (±0.386) 38.1 (±17.7) CO Citharich thys gym norhinus 0.03 0.004 (±0.063) 18.5 (±7.8) 0 Citharichthys macrops 0.20 0.04 0.02 0.13 0.090 (±0.441) 76.9 (±44.8) CO Cyclopsetta fimbriata 0.03 0.04 0.05 0.037 (±0.226) 28.9 (±14.5) 0 Syacium papillosum 0.02 0.09 0.04 0.040 (±0.269) 16.5 (±17.2) CO Scopthalmidae Scopthalmus aquosus 0.01 0.002 (±0.049) 14.0 E,C,0 Soleidae Gymnachirus melas 0.04 0.03 0.014 (±0.115) 17.1 (±7.5) 0 Cynoglossidae Symphurus spp. 0.01 0.02 0.005 (±0.067) 18.0 (±9.9) Symphurus diomedeanus 0.05 0.007 (±0.109) 22.5 (±17.7) 0 Symphurus minor 0.26 0.25 1.87 0.15 0.552 (±1.878) 23.8 (±4.2) 0 Symphurus parvus 0.05 0.02 0.012 (±0.108) 15.9 (±1.7) 0 Symphurus plagiusa 0.02 0.029 (±0.270) 59.9 (±16.5) E,C,0 Symphurus urospilus 0.42 0.24 0.43 0.28 0.257 (±0.717) 55.5 (±20.8) 0 Balistidae Aluterus schoepfi Cantherhines pullus Monacanthus ciliatus 0.02 0.01 0.02 0.005 (±0.066) 0.012 (±0.128) 0.005 (±0.069) 18.0 (±11.3) 18.5 (±1.5) 41.0 (±26.9) C 0 w s Monacanthus hispidus 0.18 0.54 1.36 0.06 0.460 (±1.314) 34.4 (±22.6) CO Monacanthus setifer 0.01 0.004 (±0.063) 91.0 (±0.0) 0 Ostraciidae Lactophyrs quadricornis 0.01 0.009 (±0.090) 67.8 (±83.6) 0 w Tetraodontidae Sphoeroides spp. 0.03 0.13 0.01 0.126 (±0.600) 15.9 (±4.0) Diodontidae Chilomycterus schoepfi 0.02 0.002 (±0.049) 17.0 E, C 278 Abstract — Age and growth estimates for salmon sharks iLamna ditropis) in the eastern North Pacific were derived from 182 vertebral centra collected from sharks ranging in length from 62.2 to 213.4 cm pre- caudal length ( PCLl and compared to previously published age and growth data for salmon sharks in the western North Pacific. Eastern North Pacific female and male salmon sharks were aged up to 20 and 17 years, respec- tively. Relative marginal increment (RMI) analysis showed that postnatal rings form annually between January and March. Von Bertalanffy growth parameters derived from vertebral length-at-age data are L^ =207.4 cm PCL, /; = 0.17/yr, and ?o=-2.3 years for females (;7 = 166), and L,=182.8 cm PCL, k = 0.2ZlyT , and tf, = -1.9 years for males {n = 16}. Age at maturity was estimated to range from six to nine years for females (median pre- caudal length of 164.7 cm PCLi and from three to five years old for males (median precaudal length of 124.0 cm PCLl. Weight-length relationships for females and males in the eastern North Pacific are W=8.2x 10-05xL2 759 (r-' = 0.99) and W=3.2x IQ-"** xL^ ^^^ (r-=0.99), respectively. Our results show that female and male salmon sharks in the eastern North Pacific possess a faster growth rate, reach sexual maturity earlier, and attain greater weight-at-length than their same-sex counterparts living in the western North Pacific. Growth and maturity of salmon sharks iLamna ditropis) in the eastern and western North Pacific, and comments on back-calculation methods Kenneth J. Goldman Alaska Department of Fish and Game 3298 Douglas Place Homer, Alaska 99603 E-mail address: ken_goldmanm'fishgame. state. ak. us John A. Musick College of William & Mary. School of Marine Science Virginia Institute of Marine Science Route 1208, Create Road Gloucester Point, Virginia 23062 Manuscript submitted 23 October 2003 to the Scientific Editor's Office. Manuscript approved for publication 15 September 2005 by the Scientific Editor. Fish. Bull. 104:278-292 (2006). The salmon shark (Lamna ditropis) is a large apex predator inhabiting the coastal and oceanic waters of the North Pacific Ocean, most commonly ranging from 65°N latitude to 35°N in the west and to 30^N in the east (Strasburg, 1958; Neave and Hana- van, 1960; Compagno, 1984, 2001; Blagoderov, 1994; Nagasawa, 1998). It is found individually and in large aggregations at sea-surface tempera- tures of 5°C to 18°C and has a depth distribution ranging from the surface to at least 150 m (Compagno, 1984, 2001). The salmon shark is a highly opportunistic predator, feeding on a wide variety of prey and sharing the highest trophic level of the food web in boreal and temperate North Pacific waters with marine mammals and seabirds (Brodeur, 1988; Nagasawa, 1998). Adult salmon sharks typically range in size from 180-210 cm precau- dal length (PCL; = 200-260 cm total length, TL) and can weigh up to 220 kg (Tanaka, 1980; JAMARC, 1980; Naga- sawa, 1998). Reported lengths of 300 cm TL and greater and weights exceed- ing 450 kg are unsubstantiated. Salmon sharks are migratory in nature. North-south migrations have been documented on both sides of the North Pacific; northern movements occur in spring and southern move- ments are occur in autumn (lino, 1939; Kosugi and Tsuchisaki, 1950; Gorbatenko and Cheblukova, 1990; Balgaderov, 1994; Nakano and Na- gasawa, 1994). However, this species is present in boreal waters through- out the year (Goldman and Human, in press; Weng et al., 2005). Salmon sharks tagged with satellite transmit- ters have shown an extremely wide range of movements throughout the eastern North Pacific — movement as far south as southern California and Hawaii (Weng et al., 2005). To date, however, no salmon shark has been recorded moving across the Pacific, but such movements are suspected to occur (Nakano and Nagasawa 1996; Goldman and Musick, in press; Weng et al., 2005) Sexual segregation is relatively common in sharks; however an ex- tremely large sex ratio difference exists in salmon sharks across the North Pacific basin (Sano, 1962; Na- gasawa. 1998; Goldman and Musick, in press). The western North Pa- cific (WNP) is male dominated and the eastern North Pacific (ENP) is female dominated. Male dominance in the WNP and female dominance in the ENP increase with increasing latitude. Larger sharks range farther north than smaller individuals, and Goldman and Musick Growth and maturity of Lamna ditropis 279 southern catches generally occur in deeper waters (Na- gasawa, 1998; Goldman and Musick'). Shark catches in Alaska waters have been reported to be as high as those from Washington, Oregon and California combined (Camhi, 1999), and there is con- cern over the amount of shark bycatch being taken (Gaichas^). In 1997, state managers closed commercial shark fishing in Alaska state waters and imposed con- servative sport-fishing limits (1 shark per person per day, 2 per year) that legally encompass federal waters too. Federal managers are currently addressing elasmo- branch management issues (Goldman, 2001). Because of the ever-increasing importance of provid- ing accurate and timely life history parameters to foster responsible management efforts, we had two main ob- jectives in our study. The first was to estimate growth parameters and the age and length at sexual matu- rity of salmon sharks in the ENP. The second was to compare our results to those previously published on salmon sharks from the WNP to elucidate any exist- ing variability in the life history parameters of this highly sexually segregated population. Lastly, because of the small number of samples for female juvenile age classes, a small overall sample size for males, and the fact that samples (for both sexes) were not obtained for each month, we applied several back-calculation meth- ods to our data (see Cailliet and Goldman, 2004 and Goldman, 2004) and provide a demonstration and brief discussion on the importance of choosing an appropriate back-calculation method. Materials and methods Salmon shark vertebrae were obtained from numer- ous sources and locations ranging geographically from southern California to Kodiak Island (»=182) between 1997 and 2002. The majority of samples came from research cruises in the Gulf of Alaska (GOA) and Prince William Sound (PWS), by accompanying sport-fishing vessels on salmon shark trips in the GOA and PWS, and from Alaska Department of Fish and Game (ADF&G) port samplers. Vertebral samples collected from British Columbia to southern California came from incidental catches and recently beached animals. Precaudal, fork, and total length (PCL, FL, and TL) were measured on a straight line and recorded along with date, location, and sex (and clasper length in males). Samples provided by additional sources included at least one of the necessary straight-line measurements, date and location of capture or wash-up, and sex. Weights were obtained from 20 females ranging from 67.1 to 213.4 cm PCL, and 7 males ranging from 69.5 to 187 cm PCL. The power function, W = aL'\ was fitted to the data (in SigmaPlot, vers. 6.0, SPSS Inc., Chicago, IL), where W = weight (kg) and L = length (cm PCL). Specimens weighed had been collected from southern California to Alaska. A likelihood ratio test was used to determine whether differences between female and male weight- length parameters were significant or if a single set of parameters described the data better (Kimura, 1980; Quinn and Deriso, 1999; Haddon, 2001) (SAS, vers. 8.0, SAS, Cary, NO. A 20- to 25-cm section of vertebrae was removed from directly in front of the first dorsal fin (above the gills) and stored frozen. We used PCL mea- surements to make direct comparisons with previously published salmon shark age and growth data from the WNP. Linear regression equations, based on our samples, were developed for converting PCL to FL and TL. Vertebral samples were thawed, cleaned of excess tissue, separated into individual centra and stored in 70% ethyl alcohol for at least 24 h. Centra were sagit- tally sectioned immediately adjacent to the center of the focus and then cut again approximately 1.5 mm off- center by using an Isomet rotary diamond saw (Buehler, Lake Bluff, IL). The sections were pressed between two pieces of Plexiglas (to prevent warping), air-dried for 24 h under a ventilation hood, and then mounted onto mi- croscope slides with a mounting medium. After drying, sections were polished with wet fine grit (320-, 400-, and 600-grit) sand paper to approximately 0.5 mm and air-dried. Sections were viewed by using a binocular dissecting microscope with transmitted light. Centrum radius (CR) and distance to the outer mar- gin of each ring were measured to the nearest 0.001 mm along a straight line from the central focus to the center of the outer margin of the corpus calcareum (Fig. lA) using a compound video microscope with the Optimus image analysis system (Media Cybernetics, vers. 6.5., Silver Spring, MD). PCL was plotted against CR to determine the proportional relationship between somatic and vertebral growth. A banding pattern was readily distinguishable in sectioned centra; wide bands were separated by distinct narrow opaque bands. This pattern occurred on both arms of the corpus calcareum and extended across the intermedialia. Each pair of wide-narrow bands was considered a growth cycle; the narrow opaque bands, hereafter referred to as "rings," were counted (Fig. lA). An angle change in the intermedialia and a ring on the corpus calcareum were present approximately 5 to 6 mm from the focus of each centrum and were con- sidered to represent a birthmark. The "prebirth ring" reported by Nagasawa (1998) was present in most speci- mens just prior to the birthmark but was not counted nor measured (Fig. lA).-^ A distinct notch was usually ' Goldman, K. J., and J. A. Musick. 2002. Unpubl. data. Alaska Department of Fish and Game, 3298 Doug- las PI., Homer, AK 99603, and Virginia Institute of Marine Science. 1208 Greate Rd. Gloucester Ft.. VA 23062. - Gaichas, S. 2002. Personal commun. NMFS Ala.ska Fish- eries Science Center, 7600 Sand Point Way NE, Building 4, Seattle WA 98115. ■^ Our aging protocol was independently developed by the senior author and a co-reader, and later found to be identical to that used for salmon sharks in the western North Pacific i H. Nakano, pers. commun.) Vertebral images were exchanged and aged by the senior author and H. Nakano (and read randomly and blindly). The results validated that aging was conducted in an identical manner between studies. 280 Fishery Bulletin 104(2) present along the inside edge of the cor- pus calcareum at each ring, providing an additional aging feature, particularly in sections where the cut excluded the radi- als of the intermedialia (Fig. IB). Annulus measurements were made from the origin of a wide band to the outer margin of the adjacent narrow band. Two readers independently aged all cen- tra two times in blind, randomized trials. This procedure allowed the calculation of within-reader precision, and between-reader precision twice. When there was a disagree- ment between readers, a final age determi- nation was made by both readers viewing the centrum together. Percent agreement {PA=lNo. agreed/No. read]xlOO). and per- cent agreement plus or minus one year (PA [±1 yr]) were calculated for length groups of 10 cm to test for precision. The criticism of percent agreement as a measure of preci- sion has been that it varies widely among species and ages within a species (Beamish and Fournier, 1981; Campana, 2001). The criticism regarding variation among species is irrelevant becauses one is only interested in the particular species one is aging (i.e., these particular species are not compared among other species), but the criticism re- garding variation of ages within a species is not only relevant, but is true. There is, however, validity in using percent agreement with individuals grouped by length as a test of precision because it does not rely on ages (which have been estimated), but rather on lengths, which are empirical values (Goldman, 2004; Cailliet and Goldman, 2004). Ages could be used if and only if, all age classes have been validated. The most commonly used methods for evaluating pre- cision among age determinations have been the average percent error (APE ) technique of Beamish and Fournier (1981) and the modification of their method by Chang (1982). However, Hoenig et al. (1995) have demonstrated (using the Beamish and Fournier, 1981, data) that there can be differences in precision that these methods ob- scure because with the APE it is assumed that the variability among observations of individual fish can be averaged over all age groups and that this variability can be expressed in relative terms. Also, APE indices do not result in values that are independent of the age estimates, do not test for systematic differences, and do not distinguish all sources of variability (such as differences in precision with age) (Hoenig et al., 1995). Good APE values appear to tell us only which reader was less variable, not which was better or if either was biased, which is more critical in knowing whether reliable ages can be produced and replicated (i.e. is the error within and between readers due to random error or a systematic bias). Campana et al. (1995) stated the importance of a separate measure for bias, and Hoenig et al. (1995) and Figure 1 lA) Sagittal section of a 10-yr-old salmon shark iLamna ditropis) vertebral centrum showing typical banding pattern. CR = centrum radius. (B) Portion of a sagittal section from a salmon shark verte- bral centrum without intermedialia, showing the distinct notching pattern (white arrows) that accompanies the banding pattern and that is used to aid in assessing ages. The 1.0-mm bar applies to both A and B. PB = prebirth ring, B = birth ring and numbers indicate rings or age in years. Evans and Hoenig (1998) suggested testing for sys- tematic differences between readers using chi-square tests of symmetry to determine whether differences are systematic (biased) or due to random error. These are statistically rigorous and effective methods for detect- ing bias (Campana, 2001) and were conducted in the present study. These techniques place all age values in contingency tables and test the hypothesis that values in a given table are symmetrical about the main diagonal (Hoenig et al, 1995; Evans and Hoenig, 1998). They can also be set up to test among all individual age classes or groups of age classes (Hoenig et al., 1995). The test statistic (the chi-square variable) will tend to be large if a systematic difference exists between the two readers. A relative marginal increment (RMI) analysis was used to verify the temporal periodicity of ring forma- tion in the vertebrae. This is a standardized marginal increment analysis whereby the margin, or growth area of a centrum from the last narrow growth ring to the centrum edge, is divided by the width of the last fully formed growth increment (Branstetter and Musick, 1994; Conrath et al., 2002). Resulting RMI values were compared to the month of capture. This analysis was performed on immature and mature sharks separately and combined. Age-zero animals were not included (be- cause they have no fully formed increments). The von Bertalanffy growth function was fitted to the vertebral age-at -length data for salmon sharks from the ENP with a nonlinear least squares regression al- gorithm (."nls" in S-Plus, Professional Release 1, Math- Goldman and Musick: Growth and maturity of Lamna ditropis 281 soft Inc., Seattle, WA) to estimate parameters. The von Bertalanffy growth function is L, =L„[l-exp(-^(^ -/(,))], where L, = length at age 7'; L^ = asymptotic or maximum length; k = the growth coefficient; and t^ = age or time when length theoretically equals zero. Growth parameters were estimated for the sexes sepa- rately and combined. A likelihood ratio test was used to determine whether differences between female and male growth parameters were significant or if a single set of growth parameters described the data better (Kimura, 1980; Quinn and Deriso, 1999; Haddon, 2001) (SAS, vers. 8.0, SAS, Gary, NO. Back-calculation is a method for describing the growth history of each individual sampled, and numerous varia- tions in methods exist (see Francis, 1990, for a thorough review and Goldman, 2004, for a description and ap- plication to elasmobranchs). Back-calculations estimate lengths-at-previous-ages for each individual and should be used if sample sizes are small and if samples have not been obtained from each month (Goldman, 2004; Cailliet and Goldman, 2004). Cailliet and Goldman (2004) stated "the proportional relationship between animal length or disk width and the radius of the vertebral centrum among different length animals within a population is used as a basis for empirical relationships regarding population and individual growth, as is the distance from the focus to each annulus within a given centrum." Hence, choosing the appropriate method (based on the re- lationship between animal length or disk width and the radius of the vertebral centrum) is critical if back-calcu- lated data are to be used for obtaining accurate life his- tory parameter estimates from growth function models. Because smaller size classes were not well represented in our female sample and because our male sample size was small, lengths at previous ages were back-calculated from centrum measurements for both sexes and fitted with the von Bertalanffy growth function. The relationship between OR and PCL for ENP salm- on sharks was used to determine the most appropriate method for back-calculating previous length-at-age. To our knowledge, however, no studies on elasmobranch fishes have examined multiple back-calculation methods for statistical or biological accuracy in relation to ver- tebral sample data. To that end, four different propor- tion methods were used and compared with our sample length-at-age data. We first applied the standard Dahl- Lea direct proportions method (Carlander, 1969): L,{LJCR,.}xCR^, where L, = length at ring 'f ; L, = length at capture; CR^ = centrum radius at capture; and CR^ = centrum radius at ring '/'. (1) Next, we applied two modified versions of the Dahl-Lea method that use parameter estimates from the specific linear and quadratic fits that described the PCL-CR rela- tionship. The linear-modified Dahl-Lea method (Francis, 1990) is L, = L. X [{a + hCR,)/{a + bCR^.)], (2) where 'a' and 'h' are the linear fitted parameter esti- mates. The quadratic-modified Dahl-Lea method (Fran- cis, 1990) is L, = L,, X [ia + bCR, + cCRf)/{a + bCR,. + cCR^ )], (3) where 'a' 'b.' and V are the quadratic-fitted parameter estimates. Kicker (1992) applauded Francis's (1990) back-calcu- lation review paper, but like Campana (1990) suggested that the point of origin of proportional back-calcula- tions should be related to a biologically derived inter- cept (i.e., length at birth). We, therefore, also applied Campana's (1990) "size-at-birth-modified" Fraser-Lee equation: L, = L +[(Ci?, -Ci?,,)x(L,,-L;j„.„, )/(C/e,.-Ci?B„,,, )], (4) where L^,,.,,, = length at birth; and ^^Biiih - centrum radius at birth. (Based on Tan aka, 1980, 62.5 cm PCL was used for Von Bertalanffy growth parameter estimates were obtained from all individual and mean back-calculated length-at-age, and from a combination of back-calculated lengths-at-age and our sample data. A likelihood ratio test was used for all scenarios to determine whether differences between female and male growth parameters were significant or if a single set of growth parameters described the data better (Kimura, 1980; Quinn and Deriso, 1999; Haddon, 2001) (SAS vers. 8.0, SAS, Gary, NC). Back-calculated length-at-age results from all four methods were examined to see which best reflected our vertebral sample data. The reproductive tracts of 64 female and 14 male salmon sharks were examined visually to assess their reproductive status. Females ranged in size from 71 to 209 cm PCL, and males ranged from 63 to 187 cm PCL. Clasper lengths were obtained from 12 of the males (from 91 to 187 cm PCL). Gross analysis of reproduc- tive tracts and maturity determinations were made ac- cording to the methods of Pratt (1988), Gilmore (1993), Pratt and Tanaka (1994), Hamlett (1999) and Hamlett and Koob (1999). Median precaudal length-at-maturity (MPCL) was determined by first coding female (n = 64) and male (n=14) maturity data into binary form, with 0=imma- ture and l=mature. The binary data were fitted with a logistic regression model ("GLM" in S-Plus, Professional Release 1, Mathsoft Inc., Seattle, WA). The median precaudal length-at-maturity was then estimated as 282 Fishery Bulletin 104(2) MPCL = -alb (a=intercept, 6 = slope) (Mollet et al., 2000; Neer and Cailliet, 2001; Con- rath and Musick, 2002; Conrath, 2004). To estimate percent maturity by length for the ENP salmon shark population, the 'a' and '6' estimates (from the "GLM" model) were substituted into the equation; percent ma- ture = 100x[(e'"+*'<'''"^"")/(l+e"'+*'<'''"«"")], and plotted against PCL. A nonparametric boot- strap (n = 1000 replications) was conducted with S-Plus and 95*^ confidence limits were obtained from the 2.5th and 97.5th empirical percentiles. Sperm storage in the oviducal gland has been documented in several shark species (Pratt, 1993). Oviducal glands were taken from six mature females, ranging in size from 180 to 192 cm PCL, caught in PWS waters in late July and late August, to de- termine the presence or absence of sperm. Samples were initially fixed in 10% forma- lin. Samples were rinsed extensively to re- move fixative and then stored in 70% etha- nol. Cross-sections along the entire length of each gland were trimmed, dehydrated in a graded series of ethanol, cleared in a limonene-based solvent, infiltrated with paraffin, and embedded in paraffin blocks. Serial sections (5 um) were prepared by using a rotary microtome, mounted onto poly-L-ly- sine-coated slides, dried, deparaffinized, rehydrated, and stained by using standard Harris hemotoxylin and eosin in order to examine general cellular detail (Hin- ton, 1990). Sections were examined for the presence of sperm by using a compound microscope at magnifica- tions ranging from 100 x to 1000 x. Results Length and weight equations One hundred sixty-six of the 182 salmon sharks in this study were female, resulting in a F:M sex ratio of 10.4/1. Although females dominated our sample, the following equations for converting PCL into FL and TL appear to work equally well for both sexes. 225 >'*' .. .^ • f^ ^ 200 ■ . *isteti^^X«* S 175- .c ^^JT-*^ (t'» ■* •* ? 150- • ^^'^ ' /V'' Quadratic fit ' ^* PCi.= (-0 583 ■ CR2) + (25 189 ■ CRl- 63 94- ZJ >• • (J J^ r^ ^ 0 94 (D ^* / Linear fit X / 75 - " / PCL =(10 553 ■ CR) + 20 964 Ji/^ f^ ^ 0 90 50 ■! ' 1 ■ ■ ' ■ ' • ■ ' ' 1 ' 5 10 15 20 Centrum radius (mm) Figure 2 Relationship between centrum radius and precaudal length for eastern North Pacific salmon sharks (Lamna ditropist showing significant fits given by linear and quadratic equations (sexes com- bined, ;! = 182i. PCL = precaudal length, CR = centrum radius. FL = 1.08 X PCL + 6.91 TL = 1.15 xPCL + 15.19 ir- (r- = 0.99; n-- = 0.97; n-- :138) 133). Weight-length relationships for female and male salmon sharks we sampled in the ENP were W= 8.2xl0-05xL2 "9 (r-' = 0.99, ??=20) and W=3.2x 10-06xL3383 (;.2 = o.99, n=l) respectively, and 4.4x 10-°^xL2875 (^2_o.99) for the sexes combined. A likelihood ratio test showed that separate equa- tions for each sex describe the data better than a single equation for the sexes combined (x^ = 12.1; df=3; P<0,01). Vertebral analysis There was a slight curvilinear relationship between cen- trum radius (CR) and shark PCL (Fig. 2). A linear regres- sion gave a significant fit to the data (PCL = ( 10.553 x Ci?)-h20.964; r- = 0.90; P<0.0001); however, a quadratic equation produced a slightly better fit ( PCL = -63. 944 -i- (25.189xCR)-(0.583xCi?2); ^-' = 0.94; P<0.0001) and a t-tesi showed the third parameter to be statistically significantly different from zero (/ = 10.32; df=181, P<0.0001). However, it was necessary to compare the mean back-calculated results from Equations 1 through 4 with the mean sample PCL data to see if the slightly better statistical fit of the quadratic equation translated into better biological accuracy for modeling growth. Percent agreement (PA) among readers was 68.1% for the first set of blind reads and 72.0% for the second set, and the within-reader PA was 79.7%f for reader one (the lead author) and 75.3%c for reader two. Percent agree- ment (±one yr) was >95%f for all reader comparisons. Agreement between and within reader age assessments was 100% until age 7 or 8, depending on the reader comparison. When grouped by 10-cm length increments, percent agreement was 100%^ for sharks <160 cm PCL, and 100% ±1 yr for sharks ^180 cm PCL (Table 1). (PA and PA ±1 for length groups were virtually the same for both sets of between-reader comparisons, but slight differences occurred above 160 and 180 cm PCL, re- spectively.) The chi-square tests of symmetry gave no indication that differences between and within readers were systematic rather than due to random error (x^ test, P>0.05 in all cases). Relative marginal increment (RMI) analysis showed that postnatal rings form annually between January and March. The smallest relative margins occurred in Goldman and Musick Growth and maturity of Lamna dilropis 283 Table 1 Percent agreement (PA) and PA ± one year, from the second set of ditropis) when placed into 10-cm precaudal length (PCD groups. 'reads," for eastern North Pacific salmon sharks (Lamna PCL(cmi Tota 1 no. read No agreed upon No agreed upon ±1 PA PA±1 60-70 5 5 5 100 100 70.1-80 2 2 2 100 100 80.1-90 0 — — — — 90.1-100 4 4 4 100 100 100.1-110 1 1 1 100 100 110.1-120 5 5 5 100 100 120.1-130 3 3 3 100 100 130.1-140 1 1 1 100 100 140.1-150 6 6 6 100 100 150.1-160 9 9 9 100 100 160.1-170 11 8 11 72.7 100 170.1-180 30 22 30 73.3 100 180.1-190 55 38 53 69.1 96.4 190.1-200 34 22 32 64.7 94.1 200.1-210 12 4 10 33.3 83.3 >210 4 1 3 25.0 75.0 n 182 131 175 — — Percent agreement — — — 72.0 96.2 February and March, followed by a consis- tent increase in RMI, and the largest relative margins occurred in December and January. There was no difference in this trend when immature and mature sharks were examined separately; therefore all ages classes were combined (Fig. 3). Vertebral age-at-length data from 166 fe- male salmon sharks provided von Bertalanffy parameters of L^=207.4 cm PCL, it = 0.17/yr, and tg = -2.3 years (Fig. 4). When the von Bertalanffy growth function was fitted to the quadratic-modified Dahl-Lea back-calculations for females, the life history parameters were similar to those produced from the sample length data (Table 2). Vertebral age data from 16 males provided von Bertalanffy parameters of L., = 182.8 cm PCL, A' = 0.23/yr, and tQ=-1.9 years (Fig. 4). Back-calculated lengths-at-age for male salmon sharks, either with or without sample data included, provided slightly lower k coefficients (=0.20), but similar L^ and /,, parameters (Table 2). Vertebral age data for the sexes combined («=182) provided von Ber- talanffy parameters of L^=204.5 cm PCL, /e = 0.18/yr , and ?||=-2.2 years. A likelihood ratio test showed that separate von Bertalanffy growth models describe the data better for each sex than one model with the sexes combined (x2=29.1; df=3; P<0.001). The quadratic-modified Dahl-Lea method (Eq. 3) rep- resented most accurately the mean sample precaudal 10 : n=2 09 ] 0 n-3 n-3 „ c 08 - 8 ° £ n-3 n=4 o n=19 ^ o u n=1l3 8 0 o Z 06 ■ o o c 8 1 0 o CT 05 : nJ E 0.4: "=15 1 ° - 0.3 : n=6 1 8 0) DC 02 n=2 n=2 8 01 : o e ° o , ' . ( )123456789 10 11 12 Month Figure 3 Result > of relative marginal increment analysis showi ng that annulu s formation occurs between January and M arch ( /!=176). length-at-age data. It produced mean back-calculated lengths-at-age within 4.2 cm of mean sample lengths- at-age for female salmon sharks >120 cm PCL and within 8.8 cm for females <120 cm PCL (Fig. 5A). When applied to males. Equation 3 produced mean back-cal- culated lengths-at-age within 10 cm of mean sample lengths-at-age for sharks <120 cm PCL and >150 cm 284 Fishery Bulletin 104(2) Table 2 Von Bertalanffy growth parameters for female, male and sexes combined for salmon sharks (Lamna dilropis ) in the eastern and western North Pacific. Numbers in parentheses are standard errors. * = not independent. i. k ^1 Eastern North Pacific (present study) Females Vertebral sample data l;i = 166) 207.4(2.5) 0.17(0.01) -2.3(0.2) Back-calculations (;i = 1814"i 205.3(0.9) 0.18(0.002) -2.0(0.03) Mean back-calculation («=21*i 210.0(1.1) 0.17(0.004) -2.1(0.09) Back-calculations with sample data (fi=1980*) 206.0(0.8) 0.18(0.002) -2.0(0.031 Males Vertebral sample data (;i = 16) 182.8(3.7) 0.23(0.03) -1.9(0.3) Back-calculations (/! = 161*) 183.9 (2.0) 0.20(0.008) -2.0(0.1) Back-calculations with sample data (/i = 177*) 184.2(1.9) 0.20(0.008) -2.0(0.1) Combined Vertebral sample data l;? = 182) 204.5(2.4) 0.18(0.01) -2.2(0.2) Western North Pacific (Tanaka, 1980) Females Vertebral sample data 203.8 0.14 -3.9 Males Vertebral sample data 180.0 0.17 -3.6 PCL. In between those lengths, however, the deviation from mean sample length-at-age increased to 16.5 cm (Fig. 5B). (The larger deviations for males were likely the result of the small sample size.) Lee's phenomenon did not occur in either sex with the quadratic-modified Dahl-Lea method. Of the 64 female salmon shark reproductive tracts examined, 55 were fully mature, and 11 of the 14 males examined were mature. Mature female reproductive n ° o o 0 B ° 200 ■ o ° LJ— l^f"^ ° 180 • _o 160 ■ Ol 140 ■ 8 8 il-^f^T^..^-^- le o^ Females (o) Males (■ ) CO 120 ■ §y^ /..= 207.4 L. = 182.8 (0 o/ /c=0.17 /( = 0.23 80 • ■/ 0 5 10 15 20 Age (yr) Figure 4 Von Bertalanffy growth curves fitted to female (?i = 166) and male (n = 16) vertebral sample data for eastern North Pacific salmon sharks {Lamna ditropis). Estimates for parameters of the von Bertalanffy growth function are summarized. tracts consisted of a well-developed right ovary (the only functional ovary in lamniform sharks) containing various sizes of ova, large oviducts and oviducal glands, and expanded, heavily striated uteri with thickened walls. In stark contrast, the right ovary from immature female reproductive tracts was considerably smaller and had no signs of developing ova, the oviducal gland was virtually indistinguishable from the oviduct, and the uteri were extremely small, thin-walled, flaccid and completely smooth. Additionally, the epigonal organ in mature females was large and ex- tended '-/,3 the length of the uteri, whereas in immature females it was small and generally did not extend much past the anterior end of the uteri. No females examined were of an "intermediate'Veproductive nature. Mature male reproductive tracts consisted of large well-developed testes, a thick epididymis lead- ing to a coiled ductus deferens, rigid (fully cal- cified) claspers and well-developed siphon sacs. Immature male reproductive tracts consisted of small testes that were partially embedded in the epigonal organ, thin epididymis and straight ductus deferens, nonrigid claspers and poorly developed siphon sacs. Estimated median precaudal length-at- maturity (MPCL) for ENP female and male salmon sharks were 164.7 cm PCL and 124.0 cm PCL, respectively (Fig. 6). The smallest mature female observed was 164.0 cm PCL, and the largest immature female observed was 176.5 cm PCL (however, all other im- mature females were sl63 cm PCL). None of Goldman and Musick Growth and maturity of Lamna ditropis 285 Table 3 Mean ages and precaudal lengths PCLi for eastern North Pacific salmon sharks iLamna dilropis 1 by location. CA = - California; seAK = southeast Al aska. Females Males CA to seAK Alaska Combined CA to seAK Alaska Combined Age ( yr) Mean 2 11 10 3 12 9 Range 0-5 5-20 0-20 0-7 6-17 0-17 PCL (cm) Mean 103.6 184.7 175.0 113.5 171.9 153.7 Range 62.2-153 144.8-213.4 62.2-213.4 63-150.5 155.4-187 63-187 n 20 146 166 5 11 16 the oviducal glands sampled from ma- ture females contained spermatozoa. The smallest mature male observed was 155.4 cm PCL, and the largest immature male was 91.0 cm PCL. Based on age estimates from vertebral sample data, ages at MP- CL range from 6 to 9 yr for females and from 3 to 5 yr for males. Ninety-five per- cent confidence limits (for percent mature) were very narrow for females <155 cm PCL and >172.5 cm PCL and for males <107.5 cm PCL and >132.5 cm PCL. The wide vertical confidence intervals around the calculated estimates of MPCL (Fig. 6) were caused by not having many observa- tions for lengths around the estimated MPCL in the sample. Age-length composition in the Eastern North Pacific The mean age and length composition of ENP salmon sharks was latitude depen- dent (Table 3). Sharks between 0 and 7 yrs old, ranging from 62.2 to 153 cm PCL, were collected from southern California to southeast Alaska, with smaller indi- viduals found in the southern part of that range. Sharks between 5 and 20 yrs old, ranging from 144.8 to 213.4 cm PCL, were collected from southeast Alaska up into the Gulf of Alaska, Prince William Sound, and the Bering Sea. Discussion Vertebral growth significantly increased with somatic growth (Fig. 2), which, along with the reliability of the RMI analysis 30 25 :■ 20 :■ 15 :■ 10 :■ 5 ■■ O Dahl-Lea □ Dahl-Lea linear-modified • Dahl-Lea quadratic-modified O Fraser-Lee -modified 190 200 210 220 JU q B O Dahl-Lea D Dahl-Lea linear- modified 20 • Dahl-Lea quadradlic-moditied '^1 ■ D O Fraser-Lee birth -modified 10 5 ; 0 4 • • -5 : -10 ; -15 : -20 • s o e o o [ • D a 9 8 • D @ • -25 -in - — 1 . 1 . 1- O O i r 1— Sample mean precaudal length-at-age (cm) Figure 5 Mean deviation, from mean sampled precaudal length, of four pro- portional back-calculation methods for (A) female, and for (B) male eastern North Pacific salmon sharks (Lamna ditropis). Data points represent mean back-calculated lengths-at-age. An x-axis point on the zero line extending from the y-axis would represent zero deviation from the sample mean length-at-age. 286 Fishery Bulletin 104(2) / 1 ' / ' 1 1 /' 1 1 1 1 1 // 1 1 1 / 1 .yji 60 80 100 120 140 160 180 200 220 Precaudal length (cm) Figure 6 Percent versus precaudal length curves for eastern North Pacific female (heavy lines) and male (light lines) salmon sharks (Lamna ditropis) with 95'?- confidence bands on percent mature. Diamond and circle show estimates of median precaudal length at maturity (MPCL) for females and males respectively. Confidence bands (dotted lines) are 2.. 5"" and 97.5'*' empirical percentiles obtained by bootstrapping. (Fig. 3 1, demonstrates that vertebral growth patterns are a reliable indicator of age in salmon sharks. Preci- sion was high between and within readers with limited differences (Table 1) that were attributable to random error. These results provided us with a high degree of confidence in the age assessments determined from the von Bertalanffy growth model (with vertebral sample data), and hence in the resulting life history parameter estimates. Tanaka (1980) and Nagasawa (1998) stated that salm- on sharks produce one ring per year, but a RMI analy- sis was not provided in their studies. Our RMI analysis verified an annual periodicity of banding patterns for salmon sharks, ranging from 73 cm to 213.4 cm PCL and encompassing ages 1 to 20 for females and ages 1 to 17 for males. The major period of growth occurs from May through November, slowing some as January ap- proaches (Fig. 3). A brief cessation, or extreme slowing, of growth (ring formation) occurs between January and March, and growth increases again in April-May. Al- though we were able to examine specimens from every month of the year, additional samples from December through April would enhance RMI precision. The similar von Bertalanffy growth parameter es- timates generated from female vertebral sample data, back-calculated data, and the combination of the two indicated that sample size was sufficiently large and encompassed the known size range of the species. Von Bertalanffy parameters for males would have been improved with a larger sample size that would have reduced the difference between values of 'k' from ver- tebral sample and back-calculated data (Table 2). More samples would likely have had little influence on L ^ for either sex because salmon sharks close to maximum length were examined in the present study. It is pos- sible that the discrepancy between female and male sample sizes could have influenced the outcome of the likelihood ratio test. However, with the differences in observed maximum length (and estimated L,) between the sexes and the small standard errors associated with the male von Bertalanffy estimates (Table 2). it is unlikely that an increased male sample size would have altered the test result. The error associated with back-calculated length ver- sus the actual length at a given age has been a focal point of papers by Campana (1990), Francis (1990) and Ricker (1992) and prompted our evaluation of several proportional back-calculation methods. There was a statistically valid reason for using the quadratic-modi- fied Dahl-Lea (Eq. 3) over the linear-modified Dahl-Lea (Eq. 2) (see "Results" section); however, the only way to cross-compare all four back-calculation methods in a biologically meaningful way was to apply all of them to the vertebral sample data. Both modified Dahl-Lea equations were more accurate in representing the mean sample length-at-age data than the standard Dahl-Lea or the size-at-birth-modified Fraser-Lee equations (Fig. 5, A and B). However, the quadratic-modified Dahl-Lea was the best predictor of prior length-at-age (i.e., best resembled our vertebral sample length-at-age data). Although these back-calculation results are, of course, dependent on the assumption that growth has not sig- nificantly changed over time, and are applicable only to salmon sharks, they demonstrate the importance of choosing the appropriate method to minimize error, which results in a greater ability to accurately model growth. Accurate growth parameters will result in more accurate demographic parameters and stock assess- ments, leading to more responsible management. Even greater confidence could be achieved if animals collected in the past were available because they would allow a direct comparison of size-at-age then and now in order to verify that the growth rate for the salmon shark has not changed. Our results show that salmon sharks in the ENP attain their maximum size at a faster rate (k) than those from the WNP (Table 2). We were unable to test the statistical significance of these differences because neither Tanaka (1980) nor Nagasawa (1998) provided point estimates and standard error values for their WNP data. There were, however, significant resultant biological differences present. Both female and male salmon sharks reach first age at sexual maturity ap- proximately 2 years earlier in the ENP than in the WNP. Female salmon sharks in the WNP mature be- tween 8 and 10 years of age (Tanaka, 1980; Nagasawa, 1998). From the reproductive tracts examined, we found that female salmon sharks in the ENP reach sexual maturity between ages 6 and 9. Although age at first maturity was earlier in the ENP, length at maturity appears to be similar; 160-180 cm PCL in the WNP, and a MPCL = 164.7 cm PCL in the ENP (Fig. 6). Male salmon sharks in the WNP mature at approximately Goldman and Musick; Growth and maturity of Lamna ditiopis 287 5 years of age and 140 cm PCL (Tanaka, 1980; Naga- sawa, 1998). From our small sample size, the logistic method provided an estimate that male salmon sharks in the ENP mature between the ages of 3 and 5, and MPCL = 124.0 cm PCL (Fig. 6). Clasper lengths did not enhance the ability to determine male length at sexu- al maturity because specimens around the calculated MPCL were not sampled. Although the MPCL for males indicates a smaller length at maturity, our confidence in the accuracy of the male MPCL estimate (due to sample size) is low (Fig. 6). The lack of spermatozoa in any of the oviducal gland samples analyzed is not surprising considering no 1am- niform shark examined to date has shown evidence of sperm storage (Pratt, 1993). However, specimens we examined were specifically chosen from the time pe- riod of late July through late August when mating was likely to have been taking place (Goldman and Human, 2005; Goldman and Musick, in press). A larger sample size taken through that time period and extending through September may resolve the question better regarding the suspected residence time of sperm in the oviducal gland of lamniform sharks — a residence time that is probably only a few days timed with ac- tual mating activity (Pratt, 1993). The large number of eggs produced to feed oophagous young (Wourms, 1977; Gilmore et al., 1983; Gilmore. 1993) could have "flushed" any stored sperm out of the oviducal gland (Pratt, 1993). In the WNP, salmon sharks are born in the spring of the year and pups range in length from 60 to 70 cm PCL (Tanaka, 1980; JAMARC, 1980; Nagasawa, 1998). Litter size in the WNP is up to 5 pups, and the ratio of male to female is 2.2:1 (Tanaka, 1980). Our data support a similar timing for pupping and for length at birth in the ENP, but no pregnant females were taken during the course of our research and therefore infor- mation on litter size and pup sex ratio is not available for the ENP. Nagasawa (1998) characterized salmon shark growth through age ten as follows; after birth, "they grow to between 90 and 105 cm PCL by the next spring ... subsequent annual growth is 10-15 cm per year up to age-4 fish." He stated, in addition, that females and males attain an average length of 173 cm PCL and 163 cm PCL, respectively, by age 10. Our data indicated that the average growth rate (cm per year) for salmon sharks in the ENP is very similar to the rate in the WNP through age 4. By age 5, ENP females are growing at a considerably faster rate than WNP females, reaching an average of 185.4 cm PCL by age 10 (Table 4). Male salmon sharks in the ENP appear to begin outgrowing their WNP counterparts by age 4 (Table 4). Salmon shark longevity appears to be similar in the ENP and WNP. Maximum observed age for females in our study was 20, whereas it was 17 in the WNP (Tanaka, 1980); however, maximum observed age for males was only 17, whereas it was 25 for the WNP. We were unable to obtain samples of male salmon sharks at or near maxi- mum age because of our small sample size. Salmon sharks in the ENP and WNP attain the same maximum length (about 215 cm PCL for females and about 190 cm PCL for males). However, WNP salmon sharks take longer to reach a given length than those in the ENP (Table 4). Additionally, the weights for the same length females and males are considerably dif- ferent except at male maximum length, which appears to be equivalent to females of similar length. Our data indicate that the weight-at-length differences between ENP and WNP salmon sharks begin at approximately 110 cm PCL for females, and at 140 cm PCL for males (Fig. 7, A and B). Although a greater sample size could enhanced precision and may alter the equations pre- sented in the present study to some degree, the data show that weight-at-length is particularly greater for adult salmon sharks in the ENP than in the WNP. As with the von Bertalanffy growth models, it is possible that the discrepancy between female and male sample sizes may have influenced the outcome of the likelihood ratio test. However, the sample-size discrepancy was not as large as with the growth curve data and standard errors were small for both sexes; therefore it is unlikely that an increased male sample size would have altered the test result. Although season and size at birth appear to be simi- lar for salmon sharks in the ENP and WNP, the loca- tions where females pup are different. A salmon shark pupping and nursery ground exists along the transi- tional boundary of the subarctic and central Pacific currents (Nakano and Nagasawa, 1996). Our data indi- cate that a second pupping and nursery ground exists, ranging from southeast Alaska to the northern end of Baja California, Mexico (central California being the most common area for ages zero and one [see Table 3] ). Ages zero through 5 were caught only between southern California and southeast Alaska, and ages 5 and above were caught only in Alaska waters. The latitudinal size segregation observed in the ENP indicates that preg- nant females may move south along the California coast to give birth in the spring. The high degree of sexual segregation across the Pacific Basin (Nagasawa, 1998; Goldman and Musick, in press), along with the age- and length-dependent latitudinal distribution are important factors in pur- suing responsible management and conservation of the salmon shark. Although latitudinal migrations and movements are documented in this species (lino, 1939; Kosugi and Tsuchisaki, 1950; Tanaka, 1980; Balga- derov, 1994; Nakano and Nagasawa, 1994; Weng et al., 2005), they are still poorly understood. Similarly, although trans-Pacific movements have been inferred from fisheries bycatch data (Goldman and Musick, in press), there has yet to be documentation of individuals moving across the North Pacific. However, the degree of the sexual segregation by itself (in regard to finding a mate) would seem to indicate that movements across the Pacific are likely. Another critical element for successful management of salmon sharks is stock structure, which is not well understood at this time; however a study of population 288 Fishery Bulletin 104(2) Table 4 Mean WNPl precaudal length (PCD at age for salmon sharks . Ranges (or individual lengths for some males) are Lamna ditropis) given for sharks in the eastern and western North Pacific (ENP and from the present study. Age Females Males WNP* mean PCL ENP mean PCL ENP range WNP* mean ind ENP ividual or range 0 62.5 65.9 62.2-71.1 62.5 63.0 1 97.5 92.1 73.0-105.0 97.5 91.0 2 110.0 115.3 112.0-119.0 110.0 — 3 122.5 123.3 118.0-128.0 122.5 118.0 4 135.0 134.0 128.0-140.0 135.0 145.0 5 141.3 147.3 144.0-153.0 139.7 — 6 147.7 158.1 149.7-164.0 144.3 155.4-157.0 7 154.0 164.1 145.9-183.0 149.0 150.5 8 160.3 175.7 164.0-185.4 153.7 163.0 9 166.7 178.4 164.0-192.0 158.3 — 10 173.0 185.4 168.0-198.0 163.0 — 11 177.4 187.0 173.0-200.0 164.2 164.0-176.0 12 181.8 186.7 180.0-193.0 165.3 187.0 13 186.2 190.8 159.9-213.0 166.5 176.0 14 190,6 192.9 183.0-208.0 167.6 177.0 15 195.0 196.7 175.0-207.1 168.8 — 16 199.4 200.1 193.0-208.3 169.9 176.0-182.0 17 203.8 208.3 203.0-213.4 171.1 187.0 18 — — — 172.2 — 19 — — — 173.4 — 20 — 205.1 200.0-210.2 174.5 — 21 — — — 175.7 — 22 — — — 176.8 — 23 — — — 178.0 — 24 — — — 179.1 — 25 — — — 180.3 — * Mean length-at length at 25 for -age for WNP males- calculated from information in Nagasawa (1998). L, was used as mean length at a ^e 17 for fe males and as mean genetics is currently underway. Current information from the western and central North Pacific implies that salmon sharks constitute a single stock, but there is no current information for the Japan Sea or the eastern North Pacific (Sano, 1962; Tanaka, 1980; Blagoderov, 1994; Nagasawa, 1998). New technologies such as ar- chival and pop-up satellite transmitters should provide documentation of movements and migrations (Weng at al., 2005) and key information as to whether regional or international conservation and management plans are needed. Stock structure may be an important factor in the dif- ferential growth rates between ENP and WNP salmon sharks. However, ecological differences between the ENP and WNP could also be responsible for the observed differences. Young salmon sharks appear to move from temperate waters of the U.S. west coast into the Alaska Gyre and the Gulf of Alaska as they approach adult- hood, which is when their growth rate begins to exceed that of their WNP counterparts. These waters are ex- tremely productive (Strub et al. 2001), and abundant food resources may be the key factor in the differences in growth rate, age-at-maturity, and weight-at-length observed between ENP and WNP salmon sharks. A high degree of variability exists in the periodicity of ring and growth band formation within and among taxonomic groups of elasmobranchs, and much of the variation observed in several lamniform sharks has not yet been explained (Branstetter, 1990; Branstetter and Musick. 1994; Wintner and Cliff, 1999). For example, Cailliet et al. (1983) stated that shortfin mako sharks (Isurus oxyrinchus) from the Pacific appear to form growth rings annually, whereas Pratt and Casey (1983) stated that Atlantic specimens appeared to produce two Goldman and Musick, Growth and maturity of Lamna ditropis 289 rings per year. However, validation of sin- gle annual ring (band pair) deposition has now been demonstrated from recaptures of oxytetracycline(OTC)-injected mako sharks (Natanson et al., in press), as well as with a new technique where bomb radiocarbon is used to validate ages in long-lived sharks (Campana et al., 2002; Ardizzone et al., in press). Lamnid sharks are endothermic (Carey et al., 1985; Lowe and Goldman, 2001), and prior to the Campana et al. (2002) and Ar- dizzone (in press) radiocarbon studies, it appeared that lamnid sharks might possess fairly high growth rates (growth coefficient, k) compared to other sharks, particularly those that grow to a large size (Musick, 1999). However, with mako sharks now ap- pearing to have a growth coefficient around 0.072/yr (sexes combined)(Cailliet et al. 1983), the validation of porbeagle shark (Lamna nasiis) age and growth (/^ = 0.06 and 0.08/yr for females and males, respective- ly) (Natanson et al. 2002), and estimates of white shark (Carcharodon carcharias) growth coefficients ranging from 0.058 to 0.071/yr (for sexes combined)(Cailliet et al., 1985; Wintner and Cliff, 1999; Malcolm et al. 2001), salmon sharks appear to be the species that possess the higher growth rates (Tanaka, 1980; present study). Salm- on sharks are not only endothermic, but also appear to be homeothermic (i.e., defend a specific body core temperature regardless of ambient temperature) (Goldman et al., 2004). This uncommon physiological trait for a shark, combined with a diet that in- cludes many lipid-rich species (e.g., salmo- nids), may influence the growth rate of this species, but endothermy does not appear to be correlated with faster growth rates in lamnid sharks (Goldman, 2002). The most recent demographic analysis has indicat- ed that salmon shark populations in the eastern and western North Pacific are stable at this time (Gold- man, 2002) and current research is beginning to as- sess standing stock and population numbers, as well as examine the stock structure of salmon sharks in the eastern and western North Pacific in order to further responsible management of the species. The differences in growth rates and strong sexual segregation across the North Pacific basin and an unknown rate of current bycatch complicate management of this species. Acknowledgments We are grateful to the many people have helped in field and lab operations. We thank J. Castro, J. Herschleb and K. Herschleb of the FV Pagan, K. Anderson, B. tMU - [A 220- y 200- / lan H Present study / ^ W=(8,2 X 10-=) X i"59 / y ibO r==0,99 / /' 140 ■ 120 ■ /y 100- y ^ an- ^/ ^ ' Tanaka 1 980 bU ■ ^^^^ " W = ( 1 04 X 1 0"*) X i.^ ^'9 40 : ^.r-^^ 20; — > — . — ^-i — : — ' — . — 1— -1 — . — 1 — 1 — . — . — . — 1 — > — . — > — 1— ^ — , — . — I— 1 — , — , — 1 — , — , — . — 1 — , — ,1 1 180 ■ B 160 ■ W-- Present study , (3 2x10e)xL33e3 / 140 ■ (-2=0.99 / 120 ■ / ^ 100 - m ■ /> 60 ■ ^ Tanaka 1980 40 ; ^^ >^^ l/V=(1,17x 10^) X L2652 20 ■ n • ..^.-^^ '-^^ Precaudal length (cm) Figure 7 Comparison of weight-length relationships of eastern and western North Pacific salmon sharks <~Lamna diiropis) (A) females, and (B) males. Western North Pacific data were obtained from Tanaka (1980) and Nagasawa (1998). Block, and L. Hulbert for assistance in the field. We thank R. Candopoulos, the crew, and many passengers aboard the FV Legend and B. Steffen (FV Sea Sound) for vertebral samples collected on sport fishing trips, and to the following Alaska Department of Fish and Game biologists for their logistic support, field help and overall assistance with our Alaska shark research: B. Bechtol, S. Meyer, C. Stock, M. Miller, D. Vincent-Lang, and D. Branshaw and D. Anderson of the RV Montague. We thank S. Anderson for his assistance with sample collection and unwavering endurance in the field, and J. Gelsleichter for the histological workup of the oviducal glands. We also thank J. DiCosimo (North Pacific Fish- ery Management Council) for her logistical support and assistance in the field. This research was supported in part by the National Geographic Society's Committee for Research and Exploration (grant no. 6714-00), Sigma-Xi, and the Lerner-Gray Fund for Marine Research. This research represents a portion of K.J. Goldman's Ph.D. dissertation conducted at the College of William and 290 Fishery Bulletin 104(2) Mary, School of Marine Science, Virginia Institute of Marine Science. This article is dedicated to the life and memory of Philip Goldman, compassionate human being and dedicated father. Literature cited Ardizzone, D., G. M. Cailliet, L. J. Natanson, A. H. Andrews, L. A. Kerr, and T. A. Brown. In press. Application of bomb radiocarbon chronologies to shortfin niako ilsurus oxyrinchus} age validation. In Age and growth of chondrichthyan fishes: New methods, techniques and analyses; special volume from symposium of the American Elasmobranch Society, July 200.5 (K. J. Goldman and J. K. Carlson, eds.l. Environ. Biol. Fish. 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. Blagoderov, A. I. 1994. Seasonal distribution and some notes on the biology of salmon shark iLainna ditropis) in the northwestern Pacific Ocean. J. Ichthyol. 34(2):115-121. [Originally published in, and translated from, Voprosy Ikhtiolo- gii.l Branstetter. S. 1990. Early life-history implications of selected carcharhi- noid and lamnoid sharks of the northwest Atlantic. In Elasmobranchs as living resources: advances in the biol- ogy, ecology, systematics and the status of the fisheries (H. L. Pratt Jr., S. H. Gruber, and T. Taniuchi, eds.), p. 17-28. NOAA Tech. Rep. NMFS 90. Branstetter, S., and J. A. Musick. 1994. Age and growth estimates for the sand tiger in the Northwestern Atlantic Ocean. Trans. Am. Fish. Soc. 123:242-254. Brodeur, R. D. 1988. Zoogeography and trophic ecology of the dominant epipelagie fishes in the northern Pacific. In The biol- ogy of the subarctic Pacific; proceedings of the Japan- United States of America seminar on the biology of micronekton of the subarctic Pacific (T. Nemoto and W. G. Percy, eds.), p. 1-27. Bull. Ocean Res. Inst.. Univ. Tokyo, no. 26 (part II). Cailliet, G. M., and K. J. Goldman. 2004. Age determination and validation in chondrich- thyan fishes. In The Biology of sharks and their rela- tives (J. Carrier, J. A. Musick and M. Heithaus, eds.) p. 399-447. CRC Press, Boca Raton. FL. Cailliet, G. M., L. J. Natanson, B. Welden, and D. Ebert. 1985. Preliminary studies on the age and growth of the white shark, Carcharodon carcharias, using vertebral bands. So. Cal. Acad. Sci., Mem. 9:49-60. 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, A/opi'as vulpinus, and shortfin mako, Isurus oxyrinchus, sharks from California waters. In Proceedings of the international workshop on age determination of oceanic pelagic fishes: tunas, billfishes, and sharks (E. D. Prince and L. M. Pulos, eds.) p. 179-188. NOAA Tech. Rep. NMFS 8. Camhi, M. 1999. Sharks on the line II: an analysis of Pacific state shark fisheries. National Audubon Soc. Publ., October 1999, 114 p. Campana, S. E. 2001. Accuracy, precision and quality control in age deter- mination, including a review of the use and abuse of age validation methods. J. Fish. Biol. 59:197-242. 1990. How reliable are growth back-calculations based on otoliths? Can. J. Fish. Aquat. Sci. 47:2219-2227. Campana, S. E., M. C. Annand, and J. I. McMillan. 1995. Graphical and statistical methods for determining the consistency of age determinations. Trans. Am. Fish. Soc. 124:131-138. Campana, S. E., L.J. Natanson, and S. Myklevoll. 2002. Bomb dating and age determination of large pelagic sharks. Can. J. Fish. Aquat. Sci. 59:450-455. Carey, F. G., Casey, J. G.. Pratt, H. L., Urquhart. D., and J. E. McCosker. 1985. Temperature, heat production, and heat exchange in lamnid sharks. So. Cal. Acad. Sci., Mem. 9:92-108. Carlander, K. D. 1969. Handbook of freshwater fishery biology, vol. 1, 752 p. Iowa Univ. Press, Ames, lA. Chang, W. Y. B. 1982. A statistical method for evaluating the reproduc- ibility of age determination. Can. J. Fish. Aquat. Sci. 39:1208-1210. Compagno, L. J. V. 1984. Sharks of the World. An annotated and illus- trated catalogue of shark species known to date. Part 1. Hexanchiformes to Lamniformes. FAO Fisheries Synopsis no. 125, vol. 4, part 1, 249 p. FAO, Rome. 2001. Sharks of the World. An annotated and illustrated catalogue of shark species known to date. Bullhead, mackerel and carpet sharks (Heterodontiformes, Lam- niformes and Orectolobiformes). FAO Species Cata- logue for Fishery Purposes, no. 1, vol. 2, 269 p. FAO, Rome. Conrath, C. L. 2004. Reproductive Biology. In Elasmobranch fisheries management techniques (J. A. Musick and R. Bonfil, eds.) p. 133-164. Asia Pacific Economic Cooperation, Singapore. [Electronic version available at http:// www.flmnh.ufl.edu/fish/organizations/ssg/EFMT2004. htm. I Conrath, C. L., and J. A. Musick. 2002. Reproductive biology of the dusky smooth-hound, Mustelus canis, in the northwest Atlantic Ocean. En- viron. Biol. Fish. 64:367-377. Conrath. C. L., J. Gelsleichter, and J. A. Musick. 2002. Age and growth of the smooth dogfish {Mustelus canis) in the northwest Atlantic Ocean. Fish. Bull. 100: 674-682. Evans, G. T, and J. M. Hoenig. 1998. Testing and viewing symmetry in contingency tables, with application to readers of fish ages. Bio- metrics 54:620-629. Francis, R. I. C. C. 1990. Back-calculation of fish length: a critical review. J. Fish Biol. 36:883-902. Gilmore, R. G. 1993. Reproductive biology of lamnoid sharks. Environ. Biol. Fish. 38:95-114. Gilmore, R. G., J. Q. Dodrill. and P. A. Linley. 1983. Reproduction and embryonic development of the sand tiger shark. Odontaspis taurus (Rafinesque). Fish. Bull. 81(2):201-225. Goldman and Musick: Growth and matunly of Lamna ditropis 291 Goldman, K. J. 2001. Sharks and shark bycatch in Alaska state and federal waters. In Ecosystems considerations for 2002 (P. Livingston, ed.), p. 60-79. North Pacific Fishery Management Council. 605 W. 4th Ave., Suite 306, Anchor- age, AK 99501-2252. 2004. Age and growth of elasmobranch fishes. /?; Elas- mobranch fisheries management techniques ij. A. Musick and R. Bonfil, eds.) p. 97-132. Asia Pacific Economic Cooperation, Singapore. [Electronic version available at http://www.flmnh.ufl.edu/fish/organizations/ssg/ EFMT2004.htm.] 2002. Aspects of age, growth, demographics and thermal biology of two lamniform shark species. Ph.D. diss, 220 p. Virginia Institute of Marine Science, School of Marine Science, College of William and Mary, Wil- liamsburg, VA. Goldman, K. J., and B. Human. 2005. lUCN red list assessment of Lo/?i;io <:/(7/op(.s. In Sharks, rays and chimaeras: the status of the chon- drichthyan fishes (S. L. Fowler, R. D. Cavenaugh, M. Camhi, G. Burgess, G. M. Cailliet, S. Fordham, C. A. Simpfendorfer, and J. A. Musick, eds.). lUCN/SSG Shark Specialist Group. lUCNSSC Shark Specialist Group, Gland. Switzerland, and Cambridge, UK. Goldman, K. J., and J. A. Musick. In press. Biology of the salmon shark, L(7;?;;!o rf/Yrop/s. In Sharks of the open ocean (E. K. Pikitch, and M. Camhi. eds. I. Blackwell Scientific, Aimes. lA. Goldman, K. J., S. D. Anderson, R. J. Latour, and J. A. Musick. 2004. Homeothermy in adult salmon sharks, Lamna ditropis. Environ. Biol. Fish. 71:403-411. Gorbatenko, K. M., and L. V. Cheblukova. 1990. Environment conditions and fish species com- position in the epipelagic zone of the Okhotsk Sea suring summer. J. Ichthyol. 30:89-100. [Originally published in, and translated from, Voprosy Ikhtiologii 30:21-30.1 Haddon, M. 2001. Modeling and quantitative methods in fisheries, 406 p. Chapman and Hall/CRC. Boca Raton, FL. Hamlett, W. C. 1999. Male reproductive system. In Sharks, skates and rays: the biology of elasmobranch fishes (W. C. Hamlett, ed.) p. 444-470. Johns Hopkins Press, Bal- timore, MD. Hamlett, W. C, and T. J. Koob. 1999. Female reproductive system. /;; Sharks, skates and rays: the biology of elasmobranch fishes ( W. C. Hamlett, ed.), p. 398-443. Johns Hopkins Press, Bal- timore, MD. Hinton. D. E. 1990. Histological techniques. In Methods for fish biology (C. B. Schreck and P. B. Moyle, eds.). p. 191-209. Am. Fish. Soc, Bethesda, MD. Hoenig, J. M., M. J. Morgan, and C. A. Brown. 1995. Analyzing differences between two age determi- nation methods by tests of symmetry. Can. J. Fish. Aquat. Sci. 52:364-368. lino, R. 1939. Migration of salmon sharks. Suisan Kenkyu-shi 34:171-173. [In Japanese.] JAMARC (Japan Marine Fishery Resources Center). 1980. The development of the shark resources in the North Pacific Ocean, fiscal years 1979 and 1980. Japan Marine Fishery Resources Center (JAMARC) Kaihalsu News (Development News) no. 11 p. 17-25, August 1980. [English translation no. 63, by Tamio Otsu, NMFS/SWFC, Honolulu, HI, June 1982.] Kimura, D. K. 1980. Likelihood methods for the von Bertalanffy growth curve. Fish. Bull. 77:765-773. Kosugi, H., and H. Tsuchisaki. 1950. On the longline fishing of salmon sharks in the Japan Sea off Hokkaido. Hokusuisi Geppo, 17(l):14-20. [In Japane.se. 1 Lowe, C. G., and K. J. Goldman. 2001. Thermal and bioenergetic telemetry of elasmo- branchs: bridging the gap. In The behavior and sensory biology of sharks and rays: state of the art and future direction. American Elasmobranch Soc. Symp., special volume in honor of D. R. Nelson (T. C. Tricas and S. H. Gruber, eds.), p. 251-266. Environ. Biol. Fish. 60. Malcolm, H., B. D. Bruce, and J. D. Stevens. 2001. A review of the biology and status of white sharks in Australian waters. Report to Environment Austra- lia, Marine Species Protection Program, 81 p. CSIRO Marine Research, Hobart, Tasmania, Australia. Mollet, H. F., G. Cliff, H. L. Pratt Jr., and J. D. Stevens. 2000. Reproductive biology of the female shortfin mako. Isurus oxyrinchus Rafinesque, 1810, with comments on the embryonic development of lamnoids. Fish. Bull. 98:299-318. Musick, J. A. 1999. Ecology and conservation of long-lived marine animals. In Life in the slow lane: ecology and conserva- tion of long-lived marine animals (J. A. Musick, ed.), p. 1-10. Am. Fish. Soc. Symposium 23, Bethesda, MD. Nagasawa, K. 1998. Predation by salmon sharks (Lamna ditropis) on Pacific salmon (Oncorhynchus spp.) in the North Pacific Ocean. Bulletin of the North Pacific Anad. Fish Com., No. 1:419-433. Nakano, H., and K. Nagasawa. 1996. Distribution of pelagic elasmobranchs caught by salmon research gillnets in the North Pacific. Fish. Sci. 62(5):860-865. Nakano, H., and K. Nagasawa. 1994. Distribution of the salmon shark (Lamna ditropis) in the North Pacific Ocean and Bering Sea. Salmon Report Series, no. 37:226-237. Nat. Res. Inst. Far Seas Fisheries, Shimizu Japan. Natanson, L. J., J. J. Mello, and S. E. Campana. 2002. Validated age and growth of the porbeagle shark {Lamna nasus) in the western North Atlantic Ocean. Fish. Bull. 100:266-278. Natanson, L. J., N. E. Kohler, D. Ardizzone, G. M. Cailliet, S. Wintner, and H. F. Mollet. In press. Validated age and growth estimates for the sShortfin mako, Isurus oxyrinchus, in the North Atlantic Ocean. In Age and growth of chondrichthyan fishes: new methods, techniques and analyses (K. J. Gold- man and J. K. Carlson, eds.). Special volume from symposium of the Am. Elasmobranch Soc. July 2005. Environ. Biol. Fish. Neave, F., and M. G. Hanavan. 1960. Seasonal distribution of some epipelagic fishes in the Gulf of Alaska region. J. Fish. Res. Board Can. 17(2):221-233. Neer, J. A., and G. M. Cailliet. 2001. Aspects of the life history of the Pacific electric ray. Torpedo californica (Ayres). Copeia 2001:842-847. 292 Fishery Bulletin 104(2) Pratt, H. L. Jr. 1988. Elasmobranch gonad structure: a description and survey. Copeia 1988:719-729. 1993. The storage of spermatozoa in the oviducal glands of western North Atlantic sharks. Environ. Biol. Fish. 38:139-149. Pratt. H. L. Jr., and J. G. Casey. 1983. Age and growth of the shortfin mako, Isuriis oxy- ri?ichus, using four methods. Can. J. Fish. Aquat. Sci. 40:1944-1957. Pratt, H. L. Jr., and S. Tanaka. 1994. Sperm storage in male elasmobranchs: a descrip- tion and survey. J. Morphol. 219:297-308. Quinn. T. J., and R. B. Deriso. 1999. Quantitative fish dynamics, 542 p. Oxford Univ. Press, New York, NY. Ricker, W. E. 1992. Back-calculation offish lengths based on propor- tionality between scale and length increments. Can. J. Fish. Aquat. Sci. 49:1018-1026. Sano, O. 1962. The investigation of salmon sharks as a preda- tor on salmon in the North Pacific, 1960. Bull. Hok- kaido Reg. Fisheries Res. Lab., Fisheries Agency, No. 22:68-82. [In Japanese.] Strasburg, D. W. 1958. Distribution, abundance, and habits of pelagic sharks in the central Pacific Ocean. Fish. Bull. 58: 335-361. Strub, P. T., H. P. Batchelder, and T. J. Weingartner. 2001. U.S. GLOBEC Northeast Pacific program: Over- view. Oceanog. 15(2):30-35. Tanaka, S. 1980. Biological investigation of Lamna ditropis in the north-western waters of the North Pacific. In Report of investigation on sharks as a new marine resource 11979), p. 59-84. Japan Marine Fishery Resource Research Center, Tokyo, Japan. [English abstract, translation by Nakaya.] Weng, K. C, P.C. Castilho, J. M. Morrissette. A. Landiera- Fernandez. D. B. Holts. R. J. Schallert. K. J. Goldman, and B. A. Block. 2005. Satellite tagging and cardiac physiology reveal niche expansion in salmon sharks. Science 310:104-106. Wintner, S.B., and G. Cliff 1999. Age and growth determination of the white shark, Carcharodon carcharias, from the east coast of South Africa. Fish. Bull. 97:153-169. Wourms, J. P. 1977. Reproduction and development in chondrichthyan fishes. Am. Zool. 17:379-410. 293 Potential causes of mortality for horseshoe crabs (Limu/us polyphemus) during the biomedical bleeding process Lenka Hurton Department of Fisheries and Wildlife Sciences and Horsesfioe Crab Research Center Virginia Polytechnic Institute and State University Blacksburg, Virginia 24061-0321 Jim Berkson National Marine Fisheries Service-Virginia Tech RTR Unit Fisheries and Wildlife Sciences and Horseshoe Crab Research Center Virginia Polytechnic Institute and State University Blacksburg, Virginia 24061-0321 E-mail address (for J Berkson, contact author) Jim Berkson g noaa gov Biomedical companies catch and bleed horseshoe crabs for the pro- duction of Limiilus amebocyte lysate (LAL), a product used for protecting public health (Berkson and Shus- ter, 1999). LAL is a clotting agent, derived solely from horseshoe crab blood cells, which is used to detect the presence of pathogenic gram- negative bacteria in injectable drugs and implantable medical and dental devices (Mikkelsen, 1988; Novitsky, 1991). In addition, LAL is used in many diagnostic tests for such ill- nesses as gram-negative bacterial meningitis and typhoid fever (Ding and Ho, 2001). Because the LAL test allows one to detect femtogram levels of endotoxin (Ding and Ho, 2001), it is the most effective test for detect- ing endotoxin contamination, and its increasing use in medical and phar- maceutical laboratories makes it a highly valued product. The biomedical industry harvest- ed approximately 260,000 horseshoe crabs in 1997 (HCTCi) for the pro- duction of LAL. By 2000, the number of horseshoe crabs bled had increased less than 1% (ASMFC-). However, approximately 25% of the horseshoe crabs landed for biomedical purposes were rejected for use, and about 45% of those rejected were rejected be- cause they were injured (ASMFC-). Some mortality is likely related to harvest injuries, but this rate of mor- tality is unknown. At current har- vest levels, mortalities from biomedi- cal collection and bleeding methods may not be negligible. Mortalities in- flicted by the biomedical industry on the horseshoe crab population come in addition to mortalities caused by the commercial fishery, where horse- shoe crabs are harvested for use as bait for eel [Anguilla rostrata) and whelk (Busycon spp.) fisheries (Walls et al., 2002). Furthermore, numer- ous migratory shorebird species feed on horseshoe crab eggs (Clark, 1996) and adults (Botton and Love- land, 1989, 1993) to fuel their migra- tion from South American winter- ing grounds to their Arctic breeding grounds (Clark, 1996). Concern has arisen over horseshoe crabs, in part because of local population declines, increased catch and effort, and the need for a superabundance of horse- shoe crab eggs for shorebirds, all of which necessitate a conservative, risk-averse management approach ( ASMFC- ). It is important to not only investigate the effects of the com- mercial fishery harvest and natural predation on horseshoe crab stocks but to comprehensively assess the impact of the biomedical industry take on horseshoe crab population viability. Reducing mortalities from the biomedical industry may aid in conserving the horseshoe crab pop- ulation, as well as in reducing the magnitude of conflict between this industry, the commercial fishery, and environmentalists. To achieve this goal, further investigation into the biomedical bleeding process and its effects is required. Throughout the typical biomedi- cal bleeding process, horseshoe crabs are subjected to a variety of potential stressors (i.e., air exposure, increased temperature, handling, blood loss, trauma, etc.). Each LAL producer has its own bleeding process involving different methods of capture, distance and method of travel to the bleeding facility, different holding times and conditions, different bleeding meth- ods, and methods of returning bled crabs that are most appropriate to that company's location and facility set-up (Walls and Berkson, 2003). An example of one version of the bio- medical bleeding process begins with the collection of the horseshoe crabs by trawling, dredge, or hand-harvest methods (HCTC). Animals may be held on the deck of a boat or in con- tainers for several hours during col- lection, transported to the bleeding facility in trucks (that may or may not be air-conditioned), held in the 1 HCTC (Horseshoe Crab Technical Committee). 1998. Status of the horseshoe crab tLimulus polyphemus) population off the Atlantic Coast, 9 p -I- figures and tables. Horseshoe Crab Technical Committee, Atlantic States Marine Fisheries Commission, 1444 Eye Street, NW, Sixth Floor, Washington, DC 20005. - ASMFC (Atlantic States Marine Fisher- ies Commission). 2002. Review of the fishery management plan for horseshoe crab iLinutlus polyphemus). Horseshoe Crab Plan Review Team, Atlantic States Marine Fisheries Commission, 1444 Eye Street, NW, Sixth Floor. Washington, DC 20005. Manuscript submitted 17 May 2004 to the Scientific Editor's Office. Manuscript approved for publication 21 July 2005 by the Scientific Editor. Fish. Bull. 104:293-298 1 2004). 294 Fishery Bulletin 104(2) cold room of a laboratory for several hours at an air temperature of 16-18"C, bled for a period of time, and then held in the coldroom or in a truck until transport back to the dock (Grogan-^; Walls'*). Horseshoe crabs that are returned to the ocean are transferred onto a boat and returned to their approximate point of capture within 72 hours of their collection (HCTC'). The numer- ous stressors to which the horseshoe crabs are exposed throughout the biomedical bleeding process likely have an effect on their mortality rates, both for horseshoe crabs used for bleeding and for those crabs collected, but rejected for use. Additional sources of stress also may arise from the bleeding methods used by some biomedical companies. The bleeding protocols at some facilities involve gravi- tationally withdrawing blood from the heart (punctured with a large gauge needle) until blood flow slows to an intermittent drip (Grogan-^; Walls'*). One study reported a range of extracted blood to be from 8.4 mL to 218.7 mL in male horseshoe crabs (Walls, 2001) and up to 267.8 mL in females (Walls"). Unfortunately, biomedical companies do not know how much blood horseshoe crabs of a given size possess and how much can be safely ex- tracted. Thus, extracting the reported upper ranges of blood volume from males and females may cause stress or mortality. Several studies have estimated mortality associated with the biomedical bleeding process to be between 8% and 20% (Rudloe, 1983; Thompson, 1998; Kurz and James-Pirri, 2002; Walls and Berkson, 2003). These studies either simulated the biomedical bleeding process (Rudloe, 1983; Kurz and James-Pirri, 2002) or moni- tored mortality of a given number of horseshoe crabs bled by a biomedical company (Thompson. 1998; Walls and Berkson, 2003). Although the results provide infor- mation on mortality rates associated with a biomedical companies' procedures, they provide no guidance on how biomedical companies may reduce mortality rates. Rather than quantifying mortality rates associated with the biomedical bleeding process as a whole, we isolated and tested certain elements of the bleeding process to gain information on the potential causes of mortality for horseshoe crabs used in the biomedical in- dustry. The objectives of this study were 1) to quantify mortality associated with blood extraction at set levels and 2 ) to quantify mortality associated with stress from simulated transport and holding procedures, combined with the stress of blood extraction at the same levels as in the first objective. ■'' Grogan, W. 2003. Personal commun. Department of Fish- eries and Wildlife Sciences, Virginia Polytechnic Institute and State Univ., Blacksburg, VA 24061. " ■* Walls, B. 2001. Personal commun. Center for Environ- mental Studies, Virginia Commonwealth Univ., 1000 West Cary Street, Box 843050, Richmond, VA 23284. •'' Walls, B. 2001. Unpubl. data. Center for Environmental Studies, Virginia Commonwealth Univ., 1000 West Cary Street, Box 843050, Richmond, VA 23284. Methods Adult horseshoe crabs were obtained from Cambrex BioScience Walkersville, Inc., a commercial biomedical company on 9 July and 28 August 2003. Horseshoe crabs were captured by using a standard trawling procedure off the coast of Ocean City, Maryland (Hata and Berk- son, 2003). After capture, horseshoe crabs were trans- ported in an air-conditioned van to the Horseshoe Crab Research Center (HCRC) at Virginia Polytechnic Insti- tute and State University in Blacksburg, Virginia. The animals were maintained in a recirculating aquaculture system in appropriate environmental conditions (Brown and Clapper, 1981) and monitored daily, but not fed. The animals were acclimated to the aquaculture conditions for one week, during which time they were tagged, sexed, and measured. Tagging involved drilling two 3/32" holes into the prosoma at its thinnest point and attaching a laminated oval fish tag (Floy Tag, Seattle, WA) with two 3/32" wide cable ties. Measurement consisted of record- ing the inter-ocular width (10), the distance between the inside margins of the horseshoe crab's compound eyes, and the prosomal width (P), which is the distance across the horseshoe crab's carapace. Horseshoe crabs used for both bleeding mortality studies had a mean 10 width of 14.0 cm and ranged from 10.5 to 19.0 cm; as well as, an average P width of 23.5 cm that ranged from 17.5 to 35.0 cm. Only animals that were uninjured (i.e., no cracked carapaces or missing legs, etc.) and free of epibionts were used in the two mortality experiments. Experiment 1 Horseshoe crabs from the July collection were used in experiment 1. From this collection, 100 males and 100 females were selected to test only the effects of various levels of blood extraction on mortality. This sample of horseshoe crabs was termed the "lower-stressed" group because they were not exposed to external stressors associated with simulated holding and transport. The selected horseshoe crabs ranged in size to provide a representative sample of animals bled by some bio- medical companies (Grogan^; Walls^). Horseshoe crabs were arbitrarily assigned to one of five bleeding treat- ments: 1) unbled crabs (control), 2) crabs bled lO'/r of their predicted total blood volume, 3) crabs bled 20% of total blood volume, 4) crabs bled 30% of total blood volume, and 5) crabs bled 40% of their predicted total blood volume. Each bleeding subgroup comprised equal numbers of males and females. Predicted blood volume was calculated by using the relationship between blood volume and 10 width (Hurton et al., 2005): H = 25.7e 0.1928 I/O) where H = hemolymph volume in mL; and 10 = interocular width in cm. The bleeding process of experiment 1 involved remov- ing horseshoe crabs from their holding tank, positioning NOTE Beikson and Hurton Potential causes of mortality for Linwius polyphenws 295 them in a bleeding rack, and disinfecting the surface of their arthrodial membrane with a 70% ethanol-soaked cotton swab. An 18-gauge needle was inserted through the arthrodial membrane into the cardiac sinus to ex- tract the predetermined amount of blood. The animals were then placed in a holding container without water for 15-20 hours. Time out of water approximated the duration that horseshoe crabs may be out of water after bleeding at some biomedical bleeding facilities (Grogan'; Walls^). Bled horseshoe crabs were also not immediately put back in water to prevent them from absorbing wa- ter and regaining their blood volume. Temperatures to which horseshoe crabs were exposed were recorded us- ing a temperature logger. Air temperature during this time was stable at 21°C. Subsequently, the animals were returned to their holding tanks and mortality was monitored for two weeks. Experiment 2 Horseshoe crabs from the August collection were used for experiment 2. Because of difficulty in obtaining the desired 200 study animals, slightly fewer ( 195) horseshoe crabs were selected for this experiment. This sample of 110 males and 85 females was used to test the effects of various levels of blood extraction on mortality in the pres- ence of external stressors. These animals represented the "higher-stressed" group because they were exposed to additional stressors to which the first group was not. The external stressors (i.e., air exposure, increased tempera- ture, etc.) were those derived from simulated transport and holding procedures during a simulated biomedical bleeding process. Because there is great variability in transport and holding procedures and conditions, we simulated a protocol that would demonstrate horseshoe crab response under relatively poor conditions. During the week of acclimation to the aquaculture system, the horseshoe crabs were held in the same con- ditions and handled in the same manner as group 1. Horseshoe crabs were assigned to one of five bleeding treatments used in experiment 1. The bleeding process of experiment 2 involved removing horseshoe crabs from their holding tank and placing them in holding contain- ers located outside where they were exposed to air, sun, and increased temperatures. During this 6-hour period, the air temperature rose from 21°C to 29°C to simulate time on the deck of a trawler. Next, the animals were moved into a small moving truck, which was not air- conditioned. This phase simulated transportation to the bleeding facility and holding time until placement into the coldroom of the laboratory. During this 4-hour pe- riod, the outdoor temperature increased to 31°C and the temperature inside the closed truck peaked at 36°C. The animals were then transferred to the HCRC's tank room where they remained for 16 hours at 21°C. This phase mimicked the holding time in a coldroom prior to bleed- ing. The animals were bled according to their assigned treatment and in the same manner as in experiment 1. During this 8-hour period, the room temperature was 22°C. Once the bleeding was completed, the horseshoe crabs were moved back into the truck for 13 hours to simulate holding time and transport to the boat. The overnight temperature inside the truck dropped from 24°C to 20"C. In the morning, the horseshoe crabs were returned to the water in their recirculating aquaculture tanks and mortality was monitored for two weeks. Statistical analyses Fisher's exact test was used to evaluate 1) differences in mortality of unbled horseshoe crabs between the higher- stressed and lower-stressed groups, and 2) differences in mortality of bled crabs between the lower-stressed and higher-stressed groups. Logistic regression was used to examine the association between mortality and bleeding in the higher-stressed group. Data were also assessed to see if an increase in the volume bled was correlated with an increase in mortality by pooling the frequency of mortalities between males and females in each bleeding treatment and testing these values against bleeding percentage by employing a regression with a fitted quadratic curve. Results Experiment 1 No horseshoe crab mortality occurred in any of the five treatments in the lower-stressed group, indicating that post-bleeding mortality in Limulus did not arise solely from the effects of blood loss. Hence, our initial hypoth- esis that substantial blood loss was the principle cause of mortality in horseshoe crabs was rejected. Experiment 2 There were a total of 14 mortalities distributed through- out the bleeding treatments among horseshoe crabs under higher stress conditions (Table 1). All mortalities occurred within the first seven days of the study. Bled horseshoe crabs had an overall mortality rate of 8.3% compared to the 2.6% mortality rate of unbled crabs, suggesting a relationship between mortality and bleed- ing under higher-stressed conditions (;?=195; P=0.0088) (Table 1). The bleeding variable was significant in the logistic regression model (P=0.G160); yet, sex was not a significant variable (n=156; P=0.6100); seven deaths in total occurred in each sex category (Table 1). Mortality rates reached 13.6% for males bled 30% of blood volume and female mortality was as high as 29.4% in the sub- group bled 40% of blood volume (Table 1). With male and female mortalities pooled, the frequency of mortality increased as the percentage of blood taken from the crabs was increased (;! = 5; P=0.006; r-=0.994) (Table 1). Comparison of the two experiments With no mortalities in the unbled treatment of the lower- stressed group and one mortality in the unbled treat- 296 Fishery Bulletin 104(2) Table 1 Comparison of mortalities observed in the h igher- stressed group between unbled horse shoe crab s and horseshoe crabs bled at different level s of total blood volume ( /!=195). <7( bled Mortality (no.l Mortality (%) Male Female Total Male Female Total 0% 1 0 1 4.5 0 2.6 10% 0 1 1 0 5.9 2.6 20% 2 0 2 9.1 0 5.1 30%^ 3 1 4 13.6 5.9 10.3 40% 1 5 6 4.5 29.4 15.4 Total 7 7 14 6.4 8.2 7.2 ment of the higher-stressed group, mortality rates were not significantly different between unbled crabs in the two bleeding experiments (?!=79; P=0.4937). Of the bled crabs, there was a mortality rate of 0% in the lower- stressed group and a rate of 8.3% in the higher-stressed group, indicating a significant difference in mortality rates between the two groups (n = 316; P<0.0001). Discussion According to our study of 395 horseshoe crabs, no impact on mortality was observed with blood extraction up to 407f of the horseshoe crab's blood volume under condi- tions of lower stressors. Any mortality associated with this maximum level of bleeding (40% of total blood volume) was so low that it was not captured by our sample size of 40 individuals per bleeding treatment. This result also may indicate that despite hypovole- mia, or low blood volume, and altered blood chemistry (Hurton, 2003), horseshoe crabs may be relatively tol- erant of the removal of a large amount of blood in the absence of other stressors. Horseshoe crab mortality, however, was significant in the presence of higher levels of stress, including external stressors (i.e.. lengthy air exposure, elevated temperatures, etc.) as applied in our study. Novitsky's (1991) aquarium studies reported that up to 30% of blood volume can be safely extracted, indicating that the animals in that study may have been similar to our lower-stress group; however, details regarding har- vest method, transport, and handling procedures were not specified (Novitsky, 1991). It is likely that, under different conditions, horseshoe crab response to blood extraction may be altered. Findings from our study indicated that a combination of external effects and sublethal hemorrhaging may result in significantly in- creased mortalities, possibly from a synergistic inter- action between the two types of stressors. Studies in other species exposed to various stressors have also indicated that mortality may be affected by a syner- gistic combination of effects from multiple stressors (Schisler et al., 2000; Schulz and Dabrowski, 2001; Hatch and Blaustein, 2003). Such results emphasize the importance of considering the cumulative impact of multiple stressors. Although the data in our study indicated a signifi- cantly increased mortality rate in the higher-stressed group as opposed to the lower-stressed group, it is im- portant to note the variability in mortality through- out the treatments. It was observed that the male and female horseshoe crabs had different and vari- able responses in the different bleeding treatments. For example, males had two and three mortalities in the 20% and 30% bleeding treatments, respectively; yet, only one mortality in the 40% treatment (Table 1). Females appeared to have a spike in mortality in the 40% treatment (Table 1). When viewing the data this way, the variability in mortality could be in part due to the health and precollection stress level of a horseshoe crab, which can differ from its counterparts. Different conditions may lead to different responses and not every individual will react exactly alike. More than likely, much of this variability is an artifact of a small sample size. Once the male and female data were pooled, total mortality appeared to proceed in an increasing trend as bleeding amount increased (Table 1). If sample size were increased enough, perhaps a strong trend would emerge in mortality between males and females, or perhaps not. Further investigation with larger sample sizes and a different experimental design could shed light on a possible trend in male and female mortality rates and whether one sex may be more susceptible to mortality than the other. The combination of multiple stressors could also affect not only mortality, but possibly the amount of blood able to be extracted by the biomedical industry. It is possible that horseshoe crabs exposed to higher levels of external stress, such as greater lengths of time to air and higher temperatures, would lose moisture from their exposed gills and possibly become dehydrated and thus have a decreased blood volume such that a smaller amount of blood is available for extraction. It is unknown how this would affect the harvest of horseshoe crab blood for the production of LAL. Dehydration rates in horse- shoe crabs are unknown, but have been examined in a freshwater crayfish {Austropotamobius pallipes) which is a facultative air breather that can survive about three days out of water (Taylor et al., 1987). Taylor et al. reported that when exposed to air (70-80% relative humidity) for 27 hours, crayfish became dehydrated and had a 25% decreased blood volume. When exposed to water-saturated air (100% relative humidity), crayfish did not have a decrease in blood volume. Similarly, horseshoe crabs exposed to water-saturated air may be less likely to become dehydrated, thereby possibly decreasing the effects of one type of stress. The results presented in the present study provide in- sight into the possible combined effects of blood extrac- tion and external stressors associated with biomedical NOTE Berkson and Hurton Potential causes of mortality for Limulus polyphemus 297 transport and holding methods on horseshoe crab mor- tality. Of the various protocols followed in the bleeding process, we simulated an example of transport, holding, and bleeding methods that provided poor conditions for the horseshoe crabs and that resulted in significant mortality rates. In the typical biomedical bleeding pro- cess, horseshoe crabs undergo time on a boat deck or in collection bins, transport, storage in a coldroom, bleed- ing, holding time, and transport back to the ocean. At each of these stages of processing, the animals are ex- posed to a number of stressors of varying magnitudes, including exposure to air for extended time periods, elevated temperatures, dehydration, hypovolemia, and likely other unknown stressors. However, these factors can be controlled to a certain extent by providing condi- tions for the horseshoe crabs that could alleviate much of the stress. For example, holding horseshoe crabs in an air-conditioned environment (i.e., during time in the truck and in the cold room), covering the animals with wet burlap while outside to keep them moist and shaded from the sun, and increasing the relative humidity of the air in the holding room could help to decrease the physiological stress the animals experience throughout the bleeding process. Because blood volume estimates have recently been determined (Hurton et al., 2005), the blood volume extracted by biomedical companies po- tentially could be optimized according to average stress conditions and attempts could be made to decrease stress as much as feasibly possible, including keeping horseshoe crabs hydrated and bleeding a prescribed amount from individuals, as opposed to withdrawing a wide range of blood volumes. Another important con- sideration is stress arising from harvest methods. Our study did not test the effects of this factor, but it may have an effect on the condition of horseshoe crabs col- lected for bleeding. Biomedical companies should in- clude the potential effects of harvest methods when determining what may be typical stressful conditions for the horseshoe crabs, or they could consider using a harvesting method that would be the least stressful on the animals. All LAL producers can potentially imple- ment some form of these recommendations, thereby al- tering bleeding process protocols to decrease the stress and mortality levels that these horseshoe crabs experi- ence. Implementation of these recommendations would decrease the impact of the biomedical industry on the potentially declining horseshoe crab population and aid in conservation of this species. Acknowledgments This work is a result of research sponsored in part by the NOAA Office of Sea Grant, U.S. Department of Com- merce, under Grant no. NA96RG0025 to the Virginia Graduate Marine Science Consortium and Virginia Sea Grant College Program. Cambrex Bio Science Walkers- ville. Inc. generously supplied us with horseshoe crabs for the study. We wish to thank Stephen Smith of the Virginia-Maryland College of Veterinary Medicine at Virginia Tech for his help in the planning of this study. We also thank Eric Hallerman of the Department of Fisheries and Wildlife Sciences of Virginia Tech for his helpful comments throughout this project. In addition, Xin Zhong and Li Wang of the Department of Statistics at Virginia Polytechnic Institute and State University were the statistical consultants for this project. We offer a special thanks to numerous colleagues for their assis- tance in transporting, tagging, and bleeding horseshoe crabs for this study. We greatly appreciate the time and effort of all involved. Literature cited Berkson, J. B., and C. N. ShusterJr. 1999. The horseshoe crab: the battle for a true multiple- use resource. Fisheries 24(11):6-10. Botton, M. L., and R. E. Loveland. 1989. Reproductive risk: high mortality associated with spawning in horseshoe crabs {Limulus polyphemus) in Delaware Bay, USA. Mar. Biol. (Berl.) 101:143-151. 1993. Predation by herring gulls and great black-backed gulls on horseshoe crabs. Wilson Bull. 105:.518-.5'21. Brown, G. G., and D. Clapper. 1981. Procedures for maintaining adults, collecting gam- etes, and culturing embryos and juveniles of the horseshoe crah, Limulus polyphemus. In Laboratory animal man- agement: marine invertebrates, p. 268-290. National Academy Press, Washington, D.C. Clark, K. 1996. Horseshoe crabs and the shoi'ebird connection. In Proceedings of the horseshoe crab forum: status of the resource. (J. Farrell and C. Martin, eds.), p. 23-25. Univ. Delaware Sea Grant College Program, Lewes. DE. Ding, J. L., and B. Ho. 2001. A new era in pyrogen testing. Trends in Biotech- nology 19:277-281. Hata, D., and J. M. Berkson. 2003. Abundance of horseshoe crabs, Limulus polyphemus, in the Delaware Bay area. Fish. Bull. 101:933-938. Hatch, A. C, and A. R. Blaustein. 2003. Combined effects of UV-B radiation and nitrate fertilizer on larval amphibians. Ecol. Appl. 13:1083- 1093. Hurton, L. 2003. Reducing post-bleeding mortality of horseshoe crabs {Limulus polyphemus) used in the biomedical indus- try. M.S. thesis, 89 p. Virginia Polytechnic Inst, and State Univ., Blacksburg, VA. Hurton, L., J. Berkson, and S. Smith. 2005. Estimation of total hemolympb volume in the horse- shoe crab Limulus polyphemus. Mar. Freshw. Behav. Physiol. 38(2):139-147. Kurz, W., and M. J. James-Pirri. 2002. The impact of biomedical bleeding on horseshoe crab, Limulus polyphemus, movement patterns on Cape Cod, Massachusetts. Mar. Freshw. Behav. Physiol. 35:261-268. Mikkelsen, T 1988. The secret in the blue blood, 125 p. Science Press, Beijing, China. Novitsky, T. J. 1991. Discovery to commercialization: the blood of the horseshoe crab. Oceanus 27(1):13-18. 298 Fishery Bulletin 104(2) Rudloe, A. 1983. The effect of heavy bleeding on mortality of the horseshoe crab, Limulus polyphemus, in the natural environment. J. Invert. Path. 42:167-176. Schisler, G. J., E. P. Bergersen, and P. G. Walker. 2000. Effects of multiple stressors on morbidity and mor- tality of fingerling rainbow trout infected with Myxobolus cerebralis. Trans. Am. Fish. Soc. 129:859-865. Schulz, R., and J. M. Dabrowski. 2001. Combined effects of predatory fish and sublethal pesticide contamination on the behavior and mor- tality of mayfly nymphs. Environ. Toxicol. Chem. 20:2537-2543. Taylor. E. W., R. Tyler-Jones, and M. G. Wheatly. 1987. The effects of aerial exposure on the distribu- tion of body water and ions in the freshwater crayfish Austropotamobius pallipes (Lereboullet). J. Exp. Biol. 128:307-322. Thompson, M. 1998. Assessments of the population biology and critical habitat for the horseshoe crab, Limulus polyphemus, in the South Atlantic Bight. M.S. thesis, 136 p. Univ. Charleston, Charleston, SC. Walls, B. A. 2001. A horseshoe crab iLimulus polyphemus) demo- graphic study. M.S. thesis, 63 p. Virginia Polytechnic Inst, and State Univ., Blacksburg, VA. Walls, B. A., and J. M. Berkson. 2003. Effects of blood extraction on horseshoe crabs {Limulus polyphemus). Fish. Bull. 101:457-459. Walls, B. A.. J. M. Berkson, and S. A. Smith. 2002. The horseshoe crab, Limulus polyphemus: 200 million years of existence, 100 years of study. Rev. Fish. Sci. 10:39-73. 299 A study of tagging methods for the sea cucumber Cucumaria frondosa in the waters off Maine Sheril Kirshenbaum 216 Libby Hall, School of Marine Sciences University of Maine Orono. Maine 04469 E mail address sheril_kirshenbaum(®umit maine edu Scott Feindel Maine Department of Marine Resources West Bootfibay Harbor, Maine 04575 Yong Chen 216 Libby Hall, School of Marine Sciences University of Maine Orono, Maine 04469 The sea cucumber fishery m waters off Maine is developing and has recently experienced great increases in land- ings, corresponding to expanding export markets. Between 1994 and 1996, reported landings ranged from one to three million pounds (Fig. 1). In 1999, reported landings were over eight million pounds and rose to over nine million in 2000 (Feindel'). Like other developing fisheries, we have little information about the biology and ecology of the sea cucumber off Maine, limited data on the fishery, and little knowledge about the key life history processes that characterize its population dynamics. Therefore, we have a limited understanding of the current status of the resource and the impacts the fishery may have on the stock. Most research done so far has been done on the red sea cucumber [Para- stichopus californicus) in the north- western United States and British Columbia (e.g., Bradbury et al., 1996; Zhou and Shirley, 1996; Phillips and Boutillier, 1998; Perry et al., 1999; Cripps and Campbell, 2000). Lim- ited research has been done along the eastern coast of North America, although some studies have added to our knowledge of Cucumaria fron- dosa. A study of the ecology and be- havior of C. frondosa was conducted in Maine by Jordan (1972). More re- cent studies have focused on fertil- ization success and feeding behavior as it relates to aquaculture (Hamel and Mercier, 1997; Medeiros-Bergen and Miles, 1997; Singh et al., 1998). These studies have improved our knowledge of C. frondosa but still do not provide enough information on growth, one of the life history processes most important for under- standing the population dynamics of sea cucumbers in waters off Maine. Because of the lack of hard tissue in sea cucumbers to lay down growth increments, tagging is the method undertaken to determine age. Tag- ging studies of sea cucumbers are difficult because external tags are frequently lost and internal tags can be shed through the body wall. Sea cucumbers have been tagged in situ with limited success (Shelley, 1981; Conand-) by using a small T-bar tag that is inserted through the body wall with a tagging gun (Harriott, 1980). Previous studies of manual tags in P. californicus in Washington indicated that the presence of these tags had no significant effect on the behavior of the sea cucumbers and did not af- fect their ability to react to shifts in the salinity of the experimental medium (FankbonerM. Typically, ne- crosis of the tissue surrounding the tag occurs and tags fall out within a few months (Morgan, 2000). Morgan suggested that this method may be adequate for tagging broodstock in captivity over short periods of time and stated that tag loss may be min- imized by ensuring that the T-bar is pushed right through the dermis and attempting to make a puncture wound that is as small and clean as possible. Schroeter et al. (2001) used tags to conduct a growth study of P. parvimensis in California by follow- ing individuals through repeated sur- veys at two sites. Only 17.6 % of 1224 tagged animals were recaptured over one year. These studies indicate that manual tags may not be an effective means for studying growth because of the high frequency of shedding tags by sea cucumbers. Previous researchers have also used fluorescent dyes to stain the calcareous plates surrounding the buccal cavity of sea cucumbers but with varied results. Successful staining has depended on the tim- ing of the injection of the dye and the deposition of calcium for growth of the mouthparts (Conand-). Other work has suggested that fluorescent marking by dyeing the ossicle of the juveniles with tetracycline may be useful (Tanaka^). However, injec- ' Feindel, S. 2002. Status of the Maine sea cucumber (Cucumaria frondosa) fishery. Report to the standing legis- lative committee on marine resources, 35 p. Department Marine Resources, West Boothbay Harbor, ME 04575. - Conand, C. 1989. Aspidochirote holo- thurians of New Caledonia lagoon: biol- ogy, ecology and exploitation. Studies and thesis, 393 p. ORSTOM, Paris. ■' Fankboner, P. V. 2002. Seasonal vis- ceral atrophy and response to salinity by Parastichopus californicus (Stimp- son): osmoregulation? SPC Beche-de-mer Information Bulletin no. 17, 5 p. Uni- versite de La Reunion, Laboratoire de biologie marine, 97715 Saint-Denis Cedex, La Reunion, France. ■* Tanaka, M. 2000. Diminution of sea cucumber Stichopus Japonicus juveniles released on artificial reefs. Bulletin Ishikawa Prefecture Fish Research Centre 2:19-29. Manuscript submitted 13 October 2004 to the Scientific Editor's Office. Manuscript approved for publication 26 July 2005 by the Scientific Editor Fish. Bull. 104:299-302 (2006). 300 Fishery Bulletin 104(2) tions are time-intensive and marks often become difficult to discern after several months. Also, sea cucumbers could not be located soon after release in some follow- up investigations. Yanagisawa (1995) successfully used a hot wire to brand the sea cucumber Stichopiis japonicus. Although most of the marked juveniles could be distinguished three months later, it seemed to be difficult to discern their marks after four months. The results of previous worldwide sea cucumber tagging studies vary in success by species, investigator, and location. It is also possible that the technique varied significantly for studies involving the same styles of tagging. Therefore, the conclusions of these previous tagging studies may not be applied to the sea cucumber population off Maine. Factors such as dermal thick- ness and seasonal fluctuations likely alter the capacity of the sea eumcumber to re- tain tags. The objective of this study was to ascer- tain whether a reliable means of tagging the Maine sea cucumber could be identified. Tagged animals would provide information on growth and movement for later stud- ies. This important biological information would be critical for sea cucumber stock assessment and management. Materials and methods The study was carried out between April and Septem- ber 2003 to determine the effectiveness of various tags. Initially, animals were tested by using 1) long and short double T-bar tags inserted with a tagging gun; 2) dart tags; 3) streamer tags; and 4) cinch-up tags. A control group was also included in which animals were punc- tured with the tagging gun to observe any mortality from the piercing. It was evident in the first minute that dart tags, streamer tags, and cinch-up tags would not stay within the cucumber body wall and these methods were not considered further. Next, 5) a rope gun was tested as a means of scarring (branding) animals. Rope guns are normally used to burn through rope by applying the hot tip to the surface of a rope. For tagging purposes, the rope gun tip was ap- plied to the dermis of C. frondosas to produce a recogniz- able scar; the process is similar to that used in branding cattle. Although the rope gun created a noticeable mark, many animals already possessed significant scarring from being dragged during harvesting or from interac- tions with fishing gear and other marine organisms, and the new scar would not stand out to an observer. The branding method was therefore also discarded. Finally, 6) liquid orange and yellow fluorescent dyes composed of visible implant elastomer (VIE) were in- 700000 ■ 600000 ■ g 500000 ■ O o OJ E, 400000 ■ w ai c ra 300000 • 200000 ■ 100000 ■ ^ rn rn — rn Historica fron dosa 1990 1991 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Year Figure 1 1 landings (metric tons) for the sea cucumber Cucumaria fishery in Maine. jected into the animals. VIE is a silicone-based mate- rial that is mixed immediately before use and quickly becomes a pliable, biocompatible solid. VIE tags were injected internally beneath the epidermal tissue within tube feet, along the body wall, or near the oral cavity. The dye did not spread and all marks were small pin points visible only under ultraviolet light. Short and long double T-bar tags and fluorescent dyes served as the remaining tagging methods studied dur- ing the course of the investigation. Seventy tagged sea cucumbers from the Gulf of Maine were observed over time (Table 1). Ten animals were injected with orange and yellow fluorescent dyes and 45 animals were punc- tured with long or short T-bar tags. Long tags had a distance of 45 mm between T-bar ends. Short tags had a distance of 19 mm between T-bar ends. Twenty-one tags punctured the body wall just under the muscle so that one end remained inside the animal (single-anchor tags). The remaining 24 passed completely through the body wall twice so that both ends of the T-bar tags were exposed (double-pass tags). Finally, 15 control animals were punctured with the tagging gun, 9 with double punctures and 6 with single punctures. Sea cucumbers were kept in tanks in the laboratories at the Maine Department of Marine Resources in West Boothbay Harbor, Maine. The presence and absence of tags was noted every day from April to September 2003 for each sample. Notation was made when tags were partially in place or no longer observable at all. NOTE Kirshenbaum et al A study of tagging methods for Cucumaria frondoso 301 Results and discussion Tagging method Fluorescent dyes Double T-bar tags Double-pass Long tag (91 Short tag (151 Control Tagging gun puncture Total The tagging stu(iy was suc- cessful in providing infor- mation on retention rates for the different tagging methods observed during the 140 days of the investi- gation (Fig. 2). Fluorescent dye injections and single- anchor long T-bar tags were the most successful means identified in tagging sea cucumber in the waters off Maine. Eighty percent with dye continued to fluoresce under the UV light and 65% of single-anchor long T-bar tags remained embedded in the sea cucumbers at the end of the investigation on day 140. The T-bar method may be the most practical for short-term research on this species considering the time and expense of fluorescent dye injection. Another advantage over injections is that the tagging gun technique is much more easily repeatable among subjects, whereas the site of injection varied among animals depending on the loca- tion of soft tissue. Double-pass T-bar tags were expelled faster than those in the single-anchor condition. With the double-pass tag, it appeared that the long portion of the tag under the animal's epidermis gradually moved upward until it was expelled outside of the skin without scarring the animal. The longer bars were pushed out the fastest and all were shed before day 50. Shorter T-bar tags took almost twice as long to reach 0% re- tention as the longer T-bars in the double-pass condition, perhaps because the shorter bars pinched the dermis tightly between tag bars, whereas animals with the longer bars were able to relax their entire body wall. After tags were shed in all T-bar conditions, it was not possible to tell which experimental condition each individual had experienced. No mortality was observed from the tags or from the control punctures. Five T-bar tagged animals experienced a punctured respiratory tree caused by the tagging gun. The respi- ratory tree is an internal organ used for breathing by pumping seawater in and out. After the initial puncture, no trace of the injury could be detected and no differ- ences in appearance or behavior were noted throughout the investigation. Although it was not possible to track which individuals had a partially punctured respiratory tree, injured individuals may have shed tags at a differ- ent rate than others in their cohorts. Internal injuries may also have affected the results of later studies of growth and movement with the use of manual tags. Table 1 Tagging study design for the sea cucumber Cucumaria frondosa off Maine. Total num- bers sampled for each method are indicated as total animals in the right column. Paren- theses identify the number of animals sampled. Total animals 10 Single-anchor Long tag Short tag 45 (6) (15) Double-puncture (9) Single-puncture (6) 15 70 - double-pass long ■ single-anchor long ■ double-pass short - single-anchor short - florescent dye - pooled 0 25 50 75 100 125 150 Number of days after tagging Figure 2 Tagging retention in the sea cucumber Cucumaria frondosa for individual-style tags over 140 days of observation in the waters off Maine. Note that florescent dye and single-anchor long T-bar tags were the most successful methods identified. With the use of a largest possible sample size and care- ful placement of the tagging gun, this potential bias can be minimized. Additionally, it is important to consider that tag retention rate in the field may vary more than results observed in the laboratory. We believe that the fluorescent dye injections and single-anchor long-T-bar tagging methods in our in- vestigation will likely prove effective for a longer-term study with additional preliminary tagging analysis. However, a more effective way to study the growth of sea cucumbers may be the combination of caging and tagging, where tagged individuals of similar size are put in the same cage to monitor a change in size. 302 Fishery Bulletin 104(2) Future investigations will test different locations for, and methods of, injecting fluorescent dye and single- anchor long T-bar tags and will help to further identify the best tagging method. When the most successful method is determined, a long-term study will allow us to understand how individual animals grow, move, and change both seasonally and annually. Acknowledgments The Northeast Consortium, Maine Sea Grant, and Maine Department of Marine Resources provided funding. Kirshenbaum was supported by a fellowship from the Kendall Foundation. Literature cited Bradbury, A., W. A. Palsson, and R. E. Pacunski. 1996. Stock assessment of the commercial sea cucumber Parastichopus californicus in the San Juan Islands, Wash- ington State, USA. J. Shellfish Res. 15:785-786. Cripps, K. E., and A. Campbell. 2000. Distribution and biomass of the red sea cucum- ber, Parastichopus californicus, found in inlets of north British Columbia. Can. Tech. Rep. Fish. Aquat. Sci. 2300, 17 p. Hamel, J. P., and A. Mercier. 1997. Diet and feeding behavior of the sea cucumber Cucumaria frondosa in the St. Lawrence estuary, eastern Canada. Can. J. Zool. 76:1194-1198. Harriott, V. J. 1980. The ecology of Holothurian fauna of Heron Reef and Moreton Bay. M.S. thesis, 153 p. Univ.a Queens- land, Brisbane. Jordan, A. J. 1972. On the ecology and behavior of Cucumaria fron- dosa (Echinodermata: Holothuroidea) at Lamoine Beach, Maine. Ph.D. diss., 74 p. Univ. Maine, Orono, Maine. Medeiros-Bergen, D. E., and E. Miles. 1997. Recruitment in the Holothurian Cucumaria frondosa in the Gulf of Maine. Invert. Reprod. Dev. 31:123-133. Morgan, A. D. 2000. Induction of spawning in the sea cucumber Holo- thuria scabra (Echinodermata: Holothuroidea). J. World Aqua. Soc. 31(2):186-194. Perry, R. I., C. J. Walters, and J. A. Boutillier. 1999. A framework for providing scientific advice for the management of new and developing invertebrate fisheries. Rev. Fish. Biol. Fish. 9:125-150. Phillips, A. C, and J. A. Boutillier 1998. Stock assessment and quota options for the sea cucumber fishery. In Part 2, Echinoderms (invertebrate working paper (no. 22151 reviewed by the Pacific Stock Assessment Review Committee [PSARC ]in 1995 (B. J. Waddell, G. E. Gillespie, and L. C. Walthers, eds.), p 147-169. Schroeter, S. C, D. C. Reed, D. J. Kushner, J. A. Estes, and D. S. Ono. 2001. The use of marine reserves in evaluating the dive fishery for the warty sea cucumber (Parastichopus par- vimensisi in California, USA. Can. J. Fish. Aquat. Sci. 58(9): 1773-1781. Shelley, C. 1981. Aspects of the distribution, reproduction, growth and fishery potential of holothurians (Beche-de-mer) in the Papuan coastal lagoon. M.S. thesis, 165 p. Univ. Papua New Guinea. Singh, R., B. A. MacDonald, P. Lawton, and M. L. H. Thomas. 1997. Feeding response of the dendrochirote sea cucum- ber Cucumaria frondosa (Echinodermata: Holothur- oidea) to changing food concentrations in the labor- atory. Can. J. Zool. 76:1842-1849. Yanagisawa, T. 1995. Sea cucumber ranching in Japan and some sug- gestions for the South Pacific. Tonga Aquaculture Workshop 387-400. Zhou, S., and T. C. Shirley. 1996. Habitat and depth distribution of the red sea cucumber Parastichopus californicus in a southeast Alaska bav. Alaska Fish. Res. Bull. 3:123-131. 303 Parameterizing probabilities for estimating age-composition distributions for mixture models Daniel K. Kimura Martin W. Dorn Alaska Fisheries Science Center National Marine Fistieries Service 7600 Sand Point Way N E Seattle, Washington 98115-6349 E-mail (for D Kimura) Dan Kimuraa noaa gov Heifetz and Fujioka parameterization Heifetz and Fujioka (1991), for a tagged fish migration problem, used a somewhat similar parameterization that guaranteed estimated param- eters would be probability distribu- tions. Suppose a distribution consisted of categories with probabilities [p I. The Heifetz-Fujioka (H-F) parameter- ization would be ^.; = vV(l-^'^'-X'''-' j,k = l,...,a-l and When estimating parameters that constitute a discrete probability distribution |p^|, it is difficult to determine how constraints should be made to guarantee that the esti- mated parameters Ip^l constitute a probability distribution (i.e., p,>0, 5:p^ = ll. For age distributions esti- mated from mixtures of length-at-age distributions, the EM (expectation- maximization) algorithm (Hassel- blad, 1966; Hoenig and Heisey, 1987; Kimura and Chikuni, 1987), restricted least squares (Clark, 1981), and weak quasisolutions (Troynikov, 2004) have all been used. Each of these meth- ods appears to guarantee that the estimated distribution will be a true probability distribution with all cat- egories greater than or equal to zero and with individual probabilities that sum to one. In addition, all these methods appear to provide a theo- retical basis for solutions that will be either maximum-likelihood estimates or at least convergent to a probability distribution. However, all these methods are pre- sented in a theoretical context that is useful in understanding the theory, but may not be suitable for actual application. Currently, most modelers will have an optimization program available. This note describes how, in a brief amount of time, the modeler can add a mixture model program to his collection of readily available programs — one that will estimate maximum-likelihood estimates for the mixture problem and will incor- porate the experimenter's familiar optimization program. To do this it is necessary to pa- rameterize the estimated probabili- ties so that they are in fact guar- anteed to constitute a probability distribution (i.e., p,>0, 2p^=l). Two such parameterizations are the mul- tinomial logit and a parameteriza- tion method described by Heifetz and Fujioka (1991). The trick with both parameterizations is that although the parameters that are actually estimated are unconstrained, these unconstrained estimates can be eas- ily transformed to constrained prob- ability estimates. Materials and methods Multinomial logit parameterization Consider parameterizing the proba- bilities using the classic logit model: a+Xr, j,k = l,...,a-l and Pa d + I'-/,) The r^ can be guaranteed to be positive by defining r =exp(.v ). The parameters that are estimated are x^lnir^). In turn the [x^] are used to estimate the \r^\, and then the Ip^j. Thus, .Tp ... , .tjj [ if are estimated on (-00, cx>), the resulting |p^| is guaran- teed to be a probability distribution. The relationship between the Ip^l and l?-^! is a unique one-to-one mapping because given any Ip^l, r=p^/p^, for J=l a-1. P„=exp(-X'>' vhere /), > 0 for ^ = 1 a-1. As above r^, can be guaranteed to be positive by defining r^ = exp(.r^) and the parameters that are estimated are x^=ln(rj). Also as above, \x ] are used to estimate the |r ), and then the Ip^K Thus, if .tj, . . . , .t^,_j are esti- mated on (-00, 00), the resulting |p^) is guaranteed to be a probability dis- tribution. The relationship between the Ip^l and {r^l is a unique one-to- one mapping because given any |p I, r^=-p,ln(p„)/(l-pj, for./=l a-l. H-F probabilities as exploitation rates Stock assessment modelers would rec- ognize the formula P., lo,. (l-exp(-Xr^)) as the exploitation rate formula where p^ takes the role of the exploitation rate ((/ ) and r takes the role of the instantaneous mortality rate F^. The formula r^ = -p^ln(p^,)/(l-p^) indicates that the instantaneous mortality rate can be solved in closed form, but this is generally not true because the nat- ural mortality rate M is generally known as an instantaneous mortality rate rather than an exploitation rate. Manuscript .submitted 20 January 200.5 to the Scientific Editor's Office. Manuscript approved for publication 2 August 2005 by the Scientific Editor. Fish. Bull. 104:303-305 (2006). 304 Fishery Bulletin 104(2) However, when exploitation rates are all known, as in some ecological modeling exercises (e.g., u=C^IN^, i.e., catch per population), then all instantaneous mortality rates can be solved in closed form. Example: As an example, the logit and H-F parameter- ization can be used to estimate the distribution mixture problem for empirical length-at-age data. To describe the mixture problem for empirical length-at-age data, let / = 1 ^ be the length category index; 7 = 1 a be the age category index; /, = the observed length frequency, for which we wish to estimate the corresponding age distribution; /", = fjf. where f=^f,. the observed length distribution; = the unknown age distribution we wish to estimate from the observed length distribution; = the observed distribution of length at age, estimated from age-length data; Ij = '^pfl,, = length distribution estimated from an age distribution and observed length- at-age distributions. Solutions to the empirical distribution mixture problem can be stated as solving for that age distribution Ip^l, which when combined with length / at agej, say l For both of these estimation problems, the estimated Ip I distribution must be constrained to sum to one. The logit or H-F parameterization simplifies and unifies the estimation of Ip^l for these or any other objective func- tion. Any multivariate function optimization program, using these parameterizations, can generate estimated probability distributions whose components are positive and will be guaranteed to sum to one. Therefore any multivariate optimization program with these param- eterizations can be used to estimate Ip I. Results Greenland turbot {Reinhardtius hippoglossoides) length- at-age data were originally used to illustrate the iterated age-length key (lALK) (Kimura and Chikuni, 1987). Table 1 Age dist ributions for Greenl and turbot iRei n hard tins hipp oglotiaoide si estimated by fitting length at age data | to 1983 lengtl i-frequency dat a (Kimura and Chikuni, 19871. E accept for the iterated age-length key (lALKl | algorithm, all estimates were arrived at by using the logit or Heifet s-Fujioka (H-Fl parameterizations which | provided ident cal results. Vlaximum lALK Vlinimuni Age yr) ikelihood algorithm x'- 4 0.0353 0.0353 0.0377 5 0.1903 0.1903 0.18.58 6 0.2281 0.2281 0.2229 7 0.1291 0.1291 0.1318 8 0.1125 0.1125 0.1117 9 0.0380 0.0380 0.0349 10 0.0525 0.0525 0.0554 11 0.0000 0.0000 0.0000 12 0.0332 0.0332 0.0383 13 0.0023 0.0023 0.0000 14 0.0204 0.0204 0.0235 15 0.0168 0.0168 0.0162 16 0.0289 0.0289 0.0248 17 0.0253 0.0253 0.0218 18 0.0680 0.0680 0.0719 19 0.0042 0.0042 0.0036 20 0.0152 0.0152 0.0198 We used the 1983 length-frequency data from that data set, along with the length-at-age data, to illustrate the methods of estimating mixture probabilities with the logit and H-F parameterizations (Table 1). Except for results from the lALK algorithm, all parameters were estimated by using an optimization program with the logit and H-F parameterizations (i.e., the latter two parameterizations gave identical results). Results also showed that maximum-likelihood estimates from either the logit or H-F parameterizations provided maximum- likelihood estimates identical to those estimated by using the TALK algorithm. Discussion It is probable that the logit and H-F parameterizations would provide nearly identical solutions for a given data set and objective function. If maximum likelihood solu- tions are unique, either parameterization should provide the maximum-likelihood estimates because the repa- rameterizations of probabilities are one-one mappings that lead to invariance properties when optimization is performed. The maximum-likelihood solutions will gen- erally be unique when all p,>0 (Kimura and Chikuni, 1987). However, difficulties in searching in multivariate spaces, and limited computational precision may cause differences in the estimates. NOTE Kimura and Dorn: Parameterizing probabilities for estimating age-composition distributions for mixture models. 305 Simple parameterizations, like direct estimates of Ip I, will allow solutions that have negative probabil- ities. Constrained solutions, which allow maximum likelihood estimates and other types of estimates, will generally make these negative components of the prob- ability distribution have zero values. Such solutions are boundary value solutions and may not be unique (Kimura and Chikuni, 1987). This is another reason it is difficult to claim unique solutions for the mixture problem. From the modeling perspective, we illustrate the use- fulness of reparameterization to impose mathematical constraints. In the context of the mixture problem the suggested parameterizations are reasonably transpar- ent and allow the modeler to use familiar software. The reason we propose the methods described in this note is not that these methods provide superior estimates to those described in the literature, but that the procedure for estimation may actually be more straightforward and transparent for modelers more interested in solu- tions than in theory. Because of its simplicity, effectiveness, and ready applicability to different objective functions, modelers may prefer optimization using the logit or H-F param- eterizations to estimate probability distributions for the mixture problem. Another advantage of these reparam- eterizations is that they can be more generally applied, for example, to estimate geographic distribution in mi- gration models (Heifetz and Fujioka, 1991; Shimada and Kimura, 1994). Acknowledgments We thank two anonymous referees for comments that helped us clarify our presentation. Literature cited Clark, W. G. 1981. Restricted lea.st-squares estimates of age compo- sition from length composition. Can. J. Fish. Aquat. Sci. 38:297-307. Hasselblad, V. 1966. Estimation of parameters for a mixture of normal distributions. Technometrics 8:431-444. Heifetz, J., and J. T. Fujioka. 1991. Movement dynamics of tagged sablefish in the northeast Pacific. Fish. Res. 11:355-374. Hoenig, J. M., and D. M. Heisey. 1987. Use of a log-linear model with the EM algorithm to correct estimates of stock composition and to convert length to age. Trans. Am. Fish. Soc. 116:232-243. Kimura, D. K., and S. Chikuni. 1987. Mixtures of empirical distributions: an itera- tive application of the age-length key. Biometrics 43:23-35. Shimada, A. M., and D. K. Kimura 1994. Seasonal movements of Pacific cod, Gadus maci'o- cephalus, in the eastern Bering Sea and adjacent waters based on tag-recapture data. Fish Bull. 92:800-816. Troynikov, V. S. 2004. Use of weak quasi-solutions of the Fredholm first- kind equation in problems with scarce data. Appl. Math. Comput. 150:855-863. 306 Morphometric differentiation in small juveniles of the pink spotted shrimp iFarfantepenaeus brasi/iensis) and the southern pink shrimp (f. notialis) in the Yucatan Peninsula, Mexico Marco A. May-Ku Uriel Ordonez-Lopez Omar Defeo Centra de Investigacion y de Estudios Avanzados, CINVESTAV Unidad Merida Carretera Antigua a Progreso Km. 6, A P 73 Cordemex, 97310 Menda, Yucatan, Mexico E-mail address (for M A Kay-Ku) mayfa^mdacinvestavmx The morphometric and morphologi- cal characters of the rostrum have been widely used to identify penaeid shrimp species (Heales et al., 1985; Ball et al., 1990; Pendrey et al., 1999). In this setting, one of the constraints in studies of penaeid shrimp popu- lations has been the uncertainty in the identification of early life history stages, especially in coastal nursery habitats, where recruits and juve- niles dominate the population (Ball et al., 1990; Perez-Castaiieda and Befeo, 2001). In the western Atlantic Ocean, Perez-Farfante (1969, 1970, 1971a) described diagnostic charac- ters of the genus Farfantepenaeus that allowed identification of indi- viduals in the range of 8-20 mm CL (carapace length) on the basis of the following morphological features: 1) changes in the structure of the pe- tasma and thelycum; 2) absence or presence of distomarginal spines in the ventral costa of the petasma; 3) the ratio between the keel height and the sulcus width of the sixth abdominal somite; 4) the shape and position of the rostrum with respect to the segments and flagellum of the antennule; and 5) the ratio between rostrum length (RL) and carapace length (RL/CL). In addition, she classified Farfantepenaeus into two groups according to the shape and position of the rostrum with respect to the segments and flagellum of the antennule and the ratio RL/CL: 1) F. duorarum and F. notialis: short rostrum, straight distally, and the proximodorsal margin convex, usu- ally extending anteriorly to the end of distal antennular segment, sometimes reaching to proximal one-fourth of broadened portion of lateral antennu- lar flagellum, with RL/CL <0.75; and 2) F. aztecus, F. brasiliensis, F. paiilen- sis, and F. subtilis: long rostrum, usu- ally almost straight along the entire length, extending anteriorly beyond the distal antennular segment, some- times reaching to the distal one-third of broadened portion of lateral anten- nular flagellum, with RL/CL >G.80. Perez-Farfante stressed that, for the recognition to species level of juveniles <10 mm CL, all the characters listed above should be considered because occasionally one alone may not prove to be diagnostic. However, the only characters that could be distinguished for small juveniles in the range 4-8 mm CL are those defined on the ros- trum. Therefore, it has been almost impossible to identify and separate small specimens of Farfantepenaeus (Perez-Farfante, 1970, 1971a; Perez- Farfante and Kensley, 1997). Pink spotted shrimp (F. brasilien- sis) and southern pink shrimp (F. notialis) have a wide geographic dis- tribution in coastal environments of the Atlantic Ocean. The pink spot- ted shrimp is distributed within the western Atlantic from Cape Hatteras, North Carolina, to Cabo Frio, Brazil (22°00'S, 42°00'W; Perez-Farfante, 1969), including the southwestern Gulf of Mexico and the Caribbean coasts of Mexico (Perez-Farfante, 1971b). The southern pink shrimp ranges from the northwestern Yu- catan Peninsula, Mexico (20°45'N, 90°25'W; Perez-Castaiieda and Be- feo, 2000), to Rio de Janeiro (Brazil) in the western Atlantic, including Cuba and the Virgin Islands in the Caribbean Sea. The species is also reported from Mauritania to Angola in the eastern Atlantic Ocean (Perez- Farfante, 1969; Perez-Farfante and Kensley, 1997). The current identification keys show that, for shrimps >8 mm CL, the pink spotted shrimp has a pro- portionately longer rostrum than the southern pink shrimp. However, there are no scientific data to sepa- rate these species at individual sizes <8 mm CL (Perez-Farfante, 1970, 1971a). This is particularly impor- tant where both penaeid shrimps are found in the same region, mainly between the Yucatan Peninsula and along the Caribbean coast of Mexico, Puerto Rico, and Colombia. In this geographic region, small juveniles of both species have been almost impos- sible to separate and have been clas- sified only to genus level (Stoner and Zimmerman, 1998; May-Kii, 1999; Perez-Castafieda and Befeo, 2001; Criales et al., 2002). In this note we provide quantitative and qualitative information that allows separation of the sympatric F. brasiliensis and F. notialis in the range of 4-8 mm CL based on shrimp collected in the Rio Lagartos and Yalahau coastal lagoons of the Yucatan Peninsula, Mexico. Materials and methods Shrimp were collected by using a Renfro beam trawl (1.6x0.5 m mouth, 1.5 m total length, and 1.0-mm mesh), which was hand hauled in two coastal lagoons of northeastern Yucatan Peninsula (Fig. 1). The Ri'o Lagartos Manuscript submitted 8 March 2005 to the Scientific Editor's Office. Manuscript approved for publication 22 August 2005 by the Scientific Editor. Fish. Bull. 104:306-310 NOTE May Ku et a\ Morphological differentiation between small [uveniles of Farfantepenaeas biasiliensis and F notialis 307 lagoon (21'^^26'-21°38'N, 87°30'-88°15'W) was sampled from November 1996 to April 1997, whereas the Yalahau lagoon (21''26'-21°36'N, 87°08'-87°29'W) was sampled from June 2001 to May 2002. These coastal lagoons are consid- ered the main nursery areas for F. brasiliensis, accounting for 80-95% of the total shrimp abun- dance, followed by F. notialis (May-Kii, 1999; May-Kii and Ordofiez-Lopez, 2000). Shrimps (4-10 mm CL) were preserved in 10% formal- dehyde, prior to examination. Rostrum length (RL; distance from the tip of the rostrum to the postorbital margin of the carapace) and carapace length (CL; distance from the postorbital margin to the posterior margin of the carapace) were measured with an ocular micrometer to the near- est 0.1 mm. Furthermore, the shape and position of the rostrum with respect to the segments and flagellum of the antennule were considered (Perez-Farfante, 1969; 1970, 1971a). Shrimps ranging from 8 to 10 mm CL were identified to species level according to the characters out- lined by Perez-Farfante (1970, 1971a). The rela- tionship between rostrum length and carapace length (RL-CL) was determined for each spe- cies (F. brasilie?isis. F. notialis) in each lagoon, and fitted to the linear function RL = a + b CL, where a and b are coefficients. One-way analysis of covariance (ANCOVA) was used to evaluate differences between species in the RL-CL rela- tionship for each site, with RL as the dependent variable and CL as the covariate. Assumptions of homoscedasticity and homogeneity of slopes (parallelism) were met after log-transformation of RL. Lastly, the RL/CL ratio was estimated to remove the differences in individual sizes between species, then plotted against CL and determined for each lagoon. A one-way analy- sis of variance (ANOVA) was conducted to test interspecific differences between mean RL/CL in each lagoon. The original RL/CL values were log-transformed to fulfill ANOVA assumptions. Results A total of 13,234 shrimps in the size range 4-10 mm CL were collected. Of the 727 organisms measured in the Rio Lagartos lagoon, 74.6% were F. brasiliensis and 25.4% F. notialis. In Yalahau lagoon, from the 12,507 organisms measured 91.7% were F. brasiliensis and 8.3% F. notialis In the two lagoons, both species showed a signifi- cant positive linear RL-CL relationship (r->0.92 P<0.001). Significant differences were detected between the species (ANCOVA: P<0.001; Fig. 2 Table 1) — F. brasiliensis having a longer rostrum than F. notialis for a given CL. The RL/CL ratio for both species ranged from 0.42 to 0.94. In both lagoons the scatter plots il — 1 1 1 1 1 45 W mr w N t ■ 2.'i" N - 1 Gulf of Mexico - ^ R io Lagartos Yalahau Yucatan ] Peninsula /{7 / 211" N - S--^ 1' Caribbean Sea ■ Figure 1 Location of Rio Lagartos and Yalahau lagoons off the Yucatan Penin- sula. The distribution of Farfantepenacuf brasiliensis and F. notialis along the Mexican coast of the Gulf of Mexico and Caribbean Sea is indicated by oblique lines. 9- 8- 7 6 5 4 I '^ £ 2-1 01 = -0 427 + 0 840 CL, f^ = 0.96 . i ."tl 1'':'m''»*»5 ' . ■;;,;i;iii"^-- ^ V •: * . ' ■'»'' .!-' :t:,Mi« RL = -0 321 + 0 679 CL; f2 = 0.94 ° ft 7- 6- 5- 4- 3- 2- B t: Ht = -0.338 + 0 853 CL: f2 = 0 96 _ ^ l]. 1 1 ijl; ■ . i i * !|: iji- s ... 1' , ! 1 '4 . si' M « i ; ipl" 8 HL = 0.049 + 0.625 CL; (-2 = 0.92 6 7 8 Carapace lengtfi (mm) 10 Figure 2 Relationships between rostrum length (RL) and carapace length (CL) fitted by the linear model RL = a + h CL for juveniles of Farfantepenaeus brasiliensis (▲) and F. notialis (0) from (A) Rio Lagartos and (B) Yalahau lagoons. All regressions are highly significant (P<0.001l. 308 Fishery Bulletin 104(2) Table 1 Parameters of the linear regression model between rostrum length (RL) and carapace length (CL) fitted for Farfantepenaeus brasiliensis and F. notialis, in the size range of 4-10 mm CL, for Rio Lagartos and Yalahau. ANCOVA results (log-transformed RL) are also shown, including tests of homogeneity of slopes and differences in RL at CL (main effect) between species, "a" and "h" represent, respectively, the intercept and slope of the regression lines. SE = standard error; ns = not significant; *** = P<0.001. a Mean(SEi b Mean(SE) r- n P Rio Lagartos F. brasiliensis -0.427 (0.055)*** 0.840(0.007)*** 0.96 542 F.notialis - -0.321(0.091)*** 0.679(0.013)*** 0.94 185 ANCOVA Homogeneity of slopes: F, ,23 = 1.212; P=0.271 Main effect: Fi -.,^ = 1.430.2; P=0.00001 Yalahau F. brasiliensis -0.338 (0.010)*** 0.853(0.002)*** 0.96 11,467 *** F. notialis 0.049 (0.032) ns 0.625(0.006)*** 0.92 1040 ANCOVA Homogeneity of slopes: F^ 10503 = 0.966; P=0.326 Main effect: Fj i25o^ = 12486.4: P=0.0001 ^ 0.9 E S O.Qi cc 0.7 0.6 0.5 A 0.9 A ' ' ;* * i ' A' ' 0.8 * *» :•'..:■. '•|'.:::;;'"i=';i''^:*:ii'*i^v''!:*^:^:; *' ! * : r 0./ 0,6 «-*- »» -4- - -,* .^J,.t .K^^.t*. '-• L'^ i 4.- 0) *^ * ^ ° 0 n 0.5 ^ ra B ;;::|:::::|::::i;;i:::;::i' 6 7 8 Carapace length (mm) 10 Figure 3 Scatter plot of rostrum length/carapace length (RL/CL) ratios against CL for juveniles of Farfantepenaeus brasiliensis (A) and F. notialis (0) from (A) Rio Lagartos and (B) Yalahau lagoons. The separation between species at RL/CL = 0.70 is indicated by a dotted line. of RL/CL ratios against CL by species fell into two distinct clouds representing dissimilar mor- phometric characteristics of the species, with the point of separation between species set at a value of 0.70 (Fig. 3). The RL/CL ratios were signifi- cantly higher in F. brasiliensis than in F. notialis, both for Rio Lagartos (ANOVA, F^ 72,5 = 1'778.33, P<0.001) and Yalahau (ANOVA Fji2505=16627.64, P<0.001). For Rio Lagartos, F. brasiliensis had a mean RL/CL ratio ±SD of 0.78 ±0.04 and a range of 0.70-0.92, whereas the mean RL/CL ratio in F. rwtialis was 0.63 ±0.04, with a range of 0.50-0.72. In Yalahau lagoon, the RL/CL ratio in F. brasil- iensis was 0.80 ±0.05 (range 0.68-0.94), where- as in F. notialis these values were 0.62 ±0.05 (0.43-0.69; Fig. 4). The above values indicate a morphometric divergence between the two species, with higher values corresponding to F. brasiliensis and lower ones to F. notialis. Small juveniles presented clear differences be- tween species in the shape and position of the rostrum with respect to the segments and the flagellum of the antennule. Farfantepeaeus brasil- iensis had a rostrum that was straight and slen- der, extending beyond the 3'^'^ antennular segment, and generally reaching the broadened portion of lateral flagellum (Fig. 5A), whereas F. notialis had a lightly convex rostrum, never extending beyond the third antennular segment (Fig. 5B). Discussion We found clear morphometrical and morphologi- cal rostrum differences between small juveniles NOTE May-Ku et al Morphological differentiation between small |uveniles of Farfantepenaeus brasiliensis and F notiolis 309 ~r ^ Max. Mm ^ 09 CD Mean± ISD C r • Mean Q) 08 . CL 5 0.7 O f, 06 c 0) • y I 1 [ -r Rostrum o o — 1— Rio Lagartos Yalahau Rio Lagartos Yalahau F brasiliensis F. notialis Figure 4 Box and whisker plots of the rostrum length/carapace length (RL/CL) ratios for Farfantepenaeus brasiliensis and F. notialis, discriminated by lagoon. (4-8 mm CL) of F. brasiliensis and F. notialis in two coastal lagoons of the Gulf of Mexico. Our results also complement those obtained by Perez-Farfante (1969; 1970, 1971a) because the unique characters of the ros- trum highlighted by that author for juveniles 28 mm CL persisted in our study for small juveniles (<8 mm CL). We also found interspecific differences in the growth pat- terns given by the RL-CL relationship (slopes) and in the RL/CL ratio (intercept differences). These results have important implications. Dall et al. (1990) mentioned that the shape and increase in size of different body struc- tures in crustaceans depend on the species, age, sex, and the surrounding environment. Our comparison of the regression lines between RL and CL of the two shrimp species analyzed in the present study, coming from Yala- hau and Rio Lagartos lagoons, showed unambiguously that the growth pattern, or degree of change in RL in relation to CL, differed between the two species at all sizes of the specimens. Different intercepts, on the other hand, implied that although the growth patterns may be the same over the size range measured (regression lines are almost parallel), the relative body proportions of both species are different. These differences are likely due to different growth rates at some earlier size (Haddon and Willis, 1995). Both types of difference imply that the relative proportions of the body, or the body shape, are distinct for the species analyzed (Haddon and Willis, 1995). An analysis of the RL-CL relationship revealed that, at a given CL, the RL in F. brasiliensis grows 1.23-1.29 times faster than in F. notialis. On the other hand, the marked differences in the RL/CL ratio imply that this morphometric character is species-specific and, at a given CL, the pink spotted shrimp has a consistently longer rostrum than southern pink shrimp, with a value of 0.70 as a point of separation between species. These results are also qualitatively observed in the shape and position of the rostrum with respect to the segments and the flagellum of the antennule in both species. CL RL A 1 -«.— -Zl.-i-4-T-^-X.x,,,,^ 1 /^"^''"^--l^^S ■"^'o antennular ^ '' segment 1 • . '. . '"'"'""••• ' 'iiy»-''^<^" ' " ''" K.^^.^^^^1 Lateral flagellum B 1 --z,._-ri-x-t^ 1 P" ■-■'"■■ ~-^^^^^^^ I """'■""'"'''"T^^^S^; Ctee^ ■ ■. ■- •■r^-"----- ■■^.j'^w!!!^ ^^.^i^i^^^^ % Figure 5 Distinguishing features of the shape of the rostrum and its position with respect to the segments and the flagellum of the antennule in Farfantepenaeus brasil- iensis (A) and F. notialis iB). CL = carapace length. RL = rostrum length. In conclusion, we successfully separated small juve- niles of two shrimp species that are found in coastal lagoons of the Gulf of Mexico. Both species can be iden- tified by easily observed and quantified rostrum char- acteristics. Acknowledgments This work is part of the Ph.D. thesis of M. A. May-Kii at CINVESTAV-Merida. Financial support from CONA- CyT Mexico (grant to M.A. May-Kii) is acknowledged. The authors are grateful to M. Ornelas-Roa from the Plankton Laboratory at CINVESTAV-Merida for her help in obtaining morphometric data. Two anonymous referees provided useful suggestions that significantly improved the manuscript. Literature cited Criales, M., C. Yeung. F. Amaya, A. Lopez, D. Jones, and W. Richards. 2002. Larval supply of fishes, shrimps, and crabs into the nursery ground of the Cienaga Grande de Santa Marta, Colombian Caribbean. Caribb. J. Sci. 38:52-65. Dall, W.. B. J. Hill, P. C. Rothlisberg, and D. J. Staples. 1990. The biology of the Penaeidae. Adv. Mar. Biol. 27:1-489. Haddon, M., and T.J. Willis. 1995. Morphometric and meristic comparison of orange roughy iHoplostethus atlanticus: Trachichthyidae) from 310 Fishery Bulletin 104(2) the Puysegur Bank and Lord Howe Rise, New Zealand, and its implications for stock structure. Mar. Biol. 123:19-27. Heales, D. S., H. G. Polzin, and D. J. Staples. 1985. Identification of the postlarvae of the commercially important Penaeus species in Australia. In Second Australian national prawn seminar (P. Rothlisberg, B. Hill and D. Staples, eds.), p. 41-46. Seminar Sponsor, Cleveland, Queensland, Australia. May-Kii, M. A. 1999. Caracteristicas biologicas y ecologicas de los juveniles de camaron del genero Farfantepenaeiis, en el area oeste de la laguna costera de Rio Lagartos, Yucatan. M.Sc. thesis, 108 p. Centre de Investigacion y de Estudios Avanzados (CINVESTAV), Unidad Merida, Merida, Mexico. May-Kii, M. A., and U. Ordonez-Lopez. 2000. The shrimp Farfantepenaeiis notialis (Decapoda: Penaeidae) in the Gulf of Mexico. Rev. Biol. Trop. 48:271. Pendrey, R. C, N. R. Loneragan, R. A. Kenyon, and D. J. Vance. 1999. Simple morphometric characters, confirmed by gel electrophoresis, separate .small juvenile banana prawns tPenaeiis indicus and P. merguiensis). Mar. Freshw. Res. 50:677-680. Perez-Castafieda, R., and O. Defeo. 2000, Population structure of the penaeid shrimp Far- fantepenaeus notialis in its new range extension for the Gulf of Mexico. Bull. Mar. Sci. 67:1069-1074. 2001. Population variability of four sympatric penaeid shrimps (Farfantepenaeiis spp.) in a tropical coastal lagoon of Mexico. Estuar. Coast. Shelf Sci. 52:6.31-641. Perez-Farfante, L 1969. Western Atlantic shrimps of the genus Penaeus. U.S. Fish Wildl. Serv.. Fish. Bull. 67:461-591. 1970. Diagnostic characters of juveniles of the shrimps Penaeus aztecus aztecus, P. duorarum, and P. brasil- iensis (Crustacea, Decapada, Penaeidae). U.S. Fish Wildl. Serv., Special Rep.— Fisheries 599, 26 p. U.S. Dep. Interior, U.S. Fish and Wildlife Serv., Bureau of Comercial Fisheries, Washington, D.C. 1971a. Caracteristicas diagnbsticas de los juveniles de Penaeus aztecus subtilis, P. duorarum notialis y P. brasiliensis (Crustacea Decapoda Penaeidae). Mem. Soc. Cienc. Nat. La Salle 30:159-182. 1971b. Range extension of the shrimp Penaeus iMeli- certus) brasiliensis Latreille, 1817 (Decapoda, Penaei- dae). Bull. Mar. Sci, 21:745-746. Perez-Farfante, I., and P. Kensley. 1997. Penaeoid and sergesteoid shrimps and prawns of the world. Key and diagnoses for the families and genera. Mem. Mus. Nat. Hist. Natur. Paris 175, 233 p. Stoner, A. W., and R. Zimmerman. 1998. Food pathways associated with penaeid shrimps in a mangrove-fringed estuary. Bull. Mar. Sci. 86: 543-551. 311 Exploring intraspecific life history patterns in sharks Jason M. Cope School of Aquatic and Fishery Sciences University of Washington 1122 NE Boat St Seattle, Washington 98105 E-mail address icopeo u Washington edu Marine ecosystems compose the major source (85%) of world fisheries pro- duction (Garcia and Newton, 1997). Although only a few fish species tend to dominate fishery catches (Jennings et al., 2001), a large diversity of fishes representing varied taxonomic levels, ecological guilds, and life histories is commonly taken. Recently, 66% of global marine resources were deter- mined to be either fully, heavily, or over-exploited (Botsford et al., 1997). Considering the current state of many fisheries, the large diversity of spe- cies taken globally, and the general lack of resources to adequately assess many stocks, it has become important to develop shortcuts that may pro- vide methods fisheries scientists can use to determine which stocks are in danger of overexploitation and which recovery plans are appropriate when biological data are limited (Stobutzki et al., 2001). Applications of life history the- ory have proven a potentially use- ful means to accomplish such tasks (Stearns, 1992; Reynolds et al., 2001). Life history traits such as maxi- mum size and age, maturity, mortal- ity, and growth are correlated among teleost fishes (Adams, 1980; Wine- miller and Rose, 1992; Gunderson, 1997; Cortes, 2000) and the relation- ships among such traits can be used to infer some general life history pat- terns. These general patterns reveal that teleost fishes with higher maxi- mum ages tend to be larger, mature later, grow more slowly, and have lower natural mortality rates (K-se- lected species, Adams, 1980), where- as teleost fishes with lower maximum ages tend to show the opposite rela- tionships (r-selected species, Adams, 1980). Correlations among traits may also allow one to approximate dif- ficult to measure life history traits from traits that are easier to mea- sure and possibly anticipate response to exploitation rates where life his- tory data are limited (Jennings et al., 1999). Applying these patterns to fisher- ies trends reveals some consistent and useful explanations. Jennings et al. (1998) found that teleost fishes from the northeast Atlantic that have decreased in abundance are gener- ally K-selected species. Jennings et al. (1999) demonstrated that tropical teleost fishes of greatest maximum sizes were most vulnerable to exploi- tation. And Rochet (2000) illustrated the limitations of life history plas- ticity to compensate for heavy fish- ing pressure among four orders of teleosts. Elasmobranchs, and particularly sharks, have also shown life history patterns similar to those of teleosts (Cortes, 2000; Frisk et al., 2001). Cortes (2000) offered three general life history patterns for sharks; 1) large litters, moderate to high lon- gevity, large size, small offspring, slow growth, 2) small litters, high longevity, large size, large offspring, slow growth and 3) small litters, low longevity, small size, small offspring, fast growth. Simplified applications of the life history patterns have also been applied to elasmobranch fisher- ies. Smith et al. (1998) demonstrated that larger, later-to-mature Pacific shark species have lower rebound po- tentials (i.e., abilities to recover from fishing pressure), whereas Frisk et al. (2001) showed a similar pattern in sharks and rays in the north At- lantic. These relationships have been recommended as particularly useful when managing data-poor elasmo- branch species (Musick et al., 2000). As indicated by the above studies, variation in life history traits and patterns among shark species is well established (Cortes, 2000) and such relationships may be useful for the management of these fishes, but it is not known how these relationships may change within a species. Specifi- cally, if and how do intraspecific life history traits of cosmopolitan species vary in different areas of the world? The spiny dogfish (Squalus acanth- ias) provides an alluring preamble to the topic: northeast Pacific spiny dogfish have been aged to 80-i- years, and females mature at around 35 years (Jones and Geen, 1977; Saun- ders and McFarlane, 1993), whereas spiny dogfish in the north Atlantic obtain a maximum age of about 40 years, maturing at 12 years (Rago et al. 1998). In the present study, I used gen- eralized linear models (GLMs) to in- vestigate whether spatial differences in life history traits, such as those seen in the spiny dogfish, reveal con- sistent patterns when compared with other spatially resolved life history information from other shark species. I then, as demonstration of potential utility, applied these models to pre- dict life history trait values for areas lacking information for two species of shark, spiny dogfish (S. acanthias) and blue shark iPiionace glauca). Materials and methods Information for five life history traits (age at maturity, longevity (maximum age), mean fecundity, maximum size, and size at maturity) from 17 shark species in six families (Appendix, Table 1) for seven general areas (North Pacific (NP); North Atlantic (NoA); Gulf of Mexico (GM); Indian Ocean (I); Central Pacific (CP); South Pacific (SP); South Atlantic (SA)) was extracted from three primary literature sources (Smith et al., 1998; Cortes, 2000, 2002). Area distinc- tions were based on those by Cortes Manuscript submitted 6 December 2004 to the Scientific Editor's Office. Manuscript approved for publication 1.5 September 2005 by the Scientific Editor. Fish. Bull. 104:311-320 12006). 312 Fishery Bulletin 104(2) Table 1 List of shark species used in the analyses Species abbreviations are used in tables and figures to simplify presentation. Scientific name Common name Abbreviation Carcharhinidae Carcharinus acronotus Blacknose shark Cac Carcharinus amblyrhynchos Grey reef shark Cal Carcharinus falciformis Silky shark Cfa Carcharinus leucas Bull shark Cle Carcharinus limbatus Blacktip shark Cli Carcharinus longimanus Oceanic whitetip shark Clo Carcharinus obscurus Dusky shark Cob Carcharinus plumbeus Sandbar shark Cpl Galeocerdo cuvier Tiger shark Gcu Prionace glauca Blue shark Pgl Rhizopnonodon taylori Australian sharpnose s hark Rta Sphyrnidae Sphyrna lewini Scalloped hammerhead shark Sle Triakidae Galeorhinus galeus Soupfin shark Gga Lamnidae Isurus oxyrinchus Mako shark lox Alophiidae Alopias superciliosus Bigeye thresher shark Asu Squatinidae Squahis acanthias Spiny dogfish Sac Squalus mitsukurii Shortspine spurdog Smi (2000). Spatial resolution to the particular ocean basins in this stu(iy was defined by the information I was able to obtain, as were the choice of species and life history traits to analyze. Although other information for these and other species may be currently available, I limited my data to those found in primary peer-reviewed litera- ture. When mean fecundity values were unavailable, the mean was assumed to be the middle value of the fecun- dity range given. Phylogenetic variance was controlled because I strictly evaluated intraspecific comparisons. Initial data exploration was performed by visualizing intraspecific gender-based pairwise comparisons by area with dot plots (Fig. 1). Within each species and gender, the outcome of each comparison (i.e., the value for a particular trait in one area greater than, less than, or equal to that of another area) was evaluated. An overall relationship among areas for each life history trait was then constructed as a composite of each pairwise result. The purpose of this exercise was to visually explore the data and assess whether any intraspecific life history patterns by area were apparent. A GLM framework was then used to construct simple models that quantitatively relate the effect of certain factors (e.g., area, gender, or taxonomic level) to a re- sponse variable — in this case to a particular life history trait. The flexibility of the GLM framework also al- lows one to consider non-normal response distributions while maintaining the advantages of linear regression (Venables and Ripley, 2002) by means of a link func- tion relating the response variable mean to the linear predictors. I had no a priori knowledge of the variance structure for each life history trait (some of which will not have variance, given that they are maximum re- corded values), so both lognormal (with an identity link) and gamma (with a log link) distributions were considered because of their appropriateness to continu- ous and nonzero data. Akaike's information criterion (AIC; Burnham and Anderson, 2002) was used to select among models, with the lowest AIC value indicative of the most appropriate model among all considered. Mod- els explored included all combinations of the following factors: area, gender, and the taxonomic levels of spe- cies, genus, or family. Models that included the interac- tion between gender and area and gender and taxonomic level were also considered. Area effects and predictions of life history values among models with the use of the lognormal and gamma distributions were similar, but models with gamma error structure resulted in the low- est standard errors for area effects; thus a gamma error structure was ultimately chosen for each model. Resultant model effects were used to compare ar- ea effects and to predict species- and gender-specific life history trait values for each area. The predicting models were then applied to two species (S. acanth- NOTE Cope: Inlraspecific life history patterns in sharks 313 Cac Ch Cob 5 - 15 - 25- ^.^■w 4 - 3 - 2 - 1 - »^ 10- 5- O' 20- 15- 10 - 5 - y:^° 0- NoA>GM 0- l>GM 0- l>NoA NP NoA G(\^ 1 CP SP SA NP NoA GIVI 1 CP SP SA NP NoA Gt^ 1 CP SP SA Gcu lox Pgi 15 ^ 25 - 15 - Age at maturity 1 1 1 NoA>GM 20 - 15 - 10 - 5 - 0 - NP>NoA 10 - 5 - 0 - NP=NoA MP NoA Glut 1 CP SP SA NP NoA GM 1 CP SP SA NP NoA Q,U 1 CP SP SA Sle Sac Smi 2b - 20- 15- T 40 - 30- \ 25 - 20 - 15 - T-.,^,^^ 10- °-^\ 20 - °,\ 10 - O.^ 5- 0- GM>CP 10 - 0 - ° NP>NoA 5 - 0 - NP>CP MP NoA GM 1 CP SP SA SIP NoA GIU1 1 CP SP SA NP NoA GIVI 1 CP SP SA Area Figure 1 Example of plots used to evaluate the pairwise life history trait (.v-axis) comparisons among areas (.v-axis) by species (indicated by the abbreviation at the top of each panel). Gender-specific age at maturity (solid triangles = females; solid circles=males) is illustrated in this example. and the qualitative outcome of the pairwise comparisons is provided within each panel. A key to the species abbreviated names is found in Table 1. Areas: NP=North Pacific; NoA = North Atlantic; GM = Gulf of Mexico; Ulndian Ocean; CP=Central Pacific; SP=South Pacific; SA=South Atlantic. ias and P. glauca) to demonstrate the calculation of life history trait values for areas with missing values. These species were chosen as examples because they are taxonomically different, are found in most of the area designations, and are represented by at least one pairwise comparison for each life history trait. Further investigation of the predictive capacity of these models to fecundity, size at maturity, and maximum size for the two species was performed by cross-validation; for each life history trait, S. acanthias or P. glauca data for one area were removed, the models were re-estimated, and predicted values for the newly missing area were calculated. Model fits to age at maturity, longevity, and male P. glauca size at maturity were not explored with cross validation because these data lacked the sufficient sample size needed to calculate the species effect once one area was removed (at least two remaining areas were needed). Results Model structure For each life history trait that contained gender as a factor, the following final GLM model with a gamma distribution and a log link was selected (i.e., had the lowest AIC value): In yA.sp.-g = In Pa + In P, + In A,,. + In P, where /3^ = area effect; (1) P., yA+gxsp = gender effect; = species effect; = gender and species interaction effect. = value of the response variable (age, litter size, or length) for each area, species, and gender, accounting for the gender-species interaction. 314 Fishery Bulletin 104(2) This model is also biologically realistic because it in- cludes the possibility that males in one species may be smaller than females and vice versa. For fecundity, the following model was chosen: ln>'A«p = lnft4 + ln/3,p. (2) An assessment of Cook's statistic for all models revealed no evidence of any highly influential data points. A subsequent analysis of residuals by species indicated, in one case, a potential departure from the assumption that both genders of all species in all areas have the same variance. Highest residuals were reported for age at maturity in S. acanthias. Whether these high residual values are truly reflective of the species or an artifact of low sample size is unknown, so I carried forth with the analysis using the above models. Area effects Despite the low and unbalanced numbers of comparisons among coarse area designations, inter-regional varia- tion and bias in sampling each life history trait, and the concomitant lack of power to resolve statistically significant relationships across all areas, a general and consistent trend emerged among the five life history traits. Intraspecifically, populations progressed from larger, longer-lived, later-to-mature populations in the northern-most latitudes to smaller, shorter-lived, and earlier-to-mature populations in the mid and southern latitudes (Fig. 2). Predicting missing life history information by area In addition to providing a comparison of the area effects on the response variables, the resultant predicting model offers a way to estimate missing life history values by area for each species and gender: ^ = e'/*.A+/*.+ft,+lnft„.sl_ (3) The factors /3 and /3,„ , in the above equation are not present in the fecundity predictions. For both S. acanthias and P. glauca (Fig. 3), the pre- dicted values mimicked the area trends of the reported values within two standard errors in all but one case (North Pacific P. glauca size at maturity was overes- timated for both genders), and provided a means to estimate values for each life history by species and gen- der for areas not yet reported. The outlying cases may indicate an area (the North Pacific) where sampling is not representative of the true population (in this case, of P. glauca) and is in need of further investigation. Cross validating models produced response variables similar to those of the full models in all but two cases (Fig. 4). In both cases (S. acanthias fecundity minus the South Atlantic, and P. glauca size at female maturity minus the North Pacific), the observed values were in opposite magnitude to that predicted. This difference could reflect either true relationships or possibly indi- cate areas that are undersampled (i.e.. North Pacific for the size at maturity for female P. glauca). Discussion Knowledge of large-scale intraspecific spatial patterning in life history traits may be important when considering the population dynamics of a species, but such large- scale patterning has seldom been formally explored. Winemiller and Rose (1992) included median and range latitude correlations in their consideration of several life history variables of North American fishes, but comparisons were only interspecific. Vila-Gispert et al. (2002) demonstrated that fishes from higher latitudes north of the equator matured latest and had the highest fecundity, whereas fishes from South America had the lowest fecundity and earliest maturation, although these comparisons were again made interspecifically. Myers et al. (2001) described the relationship between maxi- mum reproductive rate and carrying capacity among 21 stocks of Atlantic cod (Gadus morhua) in the North Atlantic using mixed effects models, but their analysis was done for only one species in a limited region. Helser and Lai (2004) also performed a similar analysis for individual growth rates in North American largemouth bass (Micropterus salmoicles) populations and found latitudinal changes in growth rate. Regarding elasmobranchs, Cortes (2000) considered trends in intraspecific reproductive traits for sharks but did not explicitly investigate the spatial patterning of those trends. Frisk et al. (2001) found regional differ- ences across five areas for the spiny dogfish using three life history measures (maximum size, and size and age at maturity), but did not specify regional patterns. The authors also performed a similar analysis with several skate species, finding no difference among areas, but they considered only interspecific patterns. Cortes and Parsons (1996) compared the demography of two Florid- ian populations of the bonnethead shark iSphyrna tibu- ro}, which included several life history measures in the life table analyses, but the small spatial resolution was inadequate to indicate large-scale spatial life history correlations within this species. Lombardi-Carlson et al. (2003) extended the bonnethead shark investigation to a larger portion of the eastern Gulf of Mexico and found latitudinal variation in maturity and size, but again the scale of this study was relatively small. Ad- ditional small scale studies on intraspecific geographic variation in reproductive parameters of sharks have been presented by Horie and Tanaka (2002), Taniuchi et al. (1993), and Yamaguchi et al. (2000). The results of the present study, specifically aimed at sharks as an example, indicate an emerging pattern for intraspecific life history variation, not unlike previously recognized interspecific patterns. Generally, there is a distinct difference in life history traits among areas — a pattern potentially useful when considering region-spe- cific population dynamics. Across taxonomic designa- tions, populations in the northern latitudes tended to be NOTE Cope Intraspecific life history patterns in sharks 315 Age at maturity Longevity -0 2- -0.6- -1.0 -0.2 -0.6 -1.0 NP NoA GM I Fecundity CP NP NoA GM I CP SP SA D ?, -0.1 -0.3 -0.5 000 -0.05 -0.15- Size at maturity NP NoA GM I CP SP SA NP NoA GM I CP SP SA E IVlaximum size -0.05 -0.15 NP NoA GM CP SP SA Figure 2 Bar plots of the area effects for each life history trait. Areas: NP=North Pacific; NoA=North Atlantic; GM = Gulf of Mexico; Ulndian Ocean; CP=Central Pacific; SP=South Pacific; SA = South Atlantic. Vertical bars represent one standard deviation. The North Pacific area is the reference area, and all other areas are standardized to it. To interpret the relationships, lower effect values in relation to effect values from another area indicate the life history trait value would also be lower. larger, to mature later in life, to have longer life spans, and to have greater fecundity compared to conspecifics in the central and southern latitudes. Populations in the North Pacific, in particular, seem to demonstrate dramatic departures in life history measures compared to conspecifics in other areas. Therefore, instead of as- suming life history information from one region should be applied to another region, the trends and predictive methods offered in the present study provide a means to extrapolate life history traits of cosmopolitan species in specific areas when only information from other areas is available; this method may prove useful for develop- ing informative priors for Bayesian analyses (Punt and Hilborn, 1997). Caveats to these results include area- specific biases (i.e., certain-size individuals susceptible to capture) and errors in sampling programs and migra- tory patterns of specific species (e.g., individuals may be found in multiple areas during different parts of their life history). Thus a proper knowledge of the biology of the species is recommended before interpreting the interpolated life history values. Other factors, such as fishing pressure, may influence regional differences in life history traits, challenging the interpretation of such patterns. Truncation of size and age classes, and reduction in age at maturity are recognized byproducts of heavy fishing (Longhurst, 1998a; Rochet, 2000). Although all the populations used in this study are and have been fished — some more intensely than others — this study assumes there is no consistent pattern to such exploitation in shark 316 Fishery Bulletin 104(2) A Squalus acanthias: Females 100- 80- 60- 40 ■ 20- 0- 140- Maturity Longevity Fecundity NP NoA GM CP SP SA NP NoA GM I CP SP SA Area B Pnonace glauca: Females 80 40 0) 20 — ♦— fvlatunty - •- Longevity * Fecundity 4 i- .-t^ NP NoA GM I CP SP SA 400 n 300 - 200 £ 100 - - fvlaximum ■' Maturity 4 f- ^-^--^- ;-- — I 1 1 1 r- NoA GM I CP SP Area SA 100 80 •t 60- $ 40 20 0-' Squalus acanthias: Males — o- Maturity - D- Longevity 1 — 1 1 — NP NoA GM S 120- c 100- CP SP SA — *- Maximum - D- Maturity NP NoA GM 1 CP SP SA Area Pnonace glauca. Males 40 n 30 — o— rvlaturity - □- Longevity 400 300 100 I \ 1 1 1 1 1 NP NoA GM I CP SP SA Maximum " °" Maturity ■K- 1 NP 1 NoA 1 GM 1 1 Area 1 CP 1 SP SA Figure 3 Example of predictive plots based on the results of the generalized linear models for each life history trait by gender for two species: (A) Squalus acanthias and ( B ) Prionace glauca. Lines represent predicted values and symbols represent values from the literature. Vertical lines represent two standard errors. NOTE Cope Intraspecific life history patterns in sharks 317 15 10 S. acanthias Fecundity NP NoA GM I CP SP SA 150 E S 100 50 Female size 0-' -r NP NoA GM I CP SP SA Male size 100 80 60 -\ 40 20- 0 NP NoA GM I CP SP SA 100 n 80 60 40 20 P. glauca Fecundity 0 400 300 200 100 I 1 1 1 1 1 1 NP NoA GM I CP SP SA Female size NP NoA GM CP SP SA 400 300 200 100 0 - Male size All - - NP ■ • ■ NoA - • SP NP NoA GM I CP SP SA Area Figure 4 Plots of model predictions derived by cross validation for S. acanthias and P. glauca. Lines representing predicted value are distinguished as follows: "AH"- all areas are included in the prediction; Lines denoted by area are predictions made without that area. Literature values are indicated as follows: triangles — fecundity, circles — size at maturity, squares — maximum size. Size at maturity for male P. glauca was not included because of insufficient data. Areas: NP=North Pacific; NoA = North Atlan- tic; GM = Gulf of Mexico; I = Indian Ocean; CP=Central Pacific; SP=South Pacific; SA=South Atlantic. populations that would explain the results. Specifically, there is no reason to believe that species in the south- ern hemisphere have been more heavily fished than conspecifics in the northern hemisphere. The results offered in the present study are based on small sample sizes in most areas, but hopefully they will bring attention to the usefulness of collecting spa- tially varying intraspecific information with the idea of constructing more robust models. Most investigations describing the patterns of shark life history traits suffer from insufficient biological resolution (either temporally or spatially) of the very parameters and subsequent relationships they attempt to explain (Smith et al., 1998; Cortes, 2000). Limited biological knowledge and 318 Fishery Bulletin 104(2) subsequent high uncertainty in the estimation of vital rates of many marine species, including elasmobranchs, is testimony to the fact that the accumulation of life history information should be a priority to biologists, fisheries scientists, and resource managers. Results from the models presented here could be used to hy- pothesize life history values in areas currently lacking information and thus be tested with further sampling in those areas. It is also hoped that the approach of- fered here may indicate areas where sampling may not be sufficient, as denoted by departures from the general model trend. Targeted sampling in that area would help resolve whether the departure is from a true area effect or species effect. As more data is gathered, it will be possible to explore other factors — such as temperature and guilds (e.g., coastal versus oceanic) — in the model structure. Once steps are made to further resolve the species and area effects, one may start to ask questions re- garding the cause of particular area effects. Poten- tial mechanisms of true coarse-scale gradation of life history traits may be contained within the general- ized characteristics of oceanic zoogeographic realms (Longhurst, 1998b), although a slightly less abstract mechanism could be found in the physical forcing events that characterize regions in the northern and southern hemisphere. Although both hemispheres demonstrate similar large-scale current and wind patterns, physi- cal forcing events tend to be stronger in the northern hemisphere (Trenberth et al., 19981. Because this study offers coarse area designations to intraspecific life his- tory variation, it is most likely a product of some macro- scale characteristics of each region. Attention should therefore be directed towards large-scale characteris- tics of each region to explain these patterns, although small-scale dynamics are important for understanding each population's specific response to local environmen- tal conditions (i.e., countergradient variation in growth rates [Conover, 1990; Conover and Present, 1990]). It is becoming increasingly important to be able to assess fish stocks with minimal data. By combining genetic data revealing differing levels of intraspecific population substructure with the increasing number of studies demonstrating localized adaptations and plas- ticity in population parameters, it is apparent that intraspecific spatial differences must be considered in species management (Aviso, 2000; Roff, 2002). Although the predictive power of this study may currently be weak because of low sample sizes, it offers a method to quantify potential spatial patterning in intraspecific life history traits that may allow responsible manage- ment of regionally data-poor species, and it may help frame future sampling protocols and studies of spatial patterns in life history traits. Acknowledgments I am grateful to Andre Punt, Joe Bizzarro, Gavin Fay, and Arni Magnusson for their many important comments regarding model structure and data use. I also thank Kristin Benshoof, Marilyn Cope, and three anonymous reviewers who contributed insightful comments that improved the presentation and clarity of this note. Literature cited Adams, P. B. 1980. Life history patterns in marine fishes and tiieir consequences for fisheries management. Fish. Bull. 78: 1-12. Avise, J. C. 2000. Phylogeography: the history and formation of spe- cies, 447 p. Harvard Univ. Press. Cambridge, MA. Botsford, L. W., J. C. Castilla, and C. H. Peterson. 1997. The management of fisheries and marine eco- systems. Science 277:509-515. Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a prac- tical information-theoretic approach, 488 p. Springer- Verlag, New York, NY. Conover, D. O. 1990. The relation between capacity for growth and length of growing season: evidence for and implica- tions of countergradient variation. Trans. Am. Fish. Soc. 119:416-430. Conover, D. O., and T. M. C. Present. 1990. Countergradient variation in growth rate: com- pensation for length of the growing season among Atlantic silversides from different latitudes. Oecologia 83:316-324. Cortes, E. 2000. Life-history patterns and correlations in sharks. Rev. Fish. 8:299-344. 2002. Incorporating uncertainty into demographic modeling: application to shark populations and their conservation. Cons. Biol. 16:048-1062. Cortes, E., and G. R. Parsons. 1996. Comparative demography of two populations of the bonnethead shark iSphyrna tiburo). Can. J. Fish. Aquat. Sci. 53:709-718. Frisk, M. G., T. J. Miller, and M. J. Fogarty. 2001. Estimation and analysis of biological parameters in elasmobranch fishes: a comparative life history study. Can. J. Fish. Aquat. Sci. 58:969-981. Garcia, S. M., and C. Newton. 1997. Current situations, trends, and prospects in world capture fisheries. Am. Fish. Soc. Symp. 20:3-27. Gunderson, D. R. 1997. Trade-off between reproductive effort and adult survival in oviparous and viviparous fishes. Can. J. Fish. Aquat. Sci. 54:990-998. Helser. T. E., and H.-L. Lai. 2004. A Bayesian hierarchical meta-analysis of fish growth: with an example for North American large- mouth bass, Micropterus salmoides. Ecol. Mod. 178:399-416. Horie, T., and S. Tanaka. 2002. Geographical variation of maturity size of the cloudy shark, Scyliorhinus torazame, in Japan. J. School of Mar. Sci. Tokai Uni. 53:111-124. Jennings, S., M. J. Kaiser, and J. D. Reynolds. 2001. Marine fisheries ecology, 417 p. Blackwell Sci- ence Ltd , Oxford, UK. NOTE Cope Intraspecific life history patterns in sharks 319 Jennings, S., J. D. Reynolds, and S. C. Mills. 1998. Life history correlates of responses to fisheries exploitation. Proc. R. Soc. Lon. B:333-339. Jennings, S.. J. D. Reynolds, and N. V. C. Polunin. 1999. Predicting the vulnerability of tropical reef fishes to exploitation with phylogenies and life histories. Con. Biol. 13(61:1466-1475. Jones, B. C and G. H. Geen. 1977. Reproduction and embryonic development of spiny dogfish iSqualus acanthias) in the Strait of Georgia, British Columbia. J. Fish. Res. Board Can. 34:1286-1292. Lombardi-Carlson, L. A., E. Cortes, G. R. Parsons, and C. A. Manire. 2003. Latitudinal variation in life-history traits of bon- nethead sharks, Sphynm tiburo. (Carcharhiniformes: Sphyrnidae). Mar. Freshw. Res. 54:875-883. Longhurst, A. 1998a. Cod: perhaps if we all stood back a bit? Fish. Res. 38:101-108. 1998b. Ecological geography of the sea, 398 p. Academic Press, London. Musick, J. A., G. Burgess, G. Cailliet, M. Camhi, and S. Fordham. 2000. Management of sharks and their relatives (Elas- mobranchiii. Fisheries 25:9-13. Myers, R. A.. B. R. MacKenzie. K. G. Bowen. and N. J. Barrowman. 2001. What is the carrying capacity offish in the ocean? A meta-analysis of population dynamics of North Atlantic cod. Can. J. Fish. Aquat. Sci. 58:1464-1476. Punt, A. E., and R. Hilborn. 1997. Fisheries stock assessment and decision analy- sis: the Bayesian approach. Rev. Fish Biol, and Fish. 7:35-63. Rago, P. J.. K. A. Sosbee, J. K. T. Brodziak, S. A. Murawski, and E. D. Anderson. 1998. Implications of recent increases in catches on the dynamics of Northwest Atlantic spiny dogfish iSqualus acanthias). Fish. Res. 39:165-181. Reynolds, J. D., S. Jennings, and N. K. Dulvy. 2001. Life histories of fishes and population response to exploitation. In Conservation of exploited species (J. D. Reynolds, G. M. Mace, K. H. Redford, and J. G. Robinson, eds.), p. 147-168. Cambridge Univ. Press, Cambridge, UK. Rochet, M.-J. 2000. A comparative approach to life-history strategies and tactics among four orders of teleost fish. ICES J. Mar. Sci. 57:228-239. Roff, D. A. 2002. Life history evolution. 527p. Sinauer Associates, Inc, Sunderland, MA. Saunders, M. W., and G. A. McFarlane. 1993. Age and length at maturity of the female spiny dogfish, Squalus acanthias, in the Strait of Georgia, British Columbia, Canada. Environ. Biol. Fishes. 38:49-57. Smith, S. E., D. W. Au, and C. Show. 1998. Intrinsic rebound potentials of 26 species of Pacific sharks. Mar. Freshw. Res. 49:663-678. Stearns, S. C. 1992. The evolution of life histories. Oxford, Oxford Univ. Press, Oxford, UK. Stobutzki, I., M. Miller, and D. Brewer. 2001. Sustainability of fishery bycatch: a process for assessing highly diverse and numerous bycatch. En- viron. Cons. 28 :167-181. Taniuchi, T., H. Tachikawa, M. Shimizu. 1993. Geographical variation in reproductive parameters of shortspine spurdog in the North Pacific. Nippon Suisan Gakkaishi 59:45-51. Trenberth, K. E., G. W.Branstator, D. Karoly, A. Kumar. N-C. Lau, and C. Ropelewski. 1998. Progress during TOGA in understanding and modeling global teleconnections associated with tropi- cal sea surface temperatures. J. Geophy. Res. 103: 14,291-14,324. Venables, W. N., and B. D. Ripley. 2002. Modern applied statistics with S. fourth addition, 495 p. Springer-Verlag, New York, NY. Vila-Gispert, A., R. Moreno-Amich, and E. Garcia-Berthou. 2002. Gradients of life-history variation: an intercon- tinental comparison of fishes. Revs Fish Biol. Fish. 12:417-427. Winemiller, K. O., and K. A. Rose. 1992. Patterns of life-history diversification in North America fishes: Implications for population regu- lation. Can. J. Fish. Aquat. Sci. 49:2196-2218. Yamaguchi, A., T. Taniuchi, and M. Shimizu. 2000. Geographic variation in reproductive parameters of the starspotted dogfish Mustelu manazo from five locations in Japan and Taiwan. Environ. Biol. Fish. 57:221-233. 320 Fishery Bulletin 104(2) Appendix Life history traits and area assignments for species used in tlie analyses Sizes are total len Jth (cm). Areas: NP= North Pacific; NoA=North Atlantic; GM=Ct alf of Mexico; I = Indian Ocean; CP= = Central Pacific; SP=South Pacific; SA=South Atlantic. Maximum size Size at maturity Age at maturity Longevity Mean Species Area F M F M F M fecundity F M Carcharinus acronotus GM 130 122 110 103 3 2.5 5 Carcharinus acronotus NoA 154 164 120 110 4 3 5 Carcharinus amblyrhynchos CP 190 185 137 132.5 5 Carcharinus amblyrhynchos I 178 168 135 130 3 Carcharinus falcifnrmis GM 308 314 235 217.5 11 Carcharin us falciformis NoA 330 234 218 10 Carcha ri n u s fa lei form is I 283 244 248 11 Carcharinus falciformis SP 250 225 200 212.5 7 Carcharinus leucas GM 285 225 215 8 Carcharinus leucas I 300 239.5 239.5 9 Carcharinus limbatus GM 191 175 156 133 7 4.5 4.9 10 9 Carcharinus limbatus NoA 202 189 156 143.5 4 Carcharinus limbatus I 247 246 208.5 201.5 7 6 6 11 10 Carcharinus longimanus NoA 260 175 12.5 Carcharinus longimanus SA 250 235 185 185 17 14 Carcharinus longimanus I 270 245 185 198 Carcharinus longimanus SP 270 251 200 6 Carcharinus longimanus CP 272 240 182 182 7 11 11 Carcharinus obscurus NoA 371 360 284 279 21 19 11 39 39 Carcharinus obscurus I 389 324 300 280 24 20.5 9.9 34 Carcharinus plumbeus CP 190 172 144 131 5.5 Carcharin us plumbeus NoA 234 226 183 180 15 15 9 32 40 Carcharinus plumbeus I 199 190.5 169 167 8 Galeocerdo cuvier NoA 450 317.5 310 10 10 55 16 15 Galeocerdo cuvier GM 450 8 7 11 9 Galeocerdo cuvier I 410 370 340 290 35 Galeocerdo cuvier SP 428 350 287 31 Prionace glauca NP 310 150 145 6 5 60 24 27 Prionace glauca NoA 327 340 221 215 6 6 13 16 Prionace glauca SP 316 312 218 32 Prionace glauca I 321.5 214 34 Rhizoprionodon taylori SP 78 69 54 56 4.5 Rhizoprionodon taylori I 66 55 45 43 5 Sphyrna lewini GM 310 300 250 180 15 10 17 12 Sphyrna lewini NP 324 305 210 198 4.1 3.8 26 14 11 Sphyrna lewini I 346 301 200 150 16.5 Galeorhinus galeus NP 195 185 180 175 Galeorhinus galeus SP 174 175 140 135 15 17 28.4 53 41 Galeorhinus galeus SA 155 148 128 117 17.5 13 23 36 36 Isurus oxyrinchus NP 351 280 16 34 Isurus oxyrinchus SP 340 270 280 195 9 Isurus oxyrinchus NoA 375 298 280 14 13.5 23 9 Isurus oxyrinchus I 333 271 266 199.5 11.5 Alopias superciliosus NoA 444 410 341 276 2 Alopias superciliosus NP 422 357 341 288 13 10 2 20 19 Squalus acanthias NP 130 103 94 78.5 35.5 19 7.1 81 50 Squalus acanthias NA 110 90 80 59.5 12.1 6 6.6 40 35 Squalus acanthias SA 95.5 78 70 63 7 Squalus acanthias SP 111 90 71.5 58 5 Squalus mitsukurii NP 114 94 97.5 70 22 12 8.8 Squalus mitsukurii I 95 81 69 60 6.4 Squalus mitsukurii CP 91 82 69 48 15 4 3.6 27 18 Squalus mitsukurii SP 104 102 85 3.5 16 14 321 Fishery Bulletin Guidelines for authors Content of manuscripts Contributions published in Fishery Bulletin describe original research in marine fishery science, fishery engi- neering and economics, as well as the areas of marine environmental and ecological sciences (including model- ing I. Although all contributions are subject to peer review, responsibility for the contents of papers rests upon the authors and not upon the editor or publisher. Submission of an article implies that the article is original and is not being considered for publication elsewhere. Manuscripts must be written in English. Authors whose native lan- guage is not English are strongly advised to have their manuscripts checked by English-speaking colleagues prior to submission. Articles may range from relatively short contributions (10-15 typed, double-spaced pages, tables and figures not included) to extensive contributions (20-30 typed pages). Notes are reports of 5 to 10 pages without an abstract and describe methods or results not supported by a large body of data. Manuscript preparation Title page should include authors' full names and mailing addresses and the senior author's telephone, fax number, and e-mail address, and a list of key words to describe the contents of the manuscript. Abstract should be limited to 150 words (one-half typed page), state the main scope of the research, and emphasize the author's conclusions and relevant findings. Do not review the methods of the study or list the contents of the paper. Because abstracts are cir- culated by abstracting agencies, it is important that they represent the research clearly and concisely. Text must be typed in 12 point Times New Roman font throughout. A brief introduction should convey the broad significance of the paper; the remainder of the paper should be divided into the following sections: Materials and methods, Results, Discussion (or Conclusions), and Acknowl- edgments. 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, abbre- viations, and technical terms should be written out in full the first time they are used. Include FAO common names for species in the list of keywords and in the introduction. Regional common names may be used throughout the rest of the text if they are different from FAO common names which can be found at http://www.fishbase.org/search. html. Follow the U.S. Government Printing Office Style Manual (1984 ed.) and the CBE Style Manual (6th ed.) for editorial style; for fish nomenclature follow the most cur- rent issue of the American Fisheries Society's Common and Scientific Names of Fishes from the United States and Canada. Dates should be written as follows: 11 Novem- ber 2000. Measurements should be expressed in metric units, e.g., 58 metric tons (t); if other units of measure- ment are used, please make this fact explicit to the reader. Write out the numbers zero through nine unless they form part of measurement units (e.g., nine fish but 9 mm). Text footnotes should be inserted in 9-point font at the bottom of the page that displays the first citation of the footnote. Footnotes should be formatted in the same manner as citations. Footnote all personal communica- tions, unpublished data, and unpublished manuscripts with full address of the communicator or author, or, as in the case of unpublished data, where the data are on file. Authors are advised to avoid references to nonstandard (gray) literature (such as internal, project, processed, or administrative reports, ICES Council Minutes, IWC Minutes or Working Papers, any "research" or "work- ing" documents, laboratory or contract reports. Man- agement Council reports, and manuscripts in review) wherever possible. If these references are used, present them as footnotes and list whether they are available at NTIS (National Technical Information Service) or at some other public depository. Cite all software and special equipment or chemical solutions used in the study, not in a footnote but within parentheses in the text (e.g., SAS, vers. 6.03, SAS Inst., Inc., Cary, NO. Literature cited comprises published works and those accepted for publication in peer-reviewed litera- ture (in press). Follow the name and year system for cita- tion format in the "Literature cited" section. If there is a sequence of citations in the text, list chronologically: (Smith, 1932; Green, 1947; Smith and Jones, 1985). Abbre- viations of serials should conform to abbreviations given in the Serial Sources for the BIOSIS Previews Database. Authors are responsible for the accuracy and complete- ness of all citations. Literature citation format: Author (last name, followed by first-name initials). Year. Title of report or manuscript. Abbreviated title of the series to which it belongs. Always include number of pages. If the authorship for a sequence of citations is identical, list works chronologically. Tables and figures — general formiat • Zeros should precede all decimal points for values less than one. • Sample size, n, should be italicized. • Capitalize the first letter of the first word in all labels within figures. • Do not use overly large font sizes in maps and for units of measurements along axes in figures. • Do not use bold fonts or bold lines in figures. • Do not place outline rules around graphs. • Do not use horizontal lines in graphs to indicate mea- surement units on axes. • Use a comma in numbers of five digits or more (e.g. 13,000 but 3000). • Maps should have a North arrow (or compass sign) and degrees latitude-longitude (e.g., 170 E) 322 Fishery Bulletin 104(2) Tables are often overused in scientific papers; it is ssldom necessary or even desirable to present all the data associated with a study. Tables should not be excessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All unusual symbols must be explained in the table legend. Other incidental comments may be footnoted with italic footnote markers. Use asterisks to indicate probability in statistical data. Do not type table legends on a separate page; place them on the same page as the table data. Do not submit tables in photo mode. Figures include line illustrations, photographs (or slides), and computer-generated graphs and must be cited in numerical order in the text. Graphics will aid in the comprehension of the text, but they should be limited to presenting patterns rather than raw data. Figures are costly to print and should be limited to six. Figures must be labeled with author's name and number of figure. Avoid placing labels vertically (except on y-axis). Figure legends should explain all symbols and abbreviations and should be double-spaced on a separate page at the end of the manuscript. Please note that we do not print graphs in color. FAILURE TO FOLLOW THESE GUIDELINES WILL DELAY PUBLICATION OF A MANUSCRIPT Copyright law does not apply to Fishery Bulletin, which falls within the public domain. However, if an author reproduces any part of an article from Fishery Bulletin in his or her work, reference to source is consid- ered correct form (e.g.. Source: Fish. Bull 97:105). Reprints are available free of charge to the senior author (50 copies) and to his or her laboratory (50 copies). Submission The Scientific Editorial Office encourages authors to submit their manuscripts as a single PDF (pre- ferred), or Word (zipped) document by e-mail to Fish- ery.Bulletin@noaa.gov. Please use the subject heading "Fishery Bulletin manuscript submission." Do not send encrypted files. For further details on electronic submis- sion, please contact the Scientific Editorial Office directly (see address below). Or you may send your manuscript on compact disc in one of the above formats along with four printed copies (one original plus three copies [stapled]) to the Scientific Editor, at the address shown below. Dr. Adam Moles Scientific Editor, Fishery Bulletin 11305 Glacier Hwy Juneau, AK 99801-8626 Once the manuscript has been accepted for publication, you will be asked to submit a final software copy of your manuscript. When requested, the text and tables should be submitted in Word or Word Rich Text Format. Figures should be sent as PDF files, Windows metafiles, tiff files, or as EPS files. Send a copy of figures in original software if conversion to any of these formats yields a degraded version. U.S. Department of Commerce Volume 104 Number 3 July 2006 Fishery Bulletin U.S. Department of Commerce Cartos M. Gutierrez Secretary National Oceanic and Atmospheric Administration Vice Admiral Conrad C. Lautenbacher Jr., USN (ret.) Under Secretary for Oceans and Atmosphere National Marine Fisheries Service William T. Hogarth Assistant Administrator for Fisheries Scientific Editor Adam Moles, Ph.D. Technical Editor Elizabeth Calvert National Marine Fisheries Service, NOAA 11305 Glacier Highway Juneau, Alaska 99801-8626 .*t<»"^co. ^ATES 0* Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE, BIN C15700 Seattle, Washington 98115-0070 The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN C 15700, Seattle, WA 98115-0070. Periodicals postage is paid at Seattle, WA. POSTMASTER: Send address changes for subscriptions to Fish- ery Bulletin, Superintendent of Docu- ments, 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 Department. Use of funds for printing of this periodical has been approved by the Director of the Office of Management and Budget. For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $55.00 domestic and $68.75 foreign. Cost per single issue: $28.00 domestic and $35.00 foreign. See back for order form. Editorial Committee Harlyn O. Halvorson, Ph.D. Ronald W. Hardy, Ph.D. Richard D. Methot, Ph.D. Theodore W. Pietsch, Ph.D. Joseph E. Powers, Ph.D. Harald Rosenthal, Ph.D. Fredric M. Serchuk, Ph.D. George Watters, Ph.D. University of Massachusetts, Boston University of Idaho, Hagerman National Marine Fisheries Service University of Washington, Seattle National Marine Fisheries Service Universitat Kiel, Germany National Marine Fisheries Service National Marine Fisheries Service Fishery Bulletin web site: www.fishbull.noaa.gov The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. It began as the Bulletin of the United States Fish Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. U.S. Department of Commerce Seattle, Washington Volume 104 Number 3 July 2006 Fishery Bulletin MBLWHOI Library JUN ') 6 ?no6 WOODS HOLE sachusetts 02543 Contents Articles Companion articles 323-331 Dowd, W. Wesley, Richard W. Brill, Peter G. Bushnell, and John A. Musick Standard and routine metabolic rates of luvenile sandbar sharks (Carcharhinus plumbeus), including the effects of body mass and acute temperature change The conclusions and opinions ex- pressed in Fishery Bulletin are solely those of the authors and do not represent the official position of the National Marine Fisher-ies Ser- vice (NOAA) or any other agency or institution. The National Marine Fisheries Service (NMFS) does not approve, recommend, or endorse any proprie- taiy 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, recom- mends, or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. 332-342 Dowd, W. Wesley, Richard W. Brill, Peter G. Bushnell, and John A. Musick Estimating consumption rates of juvenile sandbar sharks (Carcharhinus plumbeus) in Chesapeake Bay, Virginia, using a bioenergetics model 343-349 Zollett, Erika A., and Andrew J. Read Depredation of catch by bottlenose dolphins (Tursiops truncatus) in the Florida king mackerel (Scomberomorus cavalla) troll fishery 350-359 Abookire, Alisa A. Reproductive biology, spawning season, and growth of female rex sole (Glypiocephalus zachirus) in the Gulf of Alaska 360—375 Abesamis, Rene A., Angel C. Alcala, and Garry R. Russ How much does the fishery at Apo Island benefit from spillover of adult fish from the adiacent marine reserve? 376—382 Li, Zhuozhuo, Mary M. Nishimoto, Milton S. Love, and Anthony J. Gharrett Comparing the identification of southern California juvenile rockfishes (genus Sebastes spp.) by restriction site analysis of the mitochondrial ND3/ND4 region and by morphological characteristics Fishery Bulletin 104(3) Companion articles 383—390 Love, Milton S., Donna M. Schroeder, William Lenarz, Alec MacCall, Ann Scarborough Bull, and Lyman Thorsteinson Potential use of offshore marine structures in rebuilding an overfished rockfish species, bocaccio iSebasies paucispinis) 391—400 Emery, Brian M., Libe Washburn, Milton S. Love, Mary M. Nishimoto, and J. Carter Ohimann Do oil and gas platforms off California reduce recruitment of bocaccio (Sebastes paucispmis) to natural habitat? An analysis based on trajectories derived from high-frequency radar 401—414 Leis, Jeffrey M., Amanda C. Hay, Domine L. Clark, l-Shiung Chen, and Kwang-Tsao Shao Behavioral ontogeny in larvae and early luveniles of the giant trevally (.Caranx ignobilis) (Pisces: Carangidae) 415—427 Reddin, David G., Peter Downton, and Kevin D. Friedland Diurnal and nocturnal temperatures for Atlantic salmon postsmolts iSalmo salar L ) during their early marine life 428-433 He, Xi, Marc Mangel, and Alex MacCall A prior for steepness in stock-recruitment relationships, based on an evolutionary persistence principle 434—444 Kerstetter, David W., and John E. Graves Survival of white marlin (Tetrapturus albidus) released from commercial pelagic longline gear in the western North Atlantic 445—455 Gudmundson, Carolyn J., Tonya K. Zeppelin, and Rolf R. Ream Application of two methods for determining diet of northern fur seals (Ca/lorhinus ursinus) 456-464 Smith, David R., Michael J. Millard, and Sheila Eyier Abundance of adult horseshoe crabs {Limulus polyphemus) in Delaware Bay estimated from a bay-wide mark-recapture study Notes 465—467 Clark, William G, and Stephen M. Kaimmer Estimates of commercial longline selectivity for Pacific halibut (Hippoglossus stenolepis) from multiple marking experiments 468—475 Hattori, Tsutomu, Akira Nishlmura, Yoji Narimatsu, and Daiji Kitagawa Hatching date, nursery grounds, and early growth of luvenile walleye pollock iTheragra chalcogramma) off northern Japan 476—481 De Forest, Lisa G., and Morgan S. Busby Development of larval and early luvenile penpoint gunnel (Apodichthys flavidus) (family: Pholidae) 482—485 Gandini, Patricia, and Esteban Frere Spatial and temporal patterns in the bycatch of seabirds in the Argentinian longline fishery 486 Announcement of Best Papers Awards 487 Guidelines for authors 323 Abstract — Standard and routine metabolic rates (SMRs and RMRs, respectively) of juvenile sandbar sharks iCarcharhinus plumbeus) were measured over a range of body sizes (n=34) and temperatures normally associated with western Atlantic coastal nursery areas. The mean SMR Qii, (increase in metabolic rate with temperature! was 2.9 ±0.2. Heart rate decreased with increasing body mass but increased with temperature at a Q,Q of 1.8-2.2. Self-paired measures of SMR and RMR were obtained for 15 individuals. Routine metabolic rate averaged 1.8 ±0.1 times the SMR and was not correlated with body mass. Assuming the maximum metabolic rate of sandbar sharks is 1.8-2.75 times the SMR (as is observed in other elasmobranch species), sandbar sharks are using between 34V( and 100*^ of their metabolic scope just to sustain their routine continuous activity. This limitation may help to explain their slow individual and population growth rates, as well as the slow recoveries from overfishing of many shark stocks worldwide. Standard and routine metabolic rates of juvenile sandbar sharks iCarcharhinus plumbeus), including the effects of body mass and acute temperature change* W. Wesley Dowd' Richard W. Brill^ Peter G. BushnelP John A. Musick' ' Department ol Fisheries Science Virginia Institute ol Marine Science 1208 Create Road, PO, Box 1346 Gloucester Point, Virginia 23062 Present address (for W W. Dowd): Department ol Wildlife, Fish and Conservation Biology University of California Davis One Shields Avenue, Davis, California 95616 E-mail address (for W W Dowd) wwdowd gucdavisedu ^ Virignia Cooperative Marine Education and Research Program Virginia Institute of Marine Science 1208 Create Road, PO, Box 1346 Gloucester Point, Virginia 23062 ^ Department of Biological Sciences Indiana University South Bend 1700 Mishawaka Avenue South Bend, Indiana 46634 Manuscript submitted 21 August 2004 to the Scientific Editor's Office. Manuscript approved for publication 26 September 2005 by the Scientific Editor. Fish. Bull. 104:.323-331 (2006). Shark populations continue to suffer from overfishing throughout the North- west Atlantic and worldwide (Baum et al.. 2003). The sandbar shark iCarcha- rhinus plumbeus) can serve as a model for overfished coastal shark species, many of which share ecological and ecophysiological characteristics. After the rapid expansion of the Atlantic coastal commercial shark fishery in the mid-1980s, sandbar shark num- bers declined 66% by 1991 (Musick et al., 1993; Sminkey and Musick, 1995). Like many of their K-selected relatives, sandbar sharks grow slowly and mature after a minimum of 13-15 years (Casey and Natanson, 1992; Sminkey and Musick, 1995). Demo- graphic models of these species predict very slow rates of population increase even in the absence of fishing pressure, and elasticity analyses of these models demonstrate that the juvenile stage is the most critical life history stage (Sminkey and Musick, 1996; Cortes, 1999; Brewster-Geisz and Miller, 2000). It is necessary, therefore, to understand the actual and potential contributions of various juvenile nurs- ery areas to recovery of the Northwest Atlantic sandbar shark population and to recovery of other coastal shark stocks (Branstetter, 1990). Bioenergetics models can be used to assess the impacts and requirements of juvenile sharks as apex predators. Metabolic rate is the largest and most variable component of the energy bud- get for active fish species, and it is critical that it be determined accu- rately in order to construct realistic bioenergetics models (Ney, 1993). Systematic metabolic rate data for elasmobranchs are only rarely avail- able, and previous models of sandbar shark bioenergetics have relied upon metabolic rate data from unrelated species (Medved et al., 1988; Stillwell and Kohler, 1993). ' Contribution number 2720 from Depart- ment of Fisheries Science, Virginia Institute of Marine Science, College of William and Mary, 1208 Create Road, Gloucester Point, VA 23062. 324 Fishery Bulletin 104(3) The lower Chesapeake Bay, Mid-Atlantic Bight, and adjacent coastal lagoon systems serve as the primary summer nurseries for sandbar sharks in the Northwest Atlantic (Musick et al., 1993). Juvenile sandbar sharks return for four to ten years to these nursery grounds, where they enjoy the benefits of generally high food availability and limited exposure to large shark preda- tors (Musick and Colvocoresses, 1986; Grubbs et al., in press). Juvenile sandbar sharks in the nurseries are exposed to seasonal water temperature variations, as well as shorter time-scale fluctuations associated with their vertical movements and day to day variation. The minimum seasonal temperatures (=15°C) occur in mid or late May, whereas the maximum temperatures (=28°C) are reached in surface waters in July and Au- gust (Merson and Pratt, 2001). Throughout the day, sandbar sharks perform frequent vertical excursions and thus experience surface and bottom water tem- peratures that can differ by up to 5°C (Grubbs, 2001). Similarly, in Virginia's Eastern Shore lagoons, juvenile sandbar sharks venture onto broad, warm tidal flats at high tide and return to deeper, cooler channels as the tide recedes (ConrathM. To improve bioenergetics models and to define criti- cal habitat and the current suitability of nursery areas more accurately, standard (SMR) and routine metabolic rates (RMR) of juvenile sandbar sharks were measured over a relevant range of body masses (=1 to 10 kg) and temperatures (18-28°C) (Merson and Pratt, 2001). This is the first direct measurement of SMR, and the first comparison of paired SMR and RMR, in a continuously active carcharhiniform species. Materials and methods All experiments were conducted at the Virginia Insti- tute of Marine Science Eastern Shore Laboratory from June through September 2002. Sandbar sharks (57-124 cm total length; 1.025-10.355 kg) were captured by using hook and line from the surrounding tidal lagoon system and maintained in shoreside tanks (temperature 22-29°C, salinity 34-36%^). Individuals were fasted for at least 48 hours prior to use in an experiment to reduce any confounding effects of specific dynamic action (Medved, 1985). Standard metabolic rates Because sandbar sharks are continuously active obligate ram ventilators, SMR measurements were obtained from chemically immobilized and artificially ventilated animals maintained in flow-through, sealed box res- pirometers (Brill, 1979, 1987; Leonard et al., 1999). Respirometers were constructed of 0.85 cm thick acrylic, sized to accommodate the fish being studied, and cov- Conrath, C. 2004. Personal commun. Virginia Institute of Marine Science. 1208 Create Road, Gloucester Point, VA 23062. ered with black plastic to minimize visual disturbance. Aerated and filtered seawater from a constant pressure head tank passed through the mouth and over the gills of the sharks, was mixed in the chamber by a small recirculating pump, and exited the respirometer by a hose mounted at the top. Water leaving the respirom- eter was collected, re-aerated, and mixed with a small amount of fresh filtered seawater to help maintain a constant temperature. Turnover rate for the system was 20-30%/hour (Steffensen, 1989). Sharks were netted, injected with 0.4-1.8 mg/kg of the neuromuscular blocking agent pancuronium bromide through the caudal vein, and returned to the holding tank until they were unable to swim (typically 1-2 min). They were then placed supine on a moist towel and ven- tilated with aerated seawater while electrocardiogram (EKG) wire leads were inserted subcutaneously over the pectoral girdle to monitor heart rate. Individuals were also given an intramuscular injection (0.2-1.2 mg/kg) of steroid anesthetic Saffan" (alphaxalone and alph- adolone; Pitman-Moore, Uxbridge, UK) (Oswald, 1978). Two 20-gauge hypodermic needles were inserted into the dorsal musculature and used to administer supple- mentary doses of pancuronium bromide and Saffan® whenever any slight tail movement was observed. The partial pressure of oxygen (POo, mm Hg) in the seawater was measured with a polarographic oxygen electrode (Radiometer A/S, Copenhagen, Denmark) mounted in a water-jacketed cuvette (maintained at the experimental temperature) and connected to a digital oxygen meter (Cameron Instruments Company, Port Aransas, TX). All equipment was calibrated to man- ufacturer's specifications. Oxygen level in the inflow water was measured hourly, and outflow water was otherwise monitored continuously. Water temperature, PO.,, and heart rate were recorded every 10 seconds with a computerized data acquisition system (Daqbook 120 with a DBK52 thermocouple expansion card; lotech. Inc., Cleveland, OH). The inflow ventilation volume (V , L/min) was adjusted to keep oxygen extraction between 10% and 20%. Measured PO,, values were converted to oxygen content (mg 0.,/L) following Richards (1965) and Dejours (1975). Standard metabolic rate (mg OJh) was calculated by using the Fick principle (Steffensen, 1989). Because the effects of pancuronium bromide were not reversible, at the end of each experiment individuals were euthanized with a massive overdose of sodium pen- tobarbital injected into the caudal vein. Sex was then determined and they were weighed to the nearest five grams and measured (precaudal and total lengths). Standard metabolic rate data for each individual were plotted against time and averaged over all hours (range 1-7 hours) after the outflowing PO, stabilized. Standard metabolic rate measurements were obtained at 24 ±1°C on 33 of the 34 animals. In addition, 18 animals were exposed to acute temperature changes (to 18 ±1^C or 28 ±1°C, or to both). Temperature change rates averaged 4.5 ±0.6°C per hour and 6.4 ±1.1°C per hour for cool- ing and heating, respectively, although these were not statistically different it =-1.46, df=36, P=0.15). Data Dowd et al : Metabolic rates of luvenile Carcharhinus plumbeus 325 collection was not resumed until at least one hour after the chamber had equilibrated to the new temperature (Steffensen, 1989). In order to minimize any system- atic errors, the direction of temperature change was not always the same. To quantify the effects of acute temperature changes on SMR and heart rate, Qm values were calculated over the temperature ranges 18-24°C, 24-28°C, and 18-28°C by following the methods of Schmidt-Nielsen (1997). Routine metabolic rates An annular respirometer (1250 L; diameter 167 cm) was used to measure RMR (Bushnell et al., 1989; Parsons, 1990; Carlson et al., 1999). A cage (diameter 61 cm) was placed in the center to force the sharks to swim around the perimeter of the tank. Water temperature was con- trolled during an experiment at 24-26°C by adjusting room temperature. Sharks were transferred to the respirometer and allowed to recover for 30-90 minutes. The tank was sealed and data recording continued until oxygen con- tent was reduced by 15% (=two hours). The tank water was then re-oxygenated before the next measurement by pumping the seawater through a membrane oxygenator (Medtronics, Inc., Minneapolis, MN) (Steffensen et al., 1984). A complete RMR experiment consisted of one to five iterations of this process. Oxygen concentration (mg OJh) was measured with a YSI 57 oxygen meter by using a YSI 5739 polarographic electrode oxygen-temperature probe (Yellow Spring In- struments, Yellow Spring, OH). Water temperature and oxygen content were recorded at 20-second intervals with a computerized data acquisition system (model PCA-14, Dianachart, Inc., Oak Ridge, NJ). Routine metabolic rates (mg 0,,/h) were calculated from the rate of decline in dissolved oxygen (mg 0.i/(Lxmin)) and the volume of the respirometer (Steffensen, 1989). Swimming speeds (in body lengths per second, BL/s) were determined every 15-30 minutes by averaging the time required for the shark to complete three to six laps. To account for the increased costs of swimming in a circular path, recorded RMRs were corrected to straight line estimates (RMR^,) by following the method of Weihs (1981). Statistical analysis Routine and standard metabolic rate data at each tem- perature were fitted to the allometric equation MR = a-xM^ by using nonlinear, iterative Gauss-Newton regression (Brill, 1979, 1987). This technique provides more accurate estimates of the parameters than log- transformed linear regression (Glass, 1969). The likeli- hood ratio test statistic was used to test for differences in the allometric exponents (b) among temperatures and between SMR and RMR at 24-26°C. Analysis of covariance (ANCOVA) of log-transformed metabolic rate (with log-transformed mass as covariate) was also used to test the equivalence of the exponents (b) in the untransformed allometric equations. Differences in SMR 0.91 8 910 3 4 Mass (kg) Figure 1 Standard metabolic rates (mg 0.,/hr) of juvenile sandbar sharks [Carcharhinus plumbeus) as determined by flow- through box respirometry at 18 'C (D). 24"C (O), and 28 C (Tl. Lines show best-fit allometric equations at each temperature. Error bars indicate ±1 standard error. and heart rate among temperatures were evaluated by using ANCOVA. Analysis of covariance with mass as covariate was used to test for differences in mean SMR QlO's and heart rate QlO's among the three temperature ranges. The relationship between the RMR-to-SMR ratio and body mass was assessed with linear least squares regression. Statistical analyses were performed in Sta- tistica 6.1 (StatSoft, Inc., Tulsa, OK) and SAS, version 8.0 (The SAS Institute, Inc., Gary, NC), with P<0.05 taken as the limit for significance. All values reported are means ± standard error of the mean. Results Standard metabolic rates Standard metabolic rate increased with body mass (range 1.025-10.355 kg) at all three temperatures (ANCOVA, logmass, Fj -9=265.04, P< 0.001 ) (Fig. 1). The best-fit- ting allometric equations relating SMR (mg Og/h) to body mass (M, kg) were 18°C: SMi? = 65 (±15) X ?; =16, 7-2=0.71 (1) 24°C: SMi? = 120 (±17) X ?!=33, r2=0.84 (2) ^0.79 I ±0.08 1 28°C; SM7? = 207 (±28) X 7!=16, r2=0.87 (3) J^^0 63 (±0.071 Standard metabolic rate increased with temperature for each individual and overall (ANCOVA, temperature, 326 Fishery Bulletin 104(3) (b) at each temperature were not significantly differ- ent (likelihood ratio test, x| = 3.2, P=0.20; ANCOVA, logmassxtemperature interaction, F,, 5y=0.81, P=0.45). The mean SMR Qj^'s were 3.2 ±0.4 for 18-24°C {n=U), 2.5 ±0.2 for 24-28°C (n=16), and 2.9 ±0.2 for 18-28°C (?! = 13). There was no overall effect of body Fg 59=20.99, P<0. 001). However, the allometric exponents mass on SMR QmS (increases in metabolic rate with temperature) (ANCOVA, mass, F^ 36=0.04, P=0.84), but there was a significant negative correlation between mass and SMR Qj^, for 24-28°C (P=0.014, r- = 0.36; slope = -0.20 ±0.07). The temperature range did not affect mean SMR Qj,, (ANCOVA, range, ^2,36 = 1-37, P=0.27). The data sets were therefore pooled and the overall mean Q,,, was 2.9 ±0.2 (7i=43). Heart rates Heart rate was negatively correlated with body mass at all temperatures (ANCOVA, mass, Fj ^,,,=29.99, P<0.001) (Fig. 2). The relationships between heart rate and body mass at each of the three temperatures were 4 6 Mass (kg) Figure 2 Heart rates (beats/min) of juvenile sandbar sharks (Carcharhinus plumbeus) (treated with pancuronium bromide) measured during standard metabolic rate experiments at 18°C (D), 24°C (O), and 28"C (T). Solid lines represent best-fit linear regressions at each tem- perature as a function of body mass. Error bars indicate ±1 standard error. 0.91 3 4 56789 10 Mass (kg) Figure 3 Paired routine (RMR,0) and standard metabolic rates (SMR,n) (mg 0._,/hl of 1.5 juvenile sandbar sharks {Car- charhinus plumbeus) at 24-26 C. Error bars indicate ±1 standard error. The solid line depicts the best-fitting allometric equation with the fish swimming in a curved path in an annular respirometer: RMR = 213 (±38) 3^0 79(±o 111 -pjjg dashed line represents the best-fit allo- metric equation using the corrected straight-line swim- ming (RMRj,) estimates: RMR^, = 200 (±33) M"'"''" •". 18°C: Heart rate = 39.3 (±2.0) - 1.07 (±0.49) xM 24°C: Heart rate = 66.7 (±1.6) - 1.81 (±0.30) xM 28°C: Heart rate = 80.4 (±2.9) - 2.02 (±0.61) X M. n = U,P=0.05,r^=0.29 (4) /? = 29, P<0.001, r- = 0.58 (5) « = 13, P=0.01, r- = 0.50 (6) Heart rate increased with temperature for each indi- vidual and overall (ANCOVA, temperature, F.^ 5o = 64.21, P<0.001) (Fig. 2). However, the influence of body mass on heart rate did not vary among temperatures (ANCOVA, massxtemperature interaction, F,, gij = 0.69, P=0.51). The mean QjqS for heart rate were 2.2 ±0.05 for 18-24°C (n=U). 1.8 ±0.04 for 24-28°C (/j=12), and 2.1 ±0.03 for 18-28°C («=11). Heart rate Q^g was not cor- related with body mass (ANCOVA, mass, Pj 3, = 0.95, P=0.34). However, an overall significant effect of temper- ature range on heart rate Qjg was observed (ANCOVA, range, F„ .„= 4.68, P=0.02). 18-24°C and 18-28°C were significantly different from 24-28°C (P<0.001), but not from each other (P=0.08, Tukey unequal n HSD test). Routine metabolic rates Routine metabolic rate increased with increasing body mass (Fig. 3). The best-fitting allometric equation relating RMR (mg Og/h) to mass (range 1.025-7.170 kg) was RMR = 213 (± 38) xAfO'S'^^on' 71 = 16 (53 trials), (7) r2=0.82 The estimated additional costs of swimming in a curved path versus a straight line increased with body mass (range 0.8-19.9%; Fig. 3). With the straight-line swim- ming (RMR^,) estimates, the allometric equation for RMR became: RMR, = 200 (±33) x A/o" 71 = 16, 7-2 = 0.83 (8) Although the acclimation periods in the annular res- pirometer were relatively short, it has been shown that Dowd et a\ Metabolic rates of juvenile Carcharhinus plumbeus 327 sandbar sharks recover rapidly from angling stress (6-10 h; Sparge et al.-). There was no evidence of sys- tematic decreases in RMR (relating to recovery from handling stress) during individual RMR experiments, which averaged 16.2 ±2.0 hours in length. Using only the final trial for each individual, we fitted RMR to the allometric equation: RMR = 203 (±35) X Af" ■ = 16, r'- = 0.78 (9) Routine (average ±SEM) swimming speeds decreased with increasing body size according to the exponential equation speed (bl/s) = 3.54 TL"" " (r^^O.lS) (Fig. 4). In most cases the animal maintained a swimming speed and direction along the outer wall of the chamber for 5-20 minutes before turning around. Because each shark swam at a relatively constant speed, the effect of swimming speed on metabolic rate could not be determined. Paired standard and routine metabolic rates Paired SMR and RMR measurements were obtained for 15 sharks (1.025-7.170 kg) (Fig. 3). The mean ratio of RMR to SMR at 24°C was 1.8 ±0.1. The mean ratio of RMR^i to SMR was 1.6 ±0.1. There was no significant correlation between body mass and the ratio of RMR to SMR (P=0.93, r'-<0.01}. The allometric exponents for RMR and SMR at 24°C were also not significantly differ- ent (likelihood ratio test, xi=0.002, P=0.96; ANCOVA, logmassxtype interaction, F^ ^5 = 0. 33, P=0.57). Discussion Effects of body mass and temperature on SMR and RMR This study presents the first direct measures of SMR and expands the size range over which SMR and RMR have been reported for continuously active shark spe- cies (Fig. 5). Sandbar shark metabolic rate, like that of a wide variety of species, increases with increasing body mass according to the allometric equation MR = aM'>, with a 6 of ~0.71-0.79 (Schmidt-Nielsen, 1997), The effects of body mass on SMR, RMR, and RMR^, (6 in Eqs. 1-3 and 7-9) in sandbar sharks were simi- lar to published values for other elasmobranchs (e.g., Pritchard et al., 1958; DuPreez et al., 1988; Sims, 1996). The temperature independence of 6, previously reported for the lesser sandshark (Rhinobatos armulatus) and the bullray (Myliobatus aquila) (DuPreez et al., 1988), was evident overall in sandbar sharks. However, there was a significant effect of mass on SMR Qjg for the 24-28°C 2 Sparge, A. L., N. Kohler, G. Skomal, and R. Goodwin. 2001. The physiological effects of angling on post-release survivorship in juvenile sandbar sharks ^Carcharhinus plumbeus). (Abstract.) American Elasmobranch Society 17'*' Annual Meeting, State College, PA. Website:http://www. flmnh.un.edu/fish/organizations/aes/abst2001d.htm [accessed on 26 September 2005.] 70 100 110 80 90 Total length (cm) Figure 4 Mean (±1 SE) voluntary swimming speeds of sandbar sharks (Carcharhinus plumbeus) during RMR experi- ments. The solid line represents the best-fit equation: speed (BL/s) = S.bixUotal length)-'>^\ 1000 800 600 400 ^ 300 / ♦ 9 y ...■■' ■ 200 100 80 60 0.3 0.4 0.6 0.8 1 2 3 4 6 8 10 Mass (kg) Figure 5 Standard metabolic rates of active elasmobranch species and tunas as a function of body mass: Isurus oxyrinchus 18 = 0 (•, Graham et al., 1990); Sphyrna lewini 21°C (■), 26^0 (0», and 29°C (D) (Lowe, 2001); Negaprion brevirostris 22-25°C (T, Bushnell et al., 1989) and 25°C (0, Scharold and Gruber, 1991); Carcharhinus acrono- tus 28°C (♦, Carlson et al., 1999); C. plumbeus 18°C ( ), 24°C( ), and 28°C ( ) (present study); kawakawa iEuthynjius affinis) 25°C ( , Brill, 1987); yellowfin tuna iThunnus albacares) 25"C (- ■■ -, Brill, 1987); skipjack tuna iKatsuwonus pelamis) 23.5-25.5°C (- ■ -, Brill, 1979). Lines are best-fit allometric equa- tions at the stated experimental temperatures. range. These two findings appear to contradict each other, but the 24-28°C pattern may be influenced by small sample sizes at larger body masses (only 2 sharks >7.5 kg). 328 Fishery Bulletin 104(3) The effects of acute temperature change on SMR were consistent with published values for other elasmobranchs (Qi(,=2-3; e.g., DuPreez et al., 1988; Carlson and Par- sons, 1999; Miklos et al., 2003). The Q^ for the SMR in elasmobranchs has also been reported to vary with the temperature ranges assessed (Butler and Taylor, 1975; Hopkins and Cech, 1994), but this was not the case for sandbar sharks. It is important to note the distinc- tion between acute temperature changes and seasonal acclimatization when reporting Qjg values (Schmidt- Nielsen, 1997). In the present study, sandbar sharks were exposed to rapid temperature changes that mir- rored short-term temperature fluctuations experienced in the wild (SMR Qi,j=2.9 ±0.2). In seasonally acclima- tized bonnethead sharks (Sphyrna tiburo), the effect of seasonal temperature change on metabolic rate was lower (Qj„=2.29-2.39; Carlson and Parsons, 1999). Cost of activity and routine energy expenditure The SMR is never realized in fish that must swim contin- uously to maintain hydrostatic equilibrium or to venti- late their gills. Measurement of SMR and RMR in active species nevertheless allows insight into the division of metabolic costs between swimming and maintenance processes. For example, the average metabolic rate of juvenile scalloped hammerhead sharks (Sphyrna lewini) in the wild was 1.4 times the estimated SMR (Lowe, 2002). In sandbar sharks, the RMR to SMR ratio (1.8 ±0.1) and RMR^i to SMR ratio (1.6 ±0.1) were similar to those observed and estimated for several elasmobranch species (e.g., 1.5, Brett and Blackburn, 1978; 1.4, Nixon and Gruber, 1988; 1.7, Carlson et al., 1999). In other words, SMR comprises 56-63% of total metabolic rate at routine activity levels. Because the allometric expo- nents for RMR and SMR were not different at 24°C, we conclude that the RMR-to-SMR ratio (and, therefore, cost of activity) is also size independent, at least over the size range of sandbar sharks tested. Our metabolic rate data span the size and tempera- ture ranges relevant to the summer populations of ju- venile sandbar sharks in Chesapeake Bay and other western Atlantic nursery areas (Grubbs et al., in press; Merson and Pratt, 2001). Bioenergetics models require estimates of field activity and corresponding energetic costs (Lowe, 2002). The swimming speeds of sandbar sharks in the annular respirometer (mean 0.55 ±0.03 bl/s) were well within the range of activity levels ob- served in nature (Grubbs, 2001). After the application of an oxycalorific coefficient of 13.6 J/(mg O^) (Elliott and Davison, 1975), the RMR and RMR^, for a 1-kg sandbar shark at 24"C represent energy expenditures of 69.7 and 63.4 kJ/day, respectively. These values are comparable to those for the lemon shark {Negaprion brevirostris, 67.7 kJ/day; Nixon and Gruber, 1988), S. tiburo (80.2 kJ/day; Parsons, 1990), and S. lewini (96 kJ/day at ~28°C; Lowe, 2002) . The Q,„ values for SMR obtained between 18° and 28°C demonstrate that juve- nile sandbar shark metabolic demands change signifi- cantly as ambient temperature changes, both on short time scales and over the course of the summer stay in the nursery areas. Heart rates Heart rate decreased with increasing body mass but increased with temperature (Fig. 2), as it does for other ectothermic species (Schmidt-Nielsen, 1997). Heart rates of juvenile sandbar sharks were comparable to heart rates of two other shark species while swimming (Scharold et al., 1989; Scharold and Gruber, 1991), although the sandbar shark data should be interpreted with caution. Pancuronium bromide has been shown to exhibit vagolytic activity in mammals (Melnikov et al., 1999), but to our knowledge its effect on shark heart rates is unknown and would depend on the resting vagal tone. In the dogfish iScylior-hinus canicula), resting vagal tone increased with temperature between 7°C and 17°C (Taylor et al., 1977). The resting vagal tone and resulting elevation in heart rate after treatment with pancuronium bromide could be significant in sandbar sharks, especially at the higher temperatures. If so. Figure 2 may reflect the effect of temperature and body mass on intrinsic heart rate. Measuring SMR of immobilized sharks Standard metabolic rate is defined as the oxygen con- sumption of a postabsorptive-stage fish at rest, and it is considered the minimum metabolic cost of organismal maintenance (Brett and Groves, 1979). Two methods are commonly used to determine SMR. In the first, the slope of a power-performance curve relating the loga- rithm of oxygen consumption rate to relative swimming speed is extrapolated back to zero activity (e.g., Lowe, 2001). However, extrapolation does not take into account physiological differences between active and quiescent fish, specifically the induction of anaerobic metabolism during high-velocity swimming, and may misrepresent SMR (Brett and Groves, 1979; Cech, 1990). Further, swimming kinematics can be significantly altered in a swim flume, leading to overestimates of SMR (Lowe, 1996, 2001). The second option for measuring SMR is to confine the fish in a sealed or flow-through respirometer (e.g.. Brill, 1987; Hopkins and Cech, 1994). This process works well for sedentary species, but active fish will struggle in such situations, requiring the use of paralytic and sedative agents, as well as artificial ventilation. Several studies have confirmed, however, that the two methods yield identical results (SMRs and allometric exponents) in various fish species (e.g., yellowfin tuna [Thunnus albacares]; kawakawa [Euthynnus affinis]; skipjack tuna [Katsuwonus pelamis]; rainbow trout [On- chorynchus mykiss\\ American shad [Alosa sapidissima]; aholehole [Kiihlia sandvicensis] [Brill, 1979, 1987; Dew- ar and Graham, 1994; Leonard et al., 1999)). Moreover, treatment with anaesthetics has been shown to have no effect on the SMR of little skate (Raja erinacea ; Hove and Moss, 1997) or nursehound IScyliorhinus stellaris; Baumgarten-Schumann and Piiper, 1968). Dowd et al : Metabolic rates of luvenile Caichorhinus plumbeus 329 Because sandbar sharks are continuously active, we chose to measure SMR in immobilized and artificially ventilated animals in flow-through respirometers. As a check of this technique, a two-point power-performance curve was constructed by using the logarithms of self- paired SMR and RMR and the mean swimming speed of the animal in the annular respirometer. The average slope for 15 sharks (0.38 ±0.04) at 24-26°C was similar to the slopes of power-performance curves for four other ectothermic shark species (0.2, Scharold et al., 1989; 0.34, Scharold and Gruber, 1991; 0.38, Carlson et al., 1999; 0.32, Lowe, 2001), and we interpret these data as additional evidence that the technique provides accept- able results. This approach, moreover, avoids the expense and difficulties of developing a swimming tunnel large enough to accommodate juvenile sandbar sharks and may well be generally applicable for generating power perfor- mance curves in other continuously active fish species. Elasmobranch metabolic rates and the cost of growth The published SMRs of active elasmobranchs are well below those of high-energy demand teleosts (e.g., the endothermic tunas; Korsmeyer et al., 1996), with the exception of the endothermic mako shark Usurus oxy- rinchus) (Graham et al., 1990) (Fig. 5). Brill (1987, 1996) proposed that the high SMRs of tunas are an unavoidable consequence of their morphological, bio- chemical, and physiological adaptations for extremely high maximum aerobic metabolic rates, specifically that their large gill surface areas have led to high osmoregulatory costs (but see Brill et al., 2001). Most elasmobranchs, including the sandbar shark, have less than one third the gill surface area of a similar-size tuna (Muir and Hughes, 1969; Emery and Szcepan- ski, 1986). These modest mass-specific gill surface areas and corresponding low rates of oxygen delivery in ectothermic sharks likely dictate slow asymptotic growth rates (Pauly, 1981), in contrast to those of tunas, whose cardiovascular systems are adapted for meeting multiple metabolic demands (including growth) (Bushnell and Brill, 1991; Korsmeyer et al., 1996; Brill and Bushnell, 2001). Significant levels of specific dynamic action (SDA, the elevation in metabolic rate in conjunction with protein synthesis after a meal [Brown and Cameron, 1991]) can probably not be met by the cardio-respiratory systems of elasmobranchs, particularly in continuously active species such as the sandbar shark, while they sustain routine activity levels. The oxygen consumption rate fol- lowing a meal can exceed 2-3 times the SMR (DuPreez et al., 1988; Sims and Davies, 1994; Ferry-Graham and Gibb, 2001), whereas the RMR of sandbar sharks was 1.6-1.8 times SMR. Assuming that the maximum meta- bolic rate is 1.8 to 2.75 times the SMR (Scharold et al., 1989; Lowe, 2001), we determined that sandbar sharks are using between 34% and 100% of their metabolic scope just to sustain routine activity levels. Given these limitations, sandbar sharks and other active elasmobranchs probably make tradeoffs among metabolic demands at the expense of SDA or growth to remain within their available metabolic scope, or they may adjust their behavior to seek cooler waters during digestion (Matern et al., 2000). Because rapid incorpo- ration of ingested amino acids into body proteins is not possible, many elasmobranchs may have to reduce the rate of digestion or integrate SDA over a longer period (or do both). For example, estimated daily rations for several elasmobranch species average 1-2% of body weight per day (e.g.. Bush and Holland, 2002; Dowd et al., 2006), compared to 4% or more in fast growing teleosts (e.g., Olson and Boggs, 1986). Not surprisingly, sandbar sharks in the Northwest Atlantic mature only after 13-15 years and grow less than 10 cm per year during that time (Sminkey and Musick, 1995). Growth rates for many other large, active elasmobranch spe- cies are also slow (Branstetter, 1990). Future research might well focus on exploring the relationship between SDA, active metabolic rates, and metabolic scopes of slow growing, continuously active sharks. Acknowledgments We thank Mark Luckenbach, Reade Bonniwell, and the entire staff of the VIMS Eastern Shore Laboratory for their hospitality, advice, and assistance in capturing and maintaining sandbar sharks. This work was sup- ported through the National Shark Research Consor- tmm (NOAA/NMFS grant no. NA17FL2813 to JAM); an Indiana University South Bend Faculty Research Award to PGB; and VIMS GSA Mini-Grant, William & Mary Minor Research Grant, and Oceanside Conservation Co., Inc., awards to WWD. We thank Medtronics for donating the membrane oxygenator used in the RMR experiments. All procedures were approved by the College of William and Mary Research on Animal Subjects Committee. Literature cited Baum, J. K., R. A. Myers, D. G. Kehler, B. Worm, S. J. Harley, and P. A. Doherty. 2003. Collapse and conservation of shark populations in the Northwest Atlantic. Science 299:389-392. Baumgarten-Schumann, D., and J. Piiper. 1968. Gas exchange in the gills of resting unanesthe- tized dogfish (Scyliorhinus stellaris). Resp. Physiol. 5:317-325. Branstetter, S. 1990. Early life-history implications of selected carcharhi- noid and lamnoid sharks of the Northwest Atlantic. In Elasmobranchs as living resources: advances in the biology, ecology, systematics, and status of the fisheries (H. L. Pratt Jr, S. H. Gruber, and T. Taniuchi, eds.), p. 17-28. NOAA Technical Report NMFS 90. Brett, J. R., and J. M. Blackburn. 1978. Metabolic rate and energy expenditure of the spiny dogfish, Squalus acanthias. J. Fish. Res. Board Can. 35:816-821. Brett, J. R., and T. D. D. Groves. 1979. Physiological energetics. In Fish physiology 330 Fishery Bulletin 104(3) (W. S. Hoar, D. J. Randall, and J. R. Brett, eds.), p. 279-352. Academic Press, New York, NY. Brewster-Geisz, K. K., and T. J. Miller. 2000. Management of the sandbar shark, Carcharhinus plumbeiis: implications of a stage-based model. Fish. Bull. 98:236-249. Brill, R. W. 1979. The effect of body size on the standard metabolic rate of skipjack tuna, Katsuwonus pelamis. Fish. Bull. 77:494-498. 1987. On the standard metabolic rates of tropical tunas, including the effect of body size and acute temperature change. Fish. Bull. 85:25-35. 1996. Selective advantages conferred by the high per- formance physiology of tunas, billfishes, and dolphin fish. Comp. Biochem. Phys. A. 113:3-15. Brill, R. W., and R G. Bushnell. 2001. The cardiovascular system of tunas. //; Tuna — physiology, ecology and evolution (B. A. Block and E. D. Stevens, eds.), p. 79-120. Academic Press, San Diego, CA. Brill, R., Y. Swimmer, C. Taxboel, K. Cousins, and T. Lowe. 2001. Gill and intestinal Na+-K+ ATPase activity, and estimated maximal osmoregulatory costs, in three high-energy-demand teleosts: yellowfin tuna ^Thun- nus albacares), skipjack tuna i Katsuwonus pelamis), and dolphin fish iCoryphacna hippurus). Mar. Biol. 138:935-944. Brown, C. R., and J. N. Cameron. 1991. The relationship between specific dynamic action (SDA) and protein synthesis rates in the channel catfish. Physiol. Zool. 64:298-301. Bush, A., and K. Holland. 2002. Food limitation in a nursery area: estimates of daily ration in juvenile scalloped hammerheads, Sphyr- na lewini (Griffith and Smith, 18341 in Kane'ohe Bay, O'ahu, Hawai'i. J. Exp. Mar. Biol. Ecol. 278:157-178. Bushnell, R G., and R. W. Brill. 1991. Responses of swimming skipjack {Katsuwonus pelamis) and yellowfin (Thunnus albacares) tunas to acute hypoxia, and a model of their cardiovascular function. Physiol. Zool. 64:787-811. Bushnell, P. G., P. L. Lutz, and S. H. Gruber. 1989. The metabolic rate of an active, tropical elasmo- branch, the lemon shark ^Negaprion brevirostris). J. Exp. Biol. 48:279-283. Butler, P. J., and E. W. Taylor 1975. The effect of progressive hypoxia on respiration in the dogfish iScyliorhinus canicula) at different seasonal temperatures. J. Exp. Biol. 63:117-130. Carlson, J. K., C. L. Palmer, G. R. Parsons. 1999. Oxygen consumption rate and swimming efficiency of the blacknose shark, Carcharhinus acronotus. Copeia 1999:34-39. Carlson, J. K., and G. R. Parsons. 1999. Seasonal differences in routine oxygen consump- tion rates of the bonnetbead shark. J. Fish Biol. 55:876-879. Casey, J. G., and L. J. Natanson. 1992. Revised estimates of age and growth of the sandbar shark (Carcharhinus plumbeus) from the Western North Atlantic. Can. J. Fish. Aquat. Sci. 49:1474-1477. Cech, J. J., Jr. 1990. Respirometry. In Methods for fish biology (C. B. Schreck and P. B. Moyle, eds. ), p. 535-556. Am. Fish. Soc, Bethesda, MD. Cortes, E. 1999. A stochastic stage-based population model of the sandbar shark in the Western North Atlantic. In Life in the slow lane: ecology and conservation of long-lived marine animals (J. A. Musick, ed.), p. 115-136. Am. Fish. Soc. Symp. 23. Bethesda, MD. Dejours, P. 1975. Principles of comparative respiratory physiology, 253 p. Elsevier. New York, NY. Dewar, H., and J. B. Graham. 1994. Studies of tropical tuna swimming performance in a large water tunnel. 1. Energetics. J. Exp. Biol. 192:13-31. Dowd, W. W., R. W. Brill, R G. Bushnell, and J. A. Musick. 2006. Estimating consumption rates for juvenile sand- bar sharks {Carcharhinus plumbeus) in Chesapeake Bay. Virginia, using a bioenergetics model. Fish. Bull. 104:332-342. DuPreez, H. H., A. McLachlan, and J. F. K. Marais. 1988. Oxygen consumption of two nearshore marine elas- mobranchs, Rhinobatus annulatus (Muller and Henle, 1841) SinAMyliobatus aquila (Linnaeus, 17581. Comp. Biochem. Physiol. 89A:283-294. Elliott, J. M., and W. Davison. 1975. Energy equivalents of oxygen consumption in animal energetics. Oecologia 19:195-201. Emery. S. H.. and A. Szcepanski. 1986. Gill dimensions in pelagic elasmobranch fishes. Biol. Bull. 171:441-449. Ferry-Graham, L. A., and A. C. Gibb. 2001. A comparison of fasting and post-feeding met- abolic rates in a sedentary shark Cephaloscyllium ventriosum. Copeia 2001:1108-1113. Glass, N. R. 1969. Discussion of calculation of power function with special reference to respiratory metabolism in fish. J. Fish. Res. Board Can. 26:2643-2650. Graham, J. B., H. Dewar, N. C. Lai, W. R. Lowell, and S. M. Arce. 1990. Aspects of shark swimming performance deter- mined using a large water tunnel. J. Exp. Biol. 151: 175-192. Grubbs, R. D. 2001. Nursery delineation, habitat utilization, move- ments, and migration of juvenile Carcharhinus plum- beus in Chesapeake Bay, Virginia, LISA. Ph. D. diss., 223 p. College of William and Mary, Gloucester Point, VA. Grubbs, R. D., J. A. Musick, C. L. Conrath, and J. G. Romine. In press. Long-term movements, migration, and temporal delineation of a summer nursery for juvenile sandbar sharks in the Chesapeake Bay region. In Shark nurs- ery grounds of the Gulf of Mexico and the East Coast waters of the United States (C. T. McCandless, N. E. Kohler, and H. L. Pratt Jr., eds.). Am. Fish. Soc. Symp., Bethesda, MD. Hopkins, T. E., and J. J. Cech Jr. 1994. Effect of temperature on oxygen consumption of the bat ray, Myliobatis californica (Chondrichthyes, Myliobatididae). Copeia 1994:529-532. Hove, J. R., and S. A. Moss. 1997. Effect of MS-222 on response to light and rate of metabolism of the little skate Raja erinacea. Mar. Biol. 128:579-583. Korsmeyer, K. E., H. Dewar, N C. Lai, and J. B. Graham. 1996. The aerobic capacity of tunas: Adaptation for Dowd et a\ Metabolic rates of luvenile Carcharhinus plumbeui 331 multiple metabolic demands. Comp. Biochem. Physiol. 113A:17-24. Leonard, J. B., J. F. Norieka, B. Kynard, and S. D. McCormick. 1999. Metabolic rates in an anadromous clupeid, the American shad iAlosa sapidissima). J. Comp. Physiol. 169B:287-295. Lowe, C. G. 1996. Kinematics and critical swimming speed of juve- nile scalloped hammerhead sharks. J. Exp. Biol. 199:2605-2610. 2001. Metabolic rates of juvenile scalloped hammerhead sharks {Sphyrna lewuu). Mar. Biol. 139:447-453. 2002. Bioenergetics of free-ranging juvenile scalloped hammerhead sharks iSphyrna lewini) in Kane'ohe Bay, O'ahu, HI. J. E.xp. Mar. Biol. Ecol. 278:141-156. Matern, S. A., J. J. Cech Jr, and T. E. Hopkins. 2000. Diel movements of bat rays, Myliobatis califor- nica. in Tomales Bay, California: evidence for behavioral thermoregulation? Environ. Biol. Fish. 58:173-182. Medved, R. J. 1985. Gastric evacuation in the sandbar shark, Carcha- rhinus plumbeus. J. Fish Biol. 26:239-253. Medved, R. J., C. E. Stillwell and J. G. Casey. 1988. The rate of food consumption of young sandbar sharks iCarcharhinus plumbeus) in Chincoteague Bay, Virginia. Copeia 1988:956-963. Melnikov, A. L., K. Y. Malakhov, K. G. Helgesen, and D. A. Lathrop. 1999. Cardiac effects of non-depolarizing neuromus- cular blocking agents pancuronium, vecuronium, and rocuronium in isolated rat atria. Gen. Pharmacol. 33:313-317. Merson, R. R., and H. L. Pratt Jr. 2001. Distribution, movements and growth of young sandbar sharks, Carcharhinus plumbeus. in the nurs- ery grounds of Delaware Bay. Environ. Biol. Fish. 61:13-24. Miklos, P., S. M. Katzman, and J. J. Cech Jr. 2003. Effect of temperature on oxygen consumption of the leopard shark, Triakis semifaseiata (Chondrichthyes Triakidae). Environ. Biol. Fish. 66:15-18. Muir, B. S., and G. M. Hughes. 1969. Gill dimensions for three species of tunny. J. Exp. Biol. 51:271-285. Musick, J. A., S. Branstetter, and J. A. Colvocoresses. 1993. Trends in shark abundance from 1974 to 1991 for the Chesapeake Bight region of the U.S. Mid-Atlantic coast. In Conservation biology of elasmobranchs (S. Branstetter, ed.), p. 1-18. NCAA Technical Report, NMFS 115. Musick, J. A., and J. A. Colvocoresses. 1986. Seasonal recruitment of subtropical sharks in Chesapeake Bight, USA. In Workshop on recruitment in tropical coastal demersal communities (A. Yanez-Aran- cibia and D. Pauly, eds.), p. 301-311. FAG/UNESCO, Campeche, Mexico. Ney, J. J. 1993. Bioenergetics modeling today: growing pains on the cutting edge. Trans. Am. Fish. Soc. 122:736-748. Nixon, A. J., and S. H. Gruber. 1988. Diel metabolic and activity patterns of the lemon shark {Negaprion brevirostris). J. Exp. Zool. 248:1-6. Olson, R. J., and C. H. Boggs. 1986. Apex predation by yellowfin tuna {Thun- nus albacares): independent estimates from gastric evacuation and stomach contents, bioenergetics, and cesium concentrations. Can. J. Fish. Aquat. Sci. 43: 1760-1775. Oswald, R. L. 1978. Injection anesthesia for experimental studies in fish. Comp. Biochem. Phys. 60C:19-26. Parsons, G. R. 1990. Metabolism and swimming efficiency of the bonnet- head shark Sphyrna tiburo. Mar. Biol. 104:363-367. Pauly, D. 1981. The relationships between gill surface area and growth performance in fish: a generalization of von Bertalanffy's theory of growth. Meeresforschung 28.-251-282. Pritchard, A. W., E. Florey, and A. W. Martin. 1958. Relationship between metabolic rate and body size in an elasmobranch {Squalus suckleyi [S. acanthias]) and in a teleost (Ophiodon elongatus). J. Mar. Res. 17:403-411. Richards, F. A. 1965. Dissolved gases other than carbon dioxide. In Chemical cceanography (J. P. Riley and G. Skirrow, eds.), p. 197-224. Academic Press, New York, NY. Scharold, J., and S. H. Gruber 1991. Telemetered heart rate as a measure of metabolic rate in the lemon shark, Negaprion brevirostris. Copeia 1991:942-953. Scharold, J., N. C. Lai, W. R. Lowell, and J. B. Graham. 1989. Metabolic rate, heart rate, and tailbeat frequency during sustained swimming in the leopard shark Triakis semifaseiata. J. Exp. Biol. 48:223-230. Schmidt-Nielsen, K. 1997. Animal physiology: adaptation and environment, 607 p. Cambridge Univ. Press, New York, NY. Sims, D. W. 1996. The effect of body size on the standard meta- bolic rate of the lesser spotted dogfish. J. Fish Biol. 48:542-544. Sims, D. W., and S. J. Davies. 1994. Does specific dynamic action iSDAl regulate return of appetite in the lesser spotted dogfish, Seyliorhinus eaniculal J. Fish Biol. 45:341-348. Sminkey, T. R.. and J. A. Musick. 1995. Age and growth of the sandbar shark, Carcharhinus plumbeus, before and after population depletion. Copeia 1995:871-883. 1996. Demographic analysis of the sandbar shark, Careha- rhinus plumbeus, in the western North Atlantic. Fish. Bull. 94:341-347. Steffensen, J. F. 1989. Some errors in respirometry of aquatic breath- ers: How to avoid and correct for them. Fish Physiol. Biochem. 6:49-59. Steffensen, J. F., K. Johansen, and P. G. Bushnell. 1984. An automated swimming respirometer. Comp. Biochem. Physiol. 79A:437-440. Stillwell, C. E., and Kohler, N. E. 1993. Food habits of the sandbar shark Carcharhinus plumbeus off the U.S. northeast coast, with estimates of daily ration. Fish. Bull. 91:138-150. Taylor, E. W., S. Short, P. J. Butler. 1977. The role of the cardiac vagus and the response of a dogfish Seyliorhinus canicula to hypoxia. J. Exp. Biol. 70:57-75. Weihs, D. 1981. Effects of swimming path curvature on the ener- getics offish motion. Fish. Bull. 79:171-176. 332 Abstract — Using a bioenergetics model, we estimated daily ration and seasonal prey consumption rates for six age classes of juvenile sandbar sharks (Carcharhinus plumbeus) in the lower Chesapeake Bay summer nursery area. The model, incorporat- ing habitat and species-specific data on growth rates, metabolic rate, diet composition, water temperature (range 16.8-27.9 "O, and population struc- ture, predicted mean daily rations between 2.17 ±0.03 (age-0) and 1.30 ±0.02 (age-5) % body mass/day. These daily rations are higher than earlier predictions for sandbar sharks but are comparable to those for ecologi- cally similar shark species. The total nursery population of sandbar sharks was predicted to consume -124,000 kg of prey during their 4.5 month stay in the Chesapeake Bay nursery. The predicted consumption rates sup- port the conclusion that juvenile sandbar sharks exert a lesser top- down effect on the Chesapeake Bay ecosystem than do teleost piscivores and humans. Estimating consumption rates of juvenile sandbar sharks (Carcharhinus plumbeus) in Chesapeake Bay, Virginia, using a bioenergetics model* W. Wesley Dowd^ Richard W. BrilP Peter G. BushnelP John A. Musick' ' Department of Fisheries Science Virginia Institute of Marine Science 1208 Create Road, P.O. Box 1346 College of William and Mary Gloucester Point, Virginia 23062 Present address (lor W Dowd) Graduate Group in Ecology Dept Wildlife, Fish and Conservation Biology University of California One Shields Avenue Davis, California 95616 E-mail address (for WW Dowd) wwdowdia)ucdavis edu ^ Virginia Cooperative Marine Education and Research Program Virginia Institute of Marine Science 1208 Create Road, PO. Box 1346 College of William and Mary Gloucester Point, Virginia 23062 ' Department of Biological Sciences Indiana University South Bend 1700 Mishawaka Avenue South Bend, Indiana 46634 Manuscript submitted 29 October 2004 to the Scientific Editor's Office. Manuscript approved for publication 15 September 2005 by the Scientific Editor. Fish. Bull. 104:332-342 (2006). The lower Chesapeake Bay, Mid- Atlantic Bight, and adjacent coastal lagoon systems serve as the primary summer nursery areas for the North- west Atlantic Ocean sandbar shark {Carcharhinus plumbeus) population (Musick et al., 1993), Sandbar sharks are the most abundant large coastal sharks in the Mid-Atlantic Bight (Musick et al., 1993) and an impor- tant part of the commercial shark catch. After the rapid expansion of the fishery in the mid 1980s, the sandbar shark population in Virginia's coastal ocean waters declined by approxi- mately 66% by 1991 (Musick et al., 1993). Meanwhile, catch rates in the lower Chesapeake Bay, the core nurs- ery area for juvenile sandbar sharks, remained relatively stable (Musick et al., 1993). Because juvenile sandbar sharks return to the coastal or estua- rine nursery grounds for the first four to six summers of life (Sminkey and Musick, 1995; Grubbs et al., in press), these nursery grounds are vital to the life history and potential recovery of the Northwest Atlantic sandbar shark stock (Branstetter, 1990; Hoff and Musick, 1990; Sminkey and Musick, 1996; Cortes, 1999). Despite the abundance and posi- tion of elasmobranchs at the apex of many coastal and pelagic food webs, their energetic demands and the role of elasmobranchs as predators have rarely been quantified (Gruber, 1985; DuPreez et al., 1990; Sundstrom and Gruber, 1998; Lowe, 2002; Schindler et al., 2002). In the Chesapeake Bay, sandbar sharks occupy an apex posi- tion in the food web, preying upon ' Contribution number 2721 from Virginia Institute of Marine Science, College of William and Mary, Gloucester Point, VA. Dowd et al,: Consumption rates of Carcharchinus plumbeus in Chesapeake Bay 333 Table 1 Parameters, distributions, and vakies used in error analyses of the sandbar shark iCarcharhiniis pIumheus)h\oenergetics model. See text for parameter definitions. For parameters with triangular distributions, the initial estimates described in the text were assumed to be the most likely values. Parameter Distribution type Mean or most likely value SE or range Source SMRa Normal 120.0 17.3 Dowd etal. (2006) SMRb Normal 0.788 0.076 Dowd etal. (2006) Qio Normal 2.89 0.16 Dowd etal. (2006) ACT Normal 1.62 0.11 Dowd etal. (2006) SDA Triangular O.IOC 0.06-0.17C DuPreez et al. (1988), Sims and Davies (1994), Duffy (1999), Ferry-Graham and Gibb (2001) L^ Normal 164 cm 16.4' Sminkey and Musick ( 1995 ) to Normal -3.8 yr 0.38' Sminkey and Musick ( 1995 ) K Normal 0.089 0.0089' Sminkey and Musick ( 1995 ) P Normal 0.75 0.075' Sminkey and Musick ( 1995 ) F Triangular 0.20C 0.17-0.38C Wetherbee and Gruber ( 1993 ) U Triangular 0.07C 0.05-0.08C Brett and Groves (1979), Duffy (1999) ' SE was assigned by the authors to yield a coefficient of variation of 10% Isensu Bartell et al., 1986) a number of commercially important species such as menhaden (Brevoortia tyranniis), blue crabs (Callinectes sapidus), striped bass (Moroi^e saxatilis), and bluefish (Pomatomus saltatrix) (Medved and Marshall, 1981; Medved et al., 1985; Stillwell and Kohler, 1993; El- lis, 2003). Interestingly, previous ecosystem models have predicted both significant (Stevens et al., 2000) and negligible (Kitchell et al., 2002) top-down effects of changes in shark biomass on ecosystem structure, depending primarily on the trophic complexity of the system and the incidence of omnivory (Bascompte et al., 2005). Because the sandbar shark is one of the few species for which many of the necessary modeling parameters have been measured, it serves as an excellent system for assessing the bioenergetics and ecosystem role of large coastal elasmobranchs. This article has the fol- lowing objectives: 1 to construct a realistic bioenergetics model for juve- nile sandbar sharks in the Chesapeake Bay summer nursery grounds. Because previous sandbar shark models have suffered from a lack of species-spe- cific data (Medved et al., 1988; Stillwell and Kohler, 1993), we have incorporated updated species-specific and habitat-specific data. 2 to use the model to assess the role of juvenile sand- bar sharks as predators in the Chesapeake Bay to aid ecosystem modelers and fishery management efforts. 3 to test the sensitivity of the model to uncertainty in parameter estimates using error analysis to identify future research priorities (Kitchell et al., 1977). Materials and methods Study area and nursery habitat The core sandbar shark nursery area (-500-1000 km^; Grubbs and Musick, in press) in the lower, eastern Ches- apeake Bay supports a seasonal population of -10,000 individuals (Sminkey, 1994), composed almost entirely of sandbar sharks <90 cm precaudal length (PCL) (Musick et al., 1993; VIMS'). Juvenile sandbar sharks move actively throughout the nursery area, covering large activity spaces (>110 km-) and the entire water column, as shown in telemetry studies (Medved and Marshall, 1983; Grubbs, 2001). Sandbar sharks in the nursery area are exposed to both long-term and short-term changes in water tem- peratures. Juvenile sandbar sharks inhabit Chesapeake Bay at seasonal temperatures ranging from 15 to 29°C (VIMSM. During the months of July and August, a seasonal thermocline also develops in the lower Chesa- peake Bay, which sandbar sharks will cross repeatedly throughout the day (Grubbs, 2001), The magnitude of the temperature gradient from top to bottom is typically 5-6°C (VIMS', Chesapeake Bay Program^). VIMS (Virginia Institute of Marine Science) Shark Ecology Program Longline Survey. 1973-2003. Unpubl. data (as a Microsoft Excel file). [Available from J. A. Musick. 1208 Greate Road, Gloucester Point, VA 23062-1346.] ' Chesapeake Bay Program Water Quality Database. Website: http://www.chesapeakebay.net/data/index.htni (accessed on March 2003.] 334 Fishery Bulletin 104(3) Bioenergetics model Rates of anabolism, catabolism, and waste losses (Table 1) were used to construct a bioenergetics model that predicted daily energy consumption (C^, in joules per day, J/d): Cr RMR„ + SDA + G„ + F +11. (1) The model used a daily time step, consistent with the determination of daily energy ration. Due to the reporting of the daily routine metabolic rate (RMRj^), specific dynamic action (SDA), fecal losses (F). and excretions (U) as fractions of consumption (see below), we rearranged Equation 1 and solved for C,; to yield the model: RMR^+G„ a-SDA-U-F) (2) We set the immigration and emigration dates for the simulation as May 15 and September 30, respectively (VIMSi). We used the model to estimate daily energy ration for average individuals within each of six age-classes us- ing the Chesapeake Bay nursery (Musick et al., 1993). In turn, we combined energetic requirements with diet composition data to estimate rates of food consump- tion (daily ration) and predatory impact of individual sharks over the course of the summer for each age class. Finally, these individual estimates were merged with estimates of population size and age structure to esti- mate the overall predatory demand of juvenile sandbar sharks in the Chesapeake Bay nursery area. Model parameters Routine metabolic rate (RMR) Like a number of car- charhiniform species, sandbar sharks are continuously active, which leads to high daily metabolic expenditures (e.g., Carlson et al., 1999). As a result, metabolic rate is the largest and most variable component of the energy budget for these active fish (Kerr, 1982; Boisclair and Leggett, 1989). Unfortunately, because of a paucity of available data, metabolic rate parameters are often borrowed from other species (e.g., Schindler et al., 2002). Sensitivity analyses have shown that accurate metabolic rate data are needed to construct realistic bioenergetics models (Kitchell et al., 1977; Bartell et al., 1986). The allometric (size-dependent) influence on standard metabolic rate (SMR) in juvenile sandbar sharks was re- cently determined over the entire size range (42-92 cm PCL, 1-10 kg) characteristic of the Chesapeake Bay nursery area in flow-through respirometers for sharks treated with a neuromuscular blocker (Dowd et al., 2006). The best fitting allometric equation for SMR {SMR=axMh for 33 sharks at 24°C was where M = mass in kilograms; and SMR = mgOj consumed per hour. The values in parentheses are the standard errors of the allometric intercept and the allometric exponent estimates (hereafter SMRa and SMRb, respectively). Dowd et al. (2006) also determined the routine meta- bolic rate (the average oxygen consumption rate of a swimming shark) for 15 individual sandbar sharks at 24^C in an annular respirometer (diameter 1.67 m). The ratio of routine metabolic rate to SMR, corrected for the cost of swimming in a curved path in the respirometer (Weihs. 1981), averaged 1.62 ±0.11 (Dowd et al., 2006). This ratio was used in the model as a constant activ- ity multiplier (ACT) to estimate field metabolic rate (sensu Winberg, 1960; Kitchell et al., 1977; Schindler et al., 2002). The ACT used is similar to those derived from field data for subadult Negaprion hrevirostns (1.3; Sundstrbm and Gruber, 1998) and juvenile Sphyrna lewini (1.45; Lowe, 2002). The sandbar shark ACT was assumed to remain constant for all age classes and over all temperatures (Dowd et al., 2006). The effects of acute temperature changes (quantified as Qj||) on SMR for juvenile sandbar sharks (mass 1 — 10 kg) between 18" and 28°C have also been measured (Dowd et al., 2006). The overall mean Qj,, (the relative increase in metabolic rate with temperature, scaled to a 10° temperature range) was 2.89 ±0.16 (?! = 43), was consistent over the size range of sharks tested, and was statistically indistinguishable among three treat- ments (18-24°C, 24-28°C, and 18-28°C). We assumed that the SMR Qjq remained constant throughout the simulation period. For each day of the simulation, the Q,q was used to adjust the predicted SMR from Equation 3 to the simulated daily temperature (T) (equation adapted from Schmidt-Nielsen, 1997): SMRt, = W logSMB.,4+logQi ' 10 J (4) SMR„. = 120.0 (±17.3)M0 788 (±o.o76). (3) SMRj, was then multiplied by the ACT and by 24 hours to obtain the daily metabolic expenditure in mgO^/day. Finally, this value was converted to daily metabolic energy utilization (RMR^) by using the oxycalorific coef- ficient 13.59 J/mgOj (Elliott and Davison, 1975). Specific dynamic action (SDA) Specific dynamic action represents the energetic cost of incorporation of digested amino acids into new proteins (Brown and Cameron, 1991). Although SDA varies with growth rate, or the protein content of ingested food (e.g., Ross et al., 1992), most bioenergetics models set SDA as a constant fraction of consumed energy (e.g., Hewett and Johnson, 1992). Fortunately, although SDA has been measured in only a few elasmobranch species, it is typically a relatively small fraction of consumed energy (DuPreez et al., 1988; Sims and Davies, 1994; Duffy, 1999; Ferry-Graham and Gibb, 2001). As an initial estimate, we assumed SDA to be 10% of consumed energy (Schindler et al., 2002). Dowd et a\ Consumption rates of Corcharchinus plumbeus in Chesapeake Bay 335 Growth (G) Growth (G) is the change in energy stored in biomass and can be subdivided into somatic and reproductive growth outputs. We assumed the latter to be negligible because all the age classes in the sandbar shark bioenergetics model are at least 8 years from the age at maturity (Casey et al., 1985; Sminkey and Musick, 1995). We employed a von Bertalanffy growth equation (Sminkey and Musick, 1995), based on a validated ag- ing technique for sandbar sharks (Branstetter, 1987), to represent the precaudal length (PCL) of sharks of age jy (y=0-5 yr) upon immigration (or birth) on May 15: l,„ = l4i-.-^-'-"') (5) where L^ = 164 cm; K = 0.089; and fi, = -3.8 years. The PCL at emigration (L^g) was determined by (6) where p = the proportion of annual growth in PCL that occurs in the Chesapeake Bay nursery. Analysis of vertebral rings indicates that annual growth of juvenile sandbar sharks occurs in two distinct phases: one period of rapid growth in the summer nurseries during which the sharks achieve roughly 75% of their annual growth in length, followed by a period of reduced somatic growth during the winter (Sminkey and Musick, 1995). Therefore, we assumed a p of 0.75 as an initial estimate. Limited tag-return data support this seasonal growth pattern. One juvenile (67 cm total length [TL] at tagging) was recaptured 0.5 km from the tagging loca- tion within the summer nursery in September 1998 by VIMS scientists; it had grown 3 cm TL after 44 days at liberty. Similarly, a juvenile sandbar shark of similar size that had been tagged and recaptured by NMFS scientists grew 3 cm in fork length (FL) (48-51 cm FL) over 62 days at liberty between mid-July and mid-Sep- tember (Casey et al., 1985). In Delaware Bay, two sand- bar sharks recaptured during the same summer grew 3 cm FL (45 cm flat tagging and 1 cm FL) (no size given) in 40 and 47 days at liberty, respectively (Merson and Pratt, 2001). In comparison, another juvenile (66 cm TL) was tagged in Chesapeake Bay in September 1995 and recaptured by VIMS scientists during the subsequent immigration period. This shark was at liberty for 225 days and grew only 3.5 cm TL during that time. Both Medved et al. (1988) and Kohler et al. (1995) published equations relating mass to length for sandbar sharks. Because preliminary runs of the model dem- onstrated that these length-mass relationships yielded very similar results, we used the equation produced by Kohler et al. (1995) because it was derived from a larger number of individuals: Fork length (FL) is in centimeters and mass (M) is in grams. Lengths were converted from PCL to FL and vice versa by using the regression (VIMS^): FL = 1.0791 PCL + 2.78. (« = 4385; r2 = 0.99) (8) Specific growth rate (grams added per gram of body mass per day) was modeled by assuming that the mass of the shark increased by a constant proportion (x) in each of the n days of the simulation: 0=1 (9) M = 0.0109 FL3 oi24_ (7) M,j is the mass of the shark at the beginning of day D. No data exist to support an alternative pattern (e.g., growth varying with temperature or dissolved oxygen levels). The mass of the shark on the first and last day (My and Ml,-, respectively) of the simulated nursery season was determined by using Equations 5-8. Fitted val- ues for X in Equation 9 were on the order of 0.1-0.5% increases in mass per day. We used these values to calculate daily growth increments in grams per day and then multiplied by 5400 J/g of body mass (Cortes and Gruber, 1990; Lowe, 2002) to determine the daily increase in energy content. Waste loss in feces (f ) and excretions (L/) A generally accepted value for total waste loss to excretions and fecal waste for carnivorous fishes and elasmobranchs is 27 ±3% of consumed energy (C) (Brett and Groves, 1979; e.g., Sundstrom and Gruber, 1998; Lowe, 2002; Schindler et al., 2002). This value was assumed for the sandbar shark in the present study, divided into F=0.20C and t/=0.07C. Juvenile A^. brevirostris have fecal waste losses between 38.1% and 16.9% (Wetherbee and Gruber, 1993), and excretory losses average 7% of ingested energy for a number of teleosts (Brett and Groves, 1979). Water temperature data Surface and bottom water temperatures were obtained from the Chesapeake Bay Program's water quality database- for seven monitoring stations within the core sandbar shark nursery area in Chesapeake Bay for 1996-2002. Temperature measure- ments were averaged over all stations and over all years for each day of the simulation. The surface and bottom temperature readings were also averaged to obtain a mean water temperature for each day of the simulation in an average year. The simulated temperatures ranged from 16.8° to 27.9°C over the summer nursery season (mean 23.0' ±0.2°C). Diet composition data Recent data detail the ontoge- netic patterns of juvenile sandbar shark diet composition in and around Chesapeake Bay for sharks captured with longline and gillnet gears (Ellis, 2003). Diet data are represented by the index of relative importance. Index of relative importance combines the frequency, weight, and number of each prey type and is considered to have 336 Fishery Bulletin 104(3) Table 2 Diet composition data for juvenile sandbar sharks (Carcharhinu^ plumbeus) used to estimate daily rations and seasonal prey con- sumption. Prey species were grouped into four categories for each age class. Diet data, adapted from Ellis (2003), are expressed as index of relative importance. The average energetic content (J/g wet mass! of each prey type was calculated from data in Thayer etal. (1973). Category Representative species Ages 0- 1 Ages 2-3 Ages 4-5 Energy density (J/g) Teleostei Atlantic menhaden (Brevoortia tyrannus) 0.146 Summer flounder iParalicbthys dentatus) 0.292 0.463 5050 MoUusca Squids (Lo/igo spp.) 0.007 0.004 0.023 4390 Crustacea Blue crab iCallinectes sapidus) Mantis shrimp (Sqiiilla empiisa ) 0.847 0.672 0.421 4810 Elasmobranchii primarily skates {Raja spp. ) — 0.031 0.094 5400 Table 3 Cohort sizes and estimated mean seasonal prey consumption in the lower Che {Carcharhinus plumbeus) bioenergetics model. Cohort sizes are mean ±SE. sapeake Bay foi each age class in the sandbar shark Age class Initial cohort size' Indexed cohort size- Seasonal prey consumption (kg)'* Teleostei Mollusca Crustacea Elasmobranchii Total 0 2545 ±216 4377 ±1074 4236 207 24,667 — 29,110 1 2122 ±284 2626 ±645 3634 178 21,157 — 24,969 2 2083 ±398 1837±451 6684 100 15,385 716 22,885 3 1698 ±417 1698 ±417 7757 115 17,855 831 26,558 4 900 ±184 900+184 7754 380 7053 1575 16,762 5 188 ±40 188 ±40 1900 93 1728 386 4,107 Total 9537 ±313 11,627 ±2483 31,965 1073 87,844 3,508 124,391 ' Estimates are from Sminkey ( 1994). - We retained the initial cohort size estimates for ages 3-5. ■' Estimated by using mean indexed cohort size. less bias than other diet indices (Cortes, 1997). For the present study, prey species were grouped into four categories for each age class of shark: teleost fishes, mollusks, crustaceans, and elasmobranchs (Table 2). The proportion of each prey type in the diet and the mean energy content values for each category (calculated from data in Thayer et al., 1973) were used to convert daily energy ration (kj/d) to daily ration (percent body mass per day, '7(BM/d). Diet composition was assumed to remain constant during the simulation period. The average daily ration and total seasonal prey consumption were calculated for individuals of each age class. Population estimates The relative abundance and size- class composition of the seasonal nursery population were estimated from catch per unit of effort ( CPUE ) data (Musick et al., 1993; VIMSM. Sminkey (1994) used vir- tual population analysis to estimate the sandbar shark cohort sizes in the Chesapeake Bay nursery from the VIMS Shark Longline Survey data, using the standard Mustad"' 9/0 J hooks between 1989 and 1993 (Table 3). However, the standard hooks select for larger animals, yielding underestimates of abundance for ages 0-2 years. Therefore, we indexed the VIMS CPUE data for ages 0-2, using smaller Mustad"' 12/0 circle hooks against the CPUE for larger hooks for 25 longline sets between 1997 and 2002 when both gears were fished simultaneously at the two lower Chesapeake Bay survey stations. We then used this index to produce a more realistic population age structure (Table 3). The mean adjusted nursery popula- tion size was 11,627 ±2483 individuals. For simplicity, we assumed negligible mortality and zero emigration of juvenile sharks during the simula- tion period. Consequently, the revised cohort sizes were held constant throughout the simulation period. Low natural mortality rates would be expected for these sharks, particularly in light of the near absence of large coastal shark predators in the nursery (Musick et al., 1993). Tracking, tagging, and survey data all indicate that juvenile sandbar sharks remain within the nursery throughout the summer (Grubbs et al., in press; Merson and Pratt, 2001). Dowd et al Consumption rates of Carcharchmus plumbeus in Chesapeake Bay 337 Model calculations For each daily time step of the model and for each age class, RMRp and G^ were calculated as described above. These estimates were used to solve for daily consump- tion in joules in Equation 2, where SDA, U, and F are the fractions of consumption described above. These daily energy consumption estimates were summed to determine total energy consumption for an average individual of each age class during the entire stay in the Chesapeake Bay nursery. Mean daily energy ration (DERl was calculated in kJ/d. The daily energy ration was also expressed as a percentage of the average total energy content (%DER) for each day: %DER = 100 Cd (10) ' Mg+M^^^ 5400 I Finally, gross conversion efficiency (i^T, ), the fraction of consumed energy that is devoted to growth, was cal- culated for each day: (11) This value was used as a general test of the model outputs. Error analysis Static models were run by using the initial parameter estimates described above to determine point estimates of consumption. SDA and energy losses in U and F were modeled as constant fractions of consumption. The initial choices of these values, therefore, had a direct effect on the predicted consumption rates. Further, a number of the model parameters were measured with some uncertainty. A stochastic, Monte Carlo simulation routine (Crystal Ball'- 2000 Academic Edition, vers. 5.2.2, Decisioneering, Inc., Denver, CO) was used to assess this uncertainty with error analysis (Bartell et al., 1986). Error analysis is particularly useful for evalu- ating model sensitivity to parameters that enter the model in a nonlinear fashion (Bartell et al., 1986), such as the SMR allometric exponent (SMRb) and allometric constant (SMRa) and the Qj,,. The simulation randomly drew values from probability distributions for each model parameter (Table 1) for each of the 2000 Monte Carlo iterations. The model parameters were ranked in impor- tance by their relative contribution to the variance of the stochastic model outputs (Bartell et al., 1986). Results Consumption rates The model predicted mean daily energy rations (DER) increasing from 233 ±5 kJ/d (%DER = 1.95 ±0.03%) for young-of-the-year to 784 ±16 kJ/d (%DER = 1.20 Table 4 Gross conversion efficiency (Kj), daily energy ration (DER), daily ration (DR), and total seasonal prey con- sumption (C(^,) for individuals of each age-class of the sandbar shark iCarcharhinus plumbeus) in the bioener- getics model. DER and DR were averaged over the 138 days of the simulation (mean ±SEl. Age class K, DER(kJ/d) DR(7rBM/di Ct„t(kg) 0.16 233 ±5 2.17 ±0.03 6.6 0.15 333 ±7 1.89 ±0.03 9.5 0.13 442 ±9 1.67 ±0.03 12.5 0.12 555 ±11 1.52 ±0.03 15.6 0.11 669 ±14 1.39 ±0.02 18.6 0.10 784 ±16 1.30 ±0.02 21.8 ±0.02%) for an age-5 juvenile. These values correspond to prey consumption rates of 2.17 ±0.03%BM/d and 1.30 ±0.02%BM/d, respectively (Table 4). The predicted daily rations for a given age class over the course of the simulation period fluctuated with temperature because of the thermal influence on metabolic rate. During the 4.5-month stay in the Chesapeake Bay nursery area, the static model predicted total energy con- sumption of 269% of the total energy content for an age-0 shark (-32,000 kJ), declining to 165% (-108,000 kJ) for age-5 sharks. When merged with diet composition data, the model predicted that an age-0 shark would consume 6.6 kg (300% average BM) of prey per summer, and an age-5 juvenile would consume 21.8 kg (180% average BM). Therefore, the total sandbar shark population would consume 124,400 kg of prey over the course of the sum- mer in the Chesapeake Bay nursery area (Table 3). The average Kj declined quickly with age from 16.3 ±0.3% of consumed energy for age-0 sharks to 10.0 ±0.2% of consumed energy by age five. Because growth plus rou- tine metabolism comprised a constant proportion of the total energy budget in the static model, the proportion of consumption devoted to metabolism increased with age. Metabolism for age-0 sandbar sharks accounted for roughly 46% of ingested energy, increasing to 53% of the energy budget for age-5 juveniles. When growth was set to zero, we calculated the maintenance rations to be 63-80% of the rations when growth was included. Error analysis The relative contributions of each of the input param- eters to the variance of the model outputs exhibited similar patterns for all age classes (Fig. 1). The von Bertalanffy parameters predicting size at age (L^, K) had consistently high ranks for their contribution to model variance, as did those describing the allometric scaling of standard metabolic rate {SMRa, SMRb). F also contributed significantly to the variance of the model outputs for all age classes (Fig. 1). The contributions 338 Fishery Bulletin 104(3) u- AgeO F ■ SMRb ■ ^ SMRa- ■J O10 ■ P ■ 1 1 °„RMri-' 1 K ■ I ■ Linf ■ ca ACT ■ 1 r- -10 0 10 20 30 40 50 -20 -10 0 10 20 30 40 50 u - Age 2 F - 1 1 SMRa - SDA ■ Q10 ■ P ■ to - K - 1 P Z3 ^ Lint - 1 ACT - ■ Age 3 • 10 0 10 20 30 40 50 -20 10 0 10 20 30 40 50 u- ■ a Age 4 F- ^ ' 1 W^^^^^^M SDA ■ Q10 - P - to- K ■ P I H^l^^^^l Linf ■ □ ACT ■ — 1 1 1 1 1 c Age 5 ^m — ' 1 f ^^^^^^M d 1 1 1 1 1 1 ■20 -10 0 10 20 30 40 50 -20 -10 0 10 20 30 40 50 Figure 1 Results of the error analyses for the sandbar shark iCarcharhinus pluniheus) bioenergetics model for ages 0-5 years, using the eleven parameters and distributions from Table 1 in 2000 Monte Carlo simulations. The horizontal axis is the percentage contribution of the variable of interest to the variance in two model predictions: total seasonal prey consumption in kg ^C^^^, black bars) and mean daily ration (()% maturity (MLjq) of 352 mm, which is greater than the previously estimated ML^p at southern latitudes. The maximum age of collected female rex sole was 29 years, and the estimated age at 50% maturity (MAjg) in the GOA was 5.1 years. The von Bertalanffy growth model for rex sole in the GOA was significantly different from the previ- ously estimated model for rex sole off the Oregon coast. This study indicated that there are higher growth rates for rex sole in the GOA than off the Oregon coast and that there are dif- ferences in length at maturity and similarity in age at maturity between the two regions. Reproductive biology, spawning season, and growth of female rex sole (Glyptocephalus zachirus) in the Gulf of Alaska Alisa A. Abookire Kodiak Laboratory Alaska Fisheries Science Center National Marine Fisheries Service 301 Research Court Kodiak, Alaska 99615 Email. alisa.abookire(Snoaa gov Manuscript submitted 5 January 2005 to the Scientific Editor's Office. Manuscript approved for publication 19 September 2005 by the Scientific Editor. Fish. 104:350-359 (20061. Rex sole (Glyptocephalus zachirus) have a wide distribution throughout the North Pacific, ranging from cen- tral Baja California to the western Bering Sea (Mecklenburg at al. 2002). The rex sole population is made up of several stocks that are managed as four separate units: the U.S. West Coast stock which includes waters off the coasts of California, Oregon, and Washington; the British Columbia stock which is managed separately by the Canadian government; the Gulf of Alaska (GOA) stock; and the Bering Sea stock (TurnockM. The U.S. West Coast stock of rex sole provided a steady and stable commercial fishery in California between 1970 and 1989, but landings began to decline in the 1990s (Quirollo and Dewees'). Rex sole are managed as part of the "other flatfish" category for both the West Coast stock (Quirollo and Dewees-) and the Bering Sea stock (Spencer et al.3). Throughout its range, the largest commercial harvest of rex sole oc- curs in the GOA where rex sole are one of the major commercial flatfish species. Since 1988, the commercial trawl fishery for flatfish in the GOA has concentrated in the central Gulf on the continental shelf and slope east of Kodiak Island (DiCosimo and KimbalH). Development of a man- agement plan for the rex sole fishery in the GOA has undergone several phases. Prior to 1990, all flatfishes in the GOA with the exception of Pacific halibut {Hippoglossus stenolepis) were managed as one assemblage. In 1990, four flatfish categories were created for stock assessment: shallow-water complex, deep-water complex (rex sole included), flathead sole {Hippoglos- soides elassodo?!), and arrowtooth flounder [Atheresthes stomias). In 1993, rex sole were removed from the deep-water complex and managed as a separate species. In 2003, biomass estimates for rex sole in the Gulf of Alaska totaled 99,950 metric tons (t), and the accept- able biological catch (ABC ) was 9466 t (Turnock et al.'^). Annual commercial landings for GOA rex sole averaged about 3000 t during 1999-2003, ap- proximately 35% of the ABC (Turnock et al.''). The low harvest level may be due, in part, to constraints placed on ' Turnock. B. J. 2004. Personal commun. Alaska Fisheries Science Center, National Marine Fisheries Service, 7600 Sand Point Way NE. Seattle, WA 98115. - Quirollo, L. F., and C. M. Dewees. 2001. California's living marine resources: a status report. California Department of Fish and Game. Website: http:// www.dfg.ca.gov/mrd/status/rex_sole.pdf {accessed on 9 September 2004]. ^ Spencer, P. D., G. E. Walters, and T. K. Wilderbuer. 2001. Stock assessment and fishery evaluation: Bering Sea other flatfish. Website: http://www.afsc. noaa.gov/refm/docs/2001/BSoflats.pdf [accessed on 13 September 2004]. ^ DiCosimo, J., and N. Kimball. 2001. Groundfish of the Gulf of Alaska: A species profile. North Pacific Fisheries Management Council Report. Website: http://www.fakr.noaa.gov/npfmc/reports/ goaspecies2001.pdf [accessed on 3 March 2003]. •-' Turnock, B. J., T. K. Wilderbuer, and E.S.Brown. 2003. Stock assessment and fishery evaluation: Gulf of Alaska flatfish. Website: http://www.afsc. noaa.gov/refm/docs/2003/GOAflats.pdf [accessed on 10 September 2004]. Abookire Biology, spawning season, and growth of Glyptocephalus zachirus 351 the trawl flatfish fishery once the allowed limit of Pacific halibut bycatch is captured (DiCosimo and KimbalH). The life history, reproductive biology, and growth of rex sole were investigated off the Oregon coast from 1969 to 1973 (Hosie and Horton, 1977). Despite the wide geographic range of rex sole, no investigations have oc- curred elsewhere. Although rex sole in the GOA are managed as a single species unit with species-specific assessments of biomass and ABC estimates, there is no information on length or age at maturity to incorporate into analytical stock assessments. Thus, there is a substantial need for obtaining informa- tion on the reproductive biology of rex sole in the GOA for the purpose of stock-specific assessment and management. To clarify the seasonal maturation dynam- ics and reproductive biology of female rex sole in the GOA, fish were collected year-round and their ovaries were analyzed by using standard histological techniques. To date, there have been no histological investigations on rex sole maturity (Castillo, 1995). Specific goals of this investigation were to determine the spawning season, length at first maturity, length at 50% maturity (MLgj,). age at first maturity, and age at 50'7f maturity (MAr,,,). Additionally, inasmuch as possible, reproduc- tive parameters and growth were compared between female rex sole in the GOA with those off the Oregon coast (data from Hosie and Horton, 1977). Reproductive parameters determined in this study will aid the develop- ment of a species-specific assessment of target fishing mortality rates and mature female biomass. yielding a more accurate management model for the GOA stock of rex sole. Materials and methods 6aN -I- : Commercial trawls + Research trawls 0 30 60 120 55 N s Collections Rex sole were collected from 14 February 2000 to 21 October 2001 in the central GOA (from 55°30'N, 156°00'W to 59°50'N, 147°15'W; Fig. 1). Approximately half (52% ) of the samples were collected at fish process- ing plants in Kodiak, Alaska. The shore-side sampling data were dependent on the timing of the fishery, but allowed for the annual cycle of sexual maturation to be monitored. Shore-side sampling occurred in January, February, April, May, October, November, and December, and samples were obtained from boats that captured rex sole as bycatch while fishing for other species. Although there is no inherent biological difference in rex sole col- lected by commercial or research trawl, typically samples from the commercial fishery exclude young fish that are below the minimum landing size or escape through the Figure 1 Collections of female rex sole (Glyptocephalus zachirus) in the Gulf of Alaska were concentrated around Kodiak Island, Alaska. Samples were obtained from the commercial trawl fishery (denoted by open circles) and from research survey trawls (denoted by crosses). meshes. However, because of the close proximity of the fishing grounds to the town of Kodiak, catches were not sorted at sea and very small rex sole (minimum total length [TL] = 172 mm) were brought back to the pro- cessing plants where they were obtained for this study. The remaining 48% of samples were collected during research surveys in February, March, June, and July by the National Marine Fisheries Service (NMFS) and the University of Alaska Fairbanks with smaller-meshed trawls in the same waters fished by the commercial fishery. Samples were stratified by length, so that up to five females per 1-cm length group were obtained each month. For each female, total length (to the nearest mm) and weight (to the nearest 0.1 gram) were mea- sured, both sagittal otoliths were removed, and ovarian maturity was macroscopically staged. Each ovary was removed and weighed separately (to the nearest 0.1 gram). Ovaries were preserved in 10%^ buffered forma- lin for 2 months and then transferred to 70%j ethanol. All ovaries collected at sea were weighed only after preservation; ovaries collected at processing plants were weighed both when fresh and after preservation. Sagittal otoliths were removed from each fish, stored in a glycerol and thymol solution, and most (n = 557) 352 Fishery Bulletin 104(3) D E F Figure 2 Photomicrographs of histological cross-sections of rex sole iGIyptocephalus zachirus) ovaries: (A) early perinuclear (EP) oocytes and late perinuclear [EP] oocytes, with a large central nucleus ). A postspawning fish has totally spent ovaries. Early Advanced MN >50% Number of Ovary Maturity Spawning Unyolked yolked yolked orHY yolked fish in activity stage condition oocytes oocytes oocytes oocytes POFs atresia stage Inactive Immature Nonspawning + O O O O - 274 Inactive Mature Nonspawning - + O O O - 44 Inactive Mature Nonspawning - - + - O + 0 Inactive Mature Postspawning - - + - + + 96 Active Mature Nonspawning - - + O o O 26 Active Mature Spawning - - - + o O 27 Inactive Mature Postspawning - - o o + - 14 Active Mature Spawning - - > > + o 113 Total 594 sification from Hunter et al. (1992). Females were first classed as active or inactive by using the following criteria. Through histological analysis, ovaries that contained a sufficient number of AY/MN/HY oocytes for one spawning were classed as active. Active females were then classed as either spawning or nonspawning. Ovaries without AY oocytes or with major atresia of AY oocytes were classed as inactive. Inactive females were first classed as nonspawning or postspawning and then inactive nonspawning females were further classed as mature or immature (Table 1). Active spawning females showed evidence of past spawning (POP present) or imminent spawning (MN or HY present), whereas active nonspawning females had no evidence of recent or immi- nent spawning but were presumed capable of spawning in the near future. Females with ovaries containing oocytes in early stages of vitellogenesis (EY oocytes present) were considered mature but inactive. Immature females had ovaries without vitellogenic oocytes. The fraction of active females and the fraction of postspawning females was calculated for each month. The start of the spawning season was defined by the first observance of a hydrated oocyte or a POF and the end of the spawning season was defined when the last female with hydrated oocytes was observed. A one- way ANOVA was used to test for differences in the ovary wall thickness of mature females among months. Assumptions of homogeneity of variances and normal distribution of observations were met for the ANOVAs. Bonferroni all-pairwise multiple comparison tests were used after the ANOVAs to test for differences among monthly mean values. Alpha was set at 0.05 for all tests of significance. To estimate the length and age at which 50% of the female rex sole were mature (ML^Qand MA^,,,), I used the logistic regression model: Pmat = 1 / (1 -H +fcL where Pmat = the fraction of mature females per 15-mm length-class; and L = total length in millimeters. Similarly, for estimating MA5o,age (in years) was substi- tuted for length in the above equation. In each case, the equation was solved for Pmat = 0.5 to obtain the length and age at 50% maturity. The equation for the upper and lower 95% confidence limits around Pmat when Pmat = 0.5 was solved to yield 95% confidence limits around ML,-|iand MA-,,. Results were compared with existing length-at-maturity data for female rex sole off the Oregon coast by reconstructing a length-at-maturity logistic regression model from data presented by Hosie and Horton (1977), taking the log transformation of both the GOA and Oregon logistic regression curves to make them linear, and then comparing the slopes of the two lines according to Zar (1999). Direct statistical comparison with existing age-at-maturity data (e.g., the age at 50% maturity, and age at 100% maturity) for female rex sole off Oregon (Hosie and Horton, 1977) was not conducted because Hosie and Horton (1977) did not present stan- dard error or logistic regression equations. Instead, the probability that a specific-size (or specific-age) female rex sole off the coast of Oregon would be considered mature in the GOA was calculated as the fraction of mature GOA females of the specified length divided by the GOA sample size of the specified length (Zar, 1999). Size and growth The relationship between weight and length for female rex sole in the GOA was estimated with the equation 354 Fishery Bulletin 104(3) W = aL'', where W = the total fish weight in grams; and L = total length in cm. This relationship was also determined for ovary-free weight. Because fresh ovarian weights were not mea- sured for samples collected at sea, the fresh weight was estimated from a linear regression between fresh and preserved weights from the samples collected at pro- cessing plants. The linear regression OW=1.127 (OWA- 0.024 was used to convert formalin-preserved ovary weights, OWr, to estimates of fresh ovary weights, OW (0=446, ^2=0.997, P<0.0001). When fresh and preserved ovary weights differed; fresh ovaries usu- ally weighed more than ovaries preserved in formalin, and the maximum difference between fresh ovaries and formalin-preserved ovaries was 20%. Parameters of the von Bertalanffy growth model (L^, k, and t^) were estimated for female rex sole in the GOA by using nonlinear least squares regression with the equation L, = L,a- where L, is length at age t. -k (/ - l„) All length and age data were used for the GOA growth model. Statistical comparison of growth rates for female rex sole in the GOA with those from the Oregon coast were less rigorous because Hosie and Horton (1977) used mean values of length at age to calculate the von Bertalanffy growth model, thereby decreasing the variability in length at age, which was not presented. Therefore, the parameters L^, k. and tg in the GOA growth model were each statistically compared with the parameters in the Oregon growth model for females age 1-15 years (data from Hosie and Horton, 1977) with three separate Z-tests (Zar, 1999). Alpha was set at 0.05 for all tests of significance. Results Maturity Oocyte development and maturity stage were deter- mined for 594 female rex sole. Histological examination revealed that 46% of the collected females were imma- ture and 54% were mature. Both immature and mature females were collected in every month; however, no fish were sampled in August or September (Table 2>. Year- round histological analysis of ovaries confirmed that rex sole are batch spawners, as seen in the presence of POFs, features associated with recent spawning, concurrent with advanced vitellogenic oocytes. The spawning season of rex sole in the GOA was protracted; postovulatory follicles first appeared in October and hydrated oocytes occurred as late as May, indicating a spawning season that spans at least eight Female rex sole TL = 287-420 mm 1.0 08 06 04 S 0-2 X 0) to E 2 10 g ^ ■ j Jan Feb Mar Apr May Jun Jul Oct Nov Dec Female rex sole TL >420 mm ill Jan Feb Mar Apr May Jun Jul Oct Nov Dec D Nonspawning ■ Poslspawmng H Spawning D Immature Figure 3 Monthly proportion of female rex sole (Gtyptocephalus zachi- rus) maturity stages graphed separately for fish greater than size at first maturity (287-420 mm in upper graph) and fish greater than size at lOC/r maturity (>420 mm in lower graph). No data were collected in August or Septem- ber. All mature rex sole were in the postspawning stage in June and Julv. months. Active females were present from October through May and were absent in June and July (Ta- ble 2). Mature postspawning females with no healthy AY oocytes first appeared in April and increased in June and July; thus, the spawning season ended in April for some individuals, and by June all rex sole had completed spawning (Fig. 3). The high number of postspawning fish combined with the absence of spawning fish in June and July indicated that the du- ration of POFs in an ovary was several weeks. Active females had at least one of the following oocytes (AY, MN, or HY) and could have any combination of these advanced stage oocytes, or all three oocyte stages at once. These oocytes were large and stretched the ovary wall quite thin, whereas inactive postspawning females had shrunken ovaries and thicker ovary walls (Mad- dock and Burton, 1999). In May and July ovary walls in mature females were significantly thicker than in all other months sampled (ANOVA: Fg .^-^ = 28.64, P<0.0001; Fig. 4), corresponding to the high fraction of postspawning females present from May to July (Table 2, Fig. 3). Female rex sole sampled for length and maturity ranged from 166 to 552 mm TL (Fig. 5). The smallest Abookire Biology, spawning season, and growth of Glyplocephalus zachirus 355 Table 2 Numbers of I'emale rex sole iGlyptocephalus zachirus) in various histological subclasses, listed by month with years combined. Abbreviations are as follows: postspawning (PS), nonspawning (NS), spawning (S), early yolked oocytes (EY), advanced yolked oocytes (AY), migratory nucleus oocytes (MNl, and hydrated oocytes (HY). Mature PS females had ovaries that contained post- ovulatory follicles (POF), and mature S females had ovaries that contained either MN or HY oocytes or AY oocytes combined with POF. The length distribution (minimum, maximum) and the total number of females sampled in each month are given. The percent active females in each month is given, and female rex sole were active from October through May. PS females were pres- ent from April to July. No collections were made in August or September. Ovary Maturity activity stage Histological criteria Jan Feb Mar Apr May Jun Jul Oct Nov Dec Total Inactive Immature Unyolked oocytes 2 7.5 33 20 13 14 36 10 41 30 274 Mature EY; <50 % atresia 0 13 10 8 2 0 0 6 1 4 44 Mature-PS AY; >50 % atresia; POF 0 0 0 9 6 37 44 0 0 0 96 Mature-PS AY absent; POF 0 0 0 1 1 0 12 0 0 0 14 Active Mature-NS AY; POF absent 0 5 2 0 0 0 0 15 4 0 26 Mature-S MN or HY; POF absent 0 0 0 0 0 0 0 23 3 1 27 Mature-S AY, MN, or HY; POF present 5 37 10 24 2 0 0 18 16 1 113 Minimum length (mmi 353 166 250 214 215 257 176 267 172 220 Maximum length (mm I 513 529 552 547 460 503 450 476 460 452 Percent active females 71% 32% 22% 39% 8% 0 0 789 35% 6% Percent postspawning females 0 0 0 16% 29% 73% 61% 0 0 0 Total number sampled 7 130 55 62 24 51 92 72 65 36 594 mature female in the GOA was 287 mm TL. The nonlinear logistic regression model used to determine a length-at-maturity curve for females in the GOA had a very good fit (n = 594, r-' = 0.994; Fig. 6). The MLgQ estimate for rex sole in the GOA was 352 mm (95% CI: 344-360 mm; Table 3), and length at 100% maturity was 420 mm (Fig. 5). Female rex sole in the GOA matured at a significantly larger size than off the Oregon coast (slope comparison: f. = 1356, P<0.0001; Fig. 6). Hosie and Horton (1977) estimated the ML-g as 24 cm off the Oregon coast, and at this length no GOA rex sole were mature. At the length that 100% of rex sole from Oregon waters were ma- ture (30 cm; Hosie and Horton, 1977), only 15.8% (n=19) were mature in the GOA. The MLr,o for female rex sole in the GOA was 46% greater than the ML,,, for Oregon coast stock. Female rex sole collected in this study ranged in age from 1 to 29 years. The estimated MAjq was 5.1 years (95% CI: 4.7-5.5 years) (/! = 557, r2=0.988; Table 3, Fig. 7). Minimum age-at-maturity was 3 years, and all females had reached maturity by age 9, with the exception of two slow growing females who were still immature at ages 9 and 10 years (Fig. 5). These results are nearly identical to those obtained in Oregon (MA.q=5 yr, 100% mature at 9 years.; Hosie and Horton, 1977). Jan Feb IVIar Apr May June July Oct Nov Dec Figure 4 Monthly mean (±1 standard error) ovarian wall thickness (mm) for mature female rex sole iGlyptocephalus zachirus). No data were collected in August or September. Size and growth Parameters for the weight (g) to total length (cm) relationship for female rex sole in the GOA were 356 Fishery Bulletin 104(3) W=0.000762L3 6127 (n=573, r-=0.963, P<0.0001). Param- eters for the ovary- free weight to total length (cm I rela- tionship were W=0.001277L3 "8° (n = 568, r2 = 0.969, P<0.0001). Parameters of the von Bertalanffy growth model for female rex sole in the GOA were L^ = 41.824, ;fe=0.388, and ?o=-0.022 (n=556, r2=0.480, P<0.0001). All GOA parameters of the von Bertalanffy growth model differed from those off the coast of Oregon: L^ (Z = 8.01, P<0.0001), k (Z=5.61, P<0.0001), and /g (Z=2.29, P=0.0110). Female rex sole age 1-15 years grew much more quickly in the GOA than off the Oregon coast (Fig. 8). Discussion A representative sample of fish can be challenging to collect during the spawning season, because samples from spawning aggregations may produce dispropor- tionately large numbers of mature fish whereas samples 45 40 35 - 30 ■ 25- 20 15 10 5 -I D Immature (274) ■ Mature (320) ", w. S. 170 215 260 305 350 395 440 485 530 Length (mm) 100 -1 90 80 70 - 60 ■ 50 - 40 30 20 10 -I 0 D Immature (256) ■ Mature (301) dU JL n. H. 23456789 10 Age (yr) 13 14 15+ Figure 5 The number of immature and mature female rex sole {Glypto- cephalus zachirus) by total length (upper graph) and age (lower graph). Sample size is given in parentheses. Lengths are binned at 15-mm increments and "age 15+" includes all females from 15 to 29 years old. Minimum size at maturity occurred in a fisb with total length of 287 mm, and minimum age at maturity occurred at 3 years. from elsewhere may yield unusually large numbers of immature fish (Sampson and Al-Jufaily, 1999). The combination of samples from commercial trawls and ran- domly chosen trawls taken on research surveys provide a reasonable representation of the GOA population and are suitable for estimating spawning season, length and age at maturity, and growth rate. Shore-side sampling of the commercial catch enabled the annual cycle of rex sole sexual maturation to be monitored, and although a common problem in obtaining samples from commercial fishing gear is that the size-selectivity of trawls may produce biased estimates of size frequency and size at age (Sampson and Al-Jufaily, 1999), this concern was alleviated because small rex sole (minimum length = 172 mm) were collected from processing plants. One limitation in the sampling effort was that sam- ples could not be obtained in August and September. This raises two concerns. First, the onset of the GOA spawning season was estimated to be October by the presence of POFs in mature ovaries; but, in all prob- ability, it may begin earlier. Second, Hunter et al. (1992) argue that length and age at ma- turity are best estimated from fish collected prior to the spawning season, because dur- ing the spawning season the ovaries of some postspawning females are reabsorbed to the extent that they are indistinguishable from those of immature females. Their estimates of Dover sole (Microstomus pacificus) length at maturity from samples taken during the spawning season were 1.6-5.7 cm higher than estimates from samples taken prior to the onset of spawning (Hunter et al., 1992). Given that GOA collections were unavailable prior to the spawning season, estimates of ML^q in the present study may have been biased high. However, because 1) all postspawning females had POFs present and 2) there were no fe- males that had atretic yolked oocytes without POFs, it is unlikely that mature females were misclassified as immature in this study. Spawning season The spawning season for the population of female rex sole in the GOA ends in May and appears to commence in October, although it may begin earlier (see above). Castillo (1995) estimated the spawning season for rex sole in the GOA lasts from April through September, which differs markedly from results in the present study which are based on year-round histological examination of ovaries. Castillo (1995) based his estimation on the seasonal occurrence of rex sole larvae near Kodiak Island (Kendall and Dunn, 1985) rather than on collections of adult rex sole in spawning condition. Because rex sole larvae attain an exceptionally large size (up to 89 mm standard length) and have a prolonged pelagic life of Abookire; Biology, spawning season, and growth of Glyptocephalus zachirus 357 Table 3 Logistic regression model for female rex sole {Glyptocephalus zachirus) in the Gulf of Alaska: final maturity thres and age at 50% maturity (ML^q and MA,;q), sample size (n ), model parameter estimates, and the model r'-. hold s, length Variable MLjij/MAj,, 95% CI n a SE h SE Model r~ Length; mm 1 351.7 Age (years! .^.1 343.7-359.7 594 13.79 0.77 -0.0392 0.0022 4.7-5.5 557 5.29 0.45 -1.0375 0.0854 0.994 0.988 at least 12 months (Pearcy et aL, 1977), the indeterminate pelagic larval phase may have confounded Castillo's (1995) estimates. Fish may adjust their spawning season to syn- chronize larval hatching date with the season that is most favorable for larval feeding (Qasim, 1956). Because primary production at northern latitudes is restricted to a short, intense period during early summer, fishes at higher latitudes often delay and shorten their breeding season as an adaptation to synchronize the feeding of fish larvae with the zooplankton bloom (refer- ences in Qasim, 1956; Castillo, 1995). How- ever, rex sole in the GOA spawn earlier in the year and for a longer duration than those off the Oregon coast, which spawn from January to June (Hosie and Horton, 1977). Hence, rex sole do not fit the typical pattern of geographic variation found in other pleuronectids, such as English sole (Pleuronectes vetulus; Kruse and Tyler, 1983; Sampson and Al-Jufaily, 1999), Dover sole (Abookire and Macewicz, 2003), and yellowtail flounder {Limanda ferruginea; Zamarro, 1991). Maturity and growth Size and age at sexual maturity are two impor- tant parameters offish stock assessment models used to estimate spawning biomass and annual catch quotas, but these parameters may vary throughout the geographic distribution of a spe- cies (Roff, 1981; Nichol, 1997; Bromley, 2000) and can be influenced by food availability (Cas- tillo, 1995; Sampson and Al-Jufaily, 1999), oceanographic conditions (Kruse and Tyler, 1983; Brodziak and Mikus, 2000), or popula- tion size (Morgan and Colbourne, 1999). Along the west coast of North America both English sole (Sampson and Al-Jufaily, 1999) and Dover sole (Brodziak and Mikus, 2000; Abookire and Macewicz, 2003) exhibit geographic variation in length at maturity and growth. There is some indication that female rex sole mature at a smaller size in the southern por- tion of their range: off San Francisco, Cali- fornia, females were fully mature at 22.8 cm (in Hosie and Horton, 1977), and off the Or- egon coast ML^„ was 24 cm (Hosie and Horton, GOA logistic regression — OR logistic regression (Hosie and Horton, 1977) 110 160 210 260 310 360 III iitiiiiiiii ri I 410 460 rill I II I II II II ill! 510 560 Total lengtti (mm) Figure 6 Proportion of female rex sole {Glyptocephalus zachirus) that were sexually mature in the Gulf of Alaska (GOA) as a function of total length (mm). Data points along the GOA curve represent the proportion of mature females in each 15-mm length-class interval. Logistic model parameter estimates are listed in Table 3. The dashed line denotes the length at which 50% of females are mature, and the actual size (mm) of ML^q is given. A maturity curve for female rex sole off the Oregon (OR) coast (from Hosie and Horton, 1977) is graphed for comparison. Female rex sole in the GOA mature at a significantly larger size than off the Oregon coast (P<0.0001l. o 10 0.8 - 0,6 - 0,4 - 0,2 - 0,0 oooooo ooooooo 20 25 30 15 Age (yr) Figure 7 Proportion of female rex sole {Glyptocephalus zachirus) in the Gulf of Alaska that were sexually mature as a function of age (years). Datapoints along the curve represent females grouped in intervals of one year. Logistic model parameter estimates are listed in Table 3. 358 Fishery Bulletin 104(3) 1977). Correspondingly, I found the ML^g in the GOA (35 cm) was 46% greater than that off the Oregon coast. Although there is a possibility that rex sole MLr,,, in the GOA was overestimated because sampling did not occur prior to the onset of spawning, the bias cited by Hunter et al. (1992) was relatively small and can not fully account for the 11 cm difference in ML^^ between the GOA and Oregon. However, some caution must be exercised when maturity is compared between the GOA and Oregon (Hosie and Horton, 1977) because about 25 years elapsed between studies and different criteria were used to classify fish as mature. Differences in the histological criteria used to define mature individuals and reliance solely on gross anatom- ical methods can lead to a misclassification of maturity and produce differences in estimates of MLjq (Hunter et al., 1992; Zimmermann, 1997). In a comparison of histological and gross anatomical methods. Hunter et al. (1992) found 1% of active female Dover sole were misclassified as inactive and 12% were visually clas- sified as having advanced yolked oocytes and believed capable of spawning, whereas histological analysis in- dicated that they were inactive and future spawning was unlikely. Likewise, macroscopic examination of arrowtooth flounder misclassified 18.4% of maturing or spent females as immature and 4.4%i of immature females as spent (Zimmermann, 1997). Unfortunately, in the only previous investigation on rex sole maturity, Hosie and Horton (1977) limited their analysis of ova- ries to macroscopic staging. However, the enormous dif- ferences in MLgQ between the GOA stock and the West Coast stock off Oregon likely stem from biological dif- ferences between stocks rather than from inconsistent methods used in the two studies. This interpretation is consistent with the dramatic difference in length at age between the two areas (Fig. 8) that was demonstrated in this study. Given that Hosie and Horton (1977) used the otolith- surface aging technique to analyze rex sole otoliths, they may have underestimated ages and thereby biased their estimates of both age at maturity and growth rates. However, the tendency to underestimate age with the otolith-surface aging technique is greater for older fish where the outer edges of the otolith are worn (Anderl"). Surface analysis of several GOA rex sole otoliths revealed that, although break and burn is the preferred technique for aging rex sole otoliths, it is pos- sible, depending on the clarity of the annular pattern, to age females less than 10 years old accurately by using otolith surfaces (AnderF). Because the ages at 50% and 100% maturity off the Oregon coast were 5 and 9 years old, respectively (Hosie and Horton. 1977). there is a lower potential for bias caused by the sur- face aging technique because fish were <10 years old. Although these different aging methods may limit comparisons of age at maturity, the similarity in age 60 50 £■ 40 - 30 20 10 •^ o o . .O . . o . - - ° e ogo — GOA growth curve o GOA data Oregon growth curve (Hosie and Horton, 1977) 10 15 Age (yr) 20 25 30 Figure 8 The von Bertalanffy growth curve with 9.5% confidence limits for female rex sole (Glyptocephalus zachirus) in the Gulf of Alaska. Parameters of the Gulf of Alaska growth model were L^ = 41.82, k = 0.388, and ?„ = -0.022. For comparison, the growth curve for age 1-15 female re.x sole off the Oregon coast is graphed with data from Hosie and Horton (1977). Anderl, D. 2004. Personal commun. Alaska Fisheries Science Center, National Marine Fisheries Service, 7600 Sand Point Way NE. Seattle, WA 98115. at maturity between the GOA and Oregon is strongly indicated. Additionally, because the length-at-age data presented in growth curves from Oregon were only for rex sole <15 years old (Hosie and Horton, 1977), and a difference in growth is evident at young ages (Fig. 8), it seems unlikely that potential bias from otolith-surface aging methods — rather than from biological differences between stocks — could account for the markedly lower growth rate of Oregon rex sole. To be sure, it would be useful for future research to compare the reproductive biology and growth data of female rex sole in the GOA with a current study off the Oregon coast that employed similar aging techniques and histological criteria for maturity to determine if the distinct differences found in the GOA stock (i.e., larger size at maturity, similar age at maturity, faster growth rate) persist. Acknowledgments Sampling was a cooperative effort and fish were collected with the assistance of National Marine Fisheries Ser- vice, University of Alaska Fairbanks School of Fisher- ies and Ocean Sciences, and Kodiak Island processing plants (Trident Seafoods, Alaska Pacific Seafood, and Cook Inlet Processing). I especially thank A. Barns, D. Benjamin, E. Brown, H. Emberton, R. Foy, B. Hol- laday, and N. Raring. I am grateful for the histology advice from D. Nichol and B. Macewicz, and statistical advice from M. Litzow. I thank E. Acuna for preparing slides, and I thank D. Anderl, J. Lyons, and R. Katona Abookire Biology, spawning season, and growth of Glyptocephalus zachirus 359 from the Age and Growth Unit of the Alaska Fisheries Science Center for aging otoliths. I thank M. Litzow, D. Nichol, D. Somerton, B.J. Turnock, M. Wilkins, and two anonymous reviewers for providing critical reviews of the manuscript. Project funding was provided by the Resource Assessment and Conservation Engineering Division of the Alaska Fisheries Science Center, National Oceanic and Atmospheric Administration. Literature cited Abookire, A. A., and B. J. Macewicz. 2003. Latitudinal variation in reproductive biology and growth of female Dover sole ^Microatomus pacificus) in the North Pacific, with emphasis on the Gulf of Alaska stock. J. Sea Res. 50:187-197. Beamish, R. J., and D. E. Chilton. 1982. Preliminary evaluation of a method to determine the age of sablefish {Anoplopoma fiiuhna). Can. J. Fish, Aquat. Sci. 39:277-287. Brodziak, J., and R. Mikus, 2000. Variation in life history parameters of Dover sole. Microstomas pacificus, off the coasts of Wash- ington, Oregon, and northern California. Fish. Bull. 98:661-673. Bromley, P. J. 2000. Growth, se.xual maturation and spawning in cen- tral North Sea plaice tPleuronectes platessa L.), and the generation of maturity ogives from commercial catch data. J. Sea Res. 44:27-43. Castillo, G. C. 1995. Latitudinal patterns in reproductive life history traits of northeast Pacific flatfish, //) Final proceedings of the international symposium on North Pacific flatfish, p. 51-72. Alaska Sea Grant College Program Rep. no. 95-04, Univ. Alaska Fairbanks. Fairbanks, AK. Hosie, M, J., and H. F. Horton. 1977. Biology of the rex sole, Glyptocephalus zachirus, in waters off Oregon. Fish. Bull. 75:51-60. Hunter, J, R., B, J. Macewicz, N. C. Lo, and C. A. Kimbrell. 1992. Fecundity, spawning, and maturity of female Dover sole Microstomus pacificus. with an evaluation Of assump- tions and precision. Fish. Bull. 90:101-128. Kendall, A. W., and J. R. Dunn. 1985. Ichthyoplankton of the continental shelf near Kodiak Island, Alaska. NOAA Tech. Rep. NMFS 20, 89 p. Kruse, G. H., and A. J. Tyler, 1983, Simulation of temperature and upwelling effects on the English sole iParophrys vetulus) spawning season. Can. J. Fish. Aquat. Sci. 40:230-237. Lasker, R. (editor) 1985. An egg production method for estimating spawn- ing biomass of pelagic fish: Application to the northern anchovy, Engraulis nun-clax. NOAA Tech. Rep. NMFS 36, 89 p. Maddock, D. M.. and M. P. M. Burton. 1999. Gross and histological observations of ovarian development and related condition changes in American plaice. J. Fish Biol. 53:928-944. Mecklenburg, C. W., T. A. Mecklenburg, and L. K. Thorsteinson. 2002. Fishesof Alaska, 848 p. Am, Fish, Soc, Bethesda, MD, Morgan, M. J., and E. B. Colbourne. 1999. Variation in maturity-at-age and size in three populations of American plaice, ICES J, Mar. Sci. 56:673-688. Nichol, D. G. 1997. Effects of geography and bathymetry on growth and maturity of yellowfin sole, Pteuronectes asper. in the eastern Bering Sea. Fish. Bull. 95:494-503. Pearcy, W. G., M. J. Hosie, and S. L. Richardson. 1977. Distribution and duration of pelagic life of larvae of Dover sole, Microstomus pacificus: rex sole, Glypto- cephalus zachirus; and petrale sole, Eopsetta jordani, in waters off Oregon. Fish. Bull. 75:173-183. Qasim, S, Z, 1956, Time and duration of the spawning season in some marine teleosts in relation to their distribution. J. Cons. Perm. Int. Explor. Mer 21:144-155. Roff D. A. 1981. Reproductive uncertainty and the evolution of iteroparity: why don't flatfish put all their eggs in one basket? Can. J. Fish. Aquat. Sci. 38:968-977. Sampson, D. B., and S. M. Al-Jufaily. 1999. Geographic variation in the maturity and growth schedules of English sole along the U.S. west coast. J. Fish Biol. 54:1-17. Zamarro, J. 1991. Batch fecundity and spawning frequency of yel- lowtail flounder iLimanda ferruginea) on the Grand Bank. NAFO Sci. Coun. Studies 15:43-51. Zar, J. H. 1999. Biostatistical analysis, 4"^ ed., 663 p. Prentice- Hall, Inc., Upper Saddle River, NJ. Zimmermann, M, 1997. Maturity and fecundity of arrowtooth flounder, Atheresthes stomias. from the Gulf of Alaska, Fish. Bull. 95:598-611. 360 Abstract — The contribution of the no- take marine reserve at Apo Island, Philippines, to local fishery yield through "spillover" (net export of adult fish) was estimated. Spatial patterns of fishing effort, yield, and catch rates around Apo Island were documented daily in 20C3-2004. Catch rates were higher near the reserve (by a factor of 1.1 to 2.0), but fishing effort was often lowest there. Higher catch rates near the reserve were more likely due to spillover than to low fishing intensity. Lower fishing effort near the reserve may have been due to 1) weather patterns, 2) tra- ditional importance of other fishing grounds, 3) high variability in catch rates, 4) lower market value of target species, and 5) social pressures. The yield taken near the reserve was only 10% of the total yield, but the actual spillover contribution was probably much less than this. This study is one of the few to estimate the spillover contribution to overall yield and to document the responses of fishermen to spillover. How much does the fishery at Apo Island benefit from spillover of adult fish from the adjacent marine reserve? Rene A. Abesamis'- ^ Angel C. Alcala^ Garry R. Russ^ ' School of Marine Biology and Aquaculture James Cook University Townsville, Queensland 4811, Australia E-mail address (for G R Russ, contact auttior) garry russm'icu edu au ^ Silliman University Angelo King Center for Research and Environmental Management Silliman University Dumaguete City 6200, Philippines Manuscript submitted 5 November 2004 to the Scientific Editor's Office. Manuscript approved for publication 20 September 2005 by the Scientific Editor. Fish. Bull. 104:360-375 (2006). No-take marine reserves (areas closed to fishing) are now often established with an objective of sustaining or enhancing fisheries (Gell and Roberts, 2003). Reserves could achieve such objectives by eventually becoming net exporters of adult biomass, defined as "spillover," and by providing net larval export, known as the "recruitment effect" (Russ, 2002). Fisheries will probably benefit from spillover in a minor way only. Theoretical studies have indicated that potential fishery yield (yield per recruit) from spillover would be insignificant, if any yield at all, except when stock abundance out- side reserves is very low due to high fishing mortality (Polacheck, 1990; Russ et al., 1992; DeMartini, 1993). Thus, potential gains from spillover are predicted to be only moderate. There seems to be general agree- ment that the more important fishery enhancement effect of reserves would be due to net larval export (Carr and Reed, 1993; Russ, 2002). However, the establishment of reserves, particu- larly in developing countries, requires strong support from local stakehold- ers, especially fishermen (Russ and Alcala, 1996; Galal et al., 2002). There is a concern that, in some cases, support for reserves may be difficult to obtain on the pretext of enhanced fisheries through increased recruit- ment (Russ and Alcala, 1996). Net larval export may seem less convinc- ing to fishermen because of the broad spatial scale (tens to hundreds of kilo- meters) at which it will probably occur (Russ and Alcala, 1996; Russ, 2002). Yields from spillover, although prob- ably small, may play a critical role in convincing fishermen to support establishment and maintenance of reserves (Russ and Alcala, 1996). For some fishermen, adult fish "spilling- over" from reserves, especially in the case of larger fish, will appear to be a more direct and tangible benefit than larvae recruiting to fishing grounds from distant reserves. Thus, spillover may have a substantial positive psy- chological effect on the attitudes of fishermen toward reserves. However, few empirical studies have quantified the effects of spillover on fishery yields, nor the responses of fishermen to spillover, perhaps be- cause some advocates of reserves are sometimes overly optimistic about the potential benefits of spillover. Two studies in one location in Ke- nya showed that after several years of reserve protection, spillover was not enough to compensate for reduc- tion of total yield due to the creation of a large no-take reserve (McClana- han and Kaunda-Arara, 1996; Mc- Clanahan and Mangi, 2000). The reserve, Mombasa Marine Park, took away 50-60% (-6-8 km^) of the total fishing area. In this case the reserve probably occupied too large an area to supplement total fishery yield sub- stantially (McClanahan and Mangi, 2000). However, in other countries, no-take reserves as large as Mombasa Abesamis et al How much does the fishery at Apo Island benefit from spillover from a marine reserve' 361 Marine Park are rare. In the Philippines, for example, reserves that are created with fishery enhancement as a major goal are typically small (<1 km- of reef area) and occupy £25% of the available local fishing area (Alino et al., 2002). To date, only two studies of Philippine reserves, Sumilon and Apo (-25% and -10% respectively of fishing area are no-take reserves), have shown persuasively that spillover yield may affect to- tal fishery yield (Alcala and Russ, 1990; Russ et al., 2004, Alcala et al., 2005). In both cases, however, the magnitude of actual spillover yield was unclear (Russ et al., 2004, Alcala et al., 2005). On the other hand, two studies, one in St. Lucia (Roberts et al., 2001) and one in Egypt (Galal et al., 2002), demonstrated that catch rates of fishermen increased several years after the creation of networks of reserves. It was argued in these studies that the reserves increased total fishery yield because catch rates improved but fishing effort remained constant. However, these studies provided no information on total fishery yield, precluding estimation of the magnitude of spillover benefits. More empirical studies are needed to provide assessment of potential spillover effects on adjacent fisheries. Besides available fishing area, the total fishery yield for a given location will be determined by fishing in- tensity (Hilborn and Walters, 1992). In addition, theo- retical studies indicate that spillover yield will be a function of reserve size, fishing mortality rate, and demographic parameters of target species, particularly movement rates (Polacheck, 1990; Russ et al. 1992; DeMartini, 1993). However, the actual contribution of spillover to total yield will be determined by fish- ing intensity adjacent to reserves. Indeed it has been suggested that the "first" sign of spillover is the sight of fishermen fishing close to reserve boundaries (Gell and Roberts, 2003). This phenomenon would occur pre- sumably in response to higher catch rates near than far from reserves. Spillover is predicted to produce a pattern of higher abundance of target species outside but close to reserve boundaries, but lower abundance farther away (Rakitin and Kramer, 1996; Kramer and Chapman, 1999). Decreasing catch rates (an index of abundance) away from reserve boundaries have been demonstrated by experimental trap fishing outside one reserve in Barbados and one reserve in Kenya (Rakitin and Kramer, 1996; McClanahan and Mangi, 2000). Also, monitoring studies of reserves in Kenya and the Philippines have shown that catch rates of fishermen were higher closer to reserve boundaries than farther away (McClanahan and Kaunda-Arara, 1996; McClana- han and Mangi, 2000; Russ et al., 2003, 2004). In New Zealand, large catches of lobsters were found to be more common closer to Leigh Marine Reserve than farther away (Kelly et al., 2002). However, it is less clear if fishermen tend to concentrate more effort adjacent to reserves when spillover may be present. Some studies suggest that fishermen may do so (McClanahan and Kaunda-Arara. 1996; McClanahan and Mangi, 2000), whereas other studies indicate that fishermen do not (Russ et al., 2003; Wilcox and Pomeroy, 2003). Some of the evidence for fishermen preferentially fishing near reserves is anecdotal (Gell and Roberts^. Nevertheless, it is important to stress that intense fishing near the reserve may have the effect of eventually reducing catch rates there (McClanahan and Mangi, 2000). Fishing effort may not necessarily track the spa- tial distribution of fish abundance. Fishermen may not favor fishing adjacent to reserves even if catch rates there are higher. Catch rates alone may not explain the spatial distribution of fishing effort, because deci- sions by fishermen on where to fish are usually aimed at making a profit. Hence, decisions may be influenced by fishing costs, such as fuel or time used traveling to fishing areas (Hilborn and Walters, 1992). These costs will be related to the distance of fishing grounds from home ports or residences of fishermen (e.g., Wilcox and Pomeroy, 2003). If such costs are negligible, then fish- ing effort may reflect the spatial pattern of catch rates, provided that fishermen have ample information on the latter. However, this situation may be unlikely if fishermen keep information about productive areas to themselves. Furthermore, the strategies of fishermen may become highly complex in fisheries that employ a variety of fishing gears and target a multitude of species (Hilborn and Walters, 1992). This is likely to be particularly true of coral reef fisheries in develop- ing countries (Munro, 1996). Spatial distribution of fishing effort in such fisheries may also depend upon the differential value of target species. In addition, other factors that are unrelated to income may also influence the spatial pattern of fishing effort. These include weather conditions and social factors, such as local traditions or agreements among stakeholders and managers (Wilcox and Pomeroy, 2003). Consideration of the behavior of fishermen in relation to reserves may help determine if reserves are achieving their goal of improving fishery yields through spillover. Successful use of reserves to enhance fisheries would require a case by case understanding of the spatial structure of impacted fisheries, ecosystems, and human communities (Hilborn et al., 2004). The objective of this study was to estimate the upper limit of the spillover contribution of the no-take reserve at Apo Island, Philippines to the local fishery yield. The reserve at Apo Island has been protected for over 20 years (since 1982), and there is evidence to suggest that spillover is present (Russ and Alcala, 1996; Russ et al., 2003, 2004; Alcala et al., 2005). However, previous studies indicate that access by fishermen to productive areas far from the reserve during favorable weather influences the local fishery yield considerably (White and Savina, 1987; Bellwood, 1988). In the present study, daily fishing effort and yield were documented for eight months covering two monsoonal seasons. Spatial and temporal patterns of fishing effort, yield (biomass and value), and catch rates were examined among fishing 1 Gell, F., and C. M. Roberts. 2002. Unpubl. report. The fishery effects of marine reserves and fishery closures, 89 p. WWF-US, 1250 24'h Street. NW, Washington D.C. 20037. 362 Fishery Bulletin 104(3) grounds around Apo island and within a few hun- dred meters of reserve boundaries. Materials and methods Study site Apo Island (9°4'N, 123°17'E) is located in the cen- tral Philippines, about 7 km southeast of the large island of Negros (Fig. 1). It is a small volcanic island with a fringing reef area of 1.06 km- to the 60-m isobath (0.54 km^ to the 20-m isobath). The island is inhabited by a community of about 700 perma- nent residents. The traditional source of income for the community is fisheries. At present, about 100 residents are full-time or part-time fishermen who use hook and line, gill nets, spear guns, and fish traps, and target at least 60 species of reef fishes, nonreef fishes, and some invertebrates. The major- ity of local fishermen sell their catch to resident fish buyers on Apo Island. Fish buyers, in turn, bring the catch to Negros to sell in Malatapay town or Dumaguete City. In 1982, a -450 m long no-take reserve (sanctu- ary) was established informally by the local com- munity on the southeastern side of Apo Island (Fig. 1). This reserve occupies approximately 10% of the fringing reef area to the 60-m isobath (about 13% to the 20-m isobath). The local community has ef- fectively enforced protection of the no-take reserve since 1982 (Russ and Alcala, 1999). In 1985-86, the community formally approved a marine manage- ment plan that incorporated the no-take reserve (Russ and Alcala, 1999; White et al., 2002). The management plan also prohibited destructive fish- ing methods (e.g., dynamite fishing and muro-ami drive-net fishing) and spear fishing with SCUBA. Compliance by local fishermen and visitors with these regulations has been generally good. Develop- ment of local tourism enterprises has been encour- aged by the Marine Management Plan. In the early to mid-1990s, two small resorts that could accommodate recreational SCUBA divers were established on Apo Island. Recently, the local community implemented col- lection of fees for diving in the no-take reserve and at dive sites around the island. Apo Island is influenced by a northerly mainstream current that is present for most of the year (Fig. 1). This current presumably carries the food supporting planktivorous fishes (e.g., Acanthuridae and Caesionidae) that are abundant on the northern side of the island. Predatory fishes such as Carangidae are also common in this area. Apo Island is exposed to both the NE and SW monsoons (Fig. 1). Local fishermen tend to fish the northern side of the island during the SW monsoon (June to September) and interim calm months (April, May, and October; Bellwood, 1988). However, fishing intensity on the northern side of the island may be reduced during the NE monsoon (November to March; Bellwood, 1988). Philippines mainstream .^ ^ 0 I — I 1 1 — I — I 500m "A" Village/Fisti landing site 123"17'E Figure 1 Apo Island, central Philippines, showing the no-take reserve (shaded). Names of fishing grounds around the island are italicized. Fishing grounds were grouped into "northern fish- ing grounds" (N), "western fishing grounds" (W), and "fishing grounds near Apo Reserve" (NR). The reef area of each of the fishing grounds is outlined (0-20 m isobath). Within the fishing grounds near Apo Reserve, the approximate positions of marker buoys used to indicate distance from either the northern or southern boundary of the reserve are shown. The major fish landing sites were at Baybay, Ubos, and Cogon villages (stars). Large arrows indicate the direction of the monsoons. Smaller arrows indicate direction of the mainstream current. Data collection The catches of fishermen residing at Apo Island were recorded daily from 22 July 2003 to 29 February 2004. Records were kept by three fish buyers, one buyer at each of the three major fish landing sites (the houses of the fish buyers) on the island, in Baybay, Ubos, and Cogon villages (Fig. 1). The fish buyers recorded the local names and weights of each species or family offish that contributed to the catch sold by each fishermen. Catches were weighed on market scales accurate to 0.1 kg. The value of the catch sold by fishermen was estimated by multiplying the weight of each species by its average market price per kilogram in 2003. in Philippine Pesos (PHP). In addition, a resident research assistant (A. Candido) and one of the fish buyers (M. Aldeon) col- lected information from fishermen at each of the three villages through interviews every few days or weekly. The following information was gathered from fishermen: Abesamis et a\ How much does the fishery at Apo Island benefit from spillover from a marine reserve? 363 fishing grounds visited, names and number of fishers, date and times when fishing started and finished, fish- ing gear used, and composition and weights of catches (verified from records offish buyers). Interviewers were able to determine precisely where fishing was conducted because the names and locations of fishing grounds around Apo Island are common knowledge to local people (Fig. 1). However, if fishing was done near Apo Reserve (e.g., at Katipanan, Tumoy, Kanigaran, Ubos, and Kan- uran; Fig. 1), fishermen were also asked to estimate how far away they were from the boundaries of the reserve. Color-coded marker buoys, which fishermen could locate and identify easily while fishing, indicated approximate distances from reserve boundaries. These buoys were moored permanently at 100 m (blue), 200 m (red), and 300 m (yellow) from each of the northern and southern boundaries of Apo Reserve (Fig.l). The buoys were installed with the aid of a GPS receiver and fish- ermen were informed about them one week before the study started. Data were collected from 80 full-time and part-time resident fishermen (Baybay-16, Ubos-47, Cogon-17). This group comprised 70-80% of all fishermen at Apo Island. The information in the present study probably reflects accurately the actual patterns of fishing effort and yield of the local market-oriented fishery. The ma- jority of fishermen at Apo Island sell their catch to fish buyers at each of the three fish landing sites monitored (Maypa et al., 2002). The fish buyers usually purchased their fish from a group of "loyal" fishermen. Also, most fishermen live close to the houses of the fish buyers (landing sites). Thus, interviewers were able to collect data from fishermen regularly. Fishing effort and yield data were not adjusted to account for all resident fisher- men (i.e., any fishermen not included in interviews). However, data were not obtained from the following: 1) catch sold to part-time fish buyers on the island, 2) catch sold directly to the main island of Negros, 3) fishermen visiting from Negros, 4) catch sold dried, and 5) catch brought directly to homes for consumption. Items 1-4 are probably minor contributors to the total marketed yield. Maypa et al. (2002) surveyed the same fish landing sites monitored in the present study and estimated that only 10% of the total marketed yield (presumably from items 1-4) did not pass through the three major fish landing sites. On the other hand, yield from subsistence fishing (item 5) may be comparable in quantity to the marketed yield (White and Savina, 1987), but lower in monetary value. It is unlikely, how- ever, that data on fishing effort collected in this study would differ much from data on subsistence fishing. It seemed common that fishermen went to sea to catch fish both to sell and keep for personal consumption. Data analysis Targeted species were classified into five groups accord- ing to Bellwood (1988): reef-associated species (Carangi- dae and Sphyraenidae), reef planktivores (Acanthuridae, Caesionidae, and Pomacentridae), reef species (mainly Lutjanidae, Lethrinidae, Scaridae, Serranidae, Kyphosi- dae, and octopus), open water species (Belonidae, Elopi- dae, and Scombridae), and off-reef species (Lutjanidae). Fishing grounds were classified into three groups: the northern fishing grounds (Enas, Ulo, Kasorenyo, Cogon, Punta Cogon), the western fishing grounds (Largahan, Kan-upi/Boluarte, Baybay, Katipanan), and the fish- ing grounds near Apo Reserve (Tumoy, Kanigaran, Ubos. Kan-uran) (Fig.l). Data were included in the third group if fishing was done s300 m from reserve boundaries. These three fishing grounds had roughly similar surface areas to the 20-m isobath (northern fishing grounds, 17.3 ha; western fishing grounds, 14.3 ha; fishing grounds near Apo Reserve, 18.9 ha). It was assumed that most fishing was done within or just out- side the reef area enclosed by the 0- and 20-m isobaths (Fig. 1). In addition, data for the fishing grounds near Apo Reserve were classified into the following categories of distance from reserve boundaries: 0-100 m, 100-200 m, and 200-300 m. Interviews allowed collection of fishing effort data even when fishermen returned from trips without catch- ing anything (i.e., fishermen who used hook and line, gill nets, and spear guns). On average, about 20% of total fishing trips (or 23% of total fishing effort in per- son hours) returned with no catches. This figure varied considerably according to fishing gear (hook and line; 48%, gill net: 3%, spear gun: 7%) and months. However, fishing effort data for trips with zero catch were col- lected only beginning in September. Hence, the recorded fishing effort (in person hours) in July and August was adjusted by adding a correction factor in order to ac- count for fishing trips with no catches. Correction fac- tors were calculated from the equation EC ^[ERx(EZ / ET)]/[1-{EZ / ET)], (1) where EC and ER = the correction factor (expressed in person hours) and recorded monthly fishing effort, respectively, for July or August; and EZ and ET = the total fishing effort with zero catch and the total fishing effort, respectively, in September. Values for September were used because this month is within the same season as July and August (SW monsoon). Correction factors were calculated per gear (hook and line, gill net, and spear gun) and per group of fishing grounds. However, prior to adjustment, the recorded monthly fishing effort for July was multiplied by three to obtain an estimate for a 30-day period. The yield for July was also multiplied by three to obtain a 30-day yield estimate. Catch per unit of effort (CPUE, in kg/person per hour for each month was calculated by taking the average CPUE of all individual fish- ing trips made in a given month. Monthly CPUE was calculated per gear, per group of fishing grounds, and per distance from reserve boundaries (0-100, 100-200, and 200-300 m). July and August CPUE were adjusted 364 Fishery Bulletin 104(3) to account for fishing trips with zero catch, by adding correction factors (expressed in number of fishing trips) calculated by using Equation 1. Hence, correction fac- tors were the number of fishing trips with zero CPUE. Income per unit of effort (IPUE) was used as an indica- tor of economic value among fishing grounds. This was calculated in the same manner as CPUE, expressed in PHP/person per hour. IPUE for July and August were adjusted in the same manner as CPUE. ANOVA was used to determine how fishing effort for each of the three principal gear (hook and line, gill net, and spear gun [fish traps were rarely used]) varied according to seasons (SW monsoon and interim period vs. NE monsoon) and fishing grounds (northern, western, and near Apo Reserve). ANOVA was also used to determine how CPUE or IPUE varied according to fishing grounds and the three principal fishing gear and how CPUE varied according to distance from reserve boundaries (0-100, 100-200, 200-300, and >300 m) and the three principal fishing gears. Monthly estimates of each variate of interest were used as replicates in each ANOVA. Variates were transformed (log [x+1] or square root [jc-i-l]) to satisfy ANOVA assumptions. Tukey's test (Zar. 1999) was used in all post hoc analyses. Two sets of indicators of relative economic value were used besides IPUE. The first was the frequency of cap- turing high-value species, and the frequency of landing a high yield of such species, expressed in number of fishing trips. This was summarized per species group and per fishing ground. High-value species were those with the highest, or the first and second highest, price per kg within species groups. Determination of a "high yield" within high-value species depended upon the average sizes of individuals within species groups, and whether species were usually landed as individuals or as groups. The following were considered high yield for high-value species: reef associated species (Carangidae, 3 spp.), >7.0 kg; reef planktivores (Caesionidae, 2 spp.), 23.0 kg; reef species (Serranidae, 3 spp., Lutjanidae, 7 spp., Lethrinidae, 2 spp.), ^2.0 kg; open water species (Scombridae, 2 spp.), ^7.0 kg; off-reef species (Lutjani- dae, 1 sp.), ^2.0 kg. The second indicator of economic value was the probability of capturing high-value spe- cies, and the probability of landing a high yield of such species, calculated on the basis of one fishing trip. This was calculated by dividing the frequencies (the first set of indicators) by the total number of fishing trips that used appropriate fishing gear to capture high-value species. Probabilities were calculated per species group and per fishing ground. Yield from traps were excluded in this analysis. Results Seasonal patterns of fishing effort and catch composition among fishing grounds Hook-and-line fishing The majority of hook-and-line effort (73-98%) was made on the northern fishing grounds from July to December (Fig. 2A). However, during this period, hook-and-line fishing on the northern fishing grounds declined steadily (from 2302 to 50 person hours/month). It remained at low levels from January to February (35-173 person hours/month). Hook-and-line effort on the northern fishing grounds averaged 1015 ±342 (SE) person hours/month. The total hook-and-line yield from this area was 3549 kg (Table 1), dominated by reef-associated species (57%, mainly Carangidae) and reef planktivores (24%, mainly A^oso spp.). Hook-and-line effort on the western fishing grounds was much lower, averaging 56 ±22 person hours/month (Fig. 2A). How- ever, in January and February, hook-and-line effort on the western fishing grounds increased slightly (52-93 person hours/month). During this period, 38-62% of the total hook and line effort was made on the western fishing grounds, targeting an off-reef species {Aphareiis furcci (Lacepede) [Lutjanidae]). The total hook-and-line yield from the western fishing grounds was only 202 kg (Table 1), dominated by off-reef species (37%). Hook- and-line effort on the fishing grounds near Apo Reserve was the lowest among fishing grounds, averaging 33 ±9 person hours/month (Fig. 2A). It did not exhibit distinct seasonal patterns. The total hook-and-line yield near the reserve was only 166 kg (Table 1), dominated by reef species (38%, mainly octopus). Hook-and-line effort (square root [.v-i-l] transformed) differed significantly between seasons (ANOVA, Fj j,,= 13.14, P=0.002) and among fishing grounds (ANOVA, Fj ;«= 35.08, P<0.001). The season by fishing ground interac- tion was significant (ANOVA, F.^ ig = 16.34, P<0.001). During the SW monsoon and interim period (July to October), hook-and-line effort on the northern fishing grounds was significantly higher than on the western fishing grounds (Tukey's test, q.^ ,,^=12. 45, P<0.001) and on the fishing grounds near Apo Reserve (Tukey's test, (73 is = 11.84, P<0.001). During the NE monsoon (No- vember to February), hook-and-line effort did not differ significantly among fishing grounds. Hook- and-line ef- fort on the northern fishing grounds was significantly higher during the SW monsoon and interim period (July to October) than during the NE monsoon (November to February) (Tukey's test, q.,^^=9.48. P<0.001). No signifi- cant differences in hook-and-line effort between seasons were found on the western fishing grounds and on the fishing grounds near Apo Reserve. Gillnet fishing Gillnet fishing occurred mostly on the northern fishing grounds (194-466 person hours/month) from July to October (Fig. 2B). Gillnet effort on the northern fishing grounds averaged 268 ±44 person hours/month. The total gillnet yield from this area was 724 kg (Table 1), dominated by reef planktivores (76%, mainly Caesionidae). Beginning in November, gillnet effort shifted from the northern to the western fishing grounds. Gillnet effort on the western fishing grounds increased dramatically from zero in July, to 545 person hours/month in November (Fig. 2B). From November to February, the western fishing grounds accounted for 36-63% of the total gillnet effort. Gillnet effort on Abesamis et al . How much does the fishery at Apo Island benefit from spillover from a marine reserve' 365 A Hool< and line A Northern Apo Island X Western Apo Island Near Apo Reserve ^ Jul Aug Sep Oct Nov Dec Jan Feb •S- 600 1 B Gill net 400 ■ B 200 ■ 7 ^^RS Ra£.^^ Jul Aug Sep Oct Nov Dec Jan Feb Aug Sep Oct Nov Dec Jan Feb ■ SW- -M INT K- NE • Month and moonsoon Figure 2 Seasonal trends in fishing effort (left-hand side) and composition of yield (right-hand side) for (A) hook-and-line, (B) gillnet, and (C) spear gun gear used at the three fishing grounds at Apo Island. Seasons: SW=southwest monsoon, INT=calm interim period, NE = northeast monsoon. Species groups: RA=reef associated species, RP=reef planktivores, RS = reef species, OW=open water species, OR = off-reef species. Percentages of dominant species groups are indicated. Legend for graphs at the left-hand side is at the upper right-hand side. the western fishing grounds averaged 172 ±62 person hours/month. The total gillnet yield from this area was 493 kg (Table 1), dominated by reef species (64%, mainly Scaridae). Gillnet effort near Apo Reserve was much lower compared to other fishing grounds, averaging 78 ±27 person hours/month only. However, gillnet effort was relatively high near the reserve in July, October, and November (193, 155, and 157 person hours/month, respectively) (Fig. 2B). The total gillnet yield near the reserve was 318 kg (Table 1), dominated by reef species (63%, mainly Scaridae). Gillnet effort did not differ significantly with seasons (ANOVA, F^ jj, = 0.18, P=0.68), but differed significantly among fishing grounds (ANOVA, ^.,,^,=6.72, P=0.007). The season by fishing ground interaction was significant (ANOVA, ^2 18=7-66, P=0.004). During the SW monsoon or interim period (July to October), gillnet effort on the northern fishing grounds was significantly high- er than on the western fishing grounds (Tukey's test. q^ jj, = 5.68, P<0.005) and on the fishing grounds near Apo Reserve (Tukey's test, q.^ [8=4.81, P<0.01) but did not differ between the latter two fishing grounds. Dur- ing the NE monsoon (November to February), gillnet effort did not differ between the western and northern fishing grounds, but gillnet effort on the western fishing grounds was significantly higher than near Apo Reserve (<73 18=4.47, P<0.025). Gillnet effort on the northern fish- ing grounds was higher during the SW monsoon and interim period than during the NE monsoon (Tukey's test, 9.^ 18=2.96, P=~0.05 [qo 05 2 i8=2.97]). Conversely, gillnet effort on the western fishing grounds was signifi- cantly higher during the NE monsoon than during the SW monsoon and interim period (Tukey's test, (?, ih= 4.67, P<0.005). Gillnet effort near Apo Reserve did not differ significantly between seasons. Spear gun fishing Spear fishing occurred mainly on the northern fishing grounds (Fig. 2C). However, spear 366 Fishery Bulletin 104(3) Table 1 Summary of recorded fish ing effort, yield, and income of the fishery at Apo Is and from 22 July 2003 to 29 February 2004. Contributions of the three fish ing grounds see Fig. 1) are shown PHP = Philippine peso. Hook and line Gill net Spear gun Fish trap Overall Total fishing effort (person hours) 8840 4141 452 129' 13,443-' Percent contribution Northern Apo Island 91.9 51.8 63.3 100 78.6 Western Apo Island 5.0 33.1 8.4 0 13.8 Near Apo Reserve 3.1 15.1 28.3 0 7.6 Total yield (kg) 3917 1535 406 32-' 58903 Percent contribution Northern Apo Island 90.6 47.2 66.2 100 77.7 Western Apo Island 5.2 32.1 4.8 0 12.1 Near Apo Reserve 4.2 20.7 29.0 0 10.2 Total income (PHP) 242,026 68,028 23,509 137P 334,934-5 Percent contribution Northern Apo Island 90.4 49.7 67.6 100 80.6 Western Apo Island 5.5 31.6 4.2 0 10.7 Near Apo Reserve 4.1 18.7 28.2 0 8.7 ' Fishing effort in trap days = number of traps x number of day s traps were left on the reef - Excludes fish trap effort. ^ Trap yield in July not adjust ed tr 30-da\ perio( . gun effort in this area declined from August to Febru- ary (67 to 17 person hours/month, respectively). Spear gun effort on the northern fishing grounds averaged 36 ±6 person hours/month. The total spear gun yield from the northern fishing grounds was 269 kg (Table 1), dominated by reef-associated species (44%, Carangidae). Spear gun effort on the western fishing grounds was much lower, averaging only 5 ±3 person hours/month. It was highest in November (23 person hours/month) (Fig. 2C). The total spear gun yield from the western fishing grounds was only 19 kg (Table 1), dominated by reef species (62%, mainly Scaridae and octopus). Spear gun effort near the boundary of Apo Reserve was also rather low but was higher on average than on the west- ern fishing grounds (mean 15 ±5 person hours/month). It peaked in October (44 person hours/month) (Fig. 2C). The total recorded spear gun yield near the reserve was 118 kg (Table 1), dominated by reef species (56%, mainly Scaridae and octopus). Spear gun effort did not differ significantly between seasons (ANOVA, F, i,^=2.12, P=0.16) but differed sig- nificantly among fishing grounds (ANOVA, F., j„=14.26, P<0.001). The season by fishing ground interaction was not significant (ANOVA, F., i„=1.77, P=0.20). Spear gun effort on the northern fishing grounds was significantly higher than on the western fishing grounds (Tukey's test, Q'g jg=7.41, P< 0.001) and the fishing grounds near Apo Reserve (Tukey's test, ^g jg=4.96, P<0.01). Spear gun effort, however, did not differ significantly between the fishing grounds near Apo Reserve and the western fishing grounds of Apo Island. Trap fishing Bamboo fish traps were used on the north- ern fishing grounds only, and only in July (SW monsoon). The total trap effort was 126 trap days (4 fish traps set for 14 days, 7 for 10 days) with a total yield of 32 kg (Table 1). The yield was dominated by reef planktivores (85% Acanthuridae and Caesionidae). Contributions of fishing gears and fishing grounds to overall fishing effort, yield, and income Among the three principal fishing gear, most fishing effort was spent with hook and line, followed by gill nets, then with spear guns (Table 1). Hook-and-line fishing, therefore, contributed the greatest yield and highest income, accounting for 66% of the total yield and 72% of the total income recorded. Bamboo fish traps contributed the least yield and income (Table 1). All types of fishing occurred mainly on the northern fishing grounds. The northern fishing grounds accounted for 92%, 52%, 63%, and 100% of the total effort for hook-and-line gear, gill nets, spear guns, and fish traps, respectively. Accord- ingly, the northern fishing grounds accounted for the vast majority of total yield and total income for all types of fishing gears (Table 1). On the other hand, fishing effort was often lowest on the fishing grounds near Apo Reserve (Table 1). The fishing grounds near the reserve accounted for only 3% and 15%) of the total effort spent on hook-and-line and gillnet fishing, respectively. These fishing grounds contributed only 4% to the total yield and total income from hook and line fishing, and only 21% to the total Abesamis el a\ How much does the fishery at Apo Island benefit from spillover from a marine reserve' 367 yield and 19% to the total income from gill- net fishing. Only 28% of the total spear gun effort was made near Apo Reserve. However, the overall spear gun effort near the reserve was about three times higher than that on the western fishing grounds. The fishing grounds near the reserve contributed about six times the yield and seven times the in- come of spear fishing on the western fishing grounds (Table 1). Among species groups, the highest yield recorded was for reef-associated species, fol- lowed by reef planktivores, and then reef species (Table 2). Open water and off-reef species were minor contributors to overall yield. Reef-associated species together with reef planktivores accounted for 69% of the overall yield. The northern fishing grounds contributed >80% of the total yield of reef- associated species, reef planktivores, and open water species (Table 2). Much of the total yield of reef species (47%) was also taken from the northern fishing grounds. The western fishing grounds contributed 68% of the total yield of off-reef species but accounted for little of the yield for other spe- cies groups (Table 2). The fishing grounds near Apo Reserve often accounted for the smallest contribution to yield for all species groups. However, about a quarter of the to- tal recorded yield of reef species were taken near the reserve. Spatial patterns of CPUE and IPUE among fishing grounds (wJ (ft 12 1 0 08 06 04 02 00 Hook and line X n Gill net 1 Spear gun 25-1 2.0- I '"' I [^ r 05 ' I 0 0 I W , I 1,1 I I NR W NR W NR W N LU Q. D. I 40 30 20 10 0 JL NR W N NR W N Figure 3 Mean catch per unit of effort (CPUE ) and income per unit of effort (IPUE) for hook-and-line, gillnet and spear at the three fishing grounds at Apo Island. NR = near Apo Reserve, W=western Apo Island, N = northern Apo Island (see Fig. 1). Error bars are 1 stan- dard error. PHP=Philippine peso. Table 2 Composition of total recorded yield of the fishery at Apo Island from 22 July 2003 to 29 February 2004. Contributions of the three fishing grounds (see Fig. 1) are shown. Species groups: RA=reef-associated species, RP=reef planktivores, RS=reef species, OW=open water spe- cies, OR=off-reef species. Yield in July adjusted to 30-day period for each species group. Table excludes 16 kg of unidentifiable catch. RA RP RS OW OR Mean hook-and-line, gillnet, and spear gun CPUE were highest in the fishing grounds near Apo Reserve (Fig. 3). CPUE near the reserve was higher than on the northern fishing grounds by a factor of 1.5, 1.4 and 1.4 for hook-and-line, gillnet, and spear fishing, respectively. CPUE near the reserve was higher than on the western fishing grounds by a factor of 1.6, 1.1 and 2.0 for hook-and- line, gillnet, and spear fishing, respectively. However, ANOVA indicated that CPUE (log [x+l] transformed) did not differ significantly among the three groups of fishing grounds (F, 57=1.8?, P=0.16), but differed significantly among fishfng gear (F^ 5^=9.26, P<0.001). The trends in mean IPUE reflected closely those of CPUE (Fig. 3). Mean IPUE near the reserve was higher than on the northern fishing grounds by a factor of 1.4, 1.2, and 1.4 for hook-and-line, gillnet and spear fishing, respectively. It was higher than on the west- ern fishing grounds by a factor of 1.6, 1.1 and 2.2 for hook-and-line, gillnet, and spear fishing, respectively. However, ANOVA indicated that IPUE (log [.r+1] trans- formed) did not differ significantly among the three groups of fishing grounds (F., -~=0.49, P=0.62), but Total yield (kg) 2333 1772 1305 418 111 Percent contribution Northern Apo Island Western Apo Island Near Apo Reserve 3.7 5.2 24.9 16.4 24.6 92.4 85.6 46.7 80.0 7.9 3.9 9.1 28.4 3.6 67.5 differed significantly among fishing gear (F,, gy=6.77, P=0.002). Monthly mean hook-and-line, gillnet, and spear gun CPUE and IPUE near Apo Reserve were more variable than in the other fishing grounds (Fig. 4, A-C). For example, the monthly hook-and-line CPUE near Apo Reserve changed from >1.5 to <1.0 to >2.0 and to <1.5 kg/person per hour from July to October. However, from November to February, it remained at <0.5 kg/person per hour, but was zero in January (Fig. 4A). Monthly hook-and-line IPUE near Apo Reserve varied accord- ingly. It changed from >100 to <50 to >150 and to <100 PHP/person per hour from July to October, but was less than 30 PHP/person per hour from November to Febru- ary and was zero in January. In contrast, the monthly 368 Fishery Bulletin 104(3) hook- and-line CPUE and IPUE on the northern and western fishing grounds of Apo Island exhibited little variability (Fig. 4A). For most of the period between July and February, the monthly hook-and-line CPUE on the Table 3 Range of monthly mean catch per unit of effort (CPUE) and income per unit of effort (IPUE) for hook-and-line, gillnet, and spear gun gea - at the fishing grounds at Apo Island (see Fig. 1). Lowest minimum and highest maximum average values are marked by the symbol "i". Highest minimum values are marked by ± PHP = Philippine peso. Hook and line Gill net Spear gun CPUE (kg/person per hour) Northern Apo Island 0.3t-0.8 0.2'-0.7 0.7*-2.0 Western Apo Island 0.1-1.0 0.2*-1.3 0.2t-1.5 Near Apo Reserve 0.0'-2.4' 0.r-1.6t 0.5-3.9' IPUE (PHP/person per hour* Northern Apo Island 18.4^-49.00 9.6!-31.30 40.30-117.60 Western Apo Island 6.20-58.40 8.20-46.40 11.90^-57.80 Near Apo Reserve 0'-154.90' 7.80--73.70' 30.50-249.90' Northern Apo Island X Western Apo Island ^ Near Apo Reserve Hook and line 2.0 15 Jul Aug Sep B Gill net Oct Nov Dec Jan Feb CL X CL 3 Q_ U 40- 3 0. Jul Aug Sep Oct C Spear gun Nov Dec Jan Feb Q. Aug -SW- Sep Oct Nov Dec Jan Feb INTw NE 300 n 250 200 150 100 50 0 Aug -SW- Month and monsoon Figure 4 Monthly mean catch per unit of effort I CPUE) and income per unit of effort (IPUE) for (A) hook-and-line, (B) gill-net, and (C) spear at the three fishing grounds at Apo Island. Seasons: SW=southwest monsoon, INT=calm interim period, NE = northeast monsoon. PHP=Philippine peso. northern fishing grounds remained between 0.5 to 0.8 kg/person per hour, whereas IPUE remained between 30 to 50 PHP/person per hour. Monthly CPUE for the western fishing grounds remained mostly between 0.2 to 1.0 kg/person per hour, whereas monthly IPUE re- mained mostly between 15 to 60 PHP/person per hour. Fishing grounds near Apo Reserve often had the lowest minimum monthly CPUE and IPUE for all fish- ing gears, except for spear gun, which had the lowest minimum monthly CPUE and IPUE on the western fishing grounds (Table 3). However, fishing grounds near Apo Reserve always had the highest maximum monthly CPUE and IPUE for all fishing gears (Table 3, Fig. 5). Maximum month- ly CPUE near the reserve was higher by a factor of 1.2 to 3.0, depending on the fishing gear and fishing ground (Fig. 5). Maximum monthly IPUE near the re- serve was higher by a factor of 1.6 to 4.3, depending on the fishing gear and fish- ing ground. On the other hand, the highest minimum monthly CPUE and IPUE for all fishing gears were most often found only on the northern fishing grounds of Apo Island (Table 3). Spatial patterns of high-value catches among fishing grounds Capturing high-value spe- cies and landing a high yield of such species were often more frequent on the northern fishing grounds (Table 4). For example, 133 captures of high-value reef- associated species (Caran- gidae — Caranx ignobilis (Forsskal), Caranx melam- pygus (Cuvier), and Caran- goides sp. ) were made on the northern fishing grounds. In 46 of these captures, the yield was aT.O kg or 525 PHP. Six of the fishes cap- Jul Aug Sep Oct Nov Dec Jan Feb Sep Oct Nov Dec Jan Feb INT»- -NE- Abesamis et al : How much does the fishery at Apo Island benefit from spillover from a marine reserve' 369 Table 4 Frequency and probability of capturing or landing a high yield of high-value species within each species gr-oup for each of the three fishing grounds at Apo Island (see Fig. 1), Upper values: frequency (number of fishing trips) and probability (in parenthe- ses) of capturing high-value species. Lower values; frequency (number of fishing trips) and probability (in parentheses) of landing a high yield of high-value species. Highest frequencies are marked with the symbol t. Highest probabilities are marked with the symbol ±. Price per kg and prices of what were considered "high yield" are given in Philippine pesos (PHP). Species groups, families Northern Apo Island Western Apo Island Near Apo Reserve Reef-associated species (Carangidae) 75 PHP/kg £7.0 kg or PHP 525 Reef planktivores (Caesionidae) 55 PHP/kg >3.0 kg or PHP 165 Reef species (Serranidae, Lutjanidae, Lethrinidae) 70-75 PHP/kg 22.0 kg or PHP 150 Open water species (Scombridae) 75-90 PHP/kg >7.0 kg or PHP 525 Off-reef species (Lutjanidae) 75 PHP/kg >2.0kgorPHP 150 133' (0.07)? 9 (0.03) 8 (0.05) 46t (0.03)+ 1 (0.004) 4 (0.02) 123t (0.07)i 17 (0.06) 9 (0.05) 36t (0,03)+ 4 (0.02) 0 (0) 57+ (0.03) 15 (0.06) 12 (0.07) 31t (0.02) 10 (0.04) 9 (0.05) 5' (0.003) 1 (0.01)+ 0 (0) 3< (0.002)+ 0 (0) 0 (01 4 (0.003) 29' (0.23 )i 3 (0.04) 2 (0.001) 13' l0.10)i 2 (0.03) tured weighed alS kg (equivalent to a value of all25 PHP). The largest was a 31.9-kg C. ignobilis valued at 2393 PHP, which was also the biggest fish caught during our study. In contrast, only eight captures of high-value reef-associated species were made near Apo Reserve. In only four of these captures were the yields ^7.0 kg or 525 PHP. Only one of the fishes captured near the reserve weighed >15 kg. This was a 20.2-kg C. ignobilis valued at 1515 PHP. Capturing high-value species and landing a high yield of such species often had the great- est probability of occurring on the northern fishing grounds (Table 4). For example, the probability of capturing high-value reef-as- sociated species from the northern fishing grounds was 0.07. Landing a high yield of such species from this area had a probability of 0.03. In contrast, the probability of captur- ing high-value reef-associated species near Apo Reserve was 0.05. Landing a high yield of such species near the reserve had a prob- ability of 0.02. However, the probability of capturing high-value reef species (Serrani- dae— Cephalopholis, Epinephelus, Variola spp.; Lutjanidae — Aprion, Lutjanus, Macolor. Sym- phorichthys, Symphorus spp.; and Lethrinidae — Lethrinus spp.) was greatest near Apo Reserve (Table 4). The probability of landing a high yield of such spe- cies (>2.0 kg or 150 PHP) was also greatest near the Hook and line Gill net 3£ Q- ^ O E m 1 5 1 0 00 - 160 UJ ^ D ^ Q. D — Q. E C =1 o E 2 120 80 Q- X 'h. 0 1 5 12 09 06 03 00 80 60- 40 20 40 30 Spear gun NR w NR W 250 200' 100 NR W NR W JL Figure 5 Maximum monthly mean catch per unit of effort (CPUE) and income per unit of effort (IPUE) for hook-and-line, gill-net, and spear gun gear used at the three fishing grounds at Apo Island. NR = near Apo Reserve, W=western Apo Island, N = northern Apo Island (see Fig. 1). Error bars are 1 standard error. PHP=Philippine peso. reserve, despite the higher frequency of capturing high- value reef species, or landing a high yield of such spe- cies, on the northern fishing grounds than near Apo Reserve (Table 4). 370 Fishery Bulletin 104(3) A Hook and line Lil 0-100 100-200 200-300 >300 W >300 N B Gill net Ml 0-100 100-200 200-300 >300 W ;-300 N 2 5 C Spear gun ik linn 0 0 4—^ 1— r-1 1— . 0-100 100-200 200-300 >300 W >300 N Distance from boundary (m) Figure 6 Mean catch per unit of effort (CPUEl for (A) hook-and-line, (B) gillnet, and (C) spear gun gear deployed at three distances from the boundary of Apo Reserve (0-100, 100-200, 200-300 m [shaded columns]; refer to Fig. 1) and at two areas far (>300 m) from reserve boundaries (W = western Apo Island fishing ground, N=northern Apo Island fishing ground [open columns]; refer to Fig. 1). Error bars are 1 standard error. Spatial patterns of CPUE near reserve boundaries Spatial patterns of CPUE away from reserve boundaries differeci for hook-and-line, gillnet, and spear gun gear (Fig. 6). Gillnet CPUE exhibited a pattern of decrease from 0-100 to 200-300 m from the reserve boundar- ies (from 0.88 to 0.70 kg/person per hour; Fig. 6B). Gillnet CPUE further decreased >300 m from reserve boundaries (western and northern fishing grounds). In contrast, hook-and-line and spear gun CPUE was lowest near (0-100 m) reserve boundaries, but much higher at distances of 100-300 m from the boundaries (Fig. 6, A and C). Hook-and-line and spear gun CPUEs at 100-300 m from reserve boundaries were higher than farther away (>300 m) from the boundaries (western and northern fishing grounds). However, hook-and-line and spear gun CPUEs were highest at 100-200 m from reserve boundaries (Fig. 6, A and C). An ANOVA with data for 0-300 m indicated only that CPUE did not differ significantly with distance from reserve boundar- ies (F,, o„=0.73, P=0.49) or with fishing gear {F^ 39=2.52, Table 5 Frequency of fishing trips for each of the three distances from the boundaries of Apo Reserve (see Fig. 1), for each of hook and line, gill net, and spear gun. Fishing trips were made between 22 July 2003 and 29 February 2004. Number of times fished Hook and line Gill net Spear gun Distance from reserve boundaries 0-100 m 2 3 100-200 m 43 24 200-300 m 25 28 4 33 7 P=0.09). However, an ANOVA with data for 0-300 and >300 m indicated that CPUE (log [.v+1] transformed) did not differ significantly with distance from reserve boundaries (F^ -g=1.19, P=0.32) but differed significantly among fishing gear (F., -^=7.28, P=0.001). Contrasting patterns were also found in fishing ef- fort (frequency of fishing trips) at different distances from the boundaries of Apo Reserve (Table 5). Hook- and-line and spear fishing occurred most frequently at intermediate distances from reserve boundaries, but gillnet fishing occurred most frequently at the far- thest distances. However, all types of fishing occurred very infrequently within 100 m of reserve boundaries (Table 5). Discussion Results indicate that the maximum possible contribution of spillover from the no-take reserve to the overall yield and income of the fishery at Apo Island is small. If one assumes that the spillover of coral reef fishes is most likely to operate on spatial scales of hundreds of meters (Russ, 2002), and, thus, would be unlikely to affect fish- ery yields on the western and northern fishing grounds, the maximum possible contribution would be 10% of the total fishery yield, but the real value is probably much less than this. We assumed that the eight months sampled in 2003-2004 are representative of the general spatial pattern of fishing at Apo Island. Furthermore, we assumed that the spatial pattern of fishing during the four months not sampled (IVIarch to June) remains consistent with the general pattern of fishing. It can be estimated from the present study that the total fishery yield at Apo Island in 2003-2004, excluding the yield consumed locally, was around 10.4 tons/km- per year (to the 60-m isobath). Thus, spillover from Apo Reserve would have contributed very much less than 1.0 ton/km' per year in 2003-2004. Fishing effort was often lowest on the fishing grounds near the reserve. The fishery at Apo Island is primarily hook and line, targeting reef-associated species (Caran- gidae), and to a lesser extent reef planktivores iNaso Abesamis et a\ : How much does the fishery at Apo Island benefit from spillover from a marine reserve' 371 spp.) (Alcala and Luchavez. 1981; White and Savina, 1987; Bellwood, 1988; Maypa et al., 2002; present study). Ninety-two percent of total hook-and-line effort was made on the northern fishing grounds of the island. Gill nets, spear guns, and fish traps were not used as much as hook-and-line; however, fishing with these gear was also carried out mostly on the northern fishing grounds, contributing to a yield of reef-associated spe- cies and reef planktivores. Fishing near Apo Reserve contributed mainly to catch of reef species (Scaridae and octopus), which are a less important species group for the local fishery. The fishery depended heavily on access to the northern fishing grounds during the SW monsoon and calm interim period (Bellwood, 1988; pres- ent study). Results clearly showed that fishing effort on the northern fishing grounds was considerably lower during the NE monsoon, especially for hook-and-line gear and gill nets. However, fishing effort did not shift to the fishing grounds near Apo Reserve during the NE monsoon. Gillnet fishing seemed to transfer instead to the western fishing grounds, targeting reef species (Scaridae). To a much lesser degree, hook-and-line fish- ing also transferred to the western fishing grounds, targeting high-value off-reef species (Aphareus furca [Lutjanidae]). The northern fishing grounds alone ac- counted for 78% of overall yield and 81% of overall income recorded in our study (Table 1). In contrast, fishing grounds near Apo Reserve accounted for only 10% of overall yield and 9% of overall income (Table 1). However, it is highly unlikely that spillover from Apo Reserve supplied the entire yield taken from fishing grounds near the reserve. That is, the yield near the reserve was probably not composed totally of migrants from the reserve. Therefore, the actual spillover contri- bution of Apo Reserve would be much less than 10% to the overall yield and income generated by the fishery. However, CPUE of fishermen was highest on the fish- ing grounds near Apo Reserve. This pattern may have resulted from 1) spillover from the reserve or 2) lower fishing intensity near the reserve, resulting in higher abundance of fish and, thus, higher CPUE. There is far more empirical evidence supporting the spillover hypothesis than the lower fishing intensity hypothesis. Considerable evidence exists that some supplement for the fishery near the reserve has developed over time. Firstly, monitoring of Apo Reserve since 1983 has shown that fish populations inside the reserve have increased in abundance, some having tripled in density or biomass over the last two decades (Russ and Alcala, 1996, 1998, 2003; Russ et al., 2003, 2004; Alcala et al., 2005). Secondly, long-term monitoring indicates that the reserve began to export adult fish to a site open to fishing after ~8 years of reserve protection (Russ and Alcala, 1996; Russ et al., 2003, 2004). Thirdly, catch rates of some species (Acanthuridae) were found to be higher near the reserve than elsewhere around Apo Island after two decades of reserve protection (Russ et al., 2003, 2004). Fourthly, recent studies indicate that patterns of decreasing abundance of some targeted spe- cies are present across the northern boundary of the reserve (Abesamis et al., 2005). Lastly, a recent study indicates that density-dependence may be driving net emigration of adult fish (Naso vlamingii (Valenciennes)) from the reserve (Abesamis and Russ, 2005). On the other hand, no direct evidence is available to show that fishing effort near Apo Reserve has remained low over the last two decades. However, the fishing grounds near the reserve may be partially exposed to the SW and NE monsoons (Alcala and Luchavez, 1981). During the SW monsoon, the fishing grounds adjacent to the southern end of the reserve may become rough, but those ad- jacent to the northern end are usually calm (Fig. 1). The pattern is reversed during the NE monsoon. Thus, fishing effort near the reserve may be reduced for nine months of the year (June to September, November to March) because only the area close to one side of the reserve may be fished, depending on the monsoon. The role of the monsoons in limiting fishing effort near the reserve cannot be ruled out entirely. If spillover has increased CPUE near the reserve, as long-term evidence seems to indicate, then why have local fishermen not responded noticeably to the im- provement in catch rates? Weather may be important in limiting fishing effort near the reserve, but it cannot explain the low fishing intensity in this area during the calm interim months. Results show that hook-and- line and gillnet effort on the fishing grounds near the reserve in October were still considerably lower than on the northern fishing grounds (Fig. 2, A and B). An- ecdotal information also indicates that most fishermen still prefer to fish the northern fishing grounds even during April and May (Pascobello-). This information is consistent with the findings of previous studies (White and Savina, 1987; Bellwood, 1988; Maypa et al., 2002), which indicate that the general pattern of fishing effort at Apo Island has not changed much since creation of the reserve in 1982. Furthermore, a reasonable amount of fishing area (8-11 ha) near the reserve is still avail- able to fishermen during either monsoon. Given the higher catch rates, fishermen could still concentrate effort near the reserve, one side of the reserve at a time, depending on the monsoon. Higher costs (time and energy spent to paddle a small boat [banca]] are certainly not preventing fishermen from fishing adjacent to the reserve. The majority of fishermen (80%) reside near the reserve, in Baybay and Ubos villages (Fig. 1) and therefore for most fishermen the costs of fishing the northern side of Apo Island may actually be higher. It is also unlikely that fishermen are not aware of higher catch rates near the reserve. Fishermen can probably obtain good information about catch rates from the experiences of fellow fishermen, or from word-of-mouth, because the community is relatively small and tightly knit. Many fishermen are members of the same family (including cousins and uncles) or are friends with each other. Income rates per se can also be ruled out as an - Pascobello, M. 2002. Personal commun. Apo Island resi- dent and village chairman. Apo Island, Dauin Municipality, Negros Oriental, Philippines, 6217. 372 Fishery Bulletin 104(3) important influence on the spatial pattern of effort because results indicated that IPUE for all gears was also highest on the fishing grounds near the reserve. However, one cannot discount that the northern side of Apo Island may be regarded by the local community as their main traditional fishing ground. This area has probably been fished for decades, providing the income for families generation after generation. Other characteristics of the CPUE and IPUE data may also provide a partial explanation. CPUE and IPUE were more variable on the fishing grounds near Apo Reserve (Fig. 4). In fact, the fishing grounds near the reserve often had both the highest and lowest CPUE and IPUE per month for all types of fishing gears (Ta- ble 3, Fig. 5). The only exception was for spear gun, which had the lowest CPUE and IPUE on the western fishing grounds (Table 3). These findings may indi- cate that near the reserve, fishermen could obtain very high average monthly catch rates and income rates. However, they may also indicate that near the reserve, fishermen (using hooks and lines and gill nets) could obtain very low average monthly catch rates and income rates. Monthly CPUE and IPUE were least variable on the northern fishing grounds (Fig. 4). In addition, the highest minimum average monthly CPUE and IPUE were always found in this area (Table 3). Therefore, a plausible explanation for the spatial pattern of fish- ing effort at the island scale is that fishermen prob- ably tend to avoid fishing near Apo Reserve in order to make their monthly incomes more stable and avoid occasional very low catch rates. Fishermen may prefer the northern fishing grounds even if catch rates in this area are not as high as near the reserve because they are assured of obtaining higher minimum yields and a more stable income. This strategy may be a form of financial "risk aversion" (Hilborn and Walters, 1992). It is further postulated that such a risk aversion strategy could explain why total spear gun effort was higher on the fishing grounds near Apo Reserve than on the western fishing grounds (Tables 1 and 3). Also, a risk aversion strategy may partly explain why fishing with gill nets and hook and line transferred to the western fishing grounds instead of near the reserve during the NE monsoon (Fig. 2B, Table 3). Other factors may also influence the decision by fish- ermen to concentrate effort on the northern side of Apo Island. For example, the largest and most valuable fish recorded in this study, a 31.9-kg Caranx ignobilis worth almost PHP 2400, was captured from the northern fishing grounds. A high-priced catch such as this one may have a tremendous psychological impact on local fishermen. Anecdotal information seems to indicate that the income from such a yield may be enough to provide for 80% of the daily expenses of one family for one month at Apo Island (Pascobello-Rhodes'^). In com- ^ Pascobello-Rhodes, L. 2003. Personal commun. Apo Island resident. Apo Island, Dauin Municipality, Negros Oriental, Philippines, 6217. parison, the largest and most valuable fish caught near Apo Reserve, also C. ignobilis, was only two-thirds the weight (20.2 kg) and value (PHP 1515) of the largest fish caught from the northern fishing grounds. This fish was the only one caught near the reserve with a value greater than PHP 1000. Local fishermen probably consider fish of this size more common on the northern fishing grounds. During this study, five individuals of C. ignobilis (range: 15.0-19.0 kg) that were captured from the northern fishing grounds had weights similar to the largest fish caught near the reserve. Therefore, the biggest psychological impact on fishermen may not be due to spillover from the reserve. A perception by local fishermen that the northern fishing grounds are productive areas for high-value spe- cies may not only have a psychological basis. Results in- dicate that the probability of capturing high-value reef- associated species (Carangidae) and the probability of landing a high yield of such species were highest on the northern fishing grounds (Table 4). The same was true for high-value reef planktivores (Caesionidae) (Table 4). The probability of landing a high yield of high-value open water species (Scombridae) was also greatest on the northern fishing grounds (Table 4). Assuming that fishermen prefer to target high-value species, these findings indicate that fishermen have better chances of making higher incomes from the northern fishing grounds than from other areas. The better prospect of making a higher income probably influences the deci- sion by fishermen to concentrate effort on the northern side of Apo Island whenever it is accessible. Similarly, some fishermen probably shift to the western fishing grounds during the NE monsoon (the "off-season") be- cause of better chances of gaining higher incomes by targeting high-value off-reef species (A. furca [Lutjani- dae]) (Table 4). However, the results indicated that the probability of capturing high-value reef species (Serranidae, Lut- janidae, and Lethrinidae), or landing a high yield of such species, was greatest on the fishing grounds near Apo Reserve (Table 4). Fishing intensity for high-value reef species, on the other hand, appears to be greatest on the northern fishing grounds, because capturing or landing a high yield of high-value reef species was most frequent there (Table 4). Thus, it seemed that most fishermen avoided fishing for high-value reef species near Apo Reserve despite better chances of making high incomes from high-value reef species. One plausible explanation for avoiding the fishing grounds hear the reserve is that fishermen would still prefer to fish the northern fishing grounds because they can target the high-value species that come in larger sizes (Carangidae and Scombridae) or greater numbers (Caesionidae) and at the same time occasionally capture high-value reef- species. The high-value reef species (Serranidae, Lut- janidae, Lethrinidae) are unsatisfactory alternatives for fishermen because they come in smaller sizes (compared to Carangidae and Scombridae) or in smaller numbers (compared to Caesionidae), and therefore would fetch a lower price. Abesamis et al How much does the fishery at Apo Island benefit from spillover from a marine reserve? 373 On the other hand, the highest maximum monthly CPUE and IPUE were invariably found near Apo Re- serve (Fig. 5, Table 3). This result is consistent with the occasional spillover of bigger fish from the reserve. Such an effect may have a positive influence on the at- titudes of fishermen toward reserves (Russ and Alcala, 1996). However, in the case of the artisanal fishery at Apo Island, any psychological impact of large catches near the reserve on fishermen is probably attenuated by the importance of the northern fishing grounds. In other fisheries, however, occasional spillover of large adults may be important. Recreational "trophy" fish- eries, for example, may benefit directly from such an effect of no-take reserves (Bohnsack, 1998; Johnson et al., 1999; Roberts et al., 2001). In New Zealand, Kelly et al. (2002) showed that lobster catch rates (kg/trap haul) were similar close to and far from Leigh Marine Reserve, although catches around the reserve consisted of fewer individuals. The lobsters caught near the re- serve were bigger. Furthermore, the amount of money made per trap haul close to the reserve was similar to sites far from the reserve. The findings at the scale of a few hundred meters from the boundary of the reserve provide little evidence to indicate that spillover from Apo Reserve is present. In fact, the most informative result at this spatial scale was that fishermen seemed to avoid fishing very close to the reserve (i.e., within 100 m from the boundaries) (Table 5). Furthermore, catch rates seemed to be low- est closest to reserve boundaries for hook-and-line and spear fishing (Fig. 6, A and C). However, gillnet fishing seemed to have a pattern of decreasing catch rate away from reserve boundaries (Fig. 6B). Gear selectivity in- teracting with the spatial distribution of target species may explain the differences in spatial patterns of CPUE among the three gears. Gillnet fishing is probably less selective than hook-and-line and spear fishing. The pattern of fishing effort found very near to the reserve can be interpreted in two ways. Firstly, local fishermen may be well aware of the distribution of catch rates near Apo Reserve, and they adjust their fishing effort accordingly. This interpretation is supported to some degree by the correspondence between the pattern of CPUE and the pattern of fishing effort for hook-and- line and spear gun gear. Both CPUE and fishing effort for these gear were highest at intermediate distances (100-200 m), but lower at the closest (0-100 m) and farthest distances (200-300 m) from the reserve (Fig. 6, A and C, Table 5). Fishermen may know from experi- ence that hook-and-line and spear gun catch rates are low nearest the reserve (0-100 m), hence they tend to avoid fishing in this area. Experimental fishing with hooks and line, but specifically targeting A^. vlamingii, supports the contention that hook-and-line CPUE is low very close to the boundary of Apo Reserve (Abesamis and Russ, 2005). Hook-and-line CPUE for N. vlamingii was higher at intermediate distances (150-200 m) than at the closest (50-100 m) and farthest (250-300 m) distances from the reserve (Abesamis and Russ, 2005). Although movement of fish from Apo Reserve to sites at intermediate distances (100-200 m) from the re- serve has never been demonstrated directly, research in the last two decades has shown that fish populations (Acanthuridae, Carangidae, Serranidae, Lutjanidae, and Lethrinidae) have increased outside but close to the southern boundary of the reserve (about 200-250 m from that boundary) after about eight years of reserve protection (Russ and Alcala, 1996; Russ et al. 2003, 2004). This increase in populations may indicate that the present spatial distribution of hook-and-line and spear gun effort at the local scale may reflect the re- sponse of a small number of fishermen to spillover from the reserve. Secondly, local fishermen may avoid fishing very close to the boundaries of Apo Reserve in order to prevent being accused of poaching inside the reserve. The in- consistent result of a higher catch rate but lower fishing effort nearest the reserve for gillnet fishing supports this idea (Fig. 6B, Table 5). Furthermore, there may be tremendous motivation for resident fishermen not to be implicated in poaching inside the reserve. The ultimate reason for this probably lies in the smallness of the community at Apo Island. Almost everyone has known each other for most of their lives. The no-take reserve was established by the community for their own benefit (Russ and Alcala, 1999). It has been guarded and maintained by community members, many of them fishermen themselves, for most of the two decades of its existence (Russ and Alcala, 1999). It has an important role in tourism on the island, from which the local com- munity has benefited considerably in many ways (Alcala, 1998; Russ and Alcala, 1999). It is the principal factor that has made Apo Island a nationally and internation- ally recognised model for successful, community-based resource management (Alcala, 1998; Russ and Alcala 1999). Many residents are probably aware of most, if not all, of these achievements. For these reasons, it is clearly against the best interests of a resident fisherman, and his family, to be labeled by fellow community members as a threat to the security of their reserve. In conclusion, this study has shown that spillover yield from the no-take reserve at Apo Island probably contributes much less than 10% of the overall yield to the local fishery. Fishing effort was often lowest near the reserve, despite higher catch rates there. The find- ing of low fishing intensity near reserve boundaries despite indications of spillover contrasts with most em- pirical evidence collected so far (e.g., Gell and Roberts, 2003) and recent theoretical studies (e.g., Salomon et al., 2002) indicate that spillover will increase fishing effort near reserve boundaries. Fishing effort adjacent to the reserve appears to be limited by 1) weather deter- mined by the monsoons, 2) the traditional importance of the northern fishing grounds, 3) high variability of catch rates and income rates, 4) lower value of target species found near the reserve, and 5) social pressures related to the history of community management of the reserve. However, the present study has no informa- tion on how fishing effort, yield, and catch rates near the reserve have changed over the past 20 years since 374 Fishery Bulletin 104(3) reserve establishment. Despite the Apo Island fishery being one of the best studied coral reef fisheries in the world (Alcala and Luchavez, 1981; White and Savina, 1987; Bellwood, 1988; Maypa et al., 2002; Russ et al, 2004; Alcala et al., 2005), the present study is the first to quantify detailed spatial variations in fishing effort, catch, and CPUE at Apo Island. Furthermore, it is not clear if spillover yield to the local fishery has reached its full potential or not. Long-term fishery monitoring is required to answer these questions. Although the reserve probably provides limited direct benefits to the local fishery through spillover, its indirect benefits to the community at Apo Island are considerable. The reserve became the foundation for a program of fishery conservation for the whole island, which eliminated un- sustainable fishing practices such as dynamite fishing and muro-ami drive-net fishing. It played a critical role in enhancing tourism activities on the island, which had a tremendous positive effect on the standard of living of the local community (Russ and Alcala, 1999; White et al., 2002). These indirect effects of reserve establishment are arguably as important as the direct benefits of reserves to developing areas of the world. It remains to be seen whether successful reserves such as Apo Reserve have broader-scale direct benefits to fisheries through net larval export. Acknowledgments This project was funded by a Pew Fellowship in Marine Conservation and an Australian Research Council Grant (DP0209086) to G. R. Russ and A. C. Alcala. Travel to the Philippines for R. A. Abesamis was funded by the Australian Agency for International Development (AusAID). We thank Analie Candido, Mary Aldeon, Francia Candido, and Olivia Alaton for invaluable assis- tance in the collection of data. The Apo Island Protected Area Management Board, Aileen Maypa, Hilconida Calumpong, and staff at SUAKCREM are acknowledged for logistical assistance. Literature cited Abesamis, R. A., and G. R. Russ. 2005. Density-dependent spillover from a marine reserve: long-term evidence. Ecol. Appl. 15:1798-1812. Abesamis, R. A., G. R. Russ, and A.C. Alcala. 2005. Gradients of abundance of fish across no-take marine reserve boundaries: evidence from Philippine coral reefs. Aquatic Conserv. Mar. Freshw. Ecosyst. 15. Alcala, A. C. 1998. Community-based coastal resource management in the Philippines: a case study. Ocean Coast. Manag. 38:179-186. Alcala, A. C, and T. Luchavez. 1981. Fish yield of the coral reef surrounding Apo Island, Negros Oriental, Central Visayas, Philippines. Proc. 4th Int. Coral Reef Symp. 1:70-73. Alcala, A. C, and G. R. Russ. 1990. A direct test of the effects of protective management on abundance and yield of tropical marine resources. J. Cons. Int. Explor. Mer 47:40-47. Alcala, A. C, G. R. Russ, A. P. Maypa, and H. P. Calumpong. 2005. A long-term, spatially replicated, experimental test of the effect of marine reserves on local fish yields. Can. J. Fish. Aquat. Sci. 62:98-108. Alino, P. M., N. E. Palomar, H. O. Arceo, and A. T. Uychiaoco. 2002. Challenges and opportunities for marine protected area (MPA) management in the Philippines. Proc. 9"^ Int. Coral Reef Symp. 2:635-640. Bellwood, D. R. 1988. Seasonal changes in the size and composition of the fish yield from the reef around Apo Island, central Phil- ippines, with notes on methods of yield estimation. J. Fish Biol. 32:881-893. Bohnsack, J. A. 1998. Application of marine reserves to reef fisheries management. Aust. J. Ecol. 23:98-304. Carr, M. H., and D. C. Reed. 1993. Conceptual issues relevant to marine harvest refuges: examples from temperate reef fishes. Can. J. Fish. Aquat. Sci. 50:2019-2028. DeMartini, E. E. 1993. Modelling the potential of fishery reserves for managing Pacific coral reef fishes. Fish. Bull. 91: 414-427. Galal, N., R. F. G. Ormond. and O. Hassan. 2002. Effect of a network of no-take reserves in increas- ing catch per unit effort and stocks of exploited reef fish at Nabq, South Sinai, Egypt. Mar. Freshw. Res. 53:199-205. Cell, R, and C. M. Roberts. 2003. Benefits beyond boundaries: the fishery effects of marine reserves. Trends Ecol. Evol. 18: 448-455 Hilborn, R., K. Stokes, J. J. Maguire, T. Smith, L. W. Botsford, M. Mangel, J. Oresanz, A. Parma, J. Rice, J. Bell, K. L. Cochrane, S. Garcia. S. J. Hall, G. P. Kirkwood, K. Sains- bury, G. Steffanson, and C. Waiters. 2004. When can marine reserves improve fisheries management? Ocean Coast. Manag. 47: 197-205. Hilborn, R., and C. J. Walters. 1992. Quantitative fisheries stock assessment: choice, dynamics and uncertainty, 570 p. Chapman and Hall, New York, NY. Johnson, D. R., N. A. Funicelli, and J. A. Bohnsack. 1999. Effectiveness of an existing estuarine no-take fish sanctuary within the Kennedy Space Center, Florida. N. Am. J. Fish. Manag. 19:436-453. Kelly, S., D. Scott, and A. B MacDiarmid. 2002. The value of a spillover fishery for spiny lob- sters around a marine reserve in Northern New Zeal- and. Coast. Manag. 30:153-166. Kramer, D. L., and M. R. Chapman. 1999. Implications of fish home range size and reloca- tion for marine reserve function. Environ. Biol. Fishes 55:65-79. Maypa, A. P., G. R. Russ, A. C. Alcala, and H. P. Calumpong. 2002. Long-term trends in yield and catch of the coral reef fishery at Apo Island, central Philippines. Mar. Freshw. Res. 53:207-213. McClanahan, T. R., and B. Kaunda-Arara. 1996. Fishery recovery in a coral-reef marine park and its effect on the adjacent fishery. Conserv. Biol. 10:1187-1199. Abesamis et a\ How much does the fishery at Apo Island benefit from spillover from a marine reserve' 375 McClanahan, T. R., and S. Mangi. 2000. Spillover of exploitable fishes from a marine park and its effect on the adjacent fishery. Ecol. Appl. 10:1792-1805. Munro, J. L. 1996. The scope of tropical fisheries and their man- agement. In Reef fisheries (N. V. C. Polunin and C. M. Roberts, eds.), p. 1-12. Chapman and Hall, London. Polacheck, T. 1990. Year around closed areas as a management tool. Nat. Resour. Model. 4:327-354. Rakitin, A. and D. L. Kramer. 1996. Effect of a marine reserve on the distribution of coral reef fishes in Barbados. Mar. Ecol. Prog. Ser. 131:97-113 Roberts, C. M., J. A. Bohnsack, F. Gell, J. P. Hawkins, and R. Goodridge. 2001. Effects of marine reserves on adjacent fisheries. Science 294:1920-1923. Russ, G. R. 2002. Yet another review of marine reserves as reef fish- ery management tools. In Coral reef fishes: dynamics and diversity in a complex ecosystem (P. F. Sale, ed.), p. 421-443. Academic Press Inc. San Diego, CA. Russ, G. R., and A. C. Alcala. 1996. Do marine reserves export adult fish biomass? Evidence from Apo Island, Central Philippines. Mar. Ecol. Prog. Ser. 132:1-9. 1998. Natural fishing experiments in marine reserves 1983-1993: community and trophic responses. Coral Reefs 17:383-397. 1999. Management histories of Sumilon and Apo Marine Reserves, Philippines, and their influence on national marine resource policy. Coral Reefs 18:307-319. 2003. Marine reserves: rates and patterns of recovery and decline of predatory fish, 1983-2000. Ecol. Appl. 13:1553-1565. Russ, G. R., A. C. Alcala, and A. S. Cabanban. 1992. Marine reserves and fisheries management on coral reefs with preliminary modelling of the effects on yield per recruit. Proc. 7th Int. Coral Reef Symp., Guam 2:978-985. Russ, G. R., A. C. Alcala, and A. P. Maypa. 2003. Spillover from marine reserves: the case of Naso vlamingii at Apo Island, the Philippines. Mar. Ecol. Prog. Ser. 264:15-20. Russ, G. R., A. C. Alcala, A. P. Maypa, H. P. Calumpong, and A.T. White. 2004. Marine reserve benefits local fisheries. Ecol. Appl. 14:597-606. Salomon, A. K., N. R Waller, C. Mcllhagga, R. L. Yung, and C. Walters. 2002. Modelling the trophic effects of marine pro- tected area zoning policies: a case study. Aquat. Ecol. 36:85-95. White A. T., and G. C. Savina. 1987. Reef fish yield and nonreef catch of Apo Island, Negros, Philippines. Asian Mar. Biol. 4:67-76. White, A. T., C. A. Courtney, and A. Salamanca. 2002. Experience with marine protected area planning and management in the Philippines. Coast. Manag. 30:1-26. Wilcox, C, and C. Pomeroy. 2003. Do commercial fishers aggregate around marine reserves? Evidence from Big Creek Marine Ecological Reserve, Central California. N. Am. J. Fish. Manag. 23:241-250. Zar, J. H. 1999. Biostatistical analysis, 4th ed., 663 p. Prentice- Hall, Inc., Upper Saddle River, NJ. 376 Abstract — Rockfish (Sebastes spp.) juveniles are often difficult to identify by using morphological characters. This study independently applies morphological characters and a key based on mitochondrial restriction site variation to identify juvenile rockfishes collected in southern California during juvenile rockfish surveys. Twenty-four specimens of Sebastes were examined genetically without knowledge of the morphologi- cal assignment. Seventeen fish were identified genetically as S. semicinc- tus, S. goodei, S. auriculatus, S.jor- dani, S. levis, S. rastrelliger, and S. saxicola. Identities for the remaining fish were narrowed to two or three species: 1) three fish were either S. carnatus or S. chrysomelas; 2) one fish was either S. chlorosticus, S. eos, or S. rosenblatti; and 3) three fish could have been either S. hopkinsi or S. ovalis, the latter for which we now have distinguishing mitochon- drial markers. The genetic and mor- phological assignments concurred except for the identity of one fish that could only be narrowed down to S. hopkinsi or S. semicinctus by using morphological characters. Genetics excluded more species from multispe- cies groupings than did the morpho- logical approach, especially species within the subgenus Sebastomus. Species in the genetically unresolv- able groups may be similar because of recent divergence or because of inter- species introgression. Comparing the identification of southern California juvenile rockfishes (genus Sebastes spp.) by restriction site analysis of the mitochondrial ND3/ND4 region and by morphological characteristics Zhuozhuo Li^ Mary M. Nishimoto^ Milton S. Love^ Anthony J. Gharrett' ' Fisheries Division, School of Fisheries and Ocean Sciences University of Alaska Fairbanks 11120 Glacier Highway Juneau, Alaska 99801 ^ Marine Science Institute University of California Santa Barbara, California 93106-6150 E mail address (for A, J- Gtiarrett, contact author) ffajgcfluatedu Manuscript submitted 11 October 2004 to the Scientific Editor's Office. Manuscript approved for publication 21 September 200.5 by the Scientific Editor. Fish. Bull. 104:376-382 (2006). Sixty-five rockfish species (Sebastes spp.) inhabit the waters along the California coast (Moser, 1996). Within the genus, there is a high degree of similarity among many species in morphological characters. These similarities are in part due to recent divergence, but may also have resulted from convergence of congeners occupy- ing similar habitats. Identification of Sebastes (and most species) is usually based on morphology; however, this approach may fail, especially for iden- tifying sympatric species, which can be similar in coloration and overlap in morphological characters. Juvenile rockfishes are morphologically dis- tinct from larvae and adults (Kendall, 2000), and juvenile stages of many species, especially the pelagic juvenile stage, have not yet been described; only a few species have complete ontogenetic descriptions (Matarese et al.,1989; Moser. 1996). The species of a few Sebastes larvae can be determined and adults can be misidentified. Rockfishes are important ecologi- cally and some species are economi- cally valuable. Sebastes larvae are a large component of ichthyoplankton collections and rank third or fourth in abundance among all fish larvae taken during California Cooperative Fisheries Investigations (CalCOFI) surveys, which cover the entire length of the California and Baja Califor- nia coast and now survey southern California. The ability to identify Sebastes accurately and efficiently at all developmental stages will, in turn, greatly increase our knowledge of their life histories, as well as our management and conservation efforts. An increased understanding of life history variation can improve the sys- tematic descriptions of Sebastes spe- cies, which have mostly been based on the morphology of adults. The analysis of restriction site variation of polymerase chain re- action (PCR)-amplified mtDNA is an effective and relatively simple method for species identification of adult specimens of fish species. The use of restriction enzymes to cleave the DNA can reveal variation in the mtDNA sequence at specific sites, and individuals or species can often be distinguished by the presence or ab- sence of these sites. This method has advantages over sequencing in that it can be conducted relatively easily and used to survey large stretches of a DNA sequence, whereas sequencing, Li et a\ Comparison of the identification of Sebastes spp, by restriction site analysis and by morpfiological cfiaracteristics 377 because of its high cost, is usually limited to a much smaller span of DNA. Closely related species of several genera have been identi- fied with this method; for example, snap- pers (Chow et al., 1993), eels (Aoyama et al., 2000; Lin et al., 2002), billfishes (Innes et al., 1998), and rockfishes (Gharrett et al., 2001). For identification of larvae and ju- veniles, restriction site patterns of adults can be used as references to ensure the re- liability of the results. A limited database was used and a small number of samples were tested for a previous report on larval and juvenile identification of Sebastes, also based on restriction patterns of the mtDNA (Taylor, 1998). The objective of this study was to compare the results of identifying pelagic juvenile rockfishes by the application of a key for de- lineating species based on mtDNA variation of adult rockfishes (Li et al., 2006) and by using only morphological characteristics. The questions we addressed were the following: Are the results for both genetic identifica- tions and morphological identifications con- cordant? If not, how did these two methods differ? The targets of the comparison were samples of pelagic juvenile rockfishes col- lected from southern California. Both ge- netic and morphological analyses were con- ducted without knowledge of the results of the other. Materials and methods Collection of juveniles Juvenile rockfishes were collected from the coast of Santa Barbara, California in June of 1998 and 2000 (Table 1). Fish were captured at night with a modified Cobb mid-water trawl with a nominal 12.2 mxl2.2 m opening and 9-mm mesh net. Average trawling depth ranged from 18 to 29 m below the surface and above bottom depths ranged from 500 to 800 m, except for one specimen that was caught over a bottom depth of 88 m. Generally, specimens were sorted and then frozen on board the vessel. However, specimens from two hauls were sorted and immediately put in 95% ethanol aboard the vessel. Similar collection methods were used for both years. Details of the 1998 survey are provided in Nishi- moto and Washburn (2002). The standard lengths (mm) of the specimens ranged from 16 to 42 mm. Genetic identification Following collection, each fish was identified in the labo- ratory by using morphological characters and meristics and pigmentation patterns. Specimens were chosen for genetic examination based on morphological differences. Table 1 Sample identification number.s, sampling location, standard length (SL), weight, and method of tissue storage of juvenile specimens of Sebastes spp. Missing sample numbers are those for which the DNA could not be amplified: a = frozen on the boat, thawed on ice when genetic material was taken; b=put into ethanol on the boat. Sample number North West latitude longitude SL (mm) Weight !fl M60I Mspl Rsal Genetic assignment Morphological assignment 1 X D A C P S. semicinctus S. semicinctus 2 z A A g K S. hopkinsi or ovalis S. hopkinsi 3 z A K S. hopkinsi or ovalis S. hopkinsi 4 z A g g s. hopkinsi S. hopkinsi 5 z g s. hopkinsi S. hopkinsi 8 X P s. semicinctus S. semicinctus 9 z s. hopkinsi S. hopkinsi or semicinctus 15 Y D 0 ai E S. goodei S. goodei 16 Y D 0 a2 E S. goodei S. goodei 17 0 C D G f C S. saxicola S. saxicola 18 a c S.jordani S.jordani 19 B A q S. levis S. levis 20 z A g K C S. hopkinsi or ovalis S. hopkinsi 21 c C V a A B S. rcistrelliger S. rastrelliger 22 F A i C G C a B S. carnatus or chrysomelas S. carnatus or chrysomelas 23 F a i C G C a B S. carnatus or chrysomelas S. carnatus or chrysomelas 25 F A i C c a B S. carnatus or chrysomelas S. carnatus or chrysomelas 26 F D A E C S. chlorostictus or eos or rosenblatti Sebastomus 27 X D C P S. semicinctus S. semicinctus 31 F D V A A k B S. auriculatus S. auriculatus 44 Z A n g K C S. hopkinsi S. hopkinsi 45 Z A g g C S. hopkinsi S. hopkinsi 48 Z n K S. hopkinsi S. hopkinsi 49 X P S. semicinctus S. semicinctus (as in the case of S. auriculatus) can help to differentiate species with overlapping meristic ranges. Species that could not be identified from their morpho- logical features were assigned to a complex of species or a subgenus. Group 2 in Table 2 includes species that have similar pigmentation patterns during their pelagic juvenile stage. Meristics and head spination are used to identify S. atrovirens. However, the remaining three, S. cartiatus, S. caurinus, and S. chrysomelas, are assigned as a species complex because of the uncertainty in iden- tifying the species. The subgenus Sebastomus includes ten rockfish species that were found in the collection area and that cannot be distinguished by morphological characters and meristics or by pigmentation. Results The way in which the specimens were handled and pre- served affected DNA quality and limited the number of genetic and morphological species assignments that could be compared. Only the DNA of seven of the 25 specimens frozen at sea and later thawed and processed was successfully amplified, whereas the DNA of 17 of the 24 specimens preserved in 95% ethanol at the time of capture was amplified. A total of 24 specimens were identified to species or species group based on restric- tion fragment patterns of the mtDNA using one or more restriction endonucleases (Table 3) and without knowl- edge of the morphological assignment. Table 4 sum- 380 Fishery Bulletin 104(3) 3 ■a c t>D V c ^« m x) c ?! U) ■- f^ o- IS ID * o-fr V ^ T3 to 0) x K- 3 CO — "^ CO C o * ca c y) bo OJ H) ;. S O 73 rn ^4- C !m O O -73 C c bu t-^ 3 C S c 3 "> 2 '■" O CO (A CO CA CO >1 CO o o c p c c 0) a u o b o X c u D. bfl > Cb Q a o CO CO CO 3 S CO CO O 00 0> to 00 C^) ,M ,-H W ,-1 CD oi io m 00 lr~ I~- CJJ CD CD CD CO CO -^ .— T ID "— ' 1—1 I-H ■^ C^ CO (M (M (N tH 1— 1 1— 1 '-' '-' '-' (M CO — T -^ i-H i-H CJ .— < > > 3 3 be bl C C 6 6 U I-H a ^ CO a C CO CO g 2 s s ,0-^-5-5 CM LO CO cocococococooico -^ ^ c ;; ri. P -O ■« -S CO CO CO W CO CO CO CO t/3 CO CO 3 C OJ be J3 3 bo C c ^ S CO ~ o 3 ll CO en marizes the meristic data used to identify species morphologically. When a fish could not be assigned to a species, it was assigned to a species complex or subgenus. The individuals genetically identified to species were six S. hopkinsi, four S. semi- cinctus, two S. goodei, one S. auriculatus, one S. jordani, one S. levis, one S. rastrelliger, and one S. saxicola. The restriction fragment patterns of these juveniles were identical to those of previously observed adult specimens, except for the two S. goodei juveniles, each of which differed by a single difference (presum- ably a site loss in one fish and a site gain in the other) from reference specimens for Mbo I (Table 3 and Li et al., 2006). The genetic species assignment of all but one of these specimens matched the identification based on morphological characters (Table 4). The discrepant fish was identified morphologically as either S. seinicinctus or S. hopkinsi. The specimen had 14 dorsal rays, which is within the range of both species (Table 2) and was a transforming pelagic juvenile that had not yet developed the discernible body pigmenta- tion pattern of older juveniles (Table 4). At the beginning of the study, S. oralis was not included in our database. When we added restriction site information for S. ovalis to the database, we found that the species is very similar to S. hopkinsi in its haplotype profile. Consequently, three individuals morphologi- cally identified as S. hopkinsi were genetically assigned to the complex S. hopkinsi-S. ovalis. The restriction enzyme Dde I can delineate the two species; however, when we discovered the ambiguity, no samples remained for analysis. Based on meristics, the three specimens are more likely S. hopkinsi than S. ovalis. The range of counts for fin rays overlaps for the two species (Table 2). However. S. ovalis common- ly has eight or nine anal rays, and the three specimens in question possessed seven rays that occur commonly in S. hopkinsi (Table 4). Three specimens were assigned to the spe- cies complex S. carnatus-S. chrysomelas-S. caurinus according to morphological criteria. Haplotype information eliminated S. cauri- nus. Genetic analysis of the one specimen that was morphologically assigned to the sub- genus Sebastomus, which includes 13 North Pacific species, reduced the possible species to S. chlorosticus, S. eos, or S. rosenblatti. Discussion Our results demonstrated that restriction site analysis of mtDNA is a simple and effec- tive tool for identifying juveniles of Sebastes Li et al : Comparison of the identification of Sebostes spp by restriction site analysis and by morphological characteristics 381 species. The presumed intraspecific variation observed in S. hopkinsi and S. goodei did not interfere with the application of interspecific variation in species identifica- tion, and our other work (Gharrett et al., 2001; Li, 2004) indicates that intraspecific variation generally does not obscure delineation of species. The key we constructed for mtDNA restriction site variation can be especially helpful in learning about variation in the distinguish- ing morphological characters of the early life stages of Sebastes. One factor that prevented identification of some ju- venile specimens was poor DNA quality. Specimens must be properly preserved to provide high quality DNA for genetic analysis. Our success in amplifying DNA from specimens preserved in several volumes of 95% ethanol immediately after capture was substantially higher than that of specimens that had been frozen after capture and subsequently thawed for morphologi- cal studies. Several small groups of species had similar restric- tion site patterns. Consequently, there was ambiguity in the genetic analysis of those groups. Specifically, we were unable to separate S. carnatus from S. chrys- omelas, and we were unable to distinguish among S. chlorostictus, S. eos, and S. rosenblatti. However, mor- phological characteristics could not reveal a difference between S. carnatus and S. chrysomelas, nor could they distinguish any of the 13 Sebastomus species, including S. chlorostictus, S. eos, and S. rosenblatti. Sebastes carnatus and S. chrysomelas are very closely related and, as adults, differ morphologically only in body coloration. Sebastes carnatus has flesh-colored blotches on an olive-brown background, and S. chrys- omelas has yellow blotches on a black background (Love et al., 2002). They are found across similar geographic ranges, from northern California to Baja California, and are often sympatric, inhabiting rocky substrates, although they have different depth distributions (Lar- son, 1980), and their segregation is maintained by the social dominance of S. chrysomelas (Larson, 1980; Hoel- zer, 1987). The two species may also differ slightly in gill-raker counts: S. carnatus typically has 28-30, and S. chrysomelas typically has 27-28 (Hallacher and Rob- erts, 1985). Their taxonomic status as distinct species is generally recognized from their different colorations, even though no fixed biochemical or genetic differences have been reported (Seeb, 1986; Seeb and Kendall, 1991; Hunter, 1994; Alesandrini, 1997). A recent study with the use of microsatellites in the nuclear DNA and mtDNA sequence variation has provided evidence that the two species S. carnatus and S. chrysomelas are genetically divergent (Narum, 2000). Sebastes chlorostictus, S. eos, and S. rosenblatti are also morphologically similar, are present sympatrically, and are closely related (Chen, 1971; Love, 1996; Rocha- Olivares et al., 1999; Love et al., 2002). As adults, these species segregate by depth to some extent; S. rosenblatti occupies a somewhat narrower range than the other two (Love et al., 1990). Cladistic analyses indicate that S. rosenblatti is the most ancient of the three species (Rocha-Olivares et al., 1999). Estimates of divergence times of the subgenus Sebastomus may indicate that the three species are the result of the most recent spe- ciation events within the subgenus, which may have begun less than 140,000 years ago (Rocha-Olivares et al., 1999). The assignment of four specimens to multispecies groups in this study indicates the need for further ef- fort to develop markers that will delineate the species in question. Approaches include screening additional regions of the mtDNA and application of additional restriction enzymes. If additional mtDNA regions and restriction enzymes do not provide species-specific in- formation, other molecular techniques, such as the use of microsatellites, should be considered. Acknowledgments This work represents, in part, the master's thesis work of Z. Li at the University of Alaska Fairbanks. The proj- ect was supported by funding from the U.S. Geological Service (Biological Resources Division) Western Regional Office in Seattle, WA (R.W.O. 32) to AJG. Funding for MMN and MSL was provided by the Biological Resources Division under a cooperative agreement no. 1445-CA09- 95-0836 with the Marine Science Institute, University of California, Santa Barbara. Literature cited Alesandrini, S. 1997. Genetic differentiation and population structure in the gopher rockfish {Sebastes carnatus) and the black- and-yellow rockfish (S. chrysomelas) along the California coast. M.S. thesis, 33 p. Univ. California. Santa Cruz. Santa Cruz, CA. Aoyama, J., S. Watanabe, M. Nishida, and K. Tsukamoto. 2000. Discrimination of catadromous eels of genus Anguilla using polymerase chain reaction-restriction fragment length polymorphism analysis of the mito- chondrial 168 ribosomal RNA domain. Trans. Am. Fish. Soc. 129:873-878. Chen, L.-C. 1971. Systematics, variation, distribution, and biol- ogy of rockfishes of the subgenus Sebastomus (Pisces, Scorpaenidae, Sebastes). Bull. Scripps. Inst. Oceaogr. 18:1-115. Chow, S., M. E. Clarke, and P. J. Walsh. 1993. PCR-RFLP analysis on thirteen western Atlantic snappers (subfamily Lutjaninae): a simple method for species and stock identification. Fish. Bull. 91:619- 627. Gharrett, A. J., A. K. Gray, and J. Heifetz. 2001. Identification of rockfish (Sebastes spp.) from restriction site analysis of the mitochondrial ND-3/ ND-4 and 12S/16S rRNA gene regions. Fish. Bull. 99:49-62. Hallacher, L. E., and D. A. Roberts. 1985. Differential utilization of space and food by the inshore rockfishes (Scorpaenidae: Sebastes) of Carmel Bay, California. Environ. Biol. Fish. 12:91-110. 382 Fishery Bulletin 104(3) Hoelzer, G. 1987. The effect of early experience on aggression in two territorial scorpaenid fishes. Environ. Biol. Fish. 19:183-194. Hunter, K. 1994. Incipient speciation in rockfish (Sebastes carnatus and Sebastes chrysomelas). M.S. thesis, 38 p. Cali- fornia State Univ. Northridge. Northridge, CA. Innes, B. H., P. M. Grewe, and R. D. Ward. 1998. PCR-based genetic identification of marlin and other billfish. Mar. Freshw. Res. 49:383-388. Kendall, A. W., Jr. 2000. An historic review of Sebastes taxonomy and systematics. Mar. Fish. Rev. 62:1-23. Laidig, T. E., and P. B. Adams. 1991. Methods used to identify pelagic juvenile rockfish (genus Sebastes) occurring along the coast of central California. NOAA Tech. Memo. NMFS-SWFSC-166, 180 p. Laidig, T. E., K. M. Sakuma, and M. M. Nishimoto. 1996. Description of pelagic larval and juvenile Sebastes saxicola (Family Scorpaenidae), with an examination of larval growth. Fish. Bull. 94:289-299. Larson, R. J. 1980. Competition, habitat selection, and the bathymetric segregation of two rockfish iSebastes) species. Ecol. Monogr. 50:221-239. Li, Z. 2004. Phylogenetic relationship and identification of juve- niles of the genus Sebastes. M.S. thesis, 146 p. Univ. Alaska Fairbanks, Fairbanks, AK. Li, Z., A. K. Gray, M. S. Love, A. Goto, T. Asahida., and A. J. Gharrett. 2006. A key to selected rockfishes (Sebastes spp.) based on mitochondrial DNA restriction fragment analysis. Fish. Bull. 104:182-196. Lin, Y.. Y. Poh, S. Lin, and C. Tzeng. 2002. Molecular techniques to identify freshwater eels: RFLP analyses of PCR-amplified DNA fragments and allele-specific PCR from mitochondrial DNA. Zool. Studies 41:421-430. Love, M. S. 1996. Probably more than you want to know about the fishes of the Pacific coast, 386 p. Really Big Press, Santa Barbara, CA. Love, M. S.. P. Morris, M. McCrae, and R. Collins. 1990. Life history aspects of 19 rockfish species (Scor- paenidae: Sebastes) from the southern California Bight. NOAA Tech. Rep. NMFS-87, 38 p. Love, M. S., M. Yoklavich, and L. Thorsteinson (eds.l. 2002. The rockfishes of the Northeast Pacific, 405 p. Univ. California Press, Berkeley, CA. Matarese, A. C, A. W. Kendall Jr., D. M. Blood, and B. M. Vinter. 1989. Laboratory guide to early life history stages of northeast Pacific fishes. NOAA Tech. Rep. NMFS Circ. 80, 656 p. Moser, H. G. 1996. Scorpaenidae: scorpionfishes and rockfishes. Ir> The early stages of fishes in the California Current region (H. G. Moser, ed.), p. 793-795. CalCOFI Atlas no. 33. Moser, H. G., E. H. Ahlstrom, and E. M. Sandknop. 1977. Guide to the identification of scorpionfish larvae (family Scorpaenidae) in the eastern Pacific with com- parative notes on species o( Sebastes and Helicolenus from other oceans. NOAA Tech. Rep. NMFS Circ. 402, 71 p. Narum, S. 2000. Assortatative mating and genetic structure of gopher rockfish (S. carnatus) and black-and-yellow rock- fish (S. chryso/nelas): a case of incipient speciation. M.S. thesis, 90 p. LIniv. California San Diego, San Diego, CA. Nishimoto, M. M., and L. Washburn. 2002. Patterns of coastal eddy circulation and abundance of pelagic juvenile fish in the Santa Barbara Channel, California, USA. Mar. Ecol. Prog. Ser. 241:183-199. Rocha-Olivares, A., R. H. Rosenblatt, and R. D. Vetter. 1999. Molecular evolution, systematics, and zoogeog- raphy of the rockfish subgenus Sebastomus {Sebastes, Scorpaenidae) based on mitochondrial cytochrome 6 and control region sequences. Mol. Phylogenet. Evol. 11:441-458. Sakuma, K. M., and T. E. Laidig. 1995. Description of larval and pelagic juvenile chilipep- per, Sebastes goodei (family Scorpaenidae), with an exam- ination of larval growth. Fish, Bull. 93:721-731. Seeb, L. W. 1986. Biochemical systematics and evolution of the Scor- paenidae genus Sebastes. Ph.D. diss., 176 p. Univ. Washington, Seattle, WA. Seeb, L. W., and A. W. Kendall Jr. 1991. Allozyme polymorphisms permit the identifi- cation of larval and juvenile rockfishes of the genus Sebastes. Environ. Biol. Fish. 30:191-201. Taylor, C. A. 1998. Phylogenetics and larval identification of rock- fishes (genus Sebastes) using mitochondrial DNA techniques. M.S. thesis, 118 p. San Diego State Univ., San Diego, CA. 383 Abstract — Although bocaccio iSe- bastes paucispinis) was an economi- cally important rockfish species along the west coast of North America, overfishing has reduced the stock to about TAVc of its former unfished population. In 2003, using a manned research submersible, we conducted fish surveys around eight oil and gas platforms off southern California as part of an assessment of the potential value of these structures as fish habi- tat. From these surveys, we estimated that there was a minimum of 430,000 juvenile bocaccio at these eight struc- tures. We determined this number to be about 20% of the average number of juvenile bocaccio that survive annu- ally for the geographic range of the species. When these juveniles become adults, they will contribute about one percent (0.8%) of the additional amount offish needed to rebuild the Pacific Coast population. By compari- son, juvenile bocaccio recruitment to nearshore natural nursery grounds, as determined through regional scuba surveys, was low in the same year. This research demonstrates that a relatively small amount of artificial nursery habitat may be quite valuable in rebuilding an overfished species. Potential use of offshore marine structures in rebuilding an overfished rockfish species, bocaccio (Sebastes paucispinis) Milton S. Love^ Donna M. Schroeder' William Lenarz^ Alec MacCalP Ann Scarborough BulH Lyman Thorsteinson^ ' Marine Science Institute, University ol California Santa Barbara, Calilornia, 93160 E-mail address (for M, S. Love: ioveia'lifesci.ucsbedu 2 P.O. 251 Kentlield, California 94914 3 NOAA/NfVIFS Santa Cruz Laboratory Santa Cruz, California 95060 " U.S. Minerals Management Service 770 Paseo Camarillo Camarillo, California 93010 ^ Western Fisheries Research Center U.S. Geological Survey 6505 NE 65'^' St Seattle, Washington 98115 Manuscript submitted 2 September 2005 to the Scientific Editor's Office. Manuscript approved for publication 26 September 2005 by the Scientific Editor. Fish. Bull. 104:383-390 (2006). Beginning in 1995, annual surveys of fish assemblages at oil and gas plat- forms and natural reefs throughout southern California were conducted with a research submersible (summa- rized in Love et al., 1999, 2000, 2003). Many California oil and gas platforms harbor three fish assemblages: those that occupy the shell mound area sur- rounding the base of the platform; those that occupy the waters adja- cent to the platform bottom, and those that occupy the midwater. Rockfishes (genus Sebastes), of about 35 spe- cies, dominate these assemblages. The shell mound assemblage is com- posed primarily of juvenile rockfishes, dwarf rockfishes, and other species; the platform bottom assemblage is composed of adult and subadult fishes; and the midwater assemblage of most platforms (and the bottoms of some mid-depth platforms) is dominated by young-of-the-year (YOY) and older juvenile rockfishes that comprise at least 28 species. These fishes are rarely more than 20 cm long (total length). In the midwaters of most Cal- ifornia platforms, there are only low densities of predatory reef fish species, such as kelp bass (Paralabax clath- ratus) and cabezon (Scorpaenichthys marmoratus), or semipelagic, large predatory species, such as Pacific bar- racuda (Sphyraena argentea) and yel- lowtail (Seriola lalandi). The proper disposition of the approximately 6000 marine offshore oil and gas platforms and associated structures now in ser- vice worldwide is in dispute. There are 27 platforms off California and, as in other parts of the world, there is considerable debate over the ultimate fate of these structures once they are uneconomical to operate (Schroeder and Love, 2004). In this article, we focus on the role that some artificial structures play as rockfish nursery habitat off California. During some years, we have noted particularly high densities of YOY bo- caccio {Sebastes paucispinis), widow (S. entomelas), squarespot (S. hop- kinsi), and blue (S. mystinus) rock- 384 Fishery Bulletin ]04(3) Figure 1 Examples of young-of-the-year (YOY) and other juvenile fishes found in high densities at platform Gilda, October 2003. (A) YOY bocaccio iSebastes paucispinis); (B) a YOY lingcod iOphiodon elongatus); (C) one-year-old vermilion rockfish iSebastes miiiiatus). fishes around a number of these platforms. Similar high densities of young rockfishes were noted in the late 1950s at two nearshore platforms off Santa Barbara (Carlisle et al.. 1964). The densities of YOY rockfishes are usually higher at platforms than at most natural outcrops (Sehroeder et al., 2000; Love et al., 2003). As on natural outcrops, however, YOY rockfish recruitment to platforms is highly variable from year to year. In 2003, while conducting fish surveys around eight oil platforms in southern California, we ob- served high densities of YOY rockfishes (e.g., bocac- cio and widow rockfish), YOY lingcod (Ophiodon elongatus), and one-year-old vermilion rockfish (S. miniatus) (Fig. 1). Given the uncertainty regarding the role that platforms might play as fish habitat, we were interested in understanding how important young bocaccio, and by extension the platforms, might be to the stock and populations of this spe- cies in the region. Because of the severely overfished status of bocaccio and because a stock assessment model had been developed by NOAA Fisheries, we focused our attention on this species. Along the Pacific Coast of North America, bocaccio have been reduced to about 7.4% of their unfished population (MacCallM. Bocaccio are relatively long- lived (to over 50 years) and have extremely variable annual juvenile recruitment success; over natural outcrops, large year classes are found about once a decade (Tolimieri and Levin. 2005). YOY bocaccio settle from a juvenile micronektonic stage to shallow high-relief habitats (such as kelp beds) and usually migrate into deeper waters within one year. Histori- cally, the species was abundant from Oregon to at least northern Baja California (Love et al., 2002). In this article we focus on the large recruitment of YOY bocaccio observed at platforms in the Santa Barbara Channel in 2003. We estimate the mini- mum number of YOY bocaccio at eight platforms in the Santa Barbara Channel, southern California, and estimate the contribution these juveniles may make (as surviving adults) to the rebuilding of the overfished stock. We also compare the densities of YOY bocaccio around the platforms with those determined during the same year in studies on natural reefs throughout southern, and parts of central, California. Material and methods Between 10 and 15 October 2003, we surveyed the jacket (horizontal beams and vertical supports) 1 MacCall, A. D. 2003. Status of bocaccio off California in 2003. In Status of the Pacific coast groundfish fish- ery through 2003 stock assessment and fishery evalu- ation (vol. II, v + 56 p. Pacific Fishery Management Council, 7700 NE Ambassador Place, Suite 200, Portland, OR 97220. Love et al Potential use of offshore marine structures In rebuilding an overflsfied rockflsfi species 385 of eight platforms in the Santa Barbara Channel (Fig. 2) using the Delta research submersible, a 4.6-meter, 2-person vessel operated by Delta Oceanographies of ^ J Of 34 30'! Oxnard, California. In the platform mid- water, we conducted surveys along each of the platform's horizontal beams, located at ^-^^ 20- to 30-m intervals between near-sur- face waters and the bottom. The shallowest beams were situated at depths of 15-34 m; 34 k thus the uppermost parts of the platform were not surveyed. On the seafloor, fish surveys were conducted next to the plat- 34 oc form bottom and over the shell mounds that surround platforms. Because of poor water visibility, we were unable to survey either the platform bottoms or the shell mounds of four platforms (i.e.. A, B, C, and Hillhouse). We conducted belt transects on the shell mound at an average distance of approximately 7.75 m from the platform and around the platform base and horizon- tal beams at a distance of approximately 2 m from the platform, while the submers- ible maintained a speed of about 0.5 knots. Divers estimated transect lengths by first estimating velocity over a short course, using twin laser beams as an aid to estimate their lengths. Submersible surveys were conducted during daylight hours between one hour after sunrise and two hours before sunset. During each transect, the researcher made observations from a view- ing port on the starboard side of the submersible. An externally mounted hi-8-mm video camera with associ- ated lights filmed the same viewing fields as seen by the observer. Images recorded by the camera were laid down on tape. The observer identified, counted, and estimated the lengths of all fishes and verbally recorded those data on the videotape. All fishes within two meters of the submersible were counted. Fish lengths were estimated by using a pair of parallel lasers mounted on either side of the external video camera. The projected reference points were 20 cm apart and were visible both to the observer and in the video camera image. We assumed that bocaccio 20 cm or less were YOY, in accordance with growth studies of bocaccio. Many years of experience along the Pacific Coast have shown that if the Delta is moving at a constant and slow rate of speed, as in these surveys, there is very little obvious effect on most fishes, particularly rockfishes (M. Love-; O'Connell^ Yoklavich^). Certainly, we noticed virtually no movement from most of the 120'30' 120°20' 120°10' 120"00' 119"50' 119"40' 119"30' 119"20' 119'10'W I \ \ \ I I , I I Sania Barbara 2 Love, M. 2002. Unpubl. data. Marine Science Institute, University of California, Santa Barbara, CA, 93160. ■* O'Connell, V. 2004. Persona! commun. Alaska Depart- ment of Fish and Game, 304 Lake St. Room 103 Sitka, AK 99835. ^ Yoklavich, M. 2004. Personal commun. NOAA/NMFS Santa Cruz Laboratory, 110 Shaffer Road Santa Cruz, CA 95060. ^3nta San Miguel I Ventura sland^l Island _ S^— ^ x-s. Santa Cruz Island Anacapa Island Figure 2 Location of all oil and gas plat- forms in the Santa Barbara Chan- nel region of California. Platforms surveyed during October 2003 are named. fishes in our study as the research submersible passed by. Unless hidden in complex substrate, fishes as small as about 5 cm in length are readily visible within two meters of the submersible. Using the data from these surveys, we computed the densities (number offish per m-) of YOY bocaccio at the midwater horizontal beams, bottom (when surveyed), and shell mounds (when surveyed) of each platform. From these density estimates, we calculated the abun- dance of YOY bocaccio at each platform (for protocols see Love-^). These abundance estimates are conservative for four reasons. First, we conducted the platform surveys about two meters from the platform and our belt transects count only those YOY bocaccio within two meters of the submersible. Particularly in platform midwaters, densities of YOY rockfishes tend to be lower outside the platform framework (the zone traversed by the submers- ible) than inside. Second, beginning a few months after settling on platforms, the more open water species, such as bocaccio, blue, widow, squarespot, and yellowtail rockfishes, become less closely associated with structure. There appears to be an ontogenetic shift away from a close association with habitat to a wider association, where these fish wander away from the immediate vicin- ity of the platform. This shift is particularly true dur- ing years with high densities (as in 2003), when large schools wander freely within the relatively open waters of the platform jacket. Third, as noted above, at each 5 Love, M. 2005. Methods used to estimate abundance in numbers of young-of-the-year (YOY) bocaccio. Website: http://www.id.ucsb.edu/lovelab [accessed on 26 September 2005]. 386 Fishery Bulletin 104(3) platform we did not survey in the midwaters above the shallowest horizontal beam (at depths of 15-34 m). Thus, some of the potential bocaccio nursery grounds were not sampled. Fourth, for those platforms where visibility was too poor for us to sample the bottom or shell mound (A, B, C, and Hillhouse), we estimated YOY bocaccio abundances only for that part of the platform (the midwater) that could be surveyed. However, although visibility was too poor to conduct operations on the bottom, we could see high densities of bocaccio in- side the platform below the submersible at several plat- forms when we conducted midwater surveys. Thus, on these platforms, it is highly likely that the YOY bocaccio aggregation extended well below the level of our sur- veys and that these aggre- gations contained substan- tially more of these fishes than we estimated. We also estimated the level of recruitment of YOY bocaccio in 2003 to natu- ral reefs using data from a variety of sources. In this overview, we included 1) our submersible surveys con- ducted at depths at which YOY bocaccio may reason- ably recruit (down to about 100 m), 2) a scuba survey conducted by D. M. Schro- eder, and 3) data from near- shore rocky reef fish surveys conducted by other research- ers. The latter surveys were conducted by using scuba at depths of 25 m or less. Results YOY bocaccio were observed at seven of the eight plat- forms we surveyed. We observed a total of 10,785 YOY bocaccio in the survey (Table 1). We saw no YOY bocaccio at depths of 20 m and less, none at 226 m, and relatively high densities at depths between about 25 and 80 m. Mean YOY bocaccio densities varied between platforms. Platform Grace (368.0 fish/100 m2) and Gilda (95.2 Table 1 Densities of young-of-the-year bocaccio (Sebastes pauci spinis) at eight platforms in the Santa Barbara Channel, October 2003. Because of poor visibility, the zones of some piat- forms were not surveyed; these are marked N/A. Number Density Platform Platform zone surveyed Depth (m) offish (fish/100 m2) A midwater 15 0 0 32 0 0 no base or shell mound N/A N/A N/A B midwater 20 0 0 30 13 4.8 40 165 57.7 no base or shell mound N/A N/A N/A C midwater 20 0 0 30 62 17 45 127 32.1 no base or shell mound N/A N/A N/A Gail midwater 32 2 0.6 47 27 6.7 69 18 4.1 90 3 0.8 111 25 5.6 138 1 0.2 164 1 0.2 195 4 0.7 base 226 0 0 shell mound 226 0 0 Gilda midwater 25 55 13 42 24 6.2 base 62 1023 242.4 shell mound 62 1078 169.6 Grace midwater 26 4006 476.8 45 1537 193.1 68 1124 258 70 845 171 80 92 221.5 base 96 13 2.6 shell mound 96 0 0 Hillhouse midwater 20 0 0 35 506 140 no base or shell mound N/A N/A N/A Holly midwater 34 18 4 base 60 12 2.1 shell mound 60 4 0.4 Total 10,785 fish/100 m^) harbored the highest densities, whereas no YOY bocaccio were observed at Platform A (Fig. 3). Using the methods detailed in Love,'^' we estimated that there was a minimum of 433,682 YOY bocaccio at the seven platforms (Table 2). Using the model STATC (details of the model are giv- en in MacCall'), which is the current basis for fishery management under the Pacific Fisheries Management Council, we determined that the 433,682 YOY bocaccio living around these platforms constitute about 20% of Love et al Potential use of offshore marine structures in rebuilding an overfished rockfish species 387 4 Sites 0 Density (fish/100 m-') of young-of-the-year bocaccio 2003 Figure 3 Densities of YOY bocaccio (Sebaates paucispini.'i) (fish/100 m-) in southern and parts of central California in 2003. Surveys were conducted at sites that were likely YOY bocaccio recruitment habitats, primarily at nearshore kelp beds and rocky outcrops. Surveys by M. Love were conducted from the research submersible Delta; the others were conducted by scuba divers. The open stars represent platforms that were not surveyed; the closed stars represent platforms that were surveyed. the average yearly value, and 40% of the median value for the entire range of the species (based on YOY esti- mated abundances between 1990 and 2003). From the same model, we estimated that these young fish will eventually contribute slightly less (0.8%) than one per- cent of the additional amount of fish needed to rebuild the Pacific Coast stock. By comparison with most of the platforms we sur- veyed, there was low or no YOY bocaccio recruitment to natural reefs in southern and parts of central Cali- fornia during 2003 (Fig. 3). During 2003, surveys were conducted at 64 sites and no YOYs were observed at 50 of these reefs; most of the other 14 sites harbored very low densities. The highest density noted at any natural outcrop (8.9 fish/100 m^), observed at a small kelp bed at Anacapa Island, was much lower than that at five of the surveyed platforms. Discussion Table 2 Estimated numbers of young-of-the-year bocaccio (Sebastes paucispinis) inhabiting eight platforms in the Santa Barbara Channel, 2003. Note that the surveys of four platforms, Hillhou se, C, B, and A, did not nclude the bottom of the platform and its surrounding shell mound. Platform name Estimated abundance Grace 353,411 Gilda 39,854 HiUhouse (partial) 22,171 C (partial) 5977 B (partial) 5911 Gail 5521 Holly 837 A (partial) 0 Total 433,682 According to our estimates, the YOY bocaccio living around seven platforms in the Santa Barbara Channel were important to the regional bocaccio population, and it appears that a substantial portion of the YOY bocaccio 388 Fishery Bulletin 104(3) recruitment that occurred in the Santa Barbara Channel during 2003 took place at the surveyed platforms. Because it is likely that most reefs both attract and produce fishes, discussions of reef function have evolved from the single issue of attraction versus production to an evaluation of the ecological performance of fishes at natural and human-made habitats. Reviews and stud- ies that compare natural and human-made reefs as juvenile fish habitats should address: 1) a comparison of the survival rate of young fishes (including preda- tion, growth rates, and survival after emigration, if that occurs) at the two habitat types, 2) the density of recruiting juveniles at a human-made reef versus that at surrounding natural reefs, 3) the possibility that a human-made reef is drawing recruiting juvenile fishes from nearby natural reefs, 4) the source of juvenile fishes found on a human-made reef — either from the plankton or from natural reefs (discussed in a number of papers including Bohnsack et al., 1994; Carr and Hixon, 1997; Pickering and Whitmarsh, 1997). What is the survival rate of young fishes that recruit from the plankton to platforms and those that recruit to natural reefs? What appears to be a unique feature of the midwaters of many California platforms (compared to both natural reefs and to Gulf of Mexico platforms) is that many may act as juvenile (particularly YOY) fish refuges. The midwaters of platforms do not act as struc- tural refuges. That is, they are not dotted with crevices, caves, and other complexities that allow fish to hide (Hixon and Beets, 1993). Rather, platforms may afford spatial refuges because, in general, concentrations of adult reef fishes tend to be found at the bottom of plat- forms (except for those platforms that harbor almost no adult fishes at all) rather than in the midwaters. Tran- sitory and migrating piscivorous species, such as jacks and barracuda (taxa that are abundant around Gulf of Mexico platforms [Stanley and Wilson, 2000]) are also not common around most California platforms (Love et al., 2003). We have also not observed high densities of either fish-eating pinnipeds or sea birds around these structures. In addition, studies (Schroeder et al., 2005) demonstrate that predation rates on small fishes are lower in platform midwaters than at natural reefs. Several lines of evidence demonstrate that YOY bo- caccio recruit to platforms from the plankton and that the platforms do not attract previously settled fish from natural reefs. Studies of the seasonal timing of plat- form recruitment pulses and the size of the recruiting individuals are consistent with the hypothesis that young bocaccio settle on the platforms from the pe- lagic environment (Nishimoto et al., 2005; Schroeder'^; Nishimoto'). Are these platforms, then, drawing away pelagic juveniles that would have settled instead on nat- ural outcrops? Although it is possible that some pelagic ^ Schroeder, D. M. 2002. Unpubl. data. Marine Science Institute, University of California, Santa Barbara, CA 93106. ' Nishimoto, M. 2005. Personal commun. Marine Science Institute, University of California, Santa Barbara, CA 93106. juveniles that had settled on platforms would have found natural reefs, it is likely that most of these fish would not have survived in the absence of these structures. As an example, Emery et al. (2006) simulated near-surface drift trajectories from high-frequency radar current measurements to model movements of pelagic juvenile bocaccio in the vicinity of a platform off Pt. Arguello, central California. They estimated that for 1999 and 2002, a minimum of 76% and 69% of the modeled pe- lagic juvenile bocaccio would not have found natural habitat on which to settle. Although the platforms we surveyed were east of the platform examined in the Emery et al. (2006) study, it is unreasonable to conclude that these relatively small platforms are filtering the majority of pelagic juvenile bocaccio from the pelagic environment to the exclusion of natural reefs. Rather, it is possible that some platforms provide recruitment habitat for fishes that would otherwise be lost from the population. What is the fate of YOY bocaccio that settle around platforms? In general, YOY bocaccio emigrate from shallower platforms within two years. At least some of these fish migrate to natural reefs, because young bocaccio tagged around Platforms A and B in 1978 and 1979 were recaptured on natural reefs as adults as much as 150 km from the tagging sites (Hartman, 1987). Our observations of bocaccio on the deeper Plat- form Gail demonstrate that many of the YOY bocaccio that recruited to that platform from the plankton in 1999 remained in the area (migrating to the bottom in 2000) and matured into adults (Fig. 4). To summarize, the weight of evidence implies that some of the platforms in southern California produce bocaccio. During some years, young bocaccio settle from the plankton to the platforms in large numbers. This settlement may occur even when juvenile recruitment to natural reefs in the same region is low. Compared to the structure of most natural reefs, the structural complexity and high vertical profile of platforms prob- ably provide pelagic juvenile rockfishes with a relatively strong stimulus to trigger settlement. Mortality on plat- form YOY bocaccio from predation may be relatively low because of a relative scarcity of predators. There is also evidence that platforms retain pelagic bocaccio juveniles that would otherwise have been carried into inhospitable offshore waters. At least some bocaccio that emigrate from platforms survive to populate natural reefs, whereas, as in the case of one platform, some YOY bocaccio remain and mature into adults. Thus, there is the likelihood that many of the YOY bocac- cio we observed at platforms will either emigrate and seed natural reefs or will reside at the platforms and reproduce. Clearly, our research was subject to a number of as- sumptions and the ambiguities that may be associated with these assumptions. These assumptions add some degree of uncertainty both to our estimates of YOY bo- caccio abundances and their subsequent contribution to stock rebuilding. However, we consider this research to be sufficiently strong to demonstrate the importance of Love et al Potential use of offshore marine structures in rebuilding an overfished rockfish species 389 I I I I I I I 10 15 20 25 30 35 40 45 50 55 60 Estimated total length (cm) rr 65 70 Figure 4 Size-frequency histogram of all bocaccio iSebastes paucispi- nis) observed in all depths at Platform Gail from the research submersible Delta, 1995 to 2004. A relatively strong year class recruited in 1999 to the platform midwaters. We observed these fishes at the bottom in 2000 and they remained there through 2004. The near cessation of growth of these fishes between 2003 and 2004 approximates the size at which 1009f of the fish are mature (Love et al., 2002). spatial structure in driving marine population dynam- ics, where a small amount of nursery habitat, either human-made or natural, may disproportionately im- pact a widely distributed species. It also highlights the importance of understanding the potential role of any marine habitat before that habitat is altered, either purposefully or unintentionally. Acknowledgments This research was funded by the Biological Resources Division, U. S. Geological Survey (National Offshore Environmental Studies Program 1445-CA-0995-0386) based on an information need identified by the Miner- als Management Service's Pacific OCS Regions, and by the California Artificial Reef Enhancement Pro- gram (CARE). We thank Jay Carroll, Matt Craig, David Huang, David Kushner, Brent Mardian, Dan Pon- della, and Dan Reed for sharing their fish survey data with us. We thank the pilots of the submersible Delta, Chris Ijames, Joe Lily, and Dave Slater, for their very professional handling of the technical aspects of that survey, as well as the crews of the RV Cavalier and RV Valero. Linda Snook managed the survey data. Literature cited Bohnsack, J. A., D. E. Harper, D. B. McClellan, and M. Hulsbeck. 1994. Effects of reef size on colonization and assem- blage structure of fishes at artificial reefs off east- ern Florida, USA. Bull. Mar. Sci. 55:796-823. Carlisle, J. G., Jr., C. H. Turner, and E. E. Ebert. 1964. Artificial habitat in the marine environ- ment. Calif. Dep. Fish Game, Fish Bull. 124, 93 p. Carr, M. H., and M. A. Hixon. 1997. Artificial reefs: the importance of comparisons with natural reefs. Fisheries 22(4):28-33. Emery, B. M., L. Washburn. M. Love, M. M. Nishimoto, and J. C. Ohlmann. 2006. Do oil and gas platforms off California reduce recruitment of bocaccio iSebastes paucispiiiis) to natural habitat? An analysis based on trajectories derived from high frequency radar. Fish. Bull. 104:391-400. Hartman, A. R. 1987. Movement of scorpionfishes (Scorpaenidae; Sebastes and Scorpaena) in the Southern California Bight. Calif. Fish Game 73:68-79. Hixon, M. A., and J. P. Beets. 1993. Predation, prey refuges, and the structure of coral-reef fish assemblages. Ecol. Monogr. 63(1):77-101. Love, M. S., J. E. Caselle, and L. Snook. 1999. Fish assemblages on mussel mounds sur- rounding seven oil platforms in the Santa Barbara Channel and Santa Maria Basin. Bull. Mar. Sci. 65:497-513. Love, M. S., J. E. Caselle, and L. Snook. 2000. Fish assemblages around seven oil plat- forms in the Santa Barbara Channel area. Fish. Bull. 98:96-117. Love, M. S., M. Yoklavich, and L. Thorsteinson. 2002. The rockfishes of the Northeast Pacific, 405 p. Univ. California Press, Berkeley, CA. Love, M. S., D. M. Schroeder, and M. M. Nishimoto. 2003. The ecological role of oil and gas production plat- forms and natural outcrops on fishes in southern and central California: a synthesis of information. LT.S. Dep. Int., Minerals Management Service (MMS), Outer Con- tinental Shelf (OCS) Study MMS 2003-032, 127 p. Nishimoto, M. N., L. Washburn, M. S. Love, D. Schroeder, and B. Emery. 2005. Is the delivery of juvenile fishes settling on offshore platforms linked to transport by ocean currents? Ab- stract in 8"' international conference on artificial reefs and artificial habitats, Biloxi, MS, p. 46. Pickering, H., and D. Whitmarsh. 1997. Artificial reefs and fisheries exploitation: a review 390 Fishery Bulletin 104(3) of the attraction versus production debate, the influ- ence of design and its significance for policy. Fish. Res. 31:39-59. Schroeder, D. M., and M. S. Love. 2004. Ecological and political issues surrounding decom- missioning of offshore oil facilities in the Southern Cali- fornia Bight. Ocean Coast. Manage. 47:21-48. Schroeder, D. M.. M. S. Love, and M. Nishimoto. 2005. Comparative juvenile reef fish recruitment and mortality between offshore oil-gas platforms and natu- ral reefs. Abstract in S"" international conference on artificial reefs and artificial habitats, Biloxi, MS, p. 62. Schroeder, D. M., A. J. Ammann. J. A. Harding, L. A. Mac- Donald, and W. T. Golden. 2000. Relative habitat value of oil and gas production platforms and natural reefs to shallow water fish assem- blages in the Santa Maria Basin and Santa Barbara Channel, California. In Fifth California Islands sym- posium: 29 March-1 April 1999, p. 493-498. U.S. Min- erals Management Service (MMS) and Santa Barbara Museum of Natural History, Outer Continental Shelf (OCS) Study MMS 99-0038 (CD-ROM). Stanley. D. R., and C. A. Wilson. 2000. Seasonal and spatial variation in the biomass and size frequency distribution of the fish associated with oil and gas platforms in the northern Gulf of Mexico. U. S. Dep. Int., Minerals Management Ser- vice (MMS), Outer Continental Shelf (OCS) Study MMS 2000-005, 252 p. Tolimieri, N., and P. S. Levin. 2005. The roles of fishing and climate in the popu- lation dynamics of bocaccio rockfish. Ecol. Appl. 15:458-468. 391 Abstract — To investigate the pos- sibility that oil and gas platforms may reduce recruitment of rockfishes iSebastes spp.) to natural habitat, we simulated drift pathways (termed "trajectories" in our model) from an existing oil platform to nearshore habitat using current measurements from high-frequency (HF) radars. The trajectories originated at Platform Irene, located west of Point Concep- tion, California, during two recruiting seasons for bocaccio iSebastes paii- cispinis): May through August, 1999 and 2002. Given that pelagic juvenile bocaccio dwell near the surface, the trajectories estimate transport to habitat. We assumed that appropriate shallow water juvenile habitat exists inshore of the 50-m isobath. Results from 1999 indicated that 10% of the trajectories represent transport to habitat, whereas 76% represent trans- port across the offshore boundary. For 2002, 24% represent transport to habitat, and 69% represent trans- port across the offshore boundary. Remaining trajectories (14% and 7% for 1999 and 2002, respectively! exited the coverage area either northward or southward along isobaths. Deploy- ments of actual drifters (with 1-m drogues) from a previous multiyear study provided measurements origi- nating near Platform Irene from May through August. All but a few of the drifters moved offshore, as was also shown with the HF radar-derived tra- jectories. These results indicate that most juvenile bocaccio settling on the platform would otherwise have been transported offshore and perished in the absence of a platform. However, these results do not account for the swimming behavior of juvenile bocac- cio, about which little is known. Do oil and gas platforms off California reduce recruitment of bocaccio (Sebastes paucispinis) to natural habitat? An analysis based on trajectories derived from high-frequency radar Brian M. Emery' Libe Washburn'^ Milton S. Love (contact author)^ Mary M. NIshimoto' J. Carter Ohlmann" ' Institute for Computational Earth System Science University of California Santa Barbara, California 93106-3060 2 Department of Geography University of California Santa Barbara, California 93106-4060 ^ Marine Science Institute University of California Santa Barbara, California 93106-4060 E-mail address (for M S Love, contact author) lovecailifesci ucsbedu " Institute for Computational Earth System Science University of California Santa Barbara, California 93106-3060 and Scripps Institution of Oceanography La Jolla, California 92093-0213 Manuscript submitted 20 July 2004 to the Scientific Editor's Office. Manuscript approved for publication 6 October 200,5 by the Scientific Editor. Fi.sh. Bull. 104:391-400 (2006). The 27 oil and gas platforms off south- ern and central California have limited life spans. Many of these structures have been in place for over 20 years (Love et al., 20031, and it is expected that some of these platforms will be decommissioned in the near future. Because decommissioning may entail full removal of the platform, agency personnel tasked with determining the best course of action in regard to the platforms would likely benefit from an understanding of the role that platforms play as fish habitat (Schro- eder and Love, 2004). The platforms harbor high densities of many species of fishes, although species compositions vary with plat- form bottom depth (Love et al., 1999a; 1999b; Love et al., 2003). About 35 species of rockfishes (genus Sebastes) dominate the three distinct assem- blages found around many platforms in the Santa Barbara Channel and off central California: the bottom, shell mound, and midwater assemblages. Fishes around platform bottoms tend to be adult and subadult individuals. Those on the shell mounds are usual- ly adults of dwarf species or juveniles of larger taxa. The midwater assem- blages are composed almost entirely of juvenile fishes. Some of these juve- nile fishes are one and two-year old individuals, but most are young-of- the-year (YOY) rockfishes. Densities of YOY rockfishes around platforms are usually far higher than those at nearby natural reefs (Love et al., 2003). These observations have raised a concern (e.g., Krop').) that platforms ' Krop, L. 1997. Environmental user group representative, disposition panel. In Proceedings: Public work- shop, decommissioning and removal of oil and gas facilities offshore Cali- fornia: recent experiences and future deepwater challenges, September 1997 (F. Manago, and B. Williamson, eds.), p 172. Mineral Management Service OCS Study 98-0023. Coastal Research Center, Marine Science Institute, Univ. California, Santa Barbara, California, 93106. MMS Cooperative Agreement Number 14-35-0001-30761. 392 Fishery Bulletin 104(3) may reduce recruitment to natural reefs by functioning as catchments for pelagic juvenile rockfishes. To investigate the possibility that an oil platform may reduce recruitment of rockfishes to natural habitat, we simulated drift pathways (hereafter referred to as "trajectories") from an existing platform to nearshore habitat using current measurements from high-fre- quency (HF) radars. Materials and methods Species modeled Because trajectories derived from HF radar approximate transport pathways of near surface water parcels, we chose to model the movements of pelagic juvenile bocac- cio iSebastes paucispinis) that dwell near the surface during their time in the plankton. This historically important recreational and commercial fishing target in central and southern California (Love et al., 2002), is also among the shallowest dwelling juvenile fishes (Lenarz et al., 1991; Ross and Larson, 2003). Off central and northern California, parturition for bocaccio occurs from January to May and peaks in February (Love et al., 2002). Off southern California, the species has a reproductive season that spans all year, but most larvae are released from October to July, although January is the peak month. Juvenile bocaccio recruit to inshore waters from February to August off central California, although May through July is the peak season (Love et al.. 2002). The trajectory simula- tion period from May through August was chosen to span this principal recruitment season. Bocaccio range from the Alaska Peninsula to central Baja California, and adults are usually found over high relief boulder fields and rocks in 50-250 m of water (Love et al., 2002). The fish most often settle in rocky habitat covered with various algae or in sandy zones with eelgrass. Juvenile bocaccio are commonly found in drifting kelp (Mitchell and Hunter, 1970; Boehlert, 1977) indicating that the fish recruit to natural habitat encountered in offshore surface waters. For this analy- sis, we assumed that waters from the shallow subtidal to the 50-m isobath represented suitable habitat for ju- venile recruits. This choice reflects the lack of informa- tion about suitable habitat locations in our study area and likely results in overestimates of the abundance of such habitat. Annual scuba surveys and submersible surveys (1995-2001) in the Santa Barbara Channel and Santa Maria Basin regions showed that YOY bocaccio inhabit the upper 35 m around one or more platforms for each year surveyed. Platform Irene {34°36.62'N, 120°43.40'W; bottom depth 73 m) was selected for analysis because fish recruited to it each year from 1995-2001 (Love et al., 2001) and it was the site of the highest density of YOY bocaccio observed from submersible surveys dur- ing these years (Love et al., 2003). Moreover, from May through August, 1999 and 2002, Platform Irene was also in a region of good HF radar coverage, which al- lowed computation of extensive trajectory ensembles. Ocean current measurement and trajectory calculation Near-surface ocean currents were measured hourly by using an array of three HF radars (SeaSondes, manu- factured by CODAR Ocean Sensors, Ltd. of Los Altos, CA) operating at 12-13 MHz. At these frequencies, the measurement is an average over the upper 1 m of the water column (Stewart and Joy, 1974). The radars were located at Pt. Sal, Pt. Arguello, and Pt. Conception (Fig. lA). HF radars measure components of surface currents by means of a Doppler technique, at spatial increments of 1.5 km in range and 5° in azimuth from the radar location. Surface current vectors in an east-north coordi- nate system were computed on a 2-km grid by using the least square technique described by Gurgel (1994). With this technique, all current components obtained within a 3-km radius around each grid point were combined to estimate the surface current every hour. The 3-km radius limits the spatial resolution of the near-surface current fields. Emery et al. (2004) have described the processing of the HF radar data in more detail. Fur- ther discussion on the use of HF radars for measuring near surface currents is given by Paduan and Rosenfeld (1996) and Graber et al. (1997). Emery et al. (2004) assessed performance of the three HF radars by comparing them with in situ current me- ters at 5 m depth. They found that root-mean-square differences in radial speed measurements between HF radars and current meters ranged from 0.07 to 0.19 m/ s. Recent observations comparing surface currents from HF radars and drifters have indicated that differences are substantially reduced if spatial variability in cur- rent fields is accounted for (Ohlmann, 2005). The areas used for computing trajectories were off- shore of Pt. Conception and Pt. Arguello as shown in Fig. 2A for 1999 and Fig. 3A for 2002. These areas were selected to maximize the spatial coverage, and to minimize the inclusion of grid points with low temporal coverage. Variable coverage from individual radars re- sults in differences in coverage between years. Boundar- ies of nominal coverage areas were oriented along and perpendicular to isobaths. Platform Irene is about 2 km from the inshore boundaries, which lie along the 50-m isobath (Figs. 2A and 3A). At times actual radar cover- age exceeded nominal coverage boundaries as may be seen by comparing sample computed trajectories (black lines, Fig. lA) with the 2002 boundary (gray closed curve. Fig. lA). Coverage in 2000 and 2001 for May through August was inadequate for producing trajectory ensembles around Platform Irene. A new trajectory was begun at the location of Plat- form Irene every 4 hours from 1 May through 31 August for 1999 and 2002. Positions along the trajectories were determined by integrating current vectors forward in time using a fourth order Runge-Kutta algorithm. Tra- jectories ended where they encountered spatial gaps or where they reached the edge of radar coverage. Emery el al : Do oil and gas platforms reduce recruitment of Sebastes paucispinis to natural habitat' 393 35.0N- 34.8- 34.6- 34,4- 34.2- 34.0 35. ON 34.8- 34.6- 34.4- 34.2- 340 -120.8 120.6 -120.4W 120.6 -120.4W Figure 1 (A) Map of study area near Pt. Conception, California, showing trajectories derived from high-frequency (HF) radar from 1 May-31 Aug 2002. Triangles show HF radar locations and the white square shows Plat- form Irene. Gray curve superimposed on trajectories is the coverage boundary used for 2002. Labeled thin black lines are bathymetric contours. (B) Heavier black lines are 25 sample trajectories that intersect the coverage boundary. Gray curve is the same as in panel A. The trajectories in panels A and B were created from velocity time series that were interpolated with empirical orthogonal functions (EOFs). The number of trajectories reaching the coverage boundaries defined in Figures 2A and 3A were reduced by gaps in spatial and temporal radar coverage. For example, of the 670 possible trajectories in 2002, 541 (81%) ended within the radar coverage area and 129 (19%) intersected the coverage boundary. Changes in spatial coverage on diurnal and longer time scales re- sulted from several factors, such as broadcast interfer- ence, and are a characteristic of HF radars (Paduan and Rosenfeld, 1996). Gaps in the velocity time series were also caused by outages of individual radars. The average durations of these gaps were 4.4 ±22.3 h and 5.9 ±7.9 h in 1999 and 2002, respectively. Outages of individual radars also produced a few long gaps in the velocity time series for each year across the entire cov- erage area. In 1999 two long gaps occurred: one from 1800 coordinated universal time (UTC) 28 June through 2000 UTC 22 July, and a second from 2300 UTC 24 July through 2200 UTC 13 August. In 2002 a single long gap occurred from 1700 UTC 16 May through 0100 UTC 21 May. These longer gaps were not filled. Shorter gaps were filled by interpolation by using em- pirical orthogonal functions (EOFs; (Emery and Thom- son, 1998)). EOFs incorporate the underlying spatial structure of all velocities recorded at all locations where data existed at a given time. Any velocity component, u say, at grid point j may be expressed as N where t = time; £7 = the time average at locationj (computed from available data at locationj); a, = the time-varying amplitude function; q>,j - the ith spatial EOF mode at locationj; and A^ = the number of modes. The first seven modes (i.e., N=7) were used for inter- polation and explained 64% (1999) and 56% (2002) of the variance. EOF interpolation increases the number of trajectories reaching the coverage boundary to 99%. As a test, gaps were also filled with linear, spline, and moving average interpolation, but EOF interpolation resulted in the most trajectories reaching the coverage boundaries. Otherwise, results did not depend strongly on the interpolation method. The fraction of filled data with EOF interpolation compared with the total possible data was 4% in 1999, 394 Fishery Bulletin 104(3) 34.8N- 34.6- 34.4- 34,2- 34.0- t>u- 45- B ^ 40- - 35^ - 30- - 25- : 20- 15- - 10- 1 - s- — _^-r-^ 1 ' 5 ' ' ' 10 15 ' Bin number 1 1 1 1 M M 1 1 1 1 M D 10 1 1 1 1 1 1 20 30 Distance (km) 40 5C 6C 100 90 80 O 70 60 50 s 0) -40 m =5 O -30 ^ O ■20 10 0 -120.4W Figure 2 (A) Coverage boundary (indicated by heavy lines around the black square [Platform Irene]) for 1 May-31 August 1999, Numbers and tick marks around boundary identify bins corresponding to x-axis in panel B. (B) Histogram (gray bars, left-hand scale) and cumulative frequency (bold line, right-hand scale) show fraction of trajectories (in percent) intersect- ing bins around coverage boundary. Panel A shows bin locations. Distance around the coverage boundary (Fig. 2A) in kilometers is also shown (bottom scale). Arrow above histogram shows boundary location nearest Platform Irene. 12% in 2002, and 14% with the 2002 data for the 1999 coverage boundary. Here the total possible data were the number of grid points within the coverage bound- ary for either 1999 (45 grid points) or 2002 (291 grid points) multiplied by the number of hours between 1 May and 31 August minus the long gaps discussed above (2952 hours-1057 hours in 1999, 2952 hours-104 hours in 2002). Examples of 25 EOF-filled trajectories that started every 120 hours and intersected the 2002 coverage boundary are shown in Figure IB. The principal quantity used in our study to estimate how Platform Irene might affect transport to nearshore habitat was the histogram of points where trajectories crossed the boundaries of the coverage areas. To deter- mine this quantity, coverage boundaries were divided into 4-km-long segments, or bins. The first bin of each histogram was less than 4 km because distances around the coverage boundaries were not exactly divisible by 4. Bin numbers increased counter-clockwise around the boundaries starting from 1 in the southeastern corner (Figs. 2A and 3A). The smallest numbers identified bins lying along the 50-m isobath. Results Trajectories originating at Platform Irene were suffi- ciently dense to fill in much of the surrounding area. In 2002 for example, EOF-filled trajectories spread over an area of about 20 km in the cross-shore direction by 60 km in the alongshore direction (Fig. lA). North of Pt. Arguello, several trajectories crossed the 50-m isobath and some ended very near shore. South of Pt. Arguello, only a few trajectories approached the 50-m isobath. Instead, most turned southward or southwestward and moved offshore. A tendency for trajectories to align par- allel to isobaths was evident in the northern end of the ensemble, although in other areas, such as the southeast, many trajectories lay across isobaths. A histogram of points where trajectories crossed the coverage boundary for May-August 1999 exhibited a peak in bin 11 on the offshore side along the 500-m isobath (Fig. 2B, left-hand axis). Table 1 and the cu- mulative histogram (Fig. 2B, right-hand axis) showed that 76% of the trajectories crossed the offshore side corresponding to bin numbers 9-13. A second peak oc- Emery et al : Do oil and gas platforms reduce recruitment of Sebastes paucispinis to natural fiabital' 395 34,8N- 34.6- 34.4- 34.2- 34.0- 14- 12- 10- B - r ; 8- 6- 4- f. 1 -i -l 2- - - 0- 1 1 1 1 n "- " rr-- Wffi-- — -Jta^- 10 15 20 25 30 35 Bin number 100 88 75 O c 3 c 63 Q) a 50 -38 25 t. 13 — I — 20 40 — I 1 1 60 80 100 Distance (km) 120 — I — 140 — I 160 -121 -120.8 -120.6 -120.4W Figure 3 (A) Coverage boundary (indicated by heavy lines around the black square (Platform Irene]) for 1 May-31 August 2002. Numbers and tick marks around boundary identify bins corresponding to .r-axis in panel B. (B) Histogram (gray bars, left-hand scale) and cumulative frequency (bold line, right-hand scale) show fraction of trajectories (in percent) intersect- ing bins around coverage boundary. Panel A shows bin locations. Distance around the coverage boundary (Fig. 2A) in kilometers is also shown (bottom scale). Arrow above histogram shows boundary location nearest Platform Irene. Table 1 Statistics for trajectories using empirical orthogonal function (EOF) interpolation, and Scripps Institution of Oceanography | drifters. Trajectories Trajectories Trajectories Trajectories Mean Maximum Trajectories intersecting intersecting intersecting intersecting residence residence intersecting inshore offshore northern southern time in time in Coverage Total coverage boundary (Vf) boundary boundary boundary coverage area coverage area Year boundary trajectories boundary (50-m isobath) m (Ic) ['k) (h±lSD) (h) 1999 1999 590 546 10 76 9 5 19 (±12) 86 2002 2002 716 706 24 69 6 1 47 (±34) 163 2002 1999 716 712 18 66 7 8 22 (±18) 116 SIO drifters 2002 21 17 18 71 12 0 38(±21) 79 curred in bin 3 and about 10% of trajectories crossed the inshore boundary on the 50-m isobath (bins l-5l. The remaining trajectories crossed either the northern (9%) or southern (5%) sides of the coverage boundary. In 2002 the radars covered a substantially larger area (Fig. 3A), including about 50 km of the 50-m isobath, 70 km of the 500-m isobath along the offshore bound- ary, and a portion of the western entrance of the Santa Barbara Channel. The histogram of coverage bound- ary crossings for May-August 2002 also exhibited two peaks, one between bins 9-11 along the 50-m isobath and a second in bin 31 along the offshore side of the cov- erage boundary (Fig 3B). In 2002, 24% of trajectories crossed the 50-m isobath (bins 1-13), 69% crossed the 396 Fishery Bulletin 104(3) Bin numbers 10 20 30 Distance (km) 40 50 40 S 60 Figure 4 Histogram (gray bars, left-hand scale) and cumulative frequency (bold line, right-hand scale) showing percentage of 2002 tra- jectories intersecting bins from the 1999 coverage boundary. Bottom scale shows distance around the coverage boundary in kilometers. Arrow above histogram shows boundary location nearest Platform Irene. offshore side (bins 22-39), and the remainder crossed the northern side (6%) or the southern side (1%). To compare results more directly between years, a histogram of crossings was generated from the 2002 trajectories, using the 1999 coverage boundary. The 1999 coverage boundary was completely contained within the 2002 coverage boundary. With the 2002 trajectories, a peak in the histogram again occurred along the offshore boundary (Fig.4, left-hand scale), this time at bin 12 compared with bin 11 when the 1999 trajectories were used. A second, but much smaller peak occurred along the 50-m isobath at bin 3, consistent with the small peak along the 50-m iso- bath of Fig. 2B. Table 1 and the cumulative histo- gram (Fig. 4, right-hand scale) showed that 18% of trajectories crossed the 50-m isobath, 66% crossed the offshore side of the coverage boundary, and the remainder crossed either the northern (7%) or south- ern (8%) sides. The time required for trajectories to cross the cover- age boundary, defined in our study as the residence time, varied between years and mainly depended on the size of the coverage area. In 1999 the mean and standard deviation for the residence time was 19 ±12 hours, and the maximum was 86 hours (Table 1). In 2002 they were 47 ±34 hours and the maximum was 163 hours. When the 2002 trajectories were computed over the 1999 coverage boundary, residence times were comparable to the 1999 values: 22 ±18 hours and a maximum of 116 hours. Discussion Because of limitations in spatial coverage, the HF radar- derived trajectories could not be used to examine the full range of length and time scales over which actual trajectories may extend. We used a trajectory data set resulting from the release of Argos drifters in the region to examine these scales. Drifters were deployed in the Santa Barbara Channel and Santa Maria Basin at irregular intervals from October 1992 through Decem- ber 1999 as part of a circulation study conducted by the Scripps Institution of Oceanography (SIO; see Dever [1998] and Winant et al. [2003] for a description of the drifter data set). Drifter positions were obtained up to six times per day, typically for 40 days, and had a spa- tial accuracy of about 1 km. Several trajectories ended earlier when the drifters beached. No drifters were released at Platform Irene although many approached the platform after release elsewhere (130 drifters were deployed to the north of Platform Irene, and 440 were deployed to the south). To approx- imate trajectories originating at Platform Irene, all drifters released during all seasons for all years, and approaching within 10 km of the platform, were identi- fied. This distance is a compromise between proximity to the platform and ensemble size; 93 trajectories ap- proached within 10 km of Platform Irene (white circle in Fig. 5 is 10 km in radius and is centered on Platform Irene). Of these, 34 were released north and 59 south of Platform Irene. The ensemble of trajectories beginning Emery el al : Do oil and gas platforms reduce recruitment of Sebastes paucispinis to natural fiabitat' 397 within 10 km of Platform Irene (gray and black dots, Fig. 5) mainly followed the trend of southward advection by the California Current System, although a smaller number extended northward from the platform and a few reached Monterey Bay. Four trajectories entered the Santa Barbara Channel. A further sorting of the ensemble of 93 trajectories approaching within 10 km of Platform Irene to include only those during 1 May-31 August of all years pro- duced a subset of 21 trajectories (black dots. Fig. 5). Of these, 17 crossed the 2002 coverage boundary: 3 on the 50-m isobath, 12 on the offshore boundary, 2 on the northern boundary, and 0 on the southern bound- ary. Although the ensemble was small, the fraction of drifters crossing the inshore boundary of the 2002 coverage area (18%) was comparable to the fraction of HF radar-derived trajectories that did so (24%), as shown in Table 1. Most trajectories crossing the offshore boundary continued offshore and southward, consistent with advection by the California Current System. Oth- ers crossing the offshore boundary extended north of Platform Irene before turning southward or offshore. The two trajectories crossing the northern boundary re- mained near shore and crossed the 50-m isobath north of the platform. None of these 17 trajectories entered the Santa Barbara Channel. Except for three drifters that beached, all the drifters remained offshore for the duration of Argos data logging. Trajectories crossing the 50-m isobath tended to do so north of platform Irene, as shown by most of the computed trajectories and three of the SIO drifters (Figs. 2B and 3B). This movement indicates that trans- port from the platform to shallow water habitat along the mainland coast mostly occurred during times of northward, or poleward, currents. Poleward flow in the region north of Pt. Conception results from weaken- ing or reversal of the prevailing upwelling favorable winds, the so-called "relaxation" flow state described by Dever (2004), Harms and Winant (1998), and Wi- nant et al. (2003). They also described two other flow states, "upwelling" and "convergent," which produce off- shore and equatorward transport near Platform Irene. Together these flow states have a 69% probability of occurring during May-Aug (36% for upwelling and 33%> for convergent), whereas the relaxation state has a 23% probability of occurring (Winant et al. [2003J, their Table 3). For comparison, 19-30%> of trajectories (HF radar-derived plus actual drifters) crossed the inshore and northern boundaries, consistent with the relaxation probability and 70-81% crossed the offshore and southern boundaries, consistent with the upwelling plus convergent probability. The trajectories can also be used to estimate recruit survival, based on the time required for transport to habitat. Recruit survival is estimated from a simple exponential decay model: 38N- 36- 34- 32- 30- 28- iV r-'^ -124 -122 -120 ~~l -new Figure 5 Trajectories of all drifters deployed by the Scripps Insti- tution of Oceanography at various times from October 1992 through December 1999 after passing within 10 km of Platform Irene I gray dots). Also shown are all drifters after passing within 10 km of Platform Irene for only the time period 1 May-Si August for all years (black dots). White circle is centered on Platform Irene and has a radius of 10 km. P = population at time t\ and Pq = the initial population. Here, Pq represents the population of juvenile bocaccio that recruited at Platform Irene, and survival estimates are used to predict their survival in the absence of the platform. In 1999 and 2002, trajectories to habitat from Platform Irene crossed the 50-m isobath within 19-47 hours (Table 1), indicating a high percentage of sur- vival for bocaccio (96-98%). In contrast, offshore and southward drifter trajectories from the SIO drifter data indicated much lower survival. Pelagic juvenile bocaccio transported by these flows would be carried southward by the California Current and remain far from the mainland and the Northern Channel Islands (Fig. 5) P(t) = P^e-'", (3) where m = mortality (0.02 or 0.06/day, Ralston^); ■■^Ralston, S. 2004. Personal conimun. NOAA National Marine Fisheries Service, 110 Shaffer Road, Santa Cruz, CA 95060. 398 Fishery Bulletin 104(3) for at least 40 days, the nominal time the drifters were tracked. Survival after 40 days along these trajectories ranged from 9 to 45%. It is possible that juvenile bocaccio spend time away from near surface waters during their planktonic larval phase and therefore their trajectories may depend on deeper currents. Previous observations in the study re- gion show strong correlation between near surface and deeper flows, indicating that inferences from surface trajectories apply to deeper trajectories. Comparing moored current meters at 5 and 45 m depth, Dever (2004), Harms and Winant (1998), and Winant et al. (2003) generally found a high correlation between cur- rent directions, and a decreased flow speed at depth. They also found that poleward flow at 45 m occurs during the convergent and relaxation states, and equa- torward flow occurs during the upwelling state. There- fore, during relaxation and upwelling states, currents at 45 m likely have similar directions, but lower speeds compared with surface currents. During the convergent state, currents at 45 m are opposite surface currents, indicating that surface currents represent deeper cur- rents only to some shallower depth. We assumed in this trajectory analysis that juve- nile bocaccio effectively behave like passive particles. We made this assumption for two reasons: 1) this as- sumption allowed us to focus on the lower bound of their range of possible swimming behaviors; and 2) such an assumption also eliminated the need to ac- count for their actual swimming behavior in the open ocean, about which little is known. We do not assume that juvenile bocaccio do not swim; rather, we assume that they swim randomly such that their effective transport is similar to that of passive particles. The other behavioral limit of rapid, consistently directional swimming behavior would likely alter the fraction of bocaccio encountering shallow habitats versus the frac- tion being advected offshore. Flume experiments with visual cues for directional orientation demonstrate that coral reef fish in the late pelagic stage can swim up to -100 km in 8 days (Stobutzki and Bellwood, 1997); therefore behavioral modification of trajectories could be very important. Larval and pelagic juvenile fish may possess swimming and sensory abilities to overcome passive drift in currents; however, some kind of external reference is necessary for fish to detect and respond to the direction of a current. In a review of the behavior of larval and juvenile fish in the pelagic environment, Leis and McCormick (2002) pointed out that it is yet to be demonstrated that these early-stage fish in offshore "blue water" can effectively modify current-driven trajectories by orienting to cues from settlement habitat located at a scale greater than sev- eral kilometers away. A variety of near-field stimuli, such as light and temperature gradients, sound, and visible prey affect swimming behavior. Clearly more research is needed to evaluate the effects of swimming behavior of temperate reef fishes, such as bocaccio, in order to model their dispersal. We speculate, however, that the assumption of passive dispersal will remain an important lower bound on constraining effects of swimming behavior. Smoothing and interpolation in the processing of the HF radar velocity data limit the spatial resolution of current fluctuations to scales of ~6 km, the diameter of circles used to compute velocity vectors. Velocity struc- tures smaller than this scale are not resolved but may be important in determining trajectories. For example, Helbig and Pepin (2002) found that errors in modeling the spatial distributions of fish eggs in an embayment increased as spatial resolution of a circulation mod- el decreased. Assuming effects of unresolved velocity structures on smaller scales act as a diffusive process, we speculate that incorporating diffusion would cause the locations of the boundary intersections to spread to adjacent bins. In this case, peaks in the histograms (such as in Figs. 2B, 3B, and 4) would decrease as diffusion spreads boundary intersections to adjacent bins. Velocity statistics at scales of the order of a few km and smaller in our study area, however, are not available for incorporating the effects of diffusion into the trajectories. Results from actual drifters, which do contain velocity structure unresolved by the HF radars, are not very different from HF-radar-derived trajectory results (Table 1), indicating that the effects of unresolved variance are not large. For predicting settlement to habitat, trajectory improvements gained through the incorporation of smaller scale flow features may be offset by assumptions of swimming behavior and habitat location. Questions and issues have arisen in the decommis- sioning process about the regional importance of plat- form fish assemblages (Schroeder and Love, 2004). For example, does removal of a platform impact the bocaccio population? Based on our annual research submersible surveys (detailed in Love et al., 2003) conducted in 1997, 1998, 1999, and 2001, estimates of YOY bocac- cio at Platform Irene ranged from 61 (2001) to 41,000 (1999) (Lenarz-^). YOY bocaccio abundances can be even higher than those observed at Platform Irene. We have recently estimated that, during 2003, Platform Grace, located in the Santa Barbara Channel, harbored over 300,000 YOY bocaccio (Lenarz'M. Under even the most conservative parameters, this would translate into many thousands of adults (MacCalH). In addi- tion, there is evidence that some of the bocaccio that recruit to platforms as YOYs migrate, and thus seed, natural reefs. Fish tagged at Platform A, located off Summerland, CA, in the Santa Barbara Channel, were later recovered over natural reefs over 100 km to the north and south of that platform (Hartmann, 1987). In another study, recruiting bocaccio became resident on a deep-water platform and formed the highest density of adult fish observed in the Southern California Bight ■^ Lenarz, W. 2004. Personal commun. P.O. Box 251, Kent- field, CA 94914-0251. -■ MacCall, A. 2004. Personal commun. NOAA National Marine Fisheries Service, 110 Shaffer Road, Santa Cruz, CA 95060. Emery et a\ : Do oil and gas platforms reduce recruitment of Sebastes paucispinis to natural fiabitat' 399 (Love et al., 2003). Thus, bocaccio that recruit as YOYs to a platform may benefit natural reefs either through emigration to these reefs or through increased larval production. 191). Deploying the radars was made possible through cooperative agreements with the California State Parks Channel Coast District, the U.S. Coast Guard, and the U.S. Air Force. Conclusion Observations of evolving surface current patterns obtained by HF radar are used to estimate dispersal pathways for juvenile bocaccio in the vicinity of Platform Irene, an oil production platform off the central Califor- nia coast. Results indicate that most of YOY bocaccio settling around Platform Irene would not survive in the absence of the platform. Instead, prevailing currents would likely advect them offshore where they would have a very low probability of survival. Although it is possible that some individuals would encounter acceptable nurs- ery habitat on offshore banks or islands, it is likely that most would perish. Thus, the presence of Platform Irene almost certainly increases the survival of young bocaccio in the Point Conception-Point Arguello region. These results indicate that knowledge of regional ocean circulation patterns is essential for evaluating the effects of oil production platforms, or other artifi- cial habitats, on dispersal pathways of juvenile fishes. Platform location, local current patterns, and natural habitat distribution determine the balance between settlement at a specific platform and settlement on natural habitat. The approach used in this study, an analysis of trajectories derived from HF radar current measurements, can provide insights into this balance. Additional research on small-scale circulations features unresolved by the radars and on swimming behavior of juvenile rockfishes will sharpen these insights. Acknowledgments We thank S. Ralston, NMFS, Santa Cruz, for providing pelagic juvenile bocaccio mortality estimates; W. Lenarz for YOY bocaccio abundance estimates at Platforms Irene and Grace; and A. MacCall, NMFS, Santa Cruz, for projected survival of the YOY bocaccio. We benefited from helpful discussions with Edwin Beckenbach and David Siegel. Drifter data were made available by the Center for Coastal Studies at the Scripps Institution of Oceanography. David Salazar and Chris Gotschalk provided valuable support during field operations. This research was supported by the Minerals Management Service, U.S. Department of the Interior, MMS Agree- ment 14-35-0001-30758; the California Artificial Reef Enhancement Program; the W.M. Keck Foundation; the Biological Resources Division of the U.S. Geological Survey; the University of California Marine Council; the National Oceanographic and Atmospheric Administra- tion; and PISCO, the Partnership for Interdisciplinary Studies of Coastal Oceans, funded primarily by the Gordon and Betty Moore Foundation, and the David and Lucile Packard Foundation (contribution number Literature cited Boehlert G. W. 1977. Timing of the surface-to-bentliic migration in juvenile rockfish, Sebastes diploproa, off southern California. Fish. Bull. 75:887-890. Dever E. P. 2004. Objective maps of near-surface flow states near Pt. Conception, California. J. Phys. Oceanogr. 34:444-461. Dever E. P., M. C. Henderschott, and C. D. Winant. 1998. Statistical aspects of surface drifter observations of circulation in the Santa Barbara Channel. J. Geo- phys. Res. 103:24,781-724,797. Emery B. M., L. Washburn, and J. A. Harlan. 2004. Evaluating radial current measurements from CODAR high-frequency radars with moored current meters. J. Atmos. Oceanic Technol. 21:1259-1271. Emery W. J., and R. E. Thomson. 1998. Data analysis methods in physical oceanography, 634 p. Pergamon. Elsevier Science, Inc.. New York, NY. Graber H. C, B. K. Haus, R. D. Chapman, and L. K. Shay 1997. HF radar comparison with moored estimates of current speed and direction: expected differences and implications. J. Geophys. Res. 102:18,749-18,766. Gurgel K. W. 1994. Shipborne measurement of surface current fields by HF radar L'Onde Electrique 74:54-59. Harms S., and C. D. Winant. 1998. Characteristic patterns of the circulation in the Santa Barbara Channel. J. Geophys. Res. 103: 3041-3065. Hartmann A. R. 1987. Movement of scorpionfishes (Scorpaenidae: Se- bastes and Scorpaena) in the Southern California Bight. Calif Fish Game. 73:68-79. Helbig J. A., and P. Pepin. 2002. The effects of short space and time scale cur- rent variability on the predictability of passive ich- thyoplankton distributions: an analysis based on HF radar observations. Fish. Oceanogr. 11:175-188. Leis J. M., and M. I. McCormick. 2002. The biology, behavior and ecology of the pelagic, larval stage of coral reef fishes. In Coral reef fishes: dynamics and diversity in a complex ecosystem (P. F. Sale, ed.), p. 171-199. Academic Press, San Diego, CA. Lenarz W. H., R. J. Larson, and S. Ralston. 1991. Depth distributions of late larvae and pelagic juve- niles of some fishes of the California Current. Calif. Coop. Ocean. Fish. Invest. Rep. 32:41-46. Love M. S., J. Caselle, and L. Snook. 1999a. Fish assemblages around seven oil platforms in the Santa Barbara Channel area. Fish. Bull. 98: 96-117. 1999b. Fish assemblages on mussel mounds surrounding seven oil platforms in the Santa Barbara Channel and Santa Maria Basin. Bull. Mar. Sci. 65:497-513. 400 Fishery Bulletin 104(3) Love M. S., M. M. Nishimoto. and D. M. Schroeder. 2001. The ecological role of natural reefs and oil and gas production platforms on rocky reef fishes in south- ern California: 1998-1999 survey report, 115 p. U. S. Department of the Interior, U.S. Geological Survey, Biological Resources Division, Seattle, Washington, 98104. Love M. S., D. M. Schroeder, and M. M. Nishimoto. 2003. The ecological role of oil and gas production plat- forms and natural outcrops on fishes in southern and central California: a synthesis of information. U.S. Department of the Interior, U.S. Geological Survey, Biological Resources, Seattle, WA. Love M. S., M. Yoklavich, and L. Thorsteinson. 2002. The rockfishes of the northeast Pacific, 405 p. LTniv. California Press, Berkeley, CA. Mitchell C. T., and J. R. Hunter. 1970. Fishes associated with drifting kelp, Macrocystis pyrifera, off the coast of southern California and northern Baja California. Calif Fish Game 56:288-297. Ohlmann J. C, P. F. White, A. L. Sybrandy, and P. P. Niiler. 2005. GPS-Cellular drifter technology for coastal ocean observing systems. J. Atmos. Oceanic Technol. 22:1381-1388. Paduan J. D., and L. K. Rosenfeld. 1996. Remotely sensed surface currents in Monterey Bay from shore-based HF radar (coastal ocean dynamics appli- cation radar). J. Geophys. Res. 101:20,669-620, 686. Ross J. R., and R. J. Larson. 2003. Influence of water column stratification on the depth distributions of pelagic juvenile rockfishes off central California. Calif. Coop. Ocean. Fish. Invest. Rep. 44:65-75. Schroeder D. M., and M. S. Love. 2004. Ecological and political issues surrounding decom- missioning of offshore oil facilities in the Southern Cali- fornia Bight. Ocean Coast. Manag. 47:21-48 Stewart R. H., and J. W. Joy. 1974. HF radio measurements of surface currents. Deep- Sea Res. 21:1039-1049. Stobutzki I. C, and D. R. Bellwood. 1997. Sustained swimming abilities of the late pelagic stages of coral reef fishes. Mar. Ecol. Prog. Ser. 149: 31-41. Winant C. D., E. P. Dever, and M. C. Henderschott. 2003. Characteristic patterns of shelf circulation at the boundary between central and southern California. J. Geophys. Res. 108:3021. 401 Abstract — Behavior of young (8-18 mm SL) giant trevally (Caranx igno- bills), a large coral-reef-associated predator, was observed in the labo- ratory and the ocean. Size was a better predictor of swimming speed and endurance than was age. Critical speed increased with size from 12 to 40 cm/s at 2.7 cm/s for each mm increase in size. Mean scaled critical speed was 19 body lengths/s and was not size related. Swimming speed in the ocean was 4 to 20 cm/s (about half of criti- cal speed) and varied among areas, but within each area, it increased at 2 cm/s for each mm increase in size. Swimming endurance in the labora- tory increased from 5 to 40 km at 5 km for each mm increase in size. Vertical distribution changed onto- genetically: larvae swam shallower, but more variably, and then deeper with growth. Two-thirds of indi- viduals swam directionally with no ontogenetic increase in orientation precision. Larvae swam offshore off open coasts, but not in a bay. In situ observations of C. ignobilis feeding, interacting with pelagic animals, and reacting to reefs are reported. Behavioral ontogeny in larvae and early juveniles of the giant trevally {Caranx ignobilis) (Pisces: Carangidae) Jeffrey M. Leis' Amanda C. Hay' Domine L. Clark' l-Shiung Chen^ Kwang-Tsao Shao^ ' Ichthyology, Australian Museum 6 College Street Sydney, New South Wales 2010, Australia E-mail address (for J. M. Leis); |effi@austmus.gov au 2 National Museum of Marine Biology & Aquarium 2 Houwan Road Checheng, Pingtung, 944, Taiwan ^ Institute of Zoology, Academia Sinica Academia Road Nankang, Taipei, 115, Taiwan Manuscript submitted 16 December 2004 to the Scientific Editor's Office. Manuscript approved for publication 27 September 2005 by the Scientific Editor. Fi.sh. Bull. 104:401-414 (20061. The giant trevally (known as ulua aukea in Hawaii) (Caranx ignobilis (ForsskaD), is the largest species in the commercially important family Carangidae. It has a wide tropical Indo-Pacific distribution ranging from Japan to Australia and from East Africa to the Hawaiian and Mar- quesas Islands (Randall et al., 1997; Smith-Vaniz, 1999; Hoese et al., in press). Adults of this reef-associated pelagic species may be solitary or form schools On inner reefs, inshore on sand flats, or in reef channels, but primar- ily on seaward reefs to depths of 80 m (Kuiter, 1993; Myers, 1999; Hoese et al., in press). Caranx ignobilis is an important top-level carnivore in reef systems: adults and juveniles feed during the day primarily on demersal and pelagic fishes and to a lesser extent on cephalopods and crustaceans (Blaber et al., 1983; Sude- kum et al., 1991; Myers, 1999). Adult coloration varies from silver grey to almost black above and fins are usu- ally dark grey to black, but in small juveniles the body is barred (pale to dark gray) and the anal and caudal fins are yellow (Kuiter, 1993; Ran- dall et al., 1997). Caranx ignobilis is important to commercial and recre- ational fisheries and to aquaculture because of its size (maximum of 1.7 m and 62 kg), abundance, palatability, and reputation as an excellent game fish (Smith-Vaniz, 1999; Yu, 2002). The species can, however, be cigua- toxic (Myers, 1999), Little is known of the early life history of C. ignobilis. Adults, which reach maturity at about 60 cm in length and 3-4 years of age, have been reported to gather on shallow seaward reefs and offshore banks to spawn (von Westernhagen, 1974; Jo- hannes, 1981; Sudekum et al., 1991; Myers, 1999). Neither eggs nor larvae have been described, and nothing is known of the distribution or behavior of the larvae or early juveniles. Juve- niles are found in small schools over sandy inshore bottoms (Myers, 1999), but also in turbid estuaries, where they are "uncommon" (Blaber et al., 1983). Juveniles and subadults have wide tolerances for salinity and tur- bidity (Blaber et al., 1983). In Kaneo- he Bay, Hawaii (a coral-reef lagoon with estuarine features), C ignobilis juveniles less than 20 cm (<1 year old) were found on the "murky" lagoon floor (Wetherbee et al., 2004). Juve- niles were associated with lagoonal patch reefs when they were between 25 and 40 cm in length, and appar- 402 Fishery Bulletin 104(3) ently moved out of the bay when larger than 40 cm. In South African estuaries, individuals larger than 55 cm were absent (Blaber et al., 1983). The high abundance of C. ignobilis in the NW islands of Hawaii (Sudekum et al., 1991), where estuaries are absent, shows that presence of juveniles in estuaries is facultative. Recently, Taiwanese aquaculturists have developed culture methods for C. ignobilis (Yu. 2002), and this enabled us to obtain larvae and small juveniles that we used for laboratory and field observations of behavior. Behavior of early-life history stages of marine fishes can strongly influence both the survival and distribution of these small fishes (Leis and McCormick, 2002). There- fore, an understanding of the behavioral capabilities of these early-life history stages is essential to intelligent management, and to comprehend the demographics and ecology of any marine teleost fish. Aside from field studies of the behavior of wild larvae in vertical distribution (e.g., Ahlstrom, 1959; Olivar and Sabates, 1997; Flores-Coto, et al., 2001), behavior has been studied in the early-life history stages of only a few carangid fishes. There has been extensive research on cannibalism, aggressive interactions, and school- ing behaviour by using reared larvae and juveniles of Japanese amberjack (Seriola quinqueradiata) and white trevally (Pseudocaranx dentex) in the labora- tory (e.g., Masuda and Tsukamoto 1996, 1998, 1999; Sakakura and Tsukamoto 1996, 1999). We report in the present study the first observations of the develop- ment of swimming abilities, orientation abilities, and vertical distribution for C. ignobilis, using both in situ methods (Leis et al., 1996) and laboratory swimming chambers (Fisher et al., 2000). These behaviors are important to dispersal, feeding, predator interactions, and survival generally during the early life history of C. ignobilis. Materials and methods Larvae and juveniles Young C. ignobilis were obtained from a commercial aquaculture farm near Kaohsiung, Taiwan. They were identified as C. ignobilis by the farmer with reference to photographs, and the identification was subsequently confirmed by examination of preserved specimens. Caranx ignobilis eggs from an induced spawn were placed in a large outdoor earth pond (approximately 20x20xlm), and thus hatched under "natural" condi- tions. Larvae were provided with a "natural" food source (phytoplankton and zooplankton that were resident in the pond). Surface water temperature in the outdoor pond was 32°C when the larvae were collected in May 2004. Fish were obtained on two occasions at the same pond from a single cohort and from an unknown number of females; the first collection was at 20 days after hatch- ing (dab) and the second at 24 dah. The young fish were placed in oxygenated plastic bags in an insulated box and transported to the laboratory at the National Museum of Marine Biology and Aquarium (NMMBA), Renting, Taiwan, about 1 hour by road. In the laboratory the larvae were acclimated in a 40-liter aquarium filled with water from the seawater system at NMMBA. Each aquarium was fitted with an aerator, and kept ca 25°C. Twice daily, the larvae were fed with live, newly hatched brine shrimp (Artemia) nauplii and 50% of the total volume of water exchanged with fresh seawater. The aquaria were cleaned daily by suctioning debris off the bottom. Reported sizes of larvae are given in standard length (SL, in mm). Ages are reported as days after hatching (dah) and are based on the age reported by the aqua- culturist when the larvae were obtained. The nomen- clature for early life history stages of fishes is complex; there are many different systems of terminology and no consensus on which is the most appropriate. Depending on the nomenclature used, the C. ignobilis individuals we studied (Fig. 1) would be considered larvae, or ju- veniles, or as a mixture of both. In our C. ignobilis, all fin-rays were present in the smallest specimens (8 mm), and scales were present from about 14 mm, yet the preopercular spines that characterizes most carangid larvae (Leis and Carson-Ewart, 2004) were still pres- ent at 19 mm. Unlike the pelagic stages of demersal fishes, C. ignobilis does not undergo a clear ecological transition from pelagic to demersal habitat upon which one might base life history stages. We do not attempt to distinguish between larvae and juveniles, and to avoid awkward phrasing and for simplicity, we refer to the young fish that we studied as larvae on the basis that the largest individuals retained some head spines. We acknowledge that in some taxonomic systems they may be referred by other terms. Laboratory observations Multilane swimming chambers were used to measure swimming abilities (Stobutzki and Bellwood, 1994). One chamber was used to measure critical speed and a second identical chamber was used for measurements of endurance (Fisher et al., 2000). Both chambers were made of clear plexiglass and had 6 lane-ways, each 30 mm wide, 50 mm high, and 180 mm long. A black line across the lid of the chamber provided the larvae with a point of reference for orientation. Aside from the fine mesh (0.5-mm) ends, the chamber design was identical to that of Stobutzki and Bellwood (1994, 1997). Even distribution of flow was achieved by a T-piece diffuser in the header portion of the chamber. Turbu- lence in the chamber was minimized by a 40-mm-long section of flow straighteners at the start of each lane. These straighteners also minimized possible boundary layers. Previous measurements have shown that water velocity in the 5 mm area closest to the wall was not significantly different from that in the center of the chamber (Stobutzki and Bellwood, 1997; Stobutzki, 1998; Fisher et al., 2000). Water flow speed was con- trolled by turning a calibrated ball valve. Flow rates were calibrated by recording the time taken for water Leis et al. Behavioral ontogeny in larvae and early luvenile Caranx ignobilis 403 B D Figure 1 Preserved larvae of giant trevally ^Caranx ignobilis) representing the range of sizes and developmental stages included in the present study. Preopercular spination was present in all specimens. The smallest specimen had a few scutes on the peduncle, the 14.5-mm specimen had scales forming elsewhere, and the largest specimen was largely scaled. Sizes are SL (Australian Museum catalogue numbers are provided). lA) 9.6 mm (1.43435-004); (B) 11.7 mm (1.43435-003); (C) 14.5 mm (I. 43435-004); (D) 18.5 mm (1.43435-002). flowing over the outlet weir to fill a container of known volume, divided by the cross-sectional area of the cham- ber. The average of five calibrations was used as the flow speed for a given valve angle. The chambers were calibrated each time they were set up. Flow speeds in excess of 50 cm/s could be achieved with this system. The chambers were plumbed into the seawater system of the NMMBA aquaria and laboratory, which provided a continual flow of "fresh" seawater. Before measurements of swimming speed and endur- ance were recorded, the larvae were acclimated to any differences in water quality between the holding tank and swimming chamber by gradual addition of seawater from the swimming chamber system. Larvae were placed in a chamber lane and allowed to acclimate for 5 minutes at 1 cm/s. Any larva showing signs of stress during the acclimation period was removed from the experiment and replaced with another individual. Water temperature in the swimming chamber ranged from 27° to 29°C. Two swimming parameters were measured: critical speed (t/|,^i(), which measures maximum swimming speed over periods of minutes, and endurance, which measures how long larvae can swim without food or rest. For C/pj.j, tests, starting at 1.5 cm/s flow speed was in- creased by approximately 2 cm/s every 5 minutes until the larvae were unable to swim against the flow. The elapsed time when each larva drifted to the downstream mesh was recorded. Critical speed (U^^^^) of larvae was calculated with the equation of Brett (1964): U., U + U/t, X U,), where U = penultimate speed; C/, = speed increment (ca. 2 cm/s in the present study); t = time swum in the final speed increment; and t^ = the time interval for each velocity increment (5 min). The total time for a critical speed measurement was proportional to the f/^.^^ achieved, and varied from 15 404 Fishery Bulletin 104(3) Table 1 Location-specific relationships between in situ speed (cm/s) and size (mm SL) in Caranx ignobilis larvae. NA=not applicable. Location Regression Wan Li Tong Her Chen Nan Wan Bay South Nan Wan Bay North Spd = 1.69SL - 6.2 Spd = 2.55SL - 12.2 Spd = 0.12SL + 4.9 Spd = 2.05SL - 17.9 0.83 <0.002 0.97 0.11 NA NA 0.72 0.03 minutes for the slowest individual to 82 minutes for the fastest. Given the endurance of which these larvae are capable (see "Results" section), it is unlikely that the larvae would have become fatigued over such time intervals. For endurance tests, a constant speed of 10 cm/s was used. Larvae were swum until they fatigued, which was defined as the time when a larva could no longer hold its position against the current and drifted onto the downstream mesh. During daylight hours, larvae were observed regularly and the exact time of fatigue (=swimming duration) was recorded. If a larva fatigued when it was not being observed, the time at fatigue was estimated as the midpoint between the time when the larva was last seen swimming and the time when it was found no longer swimming. Chambers were set up under cover so that they were shaded throughout the day, and a fluorescent light was used for illumination at night. The actual endurance measurements of time swum were converted to distance swum, by using the flow speed data, and were reported as kilometres swum. Fatigued larvae from both experiments were removed from the chamber, euthanized and fixed in Bouin's so- lution for one hour, then placed in 70% alcohol and stored. All preserved larvae were later examined under a dissection microscope to determine standard length (SL) and state of notochord flexion. Total lateral area (TA) and propulsive area (PA) of larvae (Fisher et al., 2000) were measured by using Scion Image for Windows (Beta 4.02, Scion Corporation, Frederick, MD). PA is TA minus the head and gut. In situ observations Four sites were used for in situ observations of C. ignobi- lis in the South China Sea. at the southern tip of Taiwan (ca. 22"N, 121°E). Two study sites were used in Nan Wan Bay (21 May 2004) where the depth range was 14-23 m and another two were used on the west coast just off the peninsula that delineates the west side of Nan Wan Bay in the vicinity of Wan Li Tong and Her Chen (14 May 2004) at a depth range of 17-31 m (Table 1). At each site, observations were made at least 50 m offshore, and all observations were conducted in the morning. Larvae were transported from the laboratory to the release sites in covered buckets fitted with a battery-op- erated aerator. Ambient seawater was gradually added to the buckets to allow the larvae to acclimatize to the surrounding water conditions. Then, 50% of the water in the bucket was exchanged for "fresh" seawater every hour. The behavior of the larvae was observed following the procedures of Leis et al. (1996) and Leis and Car- son-Ewart (1997, 1998). Two scuba divers descended to a depth of 5 m where the observer diver released a larva from a small container. Once the larva chose its initial trajectory, the two divers followed. The observer diver's sole job was to follow the larva while the second diver, following the observer, recorded data. The direction the observer diver was facing when he or she released the larva was chosen at random. Each larva was used only once and, where possible, was recaptured at the end of the observation period and preserved. The size of each larva was estimated, with the aid of a ruler, before re- lease. For recaptured larvae, the estimated and actual sizes were regressed, and this relationship was used to calculate the actual size of the five individuals that were not recaptured of the 24 released. Water column depth was measured by the depth sounder on the sup- port boat at the start of each larval release. Swimming speed, depth, and swimming direction were measured In situ. We attempted to observe each larva for 10 minutes, taking measurements of swim- ming direction and depth, using a dive compass and computer, respectively, every 30 seconds. Speed was calculated from distance traveled as measured by a calibrated flowmeter over the full period of observa- tion (Leis and Carson-Ewart, 1997). Larvae were not followed deeper than 15-18 m on some dives for safety reasons; therefore observations on some individuals were curtailed. Twenty-four larvae were observed, and observations on 18 of the larvae lasted the full 10 minutes. Five larvae swam monotonically downward to below our safety depth from 3 to 6 minutes, whereas one initially ascended, before descending after 5 min- utes and exceeding the safety depth after 8 minutes. Data analysis To determine the best predictor of performance, values of critical swimming speed and endurance were regressed against SL, TA, and PA by using linear, logarithmic, power, and exponential models. The model with the greatest r^ (coefficient of determination) was used as the best description. Leis et al Behavioral ontogeny In larvae and early luvenile Coronx ignobilis 405 All bearings are given as degrees magnetic. Cardinal directions are not abbreviated, but other directions are abbreviated (e.g., SE for southeast). From these bearings and the flow- meter data, we estimated swimming direction and speed in relation to the water (not the bottom). Circular statistical procedures followed Batschelet (1981) and Zar (1996). Mean vec- tor length (r), is a measure of angular disper- sion ranging from 0 (maximum dispersion) to 1 (lack of dispersion). The Rayleigh test was used for single-sample hypotheses about direc- tional swimming. Watson-Williams test was used for hypotheses about directional swim- ming involving more than one sample. Most of the circular statistical procedures were per- formed with Oriana software (vers. 2, Kovach Computing Services, Pentraeth, Wales, UK). For all statistical tests, we report actual P values whenever possible, but consider P<0.05 to con- stitute a "significant" result. Results Larvae used in the critical speed trials were 21-30 days old, and ranged from 8.0 to 16.5 mm SL. Larvae up to 18 mm were used in the in situ observations. All fins were formed in the smallest individual and over the size range studied, the larvae were very deep bodied, primarily silver, and usually had six dark vertical bars laterally. Although the specimens used on any day did not constitute a random sample of the available sizes, they could be used to obtain a rough estimate of the growth rate over the experimental period. The range of size-at-age on any day was as much as 6.5 mm. Average growth over this period was 0.36 mm/d, and the slope of the size versus age line was significantly different from zero, but the relationship between size and age was weak (size = 0.36 age + 2.87, P<0.001, r2 = o.28). Total lateral area increased approximately as the square of SL (TA = 0.51SL~°\ r2 = 0.93), whereas propulsive area increased at a lower rate (PA = 0.34SL^^'^, r- = 0.94). Critical speed Critical speed was measured in 54 individuals in eight batches of six over 10 days. Critical speed increased with size from about 12 cm/s in the smallest individu- als (ca. 8 mm) to about 40 cm/s in the largest (16.5 mm, Fig. 2). For SL, a linear model provided the best fit for this relationship (t/„,t= 2.72SL-9.55, P<0.001, r- = 0.51, n = 54), but the difference between the linear model and the others was small (r- for the other models ranged from 0.43 to 0.49). Both TA and PA gave somewhat better fits than SL ([/„,t= 0.39TA + 5.25, /-2 = 0.58; U„^^= 0.22PA -h5.73, r- = 0.55). Size was a much better predictor of critical speed than was age {U^^^^=0.90age-0.22, P<0.01, r2 = 0.13, K = 54). 40 1 ♦ ♦ ♦ to ■o CD « 20- 5 '^r^ . , * . y=2.72x-9.55 3 10- , ^ r^=0.51 \ 8 10 12 14 16 18 Size (mm, SL) Figure 2 Critica speed of giant trevally (Caranx ignnbilis) larvae of different sizes. At any size, variation in swimming speed was large. The fastest individual within each 1-mm increment of size was considered the best performer, and these con- stituted 16.6'7f of all larvae. These nine best perform- ing larvae swam at 20 to 40 cm/s. The nine best per- forming larvae had the same increase in performance with size (2.7 cm/s per each mm of growth) as the "all- individuals" line, but were 8 cm/s faster at all sizes ([7^,„ = 2.69SL-1.46, P<0.001, r2 = 0.76). Therefore, both average performers and "best performers" increased speed at 2.7 cm/s per each mm of growh in SL. Scaled speed in terms of body lengths per second did not increase significantly with size (P=0.07, r'~ = 0.06), and the overall mean (±standard error) was 19 ±0.7 BL/s. The fastest 10% of individuals swam at a mean t/,„, of 28.2 ±0.6 BL/s. In situ speed In situ speed was measured in 24 individuals ranging in size from 8.5 to 18 mm SL. Some individuals stopped and hovered at times during the observation period, but these pauses in swimming were incorporated into the speeds measured over the observation period. Observations were made a week apart when the larvae were 24 and 31 days old, respectively, and at a total of four locations that dif- fered in depth, exposure, and bottom type. Observed in situ speeds ranged from 3 to 19 cm/s. The in situ speeds were less than the critical speed measurements (see below). Mean scaled in situ speed was 9.5 ±0.9 BL/s. Overall, there was no significant increase in speed with size (P=0.89, r-=0.001, n=24), but it was obvious that the larvae behaved differently in the different loca- tions; therefore we examined the data for each location separately. Further, we eliminated observations on one larva that simply drifted without swimming and on four individuals that quickly swam to depths greater than 18 m and thus provided less than 5 minutes of observa- tions, because short measurements of speed taken at steep angles of descent can be inaccurate. The in situ speed data on the 19 remaining individuals were used. 406 Fishery Bulletin 104(3) For each location, in situ speed increased with size, as expected, except, perhaps for the largest individual (Table 1, Fig. 3). For the three locations with three or more observations, in situ speed increased at 1.7-2.6 cm/s per each mm of growth depending on location (Table 1), and the 95% confidence interval for the three slopes overlapped broadly. However, because the height of the speed versus size lines varied (v intercept ranged from -6 to -18), the actual speed at any size differed among locations by as much as 10 cm/s. For the in situ speed data, PA provided fits with r- values that were very similar to those of SL. The r- for PA was 0.07 higher than that for SL for Nan Wan Bay North, equal for Her Chen, and 0.15 lower for Wan Li Tong. 25- 20 - ■ E > ^^^^ 3 15 ■ y ^^ ■o X ^^ 0) ./* ^^ * o. ^^'^ en a 10 - w^^'^ ^y'^ Ul c ' /* ^ * ^^ * 5 ■ A 8 10 12 14 16 18 Size (mm. SL) Figure 3 In situ speeds of giant trevally [Caranx ignohilis) larvae and juveniles of different sizes at different locations. Locations of observations were ■ Wan Li Tong, 14 May 2004; ♦ Her Chen, 14 May 2004; ▲ Nan Wan Bay North, 24 May 2004; ▲ Nan Wan Bay South, 24 May 2004. See Table 1 for details of best-fitted regression lines for each area. No line was fitted to the two points from Nan Wan Bay South. Table 2 Comparison of two measures Caranx ignohilis larvae. Dat available. of hes size-speci indicate fie speed (cm/s) in that no data were Size (mm interval SL) Mean in situ speed [n) Mean Ratio; insitulU^^, 8.1- -9.0 8.7(3) 17.5 (3) 0.50 9.1- -10.0 11.4(4) 15.3(12) 0.75 10.1- -11.0 8.1(4) 18.0(6) 0.45 11.1- -12.0 -(0) 20.2(9) — 12.1 -13.0 8.6(3) 24.5(9) 0.35 13.1- -14.0 13.7(3) 30.7(7) 0.45 14.1- -15.0 19.5(1) 29.7(4) 0.66 15.L -16.0 -(0) 36.1(2) — 16.1- -17.0 -(0) 32.7(2) — 17.1- -18.0 7.0(1) -(0) — Mean ±SE 0.52+0.15 Comparison between critical speed and in situ speed For each 1-mm (SL) increment in size, we cal- culated a mean speed for both critical speed (54 individuals) and in situ speed (19 individuals). These size-specific speeds are summarized in Table 2 (both measurements were not available for every size increment). The ratio of the two size- specific measures (/;; situlU^^^^) ranged from 0.35 to 0.75, and had a mean value of 0.52 ±0.15. The size-specific speed measures were not significantly correlated (in s(7w = 0.45f;„„-hl.45, 7-^=0.47, P=0.13, ;; = 6). Scaled in situ speed (BL/s) was one-half scaled f/^,.,^. Endurance Endurance data were available for 12 individuals ranging in size from 9 to 14 mm SL, and for two ages (21 and 24 days at the start of the endur- ance run) measured in two batches of six. The range in fish swimming times was 12.6 to 105 hours. Swimming endurance increased with size from about 5 km in the smallest individuals to about 40 km in the largest (Fig. 4). Again, a linear model provided the best fit for this relationship (£;«rf. = 5.2SL-41.5, P<0.001, r- = 0.80, n=\2). and the difference between the linear model and the others was small (r- for the other models ranged from 0.73 to 0.79). In contrast to critical speed, the fit of the relationship between size and perfor- mance was not improved by using TA or PA instead of SL (r2=0.57 and 0.62, respectively). Endurance also increased with age, but endurance data were available only for larvae of two ages (21 and 24 days). Size was a better predictor of performance than was age {End.= 6.6age-128, P=0.001, r^=().6Q, n=\2). Vertical distribution There was not a significant overall relationship between mean swimming depth and size (P=0.29, r'' = 0.05). In spite of this, larvae of different sizes differed in their depth selection behavior. The smallest larvae (8-10 and 10-12 mm) swam pri- marily between 4 and 10-12 m (Fig. 5, A and B). One larva of each of these smallest size groups did, however, swim deep monotonically. If one ignore these two deep-swimming larvae, both smallest size groups reached an overall mean depth of about 9 m Leis et al Behavioral ontogeny In larvae and early juvenile Caranx ignobilis 407 at about 5 minutes. The smaller (8-10 mm) larvae then stayed at about 9 m, whereas the larger (10-12 mm) larvae gradually ascended, reaching a mean depth of about 6 m by the end of the 10-min observation period. In con- trast, larvae of 12-14 mm were more varied in their depth ranges, some reaching the surface, whereas others remained within 2-3 m of the release depth (Fig. 5C). Larvae of 12-14 mm had no overall temporal trend in depth and maintained a mean depth of about 5-6 m throughout the study period. Several of the larvae of this group did oscillate over a depth range of several m during the 10-min observa- tion period. The largest larvae (14-18 mm) had more varied vertical-distribution ranges, lead- ing to high amplitudes (Fig. 5D). Three of the five largest larvae swam deep monotonically and the other two oscillated — one reaching the surface after an initial descent, whereas the other reached our lower dive limit ( 16 m), but only after first ascending above the release depth for the first 5 minutes. We found a weak relationship between depth amplitude (difference between deep- est and shallowest observations) and size (a?np.=0A4SL + 1.7, P=0.03, r- =0.21, n=24); larger individuals swam over a greater range of depths owing to the tendency of larger lar- vae to either oscillate or to swim deep mono- tonically. Water column depth did not influ- ence either mean swimming depth (P=0.11, r2=0.11) or amplitude (P=0.06, r- = 0.15). Orientation Sixteen of the 24 larvae observed /?; situ had directional trajectories (P<0.05, Rayleigh test, Table 3). Figure 6 represents the direction frequency distribution of a C. ignobilis larva with average directionality. The remaining analyses included only the 16 individuals for which directional trajectories were found (termed "directional larvae" or "directional individuals'" for the sake of brevity), but the same result was obtained when all 24 individuals were included. If all locations were considered together, the mean swimming directions of these 16 individuals had no overall directionality (P>0.20, Rayleigh test). In contrast, there was signifi- cant overall directionality off the west coast (Fig. 7A) where the larvae swam, on average, offshore or toward the west (269°, Rayleigh test, P=0.024). There was no indication of any overall directionality in mean swim- ming direction in Nan Wan Bay (Fig. 7B) (mean direc- tion, 145°, P=0.58, Rayleigh test). Overall swimming direction, however, differed significantly between the west coast and Nan Wan Bay (Watson-Williams test, P=0.028). We found only a limited indication that the direction- ality of the larvae changed with size (Table 3). There was no ontogenetic increase in precision of directionality 45 1 40 • _ 35 ■ 1 30 ■ y=516x-4152 ♦ ♦ r-=a80 ^^^ S 25- ra 20 • ^^^^ ♦ ♦ f 15- ^ 10- * * ^^^^^ 5 • #^^» ♦ ■ ■'''■' ' 8 9 10 11 12 13 14 15 16 Size (mm. SL) Figure 4 Endurance swimming of giant trevally iCaranx ignobilis) larvae of different sizes. Table 3 Directionality of m situ swimming for Caranx ignobilis of different sizes. Only values from directional individuals are included for r (the length of the mean vector, which ranges from 0 to 1): the higher r, the more directional was the trajectory. Size interval (mm SLi n n directional (Rayleigh test, P<0.05) Range of r Mean r 8-10 8 6 0.41-0.86 0.73 10-12 5 2 0.60-0.90 0.7.5 12-14 6 4 0.60-0.83 0.72 14-18 5 4 0.45-0.92 0.74 (measured as r, the length of the mean vector) (r-=0.05, P>0.10). Similarly, although the mean size of the eight nondirectional larvae was 11.6 mm, and that of the 16 directional larvae was 12.2 mm, this difference was not significant (^test, P=0.22). In terms of overall orienta- tion, larvae observed off the west coast were smaller by about 4 mm (^test, P=0.002) and were studied a week earlier than those studied in Nan Wan Bay; therefore it is possible that the difference in overall swimming direction between locations was due to temporal or ontogenetic factors rather than to spatial factors. At the west coast location, the overall mean direction of the five small (8-9.5 mm) directional larvae was the same (268°) as the overall mean direction of the 4 large (10-14.5 mm) directional larvae, indicating there was no ontogenetic change in directionality. Finally, within trajectories, there was no increase in directionality with time. The r value for the first half of the trajectories was not significantly different from the r value for the second half of the trajectory (Wilcoxon signed-rank test, P>0.2). Nor was there any indication that larger larvae had a greater difference in r between the first and second halves of their trajectories than did smaller larvae (r2=0.02, P»0.20). 408 Fishery Bulletin 104(3) 8-10 mm SL 0 2 4 6 8 10 12 14 16 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 10-12 mm SL 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 12-14 mm SL 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 0 1 2 14-18 mm SL ^/^-^^^ 4 ■ '^^x-'—^ 6 • \ 8 \ 10 J \ 12 - \^ 14 - \ 16 - D ^^ ""^^ ^v— 18 - 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 Time (30-sec intervals) Figure 5 In silu vertical distribution trajectories for giant trevally (Caranx ignobilis) larvae and juveniles of four size groups. Locations of observations were ■ Wan Li Tong, 14 May 2004; ♦ Her Chen, 14 May 2004; ▲ Nan Wan Bay North, 24 May 2004; A Nan Wan Bay South, 24 May 2004. (A) 8-10 mm SL larvae; (B) 10-12 mm SL larvae; (C)12-14 mm SL larvae; ID) 14-18 mm SL larvae. 2 3 4 5^ 90 Figure 6 Frequency distribution of in situ swimming directions based on bearings taken every 30 seconds for an 18-mm-SL giant trevally (Caranx Ignobilis,) at Nan Wan Bay South. Bearings are grouped into 10" intervals. The radius that pen- etrates the outer circle (thin line) is the mean direction. This significantly directional distribu- tion (Rayleigh test, P=0.000001) has an r (length of the mean vector) of 0.76, which is close to the overall mean for the 16 directional individuals (Table 3). The numbers on dashed lines represent the number of times that bearing was recorded for a larva. Other behaviors Unplanned observations were made of feeding, interactions with other pelagic animals, and reactions with the bottom. At least nine of the 24 larvae were seen feeding, grabbing at, and apparently eating small zooplankton while swim- ming. One 12-mm larva encountered a large (ca. 30-40 cm) Naso sp. in open water and changed direction by 140° over the following minute. An 11-mm C. ignobilis encountered a jellyfish and briefly (<30 s) hovered under it. Five individuals (9-13.5 mm SL) came into visual contact with relatively high-relief coral reefs (at least the divers could clearly see the reefs) at Wan Li Tong and Her Chen. The larvae showed no particu- lar interest in the reefs; two ascended slightly (0.5-1 m), two maintained a distance of 3-4 m from the reef surface as they swam over it, and one turned offshore. One 12.5-mm larva passed through a strong thermocline at 8 m that was very noticeable to the divers without slowing or giving any indication that the temperature change was sensed. Leis et al.: Behavioral ontogeny in larvae and early luvenile Caronx ignobilis 409 West Coast 270 :? — IV B Nan Wan Bay 0 Figure 7 Frequency distributions of mean in situ swimming directions for giant trevally iCaranx ignobilis) larvae at two locations. Thick lines represent mean values for individuals, and the radius (thin line) that penetrates the outer circle represents the overall mean. The numbers on dashed lines represent the number of times that bearing was recorded for a larva. (A) West coast off Wan Li Tong and Her Chen, where larvae swam offshore: overall mean=268°, r=0.59, P=0.04. (B) Nan Wan Bay, where larvae had no overall swimming direction: overall mean = 146°, r=0.'29, P=0.58. "r" is the length of the mean vector, P is for the Rayleigh test. Discussion Previous studies of the behavior of larval carangids have focused primarily on their vertical distribution in the ocean and have been based on plankton net samples and on observations of the development of schooling- related behavior in the laboratory (see references at the beginning of this article). There have been no previous studies on the ontogeny of orientation or of vertical dis- tribution determined from in situ observations, and only one article has considered the development of swimming abilities of carangid larvae. Growth rates in C. ignobilis larvae of 0.36 mm/d in the laboratory and ca. 0.3-0.8 mm/d at the aquaculture farm at Kaohsiung, where we obtained larvae for our experiment, were comparable to the few reported field growth rates for larvae of other carangids. Small (<5 mm SL) larvae of Atlantic bumper (Chloroscombrus chrysurus) grew at 0.3-0.4 mm/d in the Gulf of Mexico (Leffler and Shaw, 1992), and larvae of two species of Trachurus grew at 0.2-0.7 mm/d (Hewett et al., 1985; Jordan, 1994). The C chrysurus larvae were found at 26-30°C, whereas the two Trachurus spp. were living at about 15°C. In T. declivis (greenback horse mackerel), 8-mm larvae were about 17 days old and 12-mm larvae were about 22 days old (Jordan, 1994). In T. syminetri- cus (Pacific jack mackerel), a 10-mm larva was 40 days old and a 20-mm larva was about 57 days old (Hewitt, 1981). Growth rates for C. ignobilis in culture or in the laboratory are unlikely to be relevant to field situations, and the available field measurements of other species are either for much smaller larvae or for individuals in much cooler water and are unlikely to be applicable to C ignobilis. Therefore, it is not possible to relate the size-based performance measures reported in our study to ages of larvae in the field. Further, for the labora- tory-reared larvae that we used, size was a much better predictor of performance than was age, as is generally found to be the case (e.g., Fuiman and Higgs, 1997; Clark et al., 2005). We found that using PA instead of SL as a measure of size resulted in little or no improve- ment in the proportion of the variation in swimming performance that was explained by the linear relation- ships developed in our study. Other studies have had mixed results in this regard (Fisher et al., 2000; Clark et al., 2005). Swimming speed of C. ignobilis increased rapidly with growth for both critical speed and in situ speed, although the rate of increase was greater for critical speed. Both in terms of actual speeds and rate of in- crease with growth, critical speeds of C. ignobilis are within the range of reported values for larvae of other warm water marine perciform fishes (Fisher et al., 2000; Clark et al., 2005). For larvae of benthic fishes about to settle (ca. 8-12 mm SL), pomacentrids have higher f/^^j, values than C. ignobilis, whereas apogonids, percichthyids, sciaenids, and sparids have similar {/^.^jj values. Seriola quinqueradiata is the only other caran- 410 Fishery Bulletin 104(3) gid for which we could find information on ontogeny of swimming abilities (Sakakura and Tsukamoto, 1999), although the method of measuring swimming perfor- mance (so-called "routine speed") differed from those used in our study. Reared S. quinqueradiata larvae and juveniles of 5-40 mm TL were filmed swimming in small (30-cm diameter, 7-cm deep) laboratory tanks lacking currents and the speed was measured over 3- min intervals. Speed of S. quinqueradiata increased at about 0.4 cm/s per each mm of growth until about 10 mm TL, after which speed remained steady at about 2-2.5 cm/s. This is much slower, both in terms of rate of increase and absolute speed, than was found for either measure of swimming speed in C. ignobilis. Possibly, the slow speeds for S. quinqueradiata were due to the small containers in which the fish were confined and which lacked currents, as has been found for other spe- cies (Theilacker and Dorsey, 1980). Although we did not study larvae smaller than 8 mm SL, it is safe to assume that smaller larvae have poorer swimming abilities. Previous investigations of swimming ontogeny have revealed that critical speed increases steadily with size (Fisher et al., 2000; Clark et al., 2005), as we did in our study; therefore it seems likely that a reasonable estimate of performance of smaller larvae can be obtained by extrapolating the linear relationship in Figure 2. Thus, a larva of 5 mm (about the size when the caudal fin is formed in caran- gids; Leis and Carson-Ewart, 2004) would be expected to have critical speeds of about 4-5 cm/s. We found that critical speeds for C. ignobilis were about twice in situ speeds, and similar relationships have been found for larvae of other species at other sizes (Fisher and Wilson, 2004; Leis and Fisher, in press). Therefore, a 5-mm C. ignobilis would be expected to be able to swim at about 2 cm/s (=72 m/h) in the ocean. This speed would have limited effectiveness for directly influencing horizontal position, but would be sufficient to determine vertical distribution at scales over which fish larvae are known to migrate vertically. To put these swimming performances into an ecologi- cal context requires knowledge of ambient currents in the area of interest. If it is assumed that effective swim- mers (those with swimming speed equal to or greater than average current speed [Leis and Stobutzki, 1999]) have the ability to control their dispersal by horizontal swimming, then the size at which larvae become effec- tive swimmers is of interest to dispersal modelers. For example, at Lizard Island on the Great Barrier Reef, average current speeds are 10-15 cm/s (Frith et al., 1986). At Lizard Island, average performing C. ignobilis would become effective swimmers at about 7-9 mm SL on the basis of critical speed, and at about 11-14 mm based on in situ speed which is about half of critical speed (assuming the relationship between in situ and critical speeds found in the present study). The best performers would be able to reach these mean Lizard Island current speeds at about 4-6 mm SL, if based on critical speed (assuming the relationship in Fig. 2 applies to smaller larvae), and at 8-12 mm if based on in situ speed. Larvae that are too small to be effective swimmers may still readily influence their dispersal by other means, particularly by vertical swimming where current velocity is not uniform with depth (Sponaugle et al., 2002; Paris and Cowen, 2004). Speeds of only 1 cm/s (36 m/h) are sufficient for vertical migration, and this could be achieved by very small larvae. Indeed, vertical movements by small carangid larvae are well documented in studies where plankton nets were used (e.g., Ahlstrom, 1959; Olivar and Sabates, 1997; Flores- Coto, et al., 2001). In situ speeds of C. ignobilis are within the range of those reported for other perciform species, although most of the available values are for settlement-stage larvae of demersal reef fishes (Trnski, 2002; Leis and Fisher, in press). Further, the rate of increase in speed at 1.6-2.6 cra/s per each mm of growth is similar to that found in the three other perciform species for which there are data (0.3-2.0 cm/s per each mm of growth; Leis et al., 2006). Differences in swimming speed among areas has been reported in a variety of reef fish larvae (Leis and McCormick, 2002); therefore finding a difference for C. ignobilis was not unexpected. This variation in behavior among areas and the fact that the rate of increase in speed with growth in the field was less than the increase of U^^^^ in the laboratory clearly demonstrates the complexities of applying labo- ratory measures to the field, and how behavior in the field can vary with location and situation. The cause for differences in behavior among locations was not clear in our study, but it is only through in situ observations that such differences can be discovered. The approximately linear increase in swimming speed with growth found in C. ignobilis was similar to that reported in other species (Fisher et al., 2000; Clark et al., 2005). The ontogeny of endurance in C. ignobilis was also linear, in contrast to that of some other spe- cies, which are reported to have a strongly concave curvilinear increase in endurance with growth (Fisher et al., 2000; Clark et al., 2005). Perhaps this linearity was the result of studying C. ignobilis larger than 8 mm. In other species, preflexion and early postflexion larvae (which we did not study) have very low endur- ance, but endurance increases rapidly with growth in postflexion larvae. All our speed measurements were made over periods of 10 minutes or less, but endurance measurements showed that speeds of 10 cm/s can be maintained from several to many hours, enabling distances of up to 40 km to be traversed. The ecological relevance of labora- tory measurements of swimming endurance is difficult to assess because, on one hand, these are forced mea- sures of performance, and it is unlikely that larval fishes would swim to exhaustion. On the other hand, endurance measurements are made without rest or food, and are therefore conservative because larvae of other species of the size studied in the present study can swim at least three times farther when fed (Fisher and Bellwood, 2001; Leis and Clark, 2005). In any case, it is clear that C. ignobilis larvae can swim for extended Leis et a\ ; Behavioral ontogeny in larvae and early juvenile Caranx ignobilis 411 periods at speeds that would make them effective swim- mers in many situations. According to the relationship in Figure 4, C. ignobilis larvae of about 9 mm could swim without food or rest for about 13 hours (5 km at 10 cm/s) at a speed that is within their observed ca- pabilities in situ: this ability increases markedly with growth. Clearly, C. ignobilis larvae from a size of 1 cm or less are capable of maintaining speeds of similar magnitude to mean ambient currents for periods of time long enough to strongly influence their dispersal. Knowledge of vertical distribution of fishes is impor- tant because many things vary vertically in the ocean, including current velocity, food concentrations, and predators, all of which can strongly influence survival and dispersal. For these vertically varying factors, the conditions that larval and small juvenile fishes actually encounter are fully under their control because swim- ming abilities and sensory abilities capable of determin- ing vertical position in the water column develop at a very small size. The C. ignobilis we observed in the ocean had considerable control over their vertical distri- bution, and the depths selected changed with size. Ma- suda and Tsukamoto (1996) noted ontogenetic changes in preferred light intensity in larvae of P. dentex, and it is possible that such changes influence the selection of depth in larval C. ignobilis. Two notes of caution are necessary in interpreting the depth-selection behavior and ontogenetic changes in this behavior. The larg- est larvae were primarily observed at one site, and, it is possible that vertical-distribution behavior varies among areas (Leis, 2004). Secondly, the largest larvae tended to swim to below our safety depth and we there- fore do not know if they continued to descend, if they subsequently ascended (in accord with the oscillatory behavior observed in a number of larger individuals), or if they leveled off In a similar experiment, much larger (60-140 mm) Pseudocaranx dentex juveniles initially descended several meters before ascending to a much shallower "preferred depth," usually within 60 seconds (Kuwada et al., 2000). We did not observe such ascents following an initial descent, but we cannot rule out the possibility that they occur for individuals that swam below our safe diving depth (Fig. 5D). Orientation in the pelagic environment is difficult for fish larvae because of the movement of the water column and the scarcity of reference points in a mov- ing pelagic environment. Orientation is, however, nec- essary for the arrival of late-stage larvae at nursery habitat and can greatly influence dispersal trajectories (Shanks, 1995; Leis and Carson-Ewart, 2003). Two- thirds of the C. ignobilis larvae that we observed swam directionally, as opposed to swimming randomly. This is a somewhat smaller percentage than the 80-100% for directional individuals at the settlement stage for demersal coral reef fishes of similar size (Leis and Car- son-Ewart, 2003). Neither the proportion of individual C. ignobilis that swam directionally, nor the precision of their directional swimming, increased with growth. The lack of ontogenetic change in individual orienta- tion— in contrast to speed, endurance, and vertical dis- tribution— indicates that orientation ability develops early (at less than 8 mm SL) in C. ignobilis larvae. Of course, orientation may improve in individuals larger than those we studied. The offshore swimming direction of C. ignobilis lar- vae found off the west coast is similar to behavior re- ported for larvae of some other species at other locations (e.g., Leis et al., 1996; Leis and Carson-Ewart, 2003), but could, in fact, be orientation to the west, rather than offshore. Observations off the east coast would be required to separate the possibilities, and any such test would need to take into account the possibility of temporal factors. The lack of any directionality in Nan Wan Bay is possibly a result of the "U" shape of the bay. Such a bay, where "offshore" constitutes less than half of the possible swimming directions, and where west is onshore, may present challenges to orientation that are greater than those present off an open coast. Alternatively, a bay may offer characteristics that would induce larvae to remain, rather than swim away as they did off the more open coast. Or, older larvae may prefer inshore environments, whereas younger ones do not. The important points, however, are that larvae of 8-15 mm SL demonstrated orientated swimming in the field and that smaller larvae swam in the same direction as larger ones. We can offer only speculation as to the types of cues that C. ignobilis may use to orient their swimming, and refer the reader to other sources for this information (e.g., Kingsford et al., 2002; Leis and Mc- Cormick, 2002). Our observations on C. ignobilis sup- port other research showing that orientation behavior in fish larvae can be location-dependent (Leis et al., 1996; Stobutzki and Bellwood, 1998; Leis and Carson-Ewart, 2003), and if so, this feature add a further complication for attempts to model dispersal. Our unplanned in situ observations of behavior shed light on little-known aspects of the early life history of C. ignobilis, which are otherwise difficult to study. Feeding activity by 8-18 mm larvae of C. ignobilis was common while they were swimming, as has been found with other species (Leis and Carson-Ewart, 1998). The only larva to encounter a large pelagic fish reacted in the same way as that reported for other species, namely with the apparent goal of moving away from the poten- tial large predator before the small fish could be readily detected (Leis and Carson-Ewart, 2001). Such behavior in a reared individual with no experience with piscivo- rous fishes indicates that the behavior is not learned. Many carangids associate with jellyfish medusae when small (Shojima, 1962); therefore the brief association we observed between a jellyfish and a 10-mm C. ignobilis was not unexpected. Caranx ignobilis is considered to be reef-associated during some periods of its life history, but individuals of 9-13 mm do not seem to be inclined to associate with coral reefs. Unlike the larvae of many demersal reef fishes (Leis and Carson-Ewart, 2002), the larval C. ignobilis that we observed made no at- tempt to settle and showed no interest in coral reefs. If anything, C. ignobilis moved away from the coral reefs they encountered, although this behavior is not uncom- 412 Fishery Bulletin 104(3) mon in larvae of other species (Leis and Carson-Ewart, 1999, 2002). We observed one individual penetrate a thermocline, indicating not only a tolerance for rapid temperature change, but also, that the thermocline is not a barrier to vertical movement for C. ignobllis, as has been proposed sometimes for larvae of other species (e.g.. Gray and Kingsford, 2003). We used reared larvae in our study and it is possible that the behavior we observed may have differed from that of wild individuals. Results of studies comparing wild and reared individuals have been inconsistent. Some studies have shown large differences in behavior of reared and wild fishes, the most marked of which involve learned interactions with predators (Brown and Laland, 2001). Differences in swimming perfor- mance (in both directions) have been documented in some studies, but not others, or have been present at some developmental stages but not others (e.g., Blax- ter, 1975; Dunmall and Shreer, 200.3; Smith and Fui- man, 2004). Even if differences in predator-related behavior exist, swimming and orientation are pre- sumably less dependent on experience and learning, and there is less reason to expect them to differ as well. The possibility that reared larvae of C. ignobilis may behave differently from wild larvae cannot be entirely dismissed, but the fact that C. ignobilis larvae were reared in conditions approaching those found in the field (but with the absence of predators) — kept in large outside ponds at an aquaculture farm and fed a natural assemblage of zooplankton — would presum- ably make large differences in behavior with their wild counterparts less likely. Unfortunately, we did not have access to wild C. ignobilis, and therefore we could not make direct tests; moreover, little is known of behavior in other carangid larvae upon which we might base comparisons. Caranx ignobilis apparently schools when young (My- ers, 1999), and the size range of larvae we studied overlapped with that at which schooling begins in other carangid species (Masuda and Tsukamoto, 1998). We did not attempt to observe schooling in C. ignobilis, but it is conceivable that some of the aspects of behavior may differ between individuals that are schooling and those that are solitary. We observed strong ontogenetic changes in behavior of C. ignobilis across a size range of 8 to 18 mm SL. At any size, there was usually a wide range of perfor- mances present, and the best performers were able to swim much better than average performers. Given the very high mortality rates experienced by marine fish larvae, the average larva at any given size is unlikely to survive (Gushing, 1990), and it may be only the excep- tional performers that do survive. Therefore, it may be more appropriate to use values for the best performers than for average performers in models of dispersal or survival. Increases in swimming abilities with growth were most marked, and significant swimming abilities would probably be present, in larvae as small as 5 mm SL based on extrapolation of performance at size rela- tionships. Ontogenetic changes in vertical distribution behavior were more complex and difficult to interpret, but depth selection seemed to be more variable (both within and among individuals) in larger larvae. We found that orientation behavior of small C. ignobilis was already developed and did not increase ontogenetically, although the direction they swam differed between locations. Behavior in small C. ignobilis was complex and swimming abilities were well developed. These behaviors are capable of strongly influencing survival and dispersal and show that the larvae of some pelagic fishes have behavioral capabilities similar to those re- cently documented for larvae of demersal reef fishes (Leis and McCormick, 2002). Acknowledgments We thank the Director of NMMBA, Lee-Shing Fang, for the opportunity to work at NMMBA and his staff for their excellent cooperation, particularly Chao-Yuan Chung for assistance in the wet laboratory. Our work could not have proceeded without the assistance of Colin Wen and Kun-Ping Kan in all phases of the work. Li- Hua Chao generously introduced us to many Taiwan- ese aquaculturists and spent many hours helping us obtain larvae. Mark Brown assisted ably, especially with laboratory work. Ray-Ming Chen, our dive boat skipper, made possible our field work, and Bill Watson helped with literature. This research was supported by an ARC Discovery Grant (DP0345876), and a DST International Science Linkages Grant (IAP-IST-CG03-0043) to JML, and by the Australian Museum. Reiji Masuda made constructive comments on the manuscript. Literature cited Ahlstrom, E. H. 1959. Vertical distribution of pelagic fish eggs and larvae off California and Baja California. Fish. Bull. 60:107-146. Batschelet, E. 1981. Circular statistics in biology, 371 p. Academic Press, London. Blaber, S. J. M., and D. P. Cyprus. 1983. The biology of the Carangidae (Teleostei) in Natal estuaries. J. Fish. Biol. 22:173-188. Blaxter, J. H. S. 1975. Reared and wild fish — how do they compare? Proc. 10th Europ. Symp. Mar. Biol. 1:11-26. Brett, J. R. 1964. The respiratory metabolism and swimming perfor- mance of young sockeye salmon. J. Fish. Res. Board Can. 21:1183-1226. Brown, C and K. Laland. 2001. Social learning and life skills training for hatchery reared fish. J. Fish. Biol. 59:471-493. Clark, D. L., J. M. Leis, A. C. Hay, and T. Trnski. 2005. Swimming ontogeny of larvae of four temperate marine fishes. Mar. Ecol. Prog. Ser. 292: 287-300. Gushing, D. H. 1990. Plankton production and year class strength in Leis et aL; Behavioral ontogeny in larvae and early juvenile Caronx ignobilis 413 fish populations: an update of the match/mismatch hypothesis. Adv. Mar. Biol. 26:249-293. Dunmall, K. M., and J. F. Schreer. 2003. A comparison of the swimming and cardiac per- formance of farmed and wild Atlantic salmon, Salino satar. before and after gamete stripping. Aquaculture 220:869-882. Fisher, R., and D. R. Bellwood. 2001. Effects of feeding on the sustained swimming abili- ties of late-stage \arva\ Amphipnon melanopus. Coral Reefs 20:151-154. Fisher, R., D. R. Bellwood, and S. D. Job. 2000. Development of swimming abilities in reef fish larvae. Mar. Ecol. Prog. Ser. 202:163-173. Fisher, R., and S. K. Wilson. 2004. Maximum sustainable swimming speeds of late- stage larvae of nine species of reef fishes. J. Exp. Mar. Biol, Ecol. 312:171-186. Flores-Coto, C, R. Rivas-Vega, F. Zavala-Garcia, and J. Sanches- Robles. 2001. Vertical distribution of larval carangids in the southern Gulf of Mexico. Gulf Carib. Res. 13:1-8. Frith, C. A., J. M. Leis, and B. Goldman. 1986. Currents in the Lizard Island region of the Great Barrier Reef Lagoon and their relevance to potential movements of larvae. Coral Reefs 5:81-92. Fuiman, L. A., and D. M. Higgs. 1997. Ontogeny, growth and the recruitment process. In Early life history and recruitment in fish populations (R. C. Chamber, and E. Trippel, eds.), p. 225-249. Chap- man and Hall, London. Gray, C. A., and M. J. Kingsford. 2003. Variability in thermocline depths and strength, and relationships with vertical distributions offish larvae and mesozooplankton in dynamic coastal waters. Mar. Ecol. Prog. Ser. 247:211-224. Hewitt, R. 1981. The value of pattern in the distribution of young fish. Rapp. P.-V. Reun. Cons. Int. Explor. Mar 178: 229-236. Hewitt, R. P., G. H. Theilacker, and N. C. H. Lo. 1985. Causes of mortality in young Jack mackerel. Mar. Ecol. Prog. Ser. 26: 1-10. Hoese, D. F., D. J. Bray, G. R. Allen, C. Allen, N. Cross, and J. R. Paxton. In press. Pisces. Zoological catalogue of Australia. Australian Biological Resources Survey, Canberra, Australia. Johannes, R. E. 1981. Words of the lagoon, 245 p. Univ. California Press, Berkeley, CA. Jordan, A. R. 1994. Age, growth and back-calculated birthdate dis- tributions of larval jack mackerel, Trachurus declivis (Pisces:Carangidae), from eastern Tasmanian coastal waters. Aust. J. Mar. Freshw. Res. 45:19-33. Kingsford, M. J., J. M. Leis, A. Shanks, K. Lindeman, S. Morgan, and J. Pineda. 2002. Sensory environments, larval abilities and local self-recruitment. Bull. Mar. Sci. 70 (suppl. 1): 309-340. Kuiter, R. H. 1993. Coastal fishes of south-eastern Australia, 437 p. Crawford House, Bathurst, New South Wales, Australia. Kuwada, H., R. Masuda, S. Shiozawa, T. Kogane, K. Imaizumi, and K. Tsukamoto. 2000. Effect of fish size, handling stress and train- ing procedure on the swimming behavior of hatchery- reared striped jack: implications for stock enhancement. Aquaculture 185:245-256. Leffler. D. L., and R. F. Shaw. 1992. Age validation, growth, and mortality of larval Atlantic bumper (Carangidae: Chloroscombrus chry- surus) in the northern Gulf of Mexico. Fish. Bull. 90:711-719. Leis. J. M. 2004. Vertical distribution behaviour and its spatial variation in late-stage larvae of coral-reef fishes during the day. Mar. Freshw. Behav. Physiol. 37:65-88. Leis, J. M., and B. M. Carson-Ewart. 1997. Swimming speeds of the late larvae of some coral reef fishes. Mar. Ecol. Prog. Ser. 159:165-174. 1998. Complex behaviour by coral-reef fish larvae in open-water and near-reef pelagic environments. Environ. Biol. Fish. 53: 259-266. 1999. In situ swimming and settlement behaviour of larvae of an Indo-Pacific coral-reef fish, the Coral Trout (Pisces, Serranidae, Plectropomus leopardiis). Mar. Biol. 134:51-64. 2001. Behavioural differences in pelagic larvae of four species of coral-reef fishes between two environments: ocean and atoll lagoon. Coral Reefs 19:247-257. 2002. hi situ settlement behaviour of damselfish larvae (Pisces: Pomacentridae). J. Fish. Biol. 61:325-346. 2003. Orientation of pelagic larvae of coral-reef fishes in the ocean. Mar. Ecol. Prog. Ser. 252:239-253. 2004. The larvae of Indo-Pacific coastal fishes, 870 p. Brill, Leiden, The Netherlands. Leis, J. M., and D. L. Clark. 2005. Feeding greatly enhances endurance swimming of settlement-stage reef-fish larvae (Pomacentridae). Ich- thyol. Res. 52:185-188. Leis, J. M., and R. Fisher. In press. Swimming speed of settlement-stage reef-fish larvae measured in the laboratory and in the field: a comparison of critical speed and in situ speed. Proc. 10th Internat. Coral Reef Symp., Okinawa, Japan. Leis, J. M., A. C. Hay, and T. Tronski. 2006. In situ ontogeny of behaviour in pelagic larvae of three temperate, marine, demersal fishes. Mar. Biol. 148(3):655-669. Leis, J. M., and M. I. McCormick. 2002. The biology, behaviour and ecology of the pelagic, larval stage of coral-reef fishes. In Coral reef fishes: new insights into their ecology (P. F. Sale, ed.), p. 171-199. Academic Press, San Diego, CA. Leis, J. M., and I. C. Stobutzki. 1999. Swimming performance of late pelagic larvae of coral-reef fishes: in situ and laboratory-based measurements. In Proc. 5th Indo-Pacific Fish Conf , Noumea, 1997 (B. Seret, and J.-Y. Sire, eds.), p. 575-583. Societe Francaise d'Ichtyologie & Institut de Recherche pour le Developpment, Paris. Leis, J. M., H. P. A. Sweatman, and S. E. Reader. 1996. What the pelagic stages of coral reef fishes are doing out in blue water: daytime field observations of larval behaviour. Mar. Freshw. Res. 47:401-411. Masuda, R., and K. Tsukamoto. 1996. Morphological development in relation to photo- 414 Fishery Bulletin 104(3) taxis and rheotaxis in the striped jack, Pseudocaranx dentex. Mar. Freshw. Behav. Physiol. 28: 75-90. 1998. The ontogeny of schooling behaviour in the striped jack. J. Fish. Biol. 52:483-493. 1999. School formation and concurrent developmental changes in carangid fish with reference to dietary conditions. Environ. Biol. Fish. 56:243-252. Myers, R. F. 1999. Mirconesian reef fishes. A comprehensive guide to the coral reef fishes of Micronesia, 330 p. Coral Graphics. Guam, U.S.A. Olivar, M. P., and A. Sabates. 1997. Vertical distribution of fish larvae in the north- west Mediterranean Sea in spring. Mar. Biol. 129: 289-300. Paris, C. B., and R. K. Cowen. 2004. Direct evidence of a biophysical retention mech- anism for coral reef fish larvae. Limnol. Oceanogr. 49:1964-1979. Randall, J. E., G. R. Allen, and R. C. Steene. 1997. Fishes of the Great Barrier Reef and Coral Sea, 557 p. Crawford House, Bathurst, New South Wales, Australia. Sakakura, Y., and K. Tsukamoto. 1996. Onset and development of cannibalistic behav- iour in early life stages of yellowtail. J. Fish. Biol. 48:16-29. 1999. Ontogeny of aggressive behaviour in schools of yellowtail, Seriola quinqueradiata. Environ. Biol. Fish. 56:231-242. Shanks, A. L. 1995. Orientated swimming by megalopae of several eastern North Pacific crab species and its potential role in their onshore migration. J. Exp. Mar. Biol. Ecol. 186:1-16. Shojima, Y. 1962. On the postlarvae and juveniles of carangid fishes collected together with the Jelly-fishes. Contrib. Seikai Regional Fish. Res. Lab. 147:49-58. Smith, M. E., and L. A. Fuiman. 2004. Behavioral performance of wild-caught and labo- ratory-reared red drum Sciaenops ocellatus (Linnaeus) larvae. J. Exp. Mar. Biol. Ecol. 302:17-33. Smith-Vaniz, W. F. 1999. Carangidae. In FAO species identification guide for fishery purposes. The living marine resources of the Western Central Pacific, vol. 4, Bony fishes, part 2 (Mugilidae to Carangidae) (K. E. Carpenter, and V. H. Niem, eds.), p. 2659-2756. FAO, Rome. Sponaugle, S., R. K. Cowen, A. Shanks, S. G. Morgan, J. M. Leis, J. Pineda, G. W. Boehlert, M. J. Kingsford, K. Lindeman, C. Grimes, and J. L. Munro. 2002. Predicting self-recruitment in marine populations: biophysical correlates. Bull. Mar. Sci. 70:341-376. Stobutzki, I. C. 1998. Interspecific variation in sustained swimming ability of late pelagic stage reef fish from two families (Pomacentridae and Chaetodontidae). Coral Reefs 17:111-119. Stobutzki, L C, and D. R. Bellwood. 1994. An analysis of the sustained swimming abilities of pre- and post-settlement coral reef fishes. J. Exp. Mar. Biol. Ecol. 175:275-286. 1997. Sustained swimming abilities of the late pelagic stages of coral reef fishes. Mar. Ecol. Prog. Ser. 149: 35-41. 1998. Nocturnal orientation to reefs by late pelagic stage coral reef fishes. Coral Reefs 17:103-110. Sudekum, A. E., J. D. Parrish, R. L. Radtke, and S. Ralston. 1991. Life history and ecology of large jacks in undis- turbed, shallow, oceanic communities. Fish. Bull. 89:493-513. Theilacker, G., and K. Dorsey. 1980. Larval fish diversity, a summary of laboratory and field research. LINESCO Intergov. Oceanogr. Comm. Workshop Rep. 28:105-142. Trnski, T. 2002. Behaviour of settlement-stage larvae of fishes with an estuarine juvenile phase: in situ observations in a warm-temperate estuary. Mar. Ecol. Prog. Ser. 242:205-214. von Westernhagen, H. 1974. Observations on the natural spawning of Alec- tis indicus (Ruppell) and Caranx ignobilis (Forsk.) (Carangidae). J. Fish. Biol. 6:513-516. Wetherbee, B. M., K. N. Holland, C. G. Meyer, and C. G. Lowe. 2004. Use of a marine reserve in Kaneohe Bay, Hawaii by the giant trevally, Caranx ignobilis. Fish. Res. 67:253-263. Yu, N.-H. 2002. New challenges: breeding species in Taiwan, Vin. Fish Breeding Assoc, Republic of China, Kaoh- siung, Taiwan, 52 p. [In Chinese, but in English at www.fish.org.tw.] Zar, J. H. 1996. Biostatistical analysis, 3rd ed., 662 p. Prentice Hall, Upper Saddle River, NJ. 415 Abstract — Data storage tags (DSTs) were applied to Atlantic salmon (Salinn salar L.) smolts during their seaward migration in the spring of 2002 at a fish counting fence on Campbellton River, Newfoundland. Our objectives were to discover whether or not salmon smolts could carry DSTs and survive, whether or not useful data on thermal habitat could be obtained and inter- preted, and whether or not salmon smolts moved vertically in the water column. Data were downloaded from 15 of the recovered tags and revealed the hourly water temperatures expe- rienced by the fish for periods of 3 to 71 days. The data on the DSTs were analyzed for temperature patterns in relation to migration behavior and diurnal movement of the fish. While in the sea, the DSTs recorded night temperatures of 12.5°C, which were higher than day temperatures of 11.6'C; the record from moored recorders, however, indicated that sea temperatures actually declined at night. It is hypothesized that posts- molts avoid avian predators during daylight hours by positioning them- selves deeper in the water column and that they were pursuing prey during the deeper vertical descents or ascents noted during the periods of more rapid changes in temperature. Diurnal and nocturnal temperatures for Atlantic salmon postsmolts (Salmo salar L.) during their early marine life David G. Reddin' Peter Downton' Kevin D. Friedland^ ' Science Branch, Department ol Fisheries and Oceans 80 East White Hills Road P. O. Box 5667 St. John's, Newfoundland Canada AlC 5X1 E-mail address (for D G Reddin) ReddinDe'dfo-mpo.gc ca) ^ National Marine Fisheries Service 28 Tarzwell Drive Narragansett, Rhode Island 02882 Manuscript submitted 27 September 2004 to the Scientific Editor's Office. Manuscript approved for publication 28 September 2005 by the Scientific Editor Fish. Bull. 104:415-428 (2006). Our knowledge of the ecology of ma- rine fish, and specifically salmon, at sea has recently been enhanced by information recovered from data stor- age tags (DSTs) and tracking tags (Sturlaugsson^; LaCroix and McCurdy, 1996; Wada and Ueno^; God0 and Michalsen, 2000; Reddin et al., 2004). Direct tracking, although valuable, has a number of shortcomings, includ- ing the usually short duration of tracking because of the requirement to follow the fish. This short-term tracking often results in interrupted data streams and the added concern that the actual tagging may affect, at least in the short term, the sub- sequent behavior of the fish (LaCroix and McCurdy, 1996; Walker et al., 2000). Studies of salmon at sea over long periods require new tools and several authors have noted that DSTs are among the least expensive meth- ods of answering some of the ques- tions posed regarding the sea life of salmon. Also, the near extirpation of some stocks, particularly in the Inner Bay of Fundy area of Canada and Maine, U.S. A, requires an explana- tion of the causative factors (Boehlert, 1997; Anon.*; Friedland et al., 2001; Reddin et al., 2004). DSTs should be able to provide better information on the natural behavior of individual fish, their specific habitats, and loca- tion over the period of their residence in the sea because of the longer period that information can be collected. On both sides of the North Atlan- tic, attention has been recently fo- cused on the downturn in Atlantic salmon stocks and the potential cause or causes (Nickson^; Mills"'; DFO*^; ' Sturlaugsson, J. 1995. Migration study on homing of Atlantic salmon iSalmo salar L.I in coastal waters W-Iceland — depth movements and sea temperatures recorded at migration routes by data stor- age tags. ICES CM. (council meeting) 1995/M: 17, 13 p. - Wada, K. and Y. Ueno. 1999. Homing behavior of chum salmon determined by an archival tag. NPAFC (North Pacific Anadromous Fish Commission Doc. 425, 29 p. 3 Anon. 1998. Report of the Study Group on ocean salmon tagging experiments with data logging tags. ICES CM. (council meeting) 1998/G: 17, 34 p. ■■ Nickson, Sir D. 1991. Chairman's report, Atlantic Salmon Trust, progress report, December 1991, 1 p. Atlantic Salmon Trust, Moulin, Pitlochry, Perth- shire, Scotland PH16 5JQ. ■5 Mills, D. (ed.). 1996. Enhancement of spring salmon. Proceedings of a one- day conference held in the rooms of the Linnean Society of London; 26 January 1996, 135 p. Atlantic Salmon Trust, Moulin, Pitlochry, Perthshire PH16 5JQ Scotland. ^ DFO (Dept. of Fisheries and Oceans). 1998. Atlantic salmon abundance over- view for 1997. DFO CSAS (Canadian Stock Assessment Secretariat), Stock Status Report D0-02(1998), 22 p. 416 Fishery Bulletin 104(3) O'Neil et al."; Anon.^). Many authors have noted that information on the sea life of Atlantic salmon (Salmo salar L.) may aid in our understanding of the recent increased mortalities that occur at sea compared to those experienced in the 1970s (Mills, 1989; Reddin and Friedland, 1993; Dempson et al.^; Jacobsen, 2000; Reddin et al., 2000). In the southern end of the range of Atlantic salmon in North America, declines have been so severe that some stocks are now threatened with extinction and others are currently considered to be extirpated (Marshall et a\}°). The recent listing of seven salmon stocks in the State of Maine under the U.S. Endangered Species Act and in rivers in the upper Bay of Fundy further underscores the urgency to under- stand more completely the entire salmon's life history, including the period spent in the ocean. Although the cause (or causes) is unknown, the continued declines in abundance appear directly related to increased mortal- ity at sea, perhaps from predation (DFO, 1998; Cairns and Reddin"; O'Neil et al.'). New techniques have recently been applied to the study of salmon in the sea on both sides of the North American continent for Pacific and Atlantic salmon, and recommendations have been made on their future use (Boehlert, 1997; Anon'^). Although large scale experi- ments with complex tags are attractive, the high cost per tag and likelihood of low returns, in part due to reduced commercial fishing for Atlantic salmon, have indicated that a smaller scale experiment with inexpen- sive simple tags may be the best place to start (Anon. 3). Reddin et al. (2004) demonstrated that these tags when applied to Atlantic salmon adults could provide insight into some aspects of salmon life history. Overall, the objectives of the present study were 1) to determine if DSTs could be applied to Atlantic salmon smolts and if data could be successfully recovered, 2) to learn more about the thermal habitat of Atlantic salmon during its early marine life, and 3) to test the hypoth- esis of the commonly held assumption that salmon are mainly surface or near to surface dwellers. This ar- ticle describes the results of the migration of Atlantic salmon smolts from Campbellton River, Newfoundland, in relation to sea water temperatures recorded by DSTs in 2002. Materials and methods Experimental site Atlantic salmon smolts, defined by Allan and Ritter (1977) as juvenile salmon that are making the transi- tion from freshwater to the marine environment, were obtained in 2002 for tagging from an enumeration fence on Campbellton River, Newfoundland (Fig. 1) where in past years, high numbers were available in early spring (Downton and Reddin'-). Also, because Campbellton River has a counting fence that is suitable for obtain- ing counts of adult salmon as they re-enter freshwater, there was a strong possibility of recapturing a return- ing tagged salmon so that its tag could be recovered (Fig. 1). Smolt and adult salmon have been counted at Campbellton River annually over the last 11 years (O'Connell et al.''). In total, there were about 32,600 salmon smolts counted through the counting fence at Campbellton River in 2002, and of these, 311 (-1% of total smolts) were tagged and released with data stor- age tags (DSTs). Owing to the size and weight of the tags, only smolts over 20 cm fork length were selected for application of tags. Average length of tagged smolts was 25 cm compared to an average length of 18 cm for untagged smolts. Larger smolts were also purposely selected to increase probability of survival. ' O'Neil, S., J. Ritter, and K. Robichaud-LeBlanc. 2000. Pro- ceedings of a workshop on research strategies into the causes of declining Atlantic salmon returns to North American rivers. CSAS (Canadian Stock Assessment Secretariat), Proceedings Series 2000/18, 80 p. Department of Fisher- ies and Oceans, Government of Canada, 200 Kent Street, Stn. 12032, Ottawa, Ontario, Canada KIA 0E6. 8 Anon. 2003. Report of the Working Group on North Atlan- tic Salmon. ICES Headquarters, Copenhagen, 31 Mareh-10 April 2003. ICES CM. (council meeting) 2003/ACFM: 19, 310 p. 3 Dempson, J. B., D. G. Reddin, M. F. O'Connell, J. Helbig, C. E. Bourgeois, C. C. Mullins, T. R. Porter, G. Lilly, J. E.Carscadden, G. B. Stenson, and D. Kulka. 1998. Spatial and temporal variation in Atlantic salmon abundance in the Newfoundland-Labrador region with emphasis on factors that may have contributed to low returns in 1997. DFO, CSAS, Res. Doc. 98/114, 161 p. 1" Marshall, T. L., G. J. Chaput, P. G. Amiro. D. K. Cairns. R. A. Jones, S. F. O'Neil, and J. A. Ritter. 1999. Assessments of Atlantic salmon stocks of the Maritimes Region, 1998. DFO, CSAS Res. Doc. 99/25, 80 p. 11 Cairns, D. K., and D. G. Reddin. 2000. The potential impact of seal and seabird predation on North American Atlantic salmon. DFO, CSAS Res. Doc. 2000/012, 36 p. Tags The data storage tags (DSTs), sometimes referred to as archival tags, used in this study were iB-4 tags (Maxim/ Dallas Semiconductor Corporation, Sunnyvale, CA) and were repackaged for use on fish by Alpha Mach Devices Inc. (Mont St-Hilaire, Quebec, Canada). The iB4 tags are small microprocessor-based data loggers embedded in green urethane on which is shown (at bottom right corner of tag) a return address, an identifying number, and the offer of a reward for return of the tag. The iB4 DST records temperature over a range of -5° to 26°C 1- Downton. P. R.. and D. G. Reddin. 2004. Status of Atlan- tic salmon (Salmo salar L.) in Campbellton River, Notre Dame Bay (SFA 4), Newfoundland in 2003. DFO, CSAS Res. Doc. 2004/043, 49 p. 13 O'Connell, M. F., J. B. Dempson, C. C. Mullins, D. G. Reddin, C. E. Bourgeois, T. R. Porter, N. M. Cochrane, and D. Caines. 2003. Status of Atlantic salmon (Sa/nio salar L.) stocks of insular Newfoundland (SFAs 3-14A), 2002. CSAS Res. Doc. 2003/002, 63 p. Reddin et al : Diurnal and nocturnal temperatures for Salmo solar 417 54"N 52°N 50°N 48'N 46°N 58 W 56 W 54-W 52 W Figure 1 Map of Newfoundland. Labrador, and Quebec showing the release loca- tion at Campbellton River for Atlantic salmon {Salmo salar) smolts with data storage tags attached. with a precision of ±1°C and an accuracy of 0.15°C. The iB4 data storage tag will store up to 2048 readings but the number of data recorded by each tag depends on the length of time that the tag was at large and the sampling rate of the tag. Our iB4 tags were set to start on the day of tagging and record a temperature every hour — a sam- pling rate that would allow for 85 days of temperature recordings. Tags were calibrated by the manufacturer and checked for accuracy on return. A tag weighs 5.5 g in air and less than 2.6 g in water, is approximately 24 mm by 17 mm by 8 mm, and has a hole at either end for attachment to a fish. Tag programming, data acquisition, and data downloading were achieved by the user with a connecting clamp interfaced to the serial adapter on any standard IBM compatible computer. No correction for drift in the time function was made because of the short period over which the experiment took place (Walker et al., 2000; Reddin et al., 2004). Tags were attached to salmon smolts by using a double-wire (0.32 mm stainless steelj bridal that was passed through the dorsal muscu- lature and that was anchored to the fish by m.eans of a backing plate on the opposite side of the fish. A reward of $10 Canadian was offered for the return of the tag. The smolts were anaesthetized with clove oil, tagged, and placed in holding tanks in the river and given an approximate 10-hour recovery period before release in the evening on the downstream side of the counting fence. All smolts survived the tagging process. From 311 salmon tagged and released, 15 DSTs were recovered with data intact. All tagged fish were recovered during the year of tagging. Data analysis Temperature data recorded by the recovered DSTs were analyzed and compared to oceanographic and freshwa- ter temperature data collected during the summer of 2002. The freshwater data was recorded by a Hugrun temperature recorder (Hugrun, Sidumuli 13, Reykjavik 108, Iceland) placed in the trap at the Campbellton River counting fence, and the sea temperatures came from Hugrun recorders placed just outside the estuary of Campbellton River at 8 m depth and from another at Comfort Cove at 10 m, both reported by Colbourne^'* (Fig. 1). Periods of daylight and darkness were esti- mated from the U.S. Navy Observatory^'^ online photo- period calculator. Latitudinal and longitudinal positions ^* Colbourne, E. 2003. Physical oceanographic conditions on the Newfoundland and Labrador Shelves during 2002. DFO, CSAS Res. Doc. 200.3/020, .57 p. '^ U. S. Naval Observatory, Astronomical Observations Department. 2004. Website: http://aa.usno.navy.mil/ data/docs/RS_OneDay.html [accessed on 7 March 2006.] 418 Fishery Bulletin 104(3) used to calculate sunrise and sunset for individual fish were based on the latitude and longitude at the mouth of Campbellton River. Although it varies by date, day at the latitude of Campbellton River during summer generally occurs from 0500 hours to 2300 hours and night occurs from 2300 to 0500 hours local time. Each individual temperature record from the tags in Green- wich Mean Time (GMT) was then coded to either day or night. Some salmon smolts in Newfoundland return to their home rivers within a couple of months of migrating to sea as precocious postsmolts and as postsmolts cannot travel extensively from the area of the river mouth, so that the river mouth provides a proxy position to deter- mine day length (Downton and Reddin, 2004). Standard weeks were assigned with the Microsoft Excel function WEEKNUM. The analysis of variance technique used to test for temperature differences between fish was the PROC MIXED procedure in SAS (SAS Institute Inc., Gary, NO) The model treated tag as a random rather than a fixed variable. Day temperatures were compared to those at night and any diurnal vertical movements were considered to be reflected by sudden changes in temperatures. Newfoundland coastal waters always exhibit vertical stratification in late spring and early summer. Results Tag and data recovery In total, temperature records were recovered from 15 of the 18 tagged Atlantic salmon smolts for an overall return rate of 5.8% (Table 1). Of the 18 recovered tags, 13 came from herring nets during the same summer that the fish were released. Average days free for those DSTs recovered in herring bait nets was 14 days and all were recovered within Indian Arm into which Camp- bellton River flows. In addition, two were recovered at the counting fence, one was recovered during a swimming survey of the lower portion of the river, one was angled, and another was found by a swimmer a full year after the fish was tagged and released. The great number of tags recovered in herring gear was due to tags becoming entangled in the nets. The lon- gest period of recorded water temperatures while fish were free swimming was 71 days and the shortest was 3 days (Table 1). All tags were recovered during the first summer after their release, except the one found by the swimmer. Water temperature profiles Temperatures were recorded every 60 minutes for all recovered tags. The overall sampling frequency or the number of data points collected per tag ranged from a low of 32 for DST257 to a high of 1681 for DST45 and the second longest was for DST267 with 1373 (Table 1). The most complete records were recorded by DSTs 45 and 267 because the fish to which they were attached spent the least amount of time in freshwater and returned to freshwater after spending a minimum of eight weeks at sea. Both of these fish returned at 0+ sea age, a trait which is observed in about 15% of the salmon returning to Campbellton River in recent years (Downton et al., 2004). The water temperature profiles from the DSTs revealed a complex pattern of daily variations (Figs. 2-4). A number of points of inter- est can be made from the individual DST records. The temperature records from the DSTs when compared to local freshwater and marine temperatures allowed the time of day to be ascribed for most fish when the transi- tion from the river to sea occurred (Fig. 3). The records indicated that 8 out of 13 fish made the transition from freshwater to the sea during daylight, whereas both fish recovered at the counting fence after spending time at sea re-entered freshwater during the night. Because they had the longest period of free swimming activity recorded on the DSTs, tags 45 and 267 were selected for further analysis including comparison for freshwater and sea temperatures (Fig. 3). Within the daily variations, temperature profiles from individual DSTs showed several prominent features: 1 a period of freshwater residence during which time the freshwater and DST temperatures corresponded closely, followed by entry into the sea at which time the temperatures recorded by the DSTs diverged from the recorded freshwater temperatures (Fig. 4, A and B); 2 a period of low variability in temperature fluctua- tions at the beginning of the marine record (with a coefficient of variation of less than or equal to 10%), continuing for at least a couple of weeks (Fig. 4C); 3 after the period of low variability, a series of rapid daily changes in temperature from either warm to cool or, for a few limited cases, from cool to warm (with coefficients of variation of greater than 30%) (Fig. 4, D and E) and during this period, DST tem- peratures sometimes varied by as much as 13°C over a couple of hours; 4 a distinct trend to increasing temperatures as the summer progressed which was due to gradual heat- ing of the surface layer and deepening of the ther- mocline (Fig. 3); and 5 a series of steady temperatures near the end of the ocean period before entry into freshwater (with coef- ficients of variation of less then 12%) (Fig. 4F). Between the periods of low temperature variability at the beginning and end of the various series, there was a period of rapid changes in temperature when the fish were assumed to have dived deeper in the water column, followed by ascents to near surface. The fre- quency of these rapid changes in temperature provides some information on the number of dives per day which numbered between 1 and 3 and may be a proxy for the length of time spent feeding and chasing after prey or fleeing from predators. Reddin et al Diurnal and nocturnal temperatures for Salmo solar 419 Table 1 Release and recovery details for eighteen iB-4 data storage tags ( DSTsi applied to Atlantic bellton River, Newfoundland, 2002. N/A= data were not available. salmon (Salmo salar) smolts at Camp- DST number Number of records Date and time tag was turned on Date and time when tag was recovered Fork length (cml Return date Days free Comments 22 327 5/17/02 8:20 5/17/02 8:20 22.0 spring, 2002 15 Caught in herring bait net in Indian Arm Tag found 25 m below counting fence site 43 N/A 5/19/02 7:59 5/19/02 9:28 24.0 24 May 2002 — No fish was attached to tag 45 1681 5/19/02 8:06 5/19/02 9:36 28.0 28 Jul 2002 71 Caught at the upstream counting fence 55 N/A 5/19/02 8:30 5/19/02 10:19 27.0 03 Jul 2003 — Found in estuary by swimmer 74 N/A 5/21/02 8:46 5/21/02 10:23 23.0 spring, 2002 — Caught in herring bait net in Indian Arm 84 180 5/23/02 10:40 5/23/02 15:32 26.0 spring, 2002 8 Caught in herring bait net in Indian Arm 97 305 5/24/02 8:16 .5/24/02 10:30 24.0 spring, 2002 14 Caught in herring bait net in Indian Arm 98 588 5/24/02 8:21 5/24/02 10:30 27.5 spring, 2002 26 Caught in herring bait net in Indian Arm 103 302 5/24/02 8:36 5/24/02 10:30 26.0 spring, 2002 14 Caught in herring bait net in Indian Arm 107 634 5/24/02 9:09 5/24/02 10:30 26.0 spring, 2002 28 Caught in herring bait net in Indian Arm 129 562 5/26/02 8:21 5/26/02 11:00 25.0 spring, 2002 25 Caught in herring bait net in Indian Arm 195 213 5/30/02 9:08 5/30/02 10:48 26.0 spring, 2002 10 Caught in herring bait net in Indian Arm 238 133 6/1/02 8:12 6/1/02 10:14 25.0 spring, 2002 7 Caught in herring bait net in Indian Arm 245 129 6/2/02 7:35 6/2/02 9:04 25.0 spring, 2002 6 Caught in herring bait net in Indian Arm 247 401 6/2/02 7:39 6/2/02 9:12 24.0 19 Jun 2002 17 Angled upstream 257 32 6/2/02 8:07 6/2/02 9:56 25.0 spring, 2002 3 Caught in herring bait net in Indian Arm 267 1373 6/2/02 8:29 6/2/02 10:49 34.0 30 Jul 2002 59 Recovered by upstream counting fence 293 240 6/4/02 6:59 6/4/02 9:50 26.0 spring, 2002 11 Caught in herring bait net in Indian Arm Thermal ecology A comparison of temperatures for night and day for the two tag recoveries with detailed data indicated differ- ences between daily night and day temperatures for weeks 24 to 29 (11 June-22 July) (Fig. 5). The tagged salmon appeared to be at lower temperatures during the day than at night. Although there were only slight differences between day and night for the cumulative temperatures in some weeks (i.e., week 24 [11-17 June]), for others the differences were quite large (i.e., week 27). These differences were reflected in the overall mean of 12.5°C at night compared to the overall mean of 11.6°C during the day — values which are significantly different from each other it =-6.32, P<0.0001) and which possibly reflect temperatures at different depths. Although there was some variation among fish, this pattern was gener- 420 Fishery Bulletin 104(3) 25 20 15 10 5 0- 25- 20 15 10- 5- 0- 25 20 15 10 5 0 25 20 „ 15 O ^ 10 - 5-1 B ro ■ ^ nij;c^w>Wj^|\^^ 0- (D Q- 25- B 20 £ 15- g 10 5- 0 25 20- 15- 10- 5 0 25 20 15 10 5 0- 25 20- 15- 10- 5 0- DST129 May 30 Jun 3 Jun 7 Jun 1 1 Jun15 Jun 1 DST 22 May 19 May 21 May 23 May 25 May 27 y,/lay 29 May 31 DST 257 DST 97 May 25 May 27 May 29 May 31 DST 247 May 28 Jun 4 June Jun 8 Jun 10 Jun 12 Jun 14 Jun 16 Jun 18 DST 45 DST 238 Jun 2 Jun 3 Jun 4 Jun 5 Jun 6 Jun 7 DST 245 r I r- May 31 Jun 1 Jun 2 Jun 3 Jun 4 Jun 5 Jun 6 Jun 7 Jun 8 DST 98 May 28 Jun 1 Jun 5 Jun 9 Jun 13 Jun 17 DST 103 ^^^-'•A^jL^yJ^^^'A^,.^^ — 1 ' r- May 26 May 28 May 30 Jun 1 Jun 3 Jun 5 Jun 7 DST 293 June Jun 8 Jun 10 Jun 12 Jun 14 . .i DST 267 KM May31 Jun 14 Jun 28 Jul 12 Jul 26 Jun 8 Jun 15 Jun 22 Jun 29 Jul 6 Jul13 Jul 20 Jul 27 DST 107 ''V^Wly^^^^_,w^(^/^,.J.,,^^ May 28 Jun 1 Jun 5 Jun 9 Jun 13 Jun 17 Jun 21 Figure 2 Hourly temperatures and dates recorded by 15 data storage tags attached to Atlantic salmon iSalmo salar) postsmolts at Campbellton River, 2002. ally consistent. The cooler temperatures experienced by the fish during the day indicated that they were deeper in the water column during day than during the night. Diurnal trends in water temperatures also have to be considered because stream temperatures are generally warmer during the day than at night. Warmer tem- peratures during the day are generally the case for sea temperatures as well (^=5.00, P<0.0001) although less Reddin et al Diurnal and nocturnal temperatures for Salmo solar 421 15 10 5^ 0- '-' l^l'Iii'r'' i\"',', I' 1 1 1 1 1 1 1 1 1 r May 28 Jun 4 Junll Jun 18 Jun 25 Jul 2 Jul 9 Jul 16 Jul 23 Jul 30 Figure 3 Hourly temperatures recorded by data storage tags on Atlantic salmon (Salmo salar) postsmolts versus date for tags 45 and 267 applied at Campbellton River, 2002. Also shown are water temperatures recorded at the trap in the Campbellton River counting fence and in the sea near the Campbellton River estuary, 2002. Arrows indicate time of departure from and entry to freshwater. Periods selected for further analysis are shown and labelled alphabetically from A to F and are shown in detail in Figure 4. pronounced at greater depths. Because sea temperatures are lower at night than during the day and because the water temperatures experienced by the fish are warmer at night, it is likely that the fish are actively changing depths. Also, there were times when water temperatures declined or increased by 5°C but the temperature of the fish remained steady; this feature would indicate that the fish was actively seeking specific temperatures and changing position in the water column in order to find these temperatures. The results of the mixed-effects ANOVA with data for weeks 24 to 29 from fish tags 45 and 267 indicated significant differences in temperature for week (Fc, jgg7=251.82, P<0.0001) and night-versus-day effects (Fj 239,3=54.53, P<0.0001). Both smolts spent about 90% of the time in water temperatures ranging from 8° to about 15°C. Overall, mean temperature for day activi- ties for both fish combined was 11.6°C and for night it was 12.5°C. Therefore, for all weeks for both fish, tem- peratures experienced during the day were generally lower than during the night. Whether these differences were due to choice or simply because the fish were in different locations and experiencing local water condi- tions is unknown. The frequency distribution of water temperatures while the fish were in the sea showed a wide range, from below 0° to about 20°C (Fig. 6). This distribution was wider and included temperatures that were both colder and warmer than temperatures available in the estuary off Camp- bellton River at 8 m or at Comfort Cove at 10 m. The warmer temperatures in the estuary than those recorded by the sea recorders are possibly due to the fish being near the surface or shore (or to both factors). The DST temperatures were lower than freshwater temperatures, reflecting the sea temperatures rather than freshwater temperatures. The pattern for each fish presumably re- flects a combination of individual preferences and the availability of specific water temperatures depending on the location of the fish at sea. The temperatures recorded by each tag reflect the time (season) of entry into the sea and the time of recapture, especially for those fish caught in fisheries while the fish were still at sea. Discussion The transition from freshwater to ocean life for Atlantic salmon (whether as smolts or kelts) can have a serious 422 Fishery Bulletin 104(3) p a> Q. E May 28 May 30 Jun 12 Jun 14 Jun 16 Jun 18 28 26 24 22 20- 18 16 14 - 12 10 8 6 4 - 2 0 rv' ^ Jun 26 Jun 28 Jun 30 I^B 88 8B ^B Jul 4 Jul 6 Jul 8 Jul 22 Jul 24 Jul 26 Jul 28 Figure 4 Comparison of day and night temperatures recorded hourly by data storage tags on Atlantic salmon iSalmo salar) postsmolts for selected periods of time. Night is shown in grey and day is shown in white. Solid lines are the DST temperatures; the dashed line represents freshwater temperature and dotted line represents sea temperature at 8 m. The periods A to F are shown on Figure 3 for each tag. consequence for an individual fish, as well as be an important factor controlling year-class strength and abundance at the population level (McCormick and Saunders, 1987, McCormick et al., 1998). The trans- formation process from freshwater to salt water made by smolts is accompanied by changes in metabolic rate (Hoar, 1988) and increases in energy demands, which explain the need for the fish to immediately begin feed- ing. Levings (1994) concluded that of all the variables influencing survival of postsmolt salmon, temperature is particularly important. If they are to survive, indi- viduals must quickly adapt to their new physical envi- ronment and be able to flee predators and seek prey. In spite of its presumed importance, the ecology of salmon smolts as they enter the marine environment is largely unknown. This article provides the first recorded results of temperatures experienced by salmon smolts as they enter the sea and provides insight into how postsmolts adapt to life at sea based on ambient temperature ranges and temperature choices made by the postsmolts. The Reddin et a\ : Diurnal and nocturnal temperatures for Salmo salar 423 8 10 12 16 18 20 16 18 -2 0 2 4 Water temperature (°C) Figure 5 Difference between night and day temperatures recorded hourly by data storage tags and expressed as cumulative counts for weeks 24 to 29 from Campbellton River, 2002. Water temperatures are shown for six periods and for all day and night temperatures. Thin solid line is for daytime temperatures and thick line is for night time temperatures. In the multi-line graphs labelled "day" and "night," the square=week 24, the circle=week 25, triangle = week 26, polygon=week 27, dark pentagon=week 28 and the star=week 29. temperature profiles collected by these data storage tags provide detailed information on the thermal habitat experienced by 15 Atlantic salmon postsmolts for periods ranging from a few days to about two months at sea. Temperatures recorded ranged from below 0° to nearly 20°C, although most of the time was spent in water from 8° to 15°C, probably near the surface. Sigholt and Finstad (1990) and Handeland et al. (2003) reported that lethal sea water temperatures for both wild and farmed salmon smolts adapting to seawa- ter occurred at very low and very high temperatures. At the lower end of the range in temperatures, some mor- talities occurred at sea temperatures of 6-7°C, whereas at the higher end, mortalities occurred at temperatures 424 Fishery Bulletin 104(3) over 14°C, indicating that there may be environmental windows for successful transition of smolts into the sea at locations other than at the extreme edges of the temperature ranges documented by Power et al. (1987). Reddin and Friedland (1993) noted that during colder years with heavy ice flows along the northeast coast of Newfoundland, commercial catches and returns of salmon to rivers were lower, even though freshwater warming was occurring normally and they also noted that these low catch rates and returns may have been due to the mortality of postsmolts early in the marine phase. Our results from DSTs applied to Campbellton River smolts indicated that temperature-related mortal- ities, at least in 2002, were unlikely because the smolts spent little time in water colder than 7°C. Furthermore, because the only colder water present was deeper in the water column, salmon postsmolts were able to avoid it by remaining in the warmer surface waters. In 2002, there was little or no inshore ice at the time of the Campbellton River smolt migration as is sometimes present in other years when mortalities from cold water temperatures could occur. It is interesting to note that temperatures recorded by the DSTs for the first few weeks after the smolts entered the sea were very similar to that experienced by salmon kelts tagged with DSTs also from Campbellton River (Reddin et al., 2004). In addition, the tempera- tures from tagged kelts when compared to temperatures from tagged smolts showed similar periods of stability just after entry of the salmon into the sea, as well as abrupt changes in temperatures consistent with deep diving activities. Although we did not know the precise geographic location of either group of fish, it appears that kelts and postsmolts are found initially in water of similar temperatures and behave in similar ways with respect to temperature. Because neither the postsmolts nor kelts started their deep diving activities for four weeks after sea entry, it may be that they were recover- ing from the transition process or that they may have been near shore where deeper depths are not available (or both these reasons may apply). Resolution of these questions will have to await the results from application of geolocation and depth-sensing tags. Colbourne et al.'", Friedland et al. (2003), Beamish and Bouillon (1993), and Downton and Miller (1998) and many others have examined relationships between environmental variables and the abundance of Pacific and Atlantic salmon and other fish species with a view of providing forecasts of future abundance but without any knowledge of what thermal regimes the fish actu- ally use. Colbourne et al.'^ indicated that the goal of searching for relationships is important in order that 1*^ Colbourne, E. B., E. G. Dawe, D. G. Parsons, E. F. Murphy, W. R. Bowering, E. L. Dalley, J. T. Anderson, J. B. Demp- son, D. Orr, D. E. Stansbury, and G. P. Ennis. 2002. A preliminary review of environmental-stock relationships for some species of marine organisms in NAFO waters of the Northwest Atlantic. NAFO SCR Doc. 02/34, 21 p. Northwest Atlantic Fisheries Organization, P.O. Box 638, Dartmouth, Nova Scotia, Canada B2Y 3Y9. 20 22 24 25 Water temperature (-C) Figure 6 Frequency distributions for all temperatures recorded hourly at sea at 10 m depth (Sea temp at 10 m) and at 8 m depth (Sea temp at 8 m), and for salmon in fresh- water from the DSTs (Salmon in fresh), and for salmon from DSTs while in the sea (Salmon in sea). All sea temperatures from the DSTs were used and the fresh- water and marine temperatures are for the same time periods that the DST-tagged fish were in the sea. influences of the physical ocean environment can be used to provide more accurate forecasts of stock abun- dance and ultimately so that they can be used to create management plans for various fisheries (Bisbal and Mc- Connaha, 1998; Friedland, 1998). These predictions are currently used for both North American and European Atlantic salmon, for which abundance has been fore- casted partly on the basis of environmental information (Anon.'^). Physical conditions in the ocean have been shown to be related to mortality and growth of some other species (Brander, 1995; Dutil et al., 1999; Wata- nabe and Yatsu, 2004). Blackbourn (1993) and Down- ton and Miller (1998) have suggested that freshwater survival rates for some species of Pacific salmon are even related to SSTs experienced by potential spawners while still at sea, shortly before their return to fresh- water. These studies and their importance clearly show the need for studies on the physical oceanography and ecology of fish at sea. For Atlantic salmon, Reddin and Friedland (1993) created a profile of sea temperatures, based on research vessel catch rates, that indicated that Atlantic salmon were commonly found in water at temperatures from 4° to 10°C. Reddin et al. (2004) Reddin et al : Diurnal and nocturnal temperatures for Salmo solar 425 indicated that Atlantic salmon kelts tagged with DSTs were found where temperatures ranged from a low near 0° to over 25°C, although most of their time was spent in water of 5° to 15°C. This finding was similar to that for adults by Sturlaugsson' and Karlsson et al.'" The present study indicates that salmon postsmolts are located mostly near the surface in water temperatures ranging from 8° to about 15"C. Because the tempera- tures experienced by kelt are somewhat warmer then those used in correlation studies of salmon growth and survival, it may be that our results will encourage other researchers to review the temperatures they use in such studies. Prey items in the diet of adult Atlantic salmon are fairly well known revealing that Atlantic salmon are opportunistic feeders, feeding on whatever is abundant in the area (Reddin, 1988; Hislop and Shelton, 1993). Although food resources are still unknown for some regions, the prey species of smolts and postsmolts at sea has recently become an important area for research in the northeast Atlantic (Andreassen et al., 2001; Sal- minen et al., 2001; Rikardsen et al., 2004). Rikardsen et al.'s (2004) extensive analysis of postsmolt diet in eight fjords in Norway revealed extensive feeding on pelagic larval fish species and crustaceans, as well as substantial geographic and annual variation in prey diversity and feeding intensity. Furthermore, Levings (1994) summarized the diet of smolts and postsmolts in the estuarine and near-shore environments from publi- cations by Power and Shooner (1966), Dutil and Coutu (1988), and Hvidsten et al.^* which shows postsmolts feeding on gammarid amphipods and intertidal and land-based insects. Although diet information exists for postsmolts, the feeding mechanisms of postsmolts are unknown because there are few, if any, tracking studies in estuaries and the coastal areas. However, the new data on postsmolts early in their marine life that we have collected using DSTs can help us infer this information and the diet studies, although not specific to Newfoundland waters, indicate that salmon prey on pelagic fish larvae and crustacean species, some of which inhabit the deeper waters of the water column. Thus, it would appear that postsmolts in the present study and the kelts in a previous study (Reddin et al., 2004) were diving from surface waters to greater depths to obtain prey. The temperature profiles from the DSTs, compared to oceanographic data available from other studies (Colbourne'^; Colbourne and Fitzpatrick'-'), in- dicate that these dives may be anywhere from 25 to 50 m in depth. 1^ Karlsson, L., E. Ikonen, H. Westerberg, and J. Sturlaugsson. 1996. Use of data storage tags to study the spawning migration of Baltic salmon (Satmo salar L.) in the Gulf of Bothnia. ICES CM. (council meeting) 1996/M: 9, 15 p. 18 Hvidsten, N. A., B. O. Johnson, and C. D. Levings. 1993. Be- haviour and feeding of emigrating salmon smolts in Trondheimfjord. Res. Rep. no. 164, 17 p. [In Norwegian.] Norwegian Institute for Nature Research, Trondheim, Norway, Tungasletta 2, 7485 Trondheim, Norway. There is a tendency for postsmolts to be caught in the upper part of the water column, as evidenced by trawl- ing for postsmolts at sea (Shelton et al., 1997; Holm et al., 2000; Rikardsen et al., 2004), by acoustic tracking of postmolts in fjords (Holm et al., 2000; Moore et al.-"), and by net catches in surface waters (Dutil and Coutu, 1988; Reddin and Short, 1991; Thorisson and Sturlaugs- son-i). The temperature profiles from our DST-tagged postsmolts indicate that although most of their time is spent near the surface (and nearer the surface at night), salmon postsmolts undergo deeper dives, prob- ably in search of prey. Deeper diving activities have been reported previously for salmon kelts also from Campbellton River by Reddin et al. (2004). The present study demonstrates that postsmolts also show the same type of behavior although the frequency of dives varies somewhat from fish to fish. Holm'-- et al. (2000) noted that salmon postsmolts caught in pelagic trawls in the Norwegian Sea were all caught during surface trawls; none were caught at deeper depths. Westerberg (1982) during coastal tracking studies noted that salmon made dives of short duration to greater depths; his results are similar to the results shown in our study. Wada and Ueno (1999) listed three hypotheses to explain diving behaviour in Pacific salmon, viz. the salmon are making orientation for homing migration, for feeding, and for controlling body temperature. Red- din et al. (2004) suggested a fourth hypothesis: that the salmon are diving to avoid predators. Our conclusion is that the deep diving activities of salmon postsmolts (to as deep as 50 m) recorded by the DSTs, because of their frequent nature, are probably related to prey seeking and feeding. Reddin et al. (2004) further pointed out that there may be an energetic advantage for salmon to seek prey in cooler, deeper waters where prey are more abundant and then to return to warmer surface waters where their food can be digested more rapidly. Ecological information can also be discerned from the temperature patterns provided by the DSTs. Ogura and Ishida (1995), Wada and Ueno (1999), Walker et al. (2000) and Reddin et al. (2004) all noted a period of more stable temperatures experienced at both the beginning and at the end of the temperature time series from the DSTs and concluded that the fish remained at the same IS Colbourne, E. B., and C. Fitzpatrick. 2003. Physical oceanographic conditions in NAFO subareas 2 and 3 on the Newfoundland and Labrador Shelf during 2002. NAFO (North Atlantic Fisheries Organization) SCR Doc. 03/14, 57 p. Northwest Atlantic Fisheries Organization. P.O. Box 638, Dartmouth, Nova Scotia, Canada B2Y 3Y9. 20 Moore, A., I. C. Russell, M. Ives, E. C. E. Potter, and C. P. Waring. 1998. The riverine, estuarine and coastal migratory behaviour of wild Atlantic salmon (Salmo salar L.) smolts. ICES CM 1998/N: 16, 11 p. -1 Thorisson, K. and J. Sturlaugsson. 1995. Postsmolt of ranched Atlantic salmon (Salmo salar L.) in Iceland: IV. Competitors and predators. ICES CM. 1995/M: 12, 9 p. " Holm, M., I. Huse, E. Waatevik, K. B. Doving, J. Aure. 1982. Behaviour of Atlantic salmon smolts during the sea- ward migration. I. Preliminary report on ultrasonic track- ing in a Norwegian fjord system. ICES CM 1982/M:7, 10 p. 426 Fishery Bulletin 104(3) location in the water column for some period of time, perhaps adjusting to the presence of the tag or to life in the sea (or to both). Furthermore. Reddin et al. (2004) noted that this period of temperature stability could be also due in some instances to the fish being near shore in shallow water where large scale vertical movements and their concomitant changes in temperature were not possible. This period of restricted vertical activity seemed to cease as the fish made its return migration towards freshwater. Our study indicates that in freshwater the water temperature recorded at the trap corresponds close- ly to the temperature recorded by the DST on the fish on its departure from the river and at its return. DST temperatures in freshwater follow the diurnal rhythm of warming and cooling of the river. In the sea, there is a period of stable temperatures with a diurnal rhythm that follows the daily warming and cooling. Next, the smolts began a period of fairly rapid descents and ascents in the water column during which temperatures rose and fell much more quickly then during the natural daily cycle. Also, night temperatures were warmer than day temperatures, in contrast to the daily warming and cool- ing cycle of the sea, indicating that smolts were higher in the water column at night then during daylight hours. The differences in temperatures between day and night may reflect avoidance of avian predators as suggested by Reddin et al. (2004) and Montevecchi et al. (2002). This finding is similar to that shown for adult chum salmon by Friedland et al. (2001) but contrasts with that reported by Shelton et al. (1997) who reported that no catches of postsmolts occurred at night during surface trawling in the northeast Atlantic. Because the records of trawling at night are sparse, we recommend that more effort should be put into night trawling in light of the information from our DST-tagged postsmolts. In terms of our objectives, we have shown that Atlan- tic salmon smolts can carry DSTs, and that a sufficient number of tags with useful data on thermal ecology can be successfully recovered. The return rates were low and validated the preliminary use of the less expensive DSTs but were high enough that these experiments could be re- peated elsewhere. The use of more expensive DSTs with geoposition and more environmental sensors will require methods to recover them at sea if sufficient tags are to be available to provide meaningful results. The thermal habitat used by Atlantic salmon is shown; postsmolts are found in water with temperatures ranging from 8° to about 15°C in the spring. Lastly, although we show from the water temperature records that salmon postsmolts are frequently found near the surface, it is also evident that they make frequent deep dives of short duration. We hypothesize that these deep dives may be directly related to feeding or to evasion from predators (or to both), as has been observed for kelts by Reddin et al. (2004). Acknowledgments We acknowledge the staff at the counting fence on Camp- bellton River, who assisted with tag application and data recording. The enthusiasm of Garnet Clarke and Roger Johnson who provided several helpful ideas on tag application is greatly appreciated. The comments provided by R. Poole and two anonymous reviewers are gratefully acknowledged. Literature cited Allan, I. R. H.. and J. A. Ritter. 1977. Salmonid terminology. J. Cons. Int. Explor. Mer 37:293-299. Andreassen, P. M. R., M. B. Martinussen, N. A. Hvidsten, and S. O. Stefansson. 2001. Feeding and prey selection of wild Atlantic salmon post-smolts. J. Fish Biol. 58:1667-1679. Beamish, R. J., and D. R. Bouillon. 1993. Pacific salmon production trends in relation to climate. Can. J. Fish. Aquat. Sci. 50:1002-1016. Bisbal, G. A., and W. E. McConnaha. 1998. Consideration of ocean conditions in the management of salmon. Can. J. Fish. Aquat. Sci. 55:2178-2186. Blackbourn, D. J. 1993. Sea surface temperature and the subsequent fresh- water survival rate of some salmon stocks: a surprising link between the climate of land and sea. In Proceed- ings of the ninth annual Pacific Climate (PACLIM) Workshop, 21-24 April 1992 (K. T. Redmond, and V. L. Tharp, eds.). Tech. Rep. no. 34, p. 23-32. California Department of Water Resources, Interagency Ecological Studies Program, Pacific Grove, CA. Boehlert, G. W. 1997. Application of acoustic and archival tags to assess estuarine, nearshore, and offshore habitat utilization and movement by salmonids. NCAA Technical Memo. NMFS-SEFSC-236, 62 p. Brander, K. M. 1995. The effect of temperature on growth of Atlantic cod {Gadus morhua L.). ICES J. Mar. Sci. 52:1-10. Downton, M. W., and K. A. Miller. 1998. Relationships between Alaskan salmon catch and North Pacific climate on interannual and interdecadal time scales. Can. J. Fish. Aquat. Sci. 50:1002-1016. Dutil, J.-D., and J.-M. Coutu. 1988. Early marine life of Atlantic salmon. Salmo salar, postsmolts in the northern Gulf of St. Lawrence. Fish. Bull. 86:197-212. Dutil, J.-D., M. Castonquay, D. Gilbert, and D. Gascon. 1999. Growth, condition, and environmental relation- ships in Atlantic cod {Gadus morhua) in the northern Gulf of St. Lawrence and implications for management strategies in the Northwest Atlantic. Can. J. Fish. Aquat. Sci. 56:1818-1831. Friedland, K. D. 1998. Ocean climate influences on critical Atlantic salmon [Salmo salar) life history events. Can. J. Fish. Aquat. Sci. 55:119-130. Friedland, K. D., D. G. Reddin, J. R. McMenemy, and K. F. Drinkwater. 2003. Multidecal trends in North American Atlantic salmon (Salmo salar) stocks and climate trends relevant to survival. Can. J. Fish. Aquat. Sci. 60:363-383. Friedland, K. D., R. V. Walker, N. D. Davis, K. W. Myers, G. W. Boehlert, S. Urawa, Y. Ueno. 2001. Open-ocean orientation and return migration routes Reddin et al Diurnal and nocturnal temperatures for Salmo solar 427 of chum salmon based on temperature data from data storage tags. Mar. Ecol. Prog. Ser. 216:235-252. Godo, O. R. and K. Michalsen. 2000. Migratory behaviour of north-east Arctic cod, studied by use of data storage tags. Fish. Res. 48: 127-140. Handeland. S. O., B. T. Bjornsson. A. M. Arnesen. and S. O. Stefansson. 2003. Seawater adaptation and growth of post-smolt Atlantic salmon (Salmo salar) of wild and farmed strains. Aquaculture 220:367-384. Hislop, J. R. G., and R. G. J. Shelton 1993. Marine predators and prey of Atlantic salmon {Salmo salar L.). In Salmon in the sea and new enhance- ment strategies (D. Mills, ed.), p. 104-118. Fishing News Books, Oxford, UK. Hoar, W. S. 1988. The physiology of smolting salmonids. In Fish physiology, vol. XIB (W. S. Hoar, and D. J. Randall, eds.), p. 275-343. Academic Press, New York, NY. Holm, M., J. C. Hoist, and L. P. Hansen. 2000. Spatial and temporal distribution of post-smolts of Atlantic salmon. ICES J. Mar. Sci. 57:955-964. Jacobsen, J. A. 2000. Aspects of the marine ecology of Atlantic salmon (Salmo salar L). Ph.D. diss., 51 p. Univ. Bergen. Bergen, Norway. LaCroix, G. L., and P. McCurdy. 1996. Migratory behaviour of post-smolt Atlantic salmon during initial stages of seaward migration. J. Fish Biol. 49:1086-1101. Levings, C. D. 1994. Feeding behaviour of juvenile salmon and sig- nificance of habitat during estuary and early sea phase. Nordic J. Freshw. Res. 69:7-16. McCormick, S. D.. and R. L. Saunders. 1987. Preparatory physiological adaptions for marine life in salmonids: osmoregulation, growth, and metabolism. Am. Fish. Soc. Symp. 1:211-229. McCormick, S. D., L. P. Hansen, T. P. Quinn, and R. L. Saunders. 1998. Movement, migration and smolting of Atlantic salmon (Sa/n;o sa/ar). Can. J. Fish. Aquat. Sci. (suppl. ll 55:77-92. Mills, D. H. 1989. Ecology and management of Atlantic salmon, 351 p. Chapman and Hall, London, UK. Montevecchi, W. A., D. K. Cairns, and R. A. Myers. 2002. Predation on marine-phase Atlantic salmon (Salmo salar) by gannets (Morus bassanus) in the Northwest Atlantic. Can. J. Fish. Aquat. Sci. 59:602-612. Ogura, M. and Y. Ishida. 1995. Homing behavior and vertical movements of four species of Pacific salmon (Oncorhynchus spp.) in the central Bering Sea. Can. J. Fish. Aquat. Sci. 52:532- 540. Power, G., and G. Shooner. 1966. Juvenile salmon in the estuary and lower Nabisipi River and some results of tagging. J. Fish. Res. Board Can. 23:947-961. Power, G., M. V. Power, R. Dumas, and A. Gordon. 1987. Marine migrations of Atlantic salmon from rivers of Ungava Bay, Quebec. In American Fisheries Soci- ety symposium on common strategies in anadromous/ catadromous fishes 1:262-275. Am. Fish. Soc, Bethesda, MD. Reddin, D. G. 1988. Ocean life of Atlantic salmon (Salmo salar L.) in the Northwest Atlantic. Chapter 26, in Atlantic salmon: planning for the future (D. H. Mills, and D. J. Piggins, eds.), p. 483-511. Proceedings of the third international Atlantic salmon symposium, Biarritz, France, 21-23 October 1986. Croom Helm, London. Reddin, D. G., and K. D. Friedland. 1993. Marine environmental factors influencing the move- ment and survival of Atlantic salmon. In Salmon in the sea and new enhancement strategies (D. Mills, ed.), p. 79-103. Fishing News Books, Oxford, UK. Reddin, D. G., K. D. Friedland, P. Downton, J. B. Dempson, and C. C. Mullins. 2004. Thermal habitat experienced by Atlantic salmon kelts (Salmo salar L.) in coastal Newfoundland waters. Fish. Oceanogr. 13:24-35. Reddin, D. G., J. Helbig, A. Thomas, B. G. Whitehouse, and K. D. Friedland. 2000. Survival of Atlantic salmon (Salmo salar L.) and its relation to marine climate. In Managing wild Atlantic salmon: new challenges — new techniques; proceedings of the fifth international Atlantic salmon symposium held at Galway, Ireland, 1997 (F. G. Whoriskey, and K. E. Whelan, eds.), pp. 24-49. The Atlantic Salmon Federation, St. Andrews, News Brunswick, Canada. Reddin, D. G., and P. B. Short. 1991. Postsmolt Atlantic salmon (Salmo salar) in the Labrador Sea. Can. J. Fish. Aquat. Sci. 48:2-6. Rikardsen, A. H., M. Haugland, P. A. Bjorns, B. Finstad, R. Knudsen, J. B. Dempson, J. C. Hoist, N. A. Hvidsten, and M. Holm. 2004. Geographical differences in marine feeding of Atlantic salmon post-smolts in Norwegian fjords. J. Fish Biol. 64:1655-1679. Salminen, M., E. Erkamo, and J. Salmi. 2001. Diet of post-smolt and one-sea-winter Atlantic salmon in the Bothnian Sea, northern Baltic. J. Fish Biol. 58:16-35. Shelton, R. G. J., J. C. Hoist, W. R. Turrell, J. C. MacLean, I. S. McLaren, and N. T. Nicoll. 1997. Records of post-smolt Atlantic salmon, Salmo salar L., in the Faroe-Shetland Channel in June 1996. Fish. Res. 31:159-162. Sigholt, T., and B. Finstad. 1990. Effect of low temperature on seawater tolerance in Atlantic salmon (Salmo salar L.) smolts. Aquaculture 84:167-172. Walker, R. V., K. W. Myers, N. D. Davis, K. Y. Aydin, K. D. Fried- land, H. R. Carlson, G. W. Boehlert, S. Urawa, Y. Ueno, and G. Anma. 2000. Diurnal variation in thermal environment expe- rienced by salmonids in the North Pacific as indicated by data storage tags. Fish. Oceanogr. 9:171-186. Watanabe, C, and A. Yatsu. 2004. Effects of density-dependence and sea surface tem- perature on interannual variation in length-at-age of chub mackerel (Scomber japonicus) in the Kuroshio-Oyashio area during 1970-1997. Fish. Bull. 102:196-206. Westerberg, H. 1982. Ultrasonic tracking of Atlantic salmon (Salmo salar L.) - II. Swimming depth and temperature stratification. Drottingholm Report 60:102-120. 428 Abstract — Priors are existing infor- mation or beliefs that are needed in Bayesian analysis. Informative priors are important in obtaining the Bayesian posterior distribu- tions for estimated parameters in stock assessment. In the case of the steepness parameter l/i). the need for an informative prior is particularly important because it determines the stock-recruitment relationships in the model. However, specifications of the priors for the h parameter are often subjective. We used a simple popula- tion model to derive h priors based on life history considerations. The model was based on the evolutionary principle that persistence of any spe- cies, given its life history (i.e., natural mortality rate) and its exposure to recruitment variability, requires a minimum recruitment compensation that enables the species to rebound consistently from low critical abun- dances (iV_,). Using the model, we derived the prior probability distri- butions of the h parameter for fish species that have a range of natural mortality, recruitment variabilities, and iV, values. A prior for steepness in stock-recruitment relationships, based on an evolutionary persistence principle XiHe^ Marc Mangel^ Alec MacCall' ' Santa Cruz Laboratory Southwest Fisheries Science Center National Marine Fisheries Service Santa Cruz, California 95060 E-mail address (for X He) xi he g'noaa gov ^ Center for Stock Assessment Research Department of Applied Mathematics and Statistics University of California Santa Cruz, California 95064 Manuscript submitted 20 July 2004 to the Scientific Editor's Office. Manuscript approved for publication 30 September 2005 by the Scientific Editor. Fish. Bull. 104:428-433 (2006). Success in parameter estimation for stock assessment models often requires sufficient data and correct specification of prior distributions for the estimated parameters (Punt and Hilborn, 1997; Needle, 2002). This is especially true for successfully esti- mating stock-recruitment relation- ships with stock assessment models because there often are not sufficient data and there is considerable variabil- ity in stock-recruitment relationships. Even in the cases where there are suf- ficient data, these data can be very noisy and may not show a clear pat- tern for fitting stock-recruitment rela- tionships (Hilborn and Walters, 1992; Williams and Shertzer, 2003; Munch et al., 2005). Yet, stock-recruitment relationships are important in making fisheries management decisions, espe- cially for over-fished stocks such as widow rockfish iSebastes entome- las), bocaccio iSebastes paucisplnis), and darkblotched rockfish (Sebastes crameri) along the west coast of the United States (He et al.i; MacCall-; Punt, 2003; Rogers^). Stocks with low values of steepness have low recruit- ment compensation; therefore stocks will take a long time to rebuild from over-fished status to desired manage- ment levels even though total allow- able catches for these stocks are kept small (Mace and Doonan^). In this study, we used a simple population model to derive prior dis- tributions (referred to hereafter as "priors") for the steepness parame- ter (hereafter denoted h) (Mace and Doonan^). Prior distributions are probability distributions that repre- sent existing information about pa- rameters. The model was based on the principle that persistence of any fish species, given its life history and its exposure to recruitment variability, ' He, X., S. V. Ralston, A. D. MacCall. D. E. Pearson, and E.J. Dick. 2003. Status of the widow rockfish resource in 2003. Vol. 1: Status of the Pacific coast ground- fish fishery stock assessment and fishery evaluation, 138 p. Pacific Fishery Man- agement Council, 7700 NE Ambassador Place, Portland, OR. - MacCall, A. D. 2003. Status of bocac- cio off California in 2003. Vol. 1: Status of the Pacific coast groundfish fishery stock assessment and fishery evalua- tion, 56 p. Pacific Fishery Management Council, 7700 NE Ambassador Place, Portland, OR. '^ Rogers, J. B. 2003. Darkblotched rock- fish [Sebastes crameri) 2003 stock status and rebuilding analysis. Vol. 1; Status of the Pacific coast groundfish fishery stock assessment and fishery evaluation, 56 p. Pacific Fishery Management Council, 7700 NE Ambassador Place, Portland, OR. ^ Mace, P. M., and I. J. Doonan. 1988. A generalized bioeconomic simulation model for fish dynamics, 47 p. New Zealand Fishery Assessment Research Document 88/4. Fisheries Research Center, P.O. Box 297, Wellington, New Zealand. He et al A prior for steepness in stock-assessment relationships, based on an evolutionary persistence principle 429 would require a minimum recruitment compensation to allow the species to rebound from low abundances. We argue that distribution of the h parameter for any species could be determined from its life history and recruitment variability. Using the model, we derived the prior distributions of h for fish species that have a range of natural mortality, recruitment variabilities, and low critical abundance (A'^,) values. Methods In calculating steepness priors, we used a simple popu- lation model with a Beverton-Holt stock recruitment relationship: N„^ = N„ .e-^+- A^„ -4) a + PN,,_ (1) where A^,,, = population size at year /; M = natural mortality; a and p = recruitment parameters; and Rill = the logarithm recruitment residual at year t that follows a normal distribution of N[0, a-), where a is recruitment variability (Hil- born and Walters, 1992). This model produces recruitments that are log-normally distributed, and a correction factor of " is applied to R,,,. The correction factor is included because it is com- monly used in stock assessments. The Beverton-Holt stock recruitment relationship can be reparametrized as (Mace and Doonan^) where T = a time far into the future; n = a specific value of population size; and N^. = a critically low level of abundance, below which the population would have very high probability of extinction. At ^ = T,p{n.T\h,T\ = 1 if n>N^. and p(n,T\h,T) = 0 oth- erwise. In addition, p(n,t\h,T) satisfies the boundary condition that p(N^J\h.T) = 0 for all t. For times previous to T, p(n,t\h,T) satisfies the sto- chastic iteration equation (Mangel and Clark, 1988; Clark and Mangel, 2000) '-,[i p(n,t\h,T) = «e^+- a(h) + P(h)n e''',t + l\h,T (7) where £^ denotes the expectation over the stochastic processes associated with recruitment. We have indi- cated that the Beverton-Holt parameters depend upon steepness (as Eqs. 2-5 show, they also depend upon mortality M, which we hold to be a fixed value). Because the recruitment uncertainty is normally dis- tributed and Equation 7 cannot be evaluated analyti- cally, we used a discrete distribution for i?, (=7?|,|-^) into K ( = 61) uniformly spaced values (r^) between -3a and 3 a, so that Pr{R. expl — ^ ^exp .2 \ (8) 2(j' and N^ 1-^ Ro 4h 5h-l AhRi, where N^ = virgin abundance at equilibrium; Rq = virgin recruitment; and h = recruitment steepness. At equilibrium. (2) (3) and N, = N,e-^+R, Ro = DNo, (4) (5) where D - death rate and is equal to 1 - e^'^. We now introduce the persistence criterion. For a given h, the persistence criterion is defined as p{nJ\h,T) = PT{N,^,>N^. forall*. 1 .■ / .t: • : 1 .9 0004 - 1 ; ' CO J3 0 . : 1 O. a) > ™ 0 002 - .' ■ ' w cc ; ; 1 Sigma r= 0,6 ' ; / Sigma r=0 8 .' y Sigma r = 1 0 00 - -' Stgma r= 1 2 0 20 0 30 0 40 0 50 0 60 0 70 0 80 0 90 1 00 /? value Figure 1 Relative probability distributions of the steepness (/;) prior for four different 0 values with the parameter setting M = 0.15, N^ = 0.1N„. and T = 500 years. ability (a) are shown in Figure 1. In general, the prior probability of h increased rapidly at low h values, and then remained at constant values at high h values. As expected, when recruitment vari- ability (a) increased, h values were higher to com- pensate for higher recruitment variability (Fig. 1). Table 1 shows parameter values of logistic curves fitted to derived /; prior distributions as natural mortality (M) ranged from 0.05 to 0.7 and recruit- ment variability (o) ranged from 0.2 to 1.6. These fitted curves would be convenient to use in stock assessment models for given natural mortality and recruitment variability, and values can be interpo- lated for intermediate cases. Derived h prior curves for four A^^ values were plotted (Fig. 2). The figure shows patterns similar to those for recruitment variability, namely that as A^, values increase, /; values are higher to increase recruitment potential for the population. Derived /( prior curves for four natural mortality (M) values (M=0.1, 0.15, 0.3, and 0.5, respectively) are shown in Figure 3. As M values increase, h values also increase. 0012 - ^■^4.v^e u %v.vxrj k zwj.vsfS2x tuvw^r W4 v=.^cv,4.-tf =c cu vwJU4 a: xckt^v- i 5 0 008 - Ii prob 0) > (fl 0 004 - 0.23 (Fig. 2), in- dicating that the population would be sustainable with very low h values. Other methods for deriving h priors include using expert opinion, borrowing values from other taxo- nomically or ecologically related species, and using regional meta-analysis (Myers, 1998, 2001; Chen and Holtby, 2002; Dorn, 2002; Millar, 2002; Myers et al., 2002). For example, Myers et al. (2002) used an empirical Bayesian approach to derive prior distri- butions for steepness for nine species. As compared to our method, their method, based on a combined method of taxonomic and ecological criteria, still re- quires the collection of biological data from related species, as well as expert opinions on the life history of each species, which are somewhat "subjective" values. Dorn (2002) used Bayesian meta-analysis to derive stock-recruitment relationships for a group of west coast groundfish species. He used the results from the previous stock assessments for widow rockfish and other west-coast groundfish species and indicated that the prior distribution for steepness for widow rockfish could have a median value around 0.72 and a prob- ability of 0.0033 for /jsO.225. This result could indi- cate that our method is more conservative because our method indicates much lower median value for widow rockfish-like species. Minte-Vera et al. (2005) also sug- gested that the priors derived by Dorn (2002) might not be appropriate for stock assessments that include the same data that were used in the meta-analysis to derive priors. We believe that the method used in our study provides a scientific way of estimating the prior for steepness and avoids the pitfall of imposing a preconception of what the true distribution is thought to be. For example, it seems reasonable that very high values of h will not be likely because of tradeoffs between individual survival 432 Fishery Bulletin 104(3) Table 1 Parameter values of logistic curves fitted to derived h prior distributions for different natural mortality (M) and recruitment | variability (a). Para meters are listed in the order of 9, e,, and e, (see Eq. 10). For some combinations of natural mortality and | recruitment variability, distributions were uniform or could not be fitted. Natura 1 mortality T1 -1 i Kecruitmer variability 0.05 0.1 0.15 0.2 0.25 0.3 0.4 0.5 0.6 0.7 0.2 Uniform distribution Uniform distribution Uniform distribution Uniform distribution Uniform distribution 1.0000 669.60 0.9289 1.0000 443.00 0.8046 1.0000 306.13 0.6662 1.0000 282.66 0.5015 1.0000 278.16 0.3724 0.4 Uniform distribution 0.9994 1.0000 1.0000 1.0000 1.0000 0.9999 0.9999 0.9999 0.9997 139.90 248.40 223.60 196.43 199.3 182.22 157.82 148.63 145.48 0.9800 0.7973 0.5585 0.3739 0.2146 0.08619 0.04454 0.02264 0.01154 0.6 1.0000 1.0000 1.0000 0.9998 0.9995 0.9992 0.9988 0.9985 Uniform distribution 149.70 133.60 116.74 109.86 105.32 89.55 78.79 69.29 61.48 0.7552 0.3895 0.1791 0.08069 0.03805 0.01543 0.007602 0.005074 0.003861 0.8 1.0000 1.0000 1.0000 0.9997 0.9992 0.9990 0.9982 0.9972 0.9959 0.9943 83.00 93.13 84.25 76.29 68.73 65.53 48.84 40.51 34.03 29.12 0.9259 0.4317 0.1376 0.04540 0.02022 0.01197 0.005796 0.003745 0.003041 0.002722 1.0 1.0000 1.0000 0.9995 0.9987 0.9980 0.9970 0.9944 0.9902 0.9830 0.9709 75.30 64.42 57.52 48.77 41.44 35.57 27.15 21.58 17.71 14.97 0.9526 0.1929 0.04464 0.1700 0.009240 0.006199 0.003906 0.003076 0.002680 0.002402 1.2 1.0000 0.9995 0.9983 0.9967 0.9946 0.9915 0.9801 0.9543 0.9011 0.8085 69.36 48.16 38.71 30.90 25.38 21.26 15.75 12.64 11.01 10.34 0.5920 0.07182 0.01844 0.009036 0.005819 0.004420 0.003134 0.002243 0.001320 0.0005882 1.4 1.0000 0.9983 0.9956 0.9913 0.9835 0.9697 0.9025 0.7428 0.4973 0.2428 47.79 35.06 25.81 19.87 15.94 13.39 10.74 10.25 10.88 12.62 0.2024 0.02504 0.009163 0.005647 0.004010 0.003087 0.001373 0.0003183 0.00003683 0.0000016 1.6 0.9992 0.9955 0.9883 0.9716 0.9338 0.8585 0.5463 0.2027 0.02790 Not fitted 39.09 25.03 17.22 13.22 11.08 10.25 10.75 12.90 16.85 0.03001 0.008900 0.004923 0.003340 0.001971 0.0008686 0.00005315 0.00000079 0.00000001 and reproduction, but that tradeoff should be shown, not imposed. The methods used in the present study can be extended to address this tradeoff (i.e.. Mangel and Clark, 1988; Clark and Mangel, 2000), but that is a subject for another study. However, for many over- fished stocks that need stock-recruitment relationships for rebuilding analysis, this method is especially ap- propriate because it indicates the ability of the popula- tion to recover from low abundance. For other depleted stocks, use of prior probability distributions based on this method can provide useful information on how quickly the populations could be rebuilt under various management policies. Acknowledgments We thank Anand Patil of the University of California Santa Cruz for help in Matlab programming. Literature cited Chen, D. G., and L. B. Holtby. 2002. A regional metal-model for stock-recruitment analy- sis using an empirical Bayesian approach. Can. J. Fish. Aquat. Sci. 59:1503-1514. Clark, C. W., and M. Mangel. 2000. Dynamic state variable models in ecology: meth- ods and applications, 289 p. Oxford Univ. Press, New York, NY. Dorn. M. W. 2002. Advice on west coast rockfish harvest rate from Bayesian meta-analysis of stock-recruitment relat- ionships. No. Am. J. Fish. Manag. 22:280-300. Hakoyama, H., Y. Iwasa, and J. Nakanishi. 2000. Comparing risk factors for population extinction. J. Theor. Biol. 204:327-336. Hilborn, R.. and C. J. Walters. 1992. Quantitative fisheries stock assessment: choice, dynamics and uncertainty, 570 p. Chapman and Hall, New York, NY. He et al A prior for steepness in stock assessment relationships, based on an evolutionary persistence principle 433 MacCall, A, D. 2002. Use of known-biomass production models to deter- mine productivity of west coast groundfish stocks. N. Am. J. Fish. Manag. 22:272-279. Mangel, M., and C. W. Clark. 1988. Dynamic modeling in behavioral ecology. 308 p. Princeton Univ. Press. Princeton, NJ. Millar. R. B. 2002. Reference priors for Bayesian fisheries models. Can. J. Fish. Aquat. Sci. 59:1492-1502. Minte-Vera, C. V., T. A. Branch, I. J. Stewart, and M. W. Dorn. 2005. Practical application for meta-analysis results: avoiding the double use of data. Can. J. Fish. Aquat. Sci. 62:925-929. Munch, S. B., A. Kottas, and M. Mangel. 2005. Bayesian non-parametric analysis of stock recruitment relationships. Can. J. Fish. Aquat. Sci. 62:1808-1821. Myers, R. A. 1998. Why do environment-recruitment correlations work? Rev. Fish Biol. Fish. 8:285-305. 2001. Stock and recruitment: generalizations about maximum reproductive rate, density dependence, and variability using meta-analytic approaches. ICES J. Mar. Sci. 58:937-951 Myers, R. A., N. J. Barrowman, R. Hilborn, and D. G. Kehler. 2002. Inferring Bayesian priors with limited direct data: Applications to risk analysis. N. Am. J. Fish. Manag. 22:351-364. Needle G. L. 2002. Recruitment models: diagnosis and prognosis. Rev. Fish Biol. Fish. 11:95-2002. Punt, A. E. 2003. Evaluating the efficacy of managing West Coast groundfish resources through simulation. Fish. Bull. 101:860-873. Punt, A. E., and R. Hilborn. 1997. Fisheries stock assessment and decision analy- sis: the Bayesian approach. Rev. Fish Biol. Fish. 7:35-63. Williams, E. H., and K. W. Shertzer. 2003. Implications of life-history invariants for biological reference points used in fishery management. Can. J. Fish. Aquat. Sci. 60:710-720. 434 Abstract — To estimate postrelease survival of white marlin (Tetraptu- rus albidus) caught incidentally in regular commercial pelagic longline fishing operations targeting sword- fish and tunas, short-duration pop- up satellite archival tags (PSATs) were deployed on captured animals for periods of 5-43 days. Twenty (71.4%) of 28 tags transmitted data at the preprogrammed time, includ- ing one tag that separated from the fish shortly after release and was omitted from subsequent analyses. Transmitted data from 17 of 19 tags were consistent with survival of those animals for the duration of the tag deployment. Postrelease sur- vival estimates ranged from 63.0% (assuming all nontransmitting tags were evidence of mortality) to 89.5% (excluding nontransmitting tags from the analysis). These results indi- cate that white marlin can survive the trauma resulting from interac- tion with pelagic longline gear, and indicate that current domestic and International management measures requiring the release of live white marlin from this fishery will reduce fishing mortality on the Atlantic-wide stock. Survival of white marlin (Tetrapturus albidus) released from commercial pelagic longline gear in the western North Atlantic* David W. Kerstetter John E. Graves Virginia Institute of Marine Science College of William and Mary Route 1208 Create Road Gloucester Point, Virginia 23062 Present address ((or D, W. Kerstetter): Cooperative Institute for Marine and Atmospheric Studies Rosenstiel School for Marine and Atmospheric Science University of Miami 4600 Rickenbacker Causeway Miami, Florida 33149 E-mail address (for D W Kerstetter) dkerstetteriSrsmas miami edu Manuscript submitted 7 March 2005 to the Scientific Editor's Office. Manuscript approved for publication 6 Ocotber 2005 by the Scientific Editor. Fish. 104:434-444(2006). White marlin (Tetrapturus albidus Poey 1860) is an istiophorid billfish species widely distributed in tropi- cal and temperate waters through- out the Atlantic Ocean, including the Caribbean Sea. There is substantial international concern regarding the population levels of this species. The standing committee for research and statistics (SCRS) of the International Commission for the Conservation of Atlantic Tunas (ICCAT) last assessed the Atlantic -wide stock of white marlin in 2002 and in its continuity-case assessment the committee indicated a total biomass of approximately 12% of that necessary to produce maximum sustainable yield. It was also esti- mated that the current international fishing mortality level for this species is equivalent to more than eight times the replacement yield, contributing to further decline of the overfished stock (ICCAT, 2005). Both recreational and commercial fisheries contribute to the mortality of white marlin. A directed recre- ational fishery exists throughout the tropical and temperate Atlantic (with considerable effort off the coasts of Brazil and Venezuela), as well as off the U.S. mid- Atlantic coast, and there is a growing trend towards catch- and-release practices in all directed recreational billfish fisheries. In con- trast to the catches by this directed recreational effort, white marlin are an infrequent bycatch or a retained incidental catch of the international pelagic longline fishery, which targets tunas (Thiinnus spp.) and swordfish iXiphias gladius). Although white marlin catches in the pelagic long- line fishery are relatively rare, the fishery accounts for the majority of the total fishing mortality on this species simply because of the sheer magnitude of pelagic longline effort exerted throughout the Atlantic (IC- CAT, 2005). Both domestic and international management measures are currently in effect for white marlin. The U.S. recreational fishery is managed with a 66" lower jaw-fork length federal minimum size and a binding ICCAT recommendation that limits the an- nual U.S. recreational landings to a total of 250 blue marlin (Makaira ni- gricans) and white marlin combined (ICCAT, 2000). U.S. commercial fish- ermen have been prohibited from landing or possessing white mar- lin since the implementation of the National Marine Fisheries Service (NMFS) Fishery Management Plan for Atlantic Billfish (NMFS, 1988). ICCAT has responded twice to the decreasing biomass of white marlin and blue marlin by adopting binding Contribution 2695 from the Virginia Institute of Marine Science, College of William and Mary, Gloucester Point, VA 23062. Kerstetter and Graves: Survival of Tetropturus albidus after release from longline gear 435 recommendations requiring reductions in commercial landings by both pelagic longline and purse seine gears (ICCAT, 2000, 2001a). However, these reductions in landings by themselves may ultimately be insufficient to rebuild these two marlin stocks. Goodyear (2G02a) found that a reduction of 60% would be necessary to halt the decline of blue marlin, a species which is more abundant, larger, and presumably more robust to the trauma associated with commercial capture (Kerstet- ter et al., 2003). Given that white marlin are smaller animals, and that the stock is more depleted than that of blue marlin, even more drastic measures are likely necessary to achieve the same management goal for this species. Because the pelagic longline fishery accounts for the majority of white marlin mortality, understanding the nature of billfish interactions with this gear is criti- cal to developing effective strategies to reduce fishing mortality. Jackson and Farber (1998) reported that 56% of white marlin caught in the Venezuelan longline fishery between 1987 and 1995 were alive at the time of haulback. Data from the U.S. observer program and mandatory pelagic longline logbook records indicate that 71% of white marlin were released alive from U.S. commercial pelagic longline gear between 1996 and 1998 (Cramer'). ICCAT has long been encouraging the release of live white marlin through both binding and nonbinding resolutions (ICCAT, 1995, 1996). More re- cently, the commission has approved binding recommen- dations that require the release of all live white marlin caught by purse seine and pelagic longline vessels (IC- CAT, 1997, 2001b). However, those animals released alive must have a reasonable probability of survival for such management measures to be ultimately effective. The assessment of postrelease survival presents spe- cial problems for large pelagic fishes, which are rarely capable of being held in captivity (de Sylva et al., 2000). In general, recovery rates of billfish tagged with con- ventional streamer tags by commercial and recreational fishermen have been quite low (0.4-1.83%: Prince et al., 2003; Ortiz et al., 2003). Although this observation is consistent with high postrelease mortality, low recovery rates could also result from tag shedding and from tags that fail to transmit data (Bayley and Prince, 1994; Jones and Prince, 1998). The results of acoustic track- ing studies of various billfish species (e.g., striped mar- lin [Tetrapturus audax]: Brill et al., 1993; blue marlin: Block et al., 1992; and black marlin [Makaira indica]: Pepperell and Davis, 1999) captured on recreational gear indicate that postrelease survival over periods of a few hours to a few days is relatively high, although mortalities have been observed in short-term tracking studies. Recently, pop-up satellite archival tag (PSAT) technology has proven especially useful to study postre- ' Cramer, J. 2000. Species reported caught in the U.S. com- mercial pelagic longline and gillnet fisheries from 1996-1998. NMFS Sustainable Fisheries Division publication, SFD- 99/00-78:1-33. NOAA/NMFS Southeast Fisheries Science Center, SFD, 75 Virginia Beach Dr., Miami, FL 33149. lease survival in several larger istiophorid species, in- cluding blue marlin in the Atlantic (Graves et al., 2002; Kerstetter et al., 2003) and striped marlin in the Pacific (Domeier et al., 2003). Only recently have PSATs been attached to smaller (<40 kg) istiophorid billfishes. Horo- dysky and Graves (2005) used PSATs to evaluate the postrelease survival of white marlin from recreation- al (rod-and-reel) fishing gear and demonstrated that smaller billfish (2I6 kg estimated weight) can carry PSATs. Their work also suggested high postrelease survival rates in the recreational fishery, especially for fish caught on circle hooks. However, pelagic longline gear presents a different suite of stressors during cap- ture of an animal than does recreational gear. These differences, including long "soak times" (the length of time in each deployment of the gear that the longline is fishing), may also affect postrelease survival rates. In our study, we applied PSAT technology to estimate the short-term mortality of white marlin released alive after capture on pelagic longline gear. Materials and methods Fishing operations White marlin tagging took place off the east coast of Florida (FL), the southwest edge of Georges Bank (GB), the Yucatan Channel (YC), the Windward Passage (WP), and the Mid-Atlantic Bight (MA). These locations are all waters traditionally fished by the U.S. pelagic longline fleet. All tagging operations occurred opportunistically aboard the commercial pelagic longline fishing vessel FV Carol Ann (54' length-over-all) between June 2002 and August 2004. This vessel is typical in size and is equipped for targeting swordfish, mixed swordfish, and tuna within the U.S. coastal pelagic longline fishery. Hook types and sizes were also typical for the fishery and included 7/0 and 9/0 offset J-style hooks (ca. 15° offset; Eagle Claw model no. 9016 or Mustad model no. 7698), 16/0 non-offset circle hooks (Mustad models no. 39660 or no. 39666), and 18/0 non-offset circle hooks (Lindgren-Pitman, Inc., Pompano Beach, FL). Adjusted seasonally, individual leader lengths were 7.5 fathoms (ca. 13.7 m) in the fall northern fishery targeting tuna and 15 fathoms (ca. 27.4 m) in the spring southern fish- ery for swordfish; this adjustment is standard practice within the fleet (O'Neill-). Individual leader lengths comprised a two-fathom "tail" separated from the rest of the leader by a 28-g leaded swivel — a configuration commonly used in this fishery to reduce tangles with other leaders or the mainline. Varying the length of the lines ("buoy drops") connecting the mainline with the small buoy floats on the surface also allows the gear to fish at different depths. Many captains will use two buoy drop lengths in the beginning of a trip to ascertain 2 O'Neill, G. 2003. Personal commun. Carol Ann Sword Corporation. 629 NE 3rd Street, Dania Beach, FL 33004. 436 Fishery Bulletin 104(3) the most productive gear configuration. For our study, two buoy drop lengths were used in each set, and these drop lengths were alternated after every 30 hooks: usu- ally 5- and 2.5-fathom (ca. 9.1 and 4.5 m, respectively) lengths in the fall and 10- and 12-fathom (ca. 18.3 and 21.9 m, respectively) lengths in the spring. Electronic hook-timers (Lindgren-Pitman, Inc.; Pompano Beach, FL) were also used during many of the sets to record the time at which an animal was hooked. Bait was usually frozen squid illlex sp.), but occasionally included frozen Atlantic mackerel {Scomber scombrus) or a haphazard mixture of the two. This project consisted of both a pilot and a main study. The pilot study occurred off the east coast of Florida during June 2002 and included deployments of five PTT-100 tags (Microwave Telemetry, Inc.; Columbia, MD) and one PAT (Wildlife Computers; Redmond, WA) tag. The main study was conducted between August 2002 and August 2004 and for this study only PTT-100 HR model tags were used. Tag models The physical characteristics of all PSAT tag models used in this study were similar and included a micro- processor, a transmitter, and various environmen- tal sensors, all contained within a resin-filled carbon fiber tube. The tag is made positively buoyant by a spherical glass-bead-embedded float at the base of the antenna. It measures approximately 38 cm in length by 4 cm diameter (including antenna) and weighs between 65 and 75 g (air weight). Tags were rigged with approximately 16 cm of 400-pound test Momoi" brand (Momoi Fishing Co.; Ako City, Japan) mono- filament line attached to a large hydroscopic nylon intramuscular tag head according to the method of Graves et al. (2002). The earlier model PTT-100 tags were identical to those used by Graves et al. (2002) and Kerstetter et al. (2003) and recorded one temperature data point for every two-hour period during their five- day (A! = 3l or 30-day (n=2) deployments, as well as a pre- and postdeployment inclinometer value. The PAT tag recorded environmental data every minute during its 43-day deployment (programmed to disengage from the fish on 30 July 2002) but transmitted data as sum- mary histograms rather than discrete data points. The PAT tag possessed emergency release software as well as a mechanical device (RD-1500; Wildlife Computers, Redmond, WA) for an early emergency release before reaching a depth at which it would be crushed by ambi- ent water pressure (crush depth). The Microwave Telemetry, Inc. model PTT-100 HR satellite tag was used for the main study and consti- tuted the majority of the PSAT deployments (;!=22). This tag has similar physical attributes to those of the model PTT-100 tags previously described, but its functionality was increased by the addition of light and pressure (depth) sensors and an increased data storage capacity. The manufacturer preprogrammed all the PTT-100 HR model tags to detach themselves from the fish after ten days, and the tags were activated prior to attachment to the animal by removing a small magnet from the side of the tag. The tags sampled environmental data at approximately four-minute or two-minute intervals. White marlin tagging procedures Preparations for tagging operations were made before each haulback of the gear. Tags were either activated prior to haulback or during haulback immediately follow- ing the tagging of a fish and during preparation for tag- ging another animal. Regardless of the time of external tag activation, all PSATs were allowed to cycle through their full ten-minute computerized internal activation process prior to being attached to a fish. The captain of the vessel identified incoming white marlin on the line during the morning haulback of the gear and fish were evaluated as live or dead based on movement (or lack thereof) alongside the vessel. All live white marlin were tagged, regardless of physical condition. Fish were manually brought alongside the vessel just aft of the hauling station along the rail and held briefly by the leader until calm. The average distance between the top of the rail and the fish (free-board) on the FV Carol Ann was approximately one meter, requiring the use of a tagging pole of approximately 2 m length to reach the fish over the gunwale. The nylon anchor to the PSAT tether was carefully inserted about 5-10 cm below the midpoint of the anterior dorsal fin to a depth of about 5 cm. This location on the fish provides an opportunity for the nylon tag head to pass through the pterygiophore bones without approaching the coelemic cavity (Prince et al., 2002a). For most white marlin in this study (93'7f ), a conventional streamer tag was also attached well posterior to the PSAT. White marlin were released as soon as possible after tagging by the standard commercial protocol of cutting the leader near the hook unless the hook was readily accessible for manual removal. No animals were resus- citated after tagging. Prior to release, hook type was noted and fish lengths and weights were estimated. Disposition ("live" vs. "dead") and hook location data were collected for all white marlin caught in 2003 and 2004. For the purposes of this study, "internal" hook locations were those in which the barb of the hook was lodged posterior to the esophageal sphincter, and "external" hook locations were noted with more speci- ficity (e.g., "upper jaw"). Hooking on the body away from the mouth ("foul hooking") was considered an "external" hook location. In addition to noting hooking location, a rapid visual examination of each fish was conducted using the five-point "ACESS" scale of activ- ity, color, eye condition, stomach status, and body state (see Kerstetter et al., 2003). The tagging operation, from positive species identification to actual release from the gear, lasted less than 10 minutes. All data, including the time of day, vessel location, and surface water temperature, were recorded immediately after tagging. Kerstetter and Graves: Survival of Tetrapturus albidus after release from longline gear 437 Table 1 Summary of locations, trips, and individual sets taken on a commercial pelagic longline vessel between June 2002 and August 2004 during tagging activities. Location refers to National Marine Fisheries Service (NOAA) statistical areas: FEC = Florida East Coast, NEC = Northeast Coastal. MAB = Mid-Atlantic Bight, COM = Gulf of Mexico, and CAR = Caribbean. For hook type, OS = offset and NOS = non-offset. 2002 2003 2004 Months June August July- September January-February August Location FEC NEC MAB GOM and CAR MAB Number tagged 6 2 6 2 12 Sets with tagging 5 1 5 2 3 Bait type frozen squid frozen squid frozen squid frozen squid, frozen mackerel, or mixture frozen squid, frozen mackerel, or mixture Hook type OS 9/0 J-style and NOS 18/0 circle OS 9/0 J-style and NOS 16/0 circle OS 9/0 J-style and NOS 16/0 circle OS 9/0 J-style and NOS 16/0 circle NOS 16/0 circle Data analysis Survival of tagged animals was inferred from three types of environmental data provided by the tag: water temperature changes, depth changes, and ambient light intensity. Frequent short-scale t0.16). For the 10° offset J-style hooks, the mortality rate was 20.0% excluding nontransmitting tags, and 55.6% if nontransmitting tags were included as mor- talities. The 0° offset circle hooks had a 7.1% mortality rate if nontransmitting tags were excluded and 27.7% if 440 Fishery Bulletin 104(3) these nontransmitting tags were included as evidence of mortalities. Nine white marlin were hooked in or near the eye. Seven fish were hooked on either circle or J-style hooks through the eye socket (with no visible damage to the eyeball) and all survived for the 5- or 10-day PSAT deployments. Two PSATs were attached to animals that had been hooked with a circle hook through the eye itself. One PSAT transmitted data consistent with survival, and the other tag did not transmit data. Only one white marlin tagged in this study was foul-hooked, caught in the ventral musculature by a size 18/0 circle hook. The PSAT attached to this fish separated from the fish prematurely. Discussion The amount of data archived and transmitted varied greatly among the three models of satellite tags, as well as among the 16 transmitting PTT-100 HRs. The early model PTT-100 tags archived only 63 data points, but 100% of the archived information was transmitted, providing sufficient information to infer survival (Graves et al., 2002; Kerstetter et al., 2003). In contrast, the newer PTT-100 HR tags archived either 4500 or 9145 data points, but not all archived data were transmitted. In this study, most of these tags transmitted a rela- tively large percentage of the archived data, facilitating determination of the fate of the released white marlin. However, one tag (MA-04-08) had an unusually low data transmitting rate of 4.4% , representing 315 data points over the ten-day tag deployment. Because these data points were transmitted in 11-minute blocks (approxi- mately 9 data points each), they often included complete short-duration movements of a fish from the surface to depth. As the transmitted blocks of data were distributed haphazardly over the entire ten-day tagging period, it remained possible to determine postrelease survival from a high-resolution tag with a low data recovery rate. Prior studies of postrelease survival have used differ- ent lengths of time to ascertain the effects of capture. These have included studies focused on postrelease survival as well as others addressing long-term be- havior, movements, and habitat preferences. Graves et al. (2002) justified a five-day deployment period for blue marlin by citing reports of blue marlin recaptured within five days after being released with conventional tags from the recreational fishery, thus demonstrating a return to feeding. Kerstetter et al. (2003) adopted a similar position, although their study on blue marlin also included the deployments of two PSATs for 30 days to evaluate the possibility of delayed mortality. Do- meier et al. (2003) used a variety of deployment periods (1-12 month durations) to assess postrelease survival in striped marlin. However, the longer the PSAT deploy- ment period, the more susceptible the animal becomes to both fishing (i.e., recapture) and natural mortality, such as predation, biasing upwards the estimate of postrelease mortality (Goodyear, 2002b). In our study, we primarily used tags with a ten-day deployment period and believe that this period is suf- ficiently long to document short-term mortality. Five of seven white marlin mortalities reported in Horodysky and Graves (2005) occurred within the first six hours of release, and the other two died less than three days later. All of the mortalities inferred for the closely re- lated striped marlin by Domeier et al. (2003) occurred within six days of release, and 75% of these mortalities happened in less than two days. The two documented mortalities in the present study (GB-02-01 and MA-03- 04) occurred within 24 hours of release. Direct comparisons of estimates of postrelease sur- vival of billfishes among previous acoustic and PSAT studies are problematic. Many acoustic tracking studies had relatively short observation periods and low sample sizes, and often fish in marginal physical condition were not tagged (reviewed in Domeier et al., 2003). Even among PSAT tagging studies, nontransmitting tags have been addressed with different protocols by various authors. Neither Graves et al. (2002) nor Kerstetter et al. (2003) directly observed mortalities of PSAT-tagged blue marlin. However, in both studies a conservative approach was adopted to estimate postrelease survival by considering nontransmitting tags as representing mortalities; this approach was adopted in part because of a lack of emergency release software or mechanisms on the tags themselves that would release the PSAT prior to its sinking with a dead fish below the depth at which the tag would be crushed. Some new models of satellite tags possess such emergency release software or physical mechanisms, such as glass implosion devices (Domeier et al., 2003) or the RD-1500 metal guillotine from Wildlife Computers (Redmond, WA) that sever the tether of the tag prior to reaching the depth limit of the tag. New generations of tags are also rated to greater crush depths (ca. 2000 m) than earlier models. The PSATs used in our study, with the exception of the one PAT tag, did not possess emergency release soft- ware or physical mechanisms. Because all the animals in this study were tagged over or near waters deeper than the crush depths of the tags, any deaths of tagged white marlin could have resulted in the PSATs being destroyed prior to transmitting data while the tag re- mained attached to the sinking, moribund fish. There are several reasons why PSATs may not re- port even with emergency releases, including recovery of the tag by a noncooperative fishing vessel, internal malfunction, or biological activities. Kerstetter et al. (2004) reported on three PSAT tags that were presum- ably ingested by sharks during predation or scavenging and suggested that a number of nontransmitting tags in all PSAT studies could result from biological activity. Goodyear (2002b) noted that including nontransmitting tags as mortalities would bias mortality estimates up- wards if the failure to transmit data was due to causes other than mortality. The combination of physically more robust tags, emer- gency release capabilities, and demonstrated mortalities has led authors (e.g., Domeier et al., 2003) to specifi- Kerstetter and Graves: Survival of Tetrapturus albidus after release from longline gear 441 cally exclude tags that do not transmit data from subse- quent analyses. Because it was not possible to estimate how many such tags in this study could have been due to malfunction versus individual mortality events, we chose to conservatively estimate two postrelease mortal- ity rates: one that includes all nontransmitting tags as mortalities and another that excludes nontransmitting tags. The expensive nature of PSAT technology resulted in relatively small sample sizes and hence large confi- dence intervals for the estimated postrelease mortality rates. However, as with Horodysky and Graves (2005), simulations with the observed rates in the present study have shown that very large sample sizes (over 200 tags for each hook type) under ideal conditions would be required to reduce these estimates to within ±5% of the true value. The advent of newer tag models with features such as an emergency release will presumably result in lower nontransmitting rates for PSATs and hence more accurate estimates of postrelease survival. In this study, PSATs attached to some white marlin in marginal physical condition at the time of release re- turned data consistent with postrelease survival. These included fish MA-04-03, which was hooked through the right eyeball, and fish WP-04-01, which displayed poor, faded color and was moving so little at haulback that it initially appeared dead until careful inspec- tion. Both internal hooking and stomach eversion have been suggested as predictors of subsequent mortality for billfishes (Domeier et al., 2003). Horodysky and Graves (2005) found a 40% mortality rate for inter- nally hooked white marlin, and Domeier et al. (2003) found a 63% mortality rate for similarly hooked striped marlin. We tagged four internally hooked animals, and the one reporting tag (GB-02-01) indicated mortality shortly after release for that fish. Three white marlin with everted stomachs at haulback were tagged in this study, but only one (MA-03-04) remained attached for the duration of the deployment period and transmitted data consistent with mortality. However, the survival of a white marlin (Horodysky and Graves, 2005) and a striped marlin (Holts and Bedford, 1990) with everted stomachs indicates that billfish with everted stomachs can survive if released. White marlin captured with circle hooks demonstrat- ed a trend of lower postrelease mortality than those hooked with J-style hooks, but this relationship was not significant. This trend in mortality rate versus hook type was independent of whether nontransmitting tags were included as mortalities or excluded from analyses. Horodysky and Graves (2005) observed a significant decrease in mortality for white marlin caught on circle hooks than on J-style hooks (0% versus 35% for J-style hooks). Domeier et al. (2003) also noted a trend for a lower mortality rate among animals hooked with non- offset circle hooks (12.5% versus 29.4% for offset J-style hooks), although this relationship was not significant. The lower mortality-rate trend for white marlin caught by circle hooks than by J-style hooks presented in the present study is also consistent with the results in sev- eral other studies of pelagic fishes, such Prince et al. (2002b) for recreationally caught billfish and Skomal et al. (2002) with recreationally caught Atlantic blue- fin tuna (Thunnus thynnus), which based predictions of postrelease survival on likely injury resulting from specific hooking locations on the animals. The majority of white marlin caught with circle hooks in the present study were hooked in the mouth or jaw (?2=23) rather than internally or by foul hooking on the body (n = ^), as was also noted by Horodysky and Graves (2005) for white marlin caught in the directed recreational fishery. In the present study, low num- bers of animals caught on either hook types prevented robust comparisons of postrelease survival rates by hook type. More balanced comparisons of postrelease survival among hook types were precluded by both a limited number of expensive PSATs and the imposition of a domestic management measure that prohibited the use of J-style hooks in the U.S. pelagic longline fish- ery as of 5 August 2004 (FR, 2004). Although beyond the scope of this study, any additional changes in the fishing practices of this fishery, such as the varying lengths of "soak time" between overnight sets (sword- fish) and daylight sets (tunas), may also affect the rates of postrelease survival of white marlin. Ultimately, hooking location may be a more important factor than hook type for predicting postrelease sur- vival. Three of the four PSATs attached to internally hooked animals in this study did not transmit data, although Prince et al. (2002b) reported encapsulated hooks from istiophorid viscera, indicating that inter- nal hooking events are not necessarily fatal. The large percentage of white marlin (35.7%) hooked through the upper lateral palate into the eye or eye socket raises some concern. Istiophorid billfishes are considered to be primarily visual predators (Rivas, 1975) and dam- age to an eye would be expected to negatively affect the foraging ability of the animal. Billfish are known to have specialized muscle tissue that allows individu- als to maintain elevated brain and eye temperatures (Block, 1986), and recent work has revealed color vision in some istiophorids (Fritsches et al., 2003). Dissections of sailfish have revealed that bookings in the eye socket often cause damage to the optic nerve and surrounding ocular musculature (Jolley^). The one fish caught with a circle hook through the eye socket in Horodysky and Graves (2005) survived for the entire 10-day deploy- ment period, and in the present study, the seven ani- mals hooked through the eye socket also all survived for their entire deployment periods, as did one white marlin caught with a circle hook through the eyeball. A tagged striped marlin in Domeier et al. (2003) with a punctured eye also survived for ten days, suggesting that this condition is not necessarily fatal over short durations, and healthy swordfish have been observed ■• JoUey, J. W. 1977. The biology and fishery of Atlantic sailfish Istiophorus platypterus, from southeast Florida, 31 p. Fla. Mar. Res. Pub., contribution no. 298. Florida Dep. Natural Resources, Marine Research Laboratory, 100 Eighth Avenue SE, St. Petersburg, FL 33701. 442 Fishery Bulletin 104(3) with one healed ocular cavity (D. W. Ker- stetter, personal observ. ). We observed a high percentage of hooked white marlin with associated eye damage, specifically in conjunction with circle hooks. In contrast, Horodysky and Graves (2005) noted only one animal out of 40 hooked through the eye with a circle hook. The difference between studies may be a factor of the hook sizes used in the fisheries; the recreational fishery gener- ally uses much smaller circle hooks than the commercial pelagic longline fishery (7/0 and 9/0 sizes versus 16/0 and 18/0). Jolley^ observed that for 134 (15.8%) of 848 sailfish caught recreationally with J-style hooks, the barbs exited near the eyes, noting that the distal lateral regions of the istiophorid mouth roof (those areas underlying the eyes) are thinly-covered muscle tissue rather than bone. A hook would therefore presumably pass much more easily through this tissue to the eye than if it encountered the lower jaw. Prince et al. (2002b) considered hooking through the upper palate potentially le- thal, not only because of the opportunity for the hook to penetrate the occipital or- bit, but also because of the tendency for J-style hooks in that location to compro- mise the integrity of the cranium, making it more susceptible to infection. Two tags that did not transmit data in our study were attached to fish caught with J-style hooks in the center of the upper palate. Borucinska et al. (2002) noted that for blue sharks iPrionace glauca) wounded by fishing hooks, an injury caused by a perforating hook may lead to systemic debilitation over longer time inter- vals than that typically measured with PSAT tags. The postrelease mortality rates obtained for white marlin from Horodysky and Graves (2005) and this study also allowed the estimation of total U.S. fishing mortality for this species. For the U.S. directed recre- ational fishery, the white marlin postrelease mortality rate (35% for J-style hooks; Horodysky and Graves, 2005) was applied to estimated yearly catch data and added to "best estimates" of the U.S. recreational land- ings (Goodyear and Prince, 2003). For the pelagic long- line fishery, catch and condition at release data were obtained from the NMFS Pelagic Observer Program database (Lee''). The 55.6% postrelease mortality rate (J-style hooks, nontransmitting tags as mortalities; present study) was applied to the number of white mar- lin released alive to obtain an estimate of the number of fish that died following release. Average underesti- n Recreational landings ■ Pelagic longline dead discards D Estimated recreational postrelease mortality (35%) B Estimated pelagic longine postrelease mortality (55 6° Figure 2 Calculated white marlin (Tetrapturiis albidus) fishing mortality esti- mates in metric tons (tl for the recreational and pelagic longline fisheries of the United States. The bottom part of each bar represents the reported mortality in each fishery (recreational landings and commercial dead discards, respectively), while the top part of the bar represents the possible additional fishing mortality based on conservative assumptions of 359r postrelease mortality with J-style hooks for the recreational fishery (Horodysky and Graves, 2005) and 55.6% postrelease mortality with J-style hooks in the commercial pelagic longline fishery (present study). The solid line is the three-year running average for estimated total recreational mortality (reported and estimated postrelease mor- tality), and the dashed line is the estimated total commercial pelagic longline mortality. Lee, D. 2004. Personal commun. NOAA/NMFS Southeast Fisheries Science Center, 75 Virginia Beach Dr., Miami, FL 33149. mates of the actual white marlin fishing mortality to recreational fishery reported landings or to commercial fishery dead discards during this ten-year period were 88.6% and 61.6%, respectively. Our analysis indicates that the directed U.S. rec- reational fishery may generate higher levels of white marlin fishing mortality than the U.S pelagic long- line fishery in some years simply due to greater num- bers of animals caught (Fig. 2). Because we chose the postrelease mortality estimates based on the historic terminal gear choices of J-style hooks, these results do not account for the probable decrease in total white marlin postrelease mortality resulting from mandated (pelagic longline) and voluntary (recreational) changes in the U.S. fisheries from J-style hooks to circle hooks. However, even this estimated magnitude of actual mor- tality incurred as the result of the U.S. recreational or pelagic longline fisheries results in the international pelagic longline fishery remaining the largest source of total white marlin fishing mortality in the Atlantic (ICCAT, 2005). The results of this study clearly demonstrate that white marlin are capable of surviving the trauma as- sociated with capture by pelagic longline fishing gear. Short-term survival of released white marlin was rela- tively high whether one discounted nontransmitting tags (89.5% survival) or considered nontransmitting Kerstetter and Graves Survival of Tetropturus albidus after release from longline gear 443 tags to represent mortalities (62.9% survival). These estimates are similar in magnitude to that found for the larger blue marlin released from pelagic longline gear (79% survival; Kerstetter et al., 2003). The documented survival of white marlin indicates that current domes- tic and international management measures requiring live release from commercial pelagic longline gear will reduce fishing mortality on this species. Acknowledgments We thank Captain Greg O'Neill, the crew, and owner Vince Pyle of the FV Carol Ann for their field assistance. Andrij Horodysky (VIMS) provided logistical field sup- port, as well as a critical review of the manuscript. Paul Howey, Lissa Werbos, and Shefali Mehta (Microwave Telemetry, Inc.) assisted with the PSAT programming and resulting data. We also thank Eric Prince (NMFS) for his support of the pilot work. This project was funded in part by NOAA/NMFS Cooperative Research Program Grant no. NA03NMF4540420 and the NOAA/NMFS Southeast Fisheries Science Center. Literature cited Bayley, R. E., and E. D. Prince. 1994. A review of tag release and recapture files for Istiophoridae from the Southeast Fisheries Science Center's Cooperative Gamefish Tagging Program. Int. Comm. Cons. Atl. Tunas (ICCATi Coll. Vol. Sci. Pap., Vol. XLI:.527-.548. Block, B. A. 1986. Structure of the brain and eye heater tissue in marlins, sailfish, and spearfish. J. Morphol. 190: 169-190. Block, B. A., D. T. Booth, and F. G. Carey. 1992. Depth and temperature of the blue marlin, Mak- aira nigricans, observed by acoustic telemetry. Mar. Biol. 114:175-183. Borucinska, J., N. Kohler, L. Natanson, and G. Skomal. 2002. Pathology associated with retained fishing hooks in blue sharks {Prionace glauca) with implications on their conservation. J. Fish Disease 25(9):.515-521. Brill, R. W., D. B. Holts, R. K. C. Chang, S. Sullivan, H. Dewar, and F. G. Carey. 1993. Vertical and horizontal movements of striped marlin (Tetrapturus audax) near the Hawaiian Islands, determined by ultrasonic telemetry, with simultane- ous measurement of oceanic currents. Mar. Biol. 117:567-574. de Sylva. D. P., W. J. Richards, T. R. Capo, and J. E. Serafy. 2000. Potential effects of human activities on billfishes (Istiophoridae and Xiphiidae) in the western Atlantic Ocean. Bull. Mar. Sci. 66(1):187-198. Domeier, M. L., H. Dewar, and N. Nansby-Lucas. 2003. Mortality rate of striped marlin {Tetrapturus audax) caught with recreational tackle. Mar. Freshw. Res. 54(4):435-445. FR I Federal Register). 2004. Atlantic highly migratory species (HMS), pelagic longline fishery, final rule. 69, Fed. Reg. 40733-40758 (July 6, 2004) (to be codified at 50 C. F. R., pts. 223 and 635). Fritches, K. A., L. Litherland, N. Thomas, and J. Shand. 2003. Cone visual pigments and retinal mosaics in the striped marlin. J. Fish Biol. 63:1347-1351. Goodyear, C. P. 2002a. Biomass projections for Atlantic blue marlin: potential benefits of fishing mortality reductions. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap. 52:1502-1506. 2002b. Factors affecting robust estimates of the catch and release mortality using pop-off tag technology. In Catch and release in marine recreational fisheries (J. A. Lucy, and A. L. Studholme, eds.), p. 172-179. Am. Fish. Soc, Bethesda, MD, Goodyear, C. P., and E. D. Prince. 2003. U.S. recreational harvest of white marlin. Int. Comm. Cons, Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap., vol. 55:624-632. Graves, J. E., B. E. Luckhurst, and E. D. Prince. 2002. An evaluation of pop-up satellite tags for estimat- ing postrelease survival of blue marlin. Fish. Bull. 100:134-142. Holts, D.. and D. Bedford. 1990. Activity patterns of striped marlin in the southern California bight. In Planning for the future of billfishes (R. H. Stroud, ed.l, p. 81-93. National Coalition for Marine Conservation, Inc., Savannah, GA. Horodysky, A. Z., and J. E. Graves. 2005. Application of pop-up satellite tag technology to estimate postrelease survival of white marlin iTet- rapterus albidus) caught on circle and straight-shank ("J") hooks in the western North Atlantic recreational fishery. Fish. Bull. 103(l):84-96. ICCAT (International Commission for the Conservation of Atlantic Tunas). 1995. Resolution by ICCAT for the enhancement of research programs for billfishes (blue marlin, white marlin, sailfish and spearfish). Int. Comm. Cons. Atl. Tunas (ICCAT) Res. 95-12., 1 p. 1996. Resolution by ICCAT Regarding the Release of Live Billfish Caught by Longline. Int. Comm. Cons. Atl. Tunas (ICCAT) Res. 96-9, 1 p. 1997. Recommendation by ICCAT Regarding Atlantic Blue Marlin and Atlantic White Marlin. Int. Comm. Cons. Atl. Tunas (ICCAT I Rec. 97-9, 1 p. 2000. Recommendation by ICCAT to Establish a Plan to Rebuild Blue Marlin and White Marlin Populations. Int. Comm. Cons. Atl. Tunas (ICCAT) Rec. 00-13, 2 p. 2001a. Report ofthe fourth ICCAT billfish workshop. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap. 53:1-22. 2001b. Recommendation by ICCAT to Amend the Plan to Rebuild Blue Marlin and White Marlin Populations. Int. Comm. Cons. Atl. Tunas (ICCAT) Rec. 01-10, 2 p. 2005. Report of the Standing Committee on Research and Statistics. Int. Comm. Cons. Atl. Tunas (ICCAT), Madrid, Spain, October 4-8, 2004, 224 p. ICCAT, Madrid, Spain. Jackson, T L., and M. I. Farber. 1998. Summary of at-sea sampling of the western Atlantic Ocean, 1987-1995, by industrial longline vessels fishing out ofthe port of Cumana, Venezuela: ICCAT Enhanced Research Program for Billfish 1987-1995. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap., vol. XVVII:203-228. 444 Fishery Bulletin 104(3) Jones, C. D., and E. D. Prince. 1998. The cooperative tagging center mark-recap- ture database for Istiophoridae (1954-1975) with an analysis of the west Atlantic ICCAT billfish tagging program. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap., vol. XLVn:311-322. Kerstetter. D. W., B. E. Luckhurst, E. D. Prince, and J. E. Graves. 200.3. Use of pop-up satellite archival tags to demonstrate survival of blue marlin (Makaira nigricans) released from pelagic longline gear. Fish. Bull. 101:939-948. Kerstetter, D. W., J. Polovina, and J. E. Graves. 2004. Evidence of shark predation and scavenging of fishes equipped with pop-up satellite archival tags. Fish. Bull. 102:750-756. NMFS (National Marine Fisheries Service). 1988. The Atlantic billfish fishery management plan. 120 p. NOAA-NMFS-F/SF-Highiy Migratory Species Division, Silver Spring, MD. Ortiz, M., E. D. Prince, J. E. Serafy, D. B. Holts, K. B. Davy, J. G. Pepperell, M. B. Lowry, and J. C. Holdsworth. 2003. Global overview of the major constituent-based billfish tagging programs and their results since 1954. Mar. Freshw. Res. 54(4):489-508. Pepperell, J. G., and T. L. O. Davis. 1999. Postrelease behavior of black marlin Makaira indica caught and released using sportfishing gear off the Great Barrier Reef (Australia). Mar. Biol. 135:369-380. Prince, E. D., M. Ortiz, A. Venizelos, and D. S. Rosenthal. 2002a. In-water conventional tagging techniques devel- oped by the Cooperative Tagging Center for large, highly migratory species. In Catch and release in marine recreational fisheries (J. A. Lucy, and A. L. Studholme, eds.), p. 155-171. Am. Fish. Soc, Bethesda, MD. Prince, E. D., M. Ortiz, and A. Venizelos. 2002b. A comparison of circle hook and "J" hook perfor- mance in recreational catch-and-release fisheries for billfish. In Catch and release in marine recreational fisheries (J. A. Lucy, and A. L. Studholme, eds.), p. 66-79. Am. Fish. Soc, Bethesda, MD. Prince, E. D., C. Rivero, J. E. Serafy, C. Porch, G. P. Scott, and K. B. Davy. 2003. An update of the tag release and recapture files for Atlantic white marlin. Int. Comm. Cons. Atl. Tunas (ICCAT) Coll. Vol. Sci. Pap., vol. LV(II):578-593. Rivas, L. R. 1975. Synopsis of biological data on blue marlin, Mak- aira nigricans Lacepede, 1802. In Proceedings of the international billfish symposium, part 3, Species synopses; Kailua-Kona, Hawaii, 9-12 August 1972, p. 1-16. NOAA NMFS SSRF-675 Tech. Report. Skomal, G. B., B. C. Chase, and E. D. Prince. 2002. A comparison of circle hook and straight hook per- formance in recreational fisheries for juvenile Atlantic bluefin tuna. In Catch and release in marine recre- ational fisheries (J. A. Lucy, and A. L. Studholme, eds.), p. 57-65. Am. Fish. Soc, Bethesda, MD. 445 Abstract — Examination of hard parts recovered from scats (feces) is cur- rently the most common method for determining the diet of pinnipeds. However, large or sharp prey remains may be spewed (regurgitated) biasing prey composition and size estimations in diet studies based on scats. Percent frequency of occurrence (F0%) and age or size of selected prey remains recovered from northern fur seal (Callorhinus ursinus) scat (/i = 3444) and spew samples (n=267) collected from rookeries on St. George Island and St. Paul Island, Alaska, between 1990 and 2000 were compared to determine if a bias in prey compo- sition and age or size estimations existed between scats and spews. Overall prey composition was simi- lar between sample type and location, but the relative F0% of primary prey (a5%) varied by sample type and loca- tion. Age or size estimates of wall- eye pollock iTheragra chalcograinma) and of two species of gonatid squids (Gonatopsis borealis and Berryteuthis magister) were significantly larger in spews than in scats. Observed differ- ences in F0% and estimated age or size of prey species whose remains were found in scats and spews likely result from size-selective digestion of prey remains. Scats were biased toward smaller prey remains, whereas spews were biased toward larger prey remains and cephalopod beaks. The percent overlap between age classes of walleye pollock caught by the commercial trawl fishery and age classes of walleye pollock consumed by northern fur seals varied notice- ably between sample types for both islands (scats: St. George = 15. 59c: St. Paul=4.19f ; spews: St. George = 94.6'X; St. Paul = 89.6'7f ). These results dem- onstrate that the inclusion of multiple sampling methods allows for a more accurate assessment of northern fur seal prey occurrence and prey age and size. Application of two methods for determining diet of northern fur seals (Callorhinus ursinus) Carolyn J. Gudmundson Tonya K. Zeppelin Rolf R. Ream National Marine Mammal Laboratory Alaska Fisheries Science Center National Marine Fisheries Service, NCAA 7600 Sand Point Way N.E Seattle, Washington 98115 E-mail address (for C J Gudmundson) Carolyn Jenkinswinoaa.gov Manuscript submitted 26 October 2004 to the Scientific Editor's Office. Manuscript approved for publication 12 October 2005 by the Scientific Editor. Fish. Bull. 104:445-455(2006). Pinniped diet analysis involves the identification of prey remains recov- ered from a variety of sample types. Prey remains are obtained from stomach lavage and enema proce- dures from feces (scats) and regur- gitations (spews), and from stomachs and gastrointestinal tracts. Histori- cally, northern fur seal {Callorhinus ursinus) diet studies have relied upon analysis of stomach and intestinal contents collected from scientific takes or harvested animals (Wilke and Kenyon, 1957; Kajimura, 1984; Bigg and Fawcett, 1985; Sinclair et al., 1994) and, more recently, scat samples (Sinclair et al., 1996; Antonelis et al., 1997). Because a thorough analysis of gastrointestinal contents generally requires the sacrifice of an animal, intact gastrointestinal specimens are no longer typically used in diet analy- ses. Lavage and enema procedures are also not widely used to describe the diet of the population because sample sizes are often small and animals may require chemical immobilization, thereby increasing the risk of injury or fatality (Antonelis et al., 1987; Harvey and Antonelis, 1994). Presently, scats are the most com- monly used sample type in pinniped diet analyses because they are easy to collect, abundant, and noninvasive (Harvey, 1989; Hammond and Prime, 1990). However, many studies have demonstrated the potential biases as- sociated with this sampling method (Jobling and Breiby, 1986; Pierce and Boyle, 1991; Bowen, 2000), including the accumulation of cephalopod beaks in the stomach (Bigg and Fawcett, 1985; Gales et al., 1993; Harvey and Antonelis, 1994) and underestimation of size and frequency of occurrence of some prey species (Bigg and Faw- cett, 1985; Harvey, 1989; Tollit et al., 1997). Captive feeding studies of nu- merous pinniped species have shown that factors such as species, sex, in- dividual activity level, stomach size, gut length, prey digestibility, feeding regime, and meal size affect the de- gree of erosion and recovery of prey remains in scats (Harvey and Antone- lis, 1994; Tollit et al., 1997; Marcus et al., 1998; Bowen, 2000; Orr and Harvey, 2001). Although studies have been conducted to account for these biases through the use of correction factors (Sinclair et al., 1994; Antone- lis et al., 1997; Tollit et al., 1997), the different retention and digestive rates of prey remains in the stomach continue to be a leading criticism for diet studies in which scat samples alone are used. Spews have been analyzed for sev- eral species of pinnipeds (Gales et al., 1993; Harvey and Antonelis, 1994; Kiyota et al., 1999; Lowry and Car- retta, 1999; Kirkman et al., 2000). Spews contain prey remains such as fish bones, otoliths, and cephalopod beaks. Often these prey remains are too large to pass through the pyloric sphincter and are, therefore, regur- gitated from the stomach (Bigg and Fawcett, 1985; Jobling and Breiby, 1986). Although spewings may be found in areas where pinnipeds come ashore, they are often less abundant 446 Fishery Bulletin 104(3) than scats (Jobling and Breiby, 1986; Gales et al., 1993) and have largely been excluded from pinniped diet analyses. Spews from northern fur seals are pres- ent on summer breeding islands, but diet studies that are based on spews are limited (Kiyota et al., 1999). Our study is a comparative evaluation of diet based on scat and spew samples collected from northern fur seal breeding rook- eries on two islands in the east- ern Bering Sea. Prey remains in scat and spew samples were compared for prey species com- position, and the age or size of walleye pollock (Theragra chal- cogramma) and of two species of gonatid squids (Gonatopsis bo- realis and Berryteuthis magister) were estimated. Finally, we com- pared age classes of walleye pol- lock found in scats and spewings to age classes of walleye pollock caught by the commercial trawl fishery in management regions adjacent to the Pribilof Islands. Figure 1 Location of St. George Island and St. Paul Island, within the Pribilof Islands, Alaska. Shaded area indicates walleye pollock iTheragra chalcogramma) com- mercial trawl fishery management areas surrounding the Phibilof Islands. Materials and methods Sample collection and processing Scats and spews were collected opportunistically from rookeries on the Pribilof Islands (St. George and St. Paul), Alaska (Fig. 1), during the breeding season between 1990 and 2000. Samples collected from rook- eries during this time (late July through September) were considered to be primarily from females of breeding age (Antonelis et al., 1997). Each scat and spew sample was placed in a plastic bag and frozen until it was analyzed in the laboratory. Samples were thawed and soaked in a mild emulsifying soap solution then rinsed through nested sieves of 4.75, 1.4, 1.0, and 0.5-mm mesh. Bones, otoliths, beaks, and eye lenses were recovered from the sieved samples and stored for analysis. Bones, otoliths, and eye lenses were stored dry in vials and beaks were stored in vials containing 50% isopropyl alcohol (Antonelis et al., 1997). Recovered diagnostic fish bones, otoliths, and cephalopod beaks were identi- fied to the lowest possible taxon by comparing them to a reference collection. Distinctions, based on the mor- phological features of beaks, between some cephalopod species within the family Gonatidae were not possible. Therefore, identifications of four squid species known to occur in northern fur seal diet were categorized into two groups following Sinclair et al. (1994) and Antonelis et al. (1997) and are referred to as Gb-Bm (Gonatopsis horealis and Berryteuthis magister] and Gm-Gm (G. madokai and G. niiddendorfii). Additionally, distinctions between four gonatid species (Eogonatus tinro. G. herryi, G. pyros, and G. onyx), were difficult to determine from the morphological features of beak and, collectively, are referred to as "gonatid group I" in the present study. Prey indices Individual prey species and prey groups were analyzed according to their frequency of occurrence (FO) in scats and spews for both islands. The percent frequency of occurrence (F0%) was calculated by dividing the number of scats or spews containing a specific prey species or group by the total number of scats or spews containing identifiable prey remains. Data from each year were pooled for F0% calculations because of low sample sizes in some years. Prey species or groups with a F0% 26% when rounded to the nearest integer, for either island or sample type, were considered to be primary prey. The minimum number of individuals (MNI) of each prey type was calculated for each sample, and summed over all samples. The cephalopod MNI was estimated by using the maximum count of either upper or lower beaks in each sample. Fish MNI was estimated by us- ing the maximum count in each sample from left or right otoliths, plus half of the otoliths for which a side (right or left) could not be determined (Antonelis et al., 1997). Percent MNI (MNI'7f) was then calculated by Gudmundson et al Diet of Callorhinus ursinus 447 dividing the MNI of each prey species or group (for all samples) by the total MNI of all prey for samples that contained identifiable fish otoliths or cephalopod beaks (or both). To determine if the occurrence of prey varied by is- land or sample type (or both), the presence or absence of the primary prey (?!=number of scats with prey re- mains) was modeled as a binomial random variable by using generalized linear models with island, sample type, and the interaction of island and sample type as explanatory variables (S-PLUS 2000, Insightful Corp., Seattle, WA). The interaction was considered significant if adding the interaction significantly re- duced the deviance in the model. If the interaction was significant, islands within each sample type or sample types within each island were compared, depending on whether sample type or island reduced more of the overall deviance. If the interaction was not significant, it was removed from the model, and island and sample type effects were tested. Even though data were insuffi- cient to test for differences in prey occurrence between years, year was included in the model as an additive variable. Comparisons of prey age and size Recovered walleye pollock otoliths were assigned a con- dition grade ("good," "fair," or "poor") based on distinc- tive features such as sulcus definition, shape, chipping, breaks, and wear (Sinclair, 1988). Pollock otoliths of "good" or "fair" condition were measured lengthwise parallel to the sulcus to the nearest 0.1 mm using hand- held digital calipers. A correction factor was applied to "fair" otoliths to account for loss of otolith length as a result of digestion (Sinclair, 1988; Antonelis et al., 1997). Fork lengths of prey were estimated by using regression formulae of otolith length against body length (Frost and Lowry, 1981) and age class was estimated from fork- length-age relationships (Sinclair et al., 1994). Pollock otoliths of "poor" condition were enumerated, but not measured, and were not included in prey size compari- sons because of their high degree of erosion. Pollock otoliths recovered from scat samples processed in the early 1990s were measured, but only age class estimations were recorded in the database. Therefore, to test how the size of walleye pollock otoliths varied by sample type and island, otoliths of "good" and "fair" grades were combined into two age categories; juvenile (0-2 age) and adult (3-5-i- age) and each sample (scat or spew) was categorized as containing juvenile, adult, or mixed (juvenile and adult) pollock. Multidimensional contingency tables with island, sample type, and age category as variables were used to test interactive ef- fects among variables. A saturated model including all variables and interactions was compared with re- stricted models by using chi-square goodness-of-fit test (S-PLUS 2000, Insightful Corp., Seattle, WA). Samples were pooled if sample type or island effects were condi- tionally independent, and two-dimensional contingency tables were then used to test variables that were not independent. The size range of Gb-Bm consumed by northern fur seals was estimated by measuring the rostral length of lower beaks recovered from scat and spews. Rostral length was measured to the nearest 0.1 mm with an optical micrometer. Because cephalopod beaks are more resistant to digestion and to subsequent loss of length than are otoliths (Sinclair et al., 1996; Tollit et al., 1997), Gb-Bm lower beaks were not assigned condition grades prior to being measured. However, lower beaks showing excessive wear, such as a broken rostral tip, were not measured. To evaluate Gb-Bm prey size differences between sample type and islands, we developed (using combined samples of both species) regression equations for low- er beak rostral length (LRL) against dorsal mantle length (DML) and for DML against weight. Squid speci- mens were opportunistically collected from commercial pollock trawl fishery bycatch, research driftnets, and NOAA research vessel mid-water trawl operations be- tween 1979 and 2000. Sampling areas were broad rang- ing throughout the North Pacific, at numerous localities in the eastern Bering Sea, Gulf of Alaska, and subarctic Pacific Ocean south of the western Aleutian Islands (WalkerM. The regression of LRL against DML for Gb-Bm was developed by using 757 lower beaks (n = 482 Gb, 275 Bm) with a dorsal mantle length range of 21-386 mm and the DML-weight regression was developed by using 1676 lower beaks (/? = 1048 Gb, 628 Bm) with a dorsal mantle length range of 17-386 mm. The regression in- cluded size ranges of Gb-Bm found in northern fur seal scats and spews examined in the present study. Linear models were used to develop regressions of LRL (mm) against DML (mm) and DML against weight (grams): DML = 39.37(Li?L)- 0.50. Log(jmg/!n = 2.87(Log( DML)) -4.16. A high degree of correlation (LRL to DML P<0.001; /•'- = 0.98; SE = 0.51); DML to weight P<0.001; ^2 = 0.99; SE = 0.01) was found for both regression equations (Walker^). Gb-Bm DML data were log transformed and differences in DML between sample types and island were determined by comparing means with a two-sample f-test. Although the DML-weight regression is not used in the analysis, it is included here for future use by other researchers conducting bioenergetics studies. Prey age or size estimations were limited to wall- eye pollock and Gb-Bm because sufficient numbers of otoliths or beaks of other primary fish and cephalopod prey species were not recovered from scat and spew samples, or because regressions for the species were unavailable. Walker, W. A. 2004. NMFS database. Alaska Fisheries Science Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. 448 Fishery Bulletin 104(3) Overlap between age classes of pollock consumed by northern fur seals and pollock caught by the commercial trawl fishery Age-class distributions of walleye pollock in scat and spew samples were compared with pollock size-compo- sition data in commercial midwater trawls (National Marine Fisheries Service, North Pacific Groundfish Observer Program). Trawl data included samples col- lected from August through September 1990 to 2000 in fishery management areas surrounding the Pribilof Islands (Berger^; Fig. 1) that encompass northern fur seal foraging habitat (Robson et al., 2004). Walleye pol- lock fork length from trawls was converted to age class by following the method of Sinclair et al. (1994) for determination of overlap with the age class of pollock consumed by northern fur seals (as estimated from scat and spew samples). Results Prey abundance A total of 1127 scats and 204 spews from St. George Island and 2317 scats and 63 spews from St. Paul Island contained prey remains (Table 1). Prey species compo- sition was similar between scats and spews for each island, but the relative importance of primary prey species based on F0% and MNI% varied by sample type and island (Table 2). Primary prey species (FO^ ^5%) found in both scats and spews consisted of gonatid group I, Gb-Bm, Gm-Gm, northern smoothtongue (Leuroglos- sus schrnidti). Pacific herring (Clupea pallasi). Pacific salmon (Oncorhynchus spp.), Pacific sand lance (Ammo- dytes hexapterus). and walleye pollock (Fig. 2). Significant interactions between sample type and island were found for Gb-Bm, Gm-Gm, Pacific salmon, and walleye pollock (Table 3); therefore the F0% of these prey was compared within island or sample type. The occurrence of Gb-Bm in scats versus spews from St. George Island was significantly different— spews having a higher F0% than scat. The F0% of Gb-Bm in scats versus spews collected from St. Paul Island was not significantly different. The F0% of walleye pollock was higher for St. George Island scats than for St. George Island spews. Pollock had the highest F0% of all prey in St. Paul Island scat and spews but was not significantly different between sample types. The ¥0% of Pacific salmon was significantly higher in spews than scat samples for both islands. The ¥0% of Gm-Gm was significantly higher in St. George Island spews than in St. Paul Island spews. Of the prey for which no interaction between sample type and island was observed, differences in the ¥0% of gonatid group Table 1 Numbers of northern fur seal iCallorh iiius ursinusi scat and spew samples collected on St. George Island and St. Paul Isla nd, Pribiloflslands, Alaska rookeries during the | breeding season (late July-September) 1990- 2000. Year St. George Island St. Paul Island Scats Spews Scats Spews 1990 162 28 484 1 1991 109 5 1992 84 32 291 13 1993 10 1 1994 184 46 230 3 1995 149 7 319 17 1996 141 11 337 14 1997 120 35 80 1 1998 199 24 207 5 1999 30 4 2000 78 20 230 Total (7!) 1127 204 2317 63 2 Berger, J. 2004. NMFS. database. Alaska Fisheries Sci- ence Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. I, northern smoothtongue, and Pacific sand lance were significant between islands and between sample types (Table 3). Pacific herring was only different between islands. Year of collection had a significant effect on the occurrence of all primary prey species except gonatid group I and was included in the model as an additive variable. Although scat and spew samples were collected during the same season each year, annual comparisons between sample types were not possible because of low sample size within years (Table 1). Variations in MNIVf were observed between sample types and island for Gb-Bm and walleye pollock (Table 2). The MNI<7f of Gb-Bm was higher in spews than in scat samples for both islands. Conversely, the MNI<7r of walleye pollock was higher in scats than in spews for both islands. Scat and spew samples from St. George Island had higher MNI% of Gb-Bm than those collected from St. Paul Island. The MNI^ of walleye pollock in both sample types from St. Paul Island was greater than the MNI% of walleye pollock from St. George Is- land in scats and spews. Comparisons of prey age and size A total of 1755 walleye pollock otoliths were recovered from 173 St. George Island scat samples containing wall- eye pollock and, of these, 1484 were determined to be in "good" or "fair" condition. Of the 152 walleye pollock oto- liths recovered from 27 St. George Island spew samples, 105 were graded as "good" or "fair." From St. Paul Island, 17,827 of 20,062 walleye pollock otoliths recovered from 902 scat samples and 154 of 202 walleye pollock otoliths recovered from 23 spew samples were determined to be Gudmundson et al : Diet of Callorhmus ursinus 449 Table 2 Percent frequency of occurrence (FC/r) and percent minimum number of individuals (MNI9f ) of prey species recovered in north- ern fur seal (.Callorhinus ursinus) scat and spew samples collected on St. George Island and St. Paul Island, Pribilof Islands, Alaska, during the breeding season (late July-September), 1990-2000. Table only includes prey species with an FO%sl% within either sample type and on either island. Prey species St. George Island St. Paul Island F0% MNI% F0% MNI% Seats Spews Scats Spews Scats Spews Scats Spews Atka mackerel (Pleurogrammus monoptcrygius) 2,1 1.0 <1.0 3.6 <1.0 Bathylagidae sp. 1.8 <1.0 <1.0 <1.0 Gb-Bm (Gonatopsis borealis and Berryteuthis magister) 22.7 55.4 62.5 94.6 5.6 9.5 5.3 32.7 Gni-Gm [Gonatus madokai and Gonatus middendorfii) 7.4 14.7 2.8 <1.0 9.8 3.2 3.5 1.6 Gonatid group I iE. tinro, G. berryi, G. onyx, and G. pyros) 3.9 23.5 1.4 2.1 1.1 4.8 <1.0 2.0 Gonatidae sp. 1.4 1.0 <1.0 <1.0 <1.0 <1.0 Gonatus sp. <1.0 2.9 <1.0 <1.0 <1.0 <1.0 Greenling sp. {Hexagrammos sp.) <1.0 1 <1.0 Northern lampfish iStenobrachius nannochir) 1.0 <1.0 <1.0 <1.0 <1.0 Northern smoothtongue iLeuroglossus schmidti) 7.4 5.4 1.7 1.5 Pacific cod (Gadus macrocephalus) <1 <1.0 1 <1.0 Pacific herring tClupea harengus) 2.0 2.0 <1.0 3.9 4.8 <1.0 Pacific salmon iOnchorhynchus spp.) 14.8 29.4 <1.0 7.5 34.9 <1.0 Pacific sand lance (Ammodytes hexapterus) 2.6 <1.0 8.9 1.1 Pacific sandfish ^Trichodon trichodon) <1.0 <1.0 2 1.6 1.3 Sablefish (Anoplopoma fimbria) <1.0 1.3 Unidentifed cephalopod 2.3 4.9 <1.0 1.2 <1.0 Unidentified fish 27.6 3.9 <1.0 31.2 <1.0 Unidentified gadid 17.3 15.2 <1.0 33.4 6.3 2.1 Walleye pollock iTheragra chalcogramma) 59.6 36.8 22.9 2.1 71 69.8 82.3 63.3 Sample size (;; ) (1127) (204) (5043) 5754) (2317) (63) (19,028) (251) in "good" or "fair" condition. The occurrence of walleye pollock age classes determined from otolith measure- ments varied by sample type (Fig. 3). A multidimensional contingency analysis showed that sample type was independent of island (P= 0.087, X~ = &.2G, df=3); therefore data from St. George and St. Paul Islands were pooled to test for differences in age class of prey between sample types. Spews contained a larger percentage of adult pollock (88.0%, P<0.001, X^=2&Q.21, df=l) and a smaller percentage of juvenile pollock (9.1%, P<0.001, x- = 180.91. df=l) than scats. There was not a significant difference in the occur- rence of mixed-age pollock between scats (6.4%) and spews (4.0%) (P=0.490, x-=QAl, df=l) and only 71 of 1125 (6.3%^) samples contained both age classes. A total of 2856 and 5030 Gb-Bm lower beaks were recovered and measured from 225 scat samples and 111 spew samples, respectively, collected at St. George Island (Fig 4.). A total of 719 Gb-Bm lower beaks were re- covered and measured from 110 St. Paul Island scat samples and 47 Gb-Bm lower beaks were recovered and measured from five St. Paul Island spew samples. The mean DML of Gb-Bm in St. George Island scat and spew samples was 69 mm (SE = 0.10) and 82 mm (SE = 0.14), respectively. The mean DML of Gb-Bm in St. Paul Island scat samples was 67 mm (SE = 0.21) and in spew samples was 101 mm (SE = 6.35). Gb-Bm data were log transformed to meet assumptions of normal- ity for island and sample-type size comparisons. The DML of Gb-Bm estimated from beaks recovered from spew samples was significantly larger than the DML estimated from beaks recovered from scat samples col- lected from both islands (two-sample ^test, P<0.001 for St. George Island and St. Paul Islands). The DML of Gb-Bm from St. George Island scats was significantly larger than that form St. Paul Island scats (two-sample /-test, P<0.001). In spew samples, estimated Gb-Bm DML was significantly larger on St. Paul Island than on St. George Island (two-sample f-test, P<0.001), How- ever, the number of spews collected from the two is- lands were highly unequal (7i=204, St. George Island; « = 63, St. Paul Island). 450 Fishery Bulletin 104(3) son 70 60 50 40 30 20 A St. George island i\ iL Gonatid Gb-Bm Gm-Gm Northern Pacific group I smoothlongue herring Pacific Pacific salmon sand lace Walleye pollock B St. Paul Island :_di JZL Gonatid group I Gm-Gm Northern Pacific smoothtongue herring Prey species Pacific Pacific salmon sand face Walfeye poflock Figure 2 Percent frequency of occurrence (FO'i) of primary prey species (FO^^S^) recovered from northern fur seal iCallorhinus ursinus) scats and spews collected from rookeries on (A) St. George Island and (B) St. Paul Island, Pribilof Islands, Alaska during the breeding season (late July-September), 1990-2000. Gonatid group I includes E. tinro, G. herryi, G. onyx, and G. pyros. Overlap between age classes of pollock consumed by northern fur seals and pollock caught by the commercial trawl fishery The percent overlap between age classes of walleye pol- lock caught by the commercial trawl fishery and those consumed by northern fur seals varied between sample types for both islands. Minimal overlap was observed between age classes of pollock found in St. George Island and St. Paul Island scats and those taken by the commercial trawl fishery (St. George Island=15.5%; St. Paul Island=4.1%, Fig. 3). However, the percent- age of overlap between age classes of pollock found in St. George Island and St. Paul Island spews and the age classes of pollock taken in commercial trawls was considerable (St. George Island=94.6%; St. Paul Island=89.6%, Fig. 3). Discussion Northern fur seal diet studies conducted since the 1980s have primarily used scat samples to determine prey spe- cies and size. Despite the presence of spews on rookery and haul-out beaches, only Kiyota et al. (1999) has utilized spews to study northern fur seal diet. Kiyota et al. (1999) examined spews and scats collected from haul-outs (subadult male dominated) and found differ- ences in the prey remains recovered between sample types. We compared the diet of adult female northern fur seals using scats and spews collected from rookeries on the Pribilof Islands over an eleven-year period. We observed differences in the percent frequency of occur- rence (Fig. 2) and the estimated age or size (Figs. 3 and 4) of prey between scats and spews. Additionally, the degree of overlap between age classes of walleye pollock consumed by northern fur seals and those caught in Gudmundson et al Diet of Collorh/nus ursinus 451 commercial fishery trawls varied considerably by sample type (Fig. 3). These findings highlight the potential limitations of a single sample type, such as scats, and demonstrate that a combina- tion of scat and spew samples can provide a more complete description of northern fur seal diet. The differences in prey occurrence that were observed among northern fur seal scats and spews in our study may have resulted from the accumulation of irregularly shaped or large prey parts in the stomach (Kiyota et al., 1999). Large cephalopod beaks and fish bones accumu- late near the pyloric sphincter and are regur- gitated from the stomach, whereas small prey parts pass through the digestive system and are recovered in scats (Bigg and Fawcett, 1985), The different rates of retention of prey parts in northern fur seal stomachs would, therefore, bias scat samples toward small prey species or small bones of large prey species and spew samples toward large prey and cephalopods. For example, the F0% of Pacific salmon, a large boned fish species, was higher in spews than in scats. Kiyota et al. (1999) also observed a high- er occurrence of Pacific salmon in spews than in scats of juvenile male northern fur seals from St. Paul Island. The FOVc we observed for northern smoothtongue and Pacific sand lance also support a prey species bias between sample types; the small forage fish were among primary prey species found in scat samples but were absent from spew samples. In addition to differences observed between sample types, we found that the occurrence of primary prey species varied between islands. Specifically, the FC/f of gonatid group I, Gb- Bm, Gm-Gm, and walleye pollock varied consid- erably between sample types and islands, and significant interactions between sample type and island were observed for a number of these species. Walleye pollock was the highest occur- ring prey in St. George Island scats, but the F0% was significantly lower in spew samples. On St. Paul Island, walleye pollock was the dominant prey species, but the F09f did not differ between scats and spews. Similarly, the F0% of Gb-Bm at St. Paul Island did not differ between sample types. However, the F0% of Gb-Bm was significantly higher in spews than in scats on St. George Island. The FCXf of gona- tid group I differed both between sample types and between islands; the occurrence of this group was highest in spews, and in samples from St. George Island. Overall, cephalopods occurred at higher frequencies on St. George Island than on St. Paul Island for both sample types. The lone exception was Gm-Gm, which had a lower frequency on St. George Island for scats. Observed differences in cephalopod and walleye pollock occurrence between St. George §e s >, >> _>. s^-^ >-. c o _>v o ^ c o u 2 « C CD c CO CO 3 ; cri d d ^ '5 E^ ■~ 05 t^ CM CO ,-H rH -^ S ■ ^ 3 o k = 00 3 O T3 S . ^ ~ 2 'g oi O o o t; 3 be CO o CO 5: ^^ sj -^ 0) O CO S n fur seal -Septemb and Gona CO a. CJ CO o. to O ^ ,-H ^ ,-H T3 CO t^ o o o CM o C JO T3 C o d o o d o o d o d o o d :S 3 .« O (U ^ "5 1 1 ,_, ,_, V V ^ 1 1 V ,-H 1 1 c .S a *s c ^ S m -O C 3 0) o .^ 00 Ti- !M 00 CO 00 t^ 00 "—I -— ; CM O) Ie3 entifi seas Gona d CO d ■^" CO CO ^ 'cC uo J -a 00 _ Ta ecies i- reedin m-Gm _ Tf oo 00 Ol CJ5 00 t:^ uo 0) a. -4-) c^ CO o CD o ,-H 00 o ^ Tf o < ,-H m C^ o o c^ o 05 y] ■^ 00 d d d d d i:^ O OS 3 nee ,A1 -yte QJ tC [^ 00 o o c~ o o o o o ng occur lof Islan is and B Qh o o CO o o o o o o o o o o o o o d d d d d d d d u V V V V V V V ■C 3 "a o 1 o 1 o o o o 1 o o 1 ^ 1 ^ I ^ ,-H ^ ^ 1 ^H 1 model cc ul Island natopsis G. pyros i-H 'x > — 1 ' — 1 CO eg co 05 c~ Ti; s 00 C-; d d c^ t-; d ^ r- 05 05 IC 05 ^ t- CO O -73 g cu O c 1 T3 -a 0? >■ alize id an :Gb G.o, c o a u c ^ . C JO ^ o Results of gene St. George Isla cephalopod pre Gonatiis berrvi o. o o bo c 'C u C 3 s BO i CO C O £ CO -a c CO a t E CQ Si E O s CO c o i- cu o o CO cd CO CO 1 Oh O o O z B. a- 0. 452 Fishery Bulletin 104(3) A St. George Island ■ St George Island scats B St George Island spews D Commercial trawl fishery L 3 B St. Paul Island 80 1 ■ St Paul Island seals /u J S St- Paul Island spews 60 ■ 50 ■ D Commercial trawl fishery 40 ■ 30 ■ 20 ■ 1 10 ■ I 0 ■ - ,ln .- 1 ,fl— 1 , —fdn , Age class Figure 3 Percent minimum number of individuals (MNI9r ) of walleye pollock iTheragra chalcogramma) age classes recovered from northern fur seal (Callorhinus ursinus) scats and spews containing measurable pollock otoliths collected from rookeries on (A) St. George Island (1484 otoliths from 173 scat samples; 105 otoliths from 27 spew samples) and (B) St. Paul Island (17.827 otoliths from 902 scat samples; 154 otoliths from 23 spew samples). Pribilof Islands. Alaska, during the breeding season (late July-September). 1990-2000. Also shown is the MNI'5 of walleye pollock age classes caught by the commercial trawl fishery between August and September 1990-2000 in fishery manage- ment areas surrounding the Pribilof Islands, Alaska. Island and St. Paul Island samples are likely related to the availability of prey within reach of female fur seals, which varies according to the physical and biological environment surrounding each island (Sinclair et al.. 1994; Antonelis et al.. 1997; Robson et al., 2004). The occurrence of important prey species in scat sam- ples was consistent with previous analyses of northern fur seal scat and stomach samples. However, there were considerable differences in prey occurrence between spews samples and previous studies (Kajimura, 1984; Sinclair et al., 1994, 1996; Antonelis et al., 1997). In northern fur seal diet studies based on analysis of the entire gastrointestinal tract adult females were found to primarily consume juvenile walleye pollock and gonatid squid (Sinclair et al., 1994; 1996). Likewise, diet stud- ies based on scat samples had the highest occurrences of walleye pollock and gonatid squid (Antonelis et al., 1997). Although we observed a similar FO^ of walleye pollock in our scat samples, the ¥0% of gonatid squid and Pacific salmon were significantly higher among spews compared to previous studies based on scat and G.I. tract samples. Thus, it is likely that the importance of some prey species, such as gonatid squids and Pacific salmon, has been underestimated in previous diet stud- ies (e.g., Antonelis et al., 1997). Significant differences in the size of Gb-Bm were observed between scat and spew samples, demonstrat- ing size-related digestive biases between sample types. Gudmundson et al Diet of Callorhinus ursinus 453 15 14 13. ^^■ 11 • 10 9 8 7 6 5 4 3 2 A St. George Island ■ Scats (n=225 samples contlning 2856 beaks) g Spews (n=5030 beaks in 1 1 1 samples) JUUUi 14 (55) 24 (94) 34 (133) 4 4 (173) 54 (212) 64 (251) B St. Paul Islantj 04 0,4 (15) ll Scats (n=71 9 beaks in 1 1 0 samples) Spews (n=47 beaks in 5 samples) u (55) (94) 34 (133) 4 4 (173) 54 (212) 64 (251) LRL (DML) Figure 4 Percent minimum number of individuals (MNI%) of Gonatopsis borealis and Berry- teuthis magister (Gb-Bm) lower rostral length (LRL) and dorsal mantle length (DMLl in mm recovered from northern fur seal iCallorhinus ursinus) scats and spews con- taining measurable Gb-Bm beaks collected from rookeries on (A) St. George Island and (B) St. Paul Island, Pribilof Islands, Alaska, during the breeding season (late July-September), 1990-2000. The dorsal mantle length size of Gb-Bm estimated from lower beaks was significantly larger in spew samples than scat samples (Fig. 4). Digestive biases relating to cephalopod beak size have also been observed in northern fur seal G.I. tracts (Yonezaki et al., 2003), as well as in Australian fur seal (Arctocephalus pusillus doriferus; Gales et al., 1993) and California sea lion (Zalophus californianus; Lowry and Carretta, 1999) spews and scats. Owing to limited taxonomic resolu- tion of squid species in previous northern fur seal diet studies, direct comparisons of Gb-Bm size were not pos- sible. For instance, Sinclair et al. (1994) presented DML size ranges of beaks recovered from stomachs, but, the cephalopod groups Gb-Bm and Gm-Gm were combined for analysis and were reported as gonatid squid. We observed a greater percentage of adult walleye pollock in spew than in scat samples (Fig. 3). In ad- dition, few scat and spewing samples were found to contain otoliths from both juvenile and adult age catego- ries, further demonstrating size-related digestive biases of prey remains between sample types. The differences in walleye pollock age classes between scat and spew samples seem to indicate that size estimations of pol- lock consumed by northern fur seals have likely been underestimated in previous studies using G.I. tracts and scats (Sinclair et al., 1994, 1996). Northern fur seal diet studies based on analysis of the entire G.I. tract have shown that adult females consume primar- ily juvenile walleye pollock (Sinclair et al., 1994. 1996). However, biases associated with the retention of large prey remains are inherent in studies based on G.I. tract samples (Pierce and Boyle, 1991; Gales et al., 1993), and size-selective relationships between prey otolith size and sample type have been documented for northern fur seals (Kiyota et al., 1999; Yonezaki et al., 2003). Data that accurately describe the ages of walleye pollock consumed by northern fur seals are critical for assessing potential competition with the commercial trawl fishery, which focuses on adult pollock. Previous northern fur seal diet analyses using scat and G.I tract 454 Fishery Bulletin 104(3) samples have revealed that northern fur seals primarily consume juvenile walleye pollock (Sinclair et al.. 1994, 1996), indicating that there is limited competition be- tween northern fur seals and the commercial trawl fish- ery. When walleye pollock age class was estimated from scats collected from 1990 to 2000, we found minimal overlap between the age classes of pollock consumed by northern fur seals and age classes of pollock caught by the commercial fishery during this time (Fig. 3). How- ever, when spew samples were used to estimate age/size of fur seal prey, a high degree of overlap between age classes of pollock consumed by northern fur seals and pollock caught by the commercial fishery was observed (Fig. 3). Because of the protective measures that resulted in closures of fisheries in Steller sea lion (Eiunetopias jubatus) critical habitat, there is concern that fishery pressures in waters surrounding the Pribilof Islands may increase (NMFS^). Approximately 75% of the glob- al northern fur seal breeding population inhabits the Pribilof Islands during the summer months (Loughlin et al., 1994) and recent estimates indicate a population decline. The total number of adult males on St. George and St. Paul Islands decreased by 13.4% and 2.8%, re- spectively, from 2002 to 2003 (Towell et al., 2006; York et al., 2005). Pup production also declined by >5% per year between 1998 and 2002 on both islands (York et al., 2005). At present, the cause of decline is uncertain but there is concern that increased fishing activity in waters surrounding the Pribilof Islands may adversely affect the northern fur seal population. We found that the occurrence and the age or size of some important prey species (such as walleye pollock) has been under- estimated in previous northern fur seal diet studies in which only scat samples were used. Because some of these species are also commercially important, northern fur seal conservation and fishery management decisions should incorporate diet-related information acquired from multiple noninvasive sampling methods. This ap- proach would allow better interpretation of northern fur seal dietary requirements, thereby providing a more accurate estimation of the extent to which protective measures in and around the Pribilof Islands should be instituted. Scat and spew samples were collected simultaneously from the same rookeries on the Pribilof Islands during the breeding season; therefore we do not believe our results were affected by sampling effort. The discrepan- cies in sample sizes observed in this study may be the result of northern fur seal foraging behavior. Satellite tracking and behavioral studies of female northern fur seals have indicated that individual at-sea foraging trips may range between 6 and 10 days on average (Loughlin et al., 1987; Gentry, 1998; Robson et al., 2004). The accumulation of large cephalopod beaks and 3 NMFS (NationalMarine Fisheries Service). 2001. Steller sea lion protection measures: final supplemental environ- mental impact statement, 3 volumes. NMFS, NOAA, Alaska Region, P.O. Box 21668, Juneau, AK 99802-1668. fish bones near the pyloric sphincter during this time may have irritated the stomach, causing the fur seals to regurgitate prey remains while at sea prior to their return to the rookeries. Although we do not know the regurgitation rates of fur seals, a tendency to regurgi- tate food during active digestion would explain the dif- ferences in sample size between spew and scat samples found on rookeries of the Pribilof Islands. Differences in proportions of sample types also have been observed on northern fur seal haul-outs (Kiyota et al., 1999), as well as on California sea lion haul-outs (Lowry and Carretta, 1999) and on Australian fur seal rookeries and haul-outs (Gales et al., 1993). Despite the skewed sample sizes between scats and spews in the present study, the observed differences in F0% and estimated age or size of prey between sample types were similar for both islands and were consistent with previous pin- niped diet studies comparing scat and spew samples (e.g., Gales et al., 1993, Kiyota et al., 1999). The various sampling methods used in pinniped diet analyses each have associated biases and sources of error that must be considered. The reliance upon any one sample type in diet assessments will limit one's ability to completely describe prey composition and prey size. Although our results corroborate the findings of previous diet studies with regard to the primary prey consumed by northern fur seals, this study demon- strates that using multiple sampling methods allows for a more accurate assessment of occurrence and age or size of prey. Acknowledgments We thank W. Walker for assistance with otolith and cephalopod identification, as well as Gb-Bm regression data; J. Thomason and E. Sinclair for help with cepha- lopod identification; J. Berger for providing commercial trawl fishery catch data; S. Crockford and Pacific Iden- tifications for identification offish bones; and J. Laake for statistical advice. Reviews and helpful comments on this manuscript were provided by W. Walker. K. Call, T. Loughlin, E. Sinclair, S. Melin, W. Jenkins, and three anonymous reviewers. Literature cited Antonelis, G. A., M. S. Lowry, D. P. DeMaster, and C. H. Fiscus. 1987. Assessing northern elephant seal feeding habits by stomach lavage. Mar. Mamm. Sci. 3:308-322. Antonelis, G. A., E. H. Sinclair, R. R. Ream, and B. W. Rob.son. 1997. Inter-island variation in the diet of female northern fur seals (Callorhinus ursinus) in the Bering Sea. J. Zool. (Lend.) 242:435-451. Bigg, M. A., and I. Fawcett. 1985. Two biases in diet determination of northern fur seals (Callorhinus ursinus). In Marine mammals and fisheries (J. R. H. Beddington, R. J. H. Beverton, and D. M. Lavigne, eds.), p. 284-299. George Allen and Unwin Ltd., London. Gudmundson et al Diet of Callorhinus ursinus 455 Bowen, W. D. 2000. Reconstruction of pinniped diets: accounting for complete digestion of otoliths and cephalopod beaks. Can. J. Fish. Aquat. Sci. 57:898-905. Frost, K. J., and L. F. Lowry. 1981. Trophic importance of some marine gadids in northern Alaska and their body-size otolith relationships. Fish. Bull. 79:187-192. Gales, R., D. Pemberton, C. C. Lu, and M. R. Clarke. 1993. Cephalopod diet of the Australian fur seal: varia- tion due to location, season and sample type. Aust. J. Mar. Freshw. Res. 44:657-671. Gentry, R. L. 1998. Behavior and ecology of the northern fur seal, 392 p. Princeton Univ. Press, Princeton, NJ. Hammond, P. S., and J. H. Prime. 1990. The diet of British grey seals (Halichoerus grypus). In Population biology of sealworm iPsuedo- terranova decipiens) in relation to its intermediate and seal hosts (W. D. Bowen, ed.), p. 243-254. Can. Bull. Fish. Aquat. Sci. no. 222. Harvey, J. T. 1989. Assessment of errors associated with harbour seal (Plioca vitulina) faecal sampling. J. Zool. (Lond.) 219:101-111. Harvey, J. T., and G. A. Antonelis. 1994. Biases associated with non-lethal methods of determining the diet of northern elephant seals. Mar. Mamm. Sci. 10(21:178-187. Jobling, M., and A. Breiby. 1986. The use and abuse offish otoliths in studies of feed- ing habits of marine piscivores. Sarsia 71:265-274. Kajimura, H. 1984. Opportunistic feeding of the northern fur seal, Callorhinus ursinus, in the eastern North Pacific Ocean and eastern Bering Sea. NOAA Tech. Rep. NMFS SSRF-779, 49 p. Kirkman, S. P., W. Wilson, N. T. W. Klages, M. N. Bester, and K. Isaksen. 2000. Diet and estimated food consumption of Antarctic fur seals at Bouvetoya during summer. Polar Biol. 23:745-752. Kiyota, M., C. Kawai, and N. Baba. 1999. Estimation of diet of male northern fur seals iCallorhinus ursinus) based on analysis of fecal and regurgitated materials. Bull. Nat. Res. Inst. Far Seas Fish. no. 36, p. 1-7. Loughlin, T. R., J. L. Bengston, and R. L. Merrick. 1987. Characteristics of feeding trips of female northern fur seals. Can. J. Zool. 65(81:2079-2084. Loughlin, T. R., G. A. Antonelis, J. D. Baker, A. E. York, C. W. Fowler, R. L. DeLong, and H. W. Braham. 1994. Status of the northern fur seal population in the U.S. during 1992. In Fur seal investigations, 1992 (E. H. Sinclair, ed.), p. 9-28. NOAA Tech. Memo. NMFS-AFSC-45. Lowry, M. S., and J. V. Carretta. 1999. Market squid (Loligo opalescens) in the diet of California sea lions (Zalnphus californianus) in southern California (1981-1995). CalCOFI Rep. 40:196-207. Marcus, J., W. D. Bowen, and J. D. Eddington. 1998. Effects of meal size on otolith recovery from fecal samples of gray and harbor seal pups. Mar. Mamm. Sci. 14(4):789-802. Orr, A. J., and J. T. Harvey. 2001. Quantifying errors associated with using fecal sam- ples to determine the diet of the California sea lion (Zalo- phus californianus). Can. J. Zool. 79:1080-1087. Pierce, G. J., and P. R. Boyle. 1991. A review of methods for diet analysis in piscivo- rous marine mammals. Oceanogr. Mar. Biol. Annu. Rev. 29:409-486. Robson. B. W., M. E. Goebel, J. D. Baker, R. R. Ream, T. R. Loughlin, R. C. Francis, G. A. Antonelis, and D. P. Costa. 2004. Separation of foraging habitat among breeding sites of a colonial marine predator, the northern fur seal [Callorhinus ursinus). Can. J. Zool. 82:20-29. Sinclair, E. H. 1988. Feeding habits of northern fur seals in the eastern Bering Sea. M.Sc. thesis, 94 p. Oregon State Univ., Corvallis, OR. Sinclair, E. H.. G. A. Antonelis, B. W. Robson, R. R. Ream, and T. R. Loughlin. 1996. Northern fur seal [Callorhinus ursinus) predation on juvenile walleye pollock, Theragra chalcogramma. NOAA Tech. Rep." 126:167-178. Sinclair, E. H., T. R. Loughlin, and W. G. Pearcy. 1994. Prey selection by northern fur seals (Callorhi- nus ursinus) in the eastern Bering Sea. Fish, Bull. 92:144-156. Tollit, D. J., M. J. Steward, P. M. Thompson, G. J. Pierce, M. B. Santos, and S. Hughes. 1997. Species and size differences in the digestion of otoliths and beaks: implications for estimates of pin- niped diet composition. Can. J. Fish. Aquat. Sci. 54:105-119. Towell, R. G., R. R. Ream, and A. E. York. 2006. Decline in northern fur seal (Callorhinus ursinus) pup production on the Pribilof Islands. Mar. Mamm. Sci. 22(21:486-491. Wilke, F., and K. W. Kenyon. 1957. The food of fur seals in the eastern Bering Sea. J. Wildl. Manag. 21(2):237-238. Yonezaki, S., M. Kiyota, N. Baba, T. Koido, and A. Takemura. 2003. Size distribution of the hard remains of prey in the digestive tract of northern fur seal (Callorhinus ursinus) and related biases in diet estimation by scat analysis. Mammal Study 28:97-102. York, A. E., R. G. Towell, R. R. Ream, and C. W Fowler. 2005. Population assessment of northern fur seals on the Pribilof Islands, Alaska, 2002-2003. In Fur seal inves- tigations, 2002-2003 (J. W. Testa, ed.), p. 8-27. NOAA Tech. Memo. NMFS-AFSC-151. 456 Abstract — Estimates of the abun- dance of American horseshoe crabs (Limulus polyphemus) are important to determine egg production and to manage populations for the ener- getic needs of shorebirds that feed on horseshoe crab eggs. In 2003, over 17,500 horseshoe crabs were tagged and released throughout Delaware Bay, and recaptured crabs came from spawning surveys that were conducted during peak spawning. We used two release cohorts to test for a tempo- rary effect of tagging on spawning behavior and we adjusted the number of releases according to relocation rates from a telemetry study. The abundance estimate was 20 million horseshoe crabs (909c confidence interval: 13-28 million), of which 6.25 million (90% CI: 4.0-8.8 million) were females. The combined harvest rate for Delaware, New Jersey. Virginia, and Maryland in 2003 was 4^ (90* CI: 3-6%) of the abundance estimate. Over-wintering of adults in Delaware Bay could explain, in part, differences in estimates from ocean-trawl sur- veys. Based on fecundity of 88,000 eggs per female, egg production was 5.5x10" (90% CI: 3.5x10", 7.7x10"), but egg availability for shorebirds also depended on overlap between horse- shoe crab and shorebird migrations, density-dependent bioturbation, and wave-mediated vertical transport. Abundance of adult horseshoe crabs (Limulus polyiphemus) in Delaware Bay estimated from a bay-wide mark-recapture study David R. Smith^ Michael J. Millard^ Sheila Eyler^ ' United States Geological Survey Leetown Science Center 11649 Leetown Road Kearneysville, West Virginia 25443 E-mail address ((or D R Smith) drsmitli g'usgs gov ^ U.S. Fish and Wildlife Service Northeast Fishery Center 308 Washington Avenue Lamar, Pennsylvania 16848 3 US- Fish and Wildlife Service Maryland Fisheries Resources Office 177 Admiral Cochrane Drive Annapolis, Maryland 21401 Manuscript submitted 29 March 2005 to the Scientific Editor. Manuscript approved for publication 21 October 2005 by the Scientific Editor. Fish. Bull. 104:456-464(2006). The American horseshoe crab (Lim- ulus polyphemus), having persisted largely unaltered for millions of years, is now central to a modern manage- ment debate of conflicting interests by commercial watermen, birdwatchers, biopharmaceutical companies, and eco-tourists (Odell et al., 2005). The management controversy is most acute in the Delaware Bay region where the high abundance of horseshoe crabs has resulted in the dependence of migrant shorebirds on horseshoe crab eggs to fuel their northern migration to nesting grounds (Botton and Har- rington, 2003). Because management goals have linked the horseshoe crab fisheries to the viability of other species, such as migrant shorebirds (ASMFC), it is not sufficient to manage on the ba- sis of typical reference points, such as maximum sustainable yield, be- cause sustainable harvest is not the primary issue. In Delaware Bay, the viability of shorebirds has taken pre- cedence in decision making (Botton and Harrington, 2003; Baker et al., 2004). Effective management should reference a critical threshold of horse- shoe crab abundance that provides sufficient eggs and should sustain that threshold. Thus, abundance be- comes a critical parameter because abundance estimates are useful for predicting the egg biomass that is available to migrant shorebirds and for assessing harvest rate. Horseshoe crabs bury their eggs in beach sediment, typically 15 to 20 cm deep (Shuster and Sekiguchi, 2003). Eggs are exhumed to the beach sur- face and become available to forag- ing shorebirds through a process of entrainment in activated sediment, followed by vertical transport into surface sediments. Nest disturbance, which precedes entrainment of eggs, is predominantly due to bioturba- tion; whereas wave energy is only a contributing factor because typical estuarine waves do reach nest depth (Jackson et al., 2005). Given a rela- tionship between egg exhumation and spawning density, egg availability could be predicted from current and projected horseshoe crab abundance. ASMFC (Atlantic States Marine Fish- eries Commission). 1998. Interstate fishery management plan for horseshoe crab. Fisherv Management Report No. 32, 58 p. ASMFC, 1444 Eye Street, NW, Sixth Floor. Washington, DC 20005. Smith et al.: Abundance of Limulus polyphemus in Delaware Bay 457 Thus, studies to estimate abundance are an important step in the process of man- aging horseshoe crabs to meet the ener- getic needs of shorebirds. We used results from a bay-wide mark- recapture effort during spring 2003 to estimate horseshoe crab abundance. Un- derlying assumptions of mark-recapture methods were accounted for in our study design and were evaluated during analy- sis. Recapture effort was distributed over Delaware Bay beaches by involving par- ticipants of a bay-wide spawning survey. We related our abundance estimates to reported landings to assess the recent harvest rate and used published fecun- dity estimates to predict egg production. Materials and methods We captured, tagged, and released horse- shoe crabs from boats during two peri- ods in 2003; a prespawning-season period (abbreviated to "preseason" in this article) from 26 March to 8 May and a prepeak- spawning period ( abbreviated to "prepeak" ) from 28 to 30 May 2003. The numbers of crabs tagged were 7221 and 10,322 during the preseason period and prepeak period, respectively. The target population was adult horseshoe crabs that were present in Delaware Bay to spawn. One boat fished throughout the bay during the preseason period, and three boats fished nearshore during the prepeak period (Fig. 1). Pre- peak period captures took place within strata of equal length along the Dela- ware Bay shoreline (Fig. 1). During the prepeak period, two boats fished in New Jersey, but only one boat fished in Dela- ware because of a lack of funding. Because of the additional boat in New Jersey, fish- ing effort and the number of crabs tagged were higher in New Jersey. Any animal injured during capture was culled and not tagged. Adult males and females were tagged with standard button tags. Tags were 4.4 cm in diameter, bore a unique tag number and carried instructions on the tag for report- ing a captured tag. Further detail on tags and tagging methods are described in Brousseau et al. (2004). Recaptured tags came from the Delaware Bay spawn- ing survey during the peak period of spawning (29 May, 31 May, and 2 June 2003). Although the spawning survey was conducted during spring tides in May and June, the recapture period for abundance estimation was limited to the peak spawning period to help satisfy the assumption of population-closure during the time of mark-recapture study (Borchers et al., 2002). Spawning survey volunteers were instructed to count all horse- \ 38 45'0"N- New Jersey 39 15'0"N •3845'0"N 75°15'0"W 75"0'Q"W Figure 1 Locations in Delaware Bay where horseshoe crabs {Limulus polyphemus) were tagged and released. Size of symbols are graduated to reflect size of catches, and symbol style reflects whether releases occurred during the prespawning season period (26 March to 8 May 2003: triangle) or prepeak-spawning period (28-30 May 2003; circle). White lines perpendicular to the shoreline indicate strata within which catch effort during prepeak-spawning period was equally distributed. Black lines along the shore indicate approximate locations of spawning areas. shoe crabs in sample quadrats and record tags that they encountered inside and outside of the sample quadrats. Flashlights were used when the survey occurred after dark. During the spring tide period around the new moon (29 May to 2 June 2003), 23 beaches were sur- veyed throughout Delaware Bay (Smith and Bennett'^; Smith, D. R., and S. Bennett. 2004. Horseshoe crab spawning activity in Delaware Bay: 5 years of a standard- ized and statistically-robust survey, 10 p. Report to the ASMFC Horseshoe Crab Management Board, Atlantic States Marine Fisheries Commission, Washington DC. Website: http://www.lsc.usgs.gov/aeb/2065/isa%20report%2005.pdf [accessed on 21 October 2005]. 458 Fishery Bulletin 104(3) Fig. 1). The design of the spawning survey is described in Smith et al. (2002). Mark-recapture methods We took two approaches to estimate abundance. The first was an application of Chapman's modification of the Petersen estimator (Borchers et al., 2002). We applied the Petersen estimator separately for each of the two release periods and the three spawning survey dates. In addition, we combined the releases and recaptures for a pooled Petersen estimate. The second approach was based on an extension of a likelihood presented by Borchers et al. (2002; p. 118, Eq. 6.11). We extended the likelihood to allow for a temporary effect on spawning behavior due to the cap- ture and tagging process by including separate recap- ture probabilities for each release cohort. The extended likelihood was Ls=i^]^"]p:-ii-Ps^''-''--"' ( M \ Mos (plf'^a-pl)'^''-'"^- where N = the abundance at the start of the recap- ture period at the end of May 2003; M,^ = the number of preseason tagged animals at large at time s\ M.,^ = the number of prepeak tagged animals at large at time s (M.^=Mj^+M.,J; irij^ and m.,^ = the recaptures of preseason and prepeak tagged animals at time s; u^ = the survey count of untagged at time s; p, = the capture probability for untagged and preseason tagged animals at time s; and Pj* = the capture probability of prepeak tagged animals at time s. We also fitted a likelihood that set all recapture prob- abilities to be constant through time, i.e., p^ = p for all recapture surveys, which reduces to Borchers' original likelihood (Borchers et al., 2002: p. 118, Eq. 6.11). We used maximum likelihood methods to estimate abundance A^ and recapture probabilities {p^.pj. We used the Petersen estimate for the initial value for N and used mJM^ as the initial value for recapture probabilities. Profile-likelihood intervals were calcu- lated for the abundance estimates (Borchers et al., 2002). MathCad (vers. 12, Mathsoft Engineering and Education Inc., Cambridge, MA) and SAS (vers. 9, SAS Institute Inc., Gary, NO were used to find nu- merical solutions to the likelihood and profile-likelihood equations. The following assumptions underlie the mark-recap- ture methods that we applied (Borchers et al., 2002): 1 No emigration or mortality occurred during the period between release and recapture; 2 the tagged animals represented an adequate sample; 3 animals were captured independently of one another; 4 tags were not lost or overlooked; and 5 recapture probability depended only on recapture occasion, was equal among animals of the same sex, and was equal for tagged and untagged animals. The study was designed through the timing and dis- tribution of releases and recaptures to meet the first two assumptions. The prepeak releases and recaptures were designed to be close in time to meet the assump- tions of no emigration and no mortality. Immigration occurred during the time between preseason release and recapture; therefore we estimated the number of adults in Delaware Bay at the time of recapture, which was at the end of May in 2003 (Skalski and Robson, 1992). Some mortality occurred during the time between preseason release and recapture that we were not able to account for; however, we expect that mortality was similar for tagged and untagged animals. We ensured that both initial capture and recaptures were spatially distributed by distribut- ing the releases throughout Delaware Bay during preseason tagging and within strata during prepeak tagging effort (Fig. 1) and by distributing recapture effort systematically by means of the spawning survey (Smith et al., 2002). The third assumption could be violated if horseshoe crabs moved locally in groups. However, recaptures came from widely spaced quad- rats, so that even if the animals moved in groups, the whole group was unlikely to be recaptured within single quadrats. Although tag loss could be a sig- nificant factor over an extended period, we did not believe that significant tag loss occurred, especially over the five days from prepeak release to recapture period (28 May to 2 June 2003). In a tag-loss study conducted at the United States Geological Survey Leetown Science Center with identical tags, no tag loss over ^60 days was reported, indicating that tag loss between preseason and recapture periods would not have been significant (Crawford'^). Tags could have been overlooked during the spawning survey. Females bury themselves in beach sediment during spawning, and their tags could have been readily overlooked. In contrast, males do not bury themselves and the 4.4 cm white button tag is highly visible in daylight or when illuminated by flashlight. Nevertheless, tags on males could have been obscured when the horse- shoe crabs piled up during peak spawning. Thus, we restricted our mark-recapture analysis to male horseshoe crabs that were counted and recaptured within 1-m- quadrats when surveyors were focusing on a small area. 3 Crawford, E. 2003. Unpubl. data. USGS-Leetown Science Center, 11649 Leetown Road, Kearneysville, WV 25430 Smith et al Abundance of Limulus polyphemus in Delaware Bay 459 We limited mark-recapture estimates to males be- cause capture probability was not equal for males and females as evidenced by recapture rates. Ratios of males and females captured for the tagging study were used to estimate total abundance with the equation estimate harvest rate. To put bounds on the problem, we calculated harvest rate 1) using landings from New Jersey and Delaware and 2) using landings from the Delaware Bay regional states (New Jersey, Delaware, Virginia, and Maryland). N=N,JR„^, where iV„,= abundance of males; and i?„, = the ratio of males to the total captured. Abundance of females was estimated by subtraction, i.e., by Nf=N-N,„. We used patterns of recaptures among preseason re- leases (released from 26 March to 8 May 2003) versus prepeak releases (released from 28 to 30 May 2003) to test the assumption that tagging did not affect spawn- ing behavior. One approach was to use a contingency table analysis that included a comparison of risks of recapture between animals tagged during preseason and prepeak periods. Another approach was based on the extended likelihood presented above (Z>, ), which included a separate set of recapture probabilities for preseason and prepeak released animals. In this way, the model allowed for a temporary effect on capture probability due to tagging, such as a temporary delay in spawning. We used model comparison techniques (i.e., Akaikes information criteria and likelihood ratio tests) to compare recapture probabilities among pre- season and prepeak released animals (Burnham and Anderson, 1998). Results from a telemetry study conducted in 2004 provided information on spawning behavior of horseshoe crabs caught and released from boats. In 2004, we at- tached radio transmitters to 60 adult males throughout the Delaware Bay prior to the spawning season. When the animals came on the beach to spawn the signals from the transmitters were recognized by one or more of the 14 fixed station receivers that were arrayed along the shoreline of the bay. In this article, we report the relocation rates (proportion of radio-tagged crabs that were recognized by at least one receiver) from that study because of their relevance to the assumption that tagging does not affect spawning behavior. An important application of abundance estimates is in the calculation of harvest rates. We calculated harvest rates by dividing estimated abundance into 1) reported landings and 2) projected landings (based on recent regulations). The population that spawns in Delaware Bay disperses widely, some leaving the bay for the ocean. Calculations of harvest rate need to account for these migration patterns because landings from New Jersey and Delaware do not include landings of Delaware Bay spawning animals that were harvested in neighboring states (Virginia and Maryland). How- ever, landings from neighboring states include animals that spawn in embayments other than Delaware Bay. Thus, use of landings from Delaware Bay states alone could underestimate harvest rate and use of landings from Delaware Bay and neighboring states could over- Results A total of 5398 males and 1823 females were tagged during the preseason period (26 March to 8 May 2003), and 7091 males and 3231 females were tagged during the prepeak period (28 to 30 May 2003; Table 1). Animals that were injured during capture were culled and not tagged. Culling rates were 0.19 during the preseason period, 0.05 during the prepeak period, and 0.12 overall. During the spawning survey 22,051 males and 6675 females were counted in quadrats and examined for tags. Forty-eight tagged animals were recaptured dur- ing the spawning season surveys conducted 29 May, 31 May, and 2 June 2003. Nineteen of the recaptures were within quadrats, and 29 were between quadrats. Only three of the recaptured tags were from females. This number is not surprising; tags are difficult to detect on females because the females remain buried during spawning. The between-quadrat recaptures were not used in the estimation because we did not have a cor- responding count of unmarked animals. Using the Petersen estimator and all releases, we estimated that there were approximately 14.5 million (SE = 3.2 million) adult male horseshoe crabs in Dela- ware Bay during peak spawning at the end of May 2003. Using just the prepeak releases, we estimated 13 million adult males (SE = 3.6 million). Model comparisons indicated that recapture prob- abilities were time-specific, but the temporary effect on recapture probabilities was not supported by the data (Table 2). The maximum likelihood estimate of male abundance with model 2 (Table 2) was 16.1 million (90% CI: 9.9 to 22.3 million). The difference between the Petersen estimate and the maximum likelihood estimate could have been due to model 2 that allowed for time-specific recapture probabilities, whereas the Petersen estimator did not. The maximum likelihood estimate based on constant recapture probabilities (i.e., model 1 in Table 2), matched the Petersen estimate (14.7 million for model 1 and 14.5 million for the Pe- tersen estimate). Sex ratios were different for the preseason and pre- peak spawning periods in a pattern consistent with a sex-specific migration schedule. There was a smaller proportion of adult females in the bay during the pre- season period (0.25) than when spawning activity was near its peak (0.31), which is consistent with the known observation that males migrate earlier than females. Sex ratios observed during the prepeak period were used to estimate abundance in the bay during the peak spawning period. The assumption of equal catchability of tagged and untagged animals is an important assumption that af- 460 Fishery Bulletin 104(3) Table 1 Mark-recapture statistics for tagged horseshoe crabs \Liiiuilits polyp/iemiis) released during a prespawning season period (26 March to 8 May 2003) and a prepeak-spawning period (28 to 30 May 20031 and recaptured during spawning surveys on 29 May, 31 May. and 2 June 2003. Mj^, is the number tagged preseason and M,,^ is the number tagged before the peak at the time of the survey date (i.e., occasions); u^ is the number of untagged that were counted during the survey; and m^^ is the number of prespawning season crabs recaptured and m.,, is the number of prepeak-spawning period crabs recaptured during the survey. Number untagged along the survey portion of the beach (;/^ 1 was estimated from quadrat counts, i.e., [number untagged on beach = beach length (mix mean number per quadrat ino.lm-- ) - number recaptured on beach]. Survey portion of the beach was less than 1 km. Sex Location s Survey date M, M.-, Along beach Quadrats u. '"i. "'2s " '"i. m.,^ Males Baywide 1 29 May 5398 4828 55,750 3 4 6846 1 2 2 31 Mav 5398 7091 48.498 10 6 5117 2 1 3 2 June 5398 7091 96,718 10 12 10,088 4 8 New Jersey 1 29 May 1448 3703 25,123 0 4 3890 0 2 2 31 May 1448 5254 16.376 3 6 2113 1 1 3 2 June 1448 5254 29,630 0 9 3524 0 6 Delaware 1 29 May 3950 1132 30,631 2 0 2956 1 0 2 31 May 39.50 1837 32,129 7 0 3004 1 0 3 2 June 39.50 1837 67.099 10 3 6564 4 2 Females Baywide 1 29 May 1823 2143 18,224 0 0 2149 0 0 2 31 May 1823 3231 17,022 1 1 1712 0 0 3 2 June 1823 3231 27,339 0 1 2814 0 1 New Jersey 1 29 May 488 1567 8021 0 0 1169 0 0 2 31 May 488 2268 3993 0 1 532 0 0 3 2 June 488 2268 6859 0 0 809 0 0 Delaware 1 29 May 1335 576 10,202 0 0 980 0 0 2 31 May 1335 963 13,029 1 0 1180 0 0 3 2 June 1335 963 20,480 0 1 2005 0 1 Table 2 Comparison between three likelihood-based models for mark and recapture of horseshoe crabs tLimulus polyphemus). Each model allowed a differed pattern of variation in recapture probabilities. The models are listed in order from least to most complex. In model 1, the recapture probabilities were set to be constant for both release cohorts (prespawning season period and prepeak- spawning period) and for all three recapture occasions. Model 2 allowed recapture probabilities to be time-specific, but equal for release cohorts. Model 3 allowed recapture to be time and cohort specific. Lower values for Akaikie information criteria (AIC) and /lAIC indicate a better model fit; 4AIC = 0 is the best fitting model. Likelihood ratio test (LRT) compares two nested models, and a significant P-value indicates the more complex model is supported. Likelihood ratio test Model Variation in recapture probabilities AIC 4AIC Models compared P-value Constant Time-specific Time and cohort-specific -348512.08 -350225.72 -350220.83 1713 0 5 Models 1 and 2 Models 2 and 3 <0.0001 0.77 fects the accuracy of abundance estimates. Therefore, it was important to evaluate thoroughly whether tagged animals had different catchability than untagged ani- mals. We tested for a temporary delay in spawning by comparing relative risk of recapture among the two release periods and by fitting a likelihood model that incorporated separate recapture probabilities for each release cohort (Table 2). Evidence did not support the hypothesis that initial capture and tagging temporarily affected spawning behavior (x2=2.03, df =2, P=0.36). The difference between the two release periods in the risk of being recaptured during peak spawning was Smith et al : Abundance of Limu/us polyphemus in Delaware Bay 461 Table 3 Maximum likelihood estimates of abundance for adult horseshoe crabs iLimultis pnlyphemus) in Delaware Bay during the end of May 2003. Estimates of females and for both sexes combined ("Total") are based on mark-recapture estimates of males and sex ratios among the animals caught and released for this study. Adjusted estimates take into account the possible effect of capture on spawning by reducing releases of males by 0.88, which is an observed relocation rate for radio-tagged males. Maximum likelihood estimates Adjusted estimates based on relocation rates from radio-tagged males Abundance 90% CI Abundance 90% CI Males Females Total 16,140,000 7,350,000 23,490,000 9,910,000-22,300,000 4,520,000-10,160,000 14,430,000-32,460,000 13,730,000 8,780,000-19,400,000 6,250,000 4,000,000-8,840,000 19,980,000 12,7800,000-28,240,000 0.001 (909; CI: -0.0007 to 0.003). A model comparison test (likelihood ratio test in Table 2) did not support co- hort-specific recapture probabilities (P=0.77). Although the difference in relative risk was in the direction of a temporary delay in spawning, the difference in recap- ture probabilities tended to be in the opposite direction. On two out of the three recapture occasions, the prepeak release cohort had a higher recapture probability than the preseason cohort. The maximum likelihood estimate of abundance of males with the use of model 2 (Table 2) was 20% higher for all releases (16.1 million crabs) than for prepeak releases only (13.4 million). A tempo- rary delay in spawning would tend to cause abundance estimates based on late-releases to be higher than esti- mates including early releases, but in fact the opposite occurred. Thus, based on three lines of evidence, it is unlikely that there was a temporary delay in spawning due to the capture and tagging of males. Another way in which tagged animals may behave differently from untagged animals is that tagged ani- mals may forego spawning altogether. In 2004, we radio tagged horseshoe crabs using the same capture process that was used to tag animals in 2003. The radio-tagged horseshoe crabs were detected by radio receivers at high tide when the radio-tagged crabs emerged from the water to spawn. An array of fixed station receivers ensured nearly complete coverage of spawning habi- tat in Delaware Bay. Among radio-tagged males, we observed that age-specific relocation rates were 0.44, 1.00, and 0.74 for young, middle, and old-aged males, respectively. The frequencies of these age groups in the end of May 2003 tag releases were 0.07, 0.62, 0.31 for young, middle, and old-aged males. Thus, the average relocation rate predicted for the 2003 releases would be 0.88 (i.e., 0.44x0.07-t-lx0.62-h0.74x0. 31=0.88). There are several reasons for a failure to detect radio-tagged ani- mals, namely behavioral response, movement beyond the range of radio receivers, transmitter loss or failure, and animal mortality. It is also possible that some adults migrate but do not spawn in a given year. Nevertheless, to be conservative, we adjusted the tag releases by the observed relocation rate (0.88) and computed estimates using the reduced releases. The 12% reduction in re- leases resulted in a 15% reduction in the abundance estimates (Table 3). From currently available estimates of fecundity (88,000 eggs per female; Shuster and Bot- ton, 1985), we estimated that egg production in 2003 was 5.5x1011 (90% CI: 3.5x10" to 7.7xl0ii). The adjusted estimates of abundance were 13.7 mil- lion (90% CI: 8.8 to 19.4 million) for males and 6.25 million (90% CI: 4.0-8.8 million) for females (Table 3). Landings in New Jersey and Delaware during 2003 were 2.4% (90% CI: 2-4%) of abundance estimates (Table 4). When landings from Virginia and Maryland are included, landings during 2003 were 4% (90% CI: 3-6%) of abundance. Harvest rates were similar for males and females because the sex ratio in the landings was similar to the ratio observed in our fishery-inde- pendent catch. We believe that our fishery-independent catch was a representative sample of mature animals in the bay. We caught horseshoe crabs prior to and during the spawning migration. Thus, a comparison of pre- and postmigration catches could indicate the proportion of the population that over-wintered in the bay. We caught, on average, 18 adults per 15-min. tow on a ves- sel with two 2.3-m dredges from 25 March to 3 April 2003 prior to the spawning migration, which appeared to begin in mid-April. The catch-per-tow was 39 adults per tow during the period from 13 April to 8 May 2003 and 60 adults per tow during 28 to 30 May 2003. Thus, we were catching approximately one third (18/60 = 0.3) of the animals prior to spawning migration; this frac- tion could represent the proportion of the population that over-wintered in the bay and that did not migrate to the ocean between 2002 and 2003. Discussion There have been few attempts to estimate abundance of adult horseshoe crabs in Delaware Bay during the spawning run when the population is spatially concen- trated. Shuster and Botton (1985) estimated population size from surveys on Delaware Bay beaches. However, a large portion of the bay was not included in the target 462 Fishery Bulletin 104(3) Harvest rates calculated from 2003 landings ware Bay at the end of May 2003. To be consei radio-tagged male animals (see Table 3). NJ= Table 4 and abundance estimates of adult horseshoe crabs ^Limulus polyphemus) in Dela- ■vative, estimates were adjusted according to the observed relocation rate (0.88) for New Jersey: DE=Delaware; VA=Virginia; and MD=Maryland. Delaware Bay iNJandDE) Delaware Bay Region (NJ,DE,VA,MD) Males Females Total Total Landings' Abundance Harvest rate 318,400 13,700,000 8.8 to 19 mil 0.023 0.02 to 0.04 151.900 6,250,000 4.0 to 8.8 mil 0.024 0.02 to 0.04 470,300 19,980,000 12.8 to 28 mil 0,024 0.02 to 0.04 745,800 19,980,000 12.8 to 28 mil 0.04 0.03 to 0.06 ' ASMFC (Atlantic States Marine Fisheries Commission). 2004. 2003 Review polyphemus), 13 p. ASMFC, 1444 Eye Street, NW, Sixth Floor, Washington, D.C of the fisher.N . 20005 management pi an for horseshoe crab fLimulus population study and was excluded from the survey. Carmichael et al. (2003) used transects located on a grid and visual counts in a shallow clear-water estuary to estimate abundance of the population in Pleasant Bay, Cape Cod, Massachusetts. Other estimates of the Delaware Bay population have been based on offshore surveys during nonspawning periods when populations are dispersed and possibly mixed (Botton and Haskin, 1984; Botton and Ropes, 1987; Hata and Berkson, 2003), Our mark-recapture estimates apply to adult horseshoe crabs present in Delaware Bay during late May 2003 when spawning peaked. In previous horseshoe crab population estimates that were based on offshore surveys (Botton and Haskin, 1984; Botton and Ropes, 1987; Hata and Berkson, 2003), capture efficiency was unknown, and adults that remained in estuaries (i,e., those that did not migrate to the ocean after spawning) were not sampled. Hata and Berkson (2003) concluded that the capture efficiency for their trawl survey was intermediate between that of the trawl survey reported in Botton and Haskin (1984) and that of the hydraulic dredge survey reported in Botton and Ropes (1987). The most recent estimate of offshore abundance of 7.1 million crabs was reported by Hata and Berkson (2003). Botton and Haskin (1984) reported densities that were 1.8 to 2.9 times the densities report- ed by Hata and Berkson (2003), which would indicate a population estimate of 12 to 20 million according to the data of Botton and Haskin (1984). Botton and Ropes (1987) reported a minimum population of 2.3 to 4.1 mil- lion. The proportion of adults that remain in the Dela- ware Bay and do not migrate to the continental shelf after spawning is unknown. However, if the proportion of nonmigratory adults is sizeable (e.g., on the order of 0.3, which was indicated by our fishery-independent catches) then that, along with gear inefficiencies in trawl surveys, could explain the difference between the estimates in our present study and those of Hata and Berkson (2003). Bias due to assumption violation is another reason for differences in estimates. We designed our study to obtain a representative sample when horseshoe crabs were concentrated and to minimize the time between release and recapture periods so that only a few days separated the prepeak release and recapture occasions. We released 17,543 tagged horseshoe crabs over two periods and examined 28,738 horseshoe crabs for tags, counting them in quadrats during subsequent spawn- ing surveys (29 May, 31 May, and 2 June 2003). Low recapture rates (<1%) were consistent with a large population. However, it was important to evaluate the potential effects of assumption violations. Tag loss and tag-induced mortality were likely to be trivial because of the short period during which crabs were at large and results from field studies and laboratory experi- ments have shown no tag loss or tag-induced mortality (Crawford'^; Brousseau et al., 2004). Brousseau et al. (2004) attached combined acoustic and radio tags and standard button tags to 24 female horseshoe crabs along two beaches in Delaware Bay, and then tracked them for eight days. All 24 were detected at least once, and 20 spawned on the beach of release within eight days of release, indicating that handling and tagging had a minimal effect on spawning behavior. Tags could have been overlooked during the spawning survey, which is why we limited our analysis to males counted and recaptured within 1-m- quadrats. We evaluated the effect of capture and tagging on spawning behavior, and found no evidence that tagged males delayed spawning. Using radio-tagged horseshoe crabs, we believe it is possible that tagged crabs termi- nated spawning, and we adjusted abundance estimates based on relocation rates of radio-tagged crabs. Thus, we based our inference on abundance estimates that were adjusted downward to account for that possibility. Recapture rates were low over the short time period for our study. However, annual recapture rate was ap- proximately 4% for all tag releases and recaptures. Smith et al : Abundance of Limulus polyphemus in Delaware Bay 463 Mark-recapture studies have been frequently applied to marine species for describing migration or estimating mortality (Hoenig et al, 1998; Bacheler et al., 2005). Mark-recapture methods are used infrequently for abun- dance estimates of marine species, with the exception of anadromous species, whose spawning migration con- centrates the population and enhances opportunities for recapture (Schwarz and Taylor, 1998). Similarly, the unusual horseshoe crab spawning migration and behavior concentrated the population and made them assessable for recapture. The validity of our abundance estimates is founded on a large number of tag releas- es and animals checked for tags, a study design that ensured population closure, an adequate number of samples to represent the population, and an evaluation of the underlying assumptions. The sex ratio in our fishery-independent catch (69% M: 31% F), which we believe is a representative sample of adults in the bay at the time of peak spawning, was similar to the sex ratio in the landings (68% M: 32% F). Hata and Berkson (2003) observed a similar sex ratio among adults in an offshore trawl survey (63% M: 37% F). Although, commercial landings in 2003 were not skewed toward the harvest of females, harvest could have selected females disproportionately in past years. The horseshoe crab harvest has been reduced through a series of reductions mandated by ASMFC (ASMFC^). Although there is evidence of stock decline coinci- dent with increased landings in the past 10-20 years (ASMFC"*), the estimates presented in the present study indicate recent regulatory changes had achieved a low harvest level by 2003. Based on the abundance esti- mates reported here, harvest rate in 2003 was 0.024 (90% CI: 0.02 to 0.04) for Delaware Bay state landings and 0.04 (90% CI: 0.03 to 0.06) for Delaware Bay area landings. In 2004, additional regulations were enacted, which capped landings at 150,000 per state for Delaware and New Jersey and prohibited harvest during May and early June when migrant shorebirds stopover in Dela- ware Bay. As a result of the 2004 regulations, landings dropped to 173,023 (males and females) for Delaware and New Jersey combined, which is a 63% drop from 2003 landings. Sex ratio of the 2004 landings was 68% M: 32% F, consistent with the 2003 landings. Estimating abundance is an important step in the process of determining the current capacity for horse- shoe crab egg production in the bay and for managing for the energetic needs of shorebirds. We estimate that egg production in 2003 was 5.5x10" (90% CI: 3.5x10" to 7.7x10"). The U. S. Fish and Wildlife Service Shore- bird Technical Committee (USEWS^) estimated that a •• ASMFC (Atlantic States Marine Fisheries Commission). 2004. Horseshoe crab 2004 stock assessment report, 87 p. ASMFC, 1444 Eye Street, NW, Sixth Floor, Wash- ington, DC 20005. = USFWS (United States Fish and Wildlife Service). 2003. Del- aware Bay shorebird-horseshoe crab assessment report and peer review. U.S. Fish and Wildlife Service Migratory Bird Publication R9-03/02, 99 p. USFWS, 4401 N. Fairfax Drive, Arlington, VA 22203. population of 423,000 shorebirds would require 1.07x10" horseshoe crab eggs as they migrated through Delaware Bay. This represents approximately 20% of all egg pro- duction, which would have to be available temporally and spatially during the shorebird migration. Egg availability for shorebirds will depend also on overlap between horseshoe crab and shorebird migra- tions, density-dependent bioturbation, and wave-medi- ated vertical transport. Some important aspects of egg production and the process of making eggs available to foraging shorebirds are not well understood and have not been quantified. For example, if fecundity is found to be age- or size-related, then age or size would need to be incorporated in a calculation of egg abundance. Also, the process of bioturbation, which releases buried eggs to the beach surface, is known to be related to the density of spawning females, but has not been parameterized (Jackson et al., 2002). Determination of the dietary re- quirements of migrant shorebirds in terms of horseshoe crab eggs coupled with reliable estimates of abundance, fecundity, and bioturbation rates, will set the stage for a management of horseshoe crabs that takes into account the trophic support it provides in Delaware Bay. Acknowledgments We thank many individuals involved in the tagging study and the Delaware Bay spawning survey. Principal among them were Stewart Michels, Lome Brousseau, Mike Oates, Bruce Freeman, Carl Shuster, Susan Love, staff from the DE Coastal Program and the Delaware National Estuarine Research Reserve (DNERR), Benjie Swan, Bill Hall, Glenn Gauvry, Sherry Bennett, and representatives from Audubon, The Nature Conservancy, the Slaughter Beach folks, and the USFWS. Fund- ing came from the USGS, NJ FG&W, and Delaware Department of Natural Resources and Environmental Control (DNREC). We appreciate comments from Stew- art Michels, Greg Breese, and John Young on an early draft of the paper. Literature cited Bacheler, N. M., R. A. Wong, and J. A. Buckel. 2005. Movements and mortality rates of striped mullet in North Carolina. N. Am. J. Fish. Manag. 25: 361-373. Baker, A. J., P. M. Gonzalez, T. Piersma, L. J. Niles, I. de Lima Serrano do Nascimento, P. W. Atlkinson, N. A. Clark, C. D. T. Minton, M. K. Peck, and G. Aarts. 2004. Rapid population decline in red knots: fitness con- sequences of decreased refueling rates and late arrival in Delaware Bay. Proc. Royal Soc. B, 271:875-882. Borchers, D. L., S. T. Buckland, and W. Zucchini. 2002. Estimating animal abundance: closed populations, 314 p. Springer, London. Botton, M. L., and B. A. Harrington. 2003. Synchronies in migration: shorebirds, horseshoe crabs, and Delaware Bay. In The American horseshoe 464 Fishery Bulletin 104(3) crab (C. N. Shuster Jr., R. B. Barlow, and H. J. Brockmann, eds.), p. 5-32. Harvard Univ. Press, Cambridge, MA. Botton, M. L., and H. H. Haskin. 1984. Distribution and feeding of the horseshoe crab, Limulus polyphemus, on the continental shelf off New Jersey. Fish. Bull. 82:383-389. Botton, M. L., and J. W. Ropes. 1987. Populations of horseshoe crabs, Limulus polyphemus, on the northwestern Atlantic continental shelf. Fish. Bull. 85:805-812. Brousseau, L. J., M. Sclafani, D. R. Smith, and D. B. Carter. 2004. Acoustic and radio tracking horseshoe crabs (Limu- lus polyphemus) to assess spawning behavior and sub- tidal habitat use in Delaware Bay. N. Am. J. Fish. Manag. 24:1376-1384. Burnham, K. P., and D. R. Anderson. 1998. Model selection and inference: a practical informa- tion-theoretic approach, 353 p. Springer, New York, NY. Carmichael, R. H., D. Rutecki, and I. Valiela. 2003. Abundance and population structure of the Atlantic horseshoe crab, Limulus polyphemus, in Pleasant Bay, Cape Cod. Mar. Ecol. Prog. Ser. 246:225-239. Hata, D.. and J. Berkson. 2003. Abundance of horseshoe crabs (Limulus polyphemus) in the Delaware Bay area. Fish. Bull. 101:933-938. Hoenig, J. M., N. J. Barrowman, W. S. Hearn, and K. H. Pollock. 1998. Multiyear tagging studies incorporating fishing effort data. Can. J. Fish. Aquat. Sci. 55:1466-1476. Jackson, N. L., K. F. Nordstrom, and D. R. Smith. 2002. Geomorphic-biotic interactions on beach foreshores in estuaries. Proceedings of the International Coastal Symposium (ICS 2002 1. J. Coastal Res. Spec. Issue 36:414-424. 2005. Influence of waves and horseshoe crab spawning on beach morphology and sediment characteristics on a sandy estuarine beach. Delaware Bay, New Jersey, USA. Sedimentology 52:1097-1108. Odell, J., M. E. Mather, and R. M. Muth. 2005. A biosocial approach for analyzing environmental conflicts: a case study of horseshoe crab allocation. Bio- science 55:735-748. Schwarz, C. J., and C. G. Taylor. 1998. Use of the stratified-Petersen estimator in fish- eries management: estimating the number of pink salmon (Oncnrhyncus gorbuscha) spawners in the Fraser River. Can. J. Fish. Aquat. Sci. 55:281-296. Shuster, C. N., Jr., and M. L. Botton. 1985. A contribution to the population biology of horseshoe crabs, Limulus polyphemus, in Delaware Bay. Estuaries 8:363-372. Shuster, C. N., Jr., and K. Sekiguchi. 2003. Growing up takes about ten years and eighteen stages. In The American horseshoe crab (C. N. Shus- ter Jr., R. B. Barlow, and H. J. Brockmann, eds.), p. 103-132. Harvard Univ. Press, Cambridge, MA. Skalski, J. R., and D. S. Robson. 1992. Techniques for wildlife investigations: design and analysis of capture data. Academic Press, Inc., New York, NY. Smith, D. R., P S. Pooler, B. L. Swan, S. Michels, W. R. Hall, P. Himchak, and M. Millard. 2002. Spatial and temporal distribution of horseshoe crab (Limulus polyphemus) spawning in Delaware Bay: impli- cations for monitoring. Estuaries 25(1):115-125. 465 Estimates of commercial longline selectivity for Pacific halibut (Hippoglossus stenolepis) from multiple marking experiments William G. Clark Stephen M. Kaimmer International Pacific Halibut Commission 250 Ocean Teaching Building University of Wasfiington Seattle, Washington 98145 E-mail address (for W G Clark): bill(a/iphc Washington edu The term "selectivity" refers to the relationship between the size (or age) of a fish and its vulnerability to a given kind of fishing gear. A selectiv- ity schedule, along with other param- eters, is normally estimated in the course of fitting a stock assessment model, and the estimated schedule can have a large effect on both the estimate of present stock abundance and the choice of an appropriate har- vest rate. The form of the relationship is usually not known and not well determined by the data, and equally good model fits can often be obtained with different plausible specifications of selectivity. Choosing among the model fits and associated abundance estimates in this situation is prob- lematic (Sigler, 1999; Sullivan et al., 1999). The selectivities of different gears can be compared by fishing the gears side by side, but without knowing the size composition of the stock being fished, it is impossible to determine the form of the selectivity functions. Therefore, one has to make some as- sumptions about them in order to locate estimates (Millar and Fryer, 1999, and references therein). In this case, too, equally good fits can often be obtained with a variety of assumed forms (Huse et al., 2000; Woll et al., 2001); therefore the true form cannot be determined by simple fishing experiments. Mark-recapture data can yield di- rect and reliable estimates of selec- tivity because in this situation the size composition of the fished stock is known (Myers and Hoenig, 1997). In this note, we report estimates of the commercial longline selectivity of Pa- cific halibut (Hippoglossus stenolepis) based on the large number of mark- recapture experiments conducted by the International Pacific Halibut Commission (IPHC) in the 1960s, 1970s, and 1980s. A similar analy- sis was done by Myhre (1969), but he used data from only two experi- ments; the present study uses data from more than 100 experiments. Materials and methods Kaimmer (2000) described all IPHC tag data for all varieties of external tags in setline and trawl catches dating back to 1925. We also used tag data for all varieties of tags (except the small strap type); however, our data were for tags released during setline catches only and our data dated back to 1960, the first year of recorded data in the computer release and recovery IPHC database. The total number of tags released was over 100,000, of which more than 13,000 were recovered in the commer- cial longline fishery. About half of the releases were at systematically placed setline survey stations that covered a large part of an IPHC regulatory area (Fig. 1). The other half were at "spot" fishing locations, deliberately chosen to produce good catches, either for marking or for gathering data on the performance of different gear types. For our study, an experiment was defined as all releases of a given tag type in a given regulatory area in a given year during either survey or spot fishing operations (not both), where at least 10 fish were released. Between 1960 and 1990 there were 131 such experiments. These data are not usable for esti- mating exploitation rates or migra- tion rates because of uncertainty concerning things like recovery ef- fort and reporting rates, but they can be used to estimate commercial selectivity. In the case of a single ex- periment, a straightforward plot of short-term recovery rate by length at release will show how selectivity changes with length. The absolute recovery rates will depend on usually unknown factors (tagging, fishing, and natural mortality rates; tag loss and reporting rates), but the relative recovery rates should depend mainly on selectivity (barring large varia- tions in length with any of the un- known factors). Myers and Hoenig (1997) showed how data from many experiments can be combined to obtain a single set of selectivity estimates. To sum- marize their derivation, let n^ , be the recovery rate of fish of length / in experiment (. This rate is treated as the product of a length-specific com- mercial selectivity Sj, which is the same for all experiments, and an experiment-specific recovery rate r, that combines all the unknown fac- tors mentioned above. Thus n^j=r^-Sj and log;r, /=log7-^-i-logs,. This has the form of a generalized linear model with a log link function and a bino- mial variance; therefore the point and variance estimates can be ob- tained in standard fashion. Some rule has to be chosen for scaling the selectivities to make the model determinate. The most com- mon rule is to require that the maxi- mum selectivity be 1.0, but that can involve using a scaling factor that is poorly determined by the data if the maximum occurs in a length group with few releases and recoveries. To avoid this problem, the rule used in Manuscript submitted 10 March 2005 to the Scientific Editor's Office. Manuscript approved for publication 27 September 2005 by the Scientific Editor. Fish. Bull. 104:465-467 (2006). 466 Fishery Bulletin 104(3) 170'E 180 65'N 60 N Russia 1 70 W H- 55'N 50 N 45"N Bering Sea 4D .4C 160 W . . y 0 150 W Alaska 140W 130 W N I .^i.^" Closed -rV-' S -^>- 4B Aleutian Is- 4B 4A ^Y^""^ 3B .--.J!^ 4A Gulf of Alaska Queen Charlolte Is. xt -X^ *<* 28 4 ^^ Vancouver Is %. J" ^^'^ ' ^ 2A ■■ 60 N 55'N 50 N 45N 170°E 180 uo^w 130*W 120"W 170'W 160W 150'W Figure 1 International Pacific Halibut Commission regulatory areas. The area marked "closed" is permanently closed to directed halibut fishing. our study was to define selectivity to be 1.0 at 120 cm. Estimated selectivity could therefore exceed 1.0 at other lengths. Results Figure 2 shows the estimates of commercial length-spe- cific selectivity in areas 2B, 2C, 3A, and 3B obtained by the method of Myers and Hoenig (1997) with the use of all available data in each area. There were insufficient data in area 4 to calculate useful estimates. The esti- mates in Figure 2 were calculated by using all recoveries from each release during the first two years at liberty, including recoveries from outside the release area and recoveries from unknown locations. Estimates computed by using only recoveries from the area of release were no different from those obtained by using all of the recoveries. In all areas, commercial selectivity in the period of 1960-90 anpears to increase with length up to a maxi- mum and then decline. In area 2B, the peak occurs at about 110 cm and there is a substantial decline thereaf- ter, to around half the peak value. In Alaska (areas 2C, 3A, 3B), selectivity peaks at a much larger size (about 150 cm). Thereafter the decline is about as steep as in area 2B, but not as large because so little of the length composition remains beyond 150 cm. Recoveries from releases at spot fishing locations show a selectivity pattern similar to that for the entire dataset. The same is true of survey releases, except in area 2B where the selectivity pattern does not show a decline among larger fish. But this impression depends on a small number of recoveries, and therefore it may be false. Discussion In previous modeling of length-specific selectivity, the IPHC staff generally assumed some kind of asymp- totic function, with full selection occurring at 110-130 cm. A function of this form is consistent with video observations of halibut behavior when they are hooked (Kaimmer, 1999), and it produces satisfactory fits to the observed length compositions of survey and commercial setline catches in the annual stock assessment. It is also consistent with the conventional view that hook selectivity varies little with size among fish large enough to take the bait (Lokkeborg and Bjordal, 1992). But the large body of mark-recapture data shows a different pattern: selectivity declining after the peak at 110 cm in area 2B, and not reaching a peak until 150 cm or so in Alaska. These patterns are, in fact, quite similar to those reported by Myhre (1969). Commercial fishing selectivity reflects ground selec- tion by the fleet, as well as size selection by the gear. It is therefore possible that the selectivity of commer- cial setline gear on a given ground has the expected asymptotic form, but ground selection has the effect of NOTE Clarks and Kaimmer: Estimates of commercial longllne gear selectivity for Hippoglossus stenolepis 467 targeting certain size groups and thereby producing a different selec- tivity schedule. In area 2B, for ex- ample, the best catch rates may be achieved by targeting smaller fish; whereas in Alaska it may be more profitable to target larger fish. If the decline in selectivity in area 2B were the result of the commer- cial fishery targeting areas where the fish are smaller, one would ex- pect to see the decline in data from survey releases (which are done over the whole area), but not in data from releases at spot fishing loca- tions (most of which are customary commercial fishing locations). How- ever, the mark-recapture data show the opposite pattern, if anything; therefore ground selection does not appear to be the explanation. When length-specific selectivity is allowed to be dome-shaped in the stock assessment model (rather than forced to be asymptotic), the esti- mated commercial selectivities turn out to be quite similar in pattern to the mark-recapture estimates, including the differences among ar- eas. But the selectivities estimated for the IPHC systematic setline sur- vey are asymptotic or ramp-shaped, rather than dome-shaped. They in- dicate that ground selection by the commercial fishery really does have an effect on the form of commercial selectivity, contrary to what the mark-recapture data may indicate. Literature cited ) I 80 100 120 140 160 180 Fork lengtti (cm) Area 2B (18589 releases, 2760 recoveries) 80 100 120 140 160 180 Fork lengtti (cm) Area 3A (33952 releases, 2479 recoveries) o l.t 80 100 120 140 160 Fork lengtti (cm) ] \ 1 1 — 80 100 120 140 160 Fork length (cm) Area 2C (15220 releases, 1422 recoveries) Area 3B (9399 releases. 464 recoveries) Figure 2 Estimates of length-specific commercial selectivity (±1 standard deviation) for Pacific halibut ^Hippoglossus stenolepis) based on all releases 1960-90, by regulatory area. The scale was set by defining selectivity to be 1.0 at 120 cm, so that value has no standard deviation, and other values can and do exceed 1.0. Huse, I., S. Lokkeborg, and A. V. Soldal. 2000. Relative selectivity in trawl, longline and gilnet fisheries for cod and haddock. ICES J. Mar. Sci. 57:1271-1282. Kaimmer, S. M. 1999. Direct observations on the hooking behavior of Pacific halibut, Hippoglossus stenolepis. Fish. Bull. 97:873-883. 2000. Pacific halibut tag release programs and tag release and recovery data, 1925 through 1998. Int. Pac. Halibut Comm. Tech. Rep. 41, 31 p. Lokkeborg, S., and A. Bjordal. 1992. Species and size selectivity in longline fishing: a review. Fish. Res. 13:311-322. Millar, R. B., and R. J. Fryer. 1999. Estimating the size-selection curves of towed gears, traps, nets, and hooks. Rev. Fish. Biol. Fish. 9:89-116. Myers, R. A., and J. M. Hoenig. 1997. Direct estimates of gear selectivity from mul- tiple tagging experiments. Can. J. Fish. Aquat. Sci. 54:1-9. Myhre, R. J. 1969. Gear selection and Pacific halibut. Int. Pac. Hali- but Comm. Tech. Rep. 51, 35 p. Sigler, M. 1999. Estimation of sablefish, Anop/opoma fimbria, abun- dance off Alaska with an age-structured population model. Fish. Bull. 97:591-603. Sullivan, P. J., A. M. Parma, and W. G. Clark. 1999. The Pacific halibut stock assessment of 1997. Int. Pac. Halibut Comm. Sci. Rep. 79. Woll, A., J. Boje, R. Hoist, and A. C. Gundersen. 2001. Catch rates and hook and bait selectivity in long- line fishery for Greenland halibut {Reinhardtius hip- poglossoides. Walbaumi at East Greenland. Fish. Res. 51:237-246. 468 Hatching date, nursery grounds, and early growth of juvenile walleye pollock (Theragra chalcogramma) off northern Japan* Tsutomu HattorP Akira Nishimura^ Yoji Narimatsu' Daiji Kitagawa' ' Hachinohe Branch Tohoku National Fisheries Research Institute 25-259 Shimomekurakubo, Same, Hachinohe Aomori 031-0841, Japan E-mail address (lor T Hattori) hmadara afra affrc go ip 2 Hokkaido National Fisheries Research Institute 116 Katsurakoi, Kushiro Hokkaido 085-0802, Japan Walleye pollock (Theragra chalco- gramma) is widely distributed in the North Pacific Ocean and plays an important role in coastal subarc- tic ecosystems. The Japanese Pacific population of this species is one of the most important demersal fishes for commercial fisheries in northern Japan. The population is distributed along the Pacific coast of Hokkaido and the Tohoku area ■-~^ 08 ^ r^^^^ - i'\ P~J Hi° -J ^ W' \o^= -) \013 1 )q14 JD15 - rfCl6 JO'7 I (27 CD 20 19 V28(r)21 O22 Si! - \ > r 1 1 141 142 143 140 141 142 143 140 141 142 143 140 141 142 143 Figure 1 Sampling stations (open circles) for walleye pollock tTheragra chalcogramma) caught with a midwater trawl net in four regions off the Pacific coast of northern Japan in May 1999-2001. Arabic numerals indicate station numbers (see Table 1). TL of fresh fish = 1.0126(rL of preserved fish) + 1.7934 [n = 97, r2=0.974]. To study the daily otolith increment of our sample, we employed the methods established by Nishimura and Yamada (1984), Bailey and Stehr (1988), Yoklavich and Bailey (1990) and Bailey et al. (1996). Sagittal otoliths, hereafter called "otoliths," from the juveniles were em- bedded in epoxy resin on glass slides and ground with carbon paper (no. 800-1200) perpendicularly to the oto- lith flat plane along the long axis (frontal section). Pol- ished otoliths were selected if they included the otolith nucleus on the ground surface and were thin enough to transmit light under a light microscope. Then 1171 specimens were polished with lapping film, cleaned, and observed. Early narrow increments were counted under 600x or lOOOx magnification. As the increments increased in width, the magnification was changed to 400x. Hatching date was estimated by subtracting the number of increments from the sampling date. Daily growth-ring diameter is the index of body length at an age because a significant linear relation- ship exists between log-transformed body length and otolith length data for pollock juveniles (Nishimura and Yamada, 1988). In the present study, we measured the short otolith radius (SOR) along the long axis ob- tained from a frontal section and used the SOR from the nucleus to the 30th ring (SOR 30) as the index of early growth. SOR 30 was measured for the specimens collected in 2000, by using the otolith daily growth ring analyzing system (Ratoc System Engineering Co., Ltd., Shinjuku, Tokyo). The Mann-Whitney t/-test was used to identify differences in SOR 30 groups. We were not able to use analysis of covariance (ANCOVA) because an interaction was found by analysis of variance (ANO- VA) in size-at-age relationships among regions in 2000 (P<0.0001) and 2001 (P=0.035). Comparisons of specific growth rate (SGR) for the juveniles were made between our regions by the Mann-Whitney {/-test or ANOVA. Comparisons of means were made using Scheffe's test. SGR values of individual fish were calculated by using the following formula (Ricker, 1979): ^^^_(hiL2-hiLi)xl00 Co t^ 470 Fishery Bulletin 104(3) Table 1 Summary of midwater trawl surveys in May of 1999-2001 SST=sea surface temperature. OB= oblique tow, HZ =horizontal tow. Station no Sampling date Depth (m) SSTCC) Towing time ( min ) Method Number offish CPUE(fish/h) 1 19 May 1999 69 7.2 63 HZ 6 5.7 2 19 May 1999 69 7.2 70 HZ 129 110.6 3 21 May 1999 99 10.6 72 HZ 4 3.3 5 21 May 1999 108 10.7 84 HZ 19 13.6 6 23 May 1999 97 11.2 88 HZ 0 0 7 23 May 1999 124 11.8 74 HZ 0 0 8 23 May 1999 136 11.6 93 HZ 0 0 9 22 May 1999 106 14.7 98 HZ 0 0 10 22 May 1999 102 16.4 84 HZ 0 0 11 22 May 1999 100 17.1 49 HZ 0 0 1 20 May 2000 84 8.4 45 OB 730 973.3 7 21 May 2000 101 8.3 57 OB 173 182.1 8 21 May 2000 111 9.6 48 OB 399 498.8 9 21 May 2000 90 10.1 42 OB 0 0 10 22 May 2000 123 10.8 50 OB 6 7.2 11 22 May 2000 111 9.2 51 OB 17 20.0 12 22 May 2000 107 10.9 54 OB 69 76.7 13 25 May 2000 121 11.3 62 OB 90 87.1 14 25 May 2000 151 10.9 48 OB 14 17.5 15 25 May 2000 138 10.9 43 OB 163 227.4 16 26 May 2000 122 11.0 49 OB 57 69.8 17 26 May 2000 133 13.1 48 OB 81 101.3 18 26 May 2000 143 13.7 51 OB 24 28.2 19 26 May 2000 121 14.3 51 OB 22 25.9 20 27 May 2000 104 12.5 48 OB 62 77.5 21 27 May 2000 122 13.7 48 OB 37 46.3 22 27 May 2000 130 12.8 57 OB 29 30.5 23 27 May 2000 97 12.5 44 OB 37 50.5 24 27 May 2000 105 12.9 51 OB 11 12.9 25 28 May 2000 77 9.7 60 OB 11 11.0 26 28 May 2000 102 9.9 50 OB 35 42.0 27 29 May 2000 69 13.3 47 OB 144 183.8 28 29 May 2000 82 12.6 45 OB 15 20.0 7 20 May 2001 71 10.2 55 OB 51 55.6 10 20 May 2001 103 7.6 39 OB 2371 3647.7 11 22 May 2001 117 10.4 48 OB 1 1.3 12 22 May 2001 90 11.0 40 OB 0 0 13 22 May 2001 176 10.5 48 OB 10 12.5 14 23 May 2001 109 10.3 51 OB 11 12.9 15 23 May 2001 104 10.6 52 OB 106 122.3 16 23 May 2001 120 13.1 45 OB 117 156.0 17 24 May 2001 141 11.7 60 OB 19 19.0 18 24 May 2001 147 11.5 44 OB 6 8.2 19 24 May 2001 108 10.7 23 OB 2 5.2 20 25 May 2001 130 10.5 31 OB 0 0 21 25 May 2001 115 11.2 23 OB 2 5.2 22 25 May 2001 112 11.6 45 OB 1 1.3 23 25 May 2001 104 12.0 22 OB 0 0 24 26 May 2001 120 13.4 42 OB 0 0 25 26 May 2001 125 12.0 57 OB 0 0 26 26 May 2001 97 12.1 35 OB 0 0 27 27 May 2001 106 15.7 35 OB 0 0 28 27 May 2001 100 19.1 52 OB 0 0 29 27 May 2001 103 19.6 21 OB 0 0 NOTE Hattori et al ; Hatching date, nursery grounds, and early growth of juvenile Theiagia cholcogramma 471 ?°N 1999 / — ^ v^^ ^ 42 ^ )> ^ 41 - J^ ^1 V o \ o 4U - \ 39 - J 38 - 3X 37 - / J 36 - 35 - 1 1 N 2000 2001 X 0 (lish/h) o -50 o 50-200 o 200-500 ) 500-1000 ^ L 000-4000 140°E 141 142 143 140 141 142 143 140 141 142 143 Figure 2 Distribution of catch per unit of effort (CPUE) (fish/hourl for walleye pollock {Theragra chalcogramma) caught with a midwater trawl net in May 1999-2001 off the Pacific coast of northern Japan. where L., and L, body length at time /,, and t^ respectively. In this study, we defined L, = 4.6 mm at hatching (t^=0) according to Nishimura and Yamada (1988) and L.,= TL of fish at the number of otolith increments it.^). All tests were evaluated for significance at the P=0.05 level. Results and discussion Distribution of juveniles In 1999, juvenile pollock were collected only from Funka Bay to northern Tohoku, but no fish were sampled in southern Tohoku (Fig. 2). In 2000, juvenile pollock were caught in abundance at almost all sampling stations: highest catch per unit of effort (CPUE) was recorded at a station in Funka Bay (973 fish/h) and many juveniles were found in northern and southern Tohoku (max: 227 and 184 fish/h, respectively). In 2001, the juvenile dis- tribution was similar to that in the previous year, but few juveniles were caught in southern Tohoku. These results show interannual variability in the distribution of juvenile walleye pollock between 1999 and 2001. Inada and Murakami (1993) suggested that the pol- lock catch and distribution in the Tohoku area are strongly affected by the coastal branch of the Oyashio Current but little information is available on the rela- tionship between oceanographic conditions and juvenile distribution and abundance in this area. The southern limit of the coastal branch characterized by the position of 5°C at 100 m depth has been regarded as an index to show the strength of the current (Ogawa, 1989). Thus we examined annual changes in the position of southern limit of the coastal branch based on water tempera- ture data (TNFRP). In April, the coastal branch of the Oyashio Current mostly stagnated in northern Tohoku in 1999 and 2001 but there were some annual varia- tions, but it went far south to southern Tohoku in 2000 (Fig. 3A). It is likely that the juvenile distribution in May of the three years from 1999 to 2001 is associated with the southern limit position of the coastal branch of the Oyashio Current in April. Kitagawa et al.'- reported the recruitment index of age-0 pollock based on bottom trawl surveys in the Tohoku area. We compared the recruitment index {X, xlO'^ fish) given by Kitagawa et al.- with the latitude of the southern limit of the coastal branch of the Oyashio Current (Y) and, as a result, a significant relationship was observed in the 1995-2001 data (F=46.653X-«oi3_ r2 = 0.960, Fig. 3B). This result may indicate that abundant recruitment is brought by TNFRI (Tohoku National Fisheries Research Institute). 1997-2002. Tohoku Block Suisan Kaiyo Renraku Kaihou, vols. 26-32. TNFRI, Fisheries Oceanography Division, 3- 27-5 Shinhama-cho, Shiogama, Miyagi 985-0001, Japan. [In Japanese.] Kitagawa, D., T. Hattori, and Y. Narimatsu. 2002. Moni- toring on the demersal fish resources in the Tohoku area. In Kaiyo Monthly 34, p.793-798. Seikai Nat. Fish. Res. Inst., 1551-8 Taira-machi, Nagasaki, Nagasaki 851-2213, Japan. [In Japanese, the title was translated by the authors.] 472 Fishery Bulletin 104(3) 43 A B Latitude of the southern limit of the coastal branch of the Oyashio Current 03 CO OJ £k fi. ^ -J CD (D o -* r>j • ♦ 1995 o 1996 A 1997 ■ 1998 "■-•.^*^ --\" -m ' \ 36 o 1999 . 2000 * 2001 February March April o so.ooo loo.ooo Recruitment index (xio^ fish) Figure 3 Monthly change in the latitude of the southern limit of the coastal branch of the Oyashio Current (A), and the relationship between recruitment index of age-0 walleye pollock (Theragra chalcogramma) in the Tohoku area and the latitude of the southern limit of the coastal branch of the Oyashio Current (B). This figure is drawn according to the water temperature data from the Fisheries Oceanography Division of the Tohoku National Fisheries Research Institute (TNFRI) (see text). The data on the recruitment index are from Kitagawa et al.^ (see general text). 1999 o W 20 148 22 1.0 4.0 ^30 W 20 2000 ,95 194 185 T f 172 ^ i T 2001 1 cr 3.0 ^"2.0 45 101 108 • f f * c,.„i,, D,, East of northern souttiern .uimauay jjugaru Strait Totioku Totioku Figure 4 Specific growth rate (SGR) for walleye pollock (Theragra chalcogramma) from four regions off the Pacific coast of northern Japan in 1999-2001. Vertical bars indi- cate standard deviation and arable numerals indicate number of samples. a strong coastal branch of the Oyashio Current in the Tohoku area. Early growth of walleye pollock There was no difference in SGR between Funka Bay and northern Tohoku in 1999 ([/-test, P=0.111, Fig. 4). Also, no regional differences were found among all regions in 2001 (ANOVA. P= 0.5261. For 2000, however, there were significant differences in SGR among all four regions (ANOVA, P<0.0001) and even between all pairs (Schef- fe's test, P<0.0001). In 2000, the SOR 30 values ranged from 30-58 fjm for the juveniles collected in Funka Bay and east of Tsugaru Strait (Fig, 5). However, two different SOR 30 groups were found for the samples from northern and southern Tohoku ([/-test, P<0.0001): one with SOR 30 values ranging from 30-60 ^tm and the other with values greater than 68 fim. Fish of the former group (low-growth group) had the same SOR 30 range as those from Funka Bay and east of Tsugaru Strait. The latter group (high-growth group) was observed only in the Tohoku area. Positive linear regressions were found in size-at-age relationships for low- and high-growth groups (P<0.0001, Fig. 6) and the difference of the SGR was sig- nificant between these two groups ([/-test, P<0.0001). These results indicate that the surveys of juvenile walleye pollock in May 2000 caught two growth groups: NOTE Hattorl et a\ Hatching date, nursery grounds, and early growth of juvenile Theragra chalcogramma 473 the low-growth group that probably spawned in Funka Bay mixed with the high-growth group that was prob- ably produced locally in the Tohoku area. This mixing of the two groups could have occurred primarily be- cause of either advection or migration. Assuming that this interpretation is correct, it would be important for local fisheries managers to take account of these variations especially if the years of extra production from the Tohoku area expand the exploitable stock. Length-frequency distribution and hatching date The length-frequency distribution and hatching date composition for juvenile pollock among regions were similar in 1999 and 2001 but were different in 2000 (Figs. 7 and 8). In 2000, large fish over 40 mm TL, designated as forming part of the low-growth group, constituted the majority of the samples from Funka Bay, and much smaller fish of the same group (<40 mm TL) were widely distributed in waters from east of Tsugaru Strait to southern Tohoku. All juveniles from Funka Bay were categorized as low-growth group fish that hatched between late December and late February, and juveniles of the low-growth group that hatched after March 1 were caught widely from east of Tsugaru Strait to southern Tohoku in 2000. These results appear to indicate that late hatching fish origi- nating from Funka Bay were transported to the Tohoku area by the strong coastal branch of the Oyashio Cur- rent after March 2000. Nakatani and Sugimoto'^ showed that only fish hatch- ing in January and February survived in Funka Bay in summer, whereas pollock usually spawn between December and March. Nakatani (1998) thus assumed that late hatching fish (af- ter early March) would not be able to con- sume enough food during their settlement stage because the abundance of copepo- dids of Pseudocalanus spp., the main food for juveniles below 30 mm TL, decreases in late May and because late hatching fish cannot eat Neocalanus plumchrus due to its large size. Therefore, he sug- gested that there are two critical periods for walleye pollock during their early lar- val and settlement stages in Funka Bay. Considering these results, we suggest that when late hatching fish were transported to the Tohoku area, their survival rates were higher than those of juveniles that remained in Funka Bay. Nakatani. T., and K. Sugimoto. 1998. Re- productive strategy of walleye pollock and the environment in the south region of Hok- kaido, Pacific. In Kaiyo Monthly special vol., p. 182-186. Graduate School of Fish- eries Sciences, Hokkaido University, 3-1-1 Minato-cho, Hakodate, Hokkaido 041-8611, Japan. |In Japanese, the title was translated by the authors.) 50 10 40 30 J Funka Bay 20 A 10 A 0 40 ^^t- 30 East of Tsugaru Strait 20 ,2? 10 wt^k o 0 ^^^^^^^^^t 5 40 1 30 northern Tohoku z 20 _ 10 ^^^^J _■ 0 40 30 southern Tohoku 20 10 1 ^d 0 <30 <40 <50 <60 <70 <80 <90 <100 <110 SOR 30 (um) Figure 5 Frequency distribution of the short otolith radius from the nucleus to the 30th ring (SOR 30) for walleye pollock {Theragra chalcogramma) collected in four regions off the Pacific coast of northern Japan in 2000. A Low-growth group A High-growtti group High-growlh group Low-growth group 120 60 80 100 Number of otolith increments Figure 6 Size-at-age relationship for two growth groups of walleye pollock (Ther- agra chalcogramma) collected in 2000. Low- and high-growth groups were divided according to the SOR 30 distribution (see text and Fig. 5). The regression lines for each growth group are as follows: low-growth group: rL = 0.017+0.429xQge [r2=0.941, n = 594, P<0.0001]; high-growth group: rL = 4. 694 + 0. 486xa^e [r2=0.715, n = 152, P<0.0001]. 474 Fishery Bulletin 104(3) 1999 No survey No catch 15 25 5 45 55 65 2000 25 5 * 55 65 5 Total length (mm) 2001 Figure 7 Total length-frequency distribution for walleye pollock iTheragra chalcogramma) from four regions off the Pacific coast of northern Japan in 1999-2001. Low G=low-growth group (solid barsi; high G = high-growth group (open bars). 1999 No catch Dec. Jan, Feb Mar Apr 2000 ■ LowG ^^ DHIghG __j^|L LM-m Mar Apr 2001 Dec Jan Feb, Mar, Apr, Hatching date Figure 8 Hatching date composition for walleye pollock (Theragra chalcogramma) from four regions off the Pacific coast of northern Japan in 1999-2001. Low G=low-growth group (solid bars), high G = high-growth group (open bars). NOTE Hattori et al Hatching date, nursery grounds, and early growth of juvenile Theragra chalcogiomma 475 We suggest that, in addition to the recruitment pro- cess in Funka Bay, an important recruitment process occasionally exists in the Tohoku area, and when the year-class strength of the Japanese Pacific population of pollock is strong, as in 2000 (Yabuki^), the Tohoku area becomes important as a nursery ground for the late hatching fish originating from Funka Bay. Acknowledgments We thank the crew of the RV Wakatako Marii who endured long working hours to obtain the samples used in this study. We also thank Yasunori Sakurai of the Graduate School of Fisheries Sciences, Hokkaido Univer- sity, Yoshioki Oozeki of the National Research Institute of Fisheries Science, Fisheries Research Agency, and Kazuya Nagasawa of the Tohoku National Fisheries Research Institute, Fisheries Research Agency, for their advice and helpful discussions. We thank the anonymous reviewers who improved the draft of this article. This research was funded by the VENFISH Program of the Ministry of Agriculture, Forestry and Fisheries. Literature cited Bailey. K. M., and C. L. Stehr. 1988. The effects of feeding periodicity and ration on the rate of increment formation in otoliths of larval walleye pollock Theragra chalcogramma (Pallas). J. Exp. Mar. Biol. Ecol. 122:147-161. Bailey, K. M., A. L. Brown, M. M. Yoklavich. and K. L. Mier. 1996. Interannual variability in growth of larval and juvenile walleye pollock Theragra chalcogramma in the western Gulf of Alaska, 1983-91. Fish. Oceanogr. 5 (suppl. 1):137-147. Brodeur. R. D., M. S. Busby, and M. T. Wilson. 1995. Summer distribution of early life stages of wall- eye pollock, Theragra chalcogramma, and associated species in the western Gulf of Alaska. Fish. Bull. 93:603-618. "i Yabuki, K. 2000. Stock assessment of the Japanese Pacific population of walleye pollock. In Tohoku Sokouo Kenkyu 20, p. 41-44. Hactiinohe Branch, Tohoku Nat. Fish. Res. Inst., 25-259 Shimomekurakubo, Same, Hachinohe, Aomori 031-0841, Japan. [In Japanese, the title was translated by the authors.] Hashimoto, R., and Y. Ishito. 1991. Recruitment of the egg, larva and juvenile of wall- eye pollock and its early stage in the Northeastern coast of Japan. Bull. Tohoku Natl. Res. Inst. 53:23-38. [In Japanese.] Hayashi, K., H. Kitahama, U. Suzuki, and N. Endo. 1968. Ecology of the young Alaska pollock. Hokusuishi Geppo. Hokkaido Fish. Exp. Stn. 25:394-403. [In Japanese.] Honda, S., T Oshima. A. Nisbimura. and T Hattori. 2004. Movement of juvenile walleye pollock, Theragra chalcogramma, from a spawning ground to a nursery ground along the Pacific coast of Hokkaido, Japan. Fish. Oceanogr. 13 (suppl. 11:84-98. Inada, T, and M. Murakami. 1993. Fluctuations of walleye pollock and Pacific cod resources and bottom temperature in the waters off the Tohoku region of Japan. Sci. Rep. Hokkaido Fish. Exp. Stn. 42:229-240. Iln Japanese.] Kendall, A. W., and T Nakatani. 1992. Comparisons of early-life-history characteristics of walleye pollock Theragra chalcogramma in Shelikof Strait, Gulf of Alaska, and Funka Bay, Hokkaido, Japan. Fish. Bull. 90:129-138. Nakatani, T 1998. Survival of walleye pollock in early life stages in Funka Bay and the surrounding vicinity in Hokkaido. Mem. Fac. Fish. Hokkaido Univ. 45:64- 70. Nakatani, T, and T. Maeda. 1987. Distribution and movement of walleye pollock larvae Theragra chalcogramma in Funka Bay and the adjacent waters, Hokkaido. Nippon Suisan Gakkaishi 53:1585-1591. [In Japanese.] Nisbimura, A., and J. Yamada. 1984. Age and growth of larval and juvenile walleye pol- lock, Theragra chalcogramma (Pallas), as determined by otolith daily growth increments. J. Exp. Mar. Biol. Ecol. 82:191-205. 1988. Geographical differences in early growth of walleye pollock Theragra chalcogramma estimated by back- calculation of otolith daily growth increments. Mar. Biol. 97:459-465. Ogawa, Y. 1989. Variations in latitude at the southern limit of the First Oyashio Intrusion. Bull. Tohoku Reg. Fish. Res. Lab. 51:1-9. [In Japanese.] Ricker, W.E. 1979. Growth rate and models. In Fish physiology, vol. VIII (W. S. Hoar, D. J. Randall, and J. R. Brett, eds.), p. 677-743. Academic Press, New York, NY. Yoklavich, M. M., and K. M. Bailey. 1990. Hatching period, growth and survival of young wall- eye pollock Theragra chalcogramma as determined from otolith analysis. Mar. Ecol. Prog. Ser. 64:13-23. 476 Development of larval and early juvenile penpoInt gunnel iApodichthys flavidus) (family: Pholldae) Lisa G. De Forest' Morgan S. Busby^ ' University of Hawaii Department of Oceanograpfiy 1000 Pope Road, Honolulu, Hawaii 96822 2 National Marine Fisfieries Service National Oceanic and Atmosptienc Administration Alaska Fisheries Science Center 7600 Sand Point Way NE Seattle, Wasfiington 98115-6349 E-mail address (for M S Busby, contact author) morgan. busbyid'noaa. gov The penpoint gunnel iApodichthys flavidus) is a member of the perci- form family Pholidae. Pholids, com- monlj' referred to as gunnels, are eel-like fishes that inhabit the rocky intertidal and subtidal regions of the northern oceans and are often associated with macroalgae, such as Fuciis spp. or kelp (Watson, 1996). Gunnels are ecologically important forage fishes that form part of the diet of birds and commercially important groundfish species (Hobson and Sealy, 1985; NMFS'; Golet et al., 2000). The diet of A. flavidus and other pholids comprises primarily harpactacoid copepods, gammarid amphipods, iso- pods, and other crustaceans (Cross, 1981). Apodichthys flavidus ranges along the west coast of North America from southern California to the Gulf of Alaska (Mecklenburg et al., 2002). Adult A. flavidus are distinguished from other pholids by their total ver- tebral counts, the presence of a thick NMFS (National Marine Fisheries Service). 1998. Final environmen- tal assessment and regulatory impact review for Amendment 36 to the Fishery Management Plan for the groundfish fishery of the Bering Sea and Aleutian Islands Area and Amendment 39 to the Fishery Management Plan for ground- fish of the Gulf of Alaska to create and manage a forage fish species category, 76 p. NOAA/NMFS Alaska Regional Office PO Box 21668 Juneau, Alaska 99802-1668. and grooved first anal spine, a pre- anal length that is approximately 60% standard length (SL), and a dark green to light olive coloration (Yatsu, 1981). It is one of the largest phol- ids (up to 46 cm) and is important in the live fish trade for both home and public aquaria (Froese and Pauly'-). In late winter to early spring months (January-April), adult A. flavidus spawn in nearshore waters. A single female lays clusters of de- mersal, adhesive eggs onto substrate that a male will guard until hatching (Clemens and Wilby, 1961; Wilkie, 1966; Marliave, 1975). The eggs are 3 mm in diameter and the incubation period is approximately 2.5 months (Wilkie, 1966; Marliave, 1975). Lar- vae are about 12-13 mm total length (TL) at the time of hatching, well de- veloped, and have pigmented eyes, an elongated body, and very little to no yolk sac (Wilkie, 1966; Marliave, 1975). After about 50 days, the larvae settle as juveniles and are approxi- mately 25 mm SL (Marliave, 1975). Although A. flavidus reproduction has been well-studied, there has not been a complete description of lar- val development. Wang-^ provided a summary of life history information and pencil illustrations of early flex- ion, postflexion, and juvenile stages. Matarese et al. (1989) published an illustration of a 15.0-mm flexion lar- va and some characters distinguish- ing A. flavidus from Pholis spp. In the present study we describe development of A. flavidus from re- cently hatched larvae to newly set- tled juveniles, including some general aspects of osteological development. Larval A. flavidus are compared with larvae of other pholid species includ- ed in the genus Apodichthys by Yatsu (1981, 1985): Xererpes fucorum and Ulvicola sanctaerosae. This classifi- cation was not followed by Matarese et al. (1989) or Watson (1996) and is not followed in the present study. This work will aid in the accurate identification of A. flavidus larvae in samples taken during nearshore ichthyoplankton surveys and in eco- logical studies and will contribute to a better understanding of pholid systematics. Materials and methods We examined 58 larval and juvenile A. flavidus (11.9-42.3 mm) collected in dip-net surveys by scientists of the Alaska Fisheries Science Center (AFSC; Busby et al., 2000) and the University of Washington (LTW) from four sites; Clam Bay (47=34. 5'N, 122°32.5'W), Sequim Bay (48°2.3'N, 123°2.0'W), Iceberg Point (48°42.4'N 122°53.3'W), and Friday Harbor (48°54.5'N, 1230.7°W), all located in 2 Froese, R., and D. Pauly (eds.). 2004. Fishbase. World wide web electronic publication http://www.fishbase.org [accessed November 20041. ■' Wang, J. C. S. 1986. Fishes of the Sacramento-San Joaquin estuary and adjacent waters, California: a guide to the early life histories, 602 p. Technical Report 9 of the Interagency ecological study program for the Sacramento-San Joaquin Estuary. [Available from Eco- logical Analysts, Inc. 2150 John Glen Drive, Conco'rd, CA 94520.1 Manuscript submitted 22 June 2005 to the Scientific Editor's Office. Manuscript approved 24 October 2005 by the Scientific Editor Fish. Bull. 104:476-481 (2006). NOTE De Forest and Busby: Development of larval and early iuvenile Apodichthys flavidus 477 Puget Sound, WA, and from adjacent waters. Specimens were initially preserved in 3.59f buffered formalin solu- tion and later transferred to 70% ethanol (Busby et al., 2000). A dissecting ste- reomicroscope was used to examine pigmentation, gen- eral body size and struc- ture, and to obtain meristic counts. Morphological mea- surements were made on 55 suitable specimens by using a digital image anal- ysis system consisting of a video camera attached to a dissecting stereomicroscope and a computer with image analysis software. All mea- surements were taken from the left side of the speci- men. Standard length was used throughout the study unless otherwise indicated. During flexion stage, noto- chord length (NL) was mea- sured and recorded as SL. Measurements included standard length, head length, eye diameter, body depth, snout to anus length, and pectoral-fin length, as described by Moser (1996). To describe osteological development of A. flavi- dus, with emphasis on the development of the caudal skeleton, 12 specimens were cleared and stained by us- ing the technique described by Potthoff (1984). The terms "unossified precur- sor" or "element" are used to describe unossified ele- ments that took up alcian blue stain but not alizarin red-s stain. From the cleared and stained specimens, stages of larval development were identified from land- marks of caudal-fin development. Caudal skeletons of six specimens representing distinct stages were used to create illustrations. Developmental stage terms follow Kendall et al. (1984) and Neira et al. (1998). The flexion stage was divided into three additional stages: early-, mid- and late-flexion. Early-flexion begins at hatch- ing, mid-flexion begins with the formation of the forth hypural and epurals, and late flexion begins with the development of the flfth hypural and ends with complete notochord flexion. Nomenclature of caudal skeleton ele- ments follows Fujita (1989). 14 5 mm ^^^--—^^^^^^^ '"^^ "~~~^~'^S^^ 37.0 mm Figure 1 Developmental series of penpoint gunnel i Apodichthys flavidus)AA) Early-flexion larva. Clam Bay, 6 April 1989 (UW 104928): (B) mid-flexion larva (from Matarese et al., 1989); (C) late-flexion larva, Friday Harbor, 20 April 1994 (UW 104930); (D) postflexion larva. Clam Bay, 20 April 1989 (UW 104932): (E) juvenile, Sequim Bay, 2.5 April 1989 (UW 104934). Illustrations by Beverly Vinter. Results Morphology Apodichthys flavidus larvae are approximately 12.0-13.0 mm at hatching and in early flexion stage and have little or no yolk sac present (Fig. lA). The early-flexion stage occurs between hatching and 14.0 mm. Mid-flexion begins at approximately 14.0 mm (Fig. IB), late-flex- ion at 17.0 mm (Fig. IC), and postflexion at 20.0 mm (Fig. ID). Transformation to the juvenile stage occurs between 25.0 mm and 30.0 mm. Juveniles examined ranged from 30.1 to 42.3 mm and looked like small 478 Fishery Bulletin 104(3) Table 1 Body proportions of larval and early j uvenile penpoint gunnel {Apodichthys flavidus). Values given for each body proportion are | expressed as percentages of standard length (SL) or head length (HL): mean ± standard deviation, and range. Sample size, standard length. and body proportion Flexion Postflexion Juvenile Sample size 33 18 4 Standard length (SL) 14.9+2.1(11.9-19.2) 21.9 ±1.3 (20.0-24.0) 34.2 ±5.6 (30.1-42.3) Snout to anus length/SL 62.9+1.8(58.0-66.1) 63.3 ±1.2 (61.8-66.1) 63.5 ±1.8 (62.5-65.2) Body depth/SL 8.3 ±0.9 (6.4-9.91 9.3 ±0.7 (8.2-10.4) 11.6 ±1.1 (11.2-12.9) Head length/SL 13.2+0.7(12.1-14.8) 14.1 ±0.7(12.9-15.5) 16.4 ±1.1 (15.3-17.3) Eye diameter/HL 33.6+4.7(22.3-44.1) 27.4+3.1(23.5-35.4) 20.0+3.9(15.8-24.8) Pectoral fin length/SL 7.6 ±1.2 (5.4-10.2) 7.9 ±1.0 (6.2-10.1) 6.2 ±1.4 (4.9-7.1) adults (Fig. IE). Larvae are slender bodied throughout development and body depth increases from S'J SL during flexion to 12% SL in juveniles (Tables 1 and 2). Head, snout-to-anus, and pectoral-fin lengths are con- sistent throughout development at approximately 14%, 63%, and 7% SL, respectively. Eye diameter decreases from 34% head length (HL) in flexion larvae to 20% HL in juveniles. Pigmentation Early-flexion larvae have a few faint melanophores located dorsally on the head and nape and a single melanophore on the isthmus (Fig. lA). Along the dorsal surface of the gut, a row of large melanophores is present — irregularly spaced anteriorly and posteriorly, regularly spaced medially. Another row of smaller, evenly spaced melanophores is present along the anterior V2 to % length on the ventral surface of the gut. A row of postanal ventral melanophores (PVMs) extends from the anus to the caudal peduncle. Generally there is one PVM per myomere but in many individuals one or more are missing from the row. In addition, there are small patches of melanophores along the dorsal and ventral margins of the caudal peduncle. In mid-flexion larvae, patches of melanophores on the head, nape, isthmus, and caudal peduncle are more defined, and a row of internal melanophores is present above the notochord (Fig. IB). These melanophores develop simultaneously. In addition, the number of melanophores along the dorsal surface of the gut nearly doubles. The PVMs in late-flexion larvae are larger and more slashlike (Fig. IC). During postflexion, pigmentation previously noted now appears very faint (Fig. ID). Juvenile pigmentation resembles that of adults (Fig. IE). Most notably, a horizontal streak of melanophores extends from the snout to the anterior margins of the eye and continues from the posterior margin of the eye to the mid-operculum. The body in live specimens is a uniform bright green to olive and has white spots located above the gut and anterior por- tion of the anal fin. The gut is generally unpigmented. with the exception of a few very small irregularly spaced melanophores on the lateral surface. Osteology Head region In early-flexion larvae the maxilla, man- dible, two mandibular teeth, branchiostegal rays, and cleithrum are ossified. The premaxilla ossifies during mid-flexion. The remaining bones in the head region are not ossified until the juvenile stage. Fins Unossified precursors of five principal caudal- fin rays are present at hatching and throughout early flexion. During mid-flexion, 12 unossified precursors of pectoral-fin rays and 13 unossified caudal-fin elements are present, both first appearing at 15.0 mm. Also at this size, a few faint unossified anal-fin elements are present in the anterior portion of the anal finfold, which are not visible in unstained specimens. Beginning at 17.0 mm, unossified precursors of dorsal-fin spines are first present in the area of the dorsal finfold above vertebrae 45-50 and then develop anteriorly and posteriorly from this position. Unossified anal-fin elements are also added from anterior to posterior. In addition, the scapula, cora- coid, and radials of the pectoral fin are ossified and the adult complement caudal-fin elements is present, but still unossified (13 principal rays and 11 procurrent rays). By the end of late flexion (about 20.0 mm), elements of the dorsal, anal, and pectoral fins finally become ossified (Table 2). The single spine is the first element in the anal fin to become ossified. Pterygiophores of all fins are ossified in juveniles and the pectoral- and caudal-fin rays become ossified and branched. The number of fin elements present at any particular stage and length is somewhat variable as we noted slight differences in the numbers present between our stained and illustrated specimens. Vertebral column At hatching, all neural and haemal spines are present and ossified (Table 2). In addition, vertebral centra begin to differentiate from anterior to NOTE De Forest and Busby Development of larval and early juvenile Apodichthys flavidus 479 posterior during the early- and mid-flexion stages. At 17.0 mm, the vertebral centra are completely dif- ferentiated but remain unossified. In juveniles, the vertebral centra are completely ossified. Caudal skeleton In early flexion larvae, the noto- chord begins to bend upward and the haemal spine of the second preural centrum, the fused parhypural plus first and second hypurals, and the third hypural are present (Fig. 2A). Caudal skeleton elements begin to ossify at this stage, beginning at the base of each. Ventral elements that develop at mid-flexion are the haemal spine of the third preural centrum and the fourth hypural (Fig. 2B). Dorsally, epurals 1-3 and neural spines of the second and third preural centra form during this stage. Elements present in early- flexion are now fully ossified. At the beginning of late-flexion (about 17.0 mm), a fifth hypural and a fourth epural are present but not ossified (Fig. 2C). In the 19.2-mm late-flexion specimen examined, the third and fourth epurals were fused and the first epural was fused to the neural spine of the second preural centrum (Fig. 2D). In postflexion larvae, the distal margins of hypurals 3-5 are oriented verti- cally and the first epural separates from the neural spine of the second preural centra (Fig. 2E). All other elements in the caudal skeleton have grown and are fully ossified. In juveniles, the caudal fin has the adult form and the final element, the uroneural, is present just ventral to the second and fused third and fourth epurals (Fig. 2F). A ventral caudal radial, present from mid-flexion to postflexion (Figs. 2, C-E), is absent in juveniles. Discussion Our description of A. flavidus development can be used to distinguish larvae of this species from co-occurring species of Pholis spp., Apodichthys, Ulvicola, and Xererpes along the West Coast of the United States. As with Ulvicola sanctaerosae, A. flavidus does not have a preflexion stage and larvae hatch at an advanced developmental state (Watson, 1996). Larval Pholis spp. differ from larval A. flavi- dus by the presence of pelvic fins in the former and by differences in pigmentation. The melanophores along the dorsal surface of the gut in Pholis spp. are more numerous (about 25 vs. about 18 in postflexion larvae) and closer in spacing anteriorly (Matarese et al., 1989), and Pholis spp. do not develop an internal row of melanophores along the dorsal margin of the notochord. Apodichthys flavidus develops a series of internal melanophores along the entire length of the notochord by the mid-flexion stage. Larvae of U. sanctaerosae can be distinguished from A. flavidus by pigmentation and by the number and persistence of pectoral-fin rays in postflexion larvae and juveniles. Larval U. sanctaerosae have fainter melanophores that are more irregularly spaced along the dorsal a. in Si ^ c c 3 60 u -a -O 3 O as c/j 6 S to o. s- tc ■- CQ o CI ^ _i a) CO g a S < "a! c CC -a P ^ CO -* C<1 -* lO ■* O O ^ CTi CC "O ■* G^ Gi O^ lO lO lO CO CQ CO ■* OJ CO OJ ^* ^* ^^ ^^ ^^ o ^ ,-( o o lO iO lO lO lO LQ »0 ^ lO lO lO IC TjH lO lO 05 i-H O (TO Tf -^ O O U O Q ?S ?S rS rS rS ^ CO O CO (N N e 480 Fishery Bulletin 104(3) 12 2 mm B NPU2 EP1 gp2 NPU3 \ I J ^EP3 NC 15.0 mm EP3 PH-'HY1-2 VCR ,,„ HPU2 17.0 mm D EP1»NPU2. ^f^ NPU3*4 HPU3 / VCR HPU2 19 2 mm VCR HPU2 21.2 mm PH*HY1-2 HPU3 "hpu2 42.3 mm Figure 2 Caudal skeleton development of penpoint gunnel iApodichthys fla- vidus). (A) Early-flexion larva, Clam Bay, 27 February 1995 (UW 104943): (B) mid-flexion larva (note irregularly shaped EP2), Clam Bay, 10 May 1989 (UW 104937); (C) late-flexion larva. Clam Bay, 19 April 1988 (UW 104939); (D) late-flexion larva. Clam Bay, 15 May 1989 (UW 104940); (E) postflexion larva, Friday Harbor, 8 April 1993 (UW 104943); iF) juvenile. Iceberg Point, 18 July 1963 (UW 018016). Caudal-fin element abbreviations; EP=epural; HPU = haemal spine, preural; HY=hypural; NC = notochord; NPU = neural spine, preural; PH = parhypural; PU = preural centra; U = urostyle; UN = uraneural; VCR=ventral caudal radial. Illustrations by Lisa De Forest. surface of the gut than in A. flavidus, and in early- and mid-flexion larvae, these melanophores are restricted to the posterior V4 to V2 of the gut. Another distinguish- ing characteristic is that U. sanctaerosae does not fully develop pectoral-fin rays and the pectoral fin does not persist after the juvenile stage. A pectoral finfold is present during the larval stage of this species; however, only the uppermost pectoral-fin rays (6 or 7 vs. 14 or 15 for A. flavidus) partially develop but do not persist, and the pectoral finfold decreases in size during the latter part of development. Larvae of Xererpes fucorum can be distinguished from A. flavidus by the presence of a preflexion stage and by having fewer total (84-93 vs. 96-101) and postanal (35-40 vs. 47-52) myomeres. In addition, during the later stages of development, X. fucorum has fewer pectoral-fin rays than A. flavidus (12 vs. 14-15). Yatsu's (1985) revision of the family Pholidae placed U. sanctaerosae and X. fucorum in the genus Apodich- thys, but this classification was not followed by Mata- rese et al. (1989), Watson (1996), or in the present study. Larvae of both these species are quite similar NOTE De Forest and Busby: Development of larval and early luvenile Apodichthys flavidas 481 to A. flavichis; however, we recommend a more detailed study of larval U. sanctaerosae and X. fucoruiu, includ- ing a description of caudal skeleton development, before concluding that larval characters do, or do not, support Yatsu's (1985) classification. In particular, it would be interesting to investigate whether either species devel- ops a fourth epural that fuses with the third, or a first epural that fuses with the neural spine of the second preural centrum during flexion. In another cleared and stained individual we observed fusion of the neural spines on the third and fourth preural centra {NPU4 and NPU3, Fig. 3D). However, only one specimen of A. flavidus was examined at each of these fusions and more specimens, when available, should be cleared and stained to determine if these fusions occur in all larval A. flavidus. Presence or absence of a ventral caudal radial may also be of interest. Although it is unclear what becomes of the ventral caudal radial between postflexion and juvenile stages from our study of A. flavidus, we hypothesize that it fuses with the tip of the haemal spine of the second preural centrum. Tak- ing all of these unusual aspects of A. flavidus larval development into account, we suggest that development of U. sanctaerosae and X. fucorum should be further investigated to clarify the systematic relationships among the genera. Acknowledgments The authors thank Theodore Pietsch (University of Washington [UW]) for his help, support, and encour- agement with this project. Beverly Vinter illustrated the specimens in the developmental series. We also thank the Ichthyoplankton Laboratory of the Alaska Fishery Science Center (AFSC) and the UW Fish Collection for use of their facilities and specimens. Ann Matarese (AFSC) and William Watson (NOA A- Southwest Fish- eries Science Center) reviewed an earlier draft of the manuscript. Literature cited Busby, M. S., A. C. Matarese, and K. L. Mier. 2000. Annual, seasonal, and diel composition of larval and juvenile fishes collected by dip-net in Clam Bay, Puget Sound, Washington, from 1985 to 1995. NOAA Tech. Memo. NMFS-AFSC-111, 36 p. Clemens, W. A., and G. V. Wilby. 1961. Fishes ofthe Pacific Coast of Canada, 2"'' ed. Bull. Fish. Res. Board Can. 68, 443 p. Cross, J. N. 1981. Structure of a rocky intertidal fish assemblage. Ph.D. diss., 259 p. Univ. Washington, College of Fish- eries, Seattle, WA. Fujita, K. 1989. Nomenclature of cartilaginous elements in the caudal skeleton of teleostean fishes. Jap. J. Ichthyol. 36:22-29. Golet, G. H., K. J. Kuletz, D. D. Roby, and D. B. Irons. 2000. Adult prey choice affects chick growth and reproductive success in pigeon guillemots. The Auk 117:82-91. Hobson, K. A., and S. G, Sealy. 1985. Diving rhythms and diurnal roosting times of pelagic cormorants. Wilson Bull. 97:116-119. Kendall, A. W., Jr., E. H. Ahlstrom, and H. G. Moser. 1984. Early life history stages and their characters. /;; Ontogeny and systematics of fishes (H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. Richardson, eds.), p. 11-22. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence. KS. Marliave, J. B. 1975. The behavioral transformation from the plank- tonic larval stage of some marine fishes reared in the laboratory. Ph.D. diss., 231 p. Univ. British Columbia, Vancouver, Canada. Matarese, A. C, A. W. Kendall Jr., D. M. Blood, and B. M. Vinter. 1989. Laboratory guide to early life history stages of northeast Pacific fishes. NOAA Tech. Rep. NMFS 80, 652 p. Mecklenburg, C. W.. T. A. Mecklenburg, and L. K. Thorsteinson. 2002. Fishes of Alaska, 1037 p. Am. Fish. Soc, Bethesda, MD. Moser, H. G. (ed.) 1996. The early stages of fishes in the California Cur- rent region. Calif Coop. Oceanic Fish. Invest. Atlas 33, 1505 p. Neira, F. J., A. G. Miskiewicz, and T. Trnski. 1998. Larvae of temperate Australian fishes: laboratory guide for larval fish identification, 474 p. Univ. Western Australia Press, Western Australia, Australia. Potthoff, T. 1984. Clearing and staining techniques. In Ontogeny and systematics of fishes (H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall Jr., and S. L. Richardson, eds.), p. 35-37. Spec. Publ. 1, Am. Soc. Ichthyol. Herpetol. Allen Press, Lawrence, KS. Watson, W. 1996. Pholidae: gunnels. //; The early stages of fishes in the California Current region (H. G. Moser, ed.), p. 1120-1125. Calif Coop. Oceanic Fish. Invest. Atlas 33. Wilkie, D. W. 1966. Color pigments in the penpoint gunnel Apodich- thys flavidus and their ecological significance. M.S. thesis, 143 p. Dept. Zoology, Univ. British Columbia, Vancouver, Canada. Yatsu, A. 1981 A revision of the gunnel family Pholidae (Pisces, Blennioidei). Bull. Natl. Sci. Mus. (Tokyo) Ser. A (Zool.) 7:165-190. [In Japanese.] 1985. Phylogeny of the family Pholidae (Blennioidei) with a redescription oiPholis Scopoli. Jap. J. Ichthyol. 32:263-282. 482 Spatial and temporal patterns In the bycatch of seabirds In the Argentinian longllne fishery Patricia Gandini Esteban Frere Centre de Investigaciones de Puerto Deseado Universidad Nacional de la Patagonia Austral Conseio Nacional de Investigaciones Cientificas y Tecnicas (CONICET) and Wildlife Conservation Society Avda, Prefectura Naval s/n, cc 238, (9050) Puerto Deseado Santa Cruz, Argentina E-mail address (for P Gandini) pagandinnayahioo com ar Longline fisheries have grown throughout the world's oceans for more than 40 years. This type of fisheries has captured high-quality fish (mature individuals rather than unwanted juveniles), has had minimal destructive effects on bottom habi- tats, and has produced a low bycatch of nontargeted fish (Brothers et al., 1999). Seabirds, however, are hooked accidentally when they swallow or are snagged on the baited hooks set by commercial longline crews (Brothers, 1991; Barnes et al., 1997; Tasker et al., 2000; Belda and Sanchez 2001; Jahncke et al., 2001). Population declines of several spe- cies of albatrosses and petrels in the Southern Ocean are linked to longlining operations (Croxall and Prince, 1990; Brothers, 1991; Cherel et al., 1996). The importance of the Patagonian shelf waters as a forag- ing habitat for seabirds is well docu- mented (Cooke and Mills, 1972; Veit, 1995), particularly for black-browed albatross (Thalassajxhe melanophi'is; Gales, 1998). We estimated the mag- nitude of seabird mortality in the kingclip {Genyptei-us blacodes) fishery in the Argentine Exclusive Economic Zone (EEZ). Materials and methods Description of the fishery Demersal longline vessels target two species on the Patagonian shelf of Argentina: kingclip and Patagonian toothfish (Dissostichiis eleginoides). Six vessels operate within Argen- tina's EEZ in the South Atlantic: three vessels, Marunaka. Estela. and Magallanes II, fish mainly kingclip off Comodoro Rivadavia and Puerto Deseado and three vessels, Antarc- tic I. II and ///, target Patagonian toothfish off Ushuaia. Another twelve vessels based in the Malvinas (Falk- land) Islands and operating outside Argentina's jurisdiction target king- clip and Patagonian toothfish (Blake, 2001; Fig. 1). Within Argentina's EEZ, an artisanal longline fleet of 70 ves- sels target hake (Merluccius hitbbsi). fishing mainly in shallower waters within the San Matias Gulf. The kingclip fishery operates year round with trips lasting up to 65 days. Vessels are equipped with an autoline system; the line is 10 km in length and has up to 20,000 hooks which are baited with thawed squid at night by using a baiting machine. Offal is released strategically to at- tract birds to the opposite side of the vessel away from the main line dur- ing hauling. Data collection Observers, trained in seabird iden- tification and in quantifying mortal- ity, estimated fishing effort and the bycatch of seabirds. Data were col- lected through a special agreement between a fishing company (Argenova S.A.) and the university (Universidad Nacional de la Patagonia Austral). Observers sampled 156 sets and hauls from December 2000 to Sep- tember 2001. Sets were deployed as early as 23:00 and as late as 04:00. Observers recorded the date, num- ber of hooks deployed, position and time of the start and end of each set, water depth, and sea condition. If a bird was caught, or if we observed flapping wings or the struggle of a bird, we recorded the event. During hauling we identified and counted all drowned birds. We collected a sample of drowned birds and brought them to the laboratory to determine their sex. Observers sampled the sets of one of the three vessels targeting kingclip in Argentina's EEZ. We estimated to- tal bird catch using observed bycatch and official information, such us the number of hooks deployed from log- sheets of the other vessels. Data analysis We used the bootstrapping method to calculate the 95% confidence limits for the mean number of birds caught/1000 hooks and made 10,000 simulations (n = 156). Using logistic regression, we determined which variables best explained the "capture events." Ini- tial independent variables included date, number of hooks deployed, water depth, latitude, sea condition, time at which the setting began, and moon phase. Moon phase was scored from 0 to 14, where 0 represents a full moon and 14, a new moon. To determine the contribution of each variable to seabird mortality rates, we used a stepwise multiple regression (Zar, 1984). On sets with catch per unit of effort (CPUE) greater than 0, three separate multiple regressions were performed: 1) for total seabirds, 2) for albatrosses, and 3) for petrels. Seabirds whose sex was determined were used to test differences from the 1:1 sex ratio by using a x~ test (Zar, 1984). We compared, using this test, the number of birds of each species caught during winter, summer and autumn (Zar, 1984). Manuscript submitted 28 May 2003 to the Scientific Editor's Office. Manuscript approved for publication 24 October 2005 by the Scientific Editor. Fish. Bull. 104:482-485(2006). NOTE Gandini and Frere Bycatch of seabirds in the Argentinian longlme fishery 483 S-54- Results Bycatch of fish and seabirds The mean depth for longlining operations was 139 ±11 meters (n=156) and the mean number of hooks per set was 20.474 ±2588 U!=156). Effort for the whole fleet during the nine-month sampling period represented a total of 10,088,235 hooks. Kingclip represented 80'7f of the fish caught. Additional catch included Patago- nian toothfish (29c). Argentine hake (3%; Merluccius hubbsi). Brazilian codling (8%; Urophycis brasiliensis), blackbelly rose- fish (2'7(\ Helicolenus dactylopterus lahil- lei), skates (5%; Raja spp.) and other fish in proportions below 1% (i.e., Patagonian cod {Salilota australis); hoki (Macrorunus magellanicus); tope shark (Galeorhinus ga- lena). Patagonian smoothhound iMusfelus schmitti), among others). Only two species of seabirds were caught during setting of the longline: 12 white- chinned petrels (Procellaria aequinoctialis) and 19 black-browed albatrosses; no sea- birds were caught during hauling. Most fish- ing and bycatch occurred between 48-46°S latitude and 63-61°W longitude (Fig. 1). The observed distribution of genders was not sig- nificantly different from an expected 1:1 sex ratio for albatrosses (x^=0.06, df=l), or pe- trels (x-=0.15, df =1). Fifty-three percent of the albatrosses caught were males and 47% females (?!=17). White-chinned petrel were evenly divided: 50% were males and 50% females (n='l2). Bycatch was not homoge- nous throughout the year. The majority of the petrels were caught during autumn (x^=66.3, df=2, P<0.001), whereas albatrosses were predominantly caught during winter (■)^=bA.l, df=2, P>0.001, Fig. 2). We estimated a mean bycatch rate of 0.034 ±0.009 birds/1000 hooks which, extrapolated to the other vessels, represents a total of 343 birds caught in a nine-month period. Of these birds, 55% would be black-browed albatross and 45% white-chinned petrels. Factors affecting bycatch rates Logistic regression analysis indicated that water depth (/Jq=-0.96) and moon phase (/3q=0.24) explained most of the variability in "capture events"; the probability of a bird getting caught increased with water depth and during brighter nights. Two variables were important in explaining the variation in bycatch rate (F=3.13; df=4, 42; P=0.02): moon phase (/3=-0.327, P=0.03) and water depth (/3=0.335, P=0.02). When we analyzed the variables that explained the incidental capture for both species separately, moonlight was the most important for albatrosses (F=4.637, df=3, 21, /3=-0.52, P<0.01), 200 m ■57 W Figure 1 Map showing ports (solid triangles) used by longliners in Argentina. Open triangles indicate set positions where seabird bycatch occurred and circles indicate sets without seabird bycatch. The 200-m isobath indicates the shelf break. whereas water depth was the most important for petrels (F=2.4, df=2, 22, /J=0.584, P<0.05). Discussion This note reports the first estimation of seabird bycatch in the Argentinean EEZ outside of the CCMLAR (Com- mission for the Conservation of Antarctic Marine Living Resources) area for the Atlantic Ocean with data from onboard observers. Pelagic seabird species, such as white-chinned petrels and black-browed albatrosses, actively forage in the highly productive Patagonian shelf waters (Cherel and Klages, 1998; Prince et al., 1998; Weimerskirch et al., 1999; Gremillet et al., 2000); there- fore these species are especially vulnerable to fisheries around coastal shelf-edge habitats (Weimerskirch et al., 1999; Gremillet et al., 2000). These two species are often associated (Veit, 1995) and are similarly affected by longline fishing (Moreno et al., 1996; Berrow et al., 2000). We found the average bycatch rate to be 0.034 birds/1000 hooks, which is one or two orders of mag- nitude lower than those reported from other longline 484 Fishery Bulletin 104(3) fisheries in the South Atlantic (i.e., Brazil 0.126 birds/1000 hooks; [Neves and Olmos, 1998], Uru- guay 4.7 birds/1000 hooks; [Stagi et al., 1998]), but similar to those reported for the CCAMLR region (1998: 0.025 birds/1000 hooks; 1999: 0.07 birds/1000 hooks; 2000: 0.0014 birds/1000 hooks; [Robertson et al., 2001]). We made bycatch estimations from vessels, while voluntarily using mitigation measures to re- duce fishery-induced mortality on seabirds — such as tossing thawed bait (less bouyant than frozen bait), setting lines only at night, and splashing the net buoys when birds came near the line. In a nine-month period, 343 seabirds were incidentally caught by vessels targeting kingclip, representing an annual mortality estimate of 457 birds. Previ- ous crude estimations made for day-setting were considerably higher (2 birds/1000 hooks, Gandini and Frere, 2001). The main factor affecting seabird mortality was water depth, which is directly related to the location of the setting because water depth in- creases towards the shelf break, where white- chinned petrels forage actively throughout the year (Weimerskirch et al., 1999; Berrow et al., 2000). Sixty-five percent of the birds were caught be- tween five days after or before the full moon. A full moon affected albatrossess more significantly than it did petrels. Albatrosses forage mainly during the day, whereas white-chinned petrels forage both at night and during the day (Weimerskirch et al., 1999). Our results support the theory that darkness reduces seabird by- catch. Therefore, improving mitigation measures (i.e., using streamer lines, splashing bouys, etc.), during full moon phases, may reduce bird bycatch further. Birds were more susceptible to being caught in autumm and winter. Our findings agree with those of White et al. (2001), who mentioned that vulnerability for seabirds in Patagonian shelf waters is higher between September and June than during the rest of the year. Population impacts and recommendations to reduce bycatch of seabirds At least 500,000 pairs (nearly 86*7? of the world popula- tion) of black-browed albatrosses breed on the Malvinas and Falkland Islands and forage on the highly productive Patagonian shelf. Thus, the black-browed albatross catch estimated in this study represents 0 7- 0.7 ■ 1 06- ■ ■g 05- ■ 3 04 ■ to 0.05). Neither sex was observed to show ontogenetic dietary changes: fish of the 19-25 cm size class, and fish between 26 and 38 cm TL showed no significant differences in prey items (G=2.27; P>0.05). On the other hand, there were signifi- cant differences in the diet of S. capensis among loca- tions (x- = 44.94, P<0.05). The correspondence analysis (Fig. 2) indicated that fish from Antofagasta, Coquimbo, and Valparaiso fed on similar prey items; whereas the fishes from Talcahuano and Valdivia fed mainly on other prey (Mysidacea spp. ). Fish from the Aysen chan- nels and the Argentinia coast showed qualitatively dif- ferent diets. The diet of fish from the Aysen channels was characterized by the absence of those prey (e.g., R. typiis; P. desmarestii) recorded in the fish from the northern locations. On the other hand, fish from the Argentina coast, fed on prey that are distributed only in higher latitudes (e.g., Miinida spp.). Table 2 summarizes the IRI % for each prey item in each location. Antofagasta In this location, 29 specimens had stomach contents, from which 13 prey items were identified. The most important prey was the crab P. desmarestii (48.68% IRI), followed by the rock shrimp, R. typus (40.76% IRI), and low occurrences of shrimp Synalpheus spinifrons (4.01% IRI) and rock crab C. setosus (1.17% IRI). Coquimbo Only 10 types of prey were identified from this location. The most predominant was R. typus (92.15% 494 Fishery Bulletin 104(4) 3 ■ 6 1 / E 5 / Q 3/ ^^~~~~---.,,.^ -1 1/ ^--•7 2* Q 1 1 -3-11 3 Dim (1) Figure 2 Compositional gradients of correspondence analysis of prey items (shown in Table 1) by location. l=Antofagasta, 2 = Coquimbo, 3 =Valparaiso, 4=Talcahuano, 5=Valdivia, 6=Aysen, and 7 = Golfo Nuevo (Argentina). Dim (l) = dimension 1; Dim (2) = dimension 2. IRI), followed by the fish Chromis crusma (3.54 % IRI) and unidentified Osteichthyes (2.12 9( IRI). Polychaetes were not found in this area, and only one type of prey belonging to the phylum Mollusca was recorded. Valparaiso Nineteen prey items were identified; of those, 14 were crustaceans. In this location, three prey items were predominant: the crab C. setosus (34.32% IRI), the shrimp, R. typus (27.38% IRI), and the crab, P. des/narestii (19.09% IRI). The fishes were relatively well represented (unidentified Osteichthyes 8.49% IRI). The Mollusca (2.60% IRI) were represented by the small octopus Robsonella fontaniana. and the snail Turritella cingulata. Talcahuano A total of 24 prey items were identified from this location: 18 crustaceans, four fishes, one snail, and one polychaetae. The most important prey item was Mysidacea spp. (70.94% IRI), followed in importance by the crabs P. desmarestii (6.4% IRI) and C. setosus (5.09% IRI). Valdivia Twenty four prey items were identified from this area. Mysidacea spp. was the most important in this location (92.13% IRI). However, six of the fish species we identified were represented in this area. Aysen Fjord Despite the importance of unidentified prey items in this area, we observed an absence of most of the prey items found in the other locations. However, the trend of the predominance of crustaceans contin- ued. The most important prey item was the megalop of P. sicarius (94% IRI) and unidentified Osteichthyes (4.15%). The larval stage of P. sicaiius occurs season- ally; therefore, their importance in the diet of S. capensis from this geographical area can be associated with the sampling period (November). Golfo Nuevo In this area, 12 prey items were iden- tified. Lobster krill Miinida subriigosa (32.07% IRI), the brachyuran crab Halicarcinus planatus (23.53% IRI), and Munida gregaria ("15.84% IRI) were the most important prey items. The percentage of unidentified Osteichthyes (22.47% IRI) was higher than in other locations on the Chilean coast. Feeding strategy Figure 3 shows prey-specific abundance plotted against frequency of occurrence of prey items for S. capensis in the entire study area (Fig. 3A), and in each subarea (termed "location" in this article) (Fig. 3, B-H). This pattern of point distributions was similar for almost all locations. Most of the points (frequency of occurrence) were close to the left side of the .v axis, and they tended to be positioned towards the upper left corner of the graph (specific abundance). There was a high between- phenotype contribution (BPC) to the niche width in Antofagasta (3B), Coquimbo (3C), Valparaiso (3D), Aysen fjord (3G), and Golfo Nuevo(3H) (i.e., individual preda- tors specialized on different prey types and each food category was consumed by only a limited fraction of the predators). A more mixed feeding strategy, with varying degrees of specialization and generalization for different prey types was observed in Talcahuano (3E ) and Valdivia (3F) (i.e., a high BPC and lower within-phenotype con- tribution (WPC) to the niche width). On the other hand, predominant prey items were observed only in Coqui- mbo, Valdivia, and Aysen fjord (R. typus, Mysidacea gen. sp., and the megalops of P. sicarius, respectively); rare or occasional prey items (crustaceans, fishes, mollusks, and polychaetes) were found in all locations. Discussion To determine the dietary variability of a predator, in order to describe its feeding pattern over its entire geographical range, is a difficult task because of the need for large sample sizes in a number of locations. Additionally, fish, such as S. capensis, whose percent- ages of empty or everted stomachs are high (50% ), make an assessment of diet variability yet more difficult. We managed these difficulties by sampling fish during the same season, sampling over a similar depth range, and by using known information about the geographical distribution of prey items. Geographical variation in the diet The high percentage of crustaceans in the diet of S. cap- ensis indicates that is primarily a carcinophagous fish, Barnentos et al Differences in tfie feeding patterns of Sebastes capensis 495 100 » A B U » R lypus ^ R lypus P desmarestii 80 • ♦ P desmarestii * Mysidacea ''° t ♦Osteichthyes * t • Cselosus ♦ u ^ 100 1 c \ * Fl lypus jy 80 ■ « • R typus « C cetosus 60 - * Specific abundance ♦ ♦ ♦ • • : E K F t "0- « ♦ Mysidacea 60 ■ » * • Mysidacea 40 ■ , ► 20 • 0 - 100 r ■ U r ♦ •P. sycanus ^^ ♦ Osteichthyes ♦ ♦ M gregana wj ♦Osteichthyes M. subrugosa 80 60 * H. planatus 40 ♦ t 20 ■ ♦ • • • ♦ 0 20 40 60 80 100 0 20 40 60 80 100 Frequency of occurrence (%) Figure 3 Prey-specific abundance iPi) plotted against frequency of occurrence l'*F) of prey items for red rockfish iSebastes capensis) from the Chilean and Argentinean coasts. lAl entire geographic area. (B) Antofagasta, (C) Coquimbo. (D) Valparaiso, (E) Talcahuano, (F) Valdivia, (G) Aysen Fjord, and (Hi Golfo Nuevo. R. typiis= the rock shrimp Rhynchocinetes typus: P. desmarestii = deep red crab (PetroUsthes desmarestii): C. setosus = the rock crab Cancer setosus; M. gregraria=the lobster krill Munida gregaria; M. subriigosa =the lobster krill Munida subriigosa: H. planatus = t'he brachyuran crab Halicarcinus planatus. and secondarily, a piscivorous species. Sixty prey items were recorded in the entire study area, including Mysi- dacea (75.069f IRI), the most important prey, followed by Osteichthyes (6.29% IIR), Rhynchocinetes typus (6.03% IRI), and Petrolisthes desmarestii (4.67% IRI). The composition of prey species for S. capensis showed clear geographic variation. These changes were reflected by the relative importance of the different prey items in each of the study locations (Table 2) and were most evident in the fjords of Aysen and on the Argentinian coast, where all of the prey items found (except for Halicarcinus planatus) were unique to those locations. The crustacean R. typus was one of the most important prey items (50% IRI) between 23°S and 33°S. The rock shrimp R. typus is distributed from 10°S in Peril to 38°S on the Chilean coast (Retamal, 1981). Therefore, this prey species is not available on the southern Chilean coast (southward from 38°S). Between 37°S and 40°S, 496 Fishery Bulletin 104(4) Mysidacea was the most important prey item. In this geographic area, the mysids are also the most impor- tant prey item for other rocky bottom fishes (Gonzalez and Oyarzun, 2003). Mysids are found in large and dense patches a few centimeters from the benthos (Silva and Stuardo, 1985). The changes in prey types and their differences in IRI were also supported by a correspondence analysis (Fig. 2), which separated the sampling locations into four major areas according to recorded prey types as follows: 1) 23°S-33°S. 2) 37°S-40°S, 3) 45°S. and 4) Golfo Nuevo (43°S), Argentina. These geographic areas agree well with the proposed biogeographic units for this region by Briggs (1995) and by Camus (2001). Ac- cording to Briggs (1995). two faunistic provinces are recognized in the southeastern Pacific: the Peruvian province, which extends from Peru (ca. 10°S) down to the northern Chilean coast (ca. 30°S), and the Magel- lanic province located on the southern Chilean coast (ca. 43°S). Between these provinces a transitional zone is found, in which species typical of the Peruvian prov- ince and the Magellanic province overlap (Brattstrom and Johanssen, 1983; Lancelloti and Vasquez, 2000). Additionally, Camus (2001) recognized five areas of distributional discontinuities along the Chilean coast: (1) 30°S; (2) 33°S; (3) 37°-38° S; (4) 41°-43°S, and (5) 52°-53°S. Thus prey, such as R. typus, are distrib- uted along the Peruvian province and transitional zone, whereas prey like mysids are present mainly between latitudes 37°S and 40"S. Alternatively, Balech (1954) and Pequeiio (2000) con- sidered all of South America's southern extremity as the same biogeographic province. The records of prey items (i.e., Miinida gregaria and M. subrugosa) in Golfo Nuevo (Argentina coast) support the existence of a connection between the southeastern Pacific and the southwestern Atlantic. These prey species are also distributed along the Chilean coast from Chiloe (42°S) southward. However, these prey items were not found in Aysen fjord, which could be explained by the presence of two zones in the Aysen area with different physical and faunistic conditions: the fjord channels and a ma- rine zone (Arntz et al., 1999). Fish were sampled in the Aysen channels, where species of Miinida are absent (Retamal, 1981). Therefore, abundance and availability of prey may be a factor explaining the differences in the dietary composition of S. capensis between biogeo- graphical regions. The present study indicates the existence of geograph- ic variations in the dietary composition of S. capensis, and these variations generally agree with the zoogeo- graphic limits proposed in the literature. However, the interpretation of this result is limited because of the absence of an adequate background on the abundance and biogeography of each prey item. Feeding strategy According to Ojeda and Farina (1996), S. capensis is a nocturnal predator that feeds selectively on the decapod crustaceans R. typus and P. perlatus in the rocky subti- dal shores of the central Chilean coast. Ojeda and Fariiia suggested that these species are consumed preferentially from a large array of potentially abundant invertebrate prey items. Crustaceans are more active at night than during the day, which could explain their large rate of consumption by rockfishes that are active a night. Decapod crustaceans have the highest caloric content (Duarte et al., 1980), therefore Ojeda and Fariiia (1996) proposed that the specialized diet of S. capensis reflects an optimal foraging alimentary strategy based on the maximization of available energy. On the other hand, Moreno and Zamorano (1980) indicated that the feeding strategy of a species has a direct relationship with its morphological and behavioral adaptations that preclude being preyed upon, and these adaptations could restrict the potential prey items that they could capture. In this context, Moreno et al. (1979) and Moreno (1981) studied the bathymetric distribution of six species of fishes from the rocky subtidal shore of the Valdivian coast, including S. capensis. They established that the fishes segregate by habitats in order to take refuge from predators, mainly the sea lion Otaria flavescens; such segregation was not related to the bathymetric distribution of their most important prey items. Sebastes capensis shows buccal traits characteristic of carnivorous speciesl it has a good protrusion capacity and its teeth are inclined in a dorsal-caudal orienta- tion that prevents the escape of prey once they have been caught (Alvarado, 1985). Sebastes capensis is a poor swimmer; its anatomy conforms to that of sed- entary species in that its head and mouth are big in comparison to the rest of the body, and it has big eyes that allow an extensive angle of vision facilitating the capture of their prey and the detection of their preda- tors. Therefore, their capacity to catch prey depends mainly on characteristics of their prey. All prey species of S. capensis show similar behavioral traits (i.e., swim- ming organisms that move between rocky intertidal and subtidal zones) and S. capensis waits to capture them. The specialized feeding strategy of S. capensis may be a consequence of the availability of prey species moving in the vicinity of the refuges of this fish. The feeding strategy index, as defined by Amund- sen et al. (1996), has been used successfully for active predators like sharks (e.g., Lucifora et al., 2000; Vogler et al. 2003). However, this index does not integrate either predator behavior or the availability of prey in a given environment. Therefore, the criteria used to differentiate specialist from generalist predators must be clarified. According to the feeding strategy index of Amundsen et al. (1996), S. capensis is a specialized predator (preying mainly on crustaceans, although it consumes some prey species, and other rare prey species in almost all locations analyzed). In reality, S. capensis is an ambush predator that does not actively search for its prey. Because this fish waits in rocky refuges for mobile prey, the probability of it encountering prey depends, to a large extent, on prey abundances, prey distributions, and prey behaviors. Barrientos et al Differences in the feeding patterns of Sebostes copensis 497 Acknowledgments The authors thank to Carlos Jara for the collaboration in the identification of crustaceans prey. This study was realized under the sponsorship of D.I. D. (Direccion de Investigacion y Desarrollo) of the Universidad Austral de Chile (Project D-2003-09). M.T.G. was supported by a CONICYT fellowship of the Chilean Government. Literature cited Abrams, P. 2000. The evolution of predator-prey interactions: theory and evidence. Annu. Rev. Ecol. Syst. 3179-105. Alvarado, V. 1985. Trophic and adaptative morphological aspect of red rockfish Sebastes capensis (Gmelin, 1788). M.S. thesis, 70 p. Facultad de Ciencias, Universidad Austral de Chile, Valdivia, Chile. [In Spanish.] Amundsen, P.. H. Gabler, and F. Staldvik. 1996. A new approach to graphical analysis of feeding strategy from stomach contents data-modification of the Costello < 1990 I method. J. Fish Biol. 48:607-614. Arntz, W. E.. M. Gorny. R. Soto, M. Lardies, M. Retamal, and I. Whertmann. 1999. Species composition and distribution of decapod crustaceans in the waters off Patagonia and Tierra del Fuego, South America. Sci. Mar. 63 (suppl. 1): 303-314. Balech, E. 1954. Zoogeographic division of the South American coast. Rev. Biol. Mar. (Chile) 4: 184-195. [In Spanish.) Brattstrom, H., and A. Johanssen. 1983. Ecological and regional zoogeography of the marine benthic fauna of Chile. Sarsia 68:289-339. Briggs, J. C. 1995. Global biogeography. Developments in palaeon- tology and stratigraphy, 475 p. Elsevier Science B.V., Amsterdam, Netherlands. Camus. P. 2001. Marine biogeography of continental Chile. Rev. Chil. Hist. Nat. 74:587-617. |In Spanish.] Chen, L. 1971. Systematic, variation, distribution and biology of rockfishes of the subgenus Sebastoiuus (Pisces Scorpae- nidae, Sebastes). Bull. Scripps Inst. Oceanogr. Calif. 18:115. Cortes, E. 1997. A critical review of methods of studying fish feed- ing based on analysis of stomach contents: application to elasmobranches fishes. Can. J. Fish. Aquat. Sci. 54:726-738. Digby, P. G., and R. A. Kempton. 1987. Multivariate analysis of ecological communities, 206 p. Chapman and Hall, New York, NY. Duarte, W., F. Jara, and C. Moreno. 1980. Energetic contents of some benthonic invertebrates from the Chilean coast and annual fluctuation in Myti- lus chilensis Hupe 1854. Bol. Inst. Ocean. Sao Paulo (Brasil). 29:157-162. [In Spanish.] Gonzalez, P., and C. Oyarziin. 2003. Diet of Chilean sandperch, Pinguipes chilensis (Perciformes, Pinguipidae) in southern Chile. J. Appl. Ichthvol. 19:371-375. Hixon, M. A., and J. P. Beets. 1993. Predation, prey refuges, and the structure of coral- reef fish assemblages. Ecol. Monogr. 63:77-101. Hyslop, E. 1980. Stomach contents analysis: a review of methods and their application. J. Fish Biol. 17:411-429. Kong, I. 1985. Review of the Chilean species of Sebastes (Osteich- thyes, Scorpaeniformes, Scorpaenidae). Estud. Oceanol. 4:21-75. [In Spanish.] Lancellotti, D. A., and J. A. Vasquez. 2000. Biogeographical patterns of benthic macro- invertebrates in the Southeastern Pacific littoral. J. Biogeogr. 26:1001-1006. Lucifora, L. 0., J. L. Valero, C. S. Bremen, and M. L. Lasta. 2000. Feeding habits and prey selection by the skate Dipturus chilensis (Elasmobranchii: Rajidae) from the southwestern Atlantic. J. Mar. Biol. Assoc. UK. 80: 953-954. Moreno, C. A. 1981. Development of the studies about trophic rela- tionship in fishes from Antarctic and Sub-Antarctic rocky sub-littoral from Chile. Medio Ambiente (Chile). 5:161-174. [In Spanish. I Moreno, C. A., W. E. Duarte, and J. H. Zamorano. 1979. Latitudinal variation of number of fish species from the rocky sub-littoral: an ecological approach. Arch. Biol. Med. Experiment. (Chile). 12:169-178. [In Spanish.] Moreno, C. A., and J. H. Zamorano. 1980. Food selectivity in two benthic fish species (Mugi- loides chilensis and Calliclinus geniguttatus). Bol. Inst. Oceanogr, Sao Paulo (Brasill. 29:245-249. [In Spanish.] Ojeda, P., and J. Farina. 1996. Temporal variations in the abundance, activity, and trophic patterns of the rockfish, Sebastes capensis, off the central Chilean coast. Rev. Chil. Hist. Nat. 69:205-211. Pequefio, G. 2000. Delimitations and biogeographical relationships of the Southeastern Pacific fishes. Estud. Oceanol. 19:53-76. [In Spanish.] Pinkas, L., M. Oliphant, and Y. Iverson. 1971. Food habits of albacore, bluefish tuna and bonito in California waters. Calif. Dep. Fish, and Game. 152:1-105. Retamal, M. A. 1981. Catalogue of the decapods crustacean of Chile. Gayana Zool. 44:1-110. [In Spanish. 1 Reznick, D. N., F. H. Shaw, F. H. Rodd, and R. G. Shaw. 1997. Evaluation of the rate of evolution in natural populations of guppies iPoecilia reticulata). Science 275:1934-1937. Silva. M., and J. Stuardo. 1985. Feeding and trophic relationship between demersal fishes and the benthos of Bahia Coliumo (Concepcion, Chile). Gayana Zool. 49:77-102. [In Spanish.] Vogler, R., A. C. Milessi, and R. A. Quiiiones. 2003. Trophic ecology of Squantina Guggenheim on the continental shelf off Uruguay and northern Argentina. J. Fish Biol. 62:1254-1267. Zar, J. 1999. Biostatistical analysis. 4"' ed., 663 p. Prentice- Hall Inc., Englewood Cliffs, NJ. 498 Abstract — We investigated the use of otolith morphology to indicate the stock structure of an exploited ser- ranid coral reef fish, Plectropomua leopardus, on the Great Barrier Reef (GBR), Australia. Otoliths were mea- sured by traditional one- and two- dimensional measures (otolith length, width, area, perimeter, circularity, and rectangularity), as well as by Fourier analysis to capture the finer details of otolith shape. Variables were compared among four regions of the GBR separated by hundreds of kilometers, as well as among three reefs within each region, hundreds of meters to tens of kilometers apart. The temporal stability in otolith structure was examined by compar- ing two cohorts of fully recruited four-year-old P. leopardus collected two years before and two years after a significant disturbance in the southern parts of the GBR caused by a large tropical cyclone in March 1997. Results indicated the presence of at least two stocks of P. leopar- dus, although the structure of each stock varied depending on the cohort considered. The results highlight the importance of incorporating data from several years in studies using otolith morphology to discriminate temporary and possibly misleading signals from those that indicate persistent spatial structure in stocks. We conclude that otolith morphology can be used as an initial step to direct further research on groups of P. leopardus that have lived at least a part of their life in different environments. The use of otolith morphology to Indicate the stock structure of common coral trout (Plectropomus leopardus) on the Great Barrier Reef, Australia Mikaela A. J. Bergenius (contact author)' Gavin A. Begg^ Bruce D. Mapstone^ ' CRC (Cooperative Research Centre) Reef Research Centre and School of Marine Biology and Aquaculture James Cook University Townsville, Queensland 4811, Australia 2 CRC Reef Research Centre James Cook University Townsville, Queensland 4811, Australia 3 Antarctic Climate and Ecosystems CRC Private Bag 80 Hobart. Tasmania 7001, Australia Email address for M, Bergenius mikaela bergeniusm'icu.edu au Manuscript submitted 14 February 2005 to the Scientific Editor's Office. Manuscript approved for publication 10 November 2005 by the Scientific Editor. Fish. Bull. 104:498-511 11041. Marine fish populations are generally distributed over large geographical ranges in a heterogeneous environ- ment. Variable physical and biological processes may restrict the exchange of dispersive larvae and adults between areas within a population's species range, resulting in groups of indi- viduals that are phenotypically or genetically distinguishable. Genetic and environmental processes may also have variable effects on the pro- ductive capacity (e.g.. growth and reproduction) of individuals in dif- ferent areas. Such variations may be directly measurable with variable life history characteristics, but also indirectly measurable with phenotypic characteristics, such as meristic and morphological characteristics. Groups of individuals with different genetic or phenotypic characteristics can be defined as separate stocks. Although the precise definition of a stock has been widely debated (see re- views by MacLean and Evans, 1981; Begg, 1998; Booke. 1999), its ultimate meaning should depend on the man- agement objective related to its use. If the management objective is to protect the genetic diversity of a spe- cies, for example, genetic information should be sought. If the purpose is to prevent over-fishing and localized depletion, information about life his- tory characteristics is required. This information is needed because groups with different life history characteris- tics may respond differently to fishing pressure and therefore have different vulnerabilities to over-fishing (Cole, 1954; Adams. 1980; Jennings et al., 1998). Variations in morphological charac- teristics of otoliths have proved use- ful for identifying stocks for a range of temperate marine fishes (e.g.. Bird et al., 1986; Castonguay et al., 1991; Smith, 1992; Campana and Cassel- man, 1993; Friedland and Reddin, 1994; Begg et al.. 2001; Smith et al., 2002), but have not been examined for tropical reef fishes. Differences in morphological characteristics be- tween putative stocks indicate that the stocks have spent some periods of their lives in different environments (Begg et al.. 1999; Cadrin, 2000) and therefore have the potential to develop different life history characteristics. Otolith morphological characteristics used as indicators of stock separa- tion generally fall within one of three categories. The first category includes the traditional one-dimensional linear measurements of size-related attri- Bergenius et al Use of otolith morphology to indicate stock structure of Pleclropomus leopardus 499 butes, such as otolith length and width (e.g., Begg and Brown, 2000; Bolles and Begg, 2000) and distances between specific features on the otolith (e.g., Turan, 2000). Internal otolith measurements, such as nucleus length (e.g., Messeih, 1972; Neilson et al., 1985) and width of hyaline bands or increments (e.g., Begg et al., 2001) also fall within this category. The second cat- egory comprises two-dimensional size measurements, such as area, perimeter (e.g., Campana and Casselman, 1993; Begg and Brown, 2000; Bolles and Begg, 2000) and different shape indices, including circularity and rectangularity (e.g.. Friedland and Reddin, 1994; Begg and Brown, 2000; Bolles and Begg 2000. Tuset et al., 2003). A third, more recent morphological technique examines the two-dimensional outline of otolith shape using Fourier analysis (e.g.. Bird et al., 1986; Smith, 1992; Campana and Casselman, 1993; Begg and Brown, 2000; Smith et al., 2002). Fourier analysis produces a series of cosine and sine curves from the coordinates of a traced outline which, when added together, describe the outline of the traced form. The cosine and sine curves can be defined mathematically in a series of Fourier descriptors and used as variables to compare otolith shapes among individuals or potential stocks (Christopher and Waters, 1974; Younker and Ehrlich, 1977). Plectropomus leopardus (common coral trout) (also known as leopard coralgrouper, FishBaseM is the most important commercially and recreationally harvested reef fish on the Great Barrier Reef (GBR), Australia (Mapstone et al.'-; Williams''). Plectropomus leopardus comprises between 35% and 50% of the commercial reef line catch annually (Mapstone et al.-) and in 2004 a total allowable commercial catch (TACC) of 1300 t was implemented. Regional (hundreds of km) or inter-reef (hundreds to thousands of m) variations have been demonstrated in some life history characteristics of P. leopardus on the GBR (Begg et al., 2005), such as differ- ences in density (Ayling et al.^), reproductive strategies (Adams, 2002), size and age (Russ et al., 1995; Lou et al., 2005), and mortality (Russ et al., 1995, 1998; Map- ' FishBase. http://www.fishbase.org/search.php [accessed June 2006]. 2 Mapstone, B. D, J. P. MacKinley, and C. R. Davies. 1996. A description of the commercial reef line fishery log book data held by the Queensland Fisheries Management Authority, 480 p. Department of Primary Industries, Queensland, Brisbane. (Available from the Department of Primary Indus- tries and Fisheries, GPO Box 46, Brisbane, Queensland 4001, Australia.] ^ Williams, L. E. 2002. Queensland's fisheries resources. Current condition and trends 1988-2000, 180 p. Department of Primary Industries, Queensland, Brisbane. (Available from the Department of Primary Industries and Fisheries, GPO Box 46, Brisbane, Queensland 4001, Australia.] "Ayling, A., M. A. Samoilys, and R. Dan. 2000. Trends in common coral trout populations on the Great Barrier Reef, 36 p. Information Series QI00063. Department of Primary Industries, Queensland, Brisbane. (Available from the Department of Primary Industries and Fisheries, GPO Box 46, Brisbane, Queensland 4001, Australia.] stone et al., 2004). Current management arrangements (such as TACC, fish size limits, gear restrictions, recre- ational bag limits, and spatial and temporal closures), however, do not incorporate the localized or regional spatial structure in the life history characteristics of P. leopardus or any other exploited species on the GBR. The overall aim of this study therefore was to ex- amine the use of otolith morphology for determining the stock structure of P. leopardus on the GBR. We investigated the broad spatial scale of P. leopardus by comparing aspects of otolith morphology among fish collected from four regions of the GBR, separated by hundreds of kilometers (north to south). Otolith struc- ture was also assessed at finer spatial scales, among P. leopardus collected from neighboring reefs separated by hundreds of meters to tens of kilometers, within each of the four regions. In addition, because temporal variation in otolith shape could confound the spatial information if samples were taken from only one time, we also compared otolith morphological characteristics from two cohorts of P. leopardus with non-overlapping life histories either side of a significant environmen- tal disturbance that affected the southern half of the GBR (the large and persistent Cyclone Justin in IVIarch 1997). Spatially variable effects of the Cyclone, such as a significant drop in temperature and salinity in large parts of the GBR (AIMS'), provided us with a unique opportunity to test the temporal stability of spatial pat- terns in otolith morphology. Methods and data analysis Background Common coral trout ^Plectropomus leopardus) were col- lected as part of the Cooperative Research Centre for the Great Barrier Reef World Heritage Area (CRC Reef) Effects of Line Fishing (ELF) experiment (Campbell et al., 2001; Mapstone et al., 2004). The ELF experiment, which began in 1995 and concluded in 2006, monitored line-caught fish populations from a group of six neigh- boring reefs in each of four regions extending over 7° of latitude along the GBR (Fig. 1; Mapstone et al.'\ 1996, 1997, 2004; Davies et al.M. At the start of the experi- ' AIMS (Australian Institute of Marine Science). 2005. Un- publ. data. (Available from http://www.aims.gov.au/pages/ facilities/weather-stations/aws-ytd.html) (Accessed on 1 February 2005]. « Mapstone, B. D.. C. R. Davies, D. C. Lou, A. E. Punt, G. R. Russ, D.A.J. Ryan, A. D. M. Smith, and M. Williams. 1998. Ef- fects of line fishing experiment 1995-1997: progress report, 86 p. (Available from the CRC Reef Research Centre, PO Box 772, Townsville, Queensland 4810, Australia.] " Davies, C. R., B. D. Mapstone, A. Ayling, D. C. Lou, A. E. Punt, G. R. Russ, M. A. Samoilys, A. D. M. Smith, D. J. Welch, and M. Williams. 1998. Effects of line fishing experiment 1995-1997: project structure and operations, 28 p. [Available from the CRC Reef Research Centre, PO Box 772, Townsville, Queensland 4810, Australia.] 500 Fishery Bulletin 104(4) 120»00 E 130«00 E 140=00 E 150»00 E t 1\ V XI -. cr- South Direction ' Rocky Islets A ~^ Rocky Islets B iSKrri '%*-Dip & }f I SKtT) Figure 1 Great Barrier Reef, Australia. Study regions and reefs (1- trout tPlectropomus leopardus) were collected are shown. 1 = 3 = Mackay; 4 = Storm Cay. •4) from which common coral Lizard Island; 2 = Townsville; ment, four reefs in each region had been closed to fish- ing for 10-12 years under GBR Marine Park Zoning Plans (zoned Marine National Park B, MNP-B) and two reefs in each region had been open to fishing (zoned as General Use, GU). Two of the reefs closed to fishing remained closed during the experiment, other than to the annual research line fishing surveys. The other two closed reefs were each subjected to one year of fishing, in 1997 and 1999, after which they were closed again. The two reefs in each region that had historically been open to fishing were subjected to increased fishing pressure for one year (i.e., were temporarily opened). These reefs were then closed for five years before reverting to their original zoning status (GU). All reefs were sampled each year in the austral spring (October-December) to coincide with the peak spawn- ing period of the main target species, P. leopardus. Each reef was divided into six approximately equal- size, contiguous blocks, and sampled on a single day on each sampling occasion. Standardized commercial reef line fishing effort was distributed uniformly across two depth strata within each block. All fish caught were measured, tagged for later identification, and kept for weighing and for extraction of gonads and otoliths. For further sampling details of the ELF experiment see Davies et al.^ and Mapstone et al. (2004). Samples of P. leopardus were aged by CRC Reef staff using standard- ized methods developed by Ferreira and Russ (1994). Sample collection Otolith morphological variables were analyzed from four-year-old P. leopardus collected in 1995 and 1999 from three reefs within each of the four regions (Fig. 1; Table 1). Four years is the youngest age at which P. leopardus in all regions are fully recruited to the fishing gear used in the ELF Experiment. Although the zoning status of a reef and the level of fishing pressure on it were unlikely to have affected otolith structure, one reef from each treatment regime in the ELF experiment was included from each region to avoid potential biases in spatial variation related to particular fishing histories. Thus, within each region, samples were analyzed from one MNP-B reef that was closed to fishing, one MNP- Bergenius et al Use of otolith morphology to indicate stock structure of Plectropomus leopardus 501 Doisal Ventral Figure 2 Example of a common coral trout (Plectr'opoiniis leopardus) otolith reconstructed by using lAi all of the 128 Fourier descriptors that were collected, and (Bi the first and last 14 descriptors. Otolith orientation and measures of length and width are indicated. Table 1 Study regions, reefs, and years for which the common coral trout t Fleet Reef and examined for otolith morphology. •opomiis leopardiis) was sampled over the Great Barrier Region Reef (GBRMPA number) 1995, 1999 replicates Total length range (mm) Lizard Island Rocky Islet Reef A 1 14132a) Rocky Islet Reef B a4132b) South Direction Reef (14147) 20,6 20,6 20,12 304-495 306-481 350-468 Townsville Knife Reef (18081) Dip Reef (18039) Yankee Reef (18074) 8,4 7,12 20,8 370-451 318-462 322-464 Mackay unnamed reef (20142) Liff Reef (202961 Robertson Reefs (no2) (20136) 20, 18 20, 16 20, 11 322-456 303-427 342-514 Storm Cay Junk reef (21131) Lorries reef (21 1301 Sullivan reef (21124) 20, 15 20, 11 18,19 345-422 328-475 280-424 Total 213, 138 303-514 B reef that was open for one year of fishing an(i then closed, and one GU reef that was subjected to increased fishing for a year prior to closure for five years. Because otoliths from only four-year-old fish were analyzed in 1995 and 1999, the two cohorts included individuals with nonoverlapping life histories. Conse- quently, individuals collected in 1995 were not exposed to the influence of the unusually large and persistent Cyclone Justin that influenced the southern half of the GBR throughout March in 1997, unlike those four-year- old fish collected in 1999. Standardizing sampling by age also minimized the potential for confounding spatial variation in otolith shape with ontogenetic changes. Sagittal otoliths from up to 20 four-year-old P. leop- ardiis were sampled from each reef each year (Table 1). Otoliths from fewer than 20 fish were analyzed only if less than this number were collected from a reef. This sampling design enabled the examination of broad (re- gion) and fine (reef within region) spatial and temporal (1995 and 1999 cohorts) patterns in the otolith morph- ology of P. leopa?-dus across much of the GBR, span- ning an unusual environmental event (Cyclone Justin, 1997) that had the potential to significantly influence the results. Morphological analysis A microscope image (lOx magnification) was projected onto a computer screen by using a video camera (Pana- sonic GP-KR222E, Panasonic, Matsushita Communica- tion Industrial Co., Osaka, Japan. Whole otolith area, length, perimeter, and width (Fig. 2) and two shape indices (circularity and rectangularity) were recorded from each otolith by using the OPTIMAS image analy- sis system (OPTIMAS, vers. 6.51, Silver Spring, MD). Rectangularity was calculated as the area of the otolith divided by the area of its minimum enclosing rectangle, and circularity as the perimeter of the otolith squared divided by its area. The perimeter of the otolith was traced in a counter clockwise direction and digitized into 128 x-y equidistant coordinates by using the distal edge of the otolith rostrum as a common starting point 502 Fishery Bulletin 104(4) for the coordinates. A fast Fourier transform (FFT) was calculated as a Cartesian FFT. The Cartesian FFT uses the 128 x-y coordinates as complex numbers (a + ib), where a is the real component and Ib the imaginary component, representing the amplitudes of the cosine and sine waves, respectively. The resultant 128 set of complex numbers or descriptors were subsequently normalized for differences in otolith position by set- ting the 0"'^ descriptor to 0. and for size and rotation of the otolith by dividing all the descriptors with the first descriptor. The normalized descriptors (a' + ;b') were used to calculate the absolute value (harmonic) of each descriptor according to the following equation (Christopher and Waters, 1974): Harmonic = J a'^f + {ib') . The harmonics were then used in combination with the other morphological variables and shape indices to compare otoliths between cohorts and among regions and reefs within regions. The higher the number of equidistant points and subsequent complex numbers included in the model, the closer is the fit to the original shape. The main features of the otolith shape, however, are generally captured by the first 10-20 harmonics (e.g., Campana and Casselman, 1993; Friedland and Reddin, 1994). The minimum number of Fourier descriptors required to explain at least 90% of the recorded shape of the otoliths in our study was calculated similarly to the range-finding procedure of Smith et al. (2002). A total of 128 descriptors were collected from two randomly selected otoliths from each reef and cohort (24 in to- tal) and normalized for position, size and rotation as described above. The shape of each otolith was recon- structed (by computing the inverse FFT) by using all the descriptors and then reconstructed by using only the first and last descriptors. The Euclidian distance between the inverse FFT using all the descriptors and the inverse FFT using only the first and last descrip- tors was defined as the maximum percent error of re- construction, i.e. 100% reconstruction error (Smith et al., 2002). Because the Cartesian descriptors are asymmetrical around the middle frequency, both ends of the array are required in the reconstruction. Oto- lith shape was reconstructed, therefore, by using the first two and last two descriptors, the first three and last three descriptors, and so on until the first and last 22 descriptors were used. This range-finding test allowed us to estimate the decrease in mean percent reconstruction error by using more and more descrip- tors and it was estimated that 14 of the first and last descriptors were required for the reconstruction error to be less than 10%. See Figure 2 for a comparison of the otolith shape reconstructed from the first and last 14 descriptors and all 128 descriptors. These descrip- tors, therefore, were used in the statistical analyses to compare the spatial and temporal patterns in otolith shape of P. leopardus. Statistical methods The assumption of normality and homogeneity of vari- ance for each morphological variable was examined by using Shapiro-Wilk's and Levenes tests, respectively, and homogeneity of the group covariance matrix by Box's M test. Variables of circularity, breadth and area were log||, -transformed and Fourier harmonics 2, 4, 9, 11-14, 120-121, 114-118, and 123-127 square-root-transformed to conform to the assumption of normality and homoge- neity of variances. A relationship between otolith shape and otolith growth rate (assumed to be correlated to fish length) may confound spatial or temporal differences in oto- lith shape (Campana and Casselman, 1993). We mini- mized the potential for such effects by 1) including only fish with a fork length (FL) between 280 and 514 mm (overall FL for four-year-olds sampled during the ELF experiment in 1995 and 1999 ranged from 250 to 551 mm), and 2) standardizing morphological variables by fish FL where a significant relationship existed between the variable and FL before further analyses. The effect of FL on each morphological variable was examined by analysis of covariance (ANCOVA; Winer et al., 1991). Our primary interests in these analyses were 1) to test whether morphological variables differed with FL for any group of samples; and 2) if so, to test whether the slopes of regressions of morphological variable on FL were homogeneous among groups. If a significant re- gression was detected and homogeneous among groups, the effect of FL was removed from each measurement by using the relationship 0,. Kb 0+b.iFL-MFL. where O otA = otolith morphological measurement of fish / adjusted to mean fork length of group j; O,^ = original otolith morphological measure- ment for fish i from group /; b. = slope of the relationship 0,^:FL common to all groups; FL^^ = fork length offish / in group j; and MFL.^ = average fork length within group j. If significant slopes of the relationship differed among groups, the correction for FL was made separately for each group by using the equation above, but by replac- ing the common slope (6.) with the group-specific slope (6^1. These corrections had the effect of scaling all mor- phological variables from all otoliths to their predicted group mean FL. Multivariate analysis of variance (MANOVA; Tabach- nick and Fidell. 1983) was used to investigate the ef- fects of sex (females, males, and individuals in the process of changing sex; i.e., transitional fish) on otolith shape (MANOVA). A total of 302 of the 351 individuals had been examined for sex, of which 166 were females, 123 males and 13 transitional fish. Separate MANOVAs were computed for the one- and two-dimensional shape Bergenius et al Use of otolith morphology to indicate stock structure of Plectropomus leopordus 503 variables and the Fourier harmonics by using a three- way crossed model (outlined below) with fixed factors: cohort, region, and sex. Data were pooled across reefs within regions owing to insufficient sample numbers to test for reef-specific effects. A MANOVA was used to test for spatial and temporal differences in otolith shape. A principal component (PC; Tabachnick and Fidell, 1983) analysis was done first on the combined data set of both the shape variables and Fourier harmonics to reduce the number of variables to be incorporated in the MANOVA. The number of PCs to extract and subsequently include in the MANOVA was determined by examining the size of the eigenvalues (representing the variance explained by each PC), as well as their relative contribution to the percent vari- ance explained compared to the other eigenvalues (i.e., scree test; Tabachnick and Fidell, 1983). The latter determines the number of PCs beyond which the addi- tion of more PCs would contribute little to the variance explained by the solution (Tabachnick and Fidell, 1983). Wilk's lambda criterion was used to test for group dif- ferences in the MANOVAs. Sums of squares and de- grees of freedom of interactions were pooled when the F-ratios of interaction effects were <1. Pooling increases the degrees of freedom for the denominator and conse- quently the power of the test of remaining (unpooled) effects in the analyses. A posteriori univariate analysis of variance (ANOVA) was used to explore patterns for each of the PCs sepa- rately when significant effects were indicated in the MANOVA. The univariate linear model for the analysis of each PC was ^ijkl ■ H +C, +R^ +r(R)i^,j,+CR„ + Cr(fl ),,,„,, -He,„^ ikn where .v ;,, = the PC score for otolith / from cohort /. region j and reef k ; fi = the estimate of the population mean PC score over all cohorts, regions, reefs, and otoliths; C, = the fixed effect of cohort i averaged over regions and reefs; R I = the fixed effect of region j averaged over cohorts and reefs; r{R) I. = the random variation attributable to reef/; within regionj averaged over cohorts; and ^i(:kji ~ unexplained random variation associated with otolith / within cohort ;, region / and reef k. Tukey's honestly significant difference (HSD) test was used to determine which means differed following significant effects detected in the ANOVAs. The com- munalities (representing the proportion of the total variance of a variable accounted for by the PC) and variable loadings of the PCs that were significant in the ANOVAs were subsequently examined. A loading below 0.45 indicated that the variable explained less than 207f of the PC and therefore was not interpreted further. Finally, two forward stepwise canonical discriminant analyses (CDAs; Tabachnick and Fidell, 1983) were computed by using the shape variables and Fourier harmonics to examine the otolith shape of P. leopardus in multivariate space and to investigate whether otolith shape could be used to classify samples to spatial scale and cohort of origin. The factor used as a separating variable in the CDA depended on the significant effects determined in the MANOVA (i.e., cohort, region, or reef [region], or any interactions between these factors). The CDA was used in this way as a confirmatory technique. Wilk's lambda criterion was used to test for significant differences between the discriminant functions. Jack- knife classification was used to minimize potential bias in the reclassification of individuals. Results Slopes of the relationship between FL and several oto- lith morphological variables for P. leopardus differed among a range of spatial scales and between cohorts (Table 2). The within-group slope, therefore, was cal- culated for each group according to the level at which the slopes of the relationship differed and was used to correct for the influence of FL (Table 2). For some variables, the common between-group slope was used to correct for the influence of FL because there was a significant overall relationship between the vari- able and FL, which was homogeneous among groups (ANC OVA homogeneity of slopes test,P>0.05; Table 2). No morphological variable was significantly correlated with FL after standardization. Furthermore, shape variables and Fourier harmonics were not significantly different between otoliths of females, males or transi- tional fish (MANOVA, P>0.05). The morphological data, therefore, were pooled across sex for the remainder of the analyses. Principal component analysis (PCA) Four PCs were extracted from the analysis of the com- bined data set of shape variables and Fourier harmonics and included in the MANOVA (Table 3). The communali- ties ranged between 0.11 and 0.88, and some morphologi- cal variables were better defined by the PC solution than others (Table 3). About 44% of the total variance in the morphological data was explained by the four extracted PCs (17.9%, 10.2%, 9.9% and 5.9% by PC I, II, III, IV, respectively). A combination of higher order harmonics describing the finer details of the otolith outline, and lower order harmonics, perimeter, length, and circular- ity representing the broad shape of otoliths explained most of the variation in PC I and III (Table 3). Variation in the broader details of otolith shape also accounted for most of the variation in PC II and PC IV; otolith area, breadth, perimeter, length, and harmonic eight explained most of the variation in PC II and breadth and harmonic 127 explained most of the variation in PC IV (Table 3). 504 Fishery Bulletin 104(4) Table 2 Results of the homogeneity of slopes test for the influence offish fork length (FL) on otolith morphological variables of four-year- old Plectropomus leopardus. The level of correction represents the level at which the slopes were heterogenous. For example. the level of correction "reef( region)" means that the slopes of the relation ship jetween FL an i a morphological variable differed among reefs nested within regions a nd that the slope for each reef was ut ed to correct for the influence of FL. Level of correction "FL" means that there was a significant overall •elat ionship between FL =ind a morphological variable, but that the slopes of this relationship were homogenous among reefs(regionl. regions and cohorts. Only variables with a statistically significant relation- ship with FL are shown. Variable df F P Level of correction Logjg breadth 8,306 1.99 0.0467 reef( region ) LogjQ circularity 8,306 1.99 0.0479 reefl region ) Length 3,315 4.05 0.0075 region Log JO area 8,306 2.01 0.0452 reefl region) Perimeter 8,306 2.34 0.0187 reefl region) Rectangularity 1,341 14.43 0.0002 FL Sqrt harmonic 2 1,349 11.15 0.0009 FL Harmonic 3 1,349 10.32 0.0014 FL Sqrt harmonic 4 1,349 6.90 0.0090 FL Harmonic 5 3, 314 3.44 0.0171 cohort X region Harmonic 6 7,314 2.12 0.0410 reeflregion) Harmonic 7 1,349 13.56 0.0003 FL Sqrt harmonic 9 7,307 2.30 0.0270 cohort X reefl region ) Sqrt harmonic 13 1,349 8.05 0.0048 FL Sqrt harmonic 14 7,317 3.31 0.0021 reefl region ) Sqrt harmonic 114 1,327 8.69 0.0034 vohort Sqrt harmonic 115 7,307 2.48 0.0170 cohort X reefl region ) Sqrt harmonic 116 7,317 3.95 0.0004 reeflregion) Sqrt harmonic 117 7,317 2.89 0.0061 reeflregion) Sqrt harmonic 118 3, 314 2.90 0.0354 cohort X region Harmonic 119 1,349 6.79 0.0096 FL Sqrt harmonic 120 3,325 2.91 0.0346 region Sqrt harmonic 121 7,307 2.10 0.0435 cohort X reeflregion ) Harmonic 124 3,314 4.01 0.0080 cohort X region Harmonic 125 1,349 10.79 0.0011 FL Harmonic 126 1,349 31.76 <0.0001 FL Harmonic 127 1,327 4.37 0.0373 cohort The MANOVA of the four PCs identifie(i significant differences in the PC scores of otolith shape among regions, but were not consistent between cohorts (1995 and 1999; cohortxregion interaction. Table 4). Uni- variate ANOVAs indicated that these differences were due to variation in PC II among regions and in PC IV among reefs within regions — both spatial patterns vary- ing depending on the cohort in consideration (Table 5). Differences in PC II were due to differences between Lizard Island and Mackay regions in 1995 and between cohorts in Mackay (Fig. 3A; HSD, P<0.05). The apparent inconsistency between the significant reef within region effect in the ANOVA (PC IV) and no such reef effect detected in the MANOVA is most likely explained by the difference between two reefs in the Lizard Island re- gion, for which there was a significantly lower mean PC IV score for Reef 14132a than for Reef 14147 in 1995, whereas the reverse was observed in 1999 (Fig. 3B; HSD, P<0.05). There was also a significant difference among cohorts on Reef 14132a (Lizard Island) and Reef 20296 (Mackay Region; Fig. 3B; HSD, P<0.05). Although only explaining 18% (PC II and PC IV com- bined) of the variation in our morphological data, the PC analysis indicated that the presence of at least two stocks of P. leopardus on the GBR in 1995, one in the northern part of the GBR (Lizard Island) and one in the southern part (Mackay), whereas one homogeneous stock was indicated in 1999. Bergenlus et al Use of otolith morphology to indicate stock structure of Plectropomus leopardus 505 Table 3 Variabk' cumniunalities and loadings on the four significant principal the otolith morphological data of four-year-old Plectropomus leopardus and 1999 on the Great Barrier Reef. components (PCs) explaini collected from three reefs ng 44'« of the total variance in in each of four regions in 1995 Variable Co Loadings mmunalities PC I pen PC HI PC IV Perimeter 0.88 0.39 0.69 0.51 0.01 Length 0.85 0.07 0..57 0.71 0.14 Logi,, area 0.81 -0.05 0.80 0.32 -0.27 LogjD breadth 0.80 -0.03 0.70 -0.10 -0.55 Harmonic 127 0.78 0.18 -0.16 0.65 0.55 Log,(| circularity 0.76 0.67 0.13 0.42 0.33 Sqrt harmonic 4 0.70 0.80 0.04 -0.21 -0.09 Harmonic 6 0.64 0.72 0.18 -0.27 -0.11 Harmonic 5 0.58 0.63 -0.29 0.27 -0.15 Harmonic 8 0.51 0.17 0.57 -0.38 0.09 Harmonic 10 0.49 -0.32 0.44 -0.29 0.34 Sqrt harmonic 2 0.49 0.70 0.00 -0.04 0.03 Harmonic 3 0.49 -0.40 -0.05 0.57 0.01 Harmonic 122 0.45 0.43 0.40 -0.25 0.21 Sqrt harmonic 13 0.44 0.58 -0.08 -0.07 -0.30 Sqrt harmonic 117 0.41 0.42 -0.21 0.11 -0.42 Harmonic 7 0.41 0.54 -0.23 0.06 0.25 Sqrt harmonic 120 0.40 0.37 0.26 -0.32 0.30 Sqrt harmonic 11 0.37 0.54 -0.12 0.03 -0.25 Harmonic 119 0.37 0.45 -0.30 0.22 -0.14 Harmonic 125 0.35 -0.22 0.02 0.55 0.03 Sqrt harmonic 115 0.35 0.33 -0.12 0.22 -0.42 Sqrt harmonic 121 0.31 0.39 -0.37 0.16 -0.01 Harmonic 124 0.30 0.48 0.11 0.03 0.24 Sqrt harmonic 116 0.26 0.30 0.14 -0.35 0.18 Sqrt harmonic 118 0.26 0.49 0.02 -0.11 0.06 Sqrt harmonic 114 0.18 0.19 0.00 -0.37 0.10 Sqrt harmonic 12 0.18 0.22 -0.11 0.33 -0.05 Harmonic 126 0.16 0.32 0.18 0.04 0.13 Sqrt harmonic 9 0.14 0.09 -0.20 -0.05 0.30 Sqrt harmonic 123 0.14 0.27 -0.03 -0.17 0.18 Rectangularity 0.13 -0.25 -0.17 -0.18 0.08 Sqrt harmonic 14 0.11 0.30 -0.09 -0.02 0.10 Discriminant analyses Two stepwise canonical discriminant analyses (CDAs) were computed for the morphological variables ( Fourier harmonics and shape variables combined) in response to the significant cohort xcegion effect in the MANOVA of the PC scores. First, separate CDAs were done for each region using cohort as the separating variable and second, separate CDAs were done for each cohort by using region as the separating variable. The region- specific CDA, where cohort was used as a separating variable, showed a significant first discriminant func- tion (DF) for all four regions, accounting for 100% of the variance in the data (Table 6). Some separation was observed in the distribution of the DF I scores between the two cohorts within each region, although the tem- poral pattern was not consistent across regions (Fig. 4). Likewise, the cohort specific CDA using region as a separating variable showed some separation between the four regions in the first two significant DFs of cohort 1995 (explaining 59.9% and 30.7% of the vari- ance, respectively) and one significant function of cohort 1999 (explaining 100% of the variance. Table 6). The means of the DF I scores in the Lizard Island and 506 Fishery Bulletin 104(4) Townsville regions were distinctly greater than those of the Mackay and Storm Cay regions in 1995 O 00 CL §-(55 S -1,0 B 0 1995 O 1999 * ^ ^ I T- CM — — — — .- — — CM Lizard Island Townsville Pi/lackay Storm Cay Figure 3 Mean (A) principal component (PC) II scores per region and cohort, and (B) PC IV scores per reef within region and cohort of four-year-old common coral trout iPlectro- pomuf; leopardus). collected from three reefs in each of four regions in 1995 and 1999 on the Great Barrier Reef the Mackay and Storm Cay regions in 1999 (HSD, P<0.05; Fig. 5B). Different combinations of harmonics 9, 114, 118 de- scribing the finer details of the otolith shape, and har- monic 124 and shape variables area and circularity representing the broad details of otolith shape, were selected into the DFs to maximize the separation of cohorts within regions (Table 7). The primary predictor variables selected for separating regions in each of the cohorts were similar to those selected in the region-spe- cific CDA, although the variables were not the same in each cohort (Table 7). Between 66.79f and 76.3'7f of P. leopardus within a region could be correctly classified to their cohort of origin (Table 8). Fewer cohort specific individuals could be classified to a region however; I] ^ 0 ■ 8 - 6 - 4 - 0 nil ^m 1995 1 ■ 1 1999 ,1^ (\.w. . . B m n DFI score Figure 4 Frequencies of discriminant function (DF) I scores from otolith morphological variables (Fourier harmonics and shape variables combined) per cohort (sampled in 1995 and 1999) for regions (A) Lizard Island, (B) Towns- ville, (C) Mackay, and (D) Storm Cay for four-year-old common coral trout iPlectropomus leopardus), from the Great Barrier Reef (The data were pooled across reefs within regions.) Bergenlus et dl Use of otolith morphology to indicate stock structure of Plecliopomus leopardus 507 34.3% and 39.7% were correctly classified for cohorts sampled in 1995 and 1999, re- spectively (Table 9). Discussion Spatial and temporal variation in otolith structure Table 5 Results of ANOVAs comparing otohth principle component (PC) scores of otolith structure from two cohorts of four-year-old Ptectropomus leopardus collected from three reefs in each of four regions in 1995 and 1999 on the Great Barrier Reef Only final analyses resulting from pooled terms with F<1 are shown. P-values in bold indicate significant differences (P<0.05). Only PCs with significant differences are shown. MS=mean square. This study detected differences in the otolith shape of P. leopardi/s at different spatial and temporal scales across the Great Barrier Reef (GBR), Australia. Although the spa- tial patterns in otolith shape of P. leopardus were not always consistent among different types of shape variables, two main inferences can be made from our results. Firstly, our results indicate that otolith morphology can be useful for identifying groups of individuals of this species that are likely to have spent a significant part of their lives in different environments and therefore may indicate potential stock separation. Overall otolith shape suggested the presence of at least one southern and one northern stock along the GBR. The Storm Cay and Mackay regions consistently belonged to the south- ern stock and the Lizard Island region to the northern stock, and it was inconclusive as to which stock indi- viduals in the Townsville region belonged. It is possible that the main morphological features of otoliths are established in the larval stage and that large amounts of larval dispersal and mixing between areas results in overlap of signatures between stocks. Finer details of otolith shape, however, are most likely influenced by environmental processes during the post larval phase and may therefore provide insights to the separation of stocks during postlarval life. Although otolith shape also suggested variation at a smaller, among reef, spatial scale (hundreds to thou- Variable Factor df MS PC II PC IV Cohort 1 1.5165 1.60 0.2067 Region 3 0.8057 0.47 0.7126 Reef( region) 8 1.7209 1.82 0.0732 Cohort X region 3 3.0900 3.26 0.0217 Cohort 1 0.3246 0.11 0.7473 Region 3 3.2901 8.11 0.0083 Cohort X region 3 3.5574 1.21 0.3675 Reeft region! 8 0.40.59 0.43 0.9011 Cohort X reef ( region ) 8 2.9452 3.14 0.0020 sands of m) the significant differences in the multi- variate morphological measurement (PC IV) occurred only between two reefs in the Lizard Island region, and therefore it seems unlikely that such small-scale separation is a common phenomenon on the GBR. Nevertheless, the presence of some difference among reefs emphasizes the importance of careful and suffi- cient sampling in order to capture the range of values manifest in small-scale, presumably random, variation within regions and to avoid erroneously ascribing to regional structure what are really a reflection of local- scale variation. Given the hydrodynamic mixing over scales of kilometers, we consider it unlikely that reef- scale variations in morphological variables represent persistent environmentally induced patterns in stock structure. Table 6 Significance test of cohort-specific and i egion-specific canonical discriminant analysis (CDA) of shape variables and Fourier harmonics combined from otoliths of fou r -year-old Plectropoinus leopardus collected from three reefs in each of four regions in two cohorts, 1995 and 1999 on the Great Barrier Reef Only significant discriminant functions ( DF ) are shown DF= = discriminant function; df=degrees of freedom. Scale Factor DF Wilks' lambda X- df P Spatial scale Lizard Island Cohort I 0.7818 19.21 2 <0.0001 Townsville Cohort I 0.6863 20.89 3 0.0001 Mackay Cohort I 0.7685 26.34 4 <0.0001 Storm Cay Cohort I 0.7683 25.43 3 <0.0001 Temporal scale 1995 Region I to III 0.7573 56.72 12 <0.0001 1995 Region II to III 0.8918 23.36 6 0.0006 1999 Region I 0.91 11.95 3 0.0075 508 Fishery Bulletin 104(4) The second inference from our results is that region- al patterns in otolith shape are not always consistent among cohorts and may be subject to temporally dy- namic events such as large-scale environmental per- C/5 tl -0.4 - 0-8 06 0.4 0.2 ■ 0.0 -0.2 ■0.4 -0.6 B Lizard Island Townsville Mackay Storm Cay Figure 5 Mean ±SE of discriminant function (DF) I scores from otolith morphological variables (Fourier harmonics and shape variables combinedl per region for lA) cohort 1995, and (B) cohort 1999 in four-year-old common coral trout iPlectropomux leopardus) from the Great Barrier Reef. (The data were pooled across reefs within regions.) turbations. Ignoring such temporal signals could give misleading information about stock structure. A single homogenous stock of P. leopardiis may have been sug- gested if otolith structure based on the 1999 PCA re- sults alone were considered, whereas two poten- tial stocks would be inferred from both the PCA and CDA analyses of the 1995 data (Figs. 2A; 5A). In addition, cohorts were distinguishable within all regions based on the frequency dis- tributions of DF scores. Other investigations of the interannual stability in otolith morphology of marine fishes have found differences between years, as well as age groups (Castonguay et al., 1991; Campana and Casselman, 1993; Begg and Brown, 2000; Begg et al., 2001). Consequently, when using otolith morphology to investigate stock structure, it is preferable to compare mul- tiple cohorts with individuals of the same age, collected over several years. Such a sampling design would minimize confounding spatial variation in otolith shape with particular times of sampling and derive a time-averaged assess- ment of the spatial structure of a stock. The spatial and temporal patterns in otolith shape are in part consistent with the assess- ment (based on otolith microchemistry) of stock structure of P. leopardus on the Great Barrier Reef (Bergenius et al., 2005). Differences in otolith chemistry of P. leopardus collected at the same spatial and temporal scales as those considered in this study have indicated the presence of two or three regional stocks. Thus, although the number and potential boundaries of phenotypic stocks based on otolith chemistry and otolith structure remain uncertain, the combined results of these studies emphasize the potential presence of several stocks of P. leopardus on the GBR and indicate a north- south demarcation. Differences in both otolith Table 7 Canonical coefficient function representing correlation between the mor phological variables (Fourier harmonics and shape vari- ables combined) and significant discrimi nant functions (DFs), separatin g two cohorts (1995 andl999(offoui -year-old Plec ropo- mus leopardus within each of four region s of the Great Barriei Reef. All variables selected in the models are s hown, but var ables with loadings <0.45 are not interpreted further Lizard Island Townsville Mackay Storm Cay 1995 1995 1999 Variable DFI DFI DFI DFI DFI DFII DFI Logjo area -0.62 0.59 -0.37 Rectangularity 0.38 Logjy circularity -0.58 1 Sqrt harmonic 9 0.49 0.57 Sqrt harmonic 114 0.62 Sqrt harmonic 115 -0.24 Sqrt harmonic 118 0.84 -0.46 0.23 Sqrt harmonic 123 0.35 Sqrt harmonic 124 0.22 0.49 0.45 0.52 Sqrt harmonic 125 0.41 -0.12 0.69 Bergenius et a\ Use of otolith morphology to indicate stock structure of Pleaiopomus leopaidus 509 Correctly classified indivi dus per cohort (1995 and Barrier Reef. Table 8 duals of fuur-year-old Plectropomus leopar- 1999) in each of four regions of the Great Cohort Lizard Island Townsville Mackay Storm Cay 1995 63.79 1999 73.91 Total 66.70 71.43 83.33 76.30 78.33 70.45 75.00 76.79 63.64 71.00 shape and chemistry indicated that the movements of aduhs of P. leopardus may be limited and are consis- tent with the presumed sedentary (reef-specific) habit of most serranids after settlement (e.g.. Chapman and Kramer, 2000; Stewart and Jones, 2001). Our results confirm those of several tagging studies of P. leopardus on the GBR which showed that individuals are unlikely to move among reefs after settlement (Davies, 1995; Zeller. 1998; Zeller and Russ, 1998). Such limited in- ter-stock movement of adults means that the recovery rate of stocks after significant harvest is largely (if not solely) dependent on some unknown level of larval dis- persal. Further investigations, therefore, are required to clarify the implications of this potential stock struc- ture of P. leopardus to ensure sustainable harvests at the appropriate scale for both management and the stock structure of the fish. Reasons for temporal and spatial variation In otolith structure Little is known about what processes influence the expression of traits responsible for otolith shape (Ihssen et al., 1981). Differences in growth rate has been the main reason given for differences in linear morphological variables of otoliths (e.g., Reznick et al., 1989; Secor and Dean, 1989; Begg et al., 2001), as well as outline shape variables (e.g., Castonguay et al., 1991; Smith, 1992; Campana and Casselman, 1993; Begg and Brown, 2000). In our study, P. leopardus subsamples were restricted to a certain size range to minimize the potential of a growth rate effect that could confound spatial or cohort related differences in otolith shape. Significant differ- ences in growth rates (indicated by mean FL of four- year-olds), however, were apparent among reefs within regions in the complete sample from the ELF experiment (Bergenius et al.**), whereas there was no consistency in the patterns of variation in FL and any of the otolith morphological variables. It is possible that the large variation in FL among reefs could have masked potential Table 9 Correctly classified individua s of four year old Plectropomus leopardus per region on the GBR in each of two cohorts (1995 and 1999). Region 1995 1999 Lizard Island 39.66 17.39 Townsville 28.57 58.33 Mackay 45.00 53.49 Storm Cay 41.07 11.36 Total 39.70 34.30 differences in broad-scale regional differences in growth rates, or that some other developmental rate, such as maturation or reproductive output, were more important in influencing the otolith shape of P. leopardus. Spatial and temporal variability in otolith shape would be expected in P. leopardus populations along the GBR given the natural variability in biological and hydrodynamic factors between reefs and regions (Wolan- ski, 1994), which in turn could affect various fish de- velopmental rates. There was no monotonic latitudinal trend apparent in any of the otolith morphological vari- ables related to a temperature gradient. The maximum difference in monthly average sea surface temperature among the latitudes examined in our study is typically less than 2°C (Lough, 1994). In contrast, there is re- gional (broad spatial scale) variability in the upwelling and inflow of cool nutrient rich oceanic water across the continental shelf of the GBR (e.g., Andrews and Gen- tien, 1982; Andrews, 1983; Wolanski, 1994; Middleton et al., 1995) which could potentially influence fish de- velopmental rates through changes in food availability or "step-wise" patterns in water temperature. There is also a possibility that the temporal variabil- ity in otolith shape of P. leopardus collected before and after Cyclone Justin in 1997 was related to associated temperature or other climatic changes. Cyclone Justin was an unusually large cyclone that remained in the Coral Sea adjacent to the GBR for over three weeks during March 1997. It caused a large and rapid drop in water temperature over the southern half of the GBR (south of ~17-18°S) during this time, and bottom water temperatures dropped in just one month (March) to be- low the average winter minimum not usually reached until July (AIMS''). Our study revealed a significant dif- ference in otolith shape between nonoverlapping cohorts sampled in 1995 and 1999 in the Mackay region, which is one of the regions where the temperature changes would have been greatest. Bergenius M. A. J.. B. D. Mapstone, G. R. Russ, and G. A. Begg. 2005. Unpubl. data from the Effects of line fishing experiment (Mapstone et al., 2004). CRC Reef Research Centre, Townsville, Australia. [Data are on files at the CRC Reef Research Centre, James Cook University, Townsville, Queensland 4810, Australia.] Management and future directions Although differences in otolith shape of P. leopardus may give an indication of stock structure, an examination of the relevance of these patterns for fisheries management should be the next step. A recent increase in the total 510 Fishery Bulletin 104(4) area of the GBR Marine Park (designated as marine national park zones [or no-take zones] ) from about 5% to 33% enhances the potential for displacement of Ash- ing effort between regions. If the proposed stocks have different life history characteristics and, potentially, different fishery productivities, less productive stocks may be subjected to greater harvest levels and increased risk of local depletion. Information on vital life history characteristics (e.g., growth, mortality, and maturity) is therefore needed to examine the potential impacts of fishing and related harvest strategies for the postulated stocks. Moreover, until more is known about how larvae relate between stocks, prudent and precautionary man- agement should recognize groups of fish that function as individual units or stocks in response to harvest and management (Haddon and Willis, 1995). Otoliths are often collected during routine sampling undertaken as part of the monitoring procedures and assessments of exploited fish stocks. As a result there may be large archives of otoliths available for mor- phological analyses in most fisheries laboratories. We therefore suggest that patterns in otolith morphology provide a cost and time effective starting point for di- recting further research on groups of harvested coral reef fishes that have lived at least parts of their life in different environments. Acknowledgments Funding for this work was provided by the Australian Cooperative Research Centres Program through the Cooperative Research Centre for the Great Barrier Reef World Heritage Area, and the Fisheries Research and Development Corporation, Great Barrier Reef Marine Park Authority and James Cook University. We thank the numerous members of the CRC Reef Effects of Line Fishing Team for the collection and processing of sam- ples, especially Dong Chung Lou, Ashley Williams, and Gary Russ for aging Plect?-opomiis leopardus. Simon Rob- ertson provided guidance with the Fourier analysis. Literature cited Adams, P. B. 1980. Life history patterns in marine fishes and their consequences for fisheries management. Fish. Bull. 78:1-12. Adams, S. 2002. The reproductive biology of three species of Plec- tropomus (Serranidae) and responses to fishing. Ph. D. diss., 167 p. James Cook University, Queensland, Australia. Andrews, J. C. 1983. Thermal waves on the Queensland shelf. Aust. J. Mar. Freshw. Res. 34: 81-96. Andrews, J. C, and P. Gentien. 1982. Upwelling as a source of nutrients for the Great Barrier Reef ecosystems: a solution to Darwin's ques- tion? Mar. Ecol. Prog. Ser. 8:257-269. Begg, G. A. 1998. A review of stock identification of haddock, Melanogrammus aeglefinus, in the northwest Atlantic ocean. Mar. Fish. Rev. 60:1-1.5. Begg, G. A., and R. W. Brown. 2000. Stock identification of haddock Melanogrammus aeglefinus on Georges Bank based on otolith shape analysis. Trans. Am. Fish. Soc. 129:935-945. Begg, G. A., K. D. Friedland, and J. B. Pearce. 1999. Stock identification and its role in stock assess- ment and fisheries management: an overview. Fish. Res. 43:1-8. Begg, G. A., B. D. Mapstone, A. J. Williams, S. Adams, C. R. Davies, and D. C. Lou. 2005. Multivariate life-history indices of e.xploited coral reef fish populations used to measure the performance of no-take zones in a marine protected area. Can. J. Fish. Aquat. Sci. 62:679-692. Begg, G. A., W. J. Overholtz, and N. J. Munroe. 2001. The use of internal otolith morphometries for identification of haddock iMelanogi-ammus aeglefinus^ stocks on Georges Bank. Fish. Bull. 99:1-4. Bergenius, M. A. J., B, D. Mapstone, G. A. Begg, and C. D. Murchie. 2005. The use of otolith chemistry to determine stock structure of three epinepheline serranid coral reef fishes on the Great Barrier Reef, Australia. Fish. Res. 72:253-270. Bird, J. L., D. T Eppler, and J. D. M. Checkley. 1986. Comparisons of herring otoliths using Fourier series shape analysis. Can. J. Fish. Aquat. Sci. 43:1228-1234. Bolles, K. L., and G. A. Begg. 2000. Distinction between silver hake (Merluccius bilin- earis) stocks in the U.S. waters of the northwest Atlan- tic based on whole otolith morphometries. Fish. Bull. 98:451-462. Booke, H. E. 1999. The stock concept revisited: perspective on its history in fisheries. Fish. Res. 43:9-11. Cadrin, S. X." 2000. Advances in morphometric identification of fishery stocks. Rev. Fish Biol. Fish. 10:91-112. Campana, S. E., and J. M. Casselman. 1993. Stock discrimination using otolith shape analysis. Can. J. Fish. Aquat. Sci. 50:1062-1083. Campbell, R. A., B. D. Mapstone, and A. D. M. Smith. 2001. Evaluating large-scale experimental designs for management of coral trout on the Great Barrier Reef. Ecol. Appl. 11:1763-1777. Castonguay, M., P. Simard, and P. Gagnon. 1991. Usefulness of Fourier analysis of otolith shape for Atlantic mackerel (Scomber scombrus) stock discrimination. Can. J. Fish. Aquat. Sci. 48: 269-302. Chapman, M. R., and D. L. Kramer. 2000. Movements of fishes within and among fring- ing coral reefs in Barbados. Environ. Biol. Fish. 57:11-24. Christopher, R. A., and J. A. Waters. 1974. Fourier series as a quantitative descriptor of mio- spore shape. J. Palaeontol. 48:697-709. Cole, L. M. 1954. The populations consequences of life history phenomena. Q. Rev. Biol. 29:103-137. Davies, C. R. 1995. Patterns of movement of three species of coral Bergenius et al Use of otolith morphology to indicate stock structure of Plectiopomus leopoidus 511 reef fish on the Great Barrier Reef Ph.D. di.s.'i.. 212 p. James Cook University, Queensland. Australia. Ferreira, B. P., and G. R. Russ. 1994. Age validation and estimation of growth rate of the coral trout Plectropomus leopardus ( Lacepede 1802) from Lizard island. Northern Great Barrier Reef. Fish. Bull. 92:46-57. Friedland, K. D., and D. C. Reddin. 1994. Use of otolith morphology in stock discrimina- tions of Atlantic salmon ^Salmon salar). Can. J. Fish. Aquat. Sci. 51:91-98. Haddon, M., and T. J. Willis. 1995. Morphometric and meristic comparison of orange roughy iHoplostethus atlanticun: Trachichthyidae) from the Puysegur Bank and Lord Howe Rise, New Zealand, and its implications for stock structure. Mar. Biol. 123:19-27. 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. Jennings, S., J. D. Reynolds, and S. C. Mills. 1998. Life history correlates of responses to fisheries exploitation. Proc. R. Soc. Lond. B 265:333-339. Lou, D. C. B. D. Mapstone, G. R. Russ, C. R. Davies, and G. A. 2005. Using otolith weight-age relationships to predict age based metrics of coral reef fish populations at dif- ferent spatial scales. Fish. Res 71:279-294. Lough, J. M. 1994. Climate variation and El Nino-Southern Oscil- lation events on the Great Barrier Reef: 1958 to 1987. Coral reefs 13:181-195. MacLean, J. A., and D. O. Evans. 1981. The stock concept, discreteness of fish stock, and fisheries management. Can. J. Fish. Aquat. Sci. 38:1889-1898. Mapstone, B. D., R. A. Campbell, and A. D. M. Smith. 1996. Design of experimental investigations of the effects of line and spearfishing on the Great Barrier Reef, 86 p. Technical Report No 7. CRC Reef Research Centre Ltd, Townsville, Australia. [ISBN 1876054050] Mapstone, B. D., C. R. Davies, L. R. Little, A. E. Punt. A. D. M. Smith, F. Pantus, D. C. Lou. A. J. Williams, A. Jones, G. R. Russ, and A. D. MacDonald. 2004. The effects of line fishing on the Great Barrier Reef and evaluations of alternative potential manage- ment strategies, 205 p. Technical Report No 52. CRC Reef Research Centre, Townsville, Australia. [ISBN 18760548911 Mapstone, B. D., C. R. Davies, and J. W. Robertson. 1997. The effects of line fishing on the Great Bar- rier Reef: available evidence and future directions, p 178-192. In Proceedings of The Great Barrier Reef; science use and management a national conference, vol 1. CRC Reef Research Centre, Townsville, Australia. [ISBN 0642230234] Messeih, S. N. 1972. Use of otolith in identifying herring stocks in the southern Gulf of St. Lawrence and adjacent waters. J. Fish. Res. Board Canada 29:1113-1118. Middleton, J. H., P. Coutis, D. A. Griffin, A. Macks, A. McTag- gart, M. A. Merrifield. and G. J. Nippard. 1995. Circulation and water mass characteristics of the southern Great Barrier Reef. Aust. J. Mar. Freshw. Res. 45:1-18. Neilson, J. D., G. H. Geen, and B. Chan. 1985. Variability in dimensions of salmonid otolith nuclei: implications for stock identification and microstruc- ture interpretation. Fish. Bull. 83:81-89. Reznick. D., E. Lindbeck. and H. Bryga. 1989. Slower results in larger otoliths: an experimental test with guppies iPoecilia reticulata). Can. J. Fish. Aquat. Sci. 46:108-112. Russ, G. R., D. C. Lou, and B. P. Ferreira. 1995. A long-term study on population structure of the coral trout Plectropomus leopardus on reefs open and closed to fishing in the central Great Barrier Reef Tech- nical report No 3, 30 p. CRC Reef Research Centre, Townsville, Australia. [ISBN 1876054026] Russ, G. R., D. C. Lou, J. B. Higgs, and B. P. Ferreira. 1998. Mortality rate of a cohort of the coral trout, Plectropomus leopardus, in zones of the Great Barrier Reef Marine Park closed to fishing. Mar. Freshw. Res. 49:507-511. Secor, D. H., and J. M. Dean. 1989. Somatic growth effects on the otolith - fish size relationship in young pond-reared striped bass, Morone SaxatiUs. Can. J. Fish. Aquat. Sci. 46:113-121. Smith, K. M. 1992. Regional differences in otolith morphology of the deep slope red snapper Etelis carhunculus. Can. J. Fish. Aquat. Sci. 49:795-804. Smith, P. J., S. G. Robertson, P. L. Horn. B. Bull, O. F. Anderson, B. R. Stanton, and C. S. Oke. 2002. Multiple techniques for determining stock rela- tionships between orange roughy, Holostethus atlanti- cus, fisheries in the eastern Tasman Sea. Fish. Res. 58:119-140. Stewart, B. D., and G. P. Jones. 2001. Association between the abundance of piscivorous fishes and their prey on coral reefs: implications for prey-fish mortality. Mar. Biol. 138:383-397. Tabachnick, B. G., and L. S. Fidell. 1983. Using multivariate statistics, 509 p. Harper and Row, Publishers. New York, NY. Turan, C. 2000. Otolith shape and meristic analysis of herring iClupea harengus) in the North-East Atlantic. Arch. Fish. Mar. Res. 48:213-225. Tuset, V. M., I. J. Lozano, J. A. Gonzalez, J. F. Pertusa, and M. M. Garcia-Diaz.. 2003. Shape indices to identify regional differences in otolith morphology of comber, Serranus cabrilla. J. Appl. Ichtyol. 19:88-93. Winer, B. J., D. R, Brown, and K. M. Michels. 1991. Statistical procedures in experimental design, 3'"'' ed., 1057 p. McGraw-Hill, New York, NY. Wolanski, E. 1994. Physical oceanographic processes of the Great Bar- rier Reef, 194 p. CRC Press, Inc, Boca Raton. FL. Younker, J. L., and R. Ehrlich. 1977. Fourier biometrics: harmonic amplitudes as multi- variate shape descriptors. Syst. Zool. 26:336-342. Zeller, D. C. 1998. Spawning aggregations: patterns of movement of the coral trout Plectropomus leopardus (Serranidael. Mar. Ecol. Prog. Ser. 162:253-263. Zeller, D. C, and G. R. Russ. 1998. Marine reserves: patterns of adult movement of the coral trout (Plectropomus leopardus (Serranidae)). Can. J. Fish. Aquat. Sci. 55:917-924. 512 Abstract — The western butterfish iPcntapodus vitta) is numerous in the bycatch of prawn trawling and recre- ational fishing in Shark Bay, Western Australia. We have thus determined crucial aspects of its biological char- acteristics and the potential impact of fishing on its abundance within this large subtropical marine embayment. Although both sexes attained a maxi- mum age of 8 years, males grow more rapidly and to a larger size. Maturity is attained at the end of the first year of life and spawning occurs between October and January. The use of a Bayesian approach to combine inde- pendent estimates for total mortality, Z, and natural mortality, M, yielded slightly higher point estimates for Z than M. This result indicates that P. vitta is lightly impacted by fishing. It is relevant that, potentially, the individuals can spawn twice before recruitment into the fishery and that 73% of recreationally caught individu- als are returned live to the water. Biological characteristics and mortality of western butterfish iPentapodus vitta), an abundant bycatch species of prawn trawling and recreational fishing in a large subtropical embayment Jason C. Mant Michael J. Moran Stephen J. Newman Western Australian Fisheries and Marine Research Laboratories Department of Fisheries, Government of Western Australia 39 Northside Drive Hillarys (Boat Harbour), Western Australia 6025, Australia S. Alex Hesp Norman G. Hall Ian C. Potter (contact author) Centre for Fish and Fisheries Research School of Biological Sciences and Biotechnology Murdoch University South Street Murdoch, Western Australia 6150, Australia Email address for I C. Potter: i.potter@murdoch,edu,au Manuscript submitted 14 October 200.5 to the Scientific Editor's Office. Manuscript approved for publication 14 November 2005 by the Scientific Editor. Fish. Bull. 104:512-520(2006). The family Nemipteri(iae (threadfin breams) is found in tropical and sub- tropical waters of the Indo-west Pacific (Nelson, 1994). The western butter- fish iPentapodus vitta) (also known as "striped whiptail"') is a small benthic nemipterid, which is endemic to West- ern Australia (Russell, 1990) and very numerous in Shark Bay on the west coast of this state (Heithaus, 2004). It is also very abundant in the bycatch of the prawn trawl and recreational fisheries in this large subtropical envi- ronment. However, as with the vast majority of bycatch species in embay- ments, no estimate has been made of the mortality of P. vitta due to fishing. Furthermore, although P. vitta is very abundant in certain environments, there are no data on the age composi- tion, growth, and reproductive biol- ogy based on validated age data for a population of this species. Indeed, previous studies of the Nemipteridae have focused almost entirely on spe- cies ofNemiptei'is and Scolopsis, which are widely distributed throughout the Indo-Pacific, where they are an important component of commercial and artisinal fisheries (Sainsbury and Whitelaw, 1984; Young and Martin, 1985; Murty et al., 1992; Zacharia, 1998). The present study elucidates key aspects of the biology of P. vitta and, in particular, length-weight relation- ships, size and age compositions, growth, size and age at maturity, duration of the spawning period, and mortality. Focus is also placed on ascertaining whether this species is hermaphroditic, as is the case with certain other nemipterids (Young and Martin, 1985). The question of whether the mortality of a species with the "characteristics" of P. vitta is particularly attributable to sub- ' FishBase. http://www.fishbase.org/ search. php [accessed June 2006]. Mant et a\ Biological characteristics and mortality of Pentapodus vilta 513 stantial trawling and recreational fishing in an em- baynient has also been explored. For this purpose, we have used a model that reconciles the often imprecise and conflicting estimates of total and natural mortality that are obtained by using traditional approaches (Hall et al., 2004). Materials and methods Sampling regime 766 P. vitta. ranging from 90 to 198 mm in fork length (FL), were collected by trawling throughout Shark Bay in November and December 1997. Trawling was conducted at night by using twin 11-m prawn trawl nets with 50- mm mesh in the panels and 44-mm mesh in the codends, as are employed by commercial prawn trawl operators in Shark Bay. Each trawl lasted 20 minutes. The samples collected were used for estimating the mortality of west- ern butterfish in Shark Bay. Pentapodus vitta {/i = 339), ranging from 31 to 215 mm in fork length, were also collected in many months from 1999 to 2004 by using trawls, hook-and-line fishing, and beach seines. These samples were used to obtain aging, growth, and repro- ductive data in the study. Relationships of total length and weight to fork length The fork length (FL) of each fish, and the total length (TL) of each fish whose caudal fin was not damaged, were measured to the nearest 1 mm. The relationship between TL and FL of each sex was then calculated by using the equation TL = a + b X FL, to enable the TL of any fish with a damaged caudal fin to be estimated from its FL. The weight of each fish (W) was recorded to the nearest 1 g. The relationships between the fork length and weight of females and males were described by the equation log W = log a + 6 log FL. Analysis of covariance (ANCOVA, c<=0.05) was used to determine whether the relationships between TL and FL and between weight and FL of the males and females of P. vitta were significantly different (Zar, 1999). film. These sections were then mounted on glass slides and viewed with transmitted light under a dissecting microscope. Validation that a single growth zone is formed each year in otoliths of P. vitta was undertaken by using marginal increment analysis. The distance between the outer edge of the outermost opaque zone and the periph- ery of the otolith section (the marginal increment) was measured on the sectioned otoliths of each fish caught between June 1999 and May 2000. The marginal incre- ment was then expressed either as a proportion of the distance between the primordium and the outer edge of the opaque zone, when only one opaque zone was pres- ent, or as a proportion of the distance between the outer edges of the two outermost opaque zones, when two or more opaque zones were present. All measurements were made with the Leica Image Manager 1000 (Leica Microsystems AG, Wetzlar, Germany). Ages were assigned to each P. vitta on the basis of the number of opaque zones (annuli) in sectioned otoliths and by taking into account the dates of capture and the birth date of the fish. The birth date was desig- nated as 1 December because this date corresponded approximately to the mid-point of the period when, on the basis of gonadosomatic indices and the trends shown by gonadal and oocyte development, P. vitta spawn (see "Discussion" section). The opaque zones were counted in the region between the primordium and the proximal surface along the ventral and dorsal margins of the sulcus acousticus, where the opaque bands were usu- ally well defined and thus easily identified. The opaque zones in each otolith were counted without knowledge of the length or date of capture of the fish from which that otolith had come. The precision of the counts from 175 P. vitta otolith sections made by the senior author (JCM) on two sepa- rate occasions and between JCM's second count and a count made by another experienced reader, was as- sessed by using the index average percent error (lAPE) of Beamish and Fournier (1981). The lAPE values were low (i.e., 0.7%) for the two counts made by the senior author (JCM)) and for JCM's second count and that of the independent reader (i.e., 3.9%), demonstrating a high level of precision for the otolith readings. Because the senior author was more experienced in reading the otoliths of P. vitta, his second round of counts (and these were almost invariably the same as in the first round of counts) was used for determining the ages of individual fish. Validation of the aging method and age determination The sagittal otoliths of each P. vitta were removed, cleaned, and stored in labelled paper envelopes. Because a preliminary examination demonstrated that the growth zones were often easier to discern in sectioned than in whole otoliths, all fish were aged by using sectioned otoliths. Transverse sections of each otolith, 0.3-0.5 mm thick, were cut through the primordium by using a low- speed jewellery saw and then ground using 9-|um lapping Growth Fish caught by all methods in all months (except those obtained during extensive trawling in November and December 1997) — fish that collectively covered the full size range of P. vitta, were used to provide data for con- structing growth curves. The large samples collected by trawling in November and December 1997 were excluded because they would have introduced an excessive bias towards the larger fish at younger ages (see Ricker 514 Fishery Bulletin 104(4) [1975] for importance of limiting bias by using data from multiple sampling methods). Von Bertalanffy growth equations were fitted to the fork lengths of the female and male fish at their estimated ages at capture by using nonlinear regression in the Statistical Package for the Social Sciences software (SPSS 12.0.2 for Windows, SPSS Inc., Chicago, IL). The lengths at age of small fish whose sex could not be determined were allocated randomly to both the male and female data sets used for calculating the growth curves. The von Bertalanffy growth equation is L (1 exp" 'o'). where L, = the estimated length at age t\ L ^ = the asymptotic length; k = the rate at which L, approaches L ,; and ^1, = the hypothetical age at zero length. A likelihood-ratio test was used to compare the growth curves of the females and males of P. uitta (Cerrato, 1990). The hypothesis that a common growth curve could be fitted for the two sexes was rejected at the a= 0.05 level of significance if the test statistic, calculated as twice the difference between the log-like- lihoods obtained by fitting a common growth curve for both sexes and by fitting separate growth curves for each sex, exceeded X'jQ^- where q is the difference between the numbers of parameters (i.e., 3) in the two approaches. Reproductive biology The gonads of each fish that could be sexed macroscopi- cally were removed and weighed to the nearest 0.01 g and, in the case of ovaries, staged macroscopically by using the scheme of Laevastu (1965) and the histological characteristics of the ovaries. For histology, ovaries of a random subsample of female P. vitta from each month of the year were preserved in Bouin's fixative for 48 hours, after which they were dehydrated in an ascending series of alcohol concentrations, embedded in paraffin wax, cut into 6-;(m thick sections, and stained with either Ehrlich's haematoxylin and eosin or Mallory's trichrome (Humanson, 1972). On each histological slide, the maxi- mum and minimum diameters of 30 randomly selected oocytes, which had been sectioned through their nuclei, were recorded to the nearest 5 fim. The mean of these two measurements for each oocyte was considered to represent the diameter of that oocyte (see Foucher and Beamish, 1980). The gonadosomatic indices (GSIs) of females and males al year of age were determined from the equation W1/W2 X 100, where Wl = weight of gonad; and W2 = weight of whole fish minus the weight of gonad. Mortality Age-frequency distributions for female and male P. vitta caught during extensive trawling in November and December 1997 were used to estimate the instan- taneous rate of total mortality (Z) for both females and males. Age classes on the descending limb of the age-frequency distribution for each sex (i.e., immedi- ately after the peak in this distribution), were consid- ered fully recruited to the fishery (Ricker, 1975). The catch curve for each sex was analysed by using the assumptions that Z and the levels of annual recruit- ment are constant and that the age composition of fully recruited fish represents a random sample from a multinominal distribution with uniform selectivity from the age of full recruitment (Hall et al., 2004). The value of Z was estimated by maximizing the log-likelihood by using the SOLVER routine in Micro- soft™ Excel (vers. 2002, Microsoft Inc., Redmond, WA). The data for P. vitta were randomly resampled with replacement and analyzed to create 1000 sets of bootstrap estimates. The point estimate for Z was taken as the median of the 1000 bootstrap estimates. The 95% confidence limits were calculated as the 2.5 and 97.5 percentiles of the corresponding estimated values. The likelihood-ratio test used to compare growth curves was also used to compare the catch curves for the two sexes. Estimates of natural mortality, M, for female and male P. vitta were calculated from the relationship between natural mortality, growth, and water tem- perature (Pauly, 1980). This relationship was refitted to Pauly s data for 175 fish stocks by using SPSS 12.0.2 for Windows (SPSS Inc., Chicago, IL). The values of /; (per year) and L, (cm TL). estimated from the as- ymptotic FL in the growth curves derived for P. vitta, and the water temperature, T, were then inserted into SPSS to obtain point estimates and associated 95% confidence limits for M. thereby taking into account the uncertainty of the parameter estimates and the variation of the data around the regression line. The mean annual bottom water temperature in Shark Bay is 23.1°C (data provided by Department of Fisheries Western Australia). The values for mortality, derived for each sex by using the above approaches, were combined with the estimates of total mortality obtained from the maxi- mum recorded ages of the two sexes with Hoenig's (1983) regression model for fish and those obtained with a simulation approach (Hall et al., 2004). This method reconciles the inconsistencies among the in- dividual estimates of mortality and, through combin- ing the different values, improves the precision of the resulting estimates of natural and total mortality. To provide a more precise estimate, the simulation routine was modified slightly from that described by Hall et al. (2004) to use the number of males or females within the sample that were older than a specified age rather than the age associated with the oldest of those fish. Mant et a\ Biological characteristics and mortality of Pentapodus vitta 515 Results Relationships of total length and weight to fork length The relationships between the FL and TL of females and males, which covered essentially the full size range of P. vitta. did not differ significantly (P>0.05) and therefore the length data for the two sexes were pooled. The linear relationship between FL and TL for both sexes of P. vitta combined is described by the equation TL = -2.35 + 1.12 FL {n=344, r2=0.99). Moreover, because the relationship between weight and fork length of western butterfish did not differ signifi- cantly between males and females (P>0.05), these data for the two sexes were pooled. The relationship between weight and fork length for both sexes combined was described by the equation logW = log 1.22.V lO-"^ + 3.06 log FL (n=1026, /-- = 0.99). Validation of aging method The mean monthly marginal increments on sectioned otoliths of western butteriish with one, two, three, and four-seven opaque zones reached their maxima in Octo- ber, after which they declined markedly to their minima in either November or December and then rose progres- sively over the ensuing months (Fig. 1). The pronounced decline and then progressive rise, exhibited by the mean monthly marginal increments on the otoliths of western butterfish, irrespective of the number of opaque zones, demonstrates that a single opaque zone is formed in the otoliths of P. vitta each year and that the number of such zones can thus be used to age this species accurately. Growth The von Bertalanffy growth curve provided a good fit to the fork lengths at age of the females and males of P. vitta (Fig. 2). This good fit is reflected in the high coefficients of determination (r^) of 0.942 for females and 0.924 for males and by the values for f,, being close to zero (Table 1). The von Bertalanffy growth curves of the females and males of P. vitta in Shark Bay differed significantly (P<0.05), and the difference in L^ between females and males contributed more to the difference than that of either /; or ?,,. Females were more prevalent than males in length classes below 160 mm, whereas the opposite was true above this length (Fig. 3A). However, age-frequency distributions showed that females were more prevalent in all age classes (Fig. 3B). The overall sex ratio was 1.4 females:l male. Reproductive biology Because virtually all P. vitta became mature at the end of their first year of life, the GSIs for all fish al year 0.8 r 0.6 0.4 0.2 0 L 1.0 0.8 1 opaque zone 14 7 21 2 opaque zones S 0 L 1 10r 3 opaque zones I I I JJASONDJFIVIAM Montti Figure 1 Mean monthly marginal increments (±1 stan- dard error) for sagittal otoliths of the western butterfish (Pentapodus vitta), derived from samples collected between 1999 and 2004. Sample size is given for each month. The closed rectangles represent the summer and winter months and the open rectangles the autumn and spring months. of age were calculated. The mean monthly GSIs for female western butterfish remained low (<0.9) in winter (June to August) and early spring (September), but then increased sharply from September to reach a maximum of 4.0 in late spring (November), before declining mark- edly during mid-summer (January) (Fig. 4). Although the pronounced seasonal trend displayed by the mean monthly GSIs for males was very similar to that for females, the maximum mean monthly GSI of males was approximately one tenth of that for females. All female western butterfish that were al year old and caught between February and August (apart from a few in July) possessed ovaries at stage I or II (i.e., they were virgin or immature; Fig. 5). Females with 516 Fishery Bulletin 104(4) 250 r 200 - 150 •JvH-^ — T^' •T>t^ = 100 / Females /• n=133 50 f " Fork length (mm) o o 200 1,.^ •t^L'^"^- 150 ■/^ 100 / Males / n=169 50 - l f 0 0123456789 Age (y) Figure 2 Von Bertalanffy growth curves fitted to fork lengths at age for individual females and males of western but- terfish iPentapodus vitta). 180 150 120 - 90 60 - :[Jil Males (n=429) Females (n=590) L ^^/V^/b^>^,1 year old. Monthly sample sizes are given above each mean. that had developed to at least stage III by September would have progressed through to maturity during the ensuing months of the spawning period if it had not been removed from the population. In each month, the distribution of oocyte diameters in the mature ovaries of P. vitta produced a well-de- fined mode between -20 and 80 am, which represented chromatin nucleolar- and perinucleolar-stage oocytes. Cortical alveolar oocytes were present in some ovaries in July and September, when oocyte diameters had reached a maximum of 200 and 150 jam, respectively. These larger oocytes and also yolk granule oocytes were abundant between October and December and reached a maximum diameter of 380 \xva.. The contribution of cortical alveolar and yolk granule oocytes declined in January. Hydrated oocytes were found in several ova- ries in October, December, and January and postovu- latory follicles were present in some ovaries during November. Mortality The age-frequency distributions for both female and male western butterfish caught during extensive trawl- ing in November and December of 1997 had a modal peak at 2 years of age (Fig. 3B). Because the age at full recruitment was thus considered to be 3 years, catch too r _ oLB June n=8 100 r too r _ oL' July n=27 August n=13 100 r oil- September n=6 100 1 October n=14 Percentage frequency o o o o o o o o o o .■ ■ 1 1 ■ - November n = 47 December n=10 January n=12 100 r _ oL' February n=2 100 r _ March n=15 oLB 100 r _ oLI April n = 5 100 r _ May n=10 0 1 1 1 II I-II III IVV-VIVIIV Gor adal stage Figure 5 Histograms showing percent frequency of occurrence of each 1 of the various gonadal female western butterfi stages in 5h (Pent a - podus vitta) in each month of the | year. curve analysis was undertaken with data for the 3-i- and older age classes. The resulting point estimate for the instantaneous rate of total mortality, Z, was signifi- cantly greater (P<0.05) for males (1.15/year) than for females (0.91/year) (Table 2). Likewise, the simulation method of Hall et al. (2004), based on the number of individuals above 5 years of age (2 males and 13 females for sample sizes of 130 and 253, respectively), yielded a greater point estimate of Z for males (1.34/year) than females (0.89/year). The point estimate for Z, derived by inserting the maximum ages for female and male butterfish (8 years for both males and females), into the refitted Hoenig (1983) equation for fish, was 0.55/year for the two sexes (Table 2). The point estimates for the 518 Fishery Bulletin 104(4) instantaneous rate of natural mortality, M, derived by inserting the parameter values for western butterfish into the refitted Pauly (1980) equation, were higher for females (1.78/year) than males (1.28/year). The combined point estimate for Z derived using the method of Hall et al. (2004) was greater for males (1.14/ year) than females (0.89/year; Fig. 6). This method yielded a slightly higher point estimate of M for males (0.80/year) than females (0.71/year) (Fig. 6). 0.5 A B £■ ° K • S 0.3 o Q. . > 02 m 5 years old) Z 1.34 0.90 1.95 Refitted Hoenig ( 1983 ) fish equation z 0.55 0.19 1.52 Combined Z (Bayesian method) z 1.14 1.01 1.35 Refitted Pauly (1980) M 1.28 0.42 3.91 Combined M (Bayesian method) M 0.80 0.37 1.23 Females Catch curve analysis Z 0.91 0.78 0.99 Simulation (number offish >5 years old) Z 0.89 0.69 1.11 Refitted Hoenig (1983) fish equation Z 0.55 0.19 1.52 Combined Z (Bayesian method) z 0.89 0.82 1.01 Refitted Pauly ( 1980 ) M 1.78 0.56 5.35 Combined M (Bayesian method) M 0.71 0.39 0.97 ManI et al Biological characteristics and mortality of Pentapodus vitta 519 October and January implies that P. vitta spawns during these latter four months. This conclusion is consistent with hydrated oocytes or postovulatory follicles or a com- bination of both being found only in fish caught between October and January. Furthermore, the precipitous decline in the mean monthly GSIs between December and February and the presence of only stage-II ova- ries in February strongly indicates that spawning does not extend into this month. Because spawning activity peaked between October and January, the western but- terfish was given a birth date of 1 December when an age was assigned to each fish. Because mature ovaries of P. vitta often contained a relatively wide size range of vitellogenic oocytes and, in many cases, also either hydrated oocytes or postovulatory follicles, this species is considered a multiple spawner serisu deVlaming (1983), namely individual females release eggs on more than one occasion in a spawning season. Is there evidence that Pentapodus vitta is hermaphroditic? Our data demonstrate that, in the population of P. vitta in Shark Bay, the females are relatively more abundant than males in length classes below 160 mm, whereas the reverse is true of length classes above that length (Fig. 3A). In contrast, the prevalence of females was always greater than that of males in each age class. The trend towards an increasing prevalence of males vs. females with increasing body length can thus be attributed to the males growing faster. The fact that there was not a progressive shift towards an increase in one sex with increasing age and none of the gonads of the 178 fish examined contained both ovarian and testicular tissue provides overwhelming circumstantial evidence that P. vitta is gonochoristic. Like P. vitta, other nemipterid species also exhibit a size-related skew in their sex ratios that can be explained by differences in growth rates between the two sexes (e.g.. Young and Martin, 1985; Lau and Sadovy, 2001; Raje, 2002; Granada et al., 2004). The trend among gonochoristic nemipterids, for their males to grow larger than their females, pre- sumably reflects a selective advantage of being large during courtship or mating (Granada et al., 2004). In contrast to these gonochoristic species, some nemipterids are protogynous hermaphrodites, e.g., Scolopsis mono- gramma, S. taeniopterus and S. bilineatus (Young and Martin, 1985). Mortality The point estimates for natural mortality, M, derived for males and females of P. vitta with Pauly's (1980) equa- tion, were inconsistent with the point estimates derived for total mortality, Z. This conclusion is drawn from the fact that the point estimates for M were greater than those for Z from both Hoenig's (1983) equation for fish and catch curve analysis, thereby paralleling the situ- ation with several other species (see Hall et al., 2004), and also those derived from the simulation and Bayesian methods of Hall et al. (2004). The use of the method of Hall et al. (2004) reconciled the inconsistencies between M and Z for P. vitta and provided more precise estimates, particularly for M. Because the point estimates for Z were only slightly greater than those for M, the popula- tion of P. vitta in Shark Bay does not appear to have been subjected to heavy fishing pressure. In this context, it is relevant that, because the vast majority of P. vitta attain maturity by the end of their first year of life, most individuals will have had the opportunity to spawn at least twice before they become fully recruited into the trawl fishery. Furthermore, recreational anglers return about 73% about of the western butterfish they catch in Shark Bay live to the water (Sumner et al.^). Pentapodus vitta lives for up to eight years, but grows very rapidly early in life and attains maturity by the end of the first year of life. The males of P. vitta grow more rapidly and reach a larger size than their fe- males. The western butterfish is a gonochoristic species that, in Shark Bay, spawns between mid-spring and mid-summer. Although commercial and recreational fishermen obtain large numbers of P. vitta as bycatch in Shark Bay, this fishing does not appear to have a marked impact on the numbers of this very abundant species in this large subtropical embayment. Acknowledgments Gratitude is expressed to many colleagues in the Depart- ment of Fisheries, Western Australia, and the Centre for Fish and Fisheries Research, Murdoch University for help with sampling. Comments and constructive criti- cisms were kindly provided by Michael Mackie, Peter Stephenson, and Gary Jackson. Financial and technical assistance were provided by the Department of Fisher- ies, Western Australia, and Murdoch University. Literature cited 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. Cerrato, R. M. 1990. Interpretable statistical tests for growth com- parisons using parameters in the von Bertalanffy equation. Can. J. Fish. Aquat. Sci. 47:1416-1426. deVlaming, V. L. 1983. Oocyte development patterns and hormonal involvements among teleosts. In Control processes in fish physiology (J. C. Rankin, T. J. Pitcher, and R. T. Duggan, eds.), p. 176-199. Croom Helm, Beckenham, England. 2 Sumner, N. R., P. C. Williamson, and B. E. Malseed. 2002. A 12 month survey of recreational fishing in the Gascoyne bioregion of Western Australia during 1998-99. Fisheries Research Report No. 139, 54 p. Department of Fisheries, Western Australia. 520 Fishery Bulletin 104(4) Foucher, R. P.. and R. J. Beamish. 1980. Production of nonviable oocytes by Pacific hake, Merluccius procluctus. Can. J. Fish. Aquat. Sci. 37: 41-48. Granada, V. L., Y. Masuda, and T. Matsuoka. 2004. Age and growth of the yellowbelly threadfin bream NemipteruH hafhybius in Kagoshima Bay, southern Japan. Fish. Sci. 70:497-506. Hall, N. G., S. A. Hesp, and I. C. Potter. 2004. A Bayesian approach for overcoming inconsisten- cies in mortality estimates using, as an example, data for Acanthopagrus latus. Can. J. Fish. Aquat. Sci. 61:1202-1211. Heithaus, M. R. 2004. Fish communities of subtropical seagrass mead- ows and associated habitats in Shark Bay, Western Australia. Bull. Mar. Sci. 75:79-99. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mor- tality rates. Fish. Bull. 82:898-903. Humanson, G. L. 1972. Animal tissue techniques. W. H. Freeman, San Francisco, CA. Laevastu, T. 1965. Manual of methods in fisheries biology, 51 p. FAO, Rome. Lau, P. P. F., and Y. Sadovy. 2001. Gonad structure and sexual pattern in two thread- fin breams and possible function of the dorsal accessory duct. J. Fish. Biol. 58: 1438:1453. Murty, V. S., T. Apparao, M. Srinath, E. Vivekanandan, K. V. Somasekharan Nair, S. K. Chakraborty, S. G. Raje, and P. U. Zachariah. 1992. Stock assessment of thi-eadfin breams (Nemipterus spp.) of India. Indian J. Fish. 39:9-41. Nelson, J. S. 1994. Fishes of the world. John Wiley, New York, NY. Pauly, D. 1980. On the interrelationships between natural mor- tality, growth parameters and mean environmental temperature in 175 fish stocks. J. Cons. Int. Explor. Mer 39:175-192. Raje, S. G. 2002. Observations on the biology of Nemiptei-iia japonicus (Blochl from Veraval. Indian J. Fish. 49:433-440. Ricker, W. E. 1975. Computation and interpretation of biological sta- tistics offish populations. Bull. Fish. Res. Board Can. 191, 382 p. Russell, B.C. 1990. Nemipterid fishes of the world. FAO Fisheries Synopsis, 149 p. FAO, Rome. Sainsbury, K. J., and A. W. Whitelaw. 1984. Biology of Peron's threadfin bream, Nemipterus peronii (Valenciennes), from the North West Shelf of Australia. Aust. J Mar. Freshw. Res. 35:167-185. Stephenson, P. C, and N. Hall. 2003. Quantitative determination of the timing of otolith ring formation from marginal increments in four marine teleost species from northwestern Australia. Fish. Bull. 101:900-909. Young, P., and R. Martin. 1985. Sex ratios and hermaphroditism in nemipterid fish from northern Australia. J. Fish Biol. 26: 273-287. Zacharia. P. U. 1998. Dynamics of the threadfin bream, Nemipterus japonicus, exploited off Karnataka. Indian J. Fish. 45:265-270. Zar, J. H. 1999. Biostatistical analysis, 4"^ ed., 663 p. Prentice Hall, Upper Saddle River, NJ. 521 Abstract — Multibeam sonar mapping techniques provide detailed benthic habitat information that can be com- bined with the data on species-specific habitat preferences to provide highly accurate calculations of populations in a particular area. The amount of suitable habitat available for the endangered white abalone {Haliotis sorenseni) was quantified to aid in obtaining an accurate estimate of the number of remaining individuals at two offshore banks and one island site off the coast of southern Cali- fornia. Habitat was mapped by using multibeam sonar survey techniques and categorized by using rugosity and topographic position analysis. Abalone densities were evaluated by using a remotely operated vehicle and video transect methods. The total amount of suitable habitat at these three sites was far greater than that previ- ously estimated. Therefore, although present estimates of white abalone densities are several orders of mag- nitude lower than historic estimates, the total population is likely larger than previously reported because of the additional amount of habitat sur- veyed in this study. The use of multibeam sonar mapping techniques to refine population estimates of the endangered white abalone {Haliotis sorenseni) John Butler (contact author)' Melissa Neuman^ Deanna Pinkard' RIkk Kvitek^ Guy Cochrane* ' National Marine Fisheries Service Southwest Fisheries Science Center 8604 La Jolla Shores Drive La Jolla, California 92037 Email address for J. Butler: John.ButleriSnoaa.gov ^ National Marine Fisheries Service, Southwest Region 501 West Ocean Blvd Long Beach, California 90802 ^ California State University, Monterey Bay 100 Campus Center Seaside, California 93955 '' United States Geological Service 345 Middlefield Road Menio Park, California 94025 Manuscript submitted 2 May 2005 to the Scientific Editor's Office. Manuscript approved for publication 6 December 2006 by the Scientific Editor. Fish. Bull. 104:521-532 (2006>. Accurate classification of habitat is important for examinations of the dis- tribution and abundance of marine organisms. The potential complexity of species-habitat relationships is such that the use of a single parameter when trying to link species to specific areas has become obsolete. Multiple physical oceanographic parameters that link species to their habitat have proven to be useful, although they are not applicable to all types of marine organisms (Freeman and Rogers, 2003). The use of high-resolu- tion multibeam sonar maps for clas- sifying habitat has become a critical step in the process of estimating the distribution and abundance of marine animals that are known to be associated with particular habi- tat types. Because an increasing number of marine resources are in states of decline, information on the amount of habitat of specific types and qualities becomes crucial to making predictions for the purposes of fisher- ies management, and to aid in deci- sions concerning potential locations of marine protected areas (Kvitek et al.'; lampietro et al., in press). For species in states of decline, protection of known critical habitat may be the only approach available to enhance survival of the species. Detailed habi- tat mapping provides a large amount of information on the benthic makeup of an area in a relatively short amount of time, and further processing of data with GIS (geographic information system) can produce accurate esti- mates of actual areas categorized by various physical parameters. These pieces of information, coupled with field surveys to obtain information on animal or plant habitat associations. 1 Kvitek, R., P. lampietro, C. Bretz, K. Thomas, S. Zurita, B. Jones, and E. Morris. 2004. Hydrographic data acquisition in support of MLPA and MLMA Implementation. California Department of Fish and Game Final Report P0170015, p. 1-74. Foundation of California State University Monterey Bay, 100 Campus Center, Seaside, CA 93955. 522 Fishery Bulletin 104(4) can allow for detailed and accurate estimates of abun- dance of marine organisms. The white abalone (Haliotis sorenseni) is one of six ab- alone species (red, green, pink, pinto, black, and white) that have been exploited commercially on the west coast of North America. The species ranged historically from Morro Bay, California, USA, to Punta Rompiente, Baja California, Mexico (Geiger, 1999). Serial depletion of stocks of congeneric, relatively shallow-dwelling aba- lone species led to the exploitation of white abalone, the deepest living abalone species in the North Pacific (Karpov et al., 2000). During a ten-year period from 1969 to 1978, approximately 360,000 white abalone were harvested (Hobday et al., 2001). After 1978, the mandatory reporting of white abalone landings was no longer required, and although abalone fishing effort has remained high, there are (at present) no data on the numbers taken (Tegner, 1989). In surveys conducted in the late 1990s in suitable habitat, very low numbers of white abalone were re- ported (Davis et al., 1996, 1998; Haaker et al.^; Hobday et al., 2001). Densities at many localities were less than 1 abalone/ha, and the entire population of white abalone was estimated at less than 3000 individuals (Hobday et al., 2001). Reproductive efforts of the remaining adults are believed to have been seriously hindered by poten- tial AUee effects (AUee, 1931) because of low densi- ties. Concern for the survival of the species prompted a status review and the white abalone was listed as an endangered species by the Department of Commerce on 29 May 2001 (Federal Register. 2001). Postexploitation estimates of abundance and available white abalone habitat need revision because the spatial coverage of sampling in the 1980s and 1990s was lim- ited in the case of determining abundance estimates and virtually nonexistent in the case of determining habitat classification (Lafferty et al., 2004). Accurate estimates of density and the amount of suitable habitat are critical for making informed decisions regarding how to prioritize recovery actions and where to focus population enhancement efforts. In addition, identifi- cation of the specific locations of white abalone in the wild will help both to focus efforts to protect areas where the species remain and to locate potential brood stocks for captive breeding. Information on the recruit- ment potential of the remaining individuals cannot be directly assessed, but an examination of the temporal trends in size distributions may be helpful in address- ing questions of population viability and sustainability. The present study was conducted with recovery efforts in mind, and therefore included intensive sampling of offshore banks and islands where white abalone abun- dance was high historically. We used a combination of advanced mapping technology (multibeam sonar and side-scan sonar) and a remotely operated vehicle (ROV) to quantify the total amount of available habitat at each location, in order to examine the size distributions of surviving populations, and to derive new estimates of white abalone abundance. Materials and methods Habitat mapping Habitat was surveyed during three cruises on the NOAA ship RV David Starr Jordan at Tanner Bank (July 2002 and September 20041, Cortes Bank (July 2003), and San Clemente Island (August 2004) with multibeam (all years, Fig.l) and side-scan sonar techniques (2002 only). The multibeam sonar system, which was installed on the RV David Starr Jordan for the Tanner Bank 2002 and Cortes Bank 2003 cruises, included a Reson 8101 multibeam sonar (Reson Inc. Goleta, CA) supple- mented with a side-scan sonar option, and was used in conjunction with a Triton-Elics Isis System (Triton Elics International, Portland. OR) for data logging and sonar control. The San Clemente Island 2004 habitat survey was conducted by using the same sonar system aboard a smaller vessel, the RV VenTresca, in conjunction with ROV operations conducted aboard the RV David Starr Jordan. Delphmap and BathyPro software (Triton Elics International, Inc., Portland. OR) were used to create real-time side-scan mosaics and to generate digital elevation models. The pitch and roll motion of the vessel used for mapping was corrected for using a TSS HDMS (heading and dynamic motion sensor) (pitch, roll and heading accuracy ±0.02°; heave accuracy ±5% or 5 cm). Hypack Max software (Hypack, Inc., Middletown, CT) was used for survey planning and navigation. Position- ing information was provided by a Trimble 4700 GPS receiver and NavBeacon for receiving U.S. Coast Guard RTCM (radio technical commission for maritime com- munications) corrections. Sound velocity profile data were collected with an Applied Microsystems (Sydney, British Columbia, and CA) SVPlus sound velocimeter. Multibeam data were processed with CARIS HIPS soft- ware (CARIS USA, Ellicott City, MD), and all final GIS products were derived from shoal-biased data. Microhabitat analysis Habitat type and algal cover were examined for surveys conducted at Tanner Bank in 2002. Habitat was charac- terized by broad type (e.g., bank, seamount), modifiers (e.g., faulted, eroded), microhabitat (e.g., sand, boulders), seafloor slope (by degree), seafloor complexity (5 levels), and algal cover (species identification and four levels of coverage), after Greene et al. (1999; Table 1). 2 Haaker, P. L., D.V.Richards, and I. Tanaguchi. 2000. White abalone program October 9-25, 1999, cruise report, p. 1-17. California Department of Fish and Game, 330 Golden Shore Suite 50, Long Beach, CA 90802. Abalone surveys Abalone surveys were conducted with a Phantom HD 2-1-2 (2002 and 2003) and a Phantom DS4 (2004) Butler et al Multibeam sonar mapping techniques for estimates of Haliolis sorenseni 523 Table 1 Habitat analysis variable descriptions (Greene et influence the PCA write out are listed (a total of 12 following within the video-captured field of view: 75-100<'f. al., 1999). Only those classifications that were used frequently enough to were included in the analysis). For algal levels, coverage was defined by the absent = <10'^f, rare = 10-.35'*, present = 35-75"^"^, and highly abundant = Variable abbreviation Variable type Levels Mhab 1-4 Macro/Microhabitat l=boulders; 2 = deformed (faulted or folded); 3 = deformed (faulted or folded) + boulders; 4 = deformed (faulted or folded) + sand Slope 1 and 2 Seafloor slope l = flat (0 to 1 degree); 2 = sloping (1 to 30 degrees) Com 1 and 2 Seafloor complexity l=very low complexity; 2=low complexity Br 1-4 All brown algae l=absent; 2=rare; 3=present; 4=highly abundant L 1-4 Laminaria farlowii l = absent; 2=rare; 3=present; 4=highly abundant A 1-4 Agarum fimbriatum l=absent; 2=rare; 3=present; 4=highly abundant E 1-4 Eisenia arborea l=absent; 2=rare; 3=present; 4 = highly abundant D 1-4 Dictyotaceae spp. l = absent; 2=rare; 3 = present; 4 = highly abundant remotely operated vehicle (ROV) equipped with a for- ward-looking 12:1 zoom color video camera (all surveys) and a high resolution Nikon Coolpix digital still camera (2004 surveys only, Insite Pacific Inc., Solana Beach, CA). We recorded video footage continuously on digital tapes and overlaid footage with temperature, depth, heading, and time information from an on-screen dis- play. Two pairs of lasers were mounted near the front of the ROV in parallel: one pair to measure abalone upon sighting (10.0 cm apart), and one pair to estimate the search field of view (60 cm apart). Upon sighting an abalone, the ROV pilot maneuvered the vehicle into position to zoom in on and photograph the animal for species confirmation, ensuring that the narrow pair of lasers passed over the abalone so that a size estimate could be obtained. Empty abalone shells were also noted and identified. The ROV was tracked by using a directional hydro- phone mounted on the ship and a transponder mounted on the ROV. Positional data and all other navigational and physical data associated with the ROV dives (e.g., heading, depth, and water temperature) were recorded every two seconds. Because of the great depth range that white abalone inhabit, it was not logistically fea- sible to compare ROV transect observations with diver transect observations. It is possible that we missed abalone residing on the back sides of rocks as the ROV passed over. Sampling design and site selection Multibeam and side-scan surveys were conducted the preceding night at sites that were to be surveyed with the ROV during the next day. Transect locations were chosen randomly initially, but after preliminary surveys at Tanner Bank revealed a narrow depth range of aba- lone occurrence from 30 to 60 m and a habitat restriction to rock reef or the sand-and-rock reef interface, only areas of suitable habitat (depths ranging from 30 to 60 m [±5 m]) were surveyed. In general, we attempted to run transects for approximately two hours, and to a length of 1 km, although actual lengths were calculated after the dive. Several of the early surveys included transects of greater lengths of time and distances. Georeferenced bathymetry maps were interfaced with the ROV naviga- tional software for all dives, allowing for the ship, and in effect the ROV, to remain within a particular depth stratum and within suitable habitat. Postdive processing Videotapes and still photos were reviewed to confirm spe- cies identification, measure sizes, search for additional abalone, record search effort, examine microhabitat, and determine the search field of view. Other abalone spe- cies that were encountered were noted (red and pink). We used strip transect techniques to estimate densities, and therefore the field of view was measured for abalone sightings from the first survey (Tanner Bank, 2002), and the average (2 m [±0.85 SD], n=171) was used as the strip width for the entirety of each transect in the study. Data analysis For habitat classification a statistical approach called textural analysis (described in Cochrane and Lafferty, 2002) was used to identify complex rock areas at Tanner Bank as documented by side-scan sonar (Fig. 1, A and B). We estimated the total area of available white aba- lone habitat on each bank by calculating the amount of rocky substrate in depth intervals, as determined by multibeam sonar surveys. A combination of rugos- ity and topographic position analysis (TPI) were used 524 Fishery Bulletin 104(4) A 3J N- CA '^ ^ N SCI A 0.05). Site-specific white abalone population estimates were based on the amount of available habitat quantified from the results of multibeam sonar surveys and white abalone density estimates determined from data from ROV surveys. The total population of white abalone on Tanner Bank based on 2002 density measurements was estimated at 12,819 ±3582 (SE), whereas the total population estimate based on 2004 density measure- ments was 5883 ±3324. The total population on Cortes Bank was estimated at 7366 ±5340 (SE), and that at San Clemente Island was estimated at 1938 ±1598 (SE; Table 2). White abalone reproductive potential Eighty-nine percent of white abalone at all sites were observed as singletons, and only 6% were observed in groups. Group sizes ranged from two to five individuals and the largest group was observed at Tanner Bank in 2002. There were no sightings of abalone in groups at Tanner Bank in 2004. White abalone were not found frequently in groups during any of the surveys, and the observed number of individuals within 250-m- subsam- ples within the 40-50 m depth range at Tanner Bank in 2002 did not differ from values predicted by the negative binomial distribution (chi-square=4.9, P=0.67). The results of an analysis of distances between sight- ings indicated that less than 259c of individual abalone were located within 5 m of one another for all surveys. The majority of individuals (60%) were over 20 m away from another sighted abalone. The survey with the highest proportion of abalone located within a distance of 20 m of another individual was that at Tanner Bank in 2002. In regard to potential mating pairs, out of a possible 19,503 combinations of mating pairs, only 202 were within 30 m from one another at Tanner Bank in 2002, 4 out of 231 pairs at Cortes Bank, 3 out of 12 pairs at San Clemente Island, and 10 out of 861 pairs at Tanner Bank in 2004. Discussion Few areas in U.S. waters have been adequately mapped to reveal detailed bathymetry or substrate character- istics. Consequently, our knowledge of the amount of habitat available for benthic marine organisms is so inadequate that one can only speculate about overall population sizes. White abalone populations in south- ern California are no exception; all previous estimates of the amount of suitable habitat available for white abalone have been based on the assumption made by Thompson et al. (1993) that 3% of the sea floor between 25 m and 65 m was rocky substrate (Davis et al., 1998; Butler et al Multibeam sonar mapping techniques for estimates of Haliotis sorenseni 529 Cortes Bank San Clemente Island Area Habitat Area Habitat Density surveyed area Density surveyed area (/ha±SE) ihal thai Population (/ha±SEi (ha) (ha) Population 12.3 ±8.8 2.6 232 2853.6+2041.6 3.1 ±2.8 0.8 392 1215.2 ±1097.6 6.1 ±2.8 1.5 423 2580.3 ± 1184.4 2.6 ±1.8 2.4 278 722.8 ±500.4 4.0+2.7 0.7 483 1932 ±2114.1 0 0.7 219 0 8.0 ±3.1 4.8 1138 7365.9 ±5340.1 1.5 ±0.2 3.9 889 1938 ±1598 Hobday and Tegner, 2000; Hobday et al., 2001). Using this assumption, the total amount of white abalone habi- tat in southern California was estimated to be only 752 ha (Davis et al., 1998; Hobday et al., 2001). The results of surveys from the present study at only three sites (Tanner and Cortes Banks and San Clemente Island) revealed 3646 ha of rocky substrate between 30 and 60 m, which is much higher than estimates based on the 39c rocky habitat assumption. Thus, the total available habitat area for white abalone and other benthic dwell- ing organisms in southern California is likely far greater than previously reported. The definition of available white abalone habitat, however, may need further refinement if white aba- lone exhibit a preference for a certain type of rocky substrate. Cochrane et al. (2005) suggested that white abalone prefer the edges of reefs at the sand-rock in- terface rather than the areas in the middle of the reef. Although there is some qualitative evidence support- ing this idea, there have been no conclusive studies on specific habitat preferences. The only conclusion that can be drawn in relation to habitat preference from the present study is that white abalone are not typically found in areas with no rock (i.e., sand only). Even if white abalone were found only at the sand-rock interface, our estimates of available habitat would still far exceed past estimates. Although the time and effort involved in mapping habitat may lead to the tempta- tion to make broad generalizations to estimate habitat types and amounts, the results of this study have shown that specific habitat mapping is necessary to accurately identify the amounts of each type of habitat existing in a particular area. The depth distribution of white abalone at these three sites was grouped into three depth bins (30-40 m, 40-50 m, and 50-60 m) and was less stratified than predicted according to preliminary surveys at Tanner Bank. The number of abalone was highest in the mid- dle depth bin (40-50 m) at Tanner Bank during 2002 and 2004, but only significantly so in 2002. The high- est number of abalone observed at Cortes Bank was in the shallowest depth bin (30-40 m). Too few abalone were sighted at San Clemente Island and they were not included in the analysis. These results indicate that there are not clear, broad trends in depth distribution that can be applied to all sites, although the deepest depth bin did yield the fewest abalone sightings at all sites. It is also possible that the only realistic trends observed were those at Tanner Bank in 2002 because of the relatively large number of animals observed at this site. The white abalone population in California has been estimated to number less than 3000 (Davis et al., 1998; Hobday and Tegner, 2000; Hobday et al., 2001). In this study a total of between 15,187 and 22,123 white aba- lone was estimated at two offshore banks and one is- land location. Our population estimate is greater for two reasons: 1) the multibeam sonar maps revealed more habitat than was previously known to exist, and 2) search effort, and therefore the calculation of density, was dependent on suitable habitat, i.e., densities were not artificially deflated by including portions of surveys conducted in inappropriate habitat (e.g., sand). The highest densities (19.8 abalone/ha between 40 and 50 m) were observed at Tanner Bank in 2002. Previous surveys at Tanner Bank underestimated habi- tat area and included the entire search area without regard to depth or habitat to determine density. A re-examination of data from surveys conducted by the Delta submersible in 1999 at Tanner Bank (Haaker et al.-) yielded an overall density of 15.7 abalone/ha. This value varies only slightly from our overall density esti- 530 Fishery Bulletin 104(4) mate of 13.0 abalone/ha. Densities were lower overall in our surveys conducted at Tanner Bank in 2004, and the most significant differences were apparent in the 40-50 m depth range. These lower densities in sur- veys conducted only two years after initial surveys are noteworthy and worrisome, especially because Tanner Bank is a site where the population was thought to be relatively stable. The density of white abalone prior to commercial and recreational exploitation is poorly known, and the few reported estimates are the subject of considerable speculation. Tutschulte (1976) reported an estimate of white abalone density of 2300 abalone/ha, based on three 10-m- quadrats. Because of the small sam- ple size and lack of replicate sites, caution should be used when applying this estimate to all of California. However, it is important to note that one cluster of five white abalone was observed in the present study, which would lead to a local density close to that re- ported by Tutschulte (1976). It is also of interest to note that Shepherd et al. (2001) suggested that once populations of Haliotis laevigata reached densities below 0.25 abalone/m- (2500 abalone/ha) management action should be taken, and therefore estimates of 0.23 abalone/m^ reported by Tutschulte (1976) were not extremely high. More recently Rogers-Bennett et al. (2002) used his- torical landings data and the habitat area estimated by Davis et al. (1998) and Hobday et al. (2001) in an attempt to reconstruct past population densities. Ap- plying the 39c rocky habitat criteria referred to above (Discussion section, paragraph 1) (Thompson et al., 1993), Rogers-Bennett et al. (2002) calculated a density of 479 abalone/ha for all of California and a density of 1623 abalone/ha for San Clemente Island, where 75% of the landings were reported. These densities are lower than those reported by Tutschulte (2300 abalone/ha; 1976), but the more recent estimate is also too high because the proportion of rocky habitat is greater than 3%, as observed in the present study. The maximum density observed in the present study (19.8 abalone/ha) is only an order of magnitude less than Rogers-Ben- nett et al.'s (2002) calculation of historic density for all of California (479 abalone/ha), but is two orders of magnitude less than the estimates by Tutschulte (2300 abalone/ha, 1976). Regardless of the problem of inflated density estimates, the densities observed in the present study at San Clemente Island were less than 5 abalone/ha at all depths, which is astoundingly lower than the 1623 abalone/ha calculated by Rogers-Bennett et al. (2002). Whether the existing white abalone populations are viewed as viable depends largely on the validity of esti- mates of past population distributions and densities and on establishing confidence in current density estimates. Total abundance of individuals becomes irrelevant if these animals are distributed such that densities within a particular area are below the critical level necessary for successful reproduction. The relatively deep range of H. sorenseni is outside the range of dense macroalgal growth, and therefore we can assume that the ROV does an adequate job in identifying animals within its field of view (2 m). Based on the assumption that we are actually sighting nearly all of the animals that are within the search range, our data indicate that the majority of individuals at these sites are greater than 5 m (linear distance) from any other individuals along and between transects, and that many are over 30 m away from a potential mate. These distances well exceed what has been shown to be a critical minimum distance (<2 m) for successful spawning and fertiliza- tion in other species of abalone (e.g., Haliotis laevigata; Shepherd and Brown, 1993; Babcock and Keesing, 1999; Shepherd et al., 2001). The large distances between individuals coupled with density estimates that are several orders of magnitude lower than those necessary for a viable population (Shepherd et al., 2001) would indicate that white abalone populations are currently in a dire state. Despite generally large distances between individu- als, an analysis of dispersion within the population surveyed at Tanner Bank in 2002 showed a large pro- portion of individual sightings within 30 m of another sighting. Additionally, there was a high degree of con- tagion between individuals in 250-m'- subsamples at depths of 40-50 m. It may be true that white abalone at this site are aggregated on a larger scale, indicat- ing that certain habitats within their most prevalent depth range (40-50 m) may promote higher survival of white abalone. However, ultimately, if animals are not packed densely enough at smaller scales, successful spawning and fertilization will not occur. This study highlights the importance of establishing accurate den- sity estimates at appropriate scales for guiding assess- ments of population viability and future enhancement protocols. The mean sizes of white abalone observed in the pres- ent study varied little between sites and were compa- rable to those observed during 1999 submersible sur- veys at Tanner Bank and Cortes Bank (Behrens and Lafferty, 2005). The size distribution of white abalone did differ slightly between the two banks and island location. More notable was an apparent shift in the size distributions of animals observed at Tanner Bank in 2002 versus those observed in 2004. Unlike in 2002, there were no large individuals observed in 2004, in- dicating a possible die-off of older individuals and a lack of new, younger individuals to fill this size class. No white abalone smaller than 9.0 cm (approximately three years old; Tutschulte, 1976) were observed at any of the sites. If this were a self-sustaining popula- tion, smaller individuals indicating recruitment from several preceding year classes would be present, al- though the likelihood of sighting abalone recruits and juveniles younger than three years is not very high. Juvenile white abalone shells are mottled red in color and settlement occurs on pink crustose coraline algae- covered rocks, thus making them very cryptic during the first few years of life. Even during historic periods when white abalone density was much higher than cur- Butler et al Mullibeam sonar mapping techniques for estimates of Haliotis sorenseni 531 rent levels, juveniles were observed on extremely rare occasions (Owen^). Thus, the absence of small individu- als during ROV observations does not rule out recent recruitment to this population. Consequently, although recruitment in the last three years may be undetect- able, recruitment from the last three decades should be evident in our surveys. Conclusions The use of detailed sonar mapping techniques has proven to be an invaluable method to estimate amounts of dif- ferent habitat types for the purposes of quantifying marine organisms within their specific habitat. Specifi- cally, white abalone restoration efforts would be greatly enhanced by more surveys incorporating high-resolution bathymetric maps that would serve to better define the characteristics of suitable white abalone habitat. Our concept of what a viable white abalone population is would benefit from more accurate density estimates for other areas within the historic range and from a better understanding of how white abalone are distributed within populations. The ultimate goals of recovery and eventual removal of white abalone from the Endangered Species List de- pends upon establishing confidence in the demographic parameters that define a viable white abalone popula- tion. It has become apparent that conducting studies such as this one, with continuing efforts to improve sample standardization techniques over time, is critical to achieving recovery goals. Acknowledgments Multibeam sonar mapping during the 2002 cruise was funded by the National Fish and Wildlife Foundation through a grant to the California Department of Fish and Game. Dan Richards provided data from the 1999 Channel Islands National Park white abalone survey. John Wagner, Anthony Cossio, Ben Maurer, Scott Mau, and David Murfin piloted the ROV and provided tech- nical support. The officers and crew of the RV David Starr Jordan ran transects and survey lines. Chuck Oliver participated in cruises and, along with Nancy Lo, designed the microhabitiat analyses. Pat lampi- etro, Carrie Bretz, Kate Thomas, Jason Mansour, Tif- fany Van, Andrew Rapp, Bryan Jones, Josh Sampey, and Saori Zurita helped with multibeam sonar data acquisition, processing, and analysis. David Murfin, Lara Asato, and several anonymous reviewers pro- vided comments that improved earlier versions of this manuscript. ■* Owen, B. (abalone fisherman). 2001. P.O. Box 601, Gualala, CA 95445. Personal commun. Literature cited Allee, W. C. 1931. Animal aggregations: a study in general soci- ology. Univ. Chicago Press, Chicago, IL. Babcock, R., and J. Keesing. 1999. Fertilization biology of the abalone Haliotis laevi- gata: laboratory and field studies. Can. J. Fish. Aquat. Sci. 56(91:1668-1678. Behrens M. D., and Lafferty K. D. 2005. Size frequency measures of white abalone, impli- cation for conservation. In Sixth California islands symposium (D. K. Garcelon and C. A. Schwemm, eds.), p. 427-432. Institute for Wildlife Studies, Ventura, CA. Cochrane, G. R., J. L. Butler, and G. E. Davis. 2005. Refining estimates of potential white abalone habitat at northern Anacapa Island California using acoustic backscatter data. In Benthic habitats and the effects of fishing (P. W. Barnes and J. P. Thomas, eds.), p. 161-164. Am. Fish. Soc, Bethesda, MD. Cochrane, G. R., and K. D. Lafferty. 2002. Use of acoustic classification of sidescan sonar data for mapping benthic habitat in the northern Channel Islands, California. Cont. Shelf Res. 22:683-690. Davis, G. E.. P. L. Haaker, and D. V. Richards. 1996. Status and trends of white abalone at the Cali- fornia Channel Islands. Trans. Am. Fish. Soc. 125(1): 42-48. 1998. The perilous condition of white abalone, Hali- otis sorenseni, Bartsch, 1940. J. Shellfish Res. 17(3): 871-875. Elliott, J. M. 1977. Some methods for the statistical analysis of samples of benthic invertebrates. Freshw. Biol. Assoc. Sci. Publ. 25:1-150. Freeman, S. M., and S. I. Rogers. 2003. A new analytical approach to the characterization of macro-epibenthic habitats: linking species to the environment. Estuar. Coast. Shelf Sci. 56:749-764. Geiger, D. L. 1999. Distribution and biogeography of the recent Hali- otidae (Gastropoda: Vetigastropodai worldwide. Boll. Malacologico. 35(5-12):57-120. Greene, G. H., M. M. Yoklavich, R. M. Star, V. M. O'Connel, W. W. Wakefield, D. E. Sullivan, J. E. McRea, and G. M. Cailliet. 1999. A classification scheme for deep seafloor habi- tats. Oceanol. Acta 22:683-690. Hobday, A. J., and M. J. Tegner. 2000. Status review of white abalone Haliotis sorenseni throughout its range in California and Mexico. NOAA Tech. Memo. NMFS-SWR-035, 101 p. Hobday, A. J., M. J. Tegner, and P. L. Haaker. 2001. Over-exploitation of a broadcast spawning marine invertebrate: decline of the white abalone. Rev. Fish Biol. Fish. 10:493-514. lampietro, P. J., R. G. Kvitek, and E. Morris. In press. Recent advances in automated genus-specific marine habitat mapping enabled by high-resolution multibeam bathymetry. Mar. Technol. Soc. J. Jenness, J. 2002. Surface areas and ratios from elevation grid, vers. 1.2 for ArcView 3.2. Jenness Enterprises, Flag- staff, AZ. 532 Fishery Bulletin 104(4) Karpov, K. A., P. L. Haaker, I. K. Tanaguchi, and L. Rogers- Bennett. 2000. Serial depletion and the collapse of the California abalone (Haliotis spp.) fishery. Special Publication Can. J. Fish. Aquat. Sci. 130:11-24. Lafferty, K. D., M. D. Behrens, G. E. Davis, P. L. Haaker. D. J. Kushner, D. V. Richards, I. K. Taniguchi, and M. J. Tegner. 2004. Habitat of endangered white abalone, Haliotis sorenseni. Biol. Cons. 116:191-194. Federal Register. 2001. Final rule: endangered and threatened species: endangered status for white abalone. May 29, 2001. U.S. Federal Register 66(103):29046-29055. Govern- ment Printing Office, Washington, D.C. Rogers-Bennett, L., P. L. Haaker, T. O. Huff, and P. K. Dayton. 2002. Estimating baseline abundances of abalone in California for Restoration. CalCOFI Rep. (2002) 43:74-96. Shepherd S. A., and L. D. Brown. 1993. What is an abalone stock- implications for the role of refugia in conservation. Can. J. Fish. Aquat. Sci. 50(9):2001-2009. Shepherd S. A., K. R. Rodda, and K. M. Vargas. 2001. A chronicle of collapse in two abalone stocks with proposals for precautionary management. J. Shellfish Res. 20(2):843-856. Tegner, M. J. 1989. The California abalone fishery: production, eco- logical interactions, and prospects for the future. In Marine invertebrate fisheries (J. F. Caddy, ed.), p. 401-420. Wiley, New York, NY. Thompson, B.. J. Dixon, S. Shoeter, and D. J. Reish. 1993. Benthic Invertebrates, /n Ecology of the southern California bight (M. D. Dailey, D. J. Rei.sh. and J. W. Anderson, eds.l, p. 369-440. Univ. California Press, Berkeley, CA. Tutschulte, T. C. 1976. The comparative ecology of three sympatric abalone. Ph.D. diss., 335 p. Univ. California, San Diego, San Diego, CA. Vedder, J. G., H. G. Greene, S. H. Clarke, and M. P. Kennedy. 1986. Geologic map of the mid-southern California con- tinental margin. In California continental margin geo- logic map series, area 2 of 7 (H. G. Greene and M. P. Kennedy, eds.). Map 2A 1:250,000. California Division of Mines and Geology, Sacramento, CA. 533 Abstract— Three experiments were performed in an estuarine squid- trawl fishery in New South Wales, Australia, to test modifications to trawl nets. Lateral mesh openings were experimentally increased and physical bycatch reduction devices (BRDs) were placed in codends. These modifications aimed to reduce nontargeted catches of fish, while maintaining catches of the targeted broad squid [PhotoloUgo ethcridgei) and bottle squid tLoliolus noctihica). Compared to conventional codends made with 41-mm diamond mesh, codends made with different posterior circumferences and larger 45-mm mesh had no significant effect on the catches of any species. The best performing configurations involved the installation of BRDs designed to separate organisms according to differences in behavior. In particular, versions of a composite square-mesh panel reduced the total weight of bycatch by up to 71'7f and there was no significant effect on the catches of squid. The results are discussed in terms of the probable differences in behavior between fish and squid in codends. After this study, a square- mesh panel BRD was voluntarily adopted throughout the fishery. Experiments in gear configuration to reduce bycatch in an estuarine squid-trawl fishery James P. Scandol (contact author) Tony J. Underwood Centre for Research on Ecological Impacts of Coastal Cities Marine Ecology Laboratories, All University of Sydney Sydney. New South Wales 2006, Australia Present address for J, P, Scandol, NSW Department of Primary Industries PO Box 21 Cronulla New South Wales 2230, Australia Email address for J. P Scandol: James.Scandol.S'dpi. nsw.gov.au Matt K. Broadhurst NSW Department of Primary Industries Fisheries Conservation Technology Unit, POBoxJ321 Coffs Harbour, New South Wales 2450, Australia Manuscript submitted 26 April 200,5 to the Scientific Editor's Office. Manuscript approved for publication 6 December 2005 by the Scientific Editor. Fish. Bull. 104:533-541 (2006). The incidental catch of nontarget organisms (termed "bycatch") by com- mercial fishing gears will remain an important issue in the management of fisheries. Article 7.6.9 of the Food and Agriculture Organization's Code of Conduct for Responsible Fisher- ies (FAO, 1995) notes "States should take appropriate measures to mini- mize waste, discards, catch by lost or abandoned gear, catch of non-target species, both fish and non-fish species, and negative impacts on associated or dependent species, in particular endangered species," Several options are available for achieving these aims (Hall, 1996), but the most common strategy involves technological improvements to fishing gears that reduce unwanted mortality due to fishing (Kennelly and Broadhurst, 2002), Much research into gear design to reduce bycatch has been undertaken in demersal trawl fisheries and espe- cially those targeting shrimp (for re- views see Kennelly, 1995; Broadhurst, 2000), This research reflects the se- riousness of bycatch issues in these fisheries and, in particular, the inci- dental mortality of key species such as marine turtles and juveniles of commercially and recreationally im- portant fish. Considerably less work has been done on bycatch in other fisheries, including those targeting cephalopods. The global catch of squid has been estimated at 2,8 million tonnes (t) per year (FAO^), Squid are harvested with highly-selective jigs or less selective gill nets, seines and, more commonly, trawls (Rathjen, 1991; Morals da Cunha and Moreno, 1994; Simon et al,, 1996), Large-scale use of pelagic gill or drift nets has been restricted because of concern about incidental catches (e.g., Burke et al,, 1994; Piatt and Gould, 1994), Despite their widespread use, very little work has been done to modify towed gears so that they selectively harvest squid. In one of the few studies relevant to this issue. Glass et al. (1999) attached video cameras to the anterior sections of trawls targeting longfin squid (Lo- ligo pealeii) in the Atlantic Ocean and observed that squid positioned themselves considerably higher in the trawl than fish. It was suggested that this difference in behavior could be used to separate catches. ' Food and Agriculture Organization (FAO). 2005. FAO Fisheries Global Information System, www.fao.org/fi/ figis laccessed on 23 Nov 2005]. Aver- age annual global production of all squid species for 2000 to 200.3. 534 Fishery Bullelm 104(4) In New South Wales (NSW) Australia, arrow squid {Nototodarus gouldi), mitre squid (Photololigo sp.) and southern calamari (Sepioteuthis australis) are impor- tant in the legally retained bycatch from fish and prawn trawlers working in oceanic waters. In addition, two species of squid are specifically targeted by up to 20 trawlers with modified, single-rigged prawn trawls in Broken Bay (Fig. 1). Broad squid (P. etherldgei) (typi- cally 30-290 mm mantle length [ML] when harvested) represent the majority of the total catch (approximately 25-50 t per year since 1997), although the smaller bottle squid (Loliolus noctilitca) (30-90 mm ML when harvested) are also retained and sold, primarily as bait for recreational anglers. Additional details on the local biology of P. etheridgei can be found in O'Donnell (2004). The designs of trawls used to catch these squid vary among operators, but all are restricted by lengths of the headline (<11 m) and sweeps (<5 m) and minimal and maximal mesh openings of 40-60 mm in the body and 40-45 mm in the codend. To maximize catches of the smaller bottle squid, all operators use codends of the minimal legal mesh size of 40 mm, with anterior sections of 100 meshes in circumference attached to posterior sections of at least 150 meshes in circumfer- ence (made with thick twine), which are designed to reduce lateral-mesh openings (Broadhurst and Ken- nelly, 1996). The configuration of conventional codends in the Bro- ken Bay squid-trawl fishery means that these trawls are poorly selective and therefore, in addition to the targeted squid, they also retain large quantities of by- catch. Preliminary data from a 3-year observer-based study of catches in the early 1990s (DPI-) indicated a bycatch-to-squid ratio of up to 4:1 (by weight). Although a subset of this bycatch includes some species that the fishermen are legally permitted to retain, most of the bycatch is small, unwanted fish, including juveniles of several commercially and recreationally important spe- cies. Concerns over the mortality of these individuals and the negative impacts on stocks led us to test the utility of modifications to fish- and prawn-trawl codends for improving selection of species in squid trawls. Such experimental work has not previously been completed in squid-trawling fisheries. A series of explicit hypotheses were tested in the fishery. It was proposed that devices designed to reduce bycatch would 1) cause differential changes in catches, by reducing bycatch but causing no important reduction in catches of squid; 2A) not cause disproportionate by- catch of small or large fish (i.e., would not influence the distributions of sizes offish caught); 2B) not cause dif- ferential effects, so that subsequent incidental catches had similar mixtures of types and relative abundances offish species as before; and, 2C) not cause relatively ^ DPI (New South Wales Department of Primary Industries). Unpubl. data. 1992. [Data are on file at New South Wales Department of Primary Industries, PO Box 21, Cronulla, 2230, NSW, Australia.] Figure 1 Location of Broken Bay and the two areas trawled iPatonga, and Flint and Steel). Broken Bay is 35 km north of Sydney, New South Wales, Australia. larger catches of particular fish species, nor of juveniles of commercial species. Materials and methods This research was completed on two commercial squid- trawl grounds in Broken Bay NSW (Fig. 1) between February and September 2002 by using a chartered squid trawler (9.2 m in length). The vessel was rigged with a single, conventional 4-seam otter trawl made from 60-mm (all mesh sizes refer to stretched opening) polyethylene (PE) netting (headline of 11 m; Fig. 2) towed between 0.8 and 1.8 m/s over a sandy substratum at depths ranging from 5 to 15 m. A zipper (Buraschi S146R, Milan, Italy, 1.5 m in length) was attached at the posterior body of the trawl to facilitate changing codends (Fig. 2). Codends and BRDs examined Five codends of different design were made from dark PE netting with total lengths and anterior fishing cir- cumferences (i.e., the expected circumference of the codend during towing — [Broadhurst et al., 1999]) of 2.4 and 1.5 m, respectively (Fig. 3, A-E). All codends had zippers (Buraschi S146R, 1.5 m in length) attached at their anterior ends. Four of the codends were made within existing regulations and had anterior sections made from 40-mm knotted netting (2-mm diameter-o twisted twine) with a length and circumference of 35 Scandol et al Experiments to reduce bycatch in an estuaiine squid-trawl flsheiy 535 Length in meshes Length in meshes Top panel ,157 ; Bottom panel 1 4-mm 0 braided 60-mm mesh throughout 175 T 30 100T 20 T Wing panel I Zipper Figure 2 Plan of the 4-seam squid trawl used in the study. N B = bars. 20 T normals; T = transversals. and 100 meshes, respectively, and posterior sections with a length of approximately 1 m, but different mesh circumferences and sizes (all meshes were made from 3-mm 0 braided twine; Fig. 3, A-D). The first codend (41/150) represented commercially used designs with a posterior section of 41-mm mesh and a circumference of 150 meshes (Fig. 3A). The second codend (41/1001 had a posterior section made from the same mesh, with a circumference of 100 meshes (Fig. 3B). The third and fourth designs (45/150 and 45/100, respectively) had posterior sections with the same circumferences as those above, but were made from 45-mm mesh (Fig. 3, C and D). The fifth codend design (termed the "square-mesh codend") was made entirely of 30-mm mesh (1.4-mm 0 twisted twine) hung from the bar (i.e., the meshes were oriented so that they were square shaped) and had a total length and fishing circumference of 144 and 90 bars, respectively (Fig. 3E). Four variations of behavioral-type bycatch reduction devices (BRDs) were tested in combination with certain codends (Fig. 3, F-I). All BRDs were installed in the top anterior section of the relevant codend at a distance starting 1.1 m anterior to the last row of meshes in the posterior section according to the specifications provided by Broadhurst et al. (2002). The first BRD (the diamond BRD; see Broadhurst et al. [2004]) was a diamond- shaped opening (11x11 bars) cut into a conventional 41/150 codend (Fig. 3F). The second and third BRDs were composite square-mesh panels (CSMP; Broadhurst and Kennelly[1996]) with 90- or 75-mm mesh, respec- tively, hung from the bar in the main escape panels and were inserted into the conventional 41/150 and 41/100 codends. These modified codends were called the "90-CSMP" and "75-CSMP" codends (Fig. 3, G and H). The fourth BRD was inserted into the square-mesh codend (termed the "75-panel codend"; Fig. 31); it was a simple panel of 75-mm mesh hung from the bar (and having the same dimensions as that used in the CSMP codend). Experimental procedure Three independent experiments were completed during 2002 to identify an appropriate combination of codend and BRD that minimised bycatch whilst maintaining catches of squid. Experiment 1 (25 February-1 March 2002, five days, two replicate hauls/day) tested pre- dictions about the effect of 1) changing the posterior codend circumference within the maximal and minimal codend mesh sizes legally permitted in the fishery; and 2) using the diamond and CSMP BRDs located in the commercially used 41/150 codend. Experiment 2 (29 July-9 August 2002, 10 days, two replicate hauls/day) repeated experiment 1 above (but with additional repli- cation to increase the power to differentiate between the codends) and also tested the effectiveness of the square- 536 Fishery Bulletin 104(4) A 41/150 codend B 41/100 codend 45/150 codend D 45/100 codend rL Square-mesh codend 35 N Approx. 2.4 m 20 N 100 T 2-mm 0 twisted 40-mm mesh 150 T >• 100 T 1 100 T 100 T 150T 100 T -f 100 T 144 B 1 4-mm 0 twisted 30-mm mesh hung on the bar 90 B 3-mm o braided 41 -mm mesh 3-mm o braided 45-mm mesh F D, lamond codend G 90-CSMP codend H 75-CSMP codend 1 75-panel codend T 1.1 m 1 100T 150 T 4-mm o braided 90-mm mesh - hung on the bar 100 T 24 B - 5B 150T 2-mm 0 twisted 100T hung on the bar 4-mm 0 braided 6B CD m -J hung on the bar 100 T 4-mm 0 braided 6B hung on the bar ' m 90 B Figure 3 Schematic representation of the codends and bycatch reduction devices tested during the experiments. The 41/150 codend is the conventional design used in the fishery. N = normals; T = transversals, B = bars. mesh codend. Experiment 3 (26 August-6 September 2002, 10 days, three replicate hauls/day) compared the 41/100 and square-mesh codends (the most appropri- ate non-BRD codends; see Results section), with and without their respective 75-CSMP and 75-panel BRDs installed. In each experiment, the daily position and order of the codends being tested were determined randomly and they were alternately used in normal commercial hauls of 30-min duration between 0700 and 1500 hours. Depending on the number of codends being tested, we attempted two or three balanced replicates of each treatment on each day. Haul location was determined by the skipper and varied between the two squid trawl grounds in Broken Bay (Fig. 1) according to the pres- ence of jellyfish, Catostylus sp., which reduced the ef- ficiency of the trawl. Data collected from each replicate haul included the numbers and weights of all taxa and the sizes of commercially and recreationally important species. When large numbers (several hundred) of any species were caught, a randomly selected subsample (approximately 25-50% of the total) was measured. Fish were measured to the nearest 0.5 cm fork length; carapace lengths of crustaceans were measured to the nearest 0.1 cm. Results Bycatch and target squid The primary hypothesis for this study was that the devices designed to reduce bycatch would cause changes in catches, by reducing bycatch but causing no important reduction in the catches of squid. This hypothesis was tested by using univariate statistical methods applied to the results from all three experiments. In Experi- ment 1, the diamond and 90-CSMP codends caught less bycatch (about 74% less on average) than the other types of gear (Table 1). About 50% of the bycatch were discards, including teleost fish such as tailor ^Pomatonus saltatrix) and mulloway iArgyrosomus japoniciis). There were, however, no significant reductions in catches of squid, although catches were smaller with the diamond codend. As a result, later experiments were mostly done with larger mesh sizes. There were no effects of altering other components of the gear (circumference or mesh- size of the codend). Catches of squid were significantly different on the two days (Table lA), but there was no interaction with the type of codend; therefore differences among types of gear were consistent. Sources of varia- tion that were determined to be nonsignificant at P>0.25 Scandol et a\. Experiments to reduce bycatch in an estuanne squid-trawl fishery 537 Table 1 Analyses of catches using six types of gear from experiment 1 during February-March 2002. Data are weights of catches from Flint and Steel, fished for 2 different days (a random factor). Codend (a fixed factor) is the comparison of 41/150, 41/100, 45/150, 45/100, diamond and 90-CSMP (composite square-mesh panel) codends. ;; = 2 replicate hauls for each treatment on each day. A Analyses of variance Source of variation Bycatch Squid df Mean square F P Mean square F P Codend = C Day = D = D CxD Residual pPooled residual 5 1.51 3.16^ <0.04 0.24 0.63 1 0.46 0.96^ >0.30 2.50 6.51 5 0.39P 0.33P 12 0.51P 0.40P 17 0.48 0.38 >0.65 <0.05 B Mean untransformed weight (kg, [±standard error, n = 4]) of bycatch and squid per haul and Student-Newman- "<" indicates significant differences at P<0.05; "=" indicates P>0.05. Keuls tests; Codend Diamond 90-CSMP 41/100 41/150 45/100 45/150 Bycatch 7.7(2.3) Squid 2.2(0.6) 6.4(1.7) < 23.1(6.1) = 22.3(2.2) = 24.8(6.9) 3.5(0.8) = 2.9(0.7) = 3.9(1.2) = 3.2(1.2) 17.3(7.0) 4.3(1.2) were pooled and catches were transformed to natural logarithms to stabilize variances (Underwood, 1997). A second analysis compared the diamond and 90- CSMP with the conventional 41/50 codend in two ar- eas (in estuarine waters off Patonga, and Flint and Steel, see Fig. 1), to check consistency across fish- ing-grounds. A balanced set of replicated hauls with each type of gear was obtained. Again, the two BRDs caused large and significant reductions in bycatch compared with the conventional net. Although catches of squid were smaller, they were not significantly so and there was a large reduction in the ratio of weight of bycatch to weight of squid caught. Strikingly, al- though there was considerable variation from day to day and between locations in the amount of bycatch, there were no interactions with experimental gear used. Therefore, reductions in bycatch were consistent under the different conditions that are typical to the fishery. Experiment 2 compared catches with the four codends without BRDs, as in experiment 1, with catches with the square-mesh codend. There were complex differ- ences in the amount of bycatch caught with different types of gear, despite many catches being smaller than those in experiment 1. More bycatch was caught with the 45/150 codend than with the 41/150 codend (i.e., an effect of mesh size). This difference was, however, detected only for the 150 codend. There was also an ef- fect of circumference of the codends (41/100 caught more bycatch than did 41/150), but only for the commercial (41-mm mesh) codend. This finding explains the signifi- cant interaction between mesh-size and circumference in the analysis in Table 2. Again, although there were differences in amounts of bycatch, there were no significant or detected differ- ences in catches of squid (Table 2, A and B); therefore it was possible to achieve reductions in bycatch without any notable effect on the commercial catch. During experiment 3 catches were generally small and patchy. Data were very variable, but it was pos- sible to complete five days of trawling with the 41/100 and square-mesh codends with or without their square- mesh BRDs. There was a very marked reduction in bycatch when a BRD was added to a codend, regardless of whether there was a square-mesh panel in the codend or not (Table 3A). Again, there were no influences on the catches of squid, but the ratio of weight of bycatch to weight of squid was smaller for both types of mesh when a BRD was added (analysis of ratios in Table SB). Sizes of species in bycatch The secondary hypotheses of this study, that devices to reduce bycatch would 2A) not cause disproportionate bycatch of small or large fish; 2B) not cause differential effects; and, 2C) not cause relatively larger catches of particular species, were tested with multivariate sta- tistical methods. Analyses of sizes of bycatch were uniform in their outcome — there was no measurable difference on the size composition of species caught for any of the devices tested. As examples, two species offish were present in 538 Fishery Bulletin 104(4) Table 2 Analyses of catches from experiment 2 dur ng July-August 2002. Data are weights of catches from two locat ions (Patonga, Flint and Steel; a random factor), each sampled over four different days (randomly, nested in location). Square-mesh panels were compared with the 41/150, 41/100, 45/150 45/100 codends. There were n = 2 replicate hauls of each gear on each day in each | location. A Analyses of variance Source of variation Bycatch Squid df Mean square F P Mean square F P Treatment = T 4 1.61 0.55 Square-mesh vs. others = S 1 0.55 0.87 >0.35 0.72 1.3 >0.20 Mesh: 41 vs. 45 = M 1 1.14 1.80 >0.15 0.47 0.8 >0.35 Circumference: 100 vs. 150 = C 1 2.40 3.80 <0.05 0.30 0.5 >0.45 MxC 1 2.40 3.70 <0.05 0.57 0.71 >0.20 Location = L 1 2.17 1.13 TxL 4 0.80P 0.19P Days(L) = DiL) 6 4.9 7.10 <0.001 0.69 1.2 >0.25 TxD 24 0.72P 0.62P Residual 40 0.56P 0.50P pPooled residual 68 0.63 0.56 B Mean weight (kg, (±standard error, n = 16)) of bycatch and catch in various treatments Circumference (nc . of meshes) 100 150 100 150 Mesh 41-mmm 56.8(17.5) 28.1(5.4) 14.612.1) 11.6(1.8) 45-mm 44.6(6.7) 60.0(12.9) 13.2(1.5) 13.9(1.7) Square-mesh (30 mm) 48.7(10.5) 15.2(1.9) most hauls, were fairly abundant in bycatch, an(i were each represented by a range of sizes. Tailor iPomato- mus saltatrix) were grouped into five length classes with approximately equal numbers of fish in each class (140-145, 146-170, 171-195, 196-220, 221-280 mm); the snapper Pagrus aiiratus were also so grouped into five length classes (70-105, 106-120, 121-135, 136-155, 156-230 mm). Data were analyzed as the proportion of fish in each size class, to remove differences due to different numbers of fish caught in the various types of gear. Size classes were then used as variables in multi- variate analyses to compare the various types of gear. Using data from experiment 1, we detected no dif- ferences among types of gear in the proportional size frequencies of either tailor or snapper. The mean Bray- Curtis dissimilarities between pairs of samples were all very similar to the mean measures within samples (i.e., among replicates); therefore there was no sugges- tion of differences in size frequencies among types of gear for these species. Multidimensional scaling plots (not included, Clark and Gorley^) supported this result because the centroids of the size classes of both species were not clustered by type of gear. Composition of bycatch The types of species and relative abundances of spe- cies in the bycatch were analyzed from several times of sampling. Analyses used data without squid because numbers of squid are irrelevant to the bycatch. Typi- cally, there were no detected differences in assemblages of species in the bycatch. One exception, in Experiment 2 is described. In our study, analyses of Bray-Curtis dissimilarities showed some differences among types of gear, where the square-mesh codend caught a different assemblage. It seemed that fewer whitebait (Hyperlophus vittatus) were being caught when the square mesh was used (this was confirmed by SIMPERS analysis, which showed that numbers of whitebait were an important variable for discriminating between samples). To test the validity of this conclusion, we performed analyses, omitting numbers of whitebait. This time, all differences 3 Clarke, K. R., and R. N. Gorley. 2001. Primer version 5: user manual/tutorial, 91 p. Primer-E Ltd, Plymouth, U.K. ■• See footnote 3. Scandol et al : Experiments to reduce bycatch in an estuanne squid trawl fishery 539 Table 3 Analyses of catches from four codends 141/100. square-mesh. 75-panel, 75-CSMP Icomposite square-mesh pa nel]) from experi- ment 3 at FHnt and Steel during August-Septembe - 2002. Data are weights of catches on five days. during each of wh ch there were /; = 3 independent hauls with each type of gear A Analysis of variance Bycatch Squid Ratio bycatch/squid Mean Mean Mean df square F P square F P square F P Mesh: 41 vs. square mesh =M 1 0.9 1.4 >0.25 0.001 0 >0.95 0.3 0.8 >0.35 +BRDVS. -BRD =B 1 4.5 6.6 <0.01 0.1 0.2 >0.60 2.4 3.1 >0.15 MxB 1 2.5 3.6 >0.05 0.2 0.4 >0.50 1.0 0.4 Days =I» 4 5.7 8.4 <0.001 0.7 0.7 >0.60 2.5 6.0 <0.001 MxD 4 0.8P 0.5P 0.2P 0.4 >0.50 BxD 4 0.8P 0.6P 0.8 1.8 >0.15 MxBxD 4 0.6P 0.4P 0.2P Residual 40 0.7p 0.6P 0.5P pPooled residual 52 0.7 0.6 0.41 (48 df) B Multiple comparisons: mean weight (kg, (±stand ard error ;! = 30)) of bycatch, t quid, and ratio of bycatch/squid; * indicates a difference iStudent-Newman-Keuls test; P<0.05). Bycatch Squic Ratio- bycatch/squid -BRD + BRD -BRD +BRD -BRD -i-BRD 3.3(0.5) 2.6(0.4)* 1.6(0.2) 1.6(0.2) 6.0(3.8)* 1.6(0.41 among types of gear disappeared. Whitebait are not an important concern for management in this fishery and results like these were not obtained in the other experi- ments; therefore no further consideration was given to this matter. Results from this study indicate that the composition of assemblages of bycatch was not altered by combinations of experimental gear and there were no detectable influences on particular species. These multivariate results indicated that, with the exception of whitebait, none of the secondary hypotheses (2A, 2B, or 2C) could be rejected. These experiments did, however, have limited power to detected small ef- fects because of the large variability of the catches and the relatively limited replication completed. Discussion The results demonstrated that the general concepts used to improve species selection in fish and shrimp trawls have similar application for squid. Nevertheless, the specific designs of modifications require careful consideration, according to several fishery-related factors. One of the starting points for reducing bycatch in all towed gears involves examining simple changes to the size and shape of mesh in the codend, because this is where most of the selection is believed to occur (Wileman et al., 1996). Broadhurst and Kennelly (1996) and Broadhurst et al. (2004) showed that codends made from thick twine and with large posterior circumferences, like those used conventionally by squid trawlers in Broken Bay, have very narrow lateral mesh openings and have poor selectivity. Three of the simplest methods for improving selection in codends involve 1) reducing the posterior circumference (e.g., Broadhurst and Kennelly [1996]), 2) increasing the mesh size (e.g.. Reeves et al. [1992]), or 31 orienting meshes, typically orienting 60-90% of the size of the existing diamond-shaped mesh on the bar so that they are square shaped (e.g., Thorsteinsson [1992]; Broadhurst et al. [2004]). These three modifications were tested during the present study but, with the exception of the square- mesh codend, none were demonstrated to significantly improve species selection or influence the sizes of indi- viduals caught. The considerable temporal variability in catches and a concomitant lack of power to detect small effects may partially account for the lack of dif- ferences detected, but a more likely explanation is that the magnitude of changes to mesh openings in relation to the sizes of target and bycatch species was insuf- ficient. Given that some small fish escaped from the square-mesh codend, future work would probably benefit from further examination of increasing codend mesh openings as a means for reducing bycatch. In contrast to subtle modifications to codend mesh- es that were tested, it was demonstrated that BRDs 540 Fishery Bulletin 104(4) designed from an understanding of the behavior of fish and squid were effective in allowing fish to escape and did not significantly reduce the catches of squid. The escape of fish through these BRDs was a direct consequence of 1) their behavior and species-specific swimming abilities in the trawl (Wardle, 1983; Watson, 1989) and 2) the location of the BRDs with respect to the influences of water-flow (Broadhurst et al., 2002). More specifically, when fish enter the codend they are often herded together and herding invariably initiates movement towards the top and sides. The density of the school and any species-specific behavioral responses (Watson, 1989) determine the extent of such movement. Panels of square-shaped mesh, if strategically posi- tioned, have been demonstrated to allow actively swim- ming fish to escape the trawl through the large open- ings in the square-shaped mesh (Broadhurst, 2000). An important contributing factor is the effect of anteriorly displaced water in front of the catch, which reduces relative flow and assists small fish to maintain their position in the codend. Broadhurst et al. (2002) dem- onstrated that these effects diminish with increasing distance forward and that the optimal location for BRDs designed on fish behavior is up to 1.2 m anterior to the last row of meshes in the posterior section of the codend (i.e., the position of all BRDs examined in our study). The behavior of squid in trawls, codends, or near BRDs has not been examined in detail. Glass et al. (1999), however, observed that the reaction of schooling L. pealeii during initial detection of the gear was simi- lar to fish. Individuals orientated away and then, as a result of compensatory movements in response to shifts in their visual field (termed the optomotor response), attempted to maintain position in the mouth and body of the net. After a short period, all squid were observed to stop swimming, turn, rise in the net, and fall back toward the posterior section of the trawl and toward the codend. Their behavior in this region of the net was not documented, but, given the lack of significant reductions in the catches of squid (or other cephalopods) by prawn trawls containing behaviour-based BRDs (e.g., Broadhurst et al. 2002) and the results observed in our study, active escape is probably limited. There was a nonsignificant reduction in the weight of squid from the codends containing BRDs in experiment 1, and especially the diamond BRD, but this reduction probably occurred during retrieval of the codend. Typi- cally, once the trawl is winched to the surface, a delay of up to three minutes can occur while the vessel is stopped, the retrieval line is connected, and the codend is brought onboard. During this period, any squid in the anterior codend or extension section may wash for- ward and out of the escape opening in the BRD. Watson (1989) observed similar effects during the retrieval of prawn trawls in the Gulf of Mexico and, more recently. Brewer et al. (1998) highlighted this escape route as be- ing a major cause of loss of prawns from several BRDs in trawls used in the Gulf of Carpentaria, Australia. A simple way of minimizing the potential for such losses when using the diamond BRD would be to keep the vessel moving, thus maintaining drag on the codend, during retrieval. A simpler and more practical solu- tion would be to substitute the square-mesh BRD with panels made from 75-mm mesh hung on the bar, since the 75-CSMP and 75-panel codends tested in Experi- ment 3 showed no evidence of loss of squids. Assuming that fish escaping from the CSMP survive the process (Broadhurst et al., 1997), the results presented in this study support the adoption of the CSMP codend in Bro- ken Bay squid trawls. As a consequence of this study, and the co-operative relationship established between the NSW Department of Primary Industries and com- mercial fishermen, a square-mesh panel BRD has been voluntarily adopted throughout the Hawkesbury River squid-trawl fishery. Acknowledgments This study was funded by the Australian Research Council Linkage Scheme (Project C00106975) and the NSW Department of Primary Industries. The work would not have been possible without the technical expertise and support of Broken Bay squid fishermen Mark Peterson, Carl Blacklidge, Graeme Hillyard, Chris Stapleton, and Rolf Norington. We thank Bob Hunt, Katie O'Donnell, Simon Gartenstein, Andres Grigali- unas, Craig Myers, Sophie Diller, and David Blockley for their technical assistance. Steve Kennelly developed the original proposal for this study and suggested valuable improvements to the manuscript. Constructive com- ments from two anonymous reviewers were also greatly appreciated. Literature cited Brewer, D., N. Rawlinson, S. Eayrs, and C. Burridge. 1998. An assessment of bycatch reduction devices in a tropical Australian prawn trawl fishery. Fish. Res. 36:195-215. Broadhurst. M. K. 2000. Modifications to reduce bycatch in prawn trawls: a review and framework for development. Rev. Fish Biol. Fish. 10:27-60. Broadhurst, M. K., and S. J. Kennelly. 1996. Effects of the circumference of codends and a new design of square-mesh panel in reducing unwanted by- catch in the New South Wales oceanic prawn-trawl fish- ery, Australia. Fish. Res. 27:203-214. Broadhurst. M. K., S. J. Kennelly, and D. T. Barker. 1997. Simulated escape of juvenile sand whiting (S(7- lago ciliata, Cuvier) through square-meshes: effects on scale-loss and survival. Fish. Res. 32:51-60. Broadhurst, M. K., S. J. Kennelly, and C. A. Gray. 2002. Optimal positioning and design of behavioural-type bycatch reduction devices (BRDs) involving square-mesh panels in penaeid prawn-trawl codends. Mar. Freshw. Res. 53:813-823. Broadhurst, M. K., R. B. Larsen, S. J. Kennelly, and P. E. McShane. 1999. Use and success of composite square-mesh codends Scandol et al Experiments to reduce bycatch in an estuanne squid trawl fishery 541 in reducing bycatch and in improving size-seiectivity of prawns in Gulf St. Vincent, South Australia. Fish. Bull. 97:434-448. Broadhurst, M. K., R. B. Millar, S. J. Kennt-lly. W. G. Macbeth, D. J. Young, and C. A. Gray. 2004. Selectivity of conventional diamond- and novel square-mesh codends in an Australian estuarine penaeid- trawl fishery. Fish. Res. 67:183-194. Burke, W, T., M. Freeberg, and E. L. Miles. 1994. United Nations resolutions on driftnet fishing: An unsustainable precedent for high seas and coastal fisheries management. Ocean Dev. Int. Law 25:127- 186. FAO (Food and Agriculture Organization of the United Nations). 1995. Code of conduct for responsible fisheries, 41 p. FAO. Rome. Glass, C. W., B. Sarno, O. M. Henry, G. D. Morris, and H. A. Carr. 1999. Bycatch reduction in Massachusetts inshore squid iLoligo pealeii) trawl fisheries. Mar. Technol. Soc. J. 33:35-42. Hall, M. A. 1996. On bycatches. Rev. Fish Biol. Fish. 6:319-352. Kennelly, S. J. 1995. The issue of bycatch in Australia's demersal trawl fisheries. Rev. Fish Biol. Fish. 5:213-234. Kennelly, S. J., and M. K. Broadhurst. 2002. By-catch begone: changes in the philosophy of fishing technology. Fish Fish. 3:340-355. Morais da Cunha, M., and A. Moreno. 1994. Recent trends in the Portuguese squid fishery. Fish. Res. 21:231-241. O'Donnell, K. J. 2004. Growth and Reproduction of the Squid. Photololigo etheridgei, in the Hawkesbury River, NSW. Ph.D. diss., 221 p. Univ. Sydney, Sydney, NSW, Australia. Piatt, J. F., and P. J. Gould. 1994. Postbreeding dispersal and drift-net mortality of endangered Japanese murrelets. Auk 111:953-961. Rathjen, W. F. 1991. Cephalopod capture methods: an overview. Bull. Mar. Sci. 49:494-505. Reeves, S. A., D. W. Armstrong, R. J. Fryer, and K. A. CouU. 1992. The effects of mesh size, cod-end extension length and cod-end diameter on the selectivity of Scottish trawls and seines. ICES J. Mar. Sci. 49:279-288. Simon, F., F. Rocha, and A. Guerra. 1996. The small-scale squid hand-jig fishery off the northwestern Iberian Peninsula: application of a model based on a short survey of fishery statistics. Fish. Res. 25:253-263. Thorsteinsson, G. 1992. The use of square mesh codends in the Icelan- dic shrimp (Pandalus borealis) fishery. Fish. Res. 13:255-266. Underwood, A. J. 1997. Ecological experiments: their logical design and interpretation using analysis of variance, 504 p. Cam- bridge Univ. Press, Cambridge, England. Wardle, C. S. 1983. Fish reactions to towed fishing gears. In Experi- mental biology at sea (A. MacDonald and I. G. Priede, eds.), p. 167-195. Academic Press. New York, NY. Watson, J. W. 1989. Fish behaviour and trawl design: potential for selective trawl development. In Proceedings of the world symposium on fishing gear and fishing vessels, p. 25-29. Marine Institute, St Johns, Canada. Wileman, D. A., R. S. T. Ferro, R. Fonteyne, and R. B. Millar. 1996. Manual of methods of measuring the selectivity of towed fishing gears. ICES Cooperative Res. Rep. ICES 215, 126 p. 542 Abstract— Between 1995 and 2002, we surveyed fish assemblages at seven oil platforms off southern and central California using the manned research submersible Delta. At each platform, there is a large horizontal beam situated at or near the sea floor. In some instances, shells and sedi- ment have buried this beam and in other instances it is partially or com- pletely exposed. We found that fish species responded in various ways to the amount of exposure of the beam. A few species, such as blackeye goby iRhmogobiopts nicholsii). greenstriped rockfish iSebastes elongatus), and pink seaperch iZalemhius rosaceus) tended to avoid the beam. However, many species that typically associate with natural rocky outcrops, such as bocac- cio {S. paucispinis), cowcod (S. levis), copper (S. cauriniis). greenblotched (S. rosenblatti). pinkrose (S. simula- tor) and vermilion (S. miniatus) rock- fishes, were found most often where the beam was exposed. In particu- lar, a group of species (e.g., bocaccio, cowcod, blue (Sebastes niystinus), and vermilion rockfishes) called here the "sheltering habitat" guild, lived pri- marily where the beam was exposed and formed a crevice. This work dem- onstrates that the presence of shelter- ing sites is important in determining the species composition of man-made reefs and, likely, natural reefs. This research also indicates that adding structures that form sheltering sites in and around decommissioned plat- forms will likely lead to higher densi- ties of many species typical of hard and complex structure. The relationships between fish assemblages and the amount of bottom horizontal beam exposed at California oil platforms: fish habitat preferences at man-made platforms and (by inference) at natural reefs Milton S. Love (contact author) Marine Science Institute University of California Santa Barbara, California, 93106 Email address: loveighfesci ucsbedu Anne York PO 31375 Seattle, Washington 98103 Manuscript submitted 6 October 2004 to the Scientific Editor. Manuscript approved for publication 14 December 2005 by the Scientific Editor. Fish. Bull. 104:542-549 (20061. Understanding the habitat preferences of deeper-water (below scuba depth) fishes has proven to be elusive. Off the Pacific Coast, several studies (Stein et al., 1992; Yoklavich et al., 2000; Nasby-Lucas et al., 2002) have dem- onstrated that habitat characteristics play a major role in shaping deeper- water fish assemblages. In those stud- ies it was apparent that, although individuals of many species may be found in a number of habitats, most species showed distinct preferences. These studies clearly showed the role that hard structure plays for many species. How more subtle habitat char- acteristics, such as the presence of sheltering sites, may influence species composition was still unclear. Between 1995 and 2002 we sur- veyed fish assemblages associated with southern California oil and gas platforms. Platforms may serve at least two functions for these fishes. First, the water column around many platforms serves as a nursery ground for a suite of rockfishes {Sebastes spp.) and other fish species, often harboring higher densities of these species than do nearby natural outcrops (Love et al., 2003). Second, platform bottoms, where the platform jacket and well pipes meet the sea floor, may harbor high densities of subadult and adult fishes. Most of these fishes are rock- fishes, but lingcod iOphiodon elon- gatus). painted greenling (Oxylebius pictus), and various members of the surfperch family (Embiotocidae) may also be abundant (Love et al., 1999; Love et al., 2003). The platforms we surveyed were designed to have large, circular (1 m in diameter) horizontal beams that connect vertical and diagonal jacket elements at or near the sea floor. In some instances, these beams were buried, either by shells that had fall- en from shallow parts of the jacket or by a combination of shells and fine sediment. In other instances, beams were partially exposed (full width of beam or partial width of beam was resting on the sea floor) or completely exposed (thus leaving a gap between the beam and sea floor). During our fish surveys, we noted that fishes appeared to be patchily distributed along the platform bottom and that some species seemed to be respond- ing to the presence or absence of the beam and to the amount of space under the beam. Because the beams are all composed of similarly shaped steel and differ only in the amount of surface exposed, we hypothesized that patterns of fish associations with this structure would present insights into the role that sheltering spaces play in determining species assem- blages in both natural and man-made habitats. Love and York Relationships between fish assemblages and bottom horizontal beams of oil platforms 543 Materials and methods Field sampling Between 1996 and 2002, we surveyed fish assemblages around seven oil and gas platforms in southern and central California (Table 1, Fig. 1) using the Delta research submersible, a 4.6-meter, 2-person vessel, operated by Delta Oceanographies of Oxnard, Cali- fornia. Aboard the Delta, we conducted belt tran- sects about two meters from the platform while the submersible maintained a speed of about 0.5 knots. Surveys were conducted in fall, in order to optimize good weather and water clarity. Submersible surveys were conducted during daylight hours between one hour after sunrise and two hours before sunset. During each transect, a researcher made observations from a viewing port on the star- board side of the submersible. An externally mounted hi-8 mm video camera with lights filmed the same viewing field as seen by the observer. The researcher identified, counted, and estimated the lengths of all fishes and verbally recorded those data on the video. All fishes in a volume two meters from sea floor up- wards and two meters from the submersible outwards were counted. Fish lengths were estimated by using a pair of parallel lasers mounted on either side of the external video camera. The projected reference points were 20 centimeters apart and were visible to both the observer and the video camera. We defined the amount of beam exposure on a scale of 0-4: 0 = the beam was completely covered by shells and soft sediment and it was not visible; l=only the top of the beam (usually encrusted with invertebrates) was visible; 2=the beam was partially exposed (top and sides) — the bottom of the beam remaining in contact with the sea floor; 3=the beam was completely exposed and formed an open crevice less than 0.5 m high; and 4=the beam was completely exposed and formed an open crevice more than 0.5 m high (Fig. 2). For each fish we recorded the size of the gap with which the fish associated. An environmental monitoring system aboard the submersible continuously recorded date, time, depth, and altitude above the sea floor of the vessel. These environmental data were overlaid on the original videotape upon completion of each survey. Transect videos were reviewed aboard the research vessel or in the laboratory and observations transcribed into a database. Statistical analysis We were interested in broad patterns of species' dis- tribution among small-scale habitats. Because rare species may prefer some (nonmeasured) extreme habi- tat and thus potentially would have skewed a general picture, we did not use those species where fewer than 40 fish were seen or those that were not seen on at least five dives. This left us with 27 species (of the original a S ^ CO X o, o •= c o uo o =n ^ cfi o 5 OJ M 2 c "3 OJ m C C ■? ■- ^ o 5 a; c ? ^ a p ^ a c OJ CO QJ > -M ^ ■^ 1 o a; CO w A « =^ .'' o t- -s s-s X. II a .. CO M E S -Q d c S-H O) CJ ^ '^ ?, O to g a CO t; >< s _ rtl > f w j: -a i: -" T3 O U ^ 2 Ji X .£2 "-I a> to ij if E 5 S -2 -= J- .- .T3 to ■ — CO tr, ^ -rj ^ to — CO > CO CO U g CO) o e ffl s C CO -O r: is o > m ° S s X X a. <" c ^ ■— o — _ 03 8 o Tj t_i i; o O M QJ CJ i-H 05 lO ir^ C£) O^ O I> !ji Cj) in lO t> '-I r-* 00 Cl ■^ CD CD CJ O O O O Tf 00 O .^ t-^ --H ^ d '-i o 00 CJl 00 Tt c^i r^- ^ — ,_( ro CO uu , — 1 r>i to r- C^l 05 Ol CO CM IN tn fN ,-H fN ^ lO C~ C^J 00 ,-< Tj* ^ t- O lO CO ■* CO O 00 CD r-l C<1 "-I o o CD cd' CD 00 in o i> T— < lO i-H CO CM CO 00 in in CO CM C<1 00 o CO 00 CJ CO CJ5 o 00 ro lO o o o o CM CD 00 o ro ro 00 5 CD CO 00 CD 05 05 c5 oi CO CD 6 CO lO CM lO o CM CM c:i CO CD CM ^ X ■* CO ro CO CO 00 ro CO CO d 00 ^ o CD _ _ CJ5 -^ -* ^.. ro CO CO Tj* CM CO CO ro [> 00 I> CO CO CO ro 00 lO '^ CM 00 lO CM CO tT ^ OJ CO ■!>• 04 ^ C.l CD ^ i-H ro CO » to ^ ,„, O o ro r-' CM r-' CM o o o o ro ro O ro 0 cc o o o o ro ro o ro 0 CM ^ CM ^1 T-^ CM ^1 ra 1 1 1 1 1 1 1 CO to CO CO CO CD ro to ro ro ro ro ro ro ro ro ro ro ro ro ro ro ro ro ro ro ro g o CO a rn o to c 0) "cO O B u 0) CO ■3 CU ffi X SC X O 0 — o 544 Fishery Bulletin 104(4) Longitude ("W) t20"30' 120'20' 120"10' 12000' 119 50' 119 40' 119 30' 119 20' ll19"10' I Irene ^^ i 73m "W ■ Hidalgo -tf : 130m ^^ Figure 1 Location of platforms in the Santa Barbara Channel and Santa Maria Basin. Platforms discussed in this paper are shown as black stars and their bottom depths are indicated. The open circles indicate other platforms. 65) and 50,048 fish of the original 52,999 observations, or 95.1% offish observed on all dives. In our analyses, we separated bocaccio (Sebastes paucispinis) and ling- cod into two categories, YOY and older fishes, based on length-at-age and length-at-first-maturity data (Miller and Geibel, 1973; Cass et al., 1990; Love et al., 2002). We used chi-square goodness-of-fit tests to determine which species tended to avoid or favor certain beam habitats. Species tend to be associated with particular depth ranges, platforms are placed at fixed depths, and some species may or may not have ever been observed on some platforms. Thus, the proportion of a particular beam habitat available to a given species was deter- mined as the proportion of that habitat occurring only on those platforms where the species was observed. We hypothesized that if a particular species does not favor or avoids certain habitats, the expected number of those fish seen in particular habitats would be proportional to the amount of available habitat. In equation form; Let u^^ be the proportion of gap j ( j = 0, 1, 2, 3, or 4) available to species / (; = 1, 2..., 27) and T, be the total number individuals of species ;' observed. Then, the expected number offish / at gapj, E^^ = r/ Y T,. Under the null hypothesis for species (', ■' (O E f J,' j=o Ej,, is distributed as a chi-square random variable with n,-l degrees of freedom, where /;, is the number of Uy>0. The asymptotic assumptions for the chi-square test are not valid if the expected value of many cells is small. Cochran (1954) developed a conservative rule of thumb that the test not be used if more than 20% of the expected cell frequencies are less than five. Koehler and Larntz (1980) suggested that the chi-square test is reasonable if the total number of observations is greater than 10, the number of categories is at least 3, and the square of the number of observations is greater than 10 times the number of categories. If the goodness-of-fit hypothesis was rejected, we ex- amined the individual deviations, X^, = (O^, - £ )/(£■ )-5, which are approximately distributed as normal (0,1) random variables under the null hypothesis. Small val- ues of X^^ indicate that the species is found less often than predicted, whereas large values indicate it is found more often. Results All species satisfied both the Cochran (1954) and Koehler and Larntz (1980) criteria for the validity of the chi- square test (Table 2). The null hypothesis that species are randomly distributed among the crevice habitats was rejected (P<0.0001) for all but one species, kelp greenling (Hexagrammos decagrammus) (Table 2). We surveyed a total of 9804.1 m- of sea floor. Plat- forms varied both in the amount of horizontal beam exposed and, when exposed, the degree of gap between beam and sea floor. At each platform, there was rela- tively little annual variability in the amount of beam exposed or the size and type of gap (Fig. 3). Mean size of gap per platform over the entire study ranged from 2.5 (SD=1.1) to 0.5 m (SD = 0.4). (Fig. 3, Table 1). Platforms Gail and Grace, in the east Santa Barbara Channel, had the greatest amount of gap. In particular, almost none of the bottommost beam at Gail was completely buried and most of it was at least partially exposed. At the Love and York Relationships between fish assemblages and bottom horizontal beams of oil platforms 545 A > viii*^' ^^^^^^■^^^^^■KT 1^^^^ •-..^j*.r-' biL. ^..^LaT'^jZa ^' -^ i^S*iM^Ply^^^M^^^f ^^« *->< Figure 2 Examples of four types of bottom beam structure; (A) at least some of the beam was visible, but the full width of the beam rested on the sea floor (greenspotted rockfish, Sebastes chlorostictus); (B) the beam was partially exposed, remaining in contact with the sea floor at its bottom, not its sides (flag rockfish, S. rubrivincfus); (C) the beam was completely exposed and formed an open crevice less than 0.5 m high (cowcod, S. levis); (D) the beam was completely exposed and formed an open crevice more than 0.5 m high (vermilion rockfish, S. miniatus). Other extreme, most of the bottom beam at both Plat- forms Holly in the central Santa Barbara Channel and Harvest, off Point Conception, was completely buried. We saw little relationship between geographic location (or platform depth) and mean gap size. For instance, Platforms Hidalgo and Harvest are located within 4.6 km of each other, yet have very different beam expo- sures, as do Platforms Holly and Irene that are sited at almost the same depth. A lack of relationship probably reflects differences in oceanographic conditions, because some of these structures are found in areas where strong currents scour the bottom, whereas others are found in areas where sediments have not been disturbed. Except for kelp greenling, all species exhibited some beam habitat preference (Table 2). Species found more often where the beam was completely buried (gap 0) included greenstriped (S. elongatus), non-YOY lingcod, rosy (S. rosaceus), sharpchin (S. zacentrus), and strip- etail (S. saxicola) rockfishes, painted greenling, pink seaperch iZalembius rosaceus), and sanddabs iCithar- ichthys spp.). Those that favored the presence of the beam or some amount of exposure (exposures 1 and 2) included calico (S. dallii), copper (S. caurinus), flag (S. rubrivinctus), and pinkrose (S. simulator) rockfishes, pile perch (Rhacochilus vacca), and sharpnose seap- erch iPhanerodon atripes). Species that tended to in- habit areas where there was a gap between beam and sea floor (exposures 3 and 4) were blue (S. mystinus), brown (S. auriculatus), canary (S. pinniger), green- blotched (S. rosenblatti), halfbanded (S. semicinctus), squarespot (S. hopkinsi) and vermilion (S. miniatus) rockfishes, and both size classes of bocaccio (S. pau- cispinis). The vast majority of cowcod (S. levis) were found at beams that were scored as exposures 2 and 3. Both greenspotted rockfish (S. chlorosticus) and YOY lingcod appeared to prefer either soft bottom without beam exposure or beam exposure without a deep gap (exposures 1 and 2). 546 Fishery Bulletin 104(4) 0) , t: j: m - ^ CO 13 CO " i " z; TO ,.»■ ty O w S " fsl = S j3 CO 1 ^ i -2 - OJ > 5 ^ "2 "o "S =< '3 -o C- 0; CO OJ r^ O 3 > .ii ^ S C 3i ^ c CO >- o £ o ^- CQ § -a ^ <" V _; £ E ex o ■" "^ CO > § t; M o .S C C CO -Si: o -i V P "3 m CD O- HO t.-. " c X S to =+ ' ^ -^ n ^ CO 2 -^ ^ ^ ■"" LO -a "3 E CO ^' o a> I o g V II ^ OJ o 5 CO -a ^ X S OJ ;_ QJ CO "o M .2 ^ ax « fe g I X c X 6 X St.S a. ^ 1- ac CO X CO c c "3 ID ^ M 2 -73 ^ _3 o C^ "S J) J; CO ■" S -a w CO £ o ^ X ^ OJ 'C« QJ ■> .a to > CO ? X > m CD o ^ u 0) Cfi X Ol s 3 C 3 -n O. u cfi CO Q- o o o o o o o o o o o o o o o o o o o o o C3^ (M »0 --< CO CO 00 o o o o o 00 ^ ai O -rf - c^i GO c^ CO crj CD --H o lO CO lO CO O CO 05 (3 O C^ oi --H uo i6 <— ' (>i i>- i-l C^ .-( ^ UO CO CO CO i-H t— I CO COOOCNCNCJlCOCOO-^OOOCOCOlMTf 00 (N 05 1-H ' CO " ' lO c^i in .— I CO ojcoooco-^-^ooincDaiOi-Hc--; lOcoc-^odi-Hoocric^cocDe-ic^'cd COcO'-'O'— 'Oi'-'^-OC^t:^ -^ CN CO CO 00 o (MCOi— lOOOOOOlOiO ■^ f- lO 00 C3i UO lO CD (>i ^ lO CJ5 UO lO Gl LO t^ CTl ca •-< CO o:) ,-t CO T^ CS] lO CO LO CO --< O Tj^ 1-i T^ CO Cjj CO lO 00 1— ( o CO t— uo (M (M C^ ■^ CO 00 o CO ir- -l—i CO Oi CM lO CO rH LO c^ uo Oi LO CO CO 00 03 ^ o C- ■^ o LO 00 o '^ CD o co .— 1 CO CO CO '^ ^ o OSCOC^"^'— 'COi— 1 lO Oi o CO CO LO CO 00 CM 00 Oi 3 o + o + CD CO '^ CTi O^ 00 Ol O O IT- 00 ^ CS CO lO -^<: . ' , 1 > ir J^ - L c- T 1 r 1W6 1948 20(11) 21)02 Hidalgo (I 1 " 1 o Harvest '*- Hermosa m- n - T 1 _ T T JU . " - i- - " f _ 1 ~ - 1996 1998 2000 2002 Year "I - Holly m- (N- - — K. It moves the sampling unit to the cluster that increases the value of PiSidaia) most (if no increase is possible, the sampling unit is not moved). 2 It joins clusters. For each pair of clusters, the algo- rithm calculates the change to Pi S\ data) imposed by joining them. It joins the two clusters that cause the maximal increase (if no increase is possible, no clusters are joined). 3 It splits clusters. Using the Kullback-Leibler diver- gence between sampling units (Corander et al., 2003), the algorithm splits each cluster into maximally 20 subclusters and calculates the change to PiSldata) imposed by keeping one of the introduced subclusters as a separate new cluster, or by joining it to any of the previously existing clusters. It keeps the new configuration that improves the value of P{S\data) most (if no improvement is possible the split is not made). 4 It splits clusters into exactly two maximally homo- geneous subclusters with the Kullback-Leibler diver- gence, otherwise analogously as in step 3. 5 It re-allocates several sampling units from a cluster. In a stochastic order, for each cluster, the algorithm orders the sampling units of the cluster such that the first sampling unit is the one whose removal from the cluster would improve the marginal likelihood of the cluster most, and so on. A putative candidate for a new partition is formed by moving sampling units one by one to some other cluster, such that the P{S\data) of the resulting partition is as high as possible (these moves are performed even if a single move results in a worse solution). If at some point the total change in PiSldata) is positive, the putative candidate is accepted and operation 5 is completed. When the total change remains negative, even after re-allocation of all the sampling units in a single cluster, the putative candidate is rejected and another cluster is chosen as a target until all clusters have been considered. Corandei et al A Bayesian method for identification of stock mixtures from molecular marker data 553 The optimization operations are repeated in a varying order, until none of them improves the posterior prob- ability Pi S\ data) of the current partition. The allocation of sampling units to different clusters is based on the obtained partition, and the suitable number of clusters k is estimated from the partitions visited during the simulation. Measurement of the strength of evidence for any particular value of the partition S. given the marker data, is an intricate process, in particular for large data sets from complex stock mixtures. Theoretically, the unknown largest posterior probability may be ex- tremely small, even in situations where a particular model provides an adequate fit to the observations. An important factor explaining such a feature is the large number of possible allocations, which all have positive posterior probabilities by definition. This feature is of general concern in a Bayesian analysis that comprises vast model spaces, see, e.g., the discussion in Madigan and Raftery (1994). Because the actual estimated value of the posterior probability may be an intuitively mis- leading goodness-of-fit measure, we use an alternative strategy for characterization of the uncertainty in rela- tion to the estimated allocation. Bayes factors (e.g., Kass and Raftery, 1995) provide a computationally efficient approach to local assessment of the amount of the peak of the posterior distribution around an estimate of S. Let S* denote an alternative allocation obtained from an estimate S by moving any particular sampling unit to another putative stock. The strength of evidence in favor of placing that sampling unit in the original stock against placement in the new stock is provided by the Bayes factor Be p(data\S)Pi.S) p{data\S )P{S ) (4) fives). When only a single stock has a high conditional posterior probability, the allocation is made on a firm basis. However, when at least two sources are identified with reasonably high posterior probabilities, the genetic evidence is not conclusive enough for a classification of the particular individual to a single source. The advan- tage of the conditional posterior probabilities over Bayes factors in characterization of the classification uncer- tainty for each individual is that the former compares simultaneously all putative sources, whereas the latter provides only a pairwise judgement. The correct number of clusters needed to describe the data can be estimated from the partitions that were visited during the simulation. During the simulation the algorithm stores the marginal likelihoods and the sizes of the 30 best visited partitions, and the posterior probabilities for the different numbers of clusters can then be estimated analogously to those estimated by Corander et al. (2004). Usually, if there is a lot of mo- lecular data available (e.g., hundreds of loci have been observed) only a few of the best partitions have influ- ence on the computed posterior probabilities because the relation of marginal likelihoods between different partitions can be up to -expdOOO). If the data are sparse (e.g., only about 10-20 loci have been observed) and only partial baseline information is available, the uncertainty related to the correct number of clusters can be considerable because many partitions with dif- fering sizes and approximately equal marginal likeli- hood may be found. In these cases, to obtain a more reliable estimate of the correct number of clusters, the algorithm should be run multiple times with different upper bounds (K) in order to facilitate the identifica- tion of those partitions that have real influence in the posterior probabilities. In our implementation of the es- timation algorithm, we have included the possibility to automatically process information from multiple runs. which measures how many times more plausible the allocation S is for the particular sampling unit. When the value of Equation 4 is small, say B^ ,,. < 10 (or log,. Bggt < 2.3, Kass and Raftery, 1995), the data do not strongly support a single origin for the particular sam- pling unit. Because calculation of these Bayes factors is computationally inexpensive, they can be easily provided for every possible sampling unit or stock combination. In addition to Bayes factors, conditional posterior probabilities for the allocation of each individual over the range of different putative stocks identified through S can be used to characterize the uncertainties in the Bayesian estimate. The conditional posterior probability distribution is defined for each individual by PiS\data)-- p(rfatalS,)P(S,) "a ^p(rfatalS,)P(S,; (5) where S, denotes that the particular individual is allo- cated to the (th class of S (over the k possible alterna- Empjrical Illustration of the partition-based approach The Bayesian estimation algorithm described in the previous section is implemented in BAPS software.^ The examples considered here are produced by BAPS analyses of data simulated by using the real data from Koljonen et al. (2002), who assessed allele frequencies for nine microsatellite markers in Atlantic salmon within the Baltic Sea region. We have experimented with sev- eral simulation configurations to investigate how our method would be expected to perform under a variety of biological conditions. The five wild stocks of Atlantic salmon considered in Koljonen et al. (2002) correspond to five different rivers draining into the Baltic Sea: Tornionjoki (TornW), Simo- joki (Simo). lijoki (li), Oulujoki (Oulu), and Neva. The pairwise genetic distances (Nei's D.,, Nei et al., 1983) between these stocks underlying our simulations are ' BAPS software is freely available at URL http://www.rni. helsinki.fi/--jic/bapspage.html. Results presented here were calculated with version 3.1 (release date 5 March 2005). 554 Fishery Bulletin 104(4) given in Table 1 (reproduced from Koljonen et al., 2002). An estimate of total F^j, (Weir and Cockerham, 1984) equal to 0.07 was obtained in Koljonen et al. (2002) for these stocks on the basis of the nine microsatellite loci. The magnitude of the genetic differentiation in the underlying population is fairly small, and the pairwise distances vary considerably. Thus, we may conclude that these stocks represent a biologically challenging setting for inference about the genetic mixture in a population sample. Using the individual stock allele frequencies, we have simulated baseline individuals and catch sam- ples under the assumptions of HWE and no linkage between the loci. A wide variety of configurations, with complete and partial baseline information and different sample sizes, were tested. In the analyses involving five underlying stocks, we used Ti' = 10 as the prior upper bound, and the estimation algorithm was run 12 times for each replicate data set. For cases with only two un- derlying stocks, the upper bound was set as K = 6. Results of our simulation experiments are summa- rized in Tables 2-8. As a summary, we highlight the following aspects. Uneven proportions of stock presence in the samples do not seem to affect the inference nota- bly, even when the baseline information is only partial. The results in Table 3 are produced under a particu- larly challenging situation, where the baseline informa- tion comprises 40 individuals from a single stock only. The sample configuration then contains 40 individuals from this stock and 10 individuals from another, a pri- ori unknown stock. The results show that our method performs surprisingly well in the identification of the outgroup, given that the genetic difference between Table 1 Pairwise genetic distances (Nei's D^, Nei et al., 1983) between different Atlantic salmon stocks within the Baltic Sea region (reproduced from Table 2 in Koljonen et al., 2002). Stocks correspond to five different rivers: Tornionjoki (TornW), Simojoki (Simo). lijoki (li), Oulujoki (Oulu), and Neva. Stocks TornW Simo li Oulu Simo 0.129 li 0.068 0.125 Oulu 0.110 0.164 0.131 Neva 0.261 0.285 0.284 0.261 the two underlying stocks is not negligible. However, as the results in Table 4 illustrate, the presence of pu- tative stocks not represented by baseline information may also be masked by the baseline available for a ge- netically similar stock. Identification of putative stocks without using any baseline information may de facto be more successful under such circumstances (compare Tables 4a-4d). Therefore, we suggest that in practice both types of analyses are performed and the results compared, since this is computationally inexpensive with our method. Our results indicate that incomplete baseline information is expected to be most fruitful for the identification task when there are baseline samples available from the stocks that are genetically most similar. The baseline configurations in Table 4 can be Table 2 Allocation average percentages over 20 simu lations, when 25 individuals from each stock were present in the sample data, and the number of baseline individuals available from each stock was (A) 30 ,(B)15 and(C)5. The column with the heading "Other" refers to additional stocks inferred by the method. Stocks correspond to five different rivers Tornionjoki (TornW) Simojoki (Simo), lijoki (li) Oulujoki (Oulu), and Neva Origin Allocation TornW Simo li Oulu Neva Other A TornW 0.80 0.01 0.15 0.03 0.00 0.00 Simo 0.03 0.95 0.02 0.00 0.00 0.00 li 0.11 0.02 0.84 0.03 0.00 0.00 Oulu 0.03 0.01 0.01 0.95 0.00 0.00 Neva 0.00 0.00 0.00 0.00 1.00 0.00 B TornW 0.72 0.03 0.18 0.06 0.00 0.00 Simo 0.02 0.94 0.02 0.02 0.00 0.00 li 0.11 0.03 0.84 0.02 0.00 0.00 Oulu 0.04 0.01 0.03 0.91 0.00 0.00 Neva 0.00 0.00 0,00 0.00 0.99 0.00 C TornW 0.59 0.06 0.23 0.09 0.01 0.02 Simo 0.04 0.90 0.04 0.01 0.00 0.00 li 0.20 0.06 0.70 0.03 0.00 0.01 Oulu 0.03 0.01 0.05 0.91 0.00 0.00 Neva 0.00 0.00 0.00 0.00 0.98 0.02 Corander et al A Bayesian method for identification of stock mixtures from molecular marker data 555 Table 3 Origin identification performance when the sample consists of 40 individuals from the lijoki River lli) and 10 individuals from another stock. Stocks correspond to five different rivers: Tornionjoki (TornW), Simojoki (Simo), lijoki (li), Oulujoki (Oulu), and Neva. Baseline information was in each case available only from li I40 simulated individuals!. The numbers are based on 50 replicates of each configuration. Stock without baseline (outgroup) Percentage of correct recognition for individuals from li Percentage of correct recognition for individuals from the outgroup Percentage of replicates min. max. avg. min. max. avg. the outgroup was recognized Neva 0.95 1.00 0.9955 0.9 1.00 0.998 1.00 Oulu 0.875 1.00 0.9665 0.0 1.00 0.7440 0.86 Simo 0.85 1.00 0.9705 0.0 1.00 0.6820 0.90 TornW 0.90 1.00 0.9820 0.0 0.80 0.3660 0.34 categorized in this respect as neutral (no biasing effect; Table 4, A and B), positive (strengthens the inference; Table 4C), negative (biases the inference; Table 4, D and E). Our results indicate that commonly occurring levels (<5'*) of missing marker data do not inhibit the ability of our method to detect the correct stocks, assuming that the missing values are randomly distributed over loci and individuals (Table 5). As an overall conclusion from the simulations, it is clear that the genetic dis- similarities of the stocks matter most for identification performance. When baseline samples are available for all stocks, most individuals can be correctly assigned to their origin even when the genetic distance between the stocks is negligible (such as between Tornionjoki and lijoki rivers). Usefulness of the conditional poste- rior probabilities for characterization of the allocation uncertainty is exemplified in Table 6. The number of inferred putative stocks was in general well in accordance with the underlying true number and there was no tendency to overestimate k. However, when the number of available marker loci was decreased to five (Table 7), the probability of obtaining additional pu- tative stocks was slightly increased. Because it is widely known that the level of polymorphism of the markers affects their usefulness in origin identification, it is difficult to specify very clear boundaries with respect to the amount of loci necessary for an acceptable per- formance of any assignment method. It is important to notice that an acceptable characterization of uncertainty inherently depends on the real biological context in a particular modeling situation. As a simple rule of thumb for our method, we would suggest that N i<,Q might be regarded as an insufficient value for reliable estimation. However, when auxiliary information is available such that the sample data can be grouped before analysis (as in Table 8), the statistical power to detect correct ori- gins and k increases considerably. This situation would correspond to a geographical sampling scheme where the individuals assigned to the same sampling unit are caught simultaneously at a specific location. Discussion We have introduced a novel Bayesian method for an investigation of stock mixtures using molecular marker data by suitably modifying existing partition-based Bayesian models for estimation of genetic population structure. To enable smooth applicability, the imple- mentation is made freely available in a user-friendly software. One particular advantage of our method is the possibility of appropriately analyzing data in a situation where only partial baseline information is available for the potential stocks. Use of an analytical integration approach enhances considerably the numerical perfor- mance when the stock mixture structure is challenging (e.g., in the presence of small stocks for which no base- line samples have been collected). Contrary to the earlier Bayesian methods introduced in Corander et al. (2003, 2004), we have exploited a con- siderably less computationally intensive strategy that is based on stochastic optimization instead of MCMC simulation. To obtain stable estimates for moderate to large data sets, many long parallel MCMC chains would be needed, but the process for obtaining these chains often is not feasible under a single CPU architecture. Our intelligent search strategy, instead of the random search used in MCMC, seems to resolve this problem very efficiently. A disadvantage of stochastic optimiza- tion compared to optimization with MCMC is that a statistically consistent estimate of the number of stocks contributing to the sample cannot be derived. Neverthe- less, our novel method has performed satisfactorily in this respect under realistic sampling scenarios. We are currently exploring possibilities for using intelligent proposals in MCMC and an online-based parallel imple- mentation of the method, both of which would provide an ideal framework for biologists using molecular data in stock mixture estimation. The most relevant biological assumptions used in our approach are HWE and nonlinkage of the marker loci. The latter assumption is generally valid, at least ap- proximately, for the microsatellite markers often used in 556 Fishery Bulletin 104(4) Table 4 Allocation of 30 sampl e individuals from each of five stocks (150 Indiv duals in total) under different baseline | settings: (A) no baseline . (B) 40 baseline ind ividua Is from TornW, iC) 40 baseline individuals from TornW and Ii, (D) 15 baseline i ndividuals from TornW a nd li. (E) 15 baseline individu als from Simo. Each "C" refers to an inferred putative stock for which no baseline information | was available Stocks correspond to five different rivers: Tornionjoki (TornW), Si -nojoki (Simo), lijoki (li),0 ulujoki lOulu), and Neva Origin Allocation CI C2 C3 C4 C5 A TornW 3 1 0 3 23 Simo 1 1 0 25 3 li 0 1 0 4 25 Oulu 0 28 0 0 2 Neva 0 1 29 0 0 TornW C2 C3 C4 B TornW 28 1 1 0 Simo 2 27 1 0 li 26 4 0 0 Oulu 2 0 28 0 Neva 0 0 0 30 TornW li C3 C4 C5 C TornW 22 6 1 1 0 Simo 1 2 26 1 0 li 5 25 0 0 0 Oulu 3 1 0 26 0 Neva 0 0 0 0 30 TornW li C3 C4 D TornW 24 5 0 1 Simo 1 2 0 27 li 5 23 0 2 Oulu 29 1 0 0 Neva 0 1 29 0 Simo C2 C3 C4 C5 E TornW 12 2 1 15 0 Simo 29 1 0 0 0 li 24 0 0 5 1 Oulu 1 0 28 1 0 Neva 0 0 0 0 30 Table 5 (A) Average numbers of allocations (over 20 replicates) to the different stocks under an uneven sample size distri- bution: TornW, n = 60. Simo, )!=20, li, n = 30, Oulu, n=5. Neva, /! = 10. Stocks correspond to five different rivers: Tornionjoki (TornWi. Simojoki iSimo), lijoki (lii, Oulujoki I Oulu I, and Neva. The number of baseline individuals available from each of the five stocks was 30. The column with the heading "Other" refers to additional stocks inferred by the method. The results in (B) are otherwise based on an analogous configuration, except that 5% of the marker data was randomly set as missing values. Allocation Origin TornW Simo li Oulu Neva Other TornW Simo Ii Oulu Neva TornW Simo Ii Oulu Neva 47.5 0.6 3.1 0.3 0.0 47.4 0.6 3.5 0.3 0.0 1.2 18.9 1.2 0.0 0.1 1.7 18.4 1.1 0.1 0.1 7.9 0.3 25.3 0.1 0.1 7.2 0.8 25.0 0.1 0.0 3.2 0.2 0.5 4.7 0.0 3.6 0.3 0.5 4.6 0.1 0.0 0.1 0.0 0.0 9.9 0.2 0.0 0.0 0.0 9.9 0.3 0.0 0.0 0.0 0.0 0.2 0.1 0.1 0.0 0.1 Table 6 Average conditional posterior probabilities (over 20 repli- | cates) for allocations of inc ividuals to the different stocks under an uneven sample size distribution: TornW, n = 60. Simo, /?=20, li, ; = 30, Oulu, n = 5 , Neva, «=10. Stocks correspond to five different rivers Tornionjoki (TornW), 1 Simojoki (Simo), [ijoki (Ii ), Oulujoki (Ou u), and Neva. The number of baseline individuals available from each | stock was 30. The column with the heading Other refers to additional stocks inferred by the method . Allocation Origin TornW Simo li Oulu Neva Other TornW 0.80 0.03 0.13 0.04 0.00 0.00 Simo 0.02 0.94 0.03 0.01 0.00 0.00 li 0.12 0.04 0.82 0.03 0.00 0.00 Oulu 0.06 0.01 0.03 0.90 0.00 0.00 Neva 0.00 0.00 0.00 0.01 0.99 0.00 stock mixture analyses. Minor deviations from HWE are not expecteci to notably affect our inference method; how- ever, presence of samples from small stocks under strong inbreeding could result in an overestimation of k when there is limited baseline information available. Samples from such stocks would tend to be split into parts by the model if no baseline information about the stock allele frequencies can be used to identify the joint origin. In addition to the molecular markers, auxiliary infor- mation, such as simultaneous catch at a common geo- Corander et al A Bayesian method for Identification of stock mixtures from molecular marker data 557 Table 7 Average numbers of allocations (over 20 replicates) to the different stocks under an uneven sample size distribution: TornW, ;! = 60, Simo, n=2Q. li, ;! = 30, Oulu. ;! = 5, Neva, n = 10. Stocks correspond to five different rivers: Tornion- joki (TornWl, Simojoki iSimol, lijoki (li). Oulujoki (Oului, and Neva. The number of baseline individuals available from each of the five stocks was 30. The column with the heading "Other" refers to additional stocks inferred by the method. The marker loci used for inference were ran- domly sampled from the original nine microsatellites for each replicate; in (A) seven loci were used, in (B) five loci were used. Allocation Origin TornW Simo li Oulu Neva Other A TornW 44.5 Simo li Oulu Neva TornW Simo li Oulu Neva 0.9 2.8 0.3 0.0 37.3 1.0 3.3 0.5 0.1 3.2 17.8 1.6 0.1 0.0 4.5 16.5 2.8 0.2 0.0 7.6 0.9 24.7 0.2 0.0 9.3 1.4 21.2 0.4 0.1 4.3 0.4 0.9 4.5 0.0 5.5 0.6 2.2 4.0 0.2 0.2 0.0 0.1 0.0 10.0 0.5 0.2 0.1 0.0 9.6 0.0 0.1 0.0 0.0 0.0 3.1 0.4 0.5 0.1 0.2 graphical location, can be incorporated into the analysis. This information is incorporated by the pre-assignment of individuals in the catch data to a priori sampling units, when such are considered to be relevant for the species under investigation. Such prior information is particu- larly useful if the available molecular data are scarce because it enhances the statistical power to detect correct stock origins, as illustrated in our example analyses. Although the Bayesian method that we propose seems to be a versatile tool for stock mixture identification, certain modifications of the model would also provide fruitful extensions for a variety of biological settings. Current use of the auxiliary information necessitates that the individuals assigned to the same sampling unit represent with certainty the same origin. However, the existence of such conclusive information cannot be assumed in applications in general. There is still a pos- sibility of incorporating information about a tendency to a geographical clustering among the catch individu- als with respect to the stock origin, through a suitable modification of the prior distribution of the partitions. In general, use of biological information concerning the behaviour of a species, in combination with geographical sampling information, provides a rich area for further model development. In particular, this combination of information highlights the potential use of the Bayesian statistical framework because the relevant biological in- formation can often be efficiently incorporated through the prior distributions for the model parameters. Table 8 Allocation of the 25 sam pie individuals from each of five .stocks (125 individuals in total) under different baseline | settings: (A) no baseline (B) 15 base line individua s from each stock. (C) 15 basel ine individuals from each stock. and auxiliary biological information was introduced by considering the simulated catch dat a as sampling units of the size of five indivi duals. Stocks correspond to five different rivers: Tornionjoki (TornW), Simojoki (Simo), lijoki (li), C ulujoki (Oulu), and Nev a. Each "C" refers to | an inferred putative stock for which no baseline in forma- tion was available. Origin Allocation CI C2 C3 C4 A TornW 20 0 4 1 Simo 2 0 2 21 li 20 1 3 1 Oulu 1 0 24 0 Neva 0 24 1 0 TornW Simo li Oulu Neva B TornW 18 0 7 0 0 Simo 2 23 0 0 0 li 0 1 23 1 0 Oulu 2 0 0 23 0 Neva 0 0 0 1 24 TornW Simo li Oulu Neva C TornW 25 0 0 0 0 Simo 0 25 0 0 0 Ii 0 0 25 0 0 Oulu 0 0 0 25 0 Neva 0 0 0 0 25 Acknowledgments The authors thank Marja-Liisa Koljonen for providing data about the microsatellite allele frequencies in Baltic salmon stocks. This work was supported by the Centre of Population Genetic Analyses, University of Oulu, Finland (Academy of Finland, grant no. 53297), and by Research funds of University of Helsinki, Finland. Literature cited Balding, D. J., and R. A. Nichols. 1997. Significant genetic correlations among Caucasians at forensic DNA loci. Heredity 78:583-589. Beaumont, M. A., and B. Rannala. 2004. The Bayesian revolution in genetics. Nat. Rev. Genet. 5:251-261. Corander, J., P. Waldmann, P. Marttinen, and M. J. Sillanpaa. 2003. BAPS 2: enhanced possibilities for the analy- sis of genetic population structure. Bioinformatics 20:2363-2369. 558 Fishery Bulletin 104(4) Corander, J., P. Waldmann, and M. J. Sillanpaa. 2004. Bayesian analysis of genetic differentiation between populations. Genetics 163:367-374. Dawson, K. J., and K. Belkhir. 2001. A Bayesian approach to the identification of panmic- tic populations and the assignment of individuals. Genet. Res. 78:59-77. Fletcher, R. 1987. Practical methods of optimization, 4.50 p. Wiley, New York, NY. Gelman, A., J. B. Carlin. H. S. Stern, and D. B. Rubin. 2004. Bayesian data analysis, 2"'' ed., 668 p. Chapman & Hall/CRC, Boca Raton, LA Heuertz, M., J. F. Hausman, O. J. Hardy , G. G. Vendramin, N. Frascaria-Lacoste, and X. Vekemans. 2004. Nuclear microsatellites reveal contrasting patterns of genetic structure between western and southeastern European populations of the common ash iFraxinus excelsior L.). Evolution 58:976-988. Kalinowski, S. T. 2004. Genetic polymorphism and mixed-stock fisheries analysis. Can. J. Fish. Aquat. Sci. 61:1075-1082. Kass, R., and A. E. Raftery. 1995. Bayes factors. J. Am. Stat. Assoc. 90:773-795. Koljonen, M.-L., J. Tahtinen, M. Saisa, and J. Koskiniemi. 2002. Maintenance of genetic diversity of Atlantic salmon by captive breeding programmes and the geographic distribution of microsatellite variation. Aquaculture 212:69-92. Madigan, D., and A. E. Raftery. 1994. Model selection and accounting for model uncer- tainty in graphical models using Occam's window. J. Am. Stat. Assoc. 89:1535-1546. Maki-Petays, H., A. Zakharov, L. Viljakainen, J. Corander, and P. Pamilo. 2005. Genetic changes associated to declining populations of Formica ants in fragmented forest landscape. Mol. Ecol. 14:733-742. Mantyniemi, S., and A. Romakkaniemi. 2002. Bayesian mark-recapture estimation with an appli- cation to a salmonid smolt population. Can. J. Fish. Aquat. Sci. 59:1748-1758. Mantyniemi, S., A. Romakkaniemi, and E. Arjas. 2005. Bayesian removal estimation of a population size under unequal catchability. Can. J. Fish. Aquat. Sci. 62:291-300. Meyer, R., and R. Millar. 1999. BUGS in Bayesian stock assessment. Can. J. Fish. Aquat. Sci. 56:1078-1086. Nei, M., F. Tajima, and Y. Tateno. 1983. Accuracy of estimated phylogenetic trees from molecular data. J. Mol. Evol. 19:153-170. Bella, J., and M. Masuda. 2001. Bayesian methods for analysis of stock mixtures from genetic characters. Fish. Bull. 99:151-167. Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of population structure using multilocus genotype data. Genetics 155:945-959. Rannala, B., and J. A. Hartigan. 1996. Estimating gene flow in island populations. Genet. Res. 67:147-158. Rannala, B., and J. L. Mountain. 1997. Detecting immigration by using multilocus geno- types. Proc. Natl. Acad. Sci. USA 94: 9197-9201. Reynolds. J. H., and W. D. Templin. 2004. Detecting specific populations in mixtures. En- viron. Biol. Fish. 69:233-243. Seppa, P., N. Gyllenstrand, J. Corander, and P. Pamilo. 2004. Coexistence of the social types: Genetic popula- tion structure in the ant, fnnnica exsecta. Evolution 58:2462-2471. Smouse, P. E., R. S. Waples, and J. A. Tworek. 1990. A genetic mixture analysis for use with incom- plete source population-data. Can. J. Fish. Aquat. Sci. 47:620-634. Spiegelhalter, D. L., A. Thomas, N. Best, W. Gilks, and D. Lunn. 2003. BUGS: Bayesian inference using Gibbs samp- ling. MRC Biostatistics Unit, Cambridge, England. [Available at www.mrc-bsu.cam.ac.uk/bugs/.] Weir, B. C, and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of popula- tion structure. Evolution 38:1350-1370. 559 Abstract — Bluefish (PomalnmuK so/- tutrix) were tagged and released in Atlantic coastal areas between Mas- sachusetts and Florida from 1963 through 2003 as part of a National Marine Fisheries Service (NMFS) project and a volunteer program sponsored by the American Littoral Society ( ALS I. A total of 15,699 blue- fish were tagged by NMFS and 20,398 by ALS volunteers and A.39c (1075 NMFS tags and 464 ALS tags) were recaptured and reported. Time-at- large was limited; 65.8''?^ of the recap- tured tags were returned within two months of tagging, although nineteen of the returned tags remained at large for two years or more. Tag returns indicated seasonal migrations offish between the Middle Atlantic Bight and Florida. Three groups of bluefish are proposed for Atlantic coastal waters on the basis of tag return data and are defined by the seasonal occurrence of fish between 30 and 45 cm fork length. The northern group occupied the area from Massachusetts to Dela- ware between late spring and late fall. Bluefish in the central region between Maryland and North Carolina rep- resented a combination of seasonal transient and resident fish, as did the southern group in Florida. Mixing occurs among all three groups; and larger fish (>45 cm) spend winters in offshore areas. Estimates of von Bertalanffy growth parameters from tagging data were comparable to scale-based estimates. Swimming speeds between point of release and recapture averaged 2.6 km per day, and seasonal spikes greater than 5 km per day corresponded with periods of migration in spring and autumn. The migration patterns of bluefish iPomatomus saltatrix) along the Atlantic coast determined from tag recoveries Gary R. Shepherd (contact author) National Marine Fisheries Service 166 Water St Woods Hole, Massachusetts 02543 Email address Gary.Shepherdig'noaa.gov Joshua Moser National Marine Fisheries Service 166 Water St Woods Hole, Massachusetts 02543 David Deuel (deceased) National Marine Fisheries Service 1315 East-West Highway Silver Spring, Maryland 20910 Pam Carlsen BIdg. 18 Sandy Hook American Littoral Society Highlands, New Jersey 07732 Manuscript submitted 9 May 2005 to the Scientific Editor's Office. Manuscript approved for publication 19 December 2005 by the Scientific Editor Fish. Bull. 104:559-570 12006). Bluefish iPomatomus saltatrix) is a pelagic species with a worldwide distribution in temperate and sub- tropical oceans. In the United States, bluefish are found along the Atlantic coast from southern Florida to Cape Cod, Massachusetts, and occasion- ally as far north as Nova Scotia (Col- lette and Klein-MacPhee, 2002). The broad-scale seasonal movements of bluefish are known within the com- mercial and recreational fishing com- munities (Hersey, 1987), but details of the migratory pattern remain poorly documented in the scientific litera- ture. Tagging studies provide the most direct evidence of seasonal move- ments, but the only published account for the Atlantic coast stock is a study in Long Island Sound by Lund and Maltezos (1970). Wilk (1977) provided a description of bluefish migration that remains the accepted standard and which was based on seasonal distribution of commercial and recre- ational catches, as well as on unpub- lished results of a tagging project conducted during the 1960s by David Deuel and colleagues at the NMFS James. J. Howard Marine Sciences Laboratory (formerly known as the Sandy Hook Marine Laboratory). The proposed migration involved a north- south coastal movement between New York-New Jersey offshore waters and southeastern Florida offshore waters during the fall and a return spring migration along the same route. Larger fish (i.e. greater than three pounds) were believed to follow a more offshore pathway. The identification of distinct blue- fish stocks contributing to this mi- gratory group has been the subject of multiple investigations. The racial composition of bluefish on the Atlantic coast was investigated by Lund (1961) who concluded, primarily from dif- ferences in the number of gill rakers of small bluefish, that six races ex- isted along the coast. Lassiter (1962) found differences in first year growth on scales, which indicated that two groups of fish inhabited North Caro- lina waters. Returned tags from blue- fish tagged in the Long Island area 560 Fishery Bulletin 104(4) (Lund and Maltezos, 1970) also indicated two distinct stocks, although not necessarily the same groups as de- fined by Lassiter (1962). More recently, some scientists have concluded that either two distinct spawning groups exist (Norcross et al., 1974; Kendall and Walford, 1979) or one stock with two distinct survival periods (Hare and Cowen, 1993; Smith et al., 1994). Others, probing mitochondrial DNA (mtDNA) (Graves et al., 1993), have concluded that Atlantic coast bluefish constitute a single population. A discriminant function analysis of morpho- metric data has corroborated the one stock hypothesis despite evidence of phenotypically plastic characteristics (Austin et al., 1999). Mark-recapture experiments provide an empirical method for evaluating both migratory behavior and stock composition. In 1962, the National Marine Fisher- ies Service (NMFS) initiated a study of the migratory patterns of bluefish to obtain information on the popula- tion structure of the Atlantic coast stock; this coastwide tagging program continued until 1967. The American Littoral Society (ALS) also coordinates an annual tag- ging program by private citizens that has resulted in 20 years of tag releases for a variety of species, includ- ing bluefish. This combined tag-recapture information constitutes the largest known tagging database for this species. The goal of our study was to investigate the mi- gratory behavior of bluefish along the Atlantic coast by using the results of these previously unpublished tag- ging studies and to examine the single stock hypothesis in context of tag recovery information. Materials and methods The NMFS bluefish tagging program used several types of tags (variously colored). Field tests in 1963 indicated that a dorsal loop spaghetti tag would be more suitable, owing to a longer retention rate, than dart tags or jaw tags. This loop tag, described by Watson (1963), was closed with a pressure-fitted V-shaped clip. In addi- tion, two other types of dorsal loop tags (both closed by knotting the ends of the tubing, and one having a clear outer covering) and an internal anchor tag (5/16 by 1-1/4 inches, with a 3" streamer as described by Topp [1963]) were used. For visibility of the tags to fishermen, orange was chosen as the color for most of the dorsal loop tags and yellow for the streamer of the internal anchor tags. The fork length of each fish tagged and released was measured to the nearest cm. Recapture information was mailed to NMFS in Sandy Hook, NJ, and included data on recapture location, recapture date, fish length (usu- ally in inches), and weight (lbs and oz, when possible). The ALS program uses yellow dorsal loop tags that are inserted with a hollow needle through the muscle below the dorsal fins and tied with a simple overhand knot or, as in some recent tags, fastened with a snap- lock mechanism (Carlsen, 2000). Details on the date, location of release, fork length (to the nearest inch), and weight (lbs and oz, when possible) were furnished by the fishermen at the time of release. Information regarding tag recapture (including location, recapture date, fish length [inch] and weight) was mailed to ALS headquarters by the fishermen and the distance of the recapture location from shore was coded as inland, inshore (^3 miles), or offshore (>3 miles). Not all tags recaptured in the ALS program were reported with complete details. As a result, some of the sample sizes presented in the present study differ among analyses, depending on the available data of the recaptured tags. Also, in order to standardize data from both programs and to minimize precision error resulting from unit con- versions, fish length data are presented in centimeters followed by the inch equivalent rounded to the nearest whole number. Several types of fishing gear were used to capture bluefish in the NMFS tagging program. Fish were cap- tured by using gill nets deployed from research vessels, hook and line, commercially operated pound nets, and beach seines; bluefish tagged in the ALS program were captured with hook and line. Bluefish tagged in the NMFS program were released in areas of seasonal abundance from southeastern Flor- ida to Massachusetts, and major tagging efforts were made in southern Florida, North Carolina, Virginia, New Jersey, and New York (Fig.l). ALS tag releases were concentrated in the New York-New Jersey area, although tagging occurred from Florida to New Eng- land (Fig.l). Each ALS recapture location was catego- rized into 1 of 59 geographic areas and was assigned a latitude and longitude based on the center point of the recapture area. NMFS recaptures were assigned latitude and longitude coordinates based directly on recapture location. Distance (km) traveled was calculated as the great circle distance between the point of release and point of recapture (NMML') and there was a suitable way- point if the pathway traversed land. Swimming speed was calculated as the linear distance traveled divided by days at large. In order to allow for acclimation to the presence of the tag, fish recovered within the first three days of release were not included in analyses. Ad- ditionally, because speed calculations were influenced by the number of days at large, speed estimates per two-month increments were restricted to tags recovered within 30-day periods to minimize recoveries span- ning several months and to avoid averaging speed over various migratory phases. Swimming speeds were com- pared among months by using an analysis of variance (ANOVA) (Zar, 1974). Comparison of movements by size was made for 5-cm length classes to reduce potential bias from measurement error of recaptured fish. Bluefish growth was modeled by using the change in fork length during the time-at-large. Von Bertalanffy growth curve parameters L ^ and K were calculated by using the Fabens model (Fabens, 1965): ' NMML (National Marine Mammal Laboratory). 2004. Ex- cel geometry functions. Website: http://nmml.afsc.noaa. gov/Software/ExcelGeoFunctions/excelgeofunc.htm [accessed 4 October 2004]. Shepherd et a\ Migration routes of Pomatomus saltan ix along the Atlantic coast 561 45°N' 42 N. 39"N - 36°N 33"N 30 N - 27°N- i ME ' /NHl/^ MA VA iyihi0(l^ yini 9l4ni Northern '^,; NC ^*?- ^9 SC Cenlr.il GAf ■ • FL \* Southern NMFS Release Sites Frequency • 1-50 • 51-500 • 501-2000 #2001-5700 su.y ')lm 014m NC \ \ * • Southern <>' Central PL ALS Release Sites Frequency • 1-50 • 51-500 • 501-2000 • 2001-5700 ^-^• —I 1 72' W 69 W — 1 1 81°W 78 W 75 \V 72 W 69 W SIW 7S"W 75°W Figure 1 Distribution of release sites (by region) for bluefish {Pomatomus saltatrix) released during National Marine Fisheries Service (NMFS) (1963-1967) and American Littoral Society (ALS I (1983-2003) tagging programs. L,, + (L, L,,)(l -e-^-^'O, where L,^ = length at time of recapture; L, , = length at time of release; and • 4?, = change in time (years) between release and recapture. Parameter estimates were derived by using the SAS NONLIN procedure (SAS Inst.. Inc., Gary NC). ALS and NMFS data were modeled separately because of a differ- ence in the resolution of size measurements at release (centimeters for NMFS and inches for ALS). Bluefish at large less than 14 days, or fish with zero or negative growth, were excluded from the calculations. Geographic distributions for tag recoveries were com- pared with commercial catch locations of bluefish re- ported from the Middle Atlantic Bight. NMFS vessel logbooks from commercial fishermen in the Northeast region contain information about landings, discards, and the spatial location of catch. Bluefish catch loca- tions from 2001 to 2003 were summarized by month for all gear types. Length samples (measured to the nearest cm) collected by NMFS port agents from com- mercial landings were expanded to represent length distributions of total reported landings. Results During 1963 to 1967, the NMFS bluefish project tagged and released 15,699 bluefish; of these, 11,624 fish (74.0%) were captured with gill nets. 1393 (8.89i: ) with hook and line, 907 (5.8%) from beach seines, and 1775 (11.3%) from pound nets. Included in the hook-and-line total were 224 fish tagged by volunteer sportsmen in south Florida, New Jersey and New York (which resulted in 17 recaptures). The number of bluefish tagged and released are summarized by month and area in Table 1. From 1983 to 2003, recreational hook-and-line fishermen in the ALS tagging program caught and released 20,398 bluefish. Later recaptures totaled 1539 fish, of which 1075 were NMFS tagged and 464 were ALS tagged. The NMFS rate of tag return varied by capture meth- od, tag type, and tagging area. The highest return rates (9.6 %) were from pound net releases that could be attributed, in part, to a high recapture rate immedi- ately after tagging. The second highest percentage (6.9 %) and largest number of recaptures (802) were fish captured with gillnets. Although only 4.7%> of the fish released from hook-and-line gear were returned, over half of these released fish were from the first year of the tagging program (775 fish tagged, 16 returned, 2.1%), 562 Fishery Bulletin 104(4) Table 1 Number of bluefish iPomatoin us sa Itatnx) tags released by month and local ion of National Marine Fisheries Service ( NMFS I and 1 American Littoral Society (ALS) programs. Jan Feb Mar Apr May Ju n Jul Aug Sep Oct Nov Dec Total NMFS North New England 4 26 30 New York 196 31 1082 1859 150 3318 north New Jersey 614 614 2651 1811 5690 south New Jersey 3 1 4 Central Maryland Virginia 964 20 4 20 968 North Carolina 127 274 40 57 43 569 1110 South South Carolina north Florida 60 335 19 335 79 south Florida 510 572 1199 1008 100 425 331 4145 Total 510 572 1199 1403 1110 1111 686 3794 3739 819 425 331 15,699 ALS North New England 1 1 3 124 846 1233 1470 1845 1410 440 1 7374 New York 3 368 803 568 568 661 1143 323 2 4439 north New Jersey 76 1383 1181 814 956 961 837 411 22 6641 south New Jersey 4 90 18 18 25 26 11 192 Central Maryland 35 66 21 35 32 63 5 2 259 Virginia 8 153 122 8 6 3 5 18 14 337 North Carolina 11 21 6 4 14 10 421 124 26 637 South South Carolina Georgia 7 2 1 3 2 2 1 18 Florida 122 43 59 93 24 1 5 2 9 23 20 100 501 Total 123 44 59 201 2114 3115 2672 3072 3548 3930 1353 167 20,398 whereas the remainder of the hook-and-line releases (618 fish) resulted in 49 recaptured fish (7.99H. This difference may be attributable to less effective tags used during 1963, rather than the capture method. All ALS releases were fish caught with hook and line, but gear type was not recorded for the recaptured fish. Tag types differed in rates of return and in time fish spent at large. Return rates for NMFS tags were similar between dorsal loop tags (760 of 10,752; 7.1%) and internal anchor tags (310 of 4441; 7.0%). All NMFS tagged fish recaptured after 17 months at liberty had been tagged with internal anchor tags. Abrasion of some dorsal tags occurred after two months and appeared more severe on tags returned from Florida releases. Some abrasion occurred on the external streamer of the internal anchor tags. ALS dorsal loop tags resulted in an overall return rate of 2.2%. Return rates for the NMFS tags varied by area, from 4.5% in New York to 10.2% from Florida (Table 2). Eighty-three percent of recaptured tags that were at liberty one year («=24) were from tagging in areas north of New Jersey, and all recaptured tags at liberty 2, 3, and 4 years (10, 4, and 4 fish recaptured, respec- tively) were also from tagging in these areas (Table 2). For all areas combined, 51.3% of the NMFS recaptured fish were at liberty less than one month and 88.1% were at liberty less than five months. Among ALS tags, 50.3%f were at liberty for two months and 79.3% for five months or less. Only 4.5% of all returned fish were at liberty greater than one year and 1.2% greater than two years. The longest time-at-large for an ALS tagged bluefish was 1461 days and 1486 days for a NMFS- tagged fish. The size at release of the 36,097 tagged bluefish dif- fered between the two programs. Fork lengths (FL) ranged from 19 to 57 cm (7 to 22 inch) for NMFS and 15 to 114 cm (6 to 45 inch) for ALS (Fig. 2). The majority offish released in the NMFS program (83.1%-) were 30 to 45 cm (12 to 18 inch) and were 1-, 2-, or 3-year-olds according to estimates determined from scales from a subsample of similar size fish collected during the tag- ging operations. A greater size range of bluefish were tagged in the ALS program, and 88.2% measured 30 to 76 cm (12 to 30 inch). Size frequencies of bluefish released, and of those subsequently recaptured, differed despite the relatively short times at large (Fig. 2). Changes in length can be attributed to growth, unit conversion, and associ- ated measurement error. The NMFS recapture data set had 616 tag recoveries with appropriate fork length and time-at-large information, whereas the ALS set had 336 with these data (Fig. 3). The von Bertalanfy growth parameter estimates from these data were L^ of 100 cm and 118 cm (39 and 46 inch) for NMFS and Shepherd et al Migration routes of Pomatomus saltatnx along the Atlantic coast 563 CO fc. c o -tJ J CO Ph II CO c r/j ■•~ U. o "5 o 'cj w eg " ^1 CO CO .Si O I- ^ 2 " II ■ 2 u c |'& .^^ > CO II l> S -a c c OJ (1) o •-:) s s o CD ci: 2 T1 II C ■-5 2 3 Cb -ii n m > hn ? CO +j Oj "w z k II >- 2 t-! -o C/l c :r; CO hn o CI CI a £ S o u. 2 J3 II 3 ^ (t-i u 0) s CO -1 2: QJ li CO CD CO "-^ (M .-H O '-< 0.iai'*l>i0'^C^C0G0(M00'-i00OC0O iC CM LO'Xiooa^itDlr^-^C^loOtr-COOOOUOC-]'— I ^ CO 0^1 iC o CO CO CO >— • »c o: i^ .-t C^ (Jl O CO O O O O O CSl o iC (N CO CN O -rt* CO CD O -^ (N •-' (M CM --i tr- r- CD -^ CM CO T-H CO '-' Tf O^ O Tt' CD CO f-H CM 00 00 CM O -* "* r- CD CO '-' .-I .-H CO CO "^ CD CDOmOCO "^ COOOCOO-lCDCD'— I CO CO .— I T— I O -H ■* CD CD CM u-^ f« -^ CO o CM 5* CO t-1 o lO on ^ ^ ^ fS in C3) CD CO O r^ t^ f# on cm t^ CD to f* CO (N ^ CO CM CD CO CD in t" CO CO t- 0-) (jt M' ^ CO I— < CM o CO cn CO CO CO 00 05 CO CO .-I -* O 00 '^ ■* CD 'Ct^ CO CM ^ -^ CM CO CO 05 CM 1-H r^ ro CM CO CO OOOlr^-iOLOCO'-iC^CMCDiCiCiOCO ^ CO CM --H T-H OCMOr-iOCOCMCO"*'-t'^'-"'-fCM CM CO (M ,— 1 in (35 ■^ r- CO f% CO CM ^ CM O'-^CMCOTfiocDC-COOiO^CMCOTjHiocDt^GOOlOT-jC^CO^^in^ 1— l«-Hi— I.— I-—!!— (.— I.— ir-ii— iCMCMCMCMCMCM/h « 564 Fishery Bulletin 104(4) ALS, respectively, and K values of 0.230 and 0.206, respectively. The 95% confidence interval around L^ for the NMFS data ranged from 79 to 121 cm (31 to 48 inch) and ALS estimates ranged from 95 to 141 cm (37 to 56 inch). Confidence intervals for the K parameter ranged from 0.135 to 0.324 for the NMFS data and 0.109 to 0.303 for the ALS data. A single vector drawn between tag release and re- capture locations represents the shortest possible wa- ter route traveled between the two points. The actual route of travel likely diverts from this path; however it remains the basis for determining distance traveled. Movement of bluefish between western Long Island Sound and lower New York Bay through the East River was not indicated by the tag return data, and therefore passage of bluefish through the East River was not considered in the distance calculations. Seventy-five percent of the NMFS tags were recaptured within 100 km of the tagging location; whereas 9.7% were caught 350 km or farther from the tagging area. The maximum distance traveled was 2227 km (Chesapeake Bay to west Florida) for a fish at liberty 341 days (6.5 km per day). For the ALS recaptured fish, 59.7% were recap- tured within 100 km of the point of release and 40.3% (187) traveled beyond 100 km averaging 295 km, and 45 bluefish traveled more than 400 km. The maximum NMFS 800- 400 ■ Release (n=1 5, 699) D Recapture (n=325) Mrt^ rflJInn ALS Release ( n=20,398) r □ Recapture (n=202) --20 () -I— ^^^MJ I M|M M|l M 1,1 M 1,1 I I 1,1 M [7rn__^ 0 Fork length (cm) Figure 2 Frequency distributions for release and recapture fork lengths of bluefish (Pomatomus saltatrix) sampled in National Marine Fisheries Service (NMFS) and American Littoral Society (ALS) tagging programs. distance was 1096 km for a fish at liberty for 76 days (14.4 km/day). The number of days at liberty represents the maxi- mum time possible for a fish to travel the straight-line distance between the tagging and recapture locations. Movement was generally less than 5 km per day (84.3% of recaptures) and the overall average speed of fish recaptured in both programs was 2.6 km per day. Sev- enteen bluefish traveled more than 20 km per day and had a maximum swimming speed of 48.3 km per day for three days. The average swimming speeds for combined NMFS and ALS recaptured fish were significantly different among months (P=0.02). Speeds were highest during the autumn, peaking during October-November at 5.3 km per day, and there was another spike during April-May at 4.9 km per day (Fig. 4). These seasonal peaks were in contrast to the overall average of 2.6 km per day. Seasonal variations in swimming speed were considered indicative of periods of active migratory behavior. This annual cycle of movement was divided into four periods based on the average rate of travel for each month: winter residency (December to February), spring migration (March to May), summer residency (June to August), and autumn migration (September to November). The swimming speed and distance trav- eled during seasonal movements were dif- ferent for all bluefish size groups. Swim- ming speed offish 30-45 cm (12-18 inch) was greater than speed of fish 45-65 cm (18-26 inch) and this pattern was con- sistent across regions (Fig. 5). The speed disparity by size was influenced by the distance traveled, not by differences in time-at-large because the data were lim- ited to recoveries within 90 days. Smaller bluefish, particularly those released in the northern region, traveled farther dis- tances on average than fish in the larger size classes (Fig. 5). Movement of tagged fish among areas is a function of both the behavior of the animal and adequate time-at-large to exercise that behavior. Half (229 of 459 fish with location information) of bluefish tagged with ALS tags were recaptured beyond the state of release; the remain- ing recaptured fish were caught in the area of release (Table 3). NMFS releases had a higher proportion recaptured blue- fish within the state of release (77.1%; Table 3). Most long distance returns oc- curred during the first year after tag- ging; recoveries in years 1 through 4 af- ter tagging were generally from the area of release (62 of 100) or in the release area and adjacent areas (91 of 100). Fish recaptured from New Jersey northward accounted for 89 of the 100 recaptured ■■20 -- 10 Shepherd et al : Migration routes of Pomatomus saltotnx along the Atlantic coast 565 fish after one year at liberty, and all but two of these were from releases in this area. The higher percentage of re- captured fish from Florida may reflect greater annual fishing effort in contrast to northern areas. Seasonal recaptures of tagged bluefish along the coast (Fig. 6 1 were a function of both fishing effort and the abundance of tagged fish in the area. Over ninety- one percent of all recaptured NMFS fish (978 of 1,075) were caught in inshore ar- eas (including Chesapeake Bay, Delaware Bay, and Long Island Sound) or within one mile of the coast. Of the remaining 99 recaptures: one fish was caught 75 miles offshore and the rest within 30 miles of the coast; 47 between 1 and 6 miles, 40 between 6 and 20 miles, and 9 between 20 and 30 miles. Tagged fish released in New York and New Jersey accounted for 45 of the 50 returns over 6 miles offshore; 4 of those remaining fish were both released and recaptured off Florida, and the last fish had been re- leased in Florida and recaptured off New Jersey. All of these 45 returns released from New York and New Jersey were recovered north of Delav/are — most off New Jersey and New York. The offshore returns occurred from May through No- vember, although most of the fall recap- tured fish had been released during the same season (21 of the 23 recaptures in September, October, and November), whereas most spring and summer recap- tured fish had been at liberty for at least one year (19 of the 22 recaptures from May though July). Seasonal patterns of bluefish migra- tion, determined from information within recovered tags, generally followed one of three patterns: north-south between the northern Middle-Atlantic and Florida; a north-south pattern within the Middle- Atlantic; and year round movements be- tween inshore and offshore Florida. To simplify examination of migration pat- terns, recapture locations were divided into three regions. The area from Dela- ware to New England was designated the northern region, Maryland through North Carolina as the central region, and the area between South Carolina and southern Florida as the southern region. The single fish recovered in west- ern Florida was ignored for our analysis. A recovery matrix by region showed that greater than 90% of the bluefish recaptured within the northern or southern region had originated within the same region (Table 4). NMFS 411 - o ° • o o° ••• M - o ° °o :o - oo S^ • % .^IWr 0 Observed HI - -cftfi, •• E "^ ii 0 1 2 .1 4 S x: g o 5 40 -| ALS Mt - o « 20 - . • asL o ^ • O O 0 XOGD QDtfjffl • in -^gg5',_tf* o Observed | mmrW^ o • Predicted ^Koinoo o 0 < L ^^"^ ^ 1 1 1 1 1 I : .^ 4 5 Years at large Figure 3 Changes (observed and predicted) in fork length during time at large (years) for bluefish {Pomatomus saltatrix) sampled in National Marine Fisheries Service (NMFS) and American Littoral Society (ALSi tag- ging programs. s: 14 T (13 1 "O p . E ^ 10- T " 1 ■D 1 0) X H 1 1 T ID 1 T Q. ■"■ f 1 . t *,.**! ■ i i g C/l M % \ X \ \ ^-z \ % % \ \ \ Months Figure 4 Mean monthly swimming speeds (km/day) of bluefish iPomatomus salta- trix) ±95% confidence intervals from two combined program tag recov- eries: National Marine Fisheries Service (1963-1967) and American Littoral Society (1983-2003). However, within the central region, the percentage of recaptured fish originating in that region was only 80.37f . The remaining tags were either recaptured to the north (12.7%) or the south (7.0%). Distribution maps 566 Fishery Bulletin 104(4) of seasonal recaptured fish (Fig. 6) further highlighted the pattern of a northerly migration that began in May and was followed by southerly migration beginning in December. However, the entire coastal stock did not make the seasonal shift. Bluefish released in southern Florida were recaptured locally during the summer resi- dency period and other fish tagged in southern Florida appeared in coastal New Jersey (Table 3). Seasonal distribution data from bluefish tag returns indicated a southward movement of fish during the fall, culminating in an over-wintering aggregation near Cape Hatteras, North Carolina, and to the south. Com- mercial catch records from the Middle Atlantic for 2001 to 2003 indicated a greater seasonal inshore-offshore movement than that implied from tag recoveries. Ves- sel logbook data indicated that bluefish are present in the northern and central areas throughout the winter and have an increasingly offshore and southern distri- bution as water temperatures decrease. By March, the commercial catches increase off the coasts of Virginia and North Carolina (Fig. 7). The sizes of fish in these 16 -1 14 111 -I E g C/3 ♦ Nonh ■ Central A South Mean speed h * * * i i 100 80 - 6(1 ■ 4(1 ■ ;ii 0 ♦ North ■ Central ▲ South Days at large ^•»M ♦ « » M •^ inoo -1 ♦ North Distance SdO ■ ■ Central I 600 ■ A South , T 400 ■ ■ T 1 1 200 ■ 11 - ^ Ml.t-:*^ 1 — 1 — 1 — ♦ — 1 — 1 ^ V?' a^' v«^'■ S?' tf" kS^' -x^' # cp> Length (cm) Figure 5 Mean swimming speed, days at large, and distance traveled (±95'7r con- fidence interval) by length for bluefish (Pomatonuis saltatrix) sampled in National Marine Fisheries Service and American Littoral Society tagging programs. commercial catches were generally larger than those of tagged fish; lengths were between 25 and 70 cm and mean size was 45 cm. The commercial catch distribu- tion expanded northward beginning in April, in a direc- tion similar to the bluefish migratory route indicated by spring tag returns. Discussion A generally established hypothesis for bluefish migration is that fish undergo a seasonal coastal migration, leaving the Mid-Atlantic Bight in autumn as water temperatures decrease and moving as far south as Florida. In spring, bluefish return north where they spend the warmer months in habitat suitable for larger, mature fish — possibly in habitat extending across the continental shelf (Wilk, 1977; Fahay et al., 1999). Our hypothesis expands that migration pattern to suggest there are three bodies of fish with different migration behaviors. One group has an extensive north-south migration between New England and Florida, a second group has a migration route within the Mid- Atlantic Bight bounded to the south by North Carolina, and a third group has an inshore-offshore migration within Florida waters. The geographic ranges of these three groups overlap during at least part of the year and the seasonal areas of dis- tributions change with fish size. Bluefish from Massachusetts to Dela- ware (northern region) leave the coastal areas in autumn, and fish less than ap- proximately 45 cm (18 inch) migrate south along the coast and are found in the south Florida winter fishery. However, tag re- turn data indicate that part of this size group, as well as all fish over 45 cm (18 inch), move offshore in autumn and are distributed during winter in offshore ar- eas south of Virginia. Both size groups return to the northern area in spring and summer a year after tagging, as well as in successive years. The Mid-Atlantic area from Maryland to North Carolina (central region) appears to be a transitional area in the migratory route. A group of fish tagged in the north- ern region migrated in fall to the central region and remained throughout the win- ter, whereas other fish continued south to Florida. In the spring, a portion of the Florida fish moved back to the central region, while others continued through to the northern region. The distribution of commercial catches reported in logbooks confirmed the movement of bluefish in- to offshore waters during autumn, and movement increasingly south as winter progresses. The reported landings of large Shepherd et al Migration routes of Pomalomus saltatnx along the Atlantic coast 567 Table 3 Recapture matrix by state of release for National Marine Fisheries Service (NMFS) and American Littoral Society lALS) bluefish iPnmatomiis saltatrix) tag recoveries. NE = New England: NY = New York; NJ = New Jersey; DE/ME = Maine; MD = Maryland; VA = Virginia; NC = North Carolina; SC = South Carolina; FL = Florida. NMFS Release area Recapture area north south vIE NY NJ NJ 2 2 96 33 4 2 25 210 31 4 1 2 3 1 1 4 DE/MD VA NC SC north FL south FL west FL New England NY north NJ south NJ DE/MD VA NC SC north FL south FL 3 32 53 56 37 36 377 ALS Release area Recapture area ME MA RI CT NY NJ DE MD VA AL ME MA RI CT NY NJ DE MD VA NC FL 1 12 8 5 7 3 3 20 2 2 3 5 22 20 16 13 16 18 82 25 1 9 1 2 5 1 1 1 17 3 65 10 4 4 3 2 1 NC FL 19 bluefish fish during the Virginia-North Carolina win- ter fishery and the paucity of bluefish over 45 cm ( 18 inch) in the Florida fishery indicate that the larger fish do not continue the southern migration beyond North Carolina. Bluefish are found throughout the summer in south- ern Florida (southern region) as evidenced by tag re- turns and fishery landings, although at a low level of abundance compared with fall through spring landings. The one fall recaptured tag in New Jersey from a fish released in Florida in the spring (at liberty 17 months) indicates that fish from the southern group are found in northern areas in subsequent years. Bluefish found off Georgia and South Carolina appear to be transitory, and are present primarily in spring and fall. The rela- tive lack of tag recoveries in coastal South Carolina and Georgia implies a migratory pathway farther offshore, possibly closer to the shelf edge and Gulf Stream. Change in fish size leads to a change in the seasonal migration route. The migratory groups are best defined Table 4 Tag recovery matrix by region for combined National Marine Fisheries Service and American Littoral Society 1 tags ; data correspond to tag release and recaptures shown | in Figure 6. Release region Recapture region Recaptures North Central South North 812 67 22 901 (90'rc) n'i'c) (2%) Central 20 127 11 158 (12%) (80%) (6%) South 7 4 469 480 (1%) (1%) (98%) Recaptures 839 198 .502 1539 568 Fishery Bulletin 104(4) 42°N 38°N 34°N 30°N 26''N 42°N 38"N 34 N 30"N 26°N 91m "A^ -1 ami. i. T-T \ Jl i i Southeri Niirthcrn Release Dec-May Recaplure • Release Reeapllirc NY ;NH , ,'1._MA iCT.Bl''-:;, >iii^- Northern GA y Southern Central Release Dee-May Recapture • Release Recuplurc L,i£ Central .<.. , . ';. Southern Release p. '^^ Southern Dec-May Recaplure »^ • Release "^i, Recapture Mb i^aw.y \y\jr 91ni '^* -W-^ Mt ^ NY i MA>Z 91m KC ^4 ■! Ml) >•/ 914m Central ■<>■ NC /.. V- - sc ;>-^ / /! :;a /"',■«/) Central % FL \ Southern Northern Release Jun-Nov Recapture • Release C Recapture Central Release Southern jun_Nov Recapture • Release '.: Recapture Southern Release Jun-Nov Recapture • Release Recapture 82°W 78'W 74"W 70''W 66"W 82°W 78'W 74 W 70'W 66"W 82 'W 78'W 74 = W 70' W 66"W Figure 6 Geographic distribution of tag recoveries, by region of release and season of recapture, for bluefish iPomatomus saltatrix) sampled in National Marine Fisheries Service (NMFS) and American Littoral Society (ALS) tagging programs combined. by the seasonal aggregations of smaller fish ( less than approximately 45 cm. 18 inch FL). Juvenile bluefish use a coastal migratory route that extends farther south and offshore with increasing size. Some fish past the juvenile stage may remain in Florida for a season while others return to northern areas in the first or second spring. As growth continues, north-south migration routes become truncated — replaced by a route that keeps fish within the Mid-Atlantic Bight circuit. Because migration patterns appear to be size re- lated, growth rate will determine how long individuals maintain a particular migration behavior. Tag-based growth rate estimates (K=Q.2Z and 0.21) are similar to values reported from scale-based age studies, but the tag model resulted in larger theoretical maximum sizes of 100 and 118 cm (39 to 46 inch). Salerno et al. (2001) reported an average L, value of 87 cm and K equal to 0.26, whereas previously published estimates range from 67 to 128 cm (Lassiter. 1962; Wilk, 1977) and have K values ranging from 0.10 to 0.34 (Lassiter, 1962; Wilk 1977). The tag-based estimates of maximum size are similar in size to that of the largest fish reported in recreational landings (Salerno et al., 2001) and size of this largest fish caught in recreational landings could be considered an empirical estimate of L,. Otoliths have been shown to be preferable to scales for aging bluefish (Sipe and Chittenden, 2002). Growth parameters of bluefish from the South Atlantic, aged from otoliths (L, =101.9 cm, A:=0.10 [Barger, 1990]), compare well with our tag-based estimates of L^ but produce very different growth-rate estimates. The migration route bluefish follow from coastal New England to southern Florida, a distance of potentially >2000 km, is completed over the course of several months. Bluefish held in aquaria have been shown to travel at speeds from 40 to 60 cm/s and to have burst speeds up to 80-110 cm/s (011a et al., 1970). An ap- proximate average speed for bluefish similar in size to those tagged in the present study would be between 39 km per day (40 cm/s) and 59 km/day (60 cm/s). The majority of tagged fish moved between 0 and 5 km per day (84%), and most (97%) bluefish swam less than 20 km/day. However, several individuals averaged between 55 and 111 km per day. An average swimming speed Shepherd et al Migration routes of Pomatomus saltatiix along the Atlantic coast 569 44N 42-N 40°N 38°N 36' N 44'>N 42°N 40°N- SS^N 36"N •A' NH NY "^ t s MA BTo • \'^" '" 'I U m September Kilot-'rams o l-_S(lll O .'i()l-:(lll(l O 2(H)l-i:(«lll O 12001-2 7^1111 914 m NC October Kildgranis o 1-500 O 501-2000 O 2001-12000 O 1 200 1 -27300 .^, \ f \ ^Z^- NH y-' I, iCT R November Kilogriini^ 500 O 501-2000 O 2001-12000 O 12001-27300 VA NC December Kiloiirams o 1-500 O 501-2000 O 2001-12000 O 12001-27300 78=\V (i6"W 78"W 74' W Figure 7 Distribution of bluefish ^Pamatomua saltatnx) in the Middle Atlantic Bight based on commercial logbook catch records (2001-2003) for all gear types. greater than 50 km per day would imply that some type of passive transport supplemented active swimming movement. The offshore distribution of bluefish in the South Atlantic Bight during winter, as inferred from tags, may provide an opportunity for bluefish to use the Gulf Stream during the northern migration. Several recent studies have evaluated the stock struc- ture of bluefish along the Atlantic coast by using genetic material (Chiarella and Conover, 1990; Graves et al., 1993; Davidson, 2002) and morphometric characteris- tics (Austin et al., 1999). These studies have concluded that bluefish along the U.S. Atlantic coast comprise a single stock. The tag recovery information for bluefish illustrates differences in movement patterns among areas, but these groupings do not imply unique stock characteristics. Consequently, there is no evidence from the tag recovery information that refutes the single stock hypothesis. Implementation of a well-designed tag and release program is critical for analytical evaluation of migra- tion (Schwarz et al., 1993). Limitations associated with tag recovery must be accounted for in the design of an effective tagging program. Recovery of most tags is a function of fishing effort; therefore recaptured tags must be considered in the context of the fisheries that will provide the recoveries. Unknown variations in fish- eries over the past four decades may have influenced the patterns of bluefish tag recoveries in the NMFS and ALS programs, ultimately influencing interpreta- tion of the results. In addition, time-at-large for fish in both programs was generally less than one year which may be due to tag-induced mortality and tag loss (Henderson-Arzapalo et al., 1999). Bluefish tagged in Western Australia were generally at large for less than a year and experienced tag losses between 25% and 38% (Young et al., 1999). The NMFS and ALS programs demonstrate that tagging is a viable tool in analyses of bluefish populations, but it has limita- tions. Using the knowledge gained from these tagging programs, scientist may find that a renewed tagging effort incorporating recent technological advances will provide further insight into the migratory behavior of bluefish and, in particular, the behavior linked to environmental cues. 570 Fishery Bulletin 104(4) Acknowledgments The NMFS tagging project was conducted by David Deuel and his colleagues of NOAA Fisheries, Sandy Hook, New Jersey, Although their results were not published before Dave's death in 1994, his initiative in tagging bluefish made this study possible, for which we are grateful. We would also like to thank the American Littoral Society volunteers for data collected through their tagging pro- gram, Kristy Webber for data entry help, helpful reviews from Steve Cadrin, Mark Terceiro, and Fred Serchuk, and Ken Able for his constant encouragement to publish our tagging results. Literature cited Austin, H. M., D. Scoles, and A. J. Abell. 1999. Morphometric separation of annual cohorts within mid-Atlantic bluefish, Pomatomus saltatrix, using discriminant function analysis. Fish. Bull. 97(3):411-420. Barger, L. E. 1990. Age and growth of bluefish, Pomatomus saltatrix, from the Northern Gulf of Mexico and U.S. South Atlantic coast. Fish. Bull. 88(41:805-809. Carlsen, P. 2000. Profile of the American Littoral Society's fish tag- ging program. Fisheries 25(4):16-17. Collette, B. B., and G. Klein-MacPhee, eds. 2002. Bigelow and Schroeder's fishes of the Gulf of Maine, 3'''' ed., 748 p. Smithsonian Inst. Press, Washington, D.C. Chiarella, L. A. and D. O. Conover. 1990. Spawning season and first-year growth of adult bluefish from the New York Bight. Trans. Am. Fish. Soc. 119:455-462. Davidson, W. R. 2002. Population structure of western Atlantic bluefish (Pomatomous saltatrix). M.S. thesis, 89 p. Univ. Dela- ware, Newark, DE. Fabens, S. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. Fahay, M. P., P. L. Berrien, D. L. Johnson, and W. W. Morse. 1999. Bluefish, Pomatomus saltatrix, life history, and habitat characteristics. NOAA Tech. Memo. NMFS- NE-144, 68 p. Graves, J. E., J. R McDowell, A. M. Beardsley, and D. R. Scoles. 1993. Population genetic structure of the bluefish, Poma- tomus saltatrix, in Atlantic coastal waters. Fish. Bull. 90:469-475. Hare, J. A., and R. K.Cowen. 1993. Ecological and evolutionary implications of the larval transport and reproductive strategy of bluefish, Pomatomus saltatrix. Mar. Ecol. Prog. Ser. 98:1-16. Henderson-Arzapalo, A., P. Rago, J. Skjeveland, M. Mangold, P. Washington, J. Howe, and T. King. 1999. An evaluation of six internal anchor tags for tag- ging juvenile striped bass. N. Am. J. Fish. Manag., 19(2):482-493. Hersey, J. 1987. Blues, 224 p. Alfred A. Knopf New York, NY. Lassiter, R. R. 1962. Life history aspects of the bluefish, Pomatomus sal- tatrix ( Linnaeus), from the coast of North Carolina. M.S. thesis, 103 p. North Carolina State College, Raleigh, NC. Lund, W. A. Jr. 1961. A racial investigation of the bluefish, Pomato- mous saltatrix (Linneaus), of the Atlantic coast of North America. In Boletin del Institute Oceanografico, vol. 1, p. 73-129. Univ. de Oviente, Cumana, Venezuela. Lund, W. A., Jr, and G. C. Maltezos. 1970. Movements and migrations of the bluefish, Poma- tomus saltatrix, tagged in waters of New York and New England. Trans. Am. Fish. Soc. 99(4):719-25. Kendall, A., and L. A. Walford. 1979. Sources and distribution of bluefish. Pomatomus saltatrix, larvae and juveniles off the east coast of the United States. Fish. Bull. 77(l):213-227. Norcross, J. J., S. L. Richardson, W. Massman, and E. Joseph. 1974. Development of young bluefish, Pomatomus sal- tatrix, and distribution of eggs and young in Virginia waters. Trans. Am. Fish. Soc. 103(3):477-497. 011a, B. L., H. M. Katz, and A. L. Studholme. 1970. Prey capture and feeding motivation in the bluefish Pomatomus saltatrix. Copeia (2):360-362. Salerno, D. J., J. Burnett, and R. M. Ibara. 2001. Age, growth, maturity and spatial distribution of bluefish, Pomatomus saltatrix (Linnaeus), off the northeast coast of the United States, 1985-1996. J. Northw. Atl. Fish. Sci., Vol. 29:31-39. Schwarz, C. J., J. H. Schweigert, and A. N. Arnason. 1993. Estimating migration rates using tag-recovery data. Biometrics 49(1):177-193. Sipe, A. M. and M. E. Chittenden Jr 2002. A comparison of calcified structures for aging bluefish in the Chesapeake Bay region. Trans. Am. Fish. Soc. 131(4):783-790. Smith, W. P., P. Berrien, and T. Potoff 1994. Spawning patterns of bluefish, Pomatomus salta- trix, in the northwest continental shelf ecosystem. Bull. Mar. Sci. 54(1):8-16. Topp, R. 1963. The tagging of fishes in Florida 1962 program. Fla. State. Board Conserv. Prof Papers Serv. no. 5, 76 p. Watson, J. E. 1963. A method for tagging immature herring. U.S. Fish and Wildlife Service Spec. Sci. Report, Fisheries no. 541, 7 p. Wilk, S. J. 1977. Biological and fisheries data on bluefish, Pomato- mus saltatrix. U.S. Dep. Commer., NMFS/NOAA Tech. Ser. Rep. 11, 55 p. Young, G. C, B. S. Wise, and S. G. Ayvazian. 1999. A tagging study on tailor {Pomatomus salta- trix) in Western Australian waters: their movement, exploitation, growth and mortality. Mar. Freshw. Res. 50(7):633-64. Zar.J. H. 1974. Biostatistical analysis, 718 p. Prentice Hall, Inc., Englewood Cliffs, NJ. 571 Abstract — We evaluated light-based geulucation estimates from pop-up sat- ellite tags in high latitudes because some of the largest fisheries in the world are in areas where this tech- nique has not been assessed. Daily longitude and latitude were estimated by using two Wildlife Computers soft- ware programs: 1) Argos Message Processor (AMP), which summarizes light intensity data transmitted to satellites, and 2) Time Series Proces- sor (TSP), which uses more detailed data obtained from retrieved tags. Three experiments were conducted in the northern Gulf of Alaska using tags placed on 1) Pacific halibut in outdoor aquaria, 21 a fixed mooring line at various depths and 31 wild Pacific halibut. TSP performed better than AMP because the percentage of days with geolocation estimates was greater and the mean error magni- tude and bias were smaller for TSP and increased with depth for both programs: however, latitude errors were much greater than longitude errors at all depths. Light-based geolocation enabled us to discern basin-scale movements and showed that the Pacific halibut in our study remained within the Gulf of Alaska. We conclude that this technique pro- vides a feasible method for inferring large-scale population structure for demersal fishes in high latitudes. Evaluating light-based geolocation for estimating demersal fish movements in high latitudes Andrew C. Seitz (contact author) U S G S, Alaska Science Center 1011 E. Tudor Rd, MS 701 Anchorage, Alaska 99503 and University of Alaska Fairbanks Institute ol Marine Science PO Box 757220 Fairbanks, Alaska 99775-7220 E-mail address; aseltz@ims.uaf.edu Brenda L. Norcross University of Alaska Fairbanks Institute of Marine Science PO Box 757220 Fairbanks, Alaska 99775-7220 Derek Wilson Jennifer L. Nielsen U S G.S. Biological Resources Division Alaska Science Center 1011 E Tudor Rd, MS 701 Anchorage, Alaska 99503 Manuscript submitted 24 August 2005 to the Scientific Editor's Office. Manuscript approved for publication 20 December 200.5 by the Scientific Editor Fish. Bull. 104:571-578120061. Demersal fishes at high latitudes support some of the most lucrative fisheries in the world. An example is the Pacific halibut {Hippoglossus stenolepis) fishery off Canada and the United States. Currently, the Inter- national Pacific Halibut Commission (IPHC) manages the Pacific halibut population as a single, panmictic stock from northern California through the eastern Bering Sea based on genetic (Grant et al., 1984; Bentzen et al., 1998) and tagging data (Skud, 1977). However, Pacific halibut movements and population structure are not fully understood and mixing may be more restricted than assumed, as evidenced by a number of local depletions in recent years (Hare'). A method for estimating movements over large dis- tances is needed to improve the ability to identify populations and manage the harvest. Population structure and movement information is needed for management of several other high- latitude fisheries including Atlantic halibut (Hippoglossus hippoglossus). Pacific cod (Gadus macrocephalus) and Greenland turbot (Reinhardtius hippoglossoides) (God0 and Haug, 1988; Shimada and Kimura, 1994; Albert, 2002). New methods using information col- lected by electronic tags, which con- tain miniaturized onboard computers, are providing location estimates of demersal marine fishes (see review in Arnold and Dewar, 2001). One such method, the tidal location method, has been used to geolocate North Sea plaice (Pleuronectes platessa) (Hunter et al., 2003). This method compares the tidal range and time of high Hare, S. R. 2005. Investigation of the role of fishing in the Area 4C CPUE decline. Int. Pac. Halibut Comm. Report of Assessment and Research Activities 2004:185-198. Int. Pac. Halibut. Comm, PO Box 95009, Seattle, WA 98145- 2009. 572 Fishery Bulletin 104(4) water, as measured by the depth sensor of the electronic tag, to those predicted by tide models. Unfortunately, we are unable to use this method near Alaska because the water depth is much greater than in the North Sea. Deep water necessitates that the depth sensor of a tag have a greater range, which decreases depth resolution. Thus, tags used off Alaska have a depth resolution that is greater than the tidal range; therefore the tag cannot distinguish tidal fluctuations. Another tagging method has been used to geolocate Baltic Sea cod {Gadus morhita) (Neuenfeldt et al.-). This method is based on combined data of depth, tempera- ture, and salinity obtained by electronic tags attached to cod. Hydrographic fields obtained from hydrodynamic modeling are used as a geolocation database to identify daily locations of fish by comparison with the environ- mental data collected by each electronic tag. Unfortu- nately, the tags that we used are not available with a salinity sensor and hydrodynamic models of the area are not accurate on the bottom (Hedstrom'^). Ambient light data collected by electronic tags may be used to calculate daily estimates of latitude and longitude of fish. Geolocation by light has been imple- mented successfully on a variety of pelagic species to discern their daily position and movement patterns (Gunn and Block, 2001; Schaefer and Fuller, 2002; Itoh et al., 2003; Sibert et al., 2003). However, no studies have been conducted to evaluate light-based geolocation estimates from tags attached to demersal fish, nor from fish inhabiting high latitudes. Unfortunately, light levels in deep and high-latitude waters may be low and if the water is turbid, the light may be attenuated very quickly, thus hindering position estimates. Additionally, many demersal fishes inhabit a depth range where geolocation by light has not been evaluated at any latitude. The goal of this study was to examine the feasibility of using ambient light geolocation for estimating de- mersal fish movements in high latitudes. This was ac- complished by the following procedures: 1) by comparing daily latitude and longitude estimates from two propri- etary software types developed by Wildlife Computers, 2) by examining latitude and longitude estimates as a function of depth, and 3) by examining in situ latitude and longitude estimates of pop-up archival transmitting (PAT) tags attached to wild Pacific halibut. Materials and methods The pop-up archival transmitting tag (PAT, Wildlife Computers, Redmond, WA, vers. 2.0) is a miniature - Neuenfeldt, S., H.-H. Hinrichsen, and A. Nielsen. 2004. A method to geolocate eastern Baltic cod by using Data Storage Tags (DSTs), 14 p. Int. Coun. Explor.'Sea CM/L:06. Int. Coun. Explor. Sea, H.C. Andersens Boulevard 44-46, DK-1553, Copenhagen V, Denmark. ^ Hedstrom, K. 2005. Personal commun. Artie Region Supercomputing Center, Univ. Alaska Fairbanks, PO Box 756020, Fairbanks, AK 99775. computer that is attached externally to a fish. The tag contains a clock and sensors that collect depth, tempera- ture, and ambient light intensity data at user-specified intervals (Sibert, 2001). On a programmed date, the PAT tag disengages from the fish, floats to the surface, and transmits summaries of the recorded temperature, depth, and light data to Argos satellites; the data are then retrieved by the investigator. If the tag is retrieved, the complete archival record of temperature, depth, and ambient light data may be obtained. From October 2000 to March 2002, a pilot study was conducted to assess the feasibility of using PAT tags as a tool for identifying critical habitat of demersal fishes in high latitudes (Seitz et al., 2002, 2003). Geoposi- tion estimates were made from light data collected in three experiments in which PAT tags were attached to 1) Pacific halibut in outside aquaria, 2) a stationary mooring, and 3) wild Pacific halibut in situ. The tem- perature and depth data from the wild Pacific halibut experiment and their Argos-based final locations have been reported previously (Seitz et al., 2003). In the first experiment, two Pacific halibut were cap- tured, transported live to outside aquaria at the Alaska SeaLife Center (Seward, Alaska; 60.099°N. 149.440°W) and tagged on 18 Oct. 2000 with PAT tags programmed to record light intensity every minute. The tags were retrieved on 1 May 2001, and the longitudes and lati- tudes of Pacific halibut estimated from the tag data were compared with the known location of the Pacific halibut in the aquaria. A second experiment was conducted by using a fixed mooring to examine latitude and longitude esti- mates as a function of depth. From December 2000 to April 2002, four PAT tags were attached to a station- ary mooring line (the NOAA Alaska Observing Sys- tem's "GAK-l" mooring) in Resurrection Bay, Alaska (59.852^N, 149.330°W) at depths of 27, 57, 96 and 146 m. These tags were attached to four different current vanes on the mooring line so that their light sensors faced up. In a third experiment, to evaluate the performance of the light sensor and geolocation algorithm /;; situ. fourteen wild Pacific halibut (108-165 cm fork length) were captured, tagged, and released in November 2000, March 2001, and July 2001 from a commercial longlin- ing vessel in Resurrection Bay, AK, and off Cape Aialik, AK (for details, see Seitz et al., 2002, 2003). Light data were recovered from eight tags. PAT tags were tethered externally to each study animal by a piece of monofila- ment fishing line secured to a titanium dart that was inserted into the dorsal musculature of the fish. At a user-specified date and time, the PAT tag corroded the pin to which the tether was attached, thus releasing the tag from the animal. The tag floated to the surface and transmitted summarized data records through the Argos satellite system.^ After the tag popped-up to the ■• Service Argos, Inc. (http://www.argosinc.com). [Accessed on: 13 December 2005.1 Seltz et al Evaluating light-based geolocation for demersal fisfies in high latitudes 573 surface, its location was determined by the Dop- pler shift in the transmitted radio frequency in successive uplinks (Keating, 1995). The end- point position was the first location class (LC) estimate reported in the LCl-3 range, which all have error estimates <1.0 km. The basis of light-based geolocation is the estimation of times of sunrise and sunset. Two proprietary programs developed by the Wildlife Computers, Argos Message Processor (AMP, vers. 1.01.0007) and Time Series Pro- cessor (TSP, vers. 1.01.0008), were used to ex- tract times of sunrise and sunset from light intensity data. AMP identified daily sunrise and sunset times from light data transmit- ted through Argos satellites or directly from complete archival light records. TSP could be used only to identify sunrise and sunset times from complete archival light data from PAT tags that were physically recovered. In the next phase, another Wildlife Com- puters program. Global Position Estimator (GPE, vers. 1.01.0005), used the sunrise and sunset times to calculate the daily longitude and latitude of tags. First, we rejected days with light level curves that did not exhibit smoothly sloping light levels from high to low or low to high (Fig. 1). GPE was used to calculate longitude for the remaining data based on the local noon time of the tag (mean of the sunrise and sunset times). Estimated longitude values that were not possible for a fish released in the Gulf of Alaska were rejected from the data set. For example, an impos- sible longitude was one that placed the tag on land or outside the published range of the Pacific halibut (i.e., to the west of Hokkaido. Japan (140°E) or to the east of Santa Barbara, CA (117°W; Mecklenberg et al., 2002)). Once longitude was estimated, latitude was estimated by GPE, which used the "dav/n and dusk symmetry method" (Hill and Braun, 2001; Musyl et al., 2001). Daily latitude estimates were the theoreti- cal location of expected light levels that best matched the observed light levels measured by the tag. Lati- tude outliers were removed in the same manner as that used for longitude outliers. For all three experi- ments, the number of days with geolocation estimates was defined as the days that produced latitude and longitude estimates, after "bad" light curves (Fig. 1) and outliers were removed. For the tags with known positions in the tank and mooring experiments, we calculated bias and error magnitude based on true locations. Daily positional bias was calculated as the true position minus the estimated position (signed distance between posi- tions), and daily error magnitude was the absolute value of the bias (distance between points). For the tank experiment, we pooled the data from the two tags. Mean error magnitudes of software types were compared by using a two-tailed Ntest. For the fixed 140 120 ■ 100 80 ^,^»^- f-Ali^ '~'-^^vH, 40 20 0 0:00 ' 140 - 120 - 100 80 60 - 40 20 2 00 4:00 6 00 8 00 10 00 12:00 14:00 16:00 18:00 20:00 22:00 0 00 B A^-jA-J'-J^-M'- W iJ^^t .'VKjs,. P\ ■^^^,J V.wvvjJv_ 0:00 2:00 4 00 6 00 8:00 10:00 12 00 14:00 16:00 18:00 20:00 22:00 0 00 Time (AST) Figure 1 Examples of "good" and "bad" light curves. "Good" light curves I A) have smoothly sloping sunrise and sunset events. "Bad" light curves (B) do not have smoothly sloping sunrise and sunset events and produce outlying longitude and latitude estimates. The "good" light curve is from 2 March 2001 and the "bad" light curve is from the same tag on 10 March 2001. AST = Alaska standard time. mooring experiment, we calculated mean positional bias and mean error magnitudes for each tag and software combination. Mean biases were compared to a hypothetical bias of zero by using a two-way (tag and software) ANOVA model (vers. 8, proc GLM, SAS, Gary, NC). Mean error magnitudes were com- pared by using an ANOVA with a Tukey-Kramer test (Kramer, 1956; vers. 8 proc GLM). For both bias and error magnitude, the means are a measure of accuracy and the standard deviations are a measure of precision. For wild fish, it was impossible to know the true daily position of each fish for the duration of the experiment. However, for three of the eight tags released on wild fish, geolocation estimates were produced in the first or last six days of deployment. Therefore, we compared the estimated positions of the tags for the six days im- mediately following release of the tags and for the six days before recapture of the tags or before tags trans- mitted data to Argos satellites. All three of these tags were physically recovered and TSP produced estimates for all tags. AMP produced plausible estimates for one tag only because other estimates were rejected as outli- ers. For each comparison, we calculated the mean bias and mean error magnitude, assuming that the fish was stationary (or nearly so) during the first and last six days of the deployment. 574 Fishery Bulletin 104(4) Results All 14 tags, with the exception of one, functioned prop- erly for the duration of the three experiments. The one exception, attached to a halibut in situ, was deployed for 234 days, but it provided data for only the first 42 days because the battery failed. Tracking durations for AMP (range: 42-479 days) were always equal to or greater than the tracking durations for TSP (range: 42-348 days) because the memory for the archival data filled up before the summary data memory. In the tank experiment, TSP was a better estimator of longitude than AMP. TSP rejected fewer outliers and produced a higher percentage of days with longitude estimates (89.5%) than AMP (82.9%). Additionally, the mean longitude error magnitude for TSP (1.0° ±1.1° SD) was significantly smaller than that of AMP (2.0° ±3.2° SD) (^=.5.63, df=650, P<0.0001). Longitude errors were larger from late-fall to mid-winter in both tags when estimated by AMP, but not TSP. The mean longitude bias of TSP (-0.12° ±1.5° SD) was significantly smaller than that of AMP (-0.64° ±3.7° SD) (/ = 2.3. df=650, P=0.0215). TSP was not significantly biased and AMP had a significant mean longitude bias. In the tank experiment, TSP also produced a higher percentage of days with latitude estimates (88.2%) than AMP (81.6%). However, there was not a significant difference in the mean latitude error magnitude be- tween TSP (4.2° ±5.1° SD) and AMP (4.4° ±4.2° SD) (^=0.36, df=641, P=0.7155). The mean positional bias of TSP (-0.02° ±6.7° SD) was not significantly different (/<0.0001, df=641, P=0.9730) from that of AMP (-0.08° ±6.1° SD) and neither software type had a significant mean positional bias. In the fixed mooring experiment, TSP was a better estimator than AMP of longitude. In general, the tags produced fewer longitude estimates as depth increased, and at each depth, TSP generated more estimates than AMP (Fig. 2). The mean longitude error magnitude for both programs increased at greater depth (Fig. 3). The mean error magnitude of AMP and TSP estimates was not significantly different at 27 m and 57 m (P>0.50), but AMP estimates quickly degraded starting at 96 m (Fig. 3). For the tags at 96 m and 146 m, the mean error magnitudes for TSP estimates were significantly smaller (P<0.0001) than the AMP estimates of the same tags. The mean longitude biases of both AMP and TSP were generally to the west (positive values) of the ac- tual position of the tags, except for AMP at 96 m (Fig. 3). In several cases, the mean biases were relatively small for both AMP and TSP, however both had large variances. As with the longitude estimates in the fixed moor- ing experiment, the percentage of days with latitude estimates decreased at greater depths (Fig. 2). Unlike longitude, latitude was not estimated accurately by the tags. Mean latitude error magnitude was significantly smaller for TSP than for AMP at all depths, except 146 m (Fig. 3). The mean error magnitude for both AMP and TSP showed no relationship to increasing Percentage of days with position estimates 0 50 1 00 50 E. £ 100 150 200 TSP latitude TSP longitude AMP latitude AMP longitude Figure 2 Percentage of days with longitude and latitude estimates as a function of depth in the fixed mooring experiment. Two programs, Argos Message Processor (AMP) and Time Series Processor (TSP), were used to calculate daily longitude and latitude. depth (Fig. 3). The mean latitude biases of the tags in the fixed mooring experiment were greater than the mean longitude biases, and the biases by AMP were more variable than those of TSP (Fig. 3). Like longi- tude in the fixed mooring experiment, latitude was not estimated at 146 m during the winter and spring. The time span without geolocation estimates was longer for latitude (242 days) than for longitude (165 days). In the wild fish experiment, four tags reported on- ly to Argos satellites and geoposition was estimated from summary data by using AMP. The percentage of days with longitude estimates ranged from 0.0% to 2.3% (mean=l.l% ±1.0% SD), whereas the percentage of days with latitude estimates ranged from 0.0%' to 1.5% (mean=0.6% ±0.7% SD). The other four tags were physically recovered and geoposition was estimated by using both summary data for AMP and detailed data for TSP. For AMP, the percentage of days with longi- tude estimates ranged from 0.0% to 12.0% (mean=5.8% ±5.9% SD), whereas the percentage of days with lati- tude estimates ranged from 0.0 to 7.9% (mean=3.4% ±3.5% SD). For TSP, the percentage of days with lon- gitude estimates was higher, ranging from 9.9% to 32.3% (mean=19.7% ±9.4% SD) and days with latitude estimates ranged from 9.9% to 26.6% (mean = 16.9% ±7.2% SD). The mean error magnitude of the longitude estimates for AMP (;; = 4: 2.98° ±2.43° SD) was slightly larger than that of TSP (;;=10; 2.23° ±2.38 SD°). However, the mean error magnitude of the latitude estimates for AMP {n=4; 2.76° ±1.59° SD) was approximately half that of TSP (5.65° ±4.11° SD). The mean longitude bias Seitz et al Evaluating light-based geolocation for demersal fishes in high latitudes 575 for AMP (2.95° ±2.47° SD) was larger and to the east of that of TSP {-1.32° ±3.04° SD). The mean latitude bias was relatively small for both AMP (0.56° ±3.50° SD) and TSP (0.10° ±7.26° SD); however both had large variances and thus the estimates were not precise. In several cases, the longitude estimates were within one degree of the true position and there did not appear to be a pattern of over- or underestimating longitude. Discussion Geolocation estimates determined from ambi- ent light data in high latitudes is equally effective as in lower latitudes. Similar to results from previous geolocation evalua- tions (Welch and Eveson, 1999, 2001; Musyl et al., 2001; Teo et al.. 2004), our longitude estimates were in general more accurate and precise than latitude estimates. Therefore, longitude estimation by light is a promising technique for discerning large-scale move- ment of demersal fishes in coastal Alaska, but latitude estimation determined from light data only will not be adequate for these purposes. This study was unique in testing light- based geolocation in depths greater than 60 m. The results demonstrate the importance of evaluating geolocation by light for the entire depth range of the species of inter- est. Testing only in the near-surface waters would be misleading because the percent- age of days with estimates from tags at shallower depths was much greater than the percentage of days with estimates from tags at greater depths — the depths which halibut most frequently inhabit (Seitz et al., 2003). The accuracies of the longitude estimates in this study were comparable to those at lower latitudes and similar water depth. Errors are discussed in linear distance (Table 1) to account for the fact that a de- gree of longitude varies with latitude and to facilitate comparisons to previous studies. The longitude errors from the tank experiment were generally similar to the errors produced in a comparable experiment where tags were placed on a stationary mooring at a depth of 10 m (Welch and Eveson, 1999). The tags submerged at deeper depths in the fixed mooring experiment also showed a longitude error magnitude similar to that of location estimates from tags in the offshore region of the Gulf of Alaska at 50°N, 145°W (Musyl et al., 2001; Welch and Eveson, 2001). The longitude biases were only slightly larger than those from tags on a stationary mooring near Hawaii (Musyl et al., 2001). AMP mean error magnitude ( ) 0 5 10 15 20 25 AMP mean bias (°) -25 -15-5 5 15 25 -I 146 TSP mean error magnitude (°) 0 5 10 15 20 25 -25 TSP mean bias -15-5 5 15 25 :[:. h- 146 Figure 3 Mean (±SD) positional errors and bias in the fixed mooring experi- ment. Two programs, Argos Message Processor (AMP) and Time Series Processor (TSP), were used to calculate daily longitude (black bars) and latitude (hatched bars). Asterisks (*) indicate mean posi- tional biases that were significantly different from zero in tests with two-way ANOVA. A negative bias indicates that a position estimate was either north or east of the known position, and a positive bias indicates that a position estimate was either south or west of the known position. The minimum movement of a fish that was discerned by light-based geolocation in our experiment is the ab- solute sum of the error magnitude and bias. The sum of the error magnitudes and biases of TSP were generally smaller than those of AMP; therefore TSP was a bet- ter estimator of light-based geoposition than AMP and can be used to discern movement at a finer scale. The tank and fixed mooring experiments indicated that lon- gitude estimation by TSP is able to discern movements of approximately ±200 km for depths as great as 150 m and AMP is able to discern east-west movements of approximately ±350 km at 150 m deep. Geolocation by light will be able to discern the large-scale movements 576 Fishery Bulletin 104(4) Table 1 Linear distance of mean error magnitude for the tank. fixed mooring, and wild fish experiments. Longitude and latitude were calculated from light intensity data collected by pop-up archival transmitting tags in two programs. Argos Message Processor (AMP) and Time Series Processor (TSP). Mean error magnitude was calculated by averaging the absolute value of the difference | of the true position and the estimated position of the tag for each day of the experiment Total error was the vector distance from | the known locat ion of the tag when longitude and latitude errors were combined. The great circle formula was u sed to convert angular errors to linear distances. Depth AMP TSP Longitude Latitude Total error Longitude Latitude Total error Experiment Tag (ml (km) (km) (km) (km) (km) (km) Tank 00-0740 0 139.1 500.9 519.8 62.9 500.9 504.8 Tank 00-0741 0 83.5 480.8 488.0 48.4 445.2 447.8 Mooring 00-0822 27 66.2 873.7 876.2 41.7 505.3 507.0 Mooring 00-0826 57 74.6 696.8 700.7 74.6 310.5 319.4 Mooring 00-0806 96 241.5 1421.3 1441.7 89.6 540.9 548.3 Mooring 00-0824 146 299.3 726.8 786.0 123.5 871.5 880.2 Wild fish All tags 90-202 165.8 307.2 362.6 124.1 628.9 392.0 of Pacific halibut because this species performs spawn- ing migrations of over 1100 km (Loher^). Additionally, with recovery rates as high as 50% in area-specific con- ventional tagging experiments (Kaimmer, 2000), TSP can could be used for a large portion of tag recoveries in future experiments. At the largest scale, we were able to discern with confidence whether the wild Pacific halibut in this study were in the Gulf of Alaska or Bering Sea. Individual estimates were subject to occasional large errors and therefore caution should be practiced when using these estimates to represent the true position of fish. Examin- ing patterns in estimates is more useful for determining locations. To reach the Bering Sea (west of 157°W), a Pacific halibut would have to migrate from the Gulf of Alaska through False Pass (163. 5°W), which is the eastern-most connection between the two areas. The wild Pacific halibut in our study appeared to remain within the Gulf of Alaska, because fewer than S'/f of the longitude estimates were to the west of 163. 5°W, and those appeared to be erroneous because adjacent esti- mates did not consistently corroborate them. Trends in longitude estimates did not provide sufficient evidence to indicate that aiiy of the wild Pacific halibut swam to the Bering Sea. A variety of uncontrollable factors can cause intrinsic and extrinsic errors in geolocation estimates. The pre- dominant source of intrinsic error is refraction in the earth's atmosphere that is caused when light travels through the atmosphere and is bent by air and other molecules (Schaefer and Liller, 1990). This error limits the absolute accuracy of the estimates to a constant 0.32° longitude and a minimum of 0.7° latitude (Hill ^ Loher, T. 2005. Personal commun. Int. Pac. Halibut. Comm. PO Box 95009, Seattle, WA 98145-2009. and Braun, 2001). Extrinsically, light levels may be drastically influenced by changing external conditions, such as waves, water turbidity, diving behavior of the animal, biofouling, and cloud cover (Metcalfe, 2001). In particular, the Alaska coastal region frequently experi- ences large changes in weather systems that change cloud cover and sea-state on a daily, or even hourly, basis. One final consideration for errors is that accu- rate location estimates rely on unobstructed horizons. If the horizon is obstructed, such as by the mountains surrounding the coast of Alaska, it alters the time(s) of apparent sunrise (and sunset), thus affecting geoloca- tion estimates. The tank experiment was conducted in a deep, north-south fjord whose walls obstructed the horizon, and the fixed mooring experiment was adjacent to an island on the east and to steep coastal mountains to the west. Undoubtedly, these false horizons accounted for part of the errors and bias. One shortcoming discovered in the fixed mooring experiment was a conspicuous gap in longitude and latitude estimates from December to June at 146 m. This six-month gap was probably the result of low am- bient light levels during the winter associated with high latitudes. It is unknown why the gap lasted into the summer when ambient light drastically increased. However, for practical application in studies of Pacific halibut migration, light-based geolocation estimates will capture some individual migrations to the spawn- ing grounds as some Pacific halibut begin migrating in October and arrive on the continental slope by early November (Seitz et al., 2003). We may be able to increase the number of location estimates with some fine-tuning of both software types. Several days were rejected because of poor light read- ings. However, some days had smoothly sloping sunrise and sunset events that appeared to be sufficient for accurate geolocation estimates, but the software mis- Seitz et a\ Evaluating light-based geolocation for demersal fishes in high latitudes 577 identified sunrise and sunset. This misidentification typically occurred because there were occasional aber- rant light readings. The geolocation software identified these as sunrise, sunset, or both, and therefore gave bogus position estimates. There is an option to override these aberrant sunrise and sunset times when using TSP because the software allows manual selection of sunrise and sunset. For our study, we opted not to do this because we did not want to introduce subjectivity into sunrise and sunset times. We suggest that the soft- ware be modified by the manufacturer to select the next best times for sunrise and sunset so that the investiga- tor may reject aberrant light readings and yet allow the software to objectively choose sunrise and sunset. In future studies, we hope to improve geoposition estimates by statistically filtering (Sibert et al., 2003) or smoothing longitude estimates and by incorporating additional sensor data. For example, in conjunction with light data, tag-measured sea-surface temperature (SST) can be compared to remotely sensed SST, to signifi- cantly improve geolocation estimates (Teo et al., 2004). In the case of demersal fish that rarely, if ever, visit the sea surface, maximum daily depth can be used as representative of the total water depth in the region. We can compare the maximum daily depth sampled by an electronic tag to existing bathymetry data to estimate possible daily positions of the fish. We can then combine the geolocation estimated by light-level information with the depth information to yield a most plausible track of daily positions. Accurate description of the movement of fish is the cornerstone of sound management plans for ensuring sustainable fisheries in the future (Hunter et al., 2003). Longitude estimation determined from ambient light data may be used to examine large-scale movements of demersal fish in high latitudes. There are several types of electronic tags — some designed for fish as small as 15 cm (Arnold and Dewar, 2001). Using this technique, we can describe large-scale spatial dynamics and mi- gration of several commercially important demersal fish species. Acknowledgments This project was made possible through the expertise of Capt. Harold Kalve of the FV Rocinante during fishing operations; Dave Douglas, Dan Mulcahy DVM, and Julie Meka of USGS-Alaska Science Center; Susan Ingles, Richard Hocking, Pam Parker, Brian Mullaly, Jessica Dunning, and the staff at Alaska Sea Life Center; and Tom Weingartner, David Leech and the crew on the RV Alpha Helix from the University of Alaska Fairbanks. A special thanks to Devin Johnson for statistical assis- tance. Partial funding for this project was provided by the Exxon Valdez Oil Spill (EVOS) Trustee Council (Res- toration Project 478) and the U.S. Geological Survey- Alaska Science Center. The Gulf of Alaska (GAK) 1 mooring is funded by the EVOS Trustee Council under a separate grant (Restoration Project 340). The findings presented by the authors are their own and not neces- sarily the position of the EVOS Trustee Council. Literature cited Albert, O. T. 2002. Migration from nursery to spawning area in relation to growth and maturation of Greenland hali- but [Reinhardtius hippnglotisoides) in the Northeast Arctic. J. Northw. At). Fish. Sci. 31:1-13. Arnold, G. P., and H. Dewar. 2001. Electronic tags in marine fisheries research: a 30-year perspective. In Electronic tagging and track- ing of marine fisheries (J. R. Sibert, and J. L. Nielsen, eds.), p. 7-64. Kluwer Academic Publishers, Dordrecht, The Netherlands. Bentzen, P., J. Britt, and J. Kwon. 1998. Genetic variation in Pacific halibut [Hippoglos- sus stenolepis) detected with novel microsatellite markers. In International Pacific Halibut Committee (IPHC) Report of Assessment and Research Activities, p. 229-241. International Pacific Halibut Commission, Seattle, WA. Grant, W. S., D. J. Teel, T. Kobayashi, and C. Schmitt. 1984. Biochemical population genetics of Pacific halibut {Hippoglossiis stenolepis) and comparison with Atlantic halibut (H. hippoglossus). Can. J. Fish. Aquat. Sci. 41:1083-1088. God0, O. R., and T. Haug. 1988. Tagging and recaptures of Atlantic halibut, Hip- poglossus hippoglossus, in Norwegian waters. J. Cons. Int. Explor. Mer 44:169-179. Gunn, J., and B. Block. 2001. Advances in acoustic, archival, and satellite tagging of tunas. In Tuna: physiology, ecology and evolution (B. Block, and E. D. Stevens, eds.), p. 167-224. Academic Press, San Diego, CA. HUl.R. D., and M.J. Braun. 2001. Geolocation by light level, the next step: latitude. In Electronic tagging and tracking in narine fisheries (J. R. Sibert, and J. L. Nielsen, eds. ), p. 315-330. Kluwer Academic Pubis., Dordrecht, The Netherlands. Hunter, E., J. N. Aldridge, J. D. Metcalfe, and G. P. Arnold. 2003. Geolocation of free-ranging fish on the Euro- pean continental shelf as determined from environ- mental variables. I. Tidal location method. Mar. Biol. 142:601-609. Itoh, T., S. Tsuji, and A. Nitta. 2003. Migration patterns of young Pacific bluefin tuna {Thunnus orientalis) determined with archival tags. Fish. Bull. 101:514-534. Kaimmer, S. M. 2000. Pacific halibut tag release programs and tag release and recovery data. 1925 through 1998. International Pacific Halibut Commission Technical Report no. 41. 32 p. International Pacific Halibut Commission, Seattle, WA. Keating, K. A. 1995. Mitigating elevation-induced errors in satellite telemetry locations. J. Wildlife Manag. 59:801-808. Kramer, C. Y. 1956. Extension of multiple range tests to group means with unequal numbers of replications. Biometrics 12:307-310. 578 Fishery Bulletin 104(4) Mecklenburg, C. W., T. A. Mecklenburg, and L. K. Thorsteinson. 2002. Fishes of Alaska, 1037 p. Am. Fish. Soc. Bethesda. MD. Metcalfe, J. D. 2001. Summary report of a workshop on daylight mea- surements for geolocation in animal telemetry. In Electronic tagging and tracking in marine fisheries (J. R. Sibert, and J. L. Nielsen, eds.l. p. 331-342. Kluwer Academic Pubis.. Dordrecht, The Netherlands. Musyl, M. K., R. W. Brill, D. S. Curran, J. S. Gunn, J. R. Hartog, R. D. Hill, D. W. Welch, J. P. Eveson, C. H. Boggs, and R. E. Brainard. 2001. Ability of archival tags to provide estimates of geographical position based on light intensity. In Elec- tronic tagging and tracking in marine fisheries (J. R. Sibert, and J. L. Nielsen, eds.), p. 343-367. Kluwer Academic Pubis., Dordrecht. The Netherlands. Schaefer. K. M., and D. W. Fuller. 2002. Movements, behavior and habitat selection of bigeye tuna iThunnus obesus) in the eastern equato- rial Pacific, ascertained through archival tags. Fish. Bull. 100:765-788. Schaefer, B. E., and W. Liller. 1990. Refraction near the horizon. Publ. Astron. Soc. Pac. 102:796-805. Seitz, A., D. R. Wilson, and J. L. Nielsen. 2002. Testing pop-up satellite tags as a tool for identi- fying critical habitat for Pacific halibut iHippoglossus stenolepis) in the Gulf of Alaska. E.xxon Valdez Oil Spill Restoration Project Final Report (Restoration Project 01478), 43 p. U.S. Geological Survey, Alaska Science Center, Anchorage, AK. Seitz, A. C., D. Wilson, B. L. Norcross, and J. L. Nielsen. 2003. Pop-up archival transmitting (PAT) tags: a method to investigate the migration and behavior of Pacific halibut Hippoglossus stenolepis in the Gulf of Alaska. Alaska Fish. Res. Bull. 10(2):124-136. Shimada, A. M., and D. K. Kimura. 1994. Seasonal movements of Pacific cod, Gadus macro- cephalus. in the eastern Bering Sea and adjacent waters based on tag-recapture data. Fish. Bull. 92:800-816. Sibert, J. 2001. Electronic tagging and tracking in marine fisheries. In Electronic tagging and tracking in marine fisheries ij. R. Sibert, and J. L. Nielsen, eds.). p. 1-6. Kluwer Academic Pubis., Dordrecht, The Netherlands. Sibert, J. R., M. K. Musyl, and R. W. Brill. 2003. Horizontal movements of bigeye tuna [Thunnus obesus) near Hawaii determined by Kalman filter analy- sis of archival tagging data. Fish. Oceanogr. 12(3l: 141-151. Skud, B. E. 1977. Drift, migration, and intermingling of Pacific hali- but stocks. International Pacific Halibut Commission Scientific Report no. 63, 42 p. International Pacific Halibut Commission, Seattle, WA. Teo, S. L. H., A. Boustany, S. B. Blackwell. A. Walli, K. C. Weng, and B. A. Block. 2004, Validation of geolocation estimates based on light level and sea surface temperature from electronic tags. Mar. Ecol. Progr. Ser. 283:81-98. Welch, D. W., and J. P. Eveson. 1999. An assessment of light-based geoposition esti- mates from archival tags. Can. J. Fish. Aquat. Sci. 56:1317-1327. 2001. Recent progress in estimating geoposition using daylight. In Electronic tagging and tracking in marine fisheries (J. R. Sibert, and J. L. Nielsen, eds.), p. 369-383. Kluwer Academic Pubis., Dordrecht, The Netherlands. 579 Abstract — Information on the sea- sonal abundance and distribution of whale sharks {Rhincodon typiist is largely unknown throughout range of the species. Between 1989 and 1998, three spatially and temporally inten- sive aerial surveys were conducted by the National Marine Fisheries Service, Mississippi Laboratories that provided information on sea- sonality, distribution, and aggrega- tions of whale sharks in the northern Gulf of Mexico. Transects totaling 89.369 km were surveyed over the course of the study and a total of 1 19 whale sharks were counted during 81 sighting events. There was no sta- tistical difference in the sightings per unit of effort (SPUE) of whale sharks between the eastern and western continental slope waters of the Gulf of Mexico. In the continen- tal slope waters of the eastern Gulf of Mexico, whale sharks were more abundant during the summer than in the winter. In the western Gulf of Mexico, whale shark SPUE was significantly greater in the summer than during the fall or winter; there was no significant difference between summer and spring. There was also no significant difference in whale shark SPUE among spring, fall, and winter in the western Gulf of Mexico. Aggre- gations of whale sharks were seen only during the winter and summer, and there were significantly more individuals per aggregation during the summer. The largest aggregation of whale sharks observed during the study consisted of 23 individuals. Abundance and distribution of whale sharks (Rhincodon typus) in the northern Gulf of Mexico Carolyn M. Burks William B. Driggers III (contact author) Keith D. Mullin National Oceanic and Atmospheric Administration National Marine Fisheries Service Southeast Fisheries Science Center Mississippi Laboratories P,0 Drawer 1207 Pascagoula, Mississippi 39568 Email address for W, B, Driggers: William, Driggersiainoaa gov Manuscript submitted 8 January 2005 to the Scientific Editor's Office. Manuscript approved for publication 13 January 2006 by the Scientific Editor. Fish. Bull. 104:579-584 (2006). Whale sharks {Rhincodon typus) are distributed in all tropical and subtrop- ical marine waters of the world, with the exception of the Mediterranean Sea (Compagno, 1984). This species is considered vulnerable to extinction according to the World Conservation Union (lUCN)' partly because of a lack of information pertaining to its life history. Although the seasonality of whale sharks has been examined in two geographically discrete areas (Taylor, 1996; Duffy, 2002), no infor- mation exists for the seasonal distri- bution or relative seasonal abundance of this species over a broad spatial scale. The paucity of such information is probably attributable to logistical difficulties associated with collecting required data or to the expense of sur- veying large areas. Whale sharks aggregate in areas of high biological productivity, and seasonal abundance of whale sharks could result from increased localized prey abundance. Whale sharks feed on a variety of organisms including invertebrates and teleosts (Com- pagno, 1984). Unlike basking iCe- torhiniis ?naxi}')nis} and megamouth {Megachasma pelagios) sharks, which passively filter prey from the water column, whale sharks are capable of suction filter feeding (Colman, 1997). Although this feeding strategy en- ables whale sharks to capture a wider range of prey in terms of size, mo- bility, and diversity than other filter feeding elasmobranchs, this feeding strategy requires dense aggregations of prey in order for whale sharks to meet their energetic demands (Com- pagno, 1984). Feeding aggregations of whale sharks have been reported in the Atlantic, Indian, and Pacific Oceans and specifically in the wa- ters off Belize, Western Australia, the Galapagos Islands, Mexico, New Zealand, and Thailand (Taylor, 1996; Clark and Nelson, 1997; Colman. 1997; Eckert and Stewart, 2001; Hey- man et al., 2001; Wilson et al., 2001; Duffy, 2002). The whale shark was first described in 1828 from a type specimen col- lected off the coast of South Africa (Penrith, 1972). The first record of a whale shark in the western North Atlantic Ocean was not reported until 1902 and it was 1937 before this spe- cies was documented in the Gulf of Mexico (Gudger, 1939; Baughman and Springer, 1950; Breuer, 1954). Since 1937 several authors have reported sightings of whale sharks in the Gulf of Mexico (Gudger, 1939, 1941; Baugh- man, 1947, 1950, 1955; Gunter and Knapp, 1951; Breuer, 1954; Springer, 1957; Clark and von Schmidt, 1965; Hoffman et al., 1981; Hoffmayer et al., 2005). However, these reports are restricted to spatially discrete areas, and most are primarily anecdotal and largely based on isolated observations of few individuals. In the present study we report the seasonality, rela- 1 The World Conservation Union. 2005. http://www.redlist.org [accessed 5 Janu- ary, 2006]. 580 Fishery Bulletin 104(4) SO-N - 28"N - 26-'N 24"N 46'VV 44 W 92°W mi W ssw Sd'W 84 W Figure 1 Transect lines surveyed by aircraft during three aerial studies conducted in the northern Gulf of Mexico from 1989 to 1998. Note that some lines overlap others, particularly in the eastern Gulf of Mexico. Bathymetric contours represent the 100-. 500-, 1000- and 2000-meter contours. tive seasonal abundance, distribution, and aggregations of whale sharks observed in the northern Gulf of Mexico during three spatially and temporally intensive aerial surveys conducted by the National Marine Fisheries Service, Mississippi Laboratories. Materials and methods Between 1989 and 1998, three aerial surveys were con- ducted in the northern Gulf of Mexico with a DeHavil- land Twin Otter turbine engine aircraft. The primary objective of one of these surveys, referred to as the Upper Continental Slope (UCS) Survey, was to examine species composition, distribution, and seasonality of cetaceans in the north-central Gulf of Mexico (Mullin et al., 1994). The purpose of the other two surveys, referred to as the Gulf of Mexico Cetacean Studies I (GulfCet I) and Gulf of Mexico Cetacean Studies II (GulfCet II) (Mullin and Hoggard-) surveys, was to assess possible impacts of petroleum industry activities on cetaceans and sea turtles in the Gulf of Mexico. Standard line-transect sampling methods were used for each survey (Buckland et al., 1993). Surveys were flown at a constant altitude and air speed of 229 m and 200 km/h, respectively. The aircraft was modified with a Plexiglas observation bubble on the port and starboard sides of the fuselage to permit an unobstructed view of the area along transect lines and a lateral view to each horizon. When fauna Mullin, K. D., and W. Hoggard. 2000. Visual surveys of cetaceans and sea turtles from aircraft and ships. In OCS Study MMS 96-0027, vol. II, Cetaceans, sea turtles and sea- birds in the northern Gulf of Mexico: distribution, abundance and habitat associations (R. W. Davis, W. E. Evans, and D. Wursig, eds.), p. 111-172. Minerals Management Service, Gulf of Mexico OCS Region, New Orleans, LA 70123. of interest were sighted, the pilot circled the area until all observed fauna were identified to the lowest pos- sible taxa and the latitude and longitude of the sighting and numbers of conspecifics sighted were recorded. All survey effort was limited to waters associated with the continental slope in Beaufort sea states less than 4 and restricted to areas within the United States Exclusive Economic Zone (EEZ; Fig. 1). For all analyses and dis- cussion, seasons were defined as spring (April-June), summer (July-September), fall (October-December), and winter (January-March). The UCS survey was conducted from the summer of 1989 through the spring of 1990 and was carried out during all four seasons. The study area was lo- cated in the central northern Gulf of Mexico along the continental shelf break (~200-m isobath) south of the Mississippi River Delta and extended from the DeSoto Canyon (87°00'W) to the western edge of the Missis- sippi Trough (90°05'W). GulfCet I surveys occurred from 1992 to 1994 over continental slope waters in the western Gulf of Mexico between the United States and Mexico border (25°57'N) and the Mississippi-Alabama border (88°25'W) and were conducted during all four seasons (Mullin et al., 2004). GulfCet I transects began at the 100-m isobath and extended to the 1000-m iso- bath west of 90°00'W, and to the 2000-m isobath east of 90°00'W. GulfCet II surveys were conducted only during the winter and summer between 1996 and 1998 in the northeastern Gulf of Mexico and covered continental slope waters (100-2000 m deep) and a portion of the continental shelf (Mullin and Hoggard-). The entire spatial range of the three areas was surveyed within each season of operations. Because of the spatial overlap of the UCS and GulfCet II surveys, the central northern Gulf of Mexico and the western Gulf of Mexico were treated as a single region, referred to as "the western Burks et al : Abundance and distribution of Rhincodon typus in the northern Gulf of Mexico 581 Table T Summary of survey effort, number of whale shark [R/uncodoii 0 season, and survey. UCS=upper contintental survey: GulfCetl Cetacean Studies II. •pus = Gu sightings, and sightir If of Mexico Cetacean gs per unit of effort i SPUE I by region. Studies I: GulfCet II = Gulf of Mexico Survey and region covered Season Effort ( km 1 Number of sharks SPUE UCS — western Gulf of Mexico winter spring summer fall 4382 7615 5321 3275 1 8 14 0 0.23 1.05 2.63 0.00 GulfCet I— western Gulf of Mexico winter spring summer fall 12,645 12,645 12,942 11,756 13 20 43 7 1.03 1.58 3.32 0.60 GulfCet II— eastern Gulf of Mexico winter summer 8348 10,440 3 10 0.36 0.96 Gulf of Mexico," in this article. Mobile Bay, Alabama (88°00'W) was considered to be the dividing line be- tween the eastern and western Gulf of Mexico. Sightings per unit of effort (SPUE) were calculated to correct for unequal effort among regions by dividing the number of sightings per season by the total survey effort during the same season and then multiplying the resulting value by 1000. Because aerial surveys were limited to summer and winter in the eastern Gulf of Mexico, data from these two seasons in this region were compared (by using a ^-test) to data collected during surveys in the same two seasons in the western Gulf of Mexico to determine if mean whale shark SPUE was significantly different between the two areas. To deter- mine if there was a season of peak whale shark SPUE in the western Gulf of Mexico a one-way analysis of variance (ANOVA) and the Student-Newman-Keuls test were employed. The Mann-Whitney test was used to de- termine if there was a significant relationship between season and number of individuals per aggregation. This test was selected because of the non-Gaussian distribu- tion of these data. The location of each whale shark sighting was plotted on navigational charts, produced by the National Oceanic and Atmospheric Administra- tion (Chart numbers 11006, 11300, 11340), to examine the associations of whale sharks with bathymetric fea- tures such as reefs and salt diapirs. Statistical tests were preformed according to the methods of Zar (1999) at a significance level of PsO.05. Results Transects totaling 89,369 km were surveyed over the course of the study (Table 1). A total of 119 whale sharks were counted during 81 sighting events (Fig. 2). There was no statistical difference in the SPUE of whale sharks between the eastern and western continental slope waters of the Gulf of Mexico (/-test, df=4, f-value = 1.06. P=0.35). Because survey effort in the eastern Gulf of Mexico was limited to summer and winter, it was not possible to quantitatively analyze whale shark seasonal- ity in this region. However, results indicated that whale sharks are more abundant in the eastern Gulf of Mexico during the summer (SPUE = 0.96) than in the winter (SPUE = 0.36i. There was a statistically significant differ- ence in whale shark SPUE among seasons in the western Gulf of Mexico (ANOVA. df=7, F-ratio = 12.97, P=0.02). The Student-Newman-Keuls test indicated that whale shark SPUE was significantly greater in the summer (.f = 2.98, SD = 0.49) than during the fall (.r=0.30, SD = 0.18) or winter (x=0.63, SD=0.32). However, there was no sig- nificant difference between summer and spring (.v=1.32, SD = 0.38). There was no significant difference in whale shark SPUE among spring, fall, and winter. Of the 119 whale sharks sighted over the course of the study, 45 were observed in aggregations — an ag- gregation being defined as the presence of two or more whale sharks in close proximity to one another. Seven aggregations, ranging in size from 2 to 23 individuals, were observed. Aggregations of whale sharks were seen only during the winter and summer and there were sig- nificantly more individuals per aggregation during the summer (Mann-Whitney test, W=12.0, P=0.05). Sixty- two percent (;;=74) of whale sharks were not observed in association with conspecifics. The majority of whale sharks observed were not as- sociated with discrete areas of high bathymetric relief Thirty-four of the whale sharks sighted were in close proximity to relatively small, high relief diapiric fea- tures dominated by coralline algae (Rezak et al., 1990); Ewing Bank (28°06'N, 9r02'W), Bright Bank (27°53'N, 93°18'W) and 28 Fathom Bank (27°55'N, 93"26'W). Of the seven aggregations observed during the surveys, three were associated with the aforementioned banks including the aggregation consisting of 23 individu- als. The area of highest whale shark abundance was located in an area of approximately 16,800 km- and 582 Fishery Bulletin 104(4) 30°N 28°N - 26°N - 24"N 96°W 94"W 1)2 W <-)() W XS W Sh'W 84 W Figure 2 Locations (;!=81) of whale sharks iRhirjcodon typus) observed during aerial surveys in the northern Gulf of Mexico from 1989 to 1998. Dashed line demarks the separation between the eastern and western Gulf of Mexico as defined in our study. Area contained within the circle indicates region of highest observed whale shark abundance. Bathynietric contours represent the 100-, 500-, 1000- and 2000-meter contours. Location and season of sighting symbols: A = winter, O = spring, D = summer. 0 = fall. the center of their distribution was approximately 140 km southwest of the Mississippi River Southwest Pass (Fig. 2). Twenty-seven whale sharks were observed in this region over a four-year period. Discussion We observed whale sharks throughout the year in the Gulf of Mexico, and our study is the first to identify a broad region where whale sharks are present during all seasons and where they are perhaps resident throughout the year. The highest SPUE values that we observed occurred during the summer. Therefore, given the reported seasonality of whale sharks in the western Caribbean Sea (Heyman et al., 2001) and southeast- ern Gulf of Mexico (Gudger, 1941), it is reasonable to assume that a portion of the population migrates into the northern Gulf of Mexico during the spring and in winter moves into lower latitude waters, such as the Bay of Campeche and waters off the coast of Cuba. The only entrances to the Gulf of Mexico are through the Yucatan Channel and the Straits of Florida; therefore research with telemetry and satellite tagging in these regions would help determine the timing and routes of whale shark migrations and the period of residency within the Gulf of Mexico. Whale sharks are more abundant in the western than in the eastern Gulf of Mexico. However, because the aerial surveys were limited to two seasons in the east- ern Gulf of Mexico, little can be inferred from our data, except confirmation of the presence of whale sharks in the eastern Gulf of Mexico during the winter and the addition of summer to the seasonality of occurrence in this region. Because the survey reported in our study focused solely on continental slope waters, it is possible that whale sharks in the eastern Gulf of Mexico use continental shelf waters to a greater extent and, thus, were outside of the sampled area. Despite extensive ef- fort over the broad upper continental slope off the west coast of Florida, no whale sharks were observed south of 27'38'N. However, at higher latitudes, 13 individuals were sighted in continental slope waters at depths rang- ing from 70 to 180 m. Because the upper continental slope is broader in the eastern than in the western Gulf of Mexico, it is possible that whale shark distribution in the eastern Gulf of Mexico is more diffuse than would be expected, especially if this species is associated with steep bathymetric relief, which promotes upwelling of nutrient rich waters. During our study, two whale sharks were observed at the head of the DeSoto Can- yon, an upwelling area south of the Florida panhandle. A large aggregation (30-100 individuals) was found in this same area by Hoffmayer et al. (2005). In the western Gulf of Mexico, whale sharks were ob- served in all sampled depth strata, and the spatial distri- bution of sightings was fairly continuous along the con- tinental slope from Mobile Bay (88°00'W) to the United States/Mexico border (96°30'W: Fig. 2). Seasonally, whale sharks were present year round in the western Gulf of Mexico, except between 94°00'W and 95°35'W where they were observed only during spring and fall. Bimodality in seasonal occurrence in this region could be attributed to seasonal fluctuations in biological productivity and, thus, to prey availability resulting from ephemeral mesoscale phenomena such as Loop Current eddy formation, cur- rent reversals, and coastal jetting. For example, in the western Gulf of Mexico the Texas Current is capable of Burks et a\ : Abundance and distribution of Rhincodon typus in the northern Gulf of Mexico 583 generating localized upwelling on the continental shelf break, thus increasing primary productivity and the availability of prey species (Sahl et al., 1993). However, the Texas Current occasionally flows in a direction oppo- site to its normal eastward course because of wind forc- ing (Sahl et al., 1997). Such a reversal of direction in the Texas Current could affect productivity in this area and thus influence the seasonal abundance of whale sharks in this region. Given the stochastic nature of the physical oceanography of this region, additional research will be needed to determine if whale sharks make limited use of this area during discrete times of the year or if they are in fact present year round. From observations made by commercial mariners in transit between the southern tip of Florida and unidentified ports in Texas, Gudger (1941) reported sightings of 68 whale sharks from a circular area in the north-central Gulf of Mexico with a diameter of ap- proximately 280 km; however no specific locations were given. He concluded that the density of whale sharks in this area must be related to high prey densities. The area of highest regional abundance during the present study occurred in a circular area, bounded by 89°30'W and 91°00'W longitude, and having a diameter of approximately 165 km (Fig. 2). Three aggregations of between 30 and 100 whale sharks in this same area were reported by Hoffmayer et al. (2005). The safety fairway used by shipping traffic transiting between the eastern and western Gulf of Mexico crosses this area. It is probable, therefore, that the observations of whale sharks by commercial mariners reported by Gudger (1941) were in the same area. Because of nu- trient loading from the Mississippi and Atchafalaya rivers, primary productivity is higher in this region of the Gulf of Mexico than in any other region (Lohrenz et al., 1999) and may explain the high localized abun- dance of whale sharks in this region. Other highly productive areas where whale sharks are known be present in relatively large numbers include Ningaloo Reef, Australia, and Gladden Spit, Belize (Taylor, 1996; Heyman et al., 2001). The sighting of 23 whale sharks on 9 August 1993 was the largest aggregation observed during our study. The sharks, which were observed for about 35 minutes, were distributed over an area of approximately 2.6 km'-. Coral spawning occurred on 9 August 1993 at the East Flower Gardens (27°55'N, 93°36'W; Gittings^) which is approximately 33 km due west of the location of the ag- gregation sighting. A similar relationship between coral spawning events and increased localized abundance of whale sharks has also been noted to occur off the coast of Western Australia (Taylor, 1996). Future efforts should examine the relationship between annual coral spawning events and whale shark occurrence near the East Flower Gardens to determine if there is an annual migration of whale sharks to this area. ^ Gittings, S. R. 1993. Personal commun. Science Program Manager, NOAA Office of National Marine Sanctuaries. 1305 East West Hwy, Silver Springs, MD 20910. There are inherent difficulties associated with exam- ining the seasonality of marine organisms using aerial surveys. This is particularly true when the species of interest is not obligated to surface waters where visual observations can be made. Because the whale shark is not a surface obligate species, factors such as changes in feeding behavior resulting from seasonal variabil- ity in the vertical distribution of prey species could have introduced significant bias into our sighting data, subsequent analyses, and data interpretation. We as- sumed that no differences exist among seasons in the amount of time whale sharks spend at or in close prox- imity to the surface. Using telemetry and archival tags, Gunn et al. (1999) determined that whale sharks spend 17-53% of their time at the surface during daylight hours off the coast of Australia. Over periods ranging from 28-1144 days, Eckert and Stewart (2001) tracked vertical movements of whale sharks through the water column in the Sea of Cortez using satellite tags. They concluded that during all seasons whale sharks spend greater than 80% of their time at depths of 10 m or less during all seasons — depth well within the photic zone of continental slope waters of the Gulf of Mexico. The findings of both Gunn et al. (1999) and Eckert and Stewart (2001) support our assumption. Due to the economic value of their meat, fins, and liver oil, whale sharks have recently been designated as vulnerable to harvesting-induced extinction by the lUCN (Stewart and Wilson, 2005). Furthermore, the Convention on International Trade in Endangered Spe- cies of Wild Fauna and Flora (CITES) has placed the whale shark under Appendix II, which limits the trade of whale shark products among cooperating nations (Stewart and Wilson, 2005). Within the United States EEZ, the retention of whale sharks caught commercially or recreationally is prohibited (NMFS, 1993). However, because whale sharks are highly migratory and their movements cross numerous boundaries, detailed infor- mation on their abundance and seasonal distribution is needed to ensure their well being. Acknowledgments We thank W. Hoggard, S. O'Sullivan, and C. Roden for participating in the surveys and for their valuable suggestions. We also thank R. Avent, C. Gledhill, D. Hanisko, A. Jahncke, K. Mitchell, and K. Rademacher who assisted us in varying capacities. Comments by E. Clarke, M. Grace, T. Henwood, C. Hubard, C. Jones, K. Maze-Foley, S. Nichols, C. Roden, and the three anonymous reviewers significantly improved the man- uscript. The research was authorized under Marine Mammal Research permits 738 and 779-1339 issued to the SEFSC, and was supported by interagency agree- ments between the NMFS-SEFSC and U.S. Geological Survey, Biological Resources Division (no. 1445-IA09- 96-0009); and the NMFS-SEFSC and Minerals Manage- ment Service, Gulf of Mexico Region (nos. 15958 and 14-12-0001-30398). 584 Fishery Bulletin 104(4) Literature cited Baughman. J. L. 1947. Fishes not previously reported from Texas, with mis- cellaneous notes on other species. Copeia 1947:280. 1950. Randomnotes on Texas fishes. Tex. J. Sci. 2:117- 138. 1955. The oviparity of the whale shark, Rhincodon typus, with records of this and other fishes in Texas waters. Copeia 1955:54-55. Baughman, J. L., and S. Springer. 1950. Biological and economic notes on the sharks of the Gulf of Mexico with special reference to those of Texas and with a key for their identification. Am. Midi. Nat. 44:96-152. Breuer, J. P. 1954. The littlest biggest fish. Tex. Game Fish. 12:29. Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 1993. Distance sampling: estimating abundance of biologi- cal populations, 446 p. Chapman and Hall, London. Clark, E., and D. R. Nelson. 1997. Young whale sharks, Rhincodon typus, feeding on a copepod bloom near La Paz, Mexico. Environ. Biol. Fishes 50:63-73. Clark, E., and K. von Schmidt. 1965. Sharks of the central Gulf coast of Florida. Bull. Mar. Sci. 15:13-83. Colman, J. G. 1997. A review of the biology and ecology of the whale shark. J. Fish. Biol. 51(6):1219-1234. Compagno, L. J. V. 1984. Sharks of the world. FAO Fisheries Synopsis no. 125, part 4:210-211. Duffy, C. A. 2002. Distribution, seasonality, lengths, and feeding behavior of whale sharks {Rhincodon typus) observed in New Zealand waters. N.Z.J. Mar. Freshw. Res. 36:565-570. Eckert 8. A., and B. S. Stewart. 2001. Telemetry and satellite tracking of whale sharks, Rhincodon typus. in the Sea of Cortez, Mexico and the north Pacific Ocean. Environ. Biol. Fishes 60:299-308. Gudger, E. W. 1939. The whale shark in the Caribbean Sea and the Gulf of Mexico. Sci. Monthly 48:261-264. 1941. The food and feeding habits of the whale shark {Rhincodon typus). J. Elisha Mitchell Sci. Soc. 57:57-72. Gunn, J. S., J. D. Stevens, T. L. O. Davis, and B. M. Norman. 1999. Observations on the short-term movements and behaviour of whale sharks [Rhincodon typus) at Ning- aloo Reef, Western Australia. Mar. Biol. 135:553-559. Gunter, F.. and F. T. Knapp. 1951. Fishes, new, rare or seldom recorded from the Texas coast. Tex. J. Sci. 1:134-138. Heyman, W. D., R. T. Graham, B. Kjerve, and R. E. Johannes. 2001. Whale sharks Rhincodon typus aggregate to feed on fish spawn in Belize. Mar. Ecol. Prog. Ser. 215:275-282. Hoffman, W., T. H. Fritts, and R. P. Reynolds. 1981. Whale sharks associated with fish schools off south Texas. Northeast Gulf Sci. 5:55-57. Hoffmayer, E. R., J. S. Franks, and J. P. Shelley. 2005. Recent observations of the whale shark {Rhincodon typus) in the northcentral Gulf of Mexico. Gulf Caribb. Res. 17:117-120. Lohrenz, S. E., D. A. Wiesenburg, R. A. Arnone, and X. Chen. 1999. What controls primary production in the Gulf of Mexico? In The Gulf of Mexico large marine ecosys- tem (H. Kumpf. K. Steidinger, and K. Sherman, eds.), p. 151-170. Blackwell Science, Oxford, UK. Mullin, K. D., W. Hoggard, and L. J. Hansen. 2004. Abundance and seasonal occurrence of cetaceans in outer continental shelf and slope waters of the north- central and northwestern Gulf of Mexico. Gulf Mex. Sci. 22:62-73. Mullin, K., W. Hoggard, C. Roden. R. Lohoefener, C. Rogers, and B. Taggart. 1994. Cetaceans on the upper continental slope in the north-central Gulf of Mexico. Fish. Bull. 92:773- 786. NMFS (National Marine Fisheries Service). 1993. Fishery management plan for sharks of the Atlan- tic Ocean, 166 p. U.S. Dep. Commerce, Washington, D.C. Penrith, M. J. 1972. Earliest description and name for the whale shark. Copeia 1972:362. Rezak, R., S. R. Gittings, and T. J. Bright. 1990. Biotic assemblages and ecological controls on reefs and banks of the northwest Gulf of Mexico. Am. Zool. 30:23-35. Sahl, L. E., W. J. Merrell, and D. C. Biggs. 1993. The influence of advection on the spatial variabil- ity of nutrient concentrations on the Texas-Louisiana continental shelf. Cont. Shelf Res. 13:233-251. Sahl, L. E., D. A. Wiesenburg, and W. J. Merrell. 1997. Interactions of mesoscale features with Texas shelf and slope waters. Cont. Shelf Res. 17:117-136. Springer, S. 1957. Some observations on the behavior of schools of fishes in the Gulf of Mexico and adjacent waters. Ecol- ogy 38:166-171. Stewart, B. S., and S. G. Wilson. 2005. Threatened fishes of the world: Rhincodon typus (Smith 1828) (Rhincodontidae). Environ. Biol. Fishes 74: 184-185. Taylor, J. G. 1996. Seasonal occurrence, distribution and movements of the whale shark, Rhincodon typus, at Ningaloo Reef, Western Australia. Mar. Freshw. Res. 47:637-642. Wilson, S. G., J. G. Taylor, and A. F. Pearce. 2001. The seasonal aggregation of whale sharks at Nin- galoo Reef, Western Australia: currents, migrations and the El Nino/Southern Oscillation. Environ. Biol. Fishes 61:1-11. Zar, J. H. 1999. Biostatistical analysis, 620 p. Prentice-Hall, Inc. Englewood Cliffs. NJ. 585 Abstract — Much of the information available on the population status of a harvested fish species is obtained from landings data. When fishing restric- tions are in place, fishery-dependent data are reduced and assessments rely more heavily on fishery-indepen- dent data. Stock assessments of red porgy iPagrus pagrus) have shown a declining population and have led to a number of management measures, including a moratorium on fishing this species. To investigate how a lack of fishery-dependent data during a moratorium would affect stock assess- ment results for red porgy. we con- ducted simulations representing a range of periods of moratorium. As data were removed from the model, stock status indicators and projec- tions became increasingly variable. Projections estimated that a 12-year moratorium would be needed for stock rebuilding, but simulations showed that uncertainty surrounding stock assessment estimates would increase after three years without fishery- dependent data. Unless additional data are collected during periods of strict fishing regulations, it may be difficult to accurately assess the length of time needed for the stock to rebuild and to assess the popula- tion status. Effects of a simulated fishing moratorium on the stock assessment of red porgy iPagrus pagrus) Michelle L. Davis (contact author) Department of Fisheries and Wildlife Sciences Virginia Polytechnic Institute and State University 100 Cheatham Hall Blacksburg, Virginia 24061-0321 Email address: midavisl gvt.edu Jim Berkson National Marine Fisheries Service RTR Unit Virginia Polytechnic Institute and State University 100 Cheatham Hall Blacksburg, Virginia 24061-0321 Manuscript submitted 28 March 2005 to the Scientific Editor's Office. Manuscript approved for publication 19 January 2006 by the Scientific Editor. Fish. Bull. 104:585-592(2006). For commercial or recreational fisher- ies, a large portion of the information available for assessment of the popu- lation status of a species is obtained during harvesting operations. When harvest restrictions are in place, fish- ery-dependent data are reduced and assessments become more reliant on fishery-independent data. During a moratorium, harvest ceases completely and new methods may be required to assess the status of the stock. It is assumed that stock assessment results accurately portray the current popula- tion and that these results are used in policy decisions affecting future stock status. However, stock assessment results are not solely a product of the population being assessed, but are also a function of stock status, data sources, and sample size. Although most stock assessments include close evaluation of data sources, few assess- ments include the quantification of potential effects of changes in fish- ery-dependent data as a byproduct of changing harvest restrictions. We therefore investigated the effects of the reduced availability of fishery- dependent data on stock assessment results and management decisions. To accomplish this goal, we con- ducted model simulations for red porgy (Pagrus pagrus) that is found off the coast of the southeastern United States from North Carolina to Florida. This reef-associated fish is an important resource for commercial and recreational fisheries (Huntsman et al.. 1978; Low et al., 1985). Red porgy have an extensive native range and inhabit coastal waters on both sides of the Atlantic Ocean in both hemispheres (Pajeulo and Lorenzo, 1996; Labropoulou et al., 1999; Hood and Johnson, 2000). They are associ- ated with live-bottom reef habitats on rocky outcroppings and therefore have a patchy distribution off the south- eastern United States (Grimes et al.. 1982). Red porgy are protogynous hermaphrodites; females dominate the smaller size classes, and males occur at all ages (Manooch, 1976). Fe- males generally reach maturity from age-1 to age-2 at approximately 300 mm total length, and the majority of red porgy age-5 and older are mature males (Hood and Johnson, 2000). Red porgy have been harvested ex- tensively by three major fisheries (Fig. 1): commercial, recreational, and a fishery comprising large-scale charter boats called "headboats" (Huntsman et al., 1978). All three fisheries use mainly hook-and-line gear and target a wide variety of temperate reef fishes (Chester et al., 1984), not exclusively red porgy. In the early 1980s, the red porgy stock began showing signs of decline (Collins and Sedberry, 1991; Vaughan et al., 1992). As the stock decline became more evident, the minimum size limit was raised (Table 1) by the South Atlantic Fishery Man- agement Council (SAFMC), the group mandated with managing red porgy and other fish off the coast of North 586 Fishery Bulletin 104(4) 800 D Commercial Headboat D Recreational 200 1984 1988 Year Figure 1 Landings lin metric tons) of red porgy iPagrus pagriis) off the coast of the southeastern United States for the commercial, recreational, and headboat fishery sectors from 1972 through 2001. 5 1 4 ■ 1^ o J '■ CQ 1 ■ F/Fmsy A ~^ / ' ■ CV of F2000/FMSY 0 12 3 4 5 6 Simulated duration of moratorium (years) Figure 4 Coefficient of variation (standard deviation among 50 simu- lations for each moratorium duration divided by the meani for red porgy iPagrus pagrus) for B2oii/^M.sy ^"^d f 2000'^A/sy Fishery-dependent length and age data were removed from the stock assessment model runs for the number of years specified. 06. ^ 05 - 04 03 02 62001 /Sm 0.7 06- 05 04 03 02 01 ■ 00 B F2000/F„sv r*n 0 12 3 4 5 Simulated duration of moratorium (years) Figure 5 Mean estimates and 959; confidence intervals of (A) B.,,,,,/ B^igy and (B) ^2ooo^-^,usv f*"' ^0 model runs for each mora- torium duration for red porgy iPagrus pagrus). Fishery- dependent length and age data were removed from stock assessment model runs for the number of years specified. 590 Fishery Bulletin 104(4) In addition to the expected outcome of higher vari- ance as data were removed, we also found a slight ten- dency toward overestimating stock productivity that resulted from eliminating multiple years of data. When fishery-dependent length and age information were re- moved, the model tended to predict slightly higher bio- mass, lower fishing mortality, and faster population 075 0 1 2 3 4 5 6 Simulated duration of moratorium (years) Figure 6 Proportion of projected red porgy Pagrus pagriis biomass estimates for the year 2016 under Amendment 12 for which the biomass was greater than Bjjj,^. The mean likelihood for 50 model runs for each moratorium duration is shown (with 95% confidence intervals) and the range of the 50 model runs is shown in gray. For a management option to be recommended, there must be at least a 50% likelihood (horizontal line) of the biomass being above JB,,t,.y. Fishery-dependent length and age data were removed from stock assessment model runs for the number of years specified. 5 1 00 n 0-75 0.50 S 0.25 0.00 — Oyrs - - ■ .4yrs ~ - — 6yrs Figure 7 Stock projections (50 model runs for each moratorium duration) under Amendment 12 fishing mortality levels. Fishery-dependent length and age data were removed from stock assessment model runs for the number of years specified. For a management option to be recom- mended, there must be at least a 50% likelihood (horizontal line) of the population biomass being greater than B^g^ before the year 2016 (vertical line). recovery. This finding was also supported by Chen et al. (2003), for whom data removal led to a more opti- mistic estimation of the current fishery status and stock productivity. This bias would increase the potential for overharvesting because the population may not be as productive as the model results suggest. The results of our study are specific to the South Atlantic red porgy stock and may not be applicable to other stocks or species. Stock status indicators and productivity estimates are affected by life history parameters, population size, historical and current fishing pressure, and data availability and these will dif- fer among species. Additionally, use of an alternative stock assessment model could also affect results. However, if this bias is present for other species and stock assessment models, it could have serious implications for managers because it im- plies that the less information we have on a population, the more productive we would estimate (incorrectly) the popula- tion to be. In general, removing red porgy length and age data had only minor effects on status indicator estimates, projections, or management decisions. The signal of a declining red porgy population was so strong and relatively consistent among data sources that removing a portion of this information did not greatly af- fect results. All model runs showed an overfished population but one that is not currently undergoing overfishing, and management decisions were also rela- tively consistent. In all cases, projections showed that rebuilding to B^,^y by the year 2016 would occur with no fishing mortality or a moratorium, and fishing mortality representative of Amendment 9 would be insufficient to rebuild the stock. Differences in management decisions did exist for model runs conducted under Amendment 12 fishing mortality levels; some data-poor models incorrectly identi- fied Amendment 12 as a suitable option for rebuilding the stock. For a fish popula- tion closer to an overfished status or to a condition where overfishing is in progress, or where data sources were contradictory, data removal could affect stock assess- ment results and management decisions to a greater extent. Because red porgy is a well-studied species for which there are 30 years of information from multiple sources, we removed a relatively small proportion of the data used in the stock assessment for our model runs. Conducting these types of simulations on a lesser-studied spe- Davis and Berkson Effects of a simulated fistiing moratorium on tfie stock assessment of Pagrus pagrus 591 cies would most likely result in larger changes in sta- tus indicator estimates, projections, and management decisions because a larger proportion of information would be removed. The results of our study on red porgy identified the importance of investigating the potential effects of reducing fishery-dependent information on stock assessment results. If used before implementing management actions, this kind of data simulation study could estimate the potential impact of a lack of data before harvest restrictions are put in place that would cause data loss. Management implications These types of simulation studies aid in understanding the impacts of data quantity on stock assessments, par- ticularly with respect to uncertainty and variance. By incorporating uncertainty due to data loss into manage- ment decisions, managers will be better able to assess and manage these populations. For example, SEDAR population projections estimated that red porgy required a 12-year moratorium for the population to rebuild to ^MSY (^Fig. 3). However, our data simulations showed that variability surrounding status indicators increased when as few as three years of fishery-dependent data were removed (Fig. 4) and that a tendency toward overesti- mating productivity also increased with duration of the moratorium (Fig. 7). Therefore, if managers implemented a 12-year moratorium for red porgy (with no additional sampling), it is possible that a stock assessment con- ducted at the end of the moratorium would be unable to accurately estimate stock status and identify whether or not the population biomass had reached B(/j,y. This study highlights the importance of additional sampling during periods of strict regulations to compen- sate for the fishery-dependent data that are lost (Olney and Hoenig, 2001). Because many reef fish species are caught in the same fisheries, analysis of bycatch through a fisheries observer program would yield valuable data for a number of species. Another option for data collec- tion would be to initiate a "test fishery" after a mora- torium in order to collect fishery-dependent data in a relatively short time and then use these data to estimate stock status. A third option would be to use additional sampling during a moratorium from fishery-independent sources. Although MARMAP data are currently insuf- ficient for assessments of many reef fishes, an expanded survey (i.e., into deeper water, more gear types, etc.) could prove to be valuable in future assessments. Even without a moratorium in place, assessments of many species would undoubtedly benefit from information col- lected by a fisheries observer program or an expanded fishery-independent survey. Data collected from these sources before a period of strict regulation would im- prove the ability of scientists to detect trends because a longer time series would make population recovery more apparent than that seen without such data. The data quantity simulations conducted in our study could also be used for other species. This method al- lows managers to simulate the effects of a lack of data before enacting regulations or sampling changes that cause data loss. If a loss of information is shown to affect stock assessment results, simulations can be conducted to identify whether additional data will be needed and to test the effectiveness of additional sam- pling before implementing a sampling program. Stock assessments frequently use population projections to predict the effects of a policy on the stock. Our meth- odology is an expansion of this idea, using simulations to predict how a policy affects our ability to identify changes in the stock. By incorporating uncertainty due to data loss into management decisions, managers will be better able to assess and manage these populations and improve sampling to better fulfill future research needs. Acknowledgments This research was funded by the South Atlantic Fishery Management Council. We thank E. Williams, M. Prager, D. Vaughan, and K. Shertzer for providing modeling advice, J. Potts and P. Harris for information, J. Ney and R. Neves for guidance, and the many state and federal biologists and SAFMC staff members that assisted us in this project. This manuscript was improved by the com- ments and suggestions of M. Prager and two anonymous reviewers. We greatly appreciate the time and effort of all involved. Literature cited Chen. Y., L. Chen, and K. I. Stergiou. 2003. Impacts of data quantity on fisheries stock assessment. Aquat. Sci. 65:92-98. Chester, A. J., G. R. Huntsman, P. A. Tester, and C. S. Manooch. 1984. South Atlantic Bight reef fish communities as represented in hook-and-line catches. Bull. Mar. Sci. 34:267-279. Collins, M. R. 1990. A comparison of three fish trap designs. Fish. Res. 9:325-332. Collins, M. R., and G. R. Sedberry. 1991. Status of vermilion snapper and red porgy stocks off South Carolina. Trans. Am. Fish. See. 120:116-120. Grimes, C. B., C. S. Manooch, and G. R. Huntsman. 1982. Reef and rock outcropping fishes of the outer con- tinental shelf of North Carolina and South Carolina, and ecological notes on the red porgy and vermilion snapper. Bull. Mar. Sci. 32:277-289. 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 A. K. Johnson. 2000. Age, growth, mortality, and reproduction of red porgy, Pagrus pagrus, from the eastern Gulf of Mexico. Fish. Bull. 98:723-735. Huntsman, G. R., D. R. Colby, and R. L. Dixon. 1978. Measuring catches in the Carolina headboat fishery. Trans. Am. Fish. Soc. 107:241-245. 592 Fishery Bulletin 104(4) Labropoulou, M., A. Machias, and N. Tsimenides. 1999. Habitat selection and diet of juvenile red porgy, Pagrus pagriis (Linnaeus, 1758). Fish. Bull. 97:495- 507. Low, R. A., G. F. Ulrich, C. A. Barans, and D. A. Oakley. 1985. Analysis of catch per unit of effort and length composition in the South Carolina commercial han- dline fishery, 1976-1982. N. Am. J. Fish. Manag. 5:340-363. Manooch, C. S. 1976. Reproductive cycle, fecundity, and sex ratios of the red porgy, Pagrus pagrus (Pisces: Sparidae) in North Carolina. Fish. Bull. 74:775-781. Methot, R. M. 1989. Synthetic estimates of historical abundance and mortality for northern anchovy. Am. Fish. Soc. Symp. 6:66-82. Olney, J. E., and J. M. Hoenig. 2001. Managing a fishery under moratorium: assess- ment opportunities for Virginia's stocks of American shad. Fisheries 26:6-12. Pajeulo, J. G., and J. M. Lorenzo. 1996. Life history of the red porgy Pagrus pagrus (Tele- ostei: Sparidae) off the Canary Islands, central east Atlantic. Fish. Res. 28:163-177. Vaughan, D. S., G. R. Huntsman, C. S. Manooch. F. C. Rohde, and G. F. Ulrich. 1992. Population characteristics of the red porgy, Pagrus pagrus. stock off the Carolinas. Bull. Mar. Sci. 50:1-20. Vaughan, D. S., and M. H. Prager. 2002. Severe decline in abundance of the red porgy iPagrus pagrus) population off the southeastern LInited States. Fish. Bull. 100:351-375. Williams, E. H. 2001. Assessment of cobia, Rachycentron canadum, in the waters of the U.S. Gulf of Mexico. NOAA Technical Memo. NMFS-SEFSC-469, 54 p. Center for Coastal Fisheries and Habitat Research, 101 Pivers Island Rd., Beaufort, NC 28516. 593 Abstract — The goal of this study was to provide insight into habitat use of the inner continental shelf off southern New Jersey by summer- spawned bluefish {Pomatomus salta- trix). Throughout the August-October 1998 sampling period, a total of 1071 bluefish were collected from shelf surface waters, during which the mean density and body size was 3.7 bluefish/1000 m^ and 13.7 mm standard length (SL), respectively. Spatiotemporal variability in bluefish density was explained by an inverse relationship with Secchi depth, and body size was explained by water tem- perature and depth. Bluefish size- structure in August was bimodal and comprised larval, transitional, and juvenile stages (3-48 mm SL). Size frequencies in subsequent months were unimodal and consisted of blue- fish <25 mm SL. Synoptic sampling of multiple habitats indicated that the earliest life stages of bluefish extensively, and perhaps exclusively, use inner continental shelf surface waters. Summer-spawned bluefish numerically dominated the popula- tion across all habitats and temporal scales examined. Habitat use of the inner continental shelf off southern New Jersey by summer-spawned bluefish (Pomatomus saltatrix)* David L. Taylor (contact author) Rutgers University Institute of Marine and Coastal Sciences Marine Field Station, 800 c/o 132 Great Bay Boulevard Tuckerton, New Jersey 08087-2004 and Roger Williams University Department of Biology and Marine Biology One Old Ferry Road Bristol, Rhode Island 02809 Email; dtaylon&rwu.edu Peter M. Rowe Kenneth W. Able Rutgers University Institute of Marine and Coastal Sciences Marine Field Station, 800 c/o 132 Great Bay Boulevard Tuckerton, New Jersey 08087-2004 Manuscript submitted 14 April 2005 to the Scientific Editor's Office. Manuscript approved for publication 6 March 2006 by the Scientific Editor. Fish. Bull 104:59.3-604(2006). Bluefish [Pomatomus saltatrix) is a coastal marine species that is common in temperate and subtropi- cal waters worldwide (Juanes et al., 1996), Bluefish support extensive fisheries throughout their distribu- tion (Juanes et al., 1996), including the western north Atlantic where this species has historically accounted for the greatest catch by weight in the recreational fishery (Pottern et al., 1989). Along the eastern coast of the United States, for example, bluefish landings in 1985 contributed to over 24% of the total marine recreational catch (U.S. Department of Commerce, 1986). Since the mid-1980s, however, bluefish landings in this area have declined precipitously and have yet to rebound in the last two decades (Klein-MacPhee, 2002). As a result, the Atlantic States Marine Fisher- ies Commission (ASMFC) established bluefish as a priority research sub- ject and further recommended that research provide a better under- standing of the early life history and recruitment patterns of this species. Recruitment dynamics of bluefish along the eastern coast of the United States are intrinsically linked to an- nual spawning migrations of the adult population. Bluefish spawning migra- tions move northward into the Middle Atlantic Bight (MAB; Cape Hatteras, North Carolina, to Cape Cod, Massa- chusetts) during the spring and sum- mer following an overwintering peri- od in the South Atlantic Bight (SAB; Cape Canaveral, Florida to Cape Hat- teras) (Klein-MacPhee, 2002) or at the edge of the continental shelf in the MAB (Miller, 1969; Wilk. 1982; Shepherd et al., in press). Pelagic eggs are spawned offshore during the spring and summer as bluefish migrate northward along the conti- nental shelf. The planktonic eggs of bluefish experience a brief incubation period (48 hours; Deuel et al., 1966) that is followed by a larval stage that concludes 18-25 days after hatching and once the bluefish are 10-12 mm standard length (SL) (Hare and Cow- en, 1994). Bluefish then undergo a transitional period, ending with the * Contribution 2006-2 of the Institute of Marine and Coastal Sciences, Rutgers University, Tuckerton, NJ 08087. 594 Fishery Bulletin 104(4) onset of the juvenile stage at 34-37 mm SL (Silverman, 1975). Juveniles remain oceanic for an additional 15-45 days before passively or actively recruiting to estuarine habitats at 40-80 mm fork length (FL) (Nyman and Conover, 1988; McBride and Conover, 1991; Hare and Cowen, 1993). Along the U.S. northeastern coast, juvenile bluefish recruit to estuaries and have a consistent intra-annual and bimodal length-frequency distribution: the first cohort enters estuaries from late May to early June (spring-spawned cohort), and a second cohort enters from July to October (summer-spawned cohort) (Nyman and Conover, 1988; McBride and Conover, 1991; Mc- Bride et al., 1993; Creaser and Perkins, 1994). This is consistent with previous studies that indicate that juve- niles of both spring and summer cohorts are estuarine dependent (Juanes et al., 1996; Able and Fahay. 1998). There is recent evidence, however, that select cohorts of bluefish may exclusively use alternative habitats, such as ocean beaches, during the juvenile stage (Able et al., 2003; Wilber et al., 2003), In addition, samples of larval and juvenile bluefish from surface waters on the inner continental shelf have indicated at least two cohorts of presumably summer-spawned fish that may not recruit to estuaries (Kendall and Walford, 1979; Rowe et al.'; Taylor and Able, in press). Spring-spawned bluefish historically dominate juve- nile year-class strength and were therefore assumed to determine adult population dynamics (Nyman and Conover, 1988; McBride and Conover, 1991; Munch and Conover, 2000). This assertion, however, is based pri- marily on estimates of juvenile abundance in MAB estuaries (Nyman and Conover, 1988; McBride and Conover, 1991) and failed to consider contributions from alternative habitats, including the inner continental shelf. A more recent evaluation of bluefish collected from the MAB also has indicated a shift in the cohort- specific production of juveniles over the last decade (Conover et al., 2003), such that summer-spawned fish numerically dominate recent year classes. Despite the apparent switch in the relative dominance of the two co- horts, bluefish successfully recruiting to the adult popu- lation remain the products of spring-spawning events (Chiarella and Conover, 1990; Conover et al., 2003). The factors underlying the apparent failure of the summer- spawned cohort to contribute substantially to the adult population are unknown. This failure is due, in part, to insufficient information that would otherwise enable a critical evaluation of each cohort's role in regulating year-class strength. Attempts to determine factors af- fecting recruitment patterns should include an evalua- 1 Rowe, P. M., K. W. Able, and M. J. Miller. 2002. Distri- bution, abundance, and size of young-of-the-year bluefish (Pomatomus saltatrix) in ocean and estuarine habitats in southern New Jersey during 1999-2000, .54 p. Jacques Cousteau Technical Report no. 100-16. Rutgers Univer- sity, Institute of Marine and Coastal Sciences, Marine Field Station, 800 c/o 132 Great Bay Boulevard, Tuckerton, NJ 08087-2004. tion of the cohort dynamics of summer-spawned bluefish that utilize inner continental shelf habitats. The goal of our study was to provide greater insight into the habitat use by summer-spawned bluefish in a localized area in the MAB. In 1998, young-of-the-year (YOY) bluefish were sampled during the summer and fall on the inner continental shelf off southern New Jer- sey. Spatial and temporal abundance and distribution patterns were used to evaluate the potential importance of continental shelf waters as habitats for YOY bluefish. Observations from the 1998 study were then analyzed and compared to observations from other field surveys of habitat use by YOY bluefish in the same geographic region and time: inshore and other continental shelf waters of southern New Jersey in 1998 (Able et al., 2003). Material and methods Sampling of surface waters over the inner continental shelf Bluefish were sampled at the water surface along the inner continental shelf off the southern coast of New Jersey (Fig. 1). Specifically, a 4x4 station grid (total area -825 km-) was sampled on ten dates between 11 August and 9 October 1998. Stations were grouped into four transects that were aligned parallel to the coastline (Fig. lA). Transects closest and farthest from the coast- line were 3 km and 18 km offshore, respectively. Water depths ranged between 6 and 29 m, and the shallowest depths generally occurred near the shore. Bluefish were collected during daylight with a Methot trawl deployed at the surface (Methot, 1986; Oozeki et al., 2004). The frame trawl had a mouth area of 5 m^, leading to an 11-m-long net (6-mm mesh) that tapered to a 2-m-long plankton net (0.5-m diameter, SOO-fim mesh). One tow was performed at each of the 16 sta- tions per sampling date, with the exception of 11 Au- gust and 9 October when only four and eight stations were sampled, respectively (7; = 140 total tows). At each station, the trawl was towed for four minutes in an arc (tow speed -2.5-3.5 knots) to avoid the wake of the re- search vessel. All tows were made at the surface (0-2 m depth) with the top of the frame approximately 0.25 m out of the water. Samples were immediately processed after collection, and bluefish were preserved in 95% ethanol for subsequent laboratory analysis. A General Oceanics flowmeter (General Oceanics, Inc., Miami, FL) was attached to the Methot trawl frame along a lower corner so that the volume of water sampled could be determined (-2000 m* per tow). Sur- face salinity, surface water temperature, water depth, Secchi depth, and wind speed and direction were record- ed at all stations. Additional wind data were available as a time series (15-minute sampling interval) from a meteorological platform located at the Rutgers Univer- sity Marine Field Station (Tuckerton, NJ; Fig. lA; K. W. Able, unpubl. data). Taylor et al ; Habilat use by summer-spawned Pomatomus saltolrix on the inner continental shelf off southern New Jersey 595 Figure 1 Map of inshore and inner continental shelf habitats off southern and central New Jersey. Plus symbols ' + i demarcate the 4x4 station grid (total area -825 km-) sampled with a surface- deployed Methot trawl on the inner continental shelf between 11 August and 9 October 1998. lAi Inner continental shelf sampling sites were grouped into four transects (I-IVi aligned parallel to the coastline. Estuarine (•), inlet (O), and ocean beach iD) sites sampled with a beach seine were located at the margins of Great Bay (GBi, Little Egg Harbor iLEHi, and Little Egg Inlet (LEI). (B) Solid lines (^) demarcate six depth strata sampled with an otter trawl on the inner continental shelf. Location of the Rutgers University Marine Field Station (Tuckerton, NJ; Bi and depth contours of 10, 20, and 30 m are provided. Bluefish collected with Methot trawls were counted to estimate fish density (no. of bluefish/1000 m-M, mea- sured to the nearest mm SL, and classified into distinct ontogenetic stages based on the individuals body size and presumed scale development (Silverman, 1975; Able and Lamonaca, in press). For the purposes of our inves- tigation, several life history stages were demarcated: larval, transitional, and juvenile. The onset of scale development was recognized as the end of the larval stage and the beginning of the transitional period (12 mm SL; Silverman, 1975; Hare and Cowen, 1994). The transitional period in bluefish development, in turn, concluded with complete scale formation and the juve- nile stage began at 34 mm SL (Silverman, 1975). Spatial and temporal patterns of bluefish density and mean size were analyzed independently with multivari- ate repeated-measures analysis of variance (ANOVA) models by using station transects (transects I-IV; Fig. lA) and sampling date (day of year) as the between- subject and within-subject factors, respectively. Profile transformation of the within-subject factor was used in the repeated-measures ANOVA model, and statisti- cal significance was estimated from the Greenhouse- Geisser-adjusted probability to avoid violating the as- sumption of circularity (sphericity) of the within-subject variance-covariance matrix (von Ende, 1993), The mean density and body size of bluefish across four levels of transects and ten levels of sampling dates were con- trasted with a Ryan-Einot-Gabriel-Welsch (Ryan's Q) multiple comparison test (Day and Quinn, 1989). More- over, natural log (x-fl) and natural log (x) transforma- tions were performed on bluefish density and mean-size data, respectively, to meet assumptions of normality and homogeneity of variance. When data transforma- tions did not achieve homoscedasticity, hypotheses were rejected at alpha values lower than the P-values of Levene's test for homogeneity of variance (Underwood, 1981). The effects of several environmental parameters on the spatial and temporal distribution of bluefish density and mean size were analyzed with a stepwise multiple regression. The variables included in the regression 596 Fishery Bulletin 104(4) Table 1 Summary of field sampling protocols habitat chai acteristics, and bluefish iPoniafonuis saltatnx) catch and size data across mul- tiple areas of southern and central New Jersey in the summer and fall of 1998. Bluefish catch data are reported as catch per unit of effort (CPUE). Temperature, salir ity, and size data with in parentheses repr esent minimum and maximum values observed on a given date, and NM and NA signify "no measurement" and "not applicable, ' respectively. Additional information regarding | habitat location and sampling protocols is found in the text and Figure 1. Sampling Sampling effort Depth Temperature Salinity CPUE Size Habitat gear (no. of tows (ml rC) C/rf) ( no. per tow) (mm SL) Continental shelf (surface v/aters) 11-20 Aug 1998 Methot trawl 68 6-28 25.5(16.5-27.31 30.7(30.0-31.1) 4.4 18.7(3-48) 1-11 Sep 1998 " 32 6-29 24.0(21.8-25.5) 30.6(30.3-30.7) 2.3 10.1(5-25) 18-25 Sep 1998 " 32 6-27 22.5(21.0-23.61 30.7(30.3-31.0) 21.7 12.0(4-24) 6 Oct 1998 " 8 9-29 19.5(18.7-20.3) 30.7(30.6-30.91 0.6 18.7(18-221 Estuary 13-24 Aug 1998 Beach seine 8 0-2 25.0(24.0-26.0) 28.5(28.0-29.0) 0.3 116.0(60-172) 9 Sep 1998 " 5 " 19.9(19.5-20.5) 28.0(28.0-28.0) 0.4 111.0(104-1181 21 Sep 1998 " 5 " 27.0(27.0-27.0) NM 0.2 116.0 5 Oct-3 Nov 1998 " 15 " 16.0(10.5-19.0) 27.8(26.0-29.0) 0.1 147.0 Inlet 4-21 Aug 1998 Beach seine 9 0-2 25.7(25.0-26.0) 29.3(28.0-30.0) 3.3 83.1(59-1031 3-4 Sep 1998 5 " 25.6(25.0-26.0) 28.8(28.0-30.0) 4,0 78.9(55-113) 14-29 Sep 1998 10 " 21.7(20.0-23.0) 29.7(29.0-30.0) 0.5 96.2(91-991 13-29 Sep 1998 " 12 " 15,0(13.0-17.0) 29.3(29.0-30.01 0.0 NA Ocean beach 3-25 Aug 1998 Beach seine 32 0-2 25.4(23.0-27.0) 30.1(28.0-31.0) 14.1 73.3(51-168) 2-8 Sep 1998 " 16 " 24.4(23.0-26.0) 30.4(30.0-31.0) 53.7 76.5(53-123) 14-30 Sep 1998 " 26 " 22.6(20.0-26.0) 30.1(29.5-31.01 50.2 96.6(27-205) 6 Oct-4 Nov 1998 " 41 " 15.6(13.0-18.0) 30.0(30.0-30.0) 93.2 118.5(32-225) Continental shelf (bottom) 17-20 Aug 1998 Otter trawl 17 6-18 19.6(12.3-25.4) 30.9(28.3-31.5) 10.5 63.8(30-200) 30 Oct-4 Nov 1998 17 14.4(12.6-15.4) 31.4(30.1-31.9) 21.6 125.6(80-2201 model were sampling date (day of year), sampling time of day (sine-transformed), station distance from coast- line (km), surface temperature (°C), surface salinity (%o), Secchi depth (m; a coefficient of light extinction or water clarity), water depth (m), and cross- and long- shelf wind speed (positive east and north, respectively). The significance level for entry into the regression mod- el was set at P<0.05. Environmental variables that sig- nificantly affected the spatial and temporal distribution of bluefish density and mean size (i.e., those parameters incorporated into regression models) were also analyzed with an analysis of covariance (ANCOVA) model, with sampling date as the covariate and station transects as the discrete explanatory variable. Synoptic analysis of habitat use by summer-spawned bluefish Several other sampling programs initiated in 1998 moni- tored the distribution and size composition of YOY blue- fish across multiple habitats in southern and central New Jersey (Table 1, Fig. 1) and the data from these programs were then compared to our data from surface waters on the inner continental shelf. A complete description of the sampling protocols and schedules for monitoring YOY bluefish in these habitats is provided elsewhere (Able et al., 2003; Rowe et al.'). Habitat use by different onto- genetic stages was ascertained from the size and stage (larval, transitional, juvenile) composition of bluefish observed in estuaries, inlets, ocean beaches, and the inner continental shelf. Accordingly, length-frequency histograms of bluefish were created for each habitat over four time periods (August, early to mid-September, mid- to-late September, and from October to November) by using class intervals of 5.0 mm SL for Methot trawl and beach seine samples, and 10.0 mm SL intervals for otter trawl samples. Moreover, a length-based key derived by McBride and Conover (1991) and adopted by others (Juanes and Conover, 1995; Wilber et al., 2003) was used to delineate between spring- and summer-spawned Taylor el al Habitat use by summer spawned Pomatomus saltatnx on the inner continental shelf off southern New Jersey 597 bluefish in length-frequency histograms. Summer-spawned bluefish were defined as <100 mm SL in August, <125 mm SL from early to mid-September, <150 mm SL from mid-to-late September, <175 mm SL from October to November. Bluefish at body sizes larger then these groupings were designated as spring-spawned individuals. Results Bluefish density and mean size on the Inner continental shelf A total of 1071 bluefish were col- lected from surface waters off the southern coast of New Jersey during the survey. Throughout the August-October sampling period, bluefish were found in 71'7f ofthetows(99outof 140), during which the mean den- sity was 3.7 bluefish/1000 m^ (range = 0-57.9 bluefish/1000 m^). Bluefish density differed significantly by transect and sampling date (multivariate repeated-measures ANOVA; transect: F=9.39, df=3,12. P<0.005; date: F=16.09, df=9,88, P<0.0001). However, the tran- sect-sampling date interac- tion effect was significant and precluded direct conclusions about the main effects (mul- tivariate repeated-measures ANOVA; transect X date: F= 1.95, df=27,88, P<0.05) (Fig. 2). The interaction effect was attributed to significantly higher densities of bluefish at transect II on days 226 and 268 (in comparison to transects I, IV, and transect I on these days, respectively ( Ryan's Q multiple comparison test). Bluefish den- sities were also significantly greater at transect III on days 226 and 244 than at transect I for both dates (Ryan's Q mul- tiple comparison test). In the stepwise multiple regression, Secchi depth ex- plained the most variability in bluefish density with an r^ of 0.078 (stepwise multiple 8/11/98 day 223 N 8/12/98 day 224 8/1 4/98 ^ /7 O day 226 - bo°o ■ O o eft ° O X o 8/1 8/98 O day 230 O O '^ 8/20/98 day 232 ,, 9/25/98 ^ O^ o o o / 10/9/98 day 282 1/ Density (no. /1 000 m^) X 0 O 0.1-0.9 X O X X X O 10-5.9 Q 6,0-30.9 X C^ 31.0-60 0 T Figure 2 Spatial distribution of the density (no. offish/1000 m-^) of bluefish (Pomatomus sallatrix) collected with a Methot trawl between 11 August and 9 October 1998 off the southern coast of New Jersey. Stations were located at approximately .39 27'N, 74 ll'W. 598 Fishery Bulletin 104(4) Transect I II III IV Transect O 230 250 270 Day of year Day of year Figure 3 Spatial and temporal distribution (A and B) of Secchi depth (m), (C) water depth (m), and (D) surface water temperature (°C). Mean (±1 standard deviation) Secchi depth and water depth are displayed as a function of sta- tion transects (I-IV; Fig. lA), calculated from ten sampling dates between 11 August and 9 October 1998 (days 223-282). Means with the same upper case letters are not significantly different (Ryan's Q multiple comparison test). Secchi depth and surface water temperature are also displayed as function of station transects and sampling date (days 223-224). regression; F=11.34, df=l,135, P<0.001), and remain- ing parameters were not significant at P<0.05 and therefore not included in the model. The estimated coefficient for Secchi depth was negative (parameter estimate = -0.436), indicating an inverse relationship between bluefish density and water clarity. Before ana- lyzing spatial and temporal variations in Secchi depths with an ANCOVA model, we examined the interaction effect between transect and sampling date. There was no significant transect-date interaction effect (two-way ANOVA; transect X date: F=2.47, df=3,124, P=0.065), and the assumption of equal slopes was met in the full data set (Underwood, 1981). Secchi depth differed significantly as a function of transect position (AN- COVA; F=63.34, df=3,127, P<0.0001), whereby distance from the coastline was positively correlated with Secchi depth (Fig. 3A). Furthermore, Secchi depth significantly decreased throughout the sampling period (ANCOVA; F=30.20, df=l,127, P<0.0001) (Fig. 3B). The mean body size of bluefish collected from Au- gust to October was 13.7 mm SL (range = 3-48 mm SL) (Table 1; Fig. 4). Similar to spatial and temporal density patterns, bluefish mean body size differed sig- nificantly by transect and sampling date (multivariate repeated-measures ANOVA; transect: F=6.16. df=3,12, P<0.01; date: F=7.50, df=9,50, P<0.0001) (Fig. 4). The transect-sampling date interaction effect was again significant (multivariate repeated-measures ANOVA; transect X date: F=3.50, df=24,50, P<0.0001), thereby precluding contrasts across main effects. The interac- tion effect was caused by significantly larger bluefish at transects I and II on day 254 compared to transects III, IV, and transect III, respectively (Ryan's Q mul- tiple comparison test). On day 268, significantly larger bluefish were collected at transect IV than at all other transects (Ryan's Q multiple comparison test). Surface water temperature and water depth were the most significant factors affecting the size distribution of bluefish (stepwise multiple regression; temperature: F=14.07, df=2,94, P<0.0005; depth: P=5.50, df=2,94, P< 0.05), such that temperature and depth accounted for a partial r- of 0.129 and 0.048, respectively (cumula- tive 7-2=0.177). Moreover, estimated coefficients for both variables were negative (parameter estimates = -1.121 Taylor et a\ Habitat use by summer spawned Pomolomus sa/lalrix on the inner continental shelf off southern New Jersey 599 8/1 1/98 day 223 o N 8/14/98 O^ day 226 On O o o O X o 8/20/98 O^ day 232 X O X 10/9/98 day 282 Body size (mm SL) X no fish O 1.0-5,9 O 6-0-12.9 Q 130-33.9 o 34.0-450 Figure 4 Spatial distribution of the mean size (mm SLi of biuefish (Pomatomus saltatrix) collected with a Methot trawl between 11 August and 9 October 1998 off the southern coast of New Jersey. Stations are located at approximately 39°27'N, 74 ll'W. 600 Fishery Bulletin 104(4) and -0.254 for temperature and depth, respectively) indicating that larger bluefish were associated with relatively cool surface water and shallow water depths. Before analyzing spatial and temporal variations in temperature and depth with an ANCOVA model, we examined the transect-sampling date interaction ef- fect. There were no significant interactions between 0-6 n 0.6 0 4 ■ 5 02 tj u 0) Q. « 0,0 a 0.8 jr o 6 0.6 c (D D a- 2 0.4 LL 0,2 00 0.8 0.6 0.4 August □ Shell (surface Methot trawl) I Shelf (bottom otter trawl) n Ocean beach (seine) I Estuary and inlet (seme) I I I I I I I I nW I I I I I I I I 1 30 60 90 120 150 180 210 240 Early to mid-September jyL,,,,jib^, 30 60 90 120 150 180 210 240 Mid to late September flllU Tf^ITfTlllirTITTIlltl 30 60 90 120 150 180 210 240 October to November 0 30 60 90 120 150 180 210 240 Bluefish size (mm SL) Figure S Composite length-frequency histograms for bluefish iPomatomus saltatrix) collected from multiple habitats in summer and early fall ( 1998 ) in southern and central New Jersey. Additional information regarding habitat location and sampling protocols is found in the text. Table 1, and Figure 1. transect and date for both analyses (two-way ANOVA; depth-transect X date: F=0.72, df= 3,132, P=0.5401; temperature-transect X date: F=1.87, df=3,132, P= 0.1385). Water depth significantly increased as dis- tance from the coastline increased (ANCOVA; F=231.24, df=3,135, P<0.0001) (Fig. 3C), and surface temperature significantly decreased during the sampling period (AN- COVA; F=916.49, df=l,135, P<0.0001) (Fig. 3D). Transitional bluefish (25-34 mm SL) and juveniles (35-50 mm SL) were a major portion of the overall size composition of bluefish collected in August (Fig. 5). For example, bluefish 25-50 mm SL constituted 40.1% of the individuals collected during this time period. In subsequent months, however, these bluefish were absent from remaining survey tows. Conversely, larval bluefish (3-11 mm SL) and small transitional bluefish (12-24 mm SL) were collected on all sampling dates and were numerically dominant in August and thereafter (Fig. 5). From August to late September, length-frequency dis- tributions of bluefish <25 mm SL generally broadened in range and shifted to larger body sizes. Synoptic analysis of habitat use by summer-spawned bluefish Bluefish abundance and size-structure were compared among diverse inshore and coastal habitats in an effort to provide a more synoptic examination of summer habitat use across ontogenetic stages. In the process, bluefish were examined from shallow estuarine, inlet, and ocean beaches to surface and deeper waters on the inner continental shelf. As previously discussed, all individuals collected by Methot trawl tows from shelf surface waters were <50 mm SL, and therefore defined as summer-spawned larval, transitional, or juvenile bluefish (Fig. 5). Bluefish catches in the Great Bay estuary were low; 33 seine hauls collecting five bluefish from August to October (0.1-0.4 bluefish/haul) (Table 1). With the exception of one bluefish, all individuals in the estu- ary were designated as summer-spawned fish (Fig. 5). Relatively few bluefish were collected in seine hauls at inlet sites located at the margin of Little Egg Inlet (55 bluefish from 36 hauls) that were primarily sum- mer-spawned (Table 1; Fig. 5). In contrast to estuarine and inlet sites, substantial numbers of bluefish were collected from August to early November along ocean beaches north and south of Little Egg Inlet (6440 blue- fish from 115 hauls) (Table 1). Spring-spawned bluefish were consistently collected along ocean beaches dur- ing the survey, but abundances were low compared to those of summer-spawned individuals that represented 98.4% of the total catch (Fig. 5). Compared to the sizes of bluefish collected from surface waters of the inner continental shelf, the sizes of the overwhelming major- ity (99.4%) of fish collected at inshore sites (estuary, inlet, and ocean beaches) were >50 mm SL (i.e., the sizes of juvenile fish). The only exception to this pattern occurred along ocean beaches from mid-September to early November (Fig. 5). Taylor et al Habitat use by summer-spawned Pomatomus saltatrix on the inner continental shelf off southern New Jersey 601 Sampling of bottom waters on the inner continental shelf during the same time period collected 179 and 367 bluefish in August and October-November, respectively (Table 1; Fig. 5). The size composition of bluefish col- lected from otter trawl tows was comparable to that observed at inshore sites (particularly ocean beaches), but distinctly different from that of fish captured in continental shelf surface waters. For example, 96.5% of the bluefish captured during sampling of shelf bottom waters were >50 mm SL (mean size = 63.8 and 125.6 mm SL in August and October-November, respectively). Of the remaining individuals that were <50 mm SL. all were collected during the August otter trawl sur- vey. Spring-spawned bluefish were collected in August and October-November by otter trawl tows, but again these bluefish represented a small portion of the total sampled population (5.1%). Discussion In our study, summer-spawned larval, transitional, and juvenile bluefish used the inner continental shelf exten- sively and represented a major portion of the entire fish assemblage collected in surface waters off the southern coast of New Jersey. For example, summer-spawned bluefish were the most frequently encountered species within the surface fish assemblage and ranked fourth in numerical dominance during the summer and early fall. Moreover, patterns of bluefish density observed in this study are consistent with those seen in other ichthyoplankton surveys in the MAB. Kendall and Wal- ford (1979) found that bluefish larvae <4 mm SL were widely distributed in the MAB in August (Long Island, New York, to Virginia) and that peak concentrations occurred on the inner continental shelf in the vicinity of New Jersey and Delaware. Similarly, Smith et al. (1994) observed high densities of bluefish eggs in June and July and that centers of abundance occurred in mid-shelf waters off Delaware Bay and New Jersey (Berrien and Sibunka, 1999). Subsequent months were character- ized by relatively broad spatial distributions of bluefish larvae from the central MAB to southern New England and by dense concentrations off New Jersey in July and August, followed by decreased abundance over the entire continental shelf in October (Kendall and Walford, 1979; Smith et al., 1994). These observations from the inner continental shelf, in conjunction with the reported use of inshore habitats by summer-spawned juveniles (McBride and Conover, 1991; Able et al., 2003; Wilber et al, 2003), indicate that coastal regions of New Jersey and adjacent areas are important summer-spawning sites for bluefish and, moreover, represent an appropriate location for synoptic comparisons of bluefish abundance and size- structure across habitats. Empirical observations indicate that summer-spawned bluefish may use inner continental shelf habitats and, to a lesser extent, ocean beaches during their earliest life history stages (larval, transitional, and small ju- veniles). As surmised from this and previous studies. bluefish <50 mm SL are prominent constituents of the ichthyoplankton assemblage in the MAB (Kendall and Walford, 1979; Hare et al., 2001), but the these life stages are found rarely in estuaries and inlets. Field collections from our study indicated an abundant sup- ply of larval to small juvenile bluefish in the vicinity of Little Egg Inlet from August to October, yet these particular life stages were not observed during concur- rent sampling at estuarine sites. Moreover, larval fish assemblages were monitored over 16 years (1989-2004) inside Little Egg Inlet (Fig. 1), and bluefish were found in only 0.8% of the total 640 plankton tows performed between July and September (25 total bluefish, size range: 7.6-57.0 mm SL) (Witting et al., 1999; K. W. Able, unpubl. data). In another study, juvenile bluefish were conspicuous members of the pelagic fish assemblage in Great Bay (Fig. 1), ranking fifth in frequency of occurrence and tenth in numerical dominance (Hagan and Able, 2003). The mean body size of bluefish collected in Great Bay, however, was 87.4 mm FL — a size that indicates that bluefish may not use estuaries during the earliest life history stages. More likely, YOY bluefish recruit to estuarine and ocean beach habitats as small juveniles (40-80 njm FL), as has previously been reported (Mc- Bride and Conover, 1991; Hare and Cowen, 1993; Able et al., 2003). This assertion was reconfirmed in our study by the synoptic examination of summer-spawned bluefish across multiple habitats, where the smallest bluefish collected in estuarine, inlet, and ocean beaches averaged 47.3 mm SL. In contrast to the size-composition of bluefish inhab- iting estuarine and coastal ocean sites, bluefish in sur- face waters on the inner continental shelf were strictly larval, transitional, and small juveniles <50 mm SL (Fig. 5). The paucity of juvenile bluefish >50 mm SL off the southern coast of New Jersey may be attributed to several factors including gear avoidance, size-de- pendent depth and habitat distributions, and active or passive emigration. First, gear avoidance appears un- likely because of the speed (-2.5-3.5 knots), duration (4 minutes), and frequency (140 tows) at which tows were performed (Norcross et al., 1974). For example, the ap- proximate swimming speed of a 50-mm bluefish is 10 cm/s (011a et al., 1985; Hare and Cowen, 1993) under the assumption that the fish swims at 2 body length/s (Hunter, 1981). At this rate, it is improbable that juve- nile bluefish could actively avoid a Methot trawl being towed at 128-180 cm/s (1 knot=51.44 cm/s). Secondly, larval-to-juvenile bluefish are surface oriented (0-6 m), as indicated by collection efforts across different water depths (Norcross et al., 1974; Kendall and Wal- ford, 1979; Kendall and Naplin, 1981; Shima, 1989) and by morphometric characteristics (e.g., silver and dark blue counter-coloration) of pelagic juveniles that indicate that they are adapted for a surface oceanic existence. Sampling in our study, however, was limited to the immediate surface layer (0-2 m), and therefore the current sampling design would not detect size-de- pendent depth distributions >2 m. In Virginian coastal 602 Fishery Bulletin 104(4) waters, Norcross et al. (1974) did not collect bluefish >22 mm total length in near-surface waters and at- tributed the absence of this fish to a lack of samples at depths where larger bluefish were presumably con- centrated. In our study, continental shelf bottom waters were sampled with otter trawls (trawl vertical opening ~2 m), and the size of first occurrence was 30 mm SL. This finding indicates that juveniles may seek deeper water as body size increases and swimming ability improves (Norcross et al., 1974). Alternatively, small bluefish collected with otter trawl tows may have been concentrated in surface waters and were incidentally captured during initial gear deployment and final re- trieval (i.e., under conditions when the otter trawl was inadvertently fishing the upper water column). Third, recruitment of juvenile bluefish >50 mm SL to estua- rine or coastal habitats is another plausible explanation for their absence in continental shelf waters. As previ- ously mentioned, bluefish actively or passively migrate inshore at 40-80 mm FL (McBride and Conover, 1991; Able et al., 2003). The size at which bluefish enter inshore habitats may be the result of biological tim- ing and morphological constraints. These two factors are consistent with the fact that inshore recruitment co-occurs with a switch in diet from copepods to avail- able piscine prey (Marks and Conover, 1993; Juanes and Conover, 1995). Correspondingly, bluefish may not enter estuaries until swimming ability improves and morphological development is complete, i.e., when fin ray development and scale formation (between 34 and 37 mm SL) are complete (Silverman, 1975). Swimming ability in many fish species dramatically improves af- ter the transformation from larval to juvenile stages (Hunter, 1981; Stobutzki and Bellwood. 1994). If mor- phological development is coupled with improved swim- ming ability in bluefish, juveniles most likely have the physical ability to actively recruit to estuarine habitats from inner continental shelf waters (Shima, 1989; Hare and Cowen, 1993, 1996). This study provides cursory evidence that summer-spawned juveniles move inshore during August and early September. Furthermore, the disappearance of bluefish of 25-50 mm SL from continental shelf surface waters coincided with the ap- pearance of somewhat larger fish (-50-60 mm SL) at estuarine and coastal ocean sites in late August and early September. Along the northeastern coast of the United States (Cape May, New Jersey, to Long Island, New York), in- gress of juvenile bluefish to inshore habitats presumably occurs as two distinct episodes (McBride and Conover, 1991 and references therein); which are consistent with intra-annual and bimodal length-frequency distribu- tions. This bimodality in bluefish size-composition is a result of a first cohort recruiting inshore from late May to early June (spring-spawned cohort), and a sec- ond cohort entering the same geographic region from July to October (summer-spawned cohort) (Nyman and Conover, 1988; McBride and Conover, 1991; McBride et al., 1993). In our study, the size-composition of bluefish across habitats was multimodal, yet length-based in- formation indicated that the overwhelming majority of YOY bluefish had been summer-spawned. Similarly, re- cent investigations of the inshore and continental shelf regions of the MAB also documented the numerical dominance of summer-spawned bluefish in late summer and early fall (Able et al., 2003; Conover et al., 2003; Wilber et al., 2003). Although summer-spawned bluefish presumedly domi- nated the catches of YOY across all habitats and tem- poral scales examined in our study, the recruitment success and contribution of these cohorts to year-class strength is unknown. The abundance of these fish, compared to that of spring-spawned individuals, implies that they potentially make important contributions to bluefish year-class strength (Able et al., 2003); yet it is unknown whether the area sampled in this survey is indicative of bluefish abundance along other portions of the U.S. northeastern coast or whether this sur- vey is representative of other years. Moreover, bluefish spawned later in the season (e.g., those encountered on the shelf in late September and October) may not achieve a size permitting either movement into estuar- ies or successful seasonal migrations. These bluefish, along with relatively small juveniles, may fail to con- tribute to the adult population because of size-selective mortality and decreased survival during overwintering periods (Hare and Cowen, 1997; Sogard, 1997; Hales and Able, 2001). There is also a consensus that spring- spawned juveniles frequently dominate the emigrating population in the fall, and therefore, are the key con- tributors to year-class strength (Nyman and Conover, 1988; Chiarella and Conover, 1990; Munch and Conover. 2000). This assertion has been recently contested, how- ever, because other studies have indicated that summer- spawned bluefish contribute equally, if not exceedingly, to the YOY population (Able et al., 2003; Conover et al., 2003; Wilber et al., 2003). It is probable that bluefish population dynamics along the coastal United States are a function of the combined recruitment success of spring- and summer-spawned cohorts and that contri- butions vary annually or over decadal scales (McBride and Conover, 1991; Munch and Conover, 2000). As a result, future research must focus on broad geographi- cal areas over sufficiently long temporal periods in order to adequately resolve the contribution of the different bluefish cohorts. Acknowledgments We are grateful to the staff of the Rutgers University Marine Field Station for assistance in field collections and processing of bluefish samples, particularly Geoff Bell, David Bottinelli, Ryan Nichols, and Rob Rinaldi. We also thank Doug Clarke (U.S. Army Corps of Engi- neers) for access to bluefish data from nearshore ich- thyoplankton surveys. This study was funded through the Bluefish/Striped Bass Research Program, Rutgers University and National Marine Fisheries Service col- laborative program. Taylor et al Habitat use by summer spawned Pomalomus saltatnx on the inner continental shelf off southern New Jersey 603 Literature cited Able, K. W., and M. P. Fahay. 1998. The first year in the life of estuarine fishes in the Middle Atlantic Bight, 342 p. Rutgers Univ. Press, New Brunswick. NJ. Able, K. W.. and J. C. Lamonaca. In press. Scale formation in selected western North Atlantic flatfishes. J. Fish Biol. Able, K. W., P. Rowe, M. Burlas, and D. Byrne. 2003. Use of ocean and estuarine habitats by young- of-the-year bluefish iPomatonuis saltatrix) in the New York Bight. Fish. Bull. 101:201-214. Berrien, P. L., and J. Sibunka. 1999. Distribution patterns offish eggs in the U.S. north- east continental shelf ecosystem. 1977-1987. NOAA Tech. Rep. NMFS-145, 310 p. Chiarella, L. A., and D. O. Conover. 1990. Spawning season and first-year growth of adult bluefish from the New York Bight. Trans. Am. Fish. Soc. 119:455-462. Conover. D. O., T. Gilmore, and S. B. Munch. 2003. Estimating the relative contribution of spring- and summer-spawned cohorts to the Atlantic Coast bluefish stock. Trans. Am. Fish. Soc. 132:1117-1124. Creaser. E. P., and H. C. Perkins. 1994. The distribution, food, and age of juvenile blue- fish, Ponuitomus saltatrix. in Maine. Fish. Bull. 92: 494-508. Day, R. W., and G. P. Quinn. 1989. Comparisons of treatments after an analysis of variance in ecology. Ecol. Monogr. 59:433-463. Deuel. D. G.. J. R. Clark, and A. J. Mansuetti. 1966. Description of embryonic and early larval stages of bluefish, Pomatomus saltatnx. Trans. Am. Fish. Soc. 95(3): 264-271. Hagan. S. M.. and K. W. Able. 2003. Seasonal changes of the pelagic fish assemblage in a temperate estuary. Estuar. Coast. Shelf Sci. 56:15-29. Hales, L. S., Jr., and K. W. Able. 2001. Winter mortality, growth, and behavior of young- of-the-year of four coastal marine fishes in New Jersey (USA) waters. Mar. Biol. 139:45-54. Hare, J. A., and R. K. Cowen. 1993. Ecological and evolutionary implications of the larval transport and reproductive strategy of bluefish, Pomatomus saltatrix. Mar. Ecol. Prog. Ser. 98:1-16. 1994. Ontogeny and otolith microstructure of bluefish Pomatomus saltatrix (Pisces: Pomatomidae). Mar. Biol. 118:541-550. 1996. Transport mechanisms of larval and pelagic juve- nile bluefish '.Pomatomus saltatnx) from South Atlantic Bight spawning grounds to Middle Atlantic Bight nursery habitats. Limnol. Oceanogr. 41(6):1264-1280. 1997. Size, growth, development, and survival of the planktonic larvae of Pomatomus saltatrix (Pisces: Pomatomidae). Ecology 78:2415-2431. Hare, J. A., M. P. Fahay, and R. K. Cowen. 2001. Springtime ichthyoplankton of the slope region off the north-eastern United States of America: larval assemblages, relation to hydrography and implications for larval transport. Fish. Oceanogr. 10:164-192. Hunter, J. R. 1981. Feeding ecology and predation of marine fish larvae. In Marine fish larvae: morphology, ecology, and relation to fisheries (R. Lasker, ed.), p. 34-77. Wash- ington Sea Grant Program, Univ. Washington Press, Seattle, WA. Juanes, F., and D. O. Conover. 1995. Size-structured piscivory: advection and the linkage between predator and prey recruitment in young-of-the- year bluefish. Mar. Eeoi. Prog. Ser. 128:287-304. Juanes. F., J. A. Hare, and A. G. Miskiewicz. 1996. Comparing early life history strategies of Poma- tomus saltatrix: a global approach. Mar. Freshw. Res. 47:365-379. Kendall, A. W., and N. A. Naplin. 1981. Diel-depth distribution of summer ichthyoplankton in the Middle Atlantic Bight. Fish. Bull. 79:705-726. Kendall, A. W., Jr., and L. A. Walford. 1979. Sources and distribution of bluefish, Pomatomus saltatrix, larvae and juveniles off the east coast of the United States. Fish. Bull. 77:213-227. Klein-MacPhee, G. 2002. Bluefish: family Pomatomidae. In Bigelow and Schroeder's fishes of the Gulf of Maine (B. B. Collette, and G. Klein-MacPhee. eds. ), p. 400-406. Smithsonian Institution Press, Washington, D.C. Marks, R. E.. and D. O. Conover. 1993. Ontogenetic shift in the diet of young-of-the-year bluefish Pomatomus saltatrix during the oceanic phase of the early life history. Fish. Bull. 91:97-106. McBride, R. S., and D. O. Conover. 1991. Recruitment of young-of-the-year bluefish Poma- tomus saltatrix to the New York Bight: variation in abundance and growth of spring- and summer-spawned cohorts. Mar. Ecol. Prog. Ser. 78:205-216. McBride, R. S., J. L. Ross, and D. O. Conover. 1993. Recruitment of bluefish i.Pomatomus saltatnx) to estuaries of the South Atlantic Bight, U.S.A. Fish. Bull. 91:389-395. Methot. R. D.. Jr. 1986. Frame trawl for sampling pelagic juvenile fish. CalCOFI. Rep. 27:267-278. Miller, R. V. 1969. Continental migrations of fishes. Underw. Nat. 6:15-24. Munch, S., and D. O. Conover. 2000. Recruitment dynamics of bluefish. Pomatomus salta- trix, on the continental shelf from Cape Hatteras to Cape Cod, 1973-1995. ICES J. Mar. Sci. 57:393-402. Norcross. J. J., S. L. Richardson, W. H. Massmann. and E. B. Joseph. 1974. Development of young bluefish {Pomatomus sal- tatrix) and distribution of eggs and young in Virginian coastal water. Trans. Am. Fish. Soc. 103:477-497. Nyman, R. M., and D. O. Conover. 1988. The relation between spawning season and the recruitment of young-of-the-year bluefish, Pomatomus saltatrix, to New York. Fish. Bull. 86:237-250. OUa, B. L., A. L. Studholme, and A. J. Bejda. 1985. Behavior of juvenile bluefish Pomatomus salta- trix in vertical thermal gradients: influence of season, temperature acclimation and food. Mar. Ecol. Prog. Ser. 23:165-177. Oozeki. Y., F. Hu, H. Kubota. H. Sugisaki, and R. Kimura. 2004. Newly designed quantitative frame trawl for sampling larval and juvenile pelagic fish. Fish. Sci. 70:223-232. Pottern, G. B., M. T. Huish, and J. H. Kerby. 1989. Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (mid- 604 Fishery Bulletin 104(4) Atlantic )-bluefish. U.S. Fish and Wildlife Service Bio- logical Report No. 82(11.94), 20 p. U.S. Army Corps of Engineers TR EL-82-4, Vicksburg, MS. Shepherd, G. R., J. Moser, D. Deuel, and P. Carlson. 2006. The migration patterns of bluefish iPomatoinus saltatrix) along the Atlantic Coast determined from tag recoveries. Fish. Bull. 104:559-570. Shima, M. 1989. Oceanic transport of the early life history stages of bluefish, Pomatomus saltatrix. from Cape Hatteras to the Mid-Atlantic Bight. M. Sci. thesis, 69 p. State Univ. New York at Stony Brook, Stony Brook, NY. Silverman, M. J. 1975. Scale development in the bluefish, Pomatomus saltatrix. Trans. Am. Fish. Soc. 104:773-774. Smith, W., P. Berrien, and T. Potthoff 1994. Spawning patterns of bluefish, Pomatomus salta- trix, in the northeast continental shelf ecosystem. Bull. Mar. Sci. 54:8-16. Sogard, S. M. 1997. Size-selected mortality in the juvenile stages of tele- ost fishes: a review. Bull. Mar. Sci. 60:1129-1157. Stobutzki, I. C, and D. R. Bellwood. 1994. An analysis of the sustained swimming abilities of pre- and post-settlement coral reef fishes. J. Exp. Mar. Biol. Ecol. 175:275-286. Taylor, D. L., and K. W. Able. In press. Cohort dynamics of summer-spawned bluefish as determined by length-frequency and otolith micro- structure analyses. Trans. Am. Fish. Soc. Underwood, A. J. 1981. Techniques of analysis of variance in experimental marine biology and ecology. Ocean. Mar. Biol. Ann. Rev. 19:513-605. U.S. Department of Commerce. 1986. Marine recreational fishery statistics survey, Atlantic and Gulf coasts, 1985, 130 p. Current Fish- ery Statistics No. 8327, Washington, D.C. von Ende, C. N. 1993. Repeated-measures analysis: growth and other time-dependent measures. In Design and analysis of ecological experiments (S. M. Scheiner, and J. Gur- evitch, eds.), p. 113-137. Chapman and Hall, New York, NY. Wilber, D. H., D. G. Clarke, M. H. Burlas, H. Ruben, and R, J. Will. 2003. Spatial and temporal variability in surf zone assem- blages on the coast of northern New Jersey. Estuar. Coast. Shelf Sci. 56:291-304. Wilk, S. J. 1982. Bluefish. Pomatomus saltatrix. MESA (Mar. Eco- syst. Anal.) N.Y. Bight Atlas, Monogr. 15:86-89. Witting, D. A., K. W. Able, and M. P. Fahay. 1999. Larval fishes of a Middle Atlantic Bight estuary: assemblage structure and temporal stability. Can. J. Fish. Aquat. Sci. 56:222-230. 605 Abstract — Longfin inshore iLollgo pealeii) and northern shortfin Ullex illecebrosus) squids are considered important prey species in the North- west Atlantic shelf ecosystem. The diets of four major squid predators, bluefish [PomatomuH saltatrix), goose- fish {Lophius americanus), silver hake (Merluccius bilinearis), and summer flounder (Paralichthys dentatus), were examined for seasonal and size-based changes in feeding habits. Summer and winter, two time periods largely absent from previous evaluations, were found to be the most impor- tant seasons for predation on squid, and are also the periods when the majority of squid are landed by the regional fishery. Bluefish >450 mm, silver hake >300 mm, and summer flounder >400 mm were all found to be significant predators of squid. These same size fish correspond to age classes currently targeted for biomass expansion by management committees. This study highlights the importance of understanding how squid and predator interactions vary temporally and with changes in com- munity structure and stresses the need for multispecies management in the Northwest Atlantic. Seasonal and size-based predation on two species of squid by four fish predators on the Northwest Atlantic continental shelf Michelle D. Staudlnger Department of Natural Resources Conservation 160 Holdsworth Way University of Massachusetts, Amherst Amherst, IVlassachusetts 01003 9285 Email address mstaudimg'nreumassedu Manuscript submitted 7 November 2005 to the Scientific Editor's Office Manuscript approved for publication 9 March 2006 by the Scientific Editor. Fish. Bull. 104:605-615 (2006). Global depletion of marine predators has had dramatic effects on ecosys- tem structure and function (May et al., 1979; Jackson et al., 2001; Pauly et al., 20021. In many systems, the ramifications of such changes may not be fully realized. Groundfish declines have been linked to simultaneous increases in cephalopod landings in fifteen key Food and Agriculture Organization of the United Nations (FAO) areas (Caddy and Rodhouse, 1998). It is uncertain whether cepha- lopod populations are experiencing increased growth due to a release from predation (Piatkowski et al., 2001), or whether increased harvests are representative of the trend toward fishing species at lower trophic levels (Pauly et al., 1998). In the Northwest Atlantic, while gadids, flatfish, and other demersal species have been reduced because of overfishing (Link and Garrison, 2002), squid have risen in status from a mere bait fishery to one of the most economically impor- tant stocks in the region (Cadrin and Hatfieldi). Cephalopods have been documented as principal prey for numerous species of finfish, elasmobranchs, and marine mammals (Smale, 1996); however, in comparison to the ecological relation- ships between fish and their preda- tors, less is known about the ecological interactions between squid and their predators. In a comprehensive evalu- ation of the Northwest Atlantic food web, bluefish (Pomatomus saltatrix), goosefish {Lophius americanus), sil- ver hake (Merluccius bilinearis), and summer flounder (Paralichthys den- tatus) were ranked among the most significant predators of squid (Bow- man et al., 2000). Squid represented between 17% and 95% of the total mass consumed by these four finfish regionally. Dramatic changes in stock abundance and population structure have occurred since specimens for Bowman et al.'s study were collect- ed 25-45 years ago. Exploitation of squid has risen substantially, and the four noted predators have experienced severe depletions; stock biomass lev- els have fallen as low as 20-50% of their respective maximum sustain- able yield (B,;^.,) limits (NOAA-). Additionally, the size structure of predator populations within the com- munity has become skewed because of age-truncation. Recent evaluations of bluefish (Buckel et al., 1999a), goose- fish (Armstrong et al., 1996), silver hake (Bowman, 1984), and summer flounder (Link et al., 2002) foraging habits have been conducted; however, sampling has been restricted to one or two seasons, primarily spring and fall. Analyses that base their results on feeding habits collected during a single season (Buckel et al., 1999b), or where results are based on data that have been pooled under the as- ' Cadrin. S. X., and E. M. C. Hatfield. 1999. Stock assessment of inshore longfin squid Loligo pealeii. Northeast Fish. Sci. Cent. Ref Doc. 99-12, 107 p. - NOAA ( National Oceanic and Atmospheric Administration). 2001. Northeast Fisheries Science Center (NEFSC). Status of fishery resources off the north- eastern United States. Website: http:// www.nefse.noaa.gov/sos/spsyn/species. html (accessed on 8 March 2006]. 606 Fishery Bulletin 104(4) sumption that the diets collected in one season are a proxy for another (Overholtz et al., 2000) could overlook key periods of predation and lead to an underestimation of the total predatory demand imposed on principal prey resources, such as squid. The two primary squid species found in Northwest Atlantic waters, longfin inshore iLoligo pealeii) and northern shortfin illlex illecebrosus). are both highly migratory, and their distribution on the shelf is tempo- rally variable. Squid move between inshore (spring and summer) and offshore (fall and winter) environments seasonally (Macy and Brodziak, 2001). On a diurnal basis they move from demersal waters during the day to surface waters at night (Lange and Sissenwine, 1983). The degree of vertical movement made by squid is also known to vary seasonally. Activity is more pronounced during warmer months when the water column is strati- fied and is diminished during winter and spring when shelf waters are well mixed (Hatfield and Cadrin, 2002). Prey availability and distribution in the water column, water temperature, and other environmental factors are believed to influence diel migration patterns (Cargnelli et al., 1999). Seasonal changes in squid behavior and habitat use will also affect encounter rates with differ- ent predators in the demersal environment. Knowledge of species interactions is imperative to understand population dynamics and to manage stock recovery (Murawski, 1991). The present study provides a current assessment of the reliance on squid popula- tions in the Northwest Atlantic region by four major squid predators; bluefish, goosefish, silver hake, and summer flounder. For each predator, ontogenetic and seasonal variations in feeding patterns were evaluated. Additionally, squid abundance in the demersal environ- ment was related to predator diets as a mechanism for diurnal and seasonal changes in predation. fore adequate information was not available to evaluate this season. To compare changes in longfin squid abundance in the demersal environment at different times of day and between seasons, the relative masses of prerecruits (W,) and recruits (W^^, ) present were estimated by using the equations W, Piyear. season, time of day , = I w,p. t t f year, season, time of day i r fr ( season, tune of day ) W, Rl year, season, tune of da' v,=I W.R, I 1 1 year, season, time of day i htst (1) (2) season, time of day i where t = the index of tows made at NMFS stations in New York, New Jersey, Connecticut, Rhode Island, and Massachusetts respective to each year (2002, 2003), season (winter, spring, fall), and time of day (day, night, dawn and dusk); W, = the total mass of longfin squid caught in each tow; and P and i? = (1 - P) estimate the proportion of biomass in each of two size classes, prerecruits (<80 mm) and recruits (>80 mm), of longfin squid during each season. /", = the diel correction coefficient representing relative catch rates of longfin squid for each size class, season, and time of day as deter- mined by Hatfield and Cadrin (2002) and was standardized to 1.0 during daytime for all seasons. Materials and methods Estimating changes in the abundance of longfin inshore squid Longfin and shortfin squid are regularly caught in National Marine Fisheries Service (NMFS) bottom trawl surveys. However, catches of shortfin squid were exceptionally low over 2002 and 2003 (the time period evaluated during the present study), possibly because of poor recruitment. Catch data used in subsequent cal- culations were provided by NMFS bottom-trawl surveys (NMFS^^-SS-'-S). Sufficient information was available for longfin squid only; consequently abundance estimates were limited to this species. Furthermore, abundance surveys were not conducted during the summer; there- NMFS (National Marine Fisheries Service). 2002. Fish- ermen's report: bottom trawl survey. Cape Hatteras-SE Georges Bank: February 5-March 2,' 2002, FRY Albatross IV. 24 p. NMFS, Northeast Fisheries Science Center, 166 Water St., Woods Hole, MA 02.543. ^ NMFS (National Marine Fisheries Service). 2002. Fish- ermen's report: bottom trawl survey. Cape Hatteras-Gulf of Maine: March 5-April 25, 2002," FRV Albatross IV. 34 p. NMFS, Northeast Fisheries Science Center, 166 Water St., Woods Hole, MA 02543. 5 NMFS (National Marine Fisheries Service). 2002. Fish- ermen's report: bottom trawl survey. Cape Hatteras-Gulf of Maine: September 4-October 25, 2002, FRV Albatross IV, 31 p. NMFS, Northeast Fisheries Science Center, 166 Water St., Woods Hole, MA 02543. 6 NMFS (National Marine Fisheries Service). 2003. Resource survey report: bottom trawl survey. Cape Hatteras-Southern New England: February 4-March 1, 2003, FRV Delaware II, 19 p. NMFS, Northeast Fisheries Science Center, 166 Water St., Woods Hole, MA 02543. ' NMFS (National Marine Fisheries Service). 2003. Resource survey report, bottom trawl survev, Cape Hatteras-Gulf of Maine: March 5, 2003-April 27, 2003, FRV Delaware II. 34 p. NMFS, Northeast Fisheries Science Center. 166 Water St., Woods Hole, MA 02543. * NMFS (National Marine Fisheries Service). 2003. Resource survey report: bottom trawl survey, Cape Hatteras-Gulf of Maine: September 7-November 1, 2003, FRV Albatross IV. 35 p. NMFS, Northeast Fisheries Science Center, 166 Water St., Woods Hole, MA, 02543. Staudinger Piedation by four fish predators on two squid species on the Northwest Atlantic continental shelf 607 68-30 0 i_ New York New Jersey Atlantic Ocean e ® Legend • NMFS A NJDEP ■ Sea Grant Figure 1 Map of fishery-independent sampling locations. Circles (•) are locations sampled during the National Marine Fisheries Service annual bottom-trawl survey cruise, triangles (A) are locations sampled by the New Jersey Department of Environmental Protection, squares (■) are locations sampled during a Sea Grant sponsored cruise transecting the continental shelf along Block Island Sound Canyon. To obtain an index of relative abundance, final mass values were divided by the number of tows made during each time of day, year, and seasonal period. For Equations 1 and 2, values for P and R were es- timated from longfin squid collected regionally during the winter, spring, and fall. Mantle lengths (mm) and the corresponding body weights (g) of individual squid measured in Hunsicker (2004) were used to calculate the percent distribution of biomass in each of the two size classes during each seasonal period. These esti- mated proportions were then used to partition the total catches reported at each NMFS station into prerecruit and recruit biomass. For each seasonal and diurnal period, catches were averaged between years and used to contrast relative concentrations of longfin squid in the demersal environment. Gear biases may affect the catchability of prerecruits more than recruits; therefore relative catch rates were compared among seasons only within each size class. Sampling Bluefish, goosefish, silver hake, and summer flounder were collected from continental shelf waters off the coasts of New York, New Jersey, Connecticut, Rhode Island, and Massachusetts from February 2002 to July 2003. Samples were obtained from two main sources: from fish landed by commercial fishing boats and from scientific survey cruises. Fishery-independent collections were made in conjunction with bottom trawl surveys conducted bimonthly by the New Jersey Department of Environmental Protection (NJDEP) and seasonally (except summer) by the National Marine Fisheries Service, Northeast Fisheries Science Center (NMFS- NEFSC) (Fig. 1). Samples were also collected aboard the RV Sea Wo// (Sea Grant) during an independent survey that transected the continental shelf along the Block Island Sound Canyon (Fig. 1). The proportion of samples collected from fishery-independent sources is shown in 608 Fishery Bulletin 104(4) Figure 2 and the number of samples collected from all sources during all seasons is shown in Figure 3. In addition to the stomach contents col- lected and analyzed for the present study, diet data collected regionally were drawn from the NMFS-NEFSC bottom trawl survey database. A full description of the survey design and collection methods used by the NEFSC can be found in Azarovitz et al.'' Diet analysis Winter Spring Summer Season The majority of samples obtained from fish- ery-independent sources were processed onboard research vessels. However, only cur- sory assessments of prey type and mass could be made at sea; therefore samples were frozen and transported to the laboratory for a more thorough examination. Each fish was weighed and measured (fork length for bluefish, total length for goosefish, silver hake, and summer fiounder), stomachs were removed, prey items were weighed to the nearest 0.01 gram and identified to the level of species whenever possible. With the aid of a dissecting microscope, otoliths, beaks, and other hard parts were recovered in order to classify species. When stomach contents could not be identified because of advanced stages of digestion, they were recorded as "unidentified animal remains." As with all diet stud- ies, there is always a chance that a portion of samples are biased because fish may feed while in the net. This problem was addressed by excluding fish where prey was found (undigested) in the mouth and esophagus or where Figure 3 Frequencies (in days) of samples of bluefish iPoniatomus !taltat?'ix), goosefish iLophius americanus), silver hake [Merluccius bilinearis), and summer flounder iParalichthys dentatus) collected from fish- ery-dependent and fishery-independent sources. No samples were collected for bluefish iPomatomus saltatnx) during the winter and spring seasons. Azarovitz, T., S. Clark, L. Despres. and C. Byrne. 1997. The Northeast Fisheries Science Center bottom trawl survey program. 22 p. ICES Council Meeting 1997/Y:3.3. B Bluefish BGoosetish nSilverhake D Summer flounder 1.0 1 Winter Spnng Summer Fall Season Figure 2 The proportion of samples of bluefish iPomatomus saltatrix). goosefish {Lophius americanus), silver hake [Merluccius bilin- earis), and summer flounder iParalichthys dentatus) specimens collected from fishery-independent sources. No samples were collected for bluefish during the winter and spring seasons. there was a clear difference in digestion stage among prey items found in the stomachs. Three size classes, covering the total range of lengths offish collected, were chosen for each species as follows: bluefish — small (200 to 450 mm fork length), medium (451 to 550 mm), and large (>551 mm); goosefish — small (50 to 250 mm total length), medium (251 to 450 mm), and large (>451 mm); silver hake — small (50 to 199 mm total length), medium (200 to 299 mm), and large (>300 mm); and summer flounder — small (250 to 399 mm total length), medium (400 to 549 mm), and large (>550 mm). Seasonal time periods were defined as win- ter (December-February), spring (March-May), summer (June-August), and fall (September-November). Statistical analyses The Kruskal-Wallis test was used to test seasonal and size-based differences in predation on squid. A nonparametric test was chosen because of its robustness to assumptions of normality and skew- ness (Quinn and Keough, 2003). Preliminary anal- yses indicated that none of the data sets fitted the normal distribution and, owing to large numbers of zero values (representing the absence of squid in the diet), could not be transformed. Two sets of tests were performed on the data: first an "aggres- sive" set of tests were run by using individual fish as the sampling unit to give the maximum number of degrees of freedom; second, a "conservative" set of tests were run by using either stations (for data collected from fishery-independent sources) or cruises (for data collected from commercial fishery sources) as the sampling unit. The percent mass of squid in the diet of each fish (aggressive test) or pooled fish per station or cruise (conservative test) was determined and expressed as a proper- Staudinger: Predation by four fish predators on two squid species on the Northwest Atlantic continental shelf 609 I Winter B Spring QFall Pre-recruits B Recruits Time of day Figure 4 Relative catches of longfin inshore squid {Loligo pealeii) (A) prerecruits (s80 mm mantle length) and (B) recruits (>80 mm mantle length) predicted in demersal waters during winter, spring, and fall of 2002 and 2003 (summer sampling was not conducted). Abundances of longfin inshore squid were measured as catch per tow (kg) and adjusted by seasonal diel correction coefficients specified in Hatfield and Cadrin (2002). tion of the total mass consumed. To account for large numbers of tied rank values, an adjusted H test statistic was used to test the null hypothesis (Sokal and Rohlf, 1995). The two approaches produced similar results for the majority of tests. Therefore, only conservative test results are reported. Results Seasonal and diurnal differences in longfin inshore squid abundance Predicted catches of longfin prerecruits during 2002 and 2003 indicated that the highest catch levels occured during the fall, intermediate levels during the winter, and minimum levels during the spring (Fig. 4A). These calculated values are similar to the historical abun- dance trends previously reported in Northwest Atlantic waters. Fall catches of prerecruits over all times of day were an order of magnitude higher in comparison to winter and spring values. The greatest variation in diurnal catches of prerecruits was predicted during the fall season (Fig. 4A). Conversely, catches of longfin recruits were estimated to be at their maximum during the winter, at intermediate levels during the fall, and at minimum levels during the spring (Fig. 4B). Maximum catch rates of longfin recruits were predicted during crepuscular and daytime periods of winter (Fig. 4B). Moreover, nighttime catches of recruits were two to three times higher during winter than during all other seasons. Diet analysis Bluefish Bluefish are present in Northern Atlantic waters only during the warmer months of the year; therefore sampling was restricted to the summer and fall seasons. A total of 299 bluefish from 26 sampling dates were analyzed. Fish ranged in size from 90 to 780 mm FL (average=469.6 mm). Overall, 53% of all bluefish sampled had food items present in their stomachs, and diets were split nearly equally between squid and fish. Perciforms were the dominant piscine prey; small amounts of amphi- pods and other pelagic crustaceans made up the remain- der of the diet. A complete list of prey species recovered from all predator diets and their designated taxonomic groups can be found in Staudinger (2004). Longfin squid was the primary squid species con- sumed by bluefish. A substantial amount of shortfin squid was also found in the stomachs of large fish (Table 1). Total predation on squid was greater during the summer than in the fall (Table 2); however, season- al differences were not statistically significant possibly because of the high frequency of zero values in the da- taset (i^„rf,„sW=2-40, P=0.121, n=25). Medium and large fish were the primary consumers of squid, and small bluefish fed almost exclusively on fish (Table 1). Sig- nificant differences were found in the amount of squid consumed among all size classes of bluefish (H^^j^^f^j= 6.62, P=0.037, 7?=75). Goosefish A total of 536 goosefish stomachs were ana- lyzed of which 269 (50%) contained prey. Goosefish ranged in size from 75 to 818 mm TL (average = 341.1 mm). Goosefish were almost exclusively piscivorous; 93% of the total diet consisted of a mixture of gadiforms, clu- peiforms, perciforms, and several other fish groups. Goosefish were the only species in the present study to prey on rajiforms. Occurrence of squid in the goosefish diet was re- stricted to the winter and fall (Table 2) and differences among all seasons were significant (//g^,^,,,,p^,=15.27, P=0.002, n=A6). Although medium size goosefish were 610 Fishery Bulletin 104(4) Table 1 Size-based stomach contents of fish collected in shelf waters of the Northwest Atlantic. Values presented are percentages of mass consumed (in grams). Fish lengths Ismail, medium, and large) are measured in millimeters. "Percent feeding" means the percentage of stomachs analyzed that contained prey. Blueflsh were measured in fork length: goosefish, silverhake, and summer flounder were measured in total length. Prey type Bluefish Goosefish Small 200-450 Medium 451-550 Large 551+ Small 50-250 Medium 251-450 Large 451+ Squid ■ 1.3 75.4 42.2 — 4.9 0.5 Longfin inshore squid 1.2 44.9 18.1 — 4.9 0.5 Northern shortfin squid — 1.7 15.0 — — — Unidentified squid 0.1 28.8 9.1 — — — Fish 98.4 23.0 53.5 99.4 94.8 91.9 All other prey 0.3 1.6 4.3 0.56 0.24 7.68 Number of stomachs analyzed 103 99 113 148 153 235 Number containing prey (percent feeding) 76 ( 74 ) 51(52) 49(431 96(65) 75(49) 98(42) Mean length ( mm, SE ) 316.6(9.9) 509.8(3.3) 628.6(7.7) 124.8(3.0) 358.2(6.9) 540.0(7.11 Length range (mm) 200-438 460-547 555-780 75-241 255-450 460-818 Prey type Silver hake Summer flounder Small 50-199 Medium 200-299 Large 300 + Small 250-399 Medium 400-549 Large 550+ Squid 0.0 2.7 18.6 16.7 33.5 20.6 Longfin inshore squid 0.0 2.7 12.3 15.3 25.4 14.7 Northern shortfin squid — — 0.2 — — 5.3 Unidentified squid — 0.1 6.1 1.4 8.1 0.6 Fish 34.7 17.6 62.0 67.2 55.1 76.0 All other prey 65.33 79.63 19.33 16.13 11.40 3.44 Number of stomachs analyzed 174 566 240 316 360 154 Number containing prey (percent feeding) 132(76) 298(53) 108(45) 85(27) 91(25) 54(35) Mean length (mm, SE) 150.5(3.1) 260.4(1.5) 334.6(3.9) 346.1 (4.4) 454.9(4.1) 608.2(6.4) Length range (mm) 60-199 200-298 300-484 250-399 400-530 550-750 the only size class found to consume squid in notable amounts (Table 1), differences in the proportions of squid consumed among all size classes were not signifi- cant (//,„,,„^,,,,,=3.11. P=0.211, « = 138). All squid found in the goosefish diet were classified as longfin squid. Silver hake Of the 980 silver hake stomachs examined, over half (55%) contained prey. Silver hake ranged in size from 60 to 484 mm TL (average=248.3 mm). Feed- ing rates for silver hake were highest during the summer months and lowest during the spring and fall. Overall, silver hake fed primarily on fish and crustaceans, and lesser amounts of squid and other invertebrates. Amphi- pods and euphausiids were the dominant crustacean prey. Consumption of gadiforms, including conspecifics and unclassified osteichthyans, made up the majority of piscine prey. Predation on squid differed significantly among sea- sons (W„a'ju.sto/=1003. P=0.018, ?! = 55) and was at its maximum during winter (Table 2). Silver hake diets exhibited a pronounced shift towards squid with in- creasing size (Table 1). Small and medium hake ate negligible amounts of squid. Conversely, squid totaled nearly 20*7^ of all mass consumed by large hake. Dif- ferences in squid predation among size classes were significant 60-484 79-480 103-410 110-450 270-710 300-750 268-645 250-690 The proportion of squid consumed by summer flounder fluctuated significantly among seasons (^n(/,„j,„/=14.29, P=0.003, n = 43). Predation on squid was at its maxi- mum during winter, and elevated during the summer in comparison to spring and fall (Table 2). Although differences in squid predation among size classes were not significant 20':'( ) of the goosefish diet in the southern New England region. In a study conducted by Arm- strong et al. (1996), longfin squid were found in small goosefish (<400 mm TL) stomachs during summer and represented approximately 10% of the diet. In contrast, the present study found that squid contributed trivial amounts to the diet of goosefish across all size classes and seasons. It is unclear why such vast discrepancies were observed between these studies. One possible ex- Staudinger Predation by four fish predators on two squid species on the Northwest Atlantic continental shelf 613 planation is that the opportunistic foraging strategy of goosefish is more sensitive to variation in relative and overall abundances of prey and results in erratic food habits over various time scales. Predator-prey behavior Differences in the amount of squid consumed among species, among predator sizes, and seasonal periods indicate that aspects of predator and prey behavior must be playing a pivotal role in influencing the susceptibility of squid to predation. Seasonal inshore-offshore habitat use and diurnal vertical migrations are two mechanisms that may act to mediate predator-prey encounter rates. Other behaviors that influence encounter rates are the times of day that each predator is actively hunting and the areas of the water column being searched. Of the four predators evaluated, seasonal movements by summer flounder are most similar to the migration pat- terns of longfin squid. Summer flounder are primarily daytime feeders and spend the majority of their time on or near the seafloor (Packer et al., 1999). Conversely, silver hake are a demersal species known to be active hunters primarily in the late afternoon and evening (Bowman, 1984). Large vertical movements in the water column are not customary for silver hake (Bigelow and Schroeder, 1953), therefore nighttime feeding limits interactions during seasons when squid exhibit strong diurnal migrations. Bluefish are the only predator in this study considered to be pelagic and that actively pursue prey throughout the water column. Although bluefish have been shown to forage primarily during the day (Juanes and Conover, 1994), it is possible that there is a seasonal shift in depth at which bluefish forage (Bigelow and Schroeder, 1953). A shift from feeding at depth in the late spring and summer to feeding in the surface waters during fall may explain, in part, seasonal dif- ferences in importance of squid to bluefish because few (if any) squid are present in the upper water column during daylight. In general, daytime foraging behavior, combined with positioning in the demersal environment, create the op- timal setting for regular encounters between squid and fish predators and explain differences in the proportion or even presence of squid in the diet between species with similar geographical ranges. Implications for management Since the late 1980s, squid harvests have increased substantially during the winter. Offshore harvests have been, on average, three times greater than inshore catches between October and March (Cadrin'M. The combination of increased fishing pressure and elevated predation rates concentrated within a single season may not be sustainable for a species, such as squid, with a life span of less than a year and that has little overlap between generations (Pierce and Guerra, 1994). If not properly accounted for, this increase in total mortality on squid during the winter could affect the number of squid surviving to spawn during spring and summer and limit available biomass for predators such as bluefish and summer flounder. Another area of concern stems from potential in- creases in top-down pressure from recovering predator populations. Regionally, management has emphasized age-truncation of predators as one of the most seri- ous factors affecting summer flounder and silver hake populations. The age classes of summer flounder (age-2) (Terceiro'-) and silver hake (age-3) (Brodziak'^) tar- geted for expansion in their respective management plans, correspond to the lengths and associated ages identified in the present study as being the most vora- cious predators of squid. If management objectives are met and summer flounder and silver hake stocks are fully rebuilt, predation on squid could rise substantially as fish that have been functionally absent from the population begin increasing in abundance. Other species, such as bluefish, that have been se- verely overfished are also showing signs of recovery (ASMFC'"*). Consumption of squid by bluefish has been estimated to exceed the mass removed by the fishing industry (Buckel et al., 1999b) and is contingent on predator population abundance (Overholtz et al., 2000). If one includes summer foraging habits, total consump- tion of longfin squid by bluefish may be much greater than previously estimated. In light of this new informa- tion, a reassessment of the predatory demand imposed by bluefish and other recovering predators on squid is clearly needed before fish stocks are fully rebuilt. The simultaneous exploitation of predators and their prey, when there is limited information about species interactions, is a precarious practice that can have un- intended and potentially detrimental consequences for one or both stocks. Depletion of biomass that supports higher predators may decrease biological production. Conversely, prey populations may become overextended. A greater understanding of species relationships is nec- essary, especially in ecosystems where fishing occurs at multiple trophic levels. The present study shows that consideration of predation year-round and for a range of predator sizes is imperative to accurately assess ■> Cadrin, S. X. 2001. Status of fi.shery resources off the northeastern United States: longfin inshore squid. Website: http://www.nefsc.noaa.gov/sos/spsyn/iv/lfsquid/ laccessed on 8 March 2006]. ^2 Terceiro, M. 2003. Stock assessment of summer flounder for 2003. Northeast Fisheries Science Center Reference Document, 179 p. National Marine Fisheries Service, Northeast Fisheries Science Center, 166 Water St.. Woods Hole, MA, 02.543. ^^ Brodziak, J. 2001. Status of fishery resources off the northeastern United States: silver hake. Website: http:// www.nefsc.noaa.gov/sos/spsyn/pg/silverhake/ laccessed on 8 March 20061. '••(ASMFC) Atlantic States Marine Fisheries Commission. 2003. Website: http://www.asmfc.org laccessed on 8 March 20061. 614 Fishery Bulletin 104(4) predation pressure on squid populations. As predator populations in the Northwest Atlantic are recovered under current management initiatives, a multispecies management approach will be crucial to avoid conflicts between squid and finfish fisheries and should be ap- plied to species beyond those presented in this article. Acknowledgments I thank T. Essington. R. Armstrong, F. Juanes. R. Cer- rato, D. Conover, and four anonymous reviewers for their help with previous drafts of this manuscript; W. Kramer for querying data from the NMFS database; J. Palmer and L. Hendrickson for graciously provid- ing stock assessment information; and Inlet Seafoods, Alice's Seafood, and Soleau's Wharf and Marina for pro- viding fish for diet analyses. This research was funded by the New York Sea Grant Institute. Work was conducted at the Marine Science Research Center at Stony Brook University as part of a Master of Science thesis. Literature cited Armstrong, M. P., J. A. Musick, and J. A. Colvocoresses. 1996. Food and ontogenetic shifts in feeding of the goose- fish, Lophius americanus. J. Northw. Atl. Fish. Sci. 18:99-103. Bigelow, H. B., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv., Fish. Bull. 53, 577 p. Bowman, R. E. 1984. Food of silver hake. Merluccius bilinearis. Fish. Bull. 82:21-35. Bowman, R. E., C. E. Stillwell, W. L. Michaels, and M. D. Grosslein. 2000. Food of northwest Atlantic fishes and two common species of squid. NOAA Technical Memo. NMFS-NE- 155, 137 p. National Marine Fisheries Service, 166 Water St., Woods Hole, MA, 02543. Brodziak, J., and L. Hendrickson. 1999. An analysis of environmental effects on survey catches of squid Loligo pealei and Illex illecebrosus in the northwest Atlantic. Fish. Bull. 97:9-24. Buckel, J. A., M. J. Fogarty, and D. O. Conover. 1999a. Foraging habits of bluefish, Pomatomus saltatrix, on the U.S. east coast continental shelf Fish. Bull. 97:758-775. 1999b. Mutual prey offish and humans: a comparison of biomass consumed by bluefish, Poiuatonius saltotrix, with that harvested by fisheries. Fish. Bull. 97:776-785. Caddy, J. F., and P. G. Rodhouse. 1998. Cephalopod and groundfish landings: evidence for ecological change in global fisheries. Rev. Fish. Biol. Fish. 8:431-444. Cargnelli, L. M., S. J. Griesbach, C. McBride, C. A. Zetlin, and W. W. Morse. 1999. Essential fish habitat source document: longfin inshore squid, Loligo pealei, life history and habitat characteristics. NOAA Technical Memo. NMFS-NE-146, 27 p. National Marine Fisheries Service, 166 Water St., Woods Hole, MA, 02543. Fahay, M. P., P. L. Berrien. D. L. Johnson, and W. Morse. 1999. Essential fish habitat source document: blue- fish, Pomatomus saltatrix. life history and habitat characteristics. NOAA Technical Memo. NMFS-NE-144, 68 p. National Marine Fisheries Service, 166 Water St.. Woods Hole, MA, 02543. Gerking, S. 1994. Feeding ecology of fish, 416 p. Academic Press, Inc., San Diego, CA. Garrison, L. P., and J. S. Link. 2000. Diets of five hake species in the northeast United States continental shelf ecosystem. Mar. Ecol. Prog. Ser. 204:243-255. Hatfield, E. M. C, and S. X. Cadrin. 2002. Geographic and temporal patterns in size and maturity of the longfin inshore squid iLoligo pealei} off the northeastern United States. Fish. Bull. 100:200-213. Hunsicker, M. E. 2004. Linking size-limited predation and the patterns of piscivory of the longfin inshore squid, Loligo pealei, in the mid-Atlantic continental shelf ecosystem. M.Sci. thesis, 49 p. Stony Brook Univ., Stony Brook, NY. Jackson, J. B. C, M. X. Kirby W. H. Berger, K. A. Bjorndal, L. W. Botsford, B. J. Bourque, R. H. Bradbury, R. Cooke, J. Erlandson, J. A. Estes, T. P. Hughes. S. Kidwell, C. B. Lange, H. S. Lenihan, J. M. Pandolfi, C. H. Peterson. R. S. Steneck, M. J. Tegner, and R. R. Warner. 2001. Historical overfishing and the recent collapse of coastal ecosystems. Science 293:629-638. Juanes, F., and D. O. Conover. 1994. Rapid growth, high feeding rates, and early piscivory in young-of-the-year bluefish (Pomatomus saltatrixt. Can. J. Fish. Aquat. Sci. 51:1752-1761. Lange, A. M. T, and M. P. Sissenwine. 1983. Squid resources of the northwest Atlantic. In Advances in assessment of world cephalopod resources (J. F. Caddy, ed.), p. 21-54. FAO Fish. Tech. Pap. 231. Link, J. S., K. Bolles. and C. G. Milliken. 2002. The feeding offlatfish in the Northwest Atlantic. J. Northw. Atl. Fish. Sci. 30:1-17. Link, J. S., and L. P. Garrison. 2002. Changes in piscivory associated with fishing induced changes to the finfish community of Georges Bank. Fish. Res. 55:71-86. Macy W. K., and J. K. T. Brodziak. 2001. Seasonal maturity and size at age o{ Loligo pealeii in waters of southern New England. ICES J. Mar. Sci. 58:852-864. May R. M., J. R. Beddington, C. W. Clark, S. J. Hoh, and R. M. Laws. 1979. Management of multispecies fisheries. Science 205:267-277. Murawski, S. A. 1991. Can we manage our multispecies fisheries? Fish- eries 16(5):5-13. Overholtz, W. J., J. S. Link, and L. E. Suslowicz. 2000. Consumption of important pelagic fish and squid by predatory fish in the northeastern LISA shelf eco- system with some fishery comparisons. ICES J. Mar. Sci. 57:1147-1159. Packer, D. B., S. J. Griesbach, P. L. Berrien, C. A. Zetlin, D. L. Johnson, and W. W. Morse. 1999. Essential fish habitat source document: summer flounder, Paralichthys dentatus, life history and habitat characteristics. NOAA Tech. Memo. NMFS-NE-151, Staudinger: Predation by four fish predators on two squid species on tfie Nortfiwest Atlantic continental sfielf 615 88 p. National Marine Fisheries Service, 166 Water St.. Woods Hole, MA, 02543. Pauly, D.. V. Christensen, J. Dalsgaard, R. Froese, and F. Torres Jr. 1998. Fishing down marine food webs. Science 279: 860-863. Pauly, D., V. Christensen, S. Guenettc, T. J. Pitcher, U. R. Sumaila, C. J. Walters, R. Watson, and D. Zeller. 2002. Towards sustainability in world fisheries. Nature 418:689-695. Piatkowski, U., G. J. Pierce, and M. Morals da Cunha. 2001. Impact of cephalopods in the food chain and their interaction with the environment and fisheries: an overview. Fish. Res. 52:5-10. Pierce, P. J., and A. Guerra. 1994. Stock assessment methods used for cephalopod fisheries. Fish. Res. 21:255-285. Quinn, G. P., and M. J. Keough. 2003. Experimental design and data analysis for biolo- gists, 537 p. Cambridge Univ. Press, Cambridge, UK. Smale, M. J. 1996. Cephalopods as prey. IV: Fishes. In The role of cephalopods in the world's oceans (M. R. Clarke, ed.), p. 1067-1081. Phil. Trans. R. Soc. Lond. B. 351, Staudinger, M. D. 2004. The importance of squid to the diets of bluefish, goosefish, silver hake, and summer flounder in the south- ern New England region. M. Sci. thesis, 84 p. Stony Brook Univ., Stony Brook, NY. Sokal, R. R., and F. J. Rohlf 1995. Biometry, 3'"'' ed., 887 p. W. H. Freeman and Co., New York, NY. 616 Evidence for resource partitioning and competition in nursery estuaries by juvenile flatfish in Oregon and Washington Christopher N. Rooper (contact author) Donald R. Gunderson David A. Armstrong School of Aquatic and Fisheries Sciences University ot Washington Box 355020 Seattle, Washington 98195 Present address (for C. N Rooper) National Marine Fisheries Service Alaska Fisheries Science Center 7600 Sand Point Way NE Seattle, Washington 98115-6349 Email address for C N. Rooper: Chris. Rooperfffinoaa gov Resource partitioning among co-oc- curring species can provide evidence that density-dependent survival can be a function of interspecific competi- tion. Of the early life history stages of flatfish, the juvenile phase is thought to be the most susceptible to density- dependent mortality, and intraspe- cific competition among individuals is thought to lead to decreased sur- vival of extremely strong year classes (lies and Beverton, 2000). Juveniles of multiple flatfish species are often observed to coexist in a single nurs- ery area (e.g., Burke et al., 1991; Hen- derson and Holmes, 1991; Reichert and van der Veer, 1991; Norcross et al., 1997), but evidence for interspe- cific competition is rarely examined. We examined evidence for resource partitioning and interspecific compe- tition among juvenile flatfish species co-occurring in nursery areas. Flatfish species often have juve- nile nursery areas that are segre- gated from the adult population. In many cases, juveniles use shallow nearshore habitats as nursery ar- eas, and adults are found in deeper offshore habitats (Gibson, 1994; van der Veer et al., 2000). On the Pacific coast of North America, estuaries are typically smaller and occur less fre- quently than in other areas such as the Atlantic coast of North America. Possibly because of the limited es- tuarine habitat on the West Coast, fewer species of marine fish and invertebrates use coastal estuaries for nursery areas. In our study, we examined the spatial distribution of co-occurring juvenile flatfish popula- tions within Pacific coast estuaries. The four species examined were Eng- lish sole iPleuronectes vetulus). Pacif- ic sanddab (Citharichthys sordidus), sand sole (Psettichthys melanostic- tiis), and starry flounder iPlatichthys stellatus). Juveniles of these species are found in Oregon and Washington estuaries during the summer, and adults of each of these species con- tribute to a lucrative trawl fishery on the coastal shelf. To date, estuarine residency of juvenile English sole has been well documented (Krygier and Pearcy, 1986; Gunderson et al., 1990; Roop- er et al., 2003); however, this is the first comparison of the spatial dis- tribution of the three other species within estuaries, and is the first study to encompass such a wide region of coastal estuaries (rang- ing over 400 km of coastline). The goal of this study was to compare the summer distribution of juvenile English sole, starry flounder, sand sole, and Pacific sanddab within nursery estuaries over three years, to determine whether the four flat- fish species were spatially partition- ing the nurseries. We also examined the summer distribution in relation to the abundance of species across years and habitats in order to deter- mine the potential for interspecific competition. Materials and methods Study area This study was carried out at four estuaries on the Oregon and Wash- ington coasts: Grays Harbor, Wil- lapa Bay, Coos Bay, and Yaquina Bay (Fig. 1). Seventeen to 20 trawl sites were chosen within each estu- ary, and each trawl site was classi- fied into three habitat types: lower main channel, lower side channel and upper estuary. Habitat delineations were previously identified by principle component and discriminant function analysis (Rooper et al., 2003). The lower estuarine sites were on average deeper (depth=5.1 m), colder (13. 6°C), and more saline (28.5 ppt) than upper estuarine sites because of their prox- imity to the mouth of the estuary. On average, lower side channel sites were slightly warmer (15.6°C), shallower (depth = 3.7 m) than the lower estu- arine sites, and had extensive tide flats that surrounded the sites and were exposed during low tides. Upper estuarine sites were typically shallow (depth = 4.0 m), very warm (17.0-C), less saline (25.1 ppt), and had a large sediment size and a smaller area of surrounding tidal flats (Rooper et al., 2003). Sampling procedures Trawl surveys of all four estuaries were conducted in both June and August from 1998 through 2000. In each month all four estuaries were sampled over two and a half weeks and a total of 431 trawl hauls were completed over the three summers. The same sites were resurveyed throughout all years of sampling; however the exact trawl location varied somewhat between sampling Manuscript submitted 8 September 2004 to the Scientific Editor's Office. Manuscript approved for publication 3 November 2005 by the Scientific Editor. Fish. Bull. 104:616-622(2006). NOTE Rooper et a\ Partitioning and competition among [uvenile flatfish in nursery estuaries 617 1 2S''0'0"W ! 2T(fO-V,' 1 26'0'0"W I I 1 i:5"0'iv'\v i:4"0'0"\v '■"^r^^ ' GiLfvs Harbor Pacific Ocean fl iJiiipa Bay Washinjiton '\-.jB^_ Yaqtnna Bay^ ^^ Oregon Coos Bay 50 100 I 200 km Figure 1 Map of coastal Oregon and Washington with insets showing the four estu- aries studied: Grays Harbor (total subtidal area=8545 ha). Willapa Bay (total subtidal area = ll,200 hal, Yaquina Bay (total subtidal area=471 ha), and Coos Bay (total subtidal area=1195 ha). periods by approximately 50 m. Benthic organisms, including the four target flatfish species, were collected during daylight with a beam trawl (Gunderson and Ellis, 1986). The beam trawl had an effective opening width of 2.3 m and a height of 0.6 m. The net was towed at a speed of 0.8-0.9 m/s behind a 6.4-m research vessel with a minimum 5:1 scope of line out to depth. The distance towed averaged 139 m in 1998-2000. Trawls were made against the prevailing current over a variety of tide stages, although efforts were made to conduct trawling near the time of daylight slack low water. Sampling was not conducted at high current velocities during maximum ebb and flood tides to ensure the greatest adherence of the beam trawl to the substratum. All fish and invertebrates captured were identified to species and English sole were measured for total length (TL) to the closest 1.0 mm. Other species of flatfish were measured (TL) only in 1999 and 2000. Data analysis There was considerable variability in the data associ- ated with each sampling period that was not related to habitat use. This variability was due to other factors, such as patterns in interannual variability in recruit- ment to individual estuaries or to differences in overall abundance among the four flatfish species. To minimize the effect of this variability on the analyses, it was 618 Fishery Bulletin 104(4) 0.6 0.4 - 0.2 -0.2 -0.4 - -0.6 1 — J> — 1 'n y k T - . ^ >< \ X^ -. \ -- 1 --\ \-^' \ \ — -\ " ""' f" L -. ^' ""i Lower main channel Lower side channel Upper estuary English sole Starry flounder -.*--- Pacific sanddab -■— - Sand sole Figure 2 Average density anomaly (and standard error bars) for Eng- lish sole {Pleuronectes vetulus). Pacific sanddab iCitharichthys sordidus), sand sole (Psettichthys melanostictus), and starry flounder iPlatichthys stellatus) by habitat type across all study estuaries, months, and years. necessary to standardize the data within each estuary. We used the normalized anomaly (Zar, 1974) of flatfish densities as the dependent variable to compare use of habitats within estuaries for each species, where sd„. In this equation the subscript i refers to an individual density from a trawl survey site within an estuary- month-year combination, emy refers to the estuary- month-year combination where a trawl survey occurred, and X and sd are the mean and standard deviation of all trawl surveys occurring in that estuary-month-year combination. By using the anomaly of the densities, we reduced the variation in the data that was attributed to estuarine, seasonal, and interannual sources in order to concentrate solely on the within-estuary component of variation. For example, if the average catch of Pacific sanddab was 1000/ha for the Grays Harbor estuary in June 2000, individual trawl hauls from Grays Harbor in June 2000 with catches <1000/ha would have a nega- tive density anomaly, whereas trawl hauls with Pacific sanddab catches >1000/ha would have a positive density anomaly. The density anomalies were analyzed by using analy- sis of variance (Zar, 1974). The full model used to ana- lyze spatial distributions of each species included the main effects of habitat type, month of survey, year of survey, and estuary, as well as estuary-habitat, month- habitat, and year-habitat interaction terms. The inter- action terms were specifically included to test the hy- potheses that spatial distribution of each flatfish species was consistent across estuaries, years, and months. All significant (P<0.05) main effects and interactions were compared for significant differences among effect levels by using the least squares mean approach as a post hoc test with a Tukey's adjustment for multiple comparisons (Zar, 1974). Results Juvenile Pacific sanddab collected during the trawl surveys ranged in length from 23 to 150 mm TL. The analysis of Pacific sanddab density anomalies for habitat types within estuaries resulted in significant differences among habitats in the estuaries. There were signifi- cantly higher density anomalies in lower main channel habitats and significantly lower anomalies in upper estuary sites (Fig. 2). Lower side channels usually had NOTE Rooper et al Partitioning and competition among luvenile flatfish in nursery estuaries 619 negative density anomalies — the most notable exceptions being Willapa Bay and Grays Harbor in 1998 (Fig. 3). This pattern produced significant interactions between habitat and both year and estuary. The lengths of sand sole captured by trawling were comparable to those for Pacific sanddab, ranging from 20 to 234 mm TL. and only 12 specimens longer than 150 mm TL. Sand sole catch was highly variable across all systems, and there were 318 zero catches over the total 431 trawls. Sand sole density anomalies followed a pat- tern similar to those of Pacific sanddab. Density anoma- lies were significantly higher at lower main channel sites than at other sites within the estuary (Fig. 2). Density anomalies were similar between upper estuary sites and lower side channel sites. There were no significant interaction terms between estuary or year and habitat type; thus the pattern in spatial distribution of sand sole appeared to be consistent across estuaries and years. Starry flounder were the least common species of flat- fish captured in the study estuaries. The starry flounder captured were larger than the other species of flatfish, ranging from 49 to 376 mm TL, and 44 individuals were >150 mm TL. Density anomalies were highest at upper main channel sites (Fig. 2). As with sand sole, there were no significant interaction terms between habitat type and either estuary or year. The effect of habitat type on starry flounder densities was thus a consistent effect observed across all estuaries and years. English sole lengths ranged from 14 to 200 mm TL and only 12 individuals were greater than 150 mm TL. English sole density anomalies were significantly higher at lower side channel sites than at other estuarine loca- tions (Fig. 2). There was also a significant interaction term between habitat type and month. In August there was no significant difference between density anomalies in lower side channel and lower main channel sites, while in June density anomalies were much larger in lower side channels than at all other locations. The significant month and habitat type interaction in the analysis of English sole density anomalies was most likely due to habitat changes observed in juvenile Eng- lish sole as they grow (Rooper et al., 2003). A carry- ing capacity for large (>50 mm TL) English sole was observed in the study estuaries, and large English sole distribution expanded when densities were high (Rooper et al., 2003). The expansion of English sole distribution when densities were large corresponds to an increase in the density anomalies of Pacific sanddab in lower main channel areas (Fig. 4). Conversely, as large (> 50-mm-TL) English sole densities decreased. Pacific sanddab appeared to increase their distribution in the estuaries, as observed in higher density anomalies in lower side channels. Discussion We observed evidence for resource portioning in four species of co-occurring juvenile flatfish in west coast nursery estuaries. The analysis of density anomalies Figure 3 Pacific sanddab iCitharnhtliys sordidus> density anomalies (and standard error bars) by estuary and habitat type from 1998 (upper panel) through 2000 (lower panel). indicated that the four flatfish species exploit different areas of estuaries. The spatial segregation among juve- nile flatfish species observed in our study is consistent with results of other studies and indicates the potential for interspecific competition among the four species. Sur- veys of flatfish assemblages in Puget Sound, Washington have revealed spatial segregation of juvenile flatfish by depth, although the pattern was not consistent among sites (Thornburgh, 1980). In Thornburgh's study, spa- tial segregation may have been driven by differences in sediment type at the sites and by differences in settling times among the dominant species: English sole, sand sole. Pacific sanddab, and rock sole (Lepidopsetta bilin- ecita). Distinct patterns in spatial distribution have also been observed for juvenile flatfish assemblages beyond the Pacific Northwest. In estuaries along the California 620 Fishery Bulletin 104(4) 1,5 1 □ Lower side channel • Lower main channel A Upper estuary • • " • • 1.0 • • • ^ .. • • • • . >^ • .-•'• • E 05 ■ D . ,...---• o c CO □ ..---■' D "D .-•■■• In*. « 0.0 ■ CO n ^ □ n~n -0.5- ,1 ..J -^^4 4__' A ▲ -1,0 \ 0 200 400 600 800 1000 1200 1400 1600 English sole density (no, /ha) Figure 4 Average Pacific sanddab iCitlianchthys ^ordidus) density anomaly compared to density of English sole {Pleuronectes vetulus) >50 mm TL at each location in the study from 1998 through 2000. Bight, diamond turbot {Hypsopsetta guttulata) settle in the upper portions of bays during January and March, whereas California halibut (Paralichthys californicus) are found closer to bay entrances and the open coast, settling from March to September (Kramer, 1991). Set- tlement timing has also been found to be an important partitioning factor for common dab (Limanda limanda) and plaice (Pleuronectes platessa) in a small Scottish bay, where plaice settle earlier in the year than dab (Steele and Edwards, 1970). Depth has been identified as an important factor in partitioning nursery areas for the common dab and age-0 plaice (Edwards and Steele, 1968; Gibson, 1973). Juvenile flatfish in Alaskan inlets have also been found to segregate with depth and a number of other factors, such as temperature and sediment type (Norcross et al., 1997). Habitat type has also been found important in determining the spatial segregation of northeast Atlantic flatfish assemblages; ontogenetic shifts in habitat use occur in some species (Walsh et al., 1999; Phelan et al., 2001). These studies indicate that flatfish species that ex- ist together in nursery areas can partition resources in a number of ways according to biological (i.e., diet) or physical (i.e., depth or salinity) requirements, or ac- cording to a combination of both biological and physical requirements. Studies of shallow water habitats have generally indicated that flatfish in nursery areas are not limited by food resources (Evans, 1983). In west coast estuaries food resources can be quite high, especially in intertidal areas (Gunderson et al., 1990; Dumbauld et al., 2000). After review of diet studies documenting feeding differences among Pacific sanddab and English sole (Hogue and Carey, 1982; Thornburgh, 1980), as well as the abundance of food within the nursery area, we do believe it is unlikely that direct competition for food is solely responsible for the apparent pattern in estuarine distributions between the two species. We suggest that the segregation observed in northwest es- tuaries is most likely caused by differences in physical conditions of the habitat. The four species of flatfish observed in the present study probably partition the nursery estuaries along gradients of salinity and temperature. Starry flounder were found in upper reaches of the estuary to a greater degree than the other species — a finding that is consis- tent with their ability to tolerate low salinity (Orcutt, 1950). Starry flounder were most abundant in Yaquina Bay, where the coverage of trawl survey sites extended farther up the estuarine salinity gradient than in other estuaries. The average salinity in June was 16-18 ppt at the uppermost Yaquina Bay sites, whereas the aver- age June salinity at sites in other estuaries was always greater than 19 ppt. Salinity has been found to be an important determinant of the distribution of other flat- fish species (Coggan and Dando, 1988; Marchand, 1988; Kerstan. 1991; Gibson, 1994; Marshall and ElHot, 1998). NOTE Rooper et al Partitioning and competition among [uvenile flatfish in nursery estuaries 621 The temperature preferences of English sole may allow them to exploit areas of the estuary not used exten- sively by Pacific sanddab. Temperatures less than 17.5°C have been observed to lead to high English sole growth (Yoklavich, 1982). These temperatures were observed in lower side channels of the study estuaries. For the speckled sanddab iCitharichthys sfigmacus), a close rela- tive of the Pacific sanddab. maximum growth has been observed from 8° to 13°C (Ehrlich et al.. 1979). This temperature range is closer to the temperatures found in lower side channels, and lower than the temperature range found in lower main channels in our study area. Thus, the Pacific sanddab, if it has temperature toler- ances similar to those of its congener, may be limited to lower main channel locations in order to optimize growth rates, whereas optimal English sole growth occurs at the higher temperatures found in lower side channels. An alternative explanation for Pacific sanddab distri- bution may be found in its response to high English sole density. The changing distribution of Pacific sanddab, when confronted by high densities of English sole, pro- vides evidence that competition may occur between these two species. When English sole densities are high, Pacific sanddab seem to be limited in distribution to lower main channel locations (sites not preferred by English sole), and when English sole densities are low. Pacific sanddab appear to be more evenly distributed throughout the estuary. This pattern does not apply to starry flounder or sand sole, whose distributions re- main the same regardless of English sole density. The implication of this resource partitioning is that English sole year-class strength may be affecting the ability of Pacific sanddab to use areas of the estuaries where conditions are optimal for flatfish growth. Distinct patterns of habitat use by juvenile English sole, starry flounder, sand sole, and Pacific sanddab were observed in all four estuaries that we examined. Highest densities of Pacific sanddab and sand sole were found in lower main channel sites, highest densities of English sole were found in lower side channels, and highest densities of starry flounder were found in upper estuarine sites. These patterns of spatial segregation were consistent from 1998 through 2000. The patterns were consistent across the four estuaries encompassing 400 km of coastline. Resource partitioning is probably based primarily on temperature and salinity tolerances and growth preferences of the individual flatfish spe- cies; however, there is evidence of competition between English sole and Pacific sanddab when their spatial distributions are examined in response to their relative abundances. Acknowledgments Assistance in the field provided by Tim Loher, Jennifer Boldt, Geoff Hosack, Grace Tsai, Noble Hendrix, Neil Banas and Kirsten Holsman was greatly appreciated. This research was funded by the National Oceanic and Atmospheric Administration's Coastal Ocean Program under award no. NA96OP0238 to the University of Washington. Literature cited Burke, J. S., J. M. Miller, and D. E. Hoss. 1991. Immigration and settlement pattern of Parali- chthys dentatus and P. lethostigma in an estuarine nursery ground. North Carolina, U.S.A. Neth. J. Sea Res. 27:393-405. Coggan, R. A., and P. R. Dando. 1988. Movements of juvenile Dover sole, Solea solea iL.), in the Tamar estuary, South-western England. J. Fish Biol. 33(suppl. A):177-184. Dumbauld, B. R., E. P. Visser, D. A. Armstrong. L. Cole-Warner, K. L. Feldman, and B. E. Kauffman. 2000. Use of oyster shell to create habitat for juvenile Dungeness crab in Washington coastal estuaries: status and prospects. J. Shellfish Res. 19:379-386. Edwards, R. R. C, and J. H. Steele. 1968. The ecology of 0-group plaice and common dabs at Loch Ewe: I. Population and food. J. Exp. Mar. Biol. Ecol. 2:215-238. Ehrlich, K. F, J. S. Stephens, G. Muszynske, and J. M. Hood. 1979. Thermal behavior of the speckled sanddab, Cithanch- thys stigmaeus: laboratory and field investigations. Fish. Bull. 76:867-872. Evans, S. 1983. Production, predation and food niche segregation in a marine shallow soft-bottom community. Mar. Ecol. Prog. Ser. 10:147-157. Gibson, R. N. 1973. The intertidal movements and distribution of young fish on a sandy beach with special reference to the plaice iPIeuronectes patessa L.). J. Exp. Mar. Biol. Ecol. 12:79-102. 1994. Impact of habitat quality and quantity on the recruitment of juvenile flatfishes. Neth. J. Sea Res. 32:191-206. Gunderson, D. R., D. A. Armstrong., Y. B. Shi, and R. A. McConnaughey. 1990. Patterns of estuarine use by juvenile English sole {Parophrys vetiilus) and Dungeness crab {Cancer magister). Estuaries 13:59-71. 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. Henderson, P. A., and R. H. A. Holmes. 1991. On the population dynamics of dab, sole and floun- der within Bridgewater Bay in the Lower Severn estuary, England. Neth. J. Sea Res. 27:337-344. Hogue, E. W., and A. G. Carey Jr. 1982. Feeding ecology of 0-age flatfishes at a nursery ground on the Oregon coast. Fish. Bull. 80:555-565. lies, T. C, and R. J. H. Beverton. 2000. The concentration hypothesis: the statistical evidence. ICES J. Mar. Sci., 57:216-227. Kerstan, M. 1991. The importance of rivers as nursery grounds for 0 and 1 group flounder {Platichthys flesus L.) in comparison to the Wadden Sea. Neth. J. Sea Res. 27:353-366. Kramer. S. H. 1991. The shallow-water flatfishes of San Diego county. CalCOFI Rep. 32:128-142. 622 Fishery Bulletin 104(4) Krygier. E. E., and W. G. Pearcy. 1986. The role of estuarine and offshore nursery areas for young English sole, Parophrys vetulus Girard, of Oregon. Fish. Bull. 84:119-132. Marchand, J. 1988. Seasonal distribution, growth, and ecological role of the juvenile sole, Solea solea L., populations in the Loire estuary, France. J. Fish Biol. 33:229-233. Marshall, S., and M. Elliot. 1998. Environmental influences on the fish assemblage of the Humber estuary, U.K. Estuar. Coast. Shelf Sci. 46:175-184. Norcross, B. L., F. Muter, and B. A. Holladay. 1997. Habitat models for juvenile pleuronectids around Kodiak Island. Alaska. Fish. Bull. 95:504-520. Orcutt, H. G. 1950. The life history of the starry flounder Plafichthys stellatus (Pallasl. Cal. Div. Fish and Game, Fish. Bull. 78, 64 p. Phelan, B. A., J. P. Manderson, A. W. Stoner, and A. J. Bejda. 2001. Size-related shifts in the habitat associations of the young-of-the-year winter flounder (Pseudopleuro- nectes americanus): field observations and laboratory experiments with sediments and prey. J. Exp. Mar. Biol. Ecol. 257:297-315. Reichert, M. J. M., and H. W. van der Veer. 1991. Settlement, abundance, growth, and mortality of juvenile flatfish in a subtropical tidal estuary (Georgia, U.S.A.). Neth. J. Sea Res. 27:375-391. Rooper, C. N., D. R. Gunderson, and D. A. Armstrong. 2003. Patterns in use of estuarine habitat by juvenile English sole (Pleuronectes vetulus) in four eastern north Pacific estuaries. Estuaries 26:1142-1154. Steele, J. H., and R. R. C. Edwards. 1970. The ecology of 0-group plaice and common dabs in Loch Ewe: IV. Dynamics of the plaice and dab populations. J. Exp. Mar. Biol. Ecol. 4:174-187. Thornburgh, K. 1980. Patterns of resource utilization in flatfish communities. M.S. thesis, 90 p. Univ. Washington, Seattle, WA. van der Veer, H. W., R. Berghahn, J. M. Miller, and A. D. Rijnsdorp. 2000. Recruitment in flatfish with special emphasis on North Atlantic species: progress made by the Flatfish Symposia. ICES J. Mar. Sci. 57:202-215. Walsh, H. J., D. S. Peters, and D. P. Cyrus. 1999. Habitat utilization by small flatfishes in a North Carolina estuary. Estuaries 22:803-813. Yoklavich, M. 1982. Growth, food consumption, and conversion effi- ciency of juvenile English sole tParophrys vetulus). In Fish food habit studies, proceedings of the third Pacific workshop (G. M. Caillet and C. A. Simenstad, eds.), p. 9-105. Univ. Washington, Seattle, WA. Zar, J. H. 1974. Biostatistical analysis, 620 p. Prentice-Hall Inc., Englewood Cliffs, NJ. 623 Length-specific brood size and winter parturition in pink seaperch iZaiembius rosaceus) (Perciformes: Embiotocidae) Lori H. LaPlante California State University Long Beach Department of Biological Sciences 1250BellflowerBlvd Long Beach, California 90840 Present address: Saint Anselm College Department of Biology 100 Saint Anselm Drive Mancfiester, New Hampshire 03102 Email address llaplante a anselm edu The viviparous pink seaperch iZalem- bius rosaceus) is considered unusual among other embiotocids because of its deep-water habitat (Tarp. 1952), winter parturition (Goldberg and Tic- knor, 1977), small brood size (Baltz, 1984), and apparent lack of a posi- tive relationship between brood size and female size I Baltz, 1984). In the present study, I present further sup- port that parturition in pink seaperch occurs in winter and new evidence indicating length-specific brood sizes in the species. Pink seaperch are found in deep water (to 229 m: Miller and Lea, 1972) over soft bottom (Allen, 1982) year-round off the coasts of Califor- nia and Baja California (Eschmeyer et al., 1983). Mating occurs in spring and parturition occurs in winter (Goldberg and Ticknor, 1977)— a unique breeding schedule among surfperches. Other surfperches mate in summer or fall and give birth the following spring or summer when food for offspring is plentiful. For reasons that remain unclear, brood size for pink seaperch is considered low (mean = 3.5: Goldberg and Tic- knor, 1977) compared to other small- size (<160 mm) surfperches (reviewed in Baltz, 1984). A positive relationship between brood size and female size is gener- ally the rule among fishes (Bagenal, 1978) and is observed in all surf- perches with the exception of pink seaperch (Baltz, 1984). However, data for the species are dubious because near-term females tend to abort their young during the time of collection (Baltz and Knight, 1983; Baltz, 1984) — a behavior that can re- sult in both underestimates in brood sizes and high variability in length- specific brood size. The goal of this study was to eval- uate characteristics of the reproduc- tive biology of pink seaperch using samples taken during periods both early and late in the breeding sea- son. I predicted that estimates in length-specific brood size for females collected later in the season should be significantly smaller than those collected earlier if, in fact, near-term females have a high probability of aborting embryos. Similarly, I ex- pected to see higher variability in relationships between length-specif- ic brood size and female size in the late, compared to early, period of the breeding season. Materials and methods Specimens of female pink seaperch were collected with a 7.6-m (head- rope) otter trawl with 1.3-cm codend mesh on the Palos Verdes shelf near Point Fermin, CA (33°41'N, 118°19'W) in fall (September and October, at 54 and 60 m depth, respectively) 1994 and in winter (January, at 66 m depth) 1995. Females were distin- guished from males by having either a rounded belly or by lacking a fleshy reproductive organ on the anterior portion of the anal fin (Tarp, 1952). Gravid females collected in winter were stored separately in sealable bags to ensure that any embryos sub- sequently born were included in brood size records; females collected in fall had no evidence of premature births nor were there any premature embryos present in fall collections (see Results section). Individuals were placed on ice, frozen, and processed within 10 days. Females were dissected and the following information was recorded: standard length (SL ±1 mm), body mass (±1 g), brood size, embryo SL (±1 mm), embryo mass (±0.01 g), and evidence of premature birthing. From the appearance of females that had evidently expelled their broods while in storage bags, I recorded a female as having prematurely given birth if her ovary was devoid of embryos, flaccid, and contained traces of fresh blood. The relationships between female and brood size for fall and winter were analyzed separately by using regres- sion analyses (SPSS, vers. 10, SPSS, Inc., Chicago, IL). Females showing evidence of recent expulsion were not included in brood size analyses. Results The size distribution of females cap- tured varied between seasons (Fig. 1). The mean standard length (SL) of females in fall and winter was 125 mm (±15 mm SD, n=21) and 96 mm (±21 mm SD, ;j=76), respectively, and this difference was significant (/-test: ? = 6.82, df=101, P<0.001). All females greater than or equal to 100 mm SL (;?,,,, 1 = 26, "„.inttr=31' were sexually mature, having either embryos or evi- dence of recent expulsion of embryos. Among the mature females, the mean SL in fall and winter was 127 mm (±14 mm SD, ;!=26) and 117 mm (±9 mm SD, a? = 31), respectively, and this difference was also significant (f-test; ^=3.19, df=55, P=0.002). Both mean brood size and the re- lationship between female size and brood size varied between seasons (Fig. 2A). Mean brood sizes in fall and winter were seven (±3 SD) and three (±1 SD), respectively, and this Manuscript submitted 21 March 200.5 to the Scientific Editor's Office. Manuscript approved for publication 21 November 2005 by the Scientific Editor. Fish. Bull. 104:623-62512006). 624 Fishery Bulletin 104(4) 30 20 10 ■ Fall 1994 □ Winter 1995 JI 2 4 6 8 10 12 14 Size class (cm) Figure 1 Size distribution for mature female pink sea- perch iZalembius rosaceus) (n=10.3) collected in fall 1994 (black bars) and winter 1995 (white bars). Bars represent 1-cm size classes. difference was significant (^test: ? = 5.15, df=55, P<0.001). The incidence of premature births for winter females was 18% (seven of 38 females) compared to 0% (0 of 27 females) for fall females, and differed signifi- cantly (Fisher's exact test: P=0.04). Female size was a good predictor of brood size in fall but not in winter (fall: 7-2 = 0.39; winter: r- = 0.00, Fig. 2A). Embryos were more developed and larger in win- ter than fall, and larger females brooded larger-sized embryos (Fig. 2B). Fall embryos had a mean SL of 21.3 mm (±4.2 mm SD) and were characterized by pale pink coloration, translucent tissue, and the absence of distinct scales. Winter embryos had a mean SL of 36.4 mm (±5.8 mm SD) and had firm and opaque tissue, distinct scales, iridescent pink coloration, and appeared as miniature adults. In both fall and winter, linear regressions of embryo-on-female size were significant (fall: r'-=0.17; winter: 7-2=0.20, Fig. 2B). Discussion Both embryo characteristics and female reproductive condition provided evidence of winter parturition in pink seaperch and confirmed conclusions made by Gold- berg and Ticknor (1977). Winter embryos had advanced morphological characteristics similar to those of adults, compared to embryos examined in fall, and were similar in size to those reported by Goldberg and Ticknor ( 1977 ). Evidence of a relatively high incidence of abortion and presence of many embryos (77 = 51) in winter collections provided further evidence that parturition occurs in or around January. Winter parturition is unique among surfperches yet no studies, to the authors knowledge, have provided an explanation for the unusual breeding schedule of pink seaperch. Results from this study demonstrate a positive re- lationship between brood size and female size in pink 14 T 12 10 N O u) •o 5 O O m '^ 2 0 50 45 ? -"^ § 35 <^ 30 o >, ^ 25-1 £ UJ 20 15 A Fall O Winter Baltz (1984) ««iO QO-C O u IHJ.W5 lOSJ ^^O O OO A a AO o oo O A A A 10 B 90 100 110 120 130 140 150 160 170 Female SL (mm) Figure 2 Linear regressions of I A) brood size (fall: r- = 0.39, brood size = -l'2.9 + 0.14.r, ;i=26; winter: r-= 0.00, ;( = 31 ), and (B) embryo size (fall: r- = 0.17, embryo SL = 4.9 + O.lS.v; winter: /•■- = 0.20, embryo SL = 4.2 + 0.28.V) on female size for fall (triangles) and winter (circles) collections of pink seaperch (Zalembius rosaceus). Dashed line in A represents data from Baltz (1984). seaperch, contrary to results of previous reports. Brood size increased with female size in fall but not in winter (Fig. 2A) in this study. The lack of a relationship observed in winter was similar to previous conclusions (See Fig. 2 in Baltz, 1984). Baltz and others (Goldberg and Ticknor, 1977; Baltz and Knight, 1983) speculated that sample effects may have explained the lack of a relationship between brood size and female body size, but no clear evidence was provided. In the present study, there was a high incidence of females (18%) that had aborted all embryos (i.e., complete abortion) among winter collections, and these females were excluded from analyses. However, partial abortions may occur in embiotocids (Schultz, 1993) presenting the possibility of underestimating brood size. Females with partially aborted broods were not identified nor excluded from the present study. Therefore, if such behavior occurred among large winter females, it would provide a reasonable explanation for the lack of relation- ship observed between brood size and female size. Winter females had larger brood sizes than their fall counterparts and a higher frequency of abortion, pro- viding evidence that winter females had partial rather than complete broods. Individuals smaller than the NOTE LaPlante Brood size and parturition in Zalembius rosaceus 625 mesh size of the net were present in winter collections, probably aborted by females stressed during capture. The number of embryos (/! = 51) present in the winter collections was higher than would be expected if they had been aborted by females having shown evidence of complete abortions (n=7). Therefore, pink seaperch have length-specific brood sizes as observed in other surfperches, and this finding is supported by the more reliable fall data from the present study. Mean brood size for pink seaperch in the present study was higher than previously reported for the species. Goldberg and Ticknor (1977) reported that females had an average brood size of 3.5, which was similar to mean brood sizes obtained for winter collections in this study; however, according to arguments already presented, win- ter data likely resulted in an underestimation of brood size. Both mean and maximum fall brood sizes (max=10) were greater than winter brood sizes and greater than the brood size in samples examined in Goldberg and Ticknor 's (1977) study. If brood size estimates in the Goldberg and Ticknor (1977) study included females collected in winter, the authors probably underestimated brood size, as well. The mean brood size of pink seap- erch is similar to that of other deep water fishes (Koslow et al., 2000) but is considered relatively small compared to other embiotocids (reviewed in Baltz, 1984). An unexpected result from the present study was evi- dence that large female pink seaperch have an earlier breeding schedule than small females. Reproductively active females in fall were larger on average than win- ter females (Fig. 1), indicating that large females left breeding sites earlier or that small females arrived at breeding sites later. Although greater temporal resolu- tion of changes in size distribution would have been desirable, further evidence for size-dependent breeding was apparent from 'female- and embryo-size relations. In both fall and winter, embryos in large females were larger than embryos in small females (Fig. 2B), which would be expected if the reproductively active females in fall bred earlier than those in winter. An alterna- tive explanation for the positive relationship between female size and embryo size would be that embryos in large females develop faster than those in small females because of greater maternal investment. I did not have sufficient data on embryo development because this was not the focus of my study; however, it is a hypothesis that warrants further investigation. Delayed breed- ing by smaller females of pink seaperch is a pattern observed in other surfperches (Eigenmann and Ulrey, 1894; Hubbs, 1921; Schultz, et al., 1991) and may arise from both energetic limitations on the time of repro- duction and fitness advantages accrued by postponing reproduction and diverting additional energy towards growth (Schultz, et al., 1991). Acknowledgments Fish were collected in accordance with permits issued to California State University at Long Beach, South- ern California Coastal Water Research Project, and the County Sanitation Districts of Los Angeles. This work was supported by the California State University at Long Beach (CSULB) and Southern California Coastal Water Research Project (SCCWRP). R. Flores, T. McSparren, and K. Mickey (all of CSULB) provided invaluable discus- sions and laboratory and field assistance. I am grateful to M. J. Allen (SCCWRP), J. Cross (SCCWRP), K. Mickey, E. T. Schultz, and two anonymous reviewers for comments and suggestions on earlier drafts of this manuscript. Literature cited Allen, M. J. 1982. Functional structure of soft-bottom fish communi- ties of the southern California shelf. Ph.D. diss., 603 p. Univ. California, San Diego, La Jolla, CA. Bagenal, T. B. 1978. Aspects of fish fecundity. /» Ecology of freshwater fish production (S. D. Gerking, ed.), p. 75-101, Wiley and Sons, New York, NY. Baltz, D. M. 1984. Life history variation among female surfperches (Perciformes: Embiotocidae). Environ. Biol. Fish. 10: 159-171. Baltz, D. M., and E. E. Knight. 1983. Age-growth reproductive characteristics and seasonal depth distribution of the spotfin surfperch, Hyperprosopon anale. California Fish and Game 69: 97-104. Eigenmann, C. H., and A. B. Ulrey. 1894. A review of the Embiotocidae. Bull. U.S. Fish. Comm. 12:382-400. Eschmeyer, W. N., E. S. Herald, and H. Hammann. 1983. A field guide to Pacific Coast fishes of North America. Houghton Mifflin Company, Boston. MA. Goldberg, S. R., and W. C. Ticknor. 1977. Reproductive cycle of the pink seaperch, Zalembius rosaceus (Embiotocidae). Fish. Bull. 75:882-884. Hubbs, C. L. 1921. The ecology and life-history of Amphigonop- terus aurora and of other viviparous perches of Cali- fornia. Biol. Bull. (Woods Hole). 40:181-209. Koslow, J. A., G. W. Boehlert, J. D. M. Gordon, R. L. Haedrich, P. Lorance, and N. Parin. 2000. Continental slope and deep-sea fisheries: impli- cations for a fragile ecosystem. ICES J. Mar. Sci. 57: 548-557. Miller, D. J., and R. N. Lea. 1972. Guide to coastal marine fishes of California. Fish. Bull. 157:1-235. Schultz, E. T. 1993. Sexual size dimorphism at birth in Microme- trus minimus (Embiotocidae): a prenatal cost of reproduction. Copeia 1993:456-462. Schultz, E. T., L. M. Clifton, and R. R. Warner. 1991. Energetic constraints and size-based tactics: the adaptive significance of breeding-schedule variation in a marine fish (Embiotocidae: Micrometrus minimus). Am. Nat. 138:1408-1430. Tarp, H. F. 1952. A revision of the family Embiotocidae (the surf- perches). Fish. Bull. 88:4-99. 626 Aging fish otoliths recovered from Pacific harbor seal iPhoca vitulina) fecal samples Susan D. Riemer (contact author) Oregon Department of Fish and Wildlife 1495 East Gregory Road Central Point, Oregon 97502 Email address: SusanDRiemeriS'stateorus Robert Mikus Oregon Department of Fisfi and Wildlife 4412 Silverton Road NE Salem, Oregon 97305 Seals and sea lions are opportunis- tic predators that feed on a variety of fish and cephalopods, including some commercially and recreation- ally important species. Concerns over their interactions with commercial and sport fishing operations, and other human activities, has a long his- tory in the Pacific Northwest (Ever- itt and Beach, 1982). The perceived increase in such interactions led, in part, to the U.S. Congress amending the Marine Mammal Protection Act (MMPA) m 1994. Chief among the amendments was a call for research to determine 1) whether California sea lions (Zalophus californianus) and Pacific harbor seals ^Phoca vitulina) were affecting the recovery of listed or depleted salmonids, and 2) what broader impacts they may have on the coastal ecosystems of Oregon, Wash- ington, and California (NMFS, 1997). In this note, we begin to address the latter question. Specifically, we describe a novel application of otolith aging techniques that can be used to increase the understanding of pin- niped foraging ecology and, thus, their potential impact on the fishery resources of coastal ecosystems. The Oregon Department of Fish and Wildlife's (ODFW) Marine Mam- mal Research Program has studied harbor seal foraging habits since the mid-1980s, primarily through the collection and analysis of scat (fecal) samples (Riemer and Brown, 1997; Riemer et al.^; Wright et al.-). Our results, as well as those from other pinniped food habit studies in Oregon (Graybill, 1981; Brown and Mate, 1983; Roffe and Mate, 1984; Orr et al., 2004), indicate that these animals consume a large number of diverse prey species. Concurrent research in Oregon indicates that Pacific harbor seals have increased significantly following protection un- der the MMPA; Brown et al. (2005) estimated that the 2002 statewide population total was 10,087 indi- viduals. As noted above, these types of increases in pinniped abundance have led to more frequent interac- 1 Riemer, S. D., R. F. Brown, B. E. Wright, and M. Dhruv. 2001. Monitoring pinniped predation on salmonids at Alsea River and Rogue River. Oregon: 1997-1999. Unpubl. Contract Rep. to Pacific States Marine Fisheries Com- mission, NCAA Grant No. NA87FX0464. 38 p. [Available from OR Dept. Fi.sh. and Wild!., 7118 NE Vandenberg Ave., Corvallis, OR 97330.1 - Wright, B. E., R. F. Brown, S. D. Riemer, and A. M. Ougzin. 2002. Pinniped pre- dation on adult salmonids in the Alsea Estuary, Oregon. Unpubl. Contract Rep. to Pacific States Marine Fisheries Com- mission, NOAA Grant No. NA17FX1603, 35 p. lAvailable from OR Dept. Fish, and Wild]., 7118 NE Vandenberg Ave., Corvallis, OR 97330.] tions with coastal fish resources and hence an increased interest in the composition and abundance of prey in their diet. Fortunately, new techniques have recently been developed to analyze pinniped diets. For example, rather than using traditional methods that rely strictly on otoliths (fish ear bones) for identifying prey, many researchers now try to identify all skeletal structures recovered from scats, which provide a more complete picture of pinniped diets (Olesiuk et al., 1990; Cottrell et al., 1996; Riemer and Brown, 1997; Browne et al., 2002). In addition, research- ers are beginning to use molecular genetic methods to provide greater resolution in determining diet com- position (Purcell et al., 2004; Deagle et al., 2005; Kvitrude et al, 2005). In this note, we add to this growing list of techniques by describing a novel use of fish otolith aging techniques to further our understanding of pin- niped diets. Fisheries scientists have aged fish otoliths to aid in the management of commercial and recreational fisheries (Love et al., 2002). Marine mammal scientists, on the other hand, have re- lied upon the identification of otoliths recovered from scats to identify the prey of pinnipeds (Brown and Mate, 1983; Beach et al.^; Harvey, 1989; Pierce and Boyle. 1991). However, the age of prey has rarely been consid- ered when describing seal and sea lion diets. We believe that estimates ■'Beach, R. J., A. C. Geiger, S. J. Jeffries, S. D. Treacy, and B. L. Troutman. 1985. Marine mammals and their interactions with fisheries of the Columbia River and adjacent waters, 1980-1982. NWAFC ( Northwest Alaska Fisheries Science Center) processed rep. 85-03, 316 p. NWAFC, National Marine Fisheries Service, Northwest Region, 7600 Sand Point Way N.E., Seattle, WA 98115-0070. Manuscript submitted 1 June 2005 to the Scientific Editor's Office. Manuscript approved for publication 29 November 2005 by the Scientific Editor Fish. Bull. 104:626-630 i2006l. NOTE Riemer and Mikus: Aging fish otoliths recovered from fecal samples of Phoca vitulina 627 of the age of prey will result in a more comprehensive picture of pinniped diets and help to augment stock as- sessments that use age-specific models. We report our application of these techniques to Dover sole (Micros- tomus pacificiis) otoliths recovered from Pacific harbor seal scat samples collected in an Oregon estuary. Dover sole were selected as a case study because they are a common prey of harbor seals in Oregon (Riemer and Brown, 1997) and their otoliths have been aged success- fully in previous studies (Hagerman. 1952; Brodziak and Mikus. 2000). Materials and methods We conducted our study during the spring and summer of 1996. and year-round from 1997 through 2002, in the Alsea Estuary located near Waldport, Oregon (44°26'N, 124°3'W). The local harbor seal population in this area consisted of approximately 600 animals throughout the study period. Scat samples from this population were obtained during low tides by approaching haul-out areas on foot or by boat, and slowly moving the animals into the water. Samples were placed in individually labeled plastic bags and frozen. The number of scat samples col- lected during each trip varied depending on the number and location of animals hauled out, and on weather and ocean conditions. Scats were thawed and partially dissolved in water, then rinsed through a series of nested sieves (2 mm, 1 mm, 0.71 mm). All prey hard parts (e.g., otoliths, bones) recovered were dried and placed in individually labeled jars. Prey species were identified from all prey hard parts recovered from each sample. Dried hard parts were examined under a dissecting microscope and identified by using a comparative collection of fishes from the northeast Pacific Ocean and Oregon estuaries. Otoliths and diagnostic bones were identified, counted, and the side (left or right) was noted to estimate a minimum number of individuals (MNI) represented in each sample by following the procedures described by Lance et al.'' Otoliths selected for aging were given individual iden- tification numbers and stored in gelatin capsules. One sample with 65 similarly size otoliths was subsampled by randomly selecting the first 18 otoliths recovered. Before aging, otoliths were measured to the nearest mm for total length, width, and length of sulcus by using an optical micrometer. Degree of erosion (level 1 having the least amount of erosion and level 3 having the greatest) was recorded by following techniques described in Tollit et al. (1997). Reference photographs were taken of each ^ Lance, M. M., A. J. Orr, S. D. Riemer. M. J. Weise, and J. L. Laake. 2001. Pinniped food habits and prey identifica- tion techniques protocol. AFSC (Alaska Fisheries Science Center) Proc. Rep. 2001-04, 36 p. Alaska Fisheries Science Center. NMFS, NOAA, 7600 Sand Point Way NE, Seattle, WA 98115. otolith with a dissecting microscope and digital camera before age estimates were determined. Final otolith ages were determined by using one of two techniques. First, all fish otoliths were submerged in a dish of ethanol with a black background and an- nuli were counted under a dissecting microscope with reflected light (i.e., surface aged). Second, fish a four years were re-aged according to the method described in Pikitch and Demory (1988) (i.e., break-and-burn method). The first annulus deposited was determined to be the completion of growth for the first year follow- ing the convention for aging adult Dover sole (Chilton and Beamish, 1982). In some cases, a mark within the otolith core was counted as the initial increment because it met the identification criteria for an annual increment. Results Dover sole remains (bones and otoliths) were recovered from 296 of the 3370 harbor seal scat samples collected during the study period (Table 1). Dover sole otoliths were recovered from 132 of these scat samples. Eighty- nine of these otoliths (21%) were excluded from our analyses because of poor condition (erosion or break- age), extremely small size, or because they had been randomly subsampled. The majority of otoliths were moderately eroded (level 1: 16%, level 2: 64% and level 3: 20%). Of the 339 otoliths analyzed, 98.2% were assigned an age; 71.2%. (237) by surface aging and 28.8%- (96) by the break-and-burn method. Six otoliths were excluded from age estimates because of extreme edge wear and erosion. Forty-nine (92.5%) of the scats with more than one otolith analyzed resulted in fish of multiple ages. The majority (70.6%) of otoliths analyzed were from one-, two-, and three-year-old fish; the rest were from four to six year-old (25.8%) and seven to 12 year-old (3.6%) (Table 1). Of the juvenile fish otoliths recovered, the presence of one- and two-year-old fish peaked in July, and three-year-old fish were found most frequently in August. Three of the four oldest fish otoliths were re- covered during the summer months, and one 10-year-old fish was recovered in May. The scat samples collected during fall (September-November) primarily included two- and three-year-old fish (72.5%). Aging otoliths of Dover sole increased MNI in harbor seal scats by 13% (269 versus 238). In most cases the number of Dover sole increased by one fish per scat when using both age and otolith side to enumerate MNI. Discussion We have shown in the present study that 1) Dover sole otoliths recovered from harbor seal scats can be successfully aged, and 2) seals in the Alsea River con- sumed Dover sole that ranged in age from one to 12 years. Interestingly, the highest frequency of occur- 628 Fishery Bulletin 104(4) Table 1 Age dit tribution for Dover sole {Microstoinus pacificus) consumed by harbor seals iPh r>ca V tiilina ), based on otoliths recovered from scats collected in the Al sea Estuary, Orego T. Total number of scats collected (n) scats with Dover sole bone s and otoliths | («-Dover), and otoliths only [n -otoliths), and the lumber of Dover sole otoliths recovered (n) and successfully aged ((!- iged) were pooled monthly over all years from 1996 through 2002. Month Scat Otolith Age ( y r ) '/f' n n -Dover n -otoliths n (!-aged 1 2 3 4 5 6 7 8 9 10 11 12 Jan 33 0 0% Feb 24 1 1 3 3 1 2 0.9% Mar 103 2 0% Apr 212 12 1 0 0 0% May 230 22 11 12 11 6 2 2 1 3.3% Jun 423 26 16 38 35 4 9 9 4 5 1 2 1 10.5% Jul 347 91 40 174 118 26 32 23 18 11 6 1 1 35.4%' Aug 330 65 32 102 97 11 23 35 15 7 2 3 1 29.1% Sep 806 54 21 65 55 1 23 17 7 4 2 1 16.5% Oct 387 21 8 13 10 1 5 3 1 3.0% Nov 173 2 2 20 4 2 1 1 1.2% Dec 0 0% Total 3370 296 132 428 333 43 99 93 45 30 11 3 5 1 2 0 1 333 ' Percentage of otoliths aged by month. rence for Dover sole in our collections (ODFW'^), and the greatest range in their age (Table 1), occurred during summer — a period when Dover sole move into shallow nearshore waters (Hagerman, 1952; Markle et al., 1992). This migration into nearshore waters brings fish into contact with an increasing population of pin- niped predators. The majority of Dover sole otoliths recovered from scats were those of fish younger than four years. In contrast, Dover sole landed in Oregon's commercial fishery between 1984 and 2004 were aged to be be- tween three and 60 years old and the average age was 13 years (ODFW^). Of the 45,026 fish aged from the commercial fishery, over 50% were ages similar to those recovered from seal scat samples (four to 12 years) and had an average length of 33.4 cm. Of the fish checked, 79% were identified as mature. However, the majority (70.5%) of otoliths analyzed in our study were from one, two-, and three-year-old fish, whereas only one of the commercially caught fish was in this age range. Four- to six-year-old fish made up 0.3% of the commercially sampled fish and seven- to 12-year- olds made up the largest portion at 53.5%. This lack of younger age fish in the commercial fishery is partly due to the discarding of smaller unmarketable fish at sea. Sampson and Wood (2001) reported that from 1956 through 2000, between 5% and 14.6% of the commer- cial Dover sole catch at sea was discarded because of the small unmarketable size of the catch. What effect a consistent removal of older age fish and the discarding of younger fish by the commercial fishery, coupled with the take of younger age fish by an increasing number of predators could have on the population as a whole, is unknown. Many species of marine fishes other than Dover sole contribute to important commercial fisheries in Oregon (ODFW"). Estimating prey age from fish otoliths recov- ered from seal and sea lion scats will be useful for re- searchers when describing impacts on long-lived marine species such as rockfish, which may be at greater risk to this long-term and increasing predation by coastal pin- nipeds. Love et al. (2002) reported that rougheye rock- fish (Sebastes aleutianus) live to at least 205 years and yelloweye rockfish (S. ruberrimus) to 118 years, whereas other species such as calico (S. dalli) and squarespot (S. hopkinsi) may only live for a decade or two. These older fish in some cases also have a late maturity; for example, only half of all yelloweye rockfish are mature at an age of 22 years. As the sustainability of these species and the fisheries that target them become difficult to manage, there will be more interest in the impact of predators on these fish populations. Determining fish age has enabled biologists ' ODFW (Oregon Department of Fish and Wildlife). 1984-2004. Unpubl. data. [Data are on file at Oregon Department of Fish and Wildlife Newport-Marine Program office, 2040 S.E. Marine Science Drive, Newport, Oregon 97365.] s ODFW (Oregon Department of Fish and Wildlife). 2000. Or- egon Marine Fisheries 2000 status report, 109 p. [Available from Oregon Department of Fish and Wildlife Newport- Marine Program office, 2040 S.E. Marine Science Drive, Newport, Oregon 97365.1 NOTE Riemer and Mikus Aging fish otoliths lecoveied from fecal samples of Phoca vilulina 629 to more accurately estimate appropriate fishing levels, particularly for rockfish species (Love et al., 2002). This information can also contribute to assessing the impact of pinnipeds on these species and can improve stock assessment models. Otolith age can provide informa- tion on the size of fish consumed when the age-length composition in the population or subgroup is available (Salthaug, 2003). In addition, age estimates provide information on the reproductive maturity of the prey consumed. Aging will also provide a more accurate esti- mate of the MNI of a particular prey species consumed than an estimate based on the maximum number of left or right otoliths of a similar size, particularly because different age fish can have otoliths of approximately the same size. Researchers should note that the digestive process does cause erosion of otoliths and can result in under- estimation of fish ages. We did not attempt to deter- mine how otolith-based age estimates were affected by erosion, but a carefully designed captive feeding study would be able to address this issue in the future. How- ever, because there are few other methods to determine the age distribution of fish taken by pinnipeds, this technique is useful when describing the diet of seals and sea lions. Acknowledgments Funding for this project was provided by the National Marine Fisheries Service, Northwest Regional Office (with the assistance of the Pacific States Marine Fisher- ies Commission) and the Oregon Department of Fish and Wildlife, Marine Resources Program. We thank Bryan Wright, Robin Brown, and Monique Lance for their sup- port of this project and the editing of an earlier draft of the manuscript. We thank Patty Burke and Mark Sael- ens for their input and support. We extend our thanks to our field staff Aicha Ougzin, Mark Dhruv, and Jonathon Scordino. We also thank Mark Karnowski and William Miller for their help with data and revisions. Finally, we thank the two anonymous reviewers who provided comments on this manuscript. Literature cited Brodziak, J., and R. Mikus. 2000. Variation in life history parameters of Dover sole, Microstomas pacificus, off the coasts of Wash- ington, Oregon, and northern California. Fish. Bull. 98:661-673. Brown, R. F., and B. R. Mate. 1983. Abundance, movements and feeding habits of harbor seals, Phoca vitulina, at Netarts and Tillamook Bays, Oregon. Fish. Bull. 81:291-301. Brown, R. F, B. E. Wright, S. D. Riemer, and J. Laake. 200.5. Trends in abundance and status of harbor seals in Oregon: 1977-2003. Mar. Mamm. Sci. 21:657-670. Browne, P., J. L. Laake, and R. L. DeLong. 2002. Improving pinniped diet analyses through identification of multiple skeletal structures in fecal samples. Fish. Bull. 100:423-433 Chilton, D. E., and R. Beamish. 1982. Age determination methods for fishes studied by the Groundfish Program at the Pacific Biologi- cal Station. Can. Spec. Publ. Fish. Aquat. Sci. 60., 102 p. Cottrell, P. E., A. W. Trites, and E. H. Miller. 1996. Assessing the use of hard parts in faeces to identify harbour seal prey: results of captive-feeding trials. Can. J. Zool. 74:875-880. Deagle, B. E., D. J. Tollit, S. N. Jarman, M. A. Hindell, A. W. Trites, and N. J. Gales. 2005. Molecular scatology as a tool to study diet: Analysis of prey DNA in scats from captive Steller sea lions. Mol. Ecol.'l4(6):1831-1842. Everitt, R. D., and R. J. Beach. 1982. Marine mammal fisheries interactions in Oregon and Washington: an overview. Trans. N. Am. Wildl. Nat. Resour. Conf 47:265-277. Graybill, M. R. 1981. Haulout patterns and diet of harbor seals Phoca vitulina, in Coos County, Oregon. M.S. thesis, 55 p. Univ. Oregon, Eugene. OR. Hagerman. F. B. 1952. The biology of the Dover sole, (Microstomus pacificus). Calif Dep. Fish. Game. Bull. 85. 48 p. Harvey, J. T. 1989. Assessment of errors associated with harbour seal {Phoca vitulina) faecal sampling. J. Zool., Lond. 219:101-111. Kvitrude, M. A., S. D. Riemer, R. F. Brown, M. R. Bellinger, and M. A. Banks. 2005. Pacific harbor seals i Phoca vitulina) and salmon: Genetics presents hard numbers for elucidating preda- tor-prey dynamics. Marine Biol., Springer Berlin Hei- delberg. 147(6):1459-1466. Love, M. S., M. Yoklavick, and L. Thorsteinson. 2002. The rockfishes of the northeast Pacific, 404 p. Univ. California Press Berkeley, CA. Markle, D. F., P. M Harris, and C. L. Toole. 1992. Metamorphosis and an overview of early-life his- tory stages in Dover sole, Microstomus pacificus. Fish. Bull. 90:285-301. NMFS (National Marine Fisheries Service) 1997. Investigation of scientific information on the impacts of California sea lions and Pacific harbor seals on salmonids and on the coastal ecosystems of Wash- ington, Oregon and California. NOAA Tech. Memo NMFS-NWFSC-28, 172 p. Olesiuk, P. F, M. A. Bigg, G. M. Ellis, S. J. Crockford, and R. J. Wigen. 1990. An assessment of the feeding habits of harbour seals {Phoca vitulina) in the Strait of Georgia, British Columbia, based on fecal analysis. Can. Tech. Rep. Fish. Aquat. Sci. 1730, 135 p. Orr, A. J., A. S. Banks, S. Mellman, H. R. Huber, and R. L. DeLong. 2004. Examination of the foraging habits of Pacific harbor seal {Phoca vitulina richardsi) to describe their use of the Umpqua River, Oregon, and their predation on salmonids. Fish. Bull. 102:108-117. Pierce, G. J., and P. R. Boyle. 1991. A review of methods for diet analyses in piscivorus marine mammals. Oceanogr. Mar. Biol. Ann. Rev. 29:409-486. 630 Fishery Bulletin 104(4) Pikitch, E. K., and R. L. Demory. 1988. Assessment of scales as a means of aging Dover sole. Trans. Am. Fish. Soc. 117:345-349. Purcell, M., G. Mackey, E. LaHood, H. Huber, and L. Park. 2004. Molecular methods for the genetic identification of salmonid prey from Pacific harbor seal iPhoca vituUna nchardsi) scat. Fish. Bull. 102:213-220. Riemer, S. D., and R. F. Brown. 1997. Prey of pinnipeds at selected sites in Oregon iden- tified by fecal (scat) analysis, 1983-1996. Or. Dept. Fish Wildl. Tech. Rep. 97-6-02, 34 p. Roffe, T. J., and B. R. Mate. 1984. Abundances and feeding habits of pinnipeds in the Rogue River, Oregon. J. Wildl. Manage. 48(4): 1262-1274. Salthaug, A. 2003. Dynamic age-length keys. Fish. Bull. 101:451-456 Sampson, D. B. and C. Wood. 2001. Stock status of Dover sole off the U.S. West Coast in 2000. Appendix (7? Status of the Pacific Coast ground- fish fishery through 2001 and acceptable biological catches for 2002: stock assessment and fishery evalu- ation, 110 p. Pacific Fishery Management Council, Portland, OR. Tollit, D. J., M. J. Steward, P. M. Thompson, G. J. Pierce. M. B. Santos, and S. Hughes. 1997. Species and size differences in the digestion of otoliths and beaks: implications for estimates of pin- niped diet composition. Can. J. Fish. Aquat. Sci. 54: 105-119. 631 Habitat, age, and diet of a forage fish in southeastern Alaska: Pacific sandfish (Trichodon trichodon) John F. Thedinga (contact author) Scott W. Johnson Donald G. Mortensen Auke Bay Laboratory 11305 Glacier Highway Alaska Fisheries Science Center National Marine Fisheries Service Juneau, Alaska 99801-8626 Email address for J F Thedinga John Thedinga(Snoaa gov Forage fish are an important part of Alaska's marine ecosystems and coastal areas. Forage fish are a criti- cal food source for numerous ground- fish, marine mammals, and seabirds (Wespestad'; Allen and Smith, 1988: Paul et al., 1997; Yang and Nelson, 2000; Mecklenburg et al.. 2002). Little is known, however, about the life history characteristics or habitat of many forage fish species in Alaska, including Pacific sandfish [Trichodon trichodon; Fig. 1). Only two articles have been published on the life history characteristics of Pacific sandfish in Alaska. Paul et al. (1997) investigated size-weight-age profiles, size at matu- rity, and fecundity of Pacific sandfish in the northern Gulf of Alaska, and Bailey et al. (1983) examined size and diet of juvenile (<55 mm fork length [FL]) Pacific sandfish in southeastern Alaska. Some Pacific sandfish catch data are also available for the Bering Sea, Prince William Sound, and southeastern Alaska (Isakson et al., 1971; Orsi and Landingham, 1985; Allen and Smith, 1988; Brodeur and Livingston, 1988; Sturdevant et al.-, Orsi et al., 2000). Pacific sandfish burrow into sand, usually at depths shallower than 150 m, and can reach a maximum size of about 300 mm (Marliave, 1980; Mecklenburg et al., 2002). Pacific sandfish are commonly found in nearshore waters of the southeastern Bering Sea and Gulf of Alaska. There is no commercial fishery for Pacific sandfish in Alaska, but sailfin sandfish (Arctoscopus ja- poniciis) are commercially fished and cultured in Japan and Korea (Okiya- ma, 1990). In particular, information is scarce on the biology and habitat of Pacific sandfish, especially for southeastern Alaska. Shoreline de- velopment and global climate change (e.g., increased water temperature and sea level) may adversely affect Pacific sandfish populations because of the relatively specialized nearshore spawning sites and one-year incuba- tion period of this species (Marliave, 1980). 1 Wespestad, V. G. 1987. Population dynamics of Pacific herring (Clupea pallasii). capelin (Maltotus villosiis), and other coastal pelagic fishes in the eastern Bering Sea. In Forage fishes of the southeastern Bering Sea; proceed- ings of a conference, November 1986, Anchorage, AK, p. 55-60. U.S. Dep. Interior, Minerals Management Service, OCS Study MMS 87-0017. 2 Sturdevant, M. V., T. M. Willette, S. C. Jewett. E. Debevec, L. B. Hulbert, and A. L. J. Erase. 1999. Forage fish diet overlap, 1994-1996, APEX Project: Alaska predator ecosystem experiment in Prince William Sound and the Gulf of Alaska, 103 p. Exxon Valdez Oil Spill Restoration Project Final Report (Restoration Project 98163C), Auke Bay Laboratory, National Marine Fisheries Service, 11305 Glacier Highway, Juneau, Alaska. The focus of our study was to pro- vide new information on the general biology of a little known forage fish species. Objectives were to determine habitat preference, age, size, and di- et of Pacific sandfish. To accomplish this, from 2001 to 2004, we captured Pacific sandfish with a beach seine in July and March and with a mid-wa- ter trawl in May near The Brothers Islands in southeastern Alaska. Materials and methods Fish capture and habitat Pacific sandfish were captured with a beach seine at The Brothers Islands in southeastern Alaska (Fig. 2). We seined 10 sites in summer (July 2001-2003) and in winter (March 2002-2004) in a variety of near- shore habitat types (Table 1). Habi- tats sampled included steep bedrock outcroppings, rocky bottoms with understory kelps (e.g., Laminaria), eelgrass {Zostera marina), and sand beaches. We used a 37-m variable- mesh beach seine that tapered from 5 m wide at the center to 1 m wide at the ends. Outer panels were each 10 m of 32-mm stretch mesh, inter- mediate panels were each 4 m of 6- mm square mesh, and the bunt was 9 m of 3.2-mm square mesh. We set the seine as a "round haul" by hold- ing one end on the beach, backing around in a skiff with the other end to the beach about 18 m from the starting point, and pulling the seine onto shore. The seine had a lead line and a float line so that the bottom contacted the substratum and the top floated on the surface. All seine sites were sampled during daylight and within two hours of low tide (range -i-l.O to -1.5 m below mean lower low water). After retrieval of the net, the entire catch was sorted, identified to species, counted, and a subsample was measured for length. Manuscript submitted 16 May 2005 to the Scientific Editor's Office. Manuscript approved for publication 1 December 2005 by the Scientific Editor. Fish. Bull. 104:6.31-637 (2006). 632 Fishery Bulletin 104(4) Figure 1 Pacific sandfish (Trichndon trichodoni captured in southeastern Alaska. Table 1 Habitat characteristics of sites (see Fig. 2) at The Brothers Islands in southeastern Alaska where Pacific sandfish iTrichodon trichodon) were captured with a beach seine from 2001 to 2004. Substrata are listed in decreasing order of abundance from left to right. For vegetation, algae were mostly Ulva, eelgrass was Zostera marina, and kelps were mostly Laminariales (e.g., Lami- naria,Alaria, and Cymathere). PSS = practical salinity scale. Catch = number of sandfish. Site Substrate Vegetation Slope Temperature CO Salinity (PSSi Catch July March 2001-03 2002-04 1 sand, shell algae gentle 4.0-9.5 32-33 6 110 2 large cobble, small cobble, gravel, shell kelp moderate 9.0 33 14 0 3 large cobble, sand kelp gentle 12.0 24 1 0 4 sand, shell, gravel, small cobble eelgrass gentle 12.0 33 1 0 5 small cobble, large cobble, sand, gravel kelp moderate 4.9-9.5 33 2 1 6 bedrock outcrop kelp steep 4.2-10.0 31-33 900 300 7 bedrock outcrop kelp steep 4.2-10.0 31-33 68 500 8 bedrock outcrop kelp steep 4.2 33 0 1120 9 bedrock outcrop kelp steep 4.0 33 0 5000 10 bedrock outcrop kelp steep 4.0-10.0 30-33 Total 35 1027 7200 14,231 The number of larvae (s30 mm) in large catches in winter was estimated visually. Water temperature and salinity (practical salinity scale; PSS) were measured at each site at a depth of 20 cm with a thermometer and a hand-held refractometer. Pacific sandfish were also captured with a mid-water rope trawl in May 2001 in Frederick Sound, midway between the mouth of Pybus Bay and The Brothers Islands, and in May 2003 in Pybus Bay (about 12 km from The Brothers Islands) (Fig. 2). In 2001, the midwa- ter trawl was a 164 Nordic rope trawl with 1.5-m-' alloy doors; it was 7 m high and 17 m wide and was equipped with a 19-mm codend liner. In 2003, because a different vessel was used, the trawl was a mesh wing 25/21/64 NOTE Thedmga el al Habitat, age, and diet of Tnchodon tnchodon 633 134 02 W 1 133,94 1 33,86 1 Admiralty Island ^ ■t The Brothers Islands * A "'"" ,^6 '*-'^ «,10 ? Pybws , , ^ Bay O / ^ a^^_ N S C^' / t t ^C Alaska ^^^^ ^\ 0 500 1.000 km \ 111 \ "' ~^o /^ ,'"'isiuay area ® Beach seine A Mid-water trawl 6 km 1 0 1 3 1 1 1 - 57,30 N - 57,26 Figure 2 Location of sites (see Table 1 for habitat characteristics) sampled for Pacific sandfish {Trichodon tnchodon) near The Brothers Islands, southeastern Alaska, 2001 to 2004. Sampling methods included beach seining and mid-water trawling. trawl with 3.0-m- alloy doors: it was 11 m high and 29 m wide and was equipped with a 32-mm mesh codend liner. In Frederick Sound, water depth was about 74 m, and the trawl was fished at a depth of 14 m. In Pybus Bay, water depth was about 70 m and the trawl was fished at depths from 50 to 64 m about 400 m offshore. Water temperature and salinity were obtained from deployment of a CTD (conductivity-temperature-depth) profiler in Pybus Bay in 2003. Age and size A sample of Pacific sandfish was frozen and brought to the laboratory where sagittal otoliths were removed and stored in 95% ethanol. Fish used for otolith collec- tions covered a range of sizes in the catch (Table 2). Fish were measured to the nearest mm FL. and larvae were measured to the nearest mm total length (TL). A sample of fish was individually weighed to the nearest gram. Because a sample of larvae captured with a seine and preserved (lOVf formalin) in 2002 was not measured until 2004, a shrinkage factor of 10.3% was applied to the lengths to adjust for storage in formalin (Marliave, 1980). The left otolith of each fish was cut into thin sections by using standard methods described in the otolith manual for microstructural examination (Secor et al., 1991). The otolith was attached to a glass slide, medial side down, with thermal resin. The surface of the otolith was then ground to the primordium with 1000- grit silicon carbide paper. Care was taken to leave the edge of the otolith intact and not to grind through the primordium in the center. The otolith was then polished with an 8000- and 12,000-grit cloth on both sides to reveal the internal microstructure. Otoliths collected in 2001 were aged separately by two biologists. When age estimates differed, otoliths were re-examined together by both readers, and an age was agreed upon. Otoliths collected in 2002 were aged by one biologist three dif- ferent times. If age readings differed between any of the three readings, the otolith was re-examined and a best estimate of the age was determined. For both years, otoliths were examined without knowledge of fish size or collection date. Diet A sample of Pacific sandfish (>age-0) was preserved in 10% formalin and analyzed for stomach contents. Preserved fish were measured to the nearest mm FL, weighed to the nearest 0.1 g, and stomachs were excised. Stomachs were weighed before and after removal of contents to obtain an estimate of wet weight by sub- traction. Contents were examined with a dissecting microscope and identified to major taxonomic groups. 634 Fishery Bulletin 104(4) Table 2 Size and age based on surface observations of annual rings on otoliths of Pacific sandfish \Tnchodon trichodon ) from nearshore waters of The Brothers Islands, and from Pybus Bay and Frederick Sound, southeastern Alaska from 2001 to 2003. Fish from Frederick Sound and Pybus Bay were captured with a mid-water trawl, whereas all other fish were captured with a beach seine. Asterisk denotes total length (mm), n = number of sandfish in the sample. Date Location n Fork length (mm) Mean weight (g) Age (yr) Mean Range July 2001 The Brothers Islands 22 80 70-95 6.4 1 May 2001 Frederick Sound 1 120 — 24.0 4 May 2001 Frederick Sound 1 190 — 82.0 6 March 2002 The Brothers Islands 9 16* 14-18* — 0 July 2002 The Brothers Islands 24 70 55-79 4.3 1 July 2002 The Brothers Islands 1 135 — 32.0 5 May 2003 Pybus Bay 1 121 — 23.5 4 May 2003 Pybus Bay 10 141 129-152 37.2 5 May 2003 Pybus Bay 7 160 156-166 58.3 6 Frequency of occurrence (% FO) was determine(i for each taxonomic group based on the percentage of stom- achs in which that taxonomic group occurred. Percent volume (% Vol) of each taxonomic group was estimated visually and total weights per taxon were estimated by multiplying the percent volume by the total stomach weight. Results Fish catch and habitat Total catch of Pacific sandfish was 15,431 fish. Most (99%) fish were captured with a seine: 1027 fish in July (total for all years) and 14,231 larvae in March (total for all years; Table 1). Total trawl catch was 173 fish; 2 fish in Frederick Sound in May 2001 and 171 fish in Pybus Bay in May 2003. Most Pacific sandfish captured with a seine in July and March each year were found adjacent to steep bed- rock outcroppings with attached kelp (Table 1). Water temperatures in summer and winter for all years were about 10°C and 4°C, and salinity ranged from 30 to 33 PSS. Larvae were captured in March at many of the same sites where juveniles were captured in July (sites 1, 9-11, 14; Table 1; Fig. 2). The few adult Pacific sandfish (>130 mm FL) captured with a seine in July 2002 and 2003 were found exclusively in low gradi- ent, sandy habitat (site 1; Fig. 2). Most adult Pacific sandfish were captured in Pybus Bay with a trawl at least 400 m offshore of any land mass and at depths between 14 and 64 m; at the depth (50 m) most fish were captured, temperature was 6.5°C and salinity was 31.5 PSS. Age and size Nearly all Pacific sandfish captured with a seine in July (all years) were juveniles; only larvae were captured in March (all years). Approximately 14,000 Pacific sandfish larvae were captured in March (Table 1); mean TL was 16.2 mm (13.1 to 23.2 mm; Fig. 3). Of nine larvae aged from March 2002, all were age-0 (Table 2). Mean FL of Pacific sandfish captured with a seine in July was 86 mm (62 to 150 mm; Fig. 3). Of 47 Pacific sandfish aged from seine catches in July, 46 were age-1 (mean FL=75 mm) and one was age-5 (FL = 135 mm) (Table 2). All trawl-caught Pacific sandfish were adults or sub- adults. Mean FL of Pacific sandfish from trawl catches was 150 mm (120 to 190 mm; Fig. 3). The 21 fish aged from trawl catches in May (both years) were age-4, age- 5, and age-6 (mean FL = 121, 141, and 164 mm) (Table 2). No age-2 or age-3 fish were captured in trawls. Diet From analysis of 195 stomachs, we found that Pacific sandfish were both planktivorous and piscivorous. Percent weight of stomach contents for age-1 Pacific sandfish varied between years (Fig. 4). For both 2001 and 2002 combined, decapods, fish, and euphausiids accounted for about 90% of stomach content weight. Fish made up 50% of the weight in 2001, but only 7%. in 2002. The other two heaviest items, decapods and euphausiids, ranged from 30% to 50% and from 17% to 25% of the weight in 2001 and 2002, respectively. Diet of adult Pacific sandfish (2003; 2age-4) consisted of four food items and was dominated by fish and deca- pods (Fig. 4). Fish accounted for 75%c of the weight of stomach contents for 2age-4 fish. Frequency of occur- NOTE Thedinga et a\ : Habitat, age, and diet of Tnchodon tnchodon 635 13-14 15-16 17-18 19-20 21-22 23-24 Total length (mm) Beach seine Summer 60-69 70-79 80-89 90-99 100-112 130-150 Fork length (mm) 50 40- 30- 20 n=120 FL=150 Mid-water trawl Spring 120-129 130-139 140-149 150-159 160-169 170-190 Fork length (mm) Figure 3 Length frequency of Pacific sandfish (Trichndon tnchodon) near The Brothers Islands, south- eastern Alaska. (A) Larvae were captured in March. 2002-2004, (B) juveniles and subadults were captured in July, 2001-2003, and (C) sub- adults and adults were captured in May 2001 and May 2003. Total length was measured for larvae and fork length for juveniles, subadults. and adults. rence of food items in stomachs varied among years for age-1 Pacific sandfish (Fig. 4). Occurrence of fish in stomachs of age-1 fish ranged from 72% in 2001 to 19% in 2002, and occurrence of larvacea ranged from 62% in 2001 to 1% in 2002. Nearly all identifi- able fish in juvenile stomachs were gadids. For 2001 and 2002 combined, decapod larvae occurred in about 98% of stomachs, euphausiids in 74%, and amphipods in 65%. For adults (2003; 2age-4), fish were the most 100 80- 60 ■ 40 20 Age 1 n=96 Age 1 n=68 DFish ■ Decapods S Euphausiids D Calanoids H Amphipods Q Larvacea H Other 2001 2002 2003 100 80 60 40 20 Age 1 1 1 Age 1 B "" pi n=68 r~ 1 ' ^ Age>4 n=31 I § - i g - i r- "".~ I 1 ::ii^ a § m 1 n 2001 2002 2003 Figure 4 (A) Percent weight of stomach contents and (B) percent frequency of occurrence of prey items in stomachs of Pacific sandfish (Tnchodon trichodon ) near The Brothers Islands, southeastern Alaska. Data for 2001 and 2002 represent age-1 fish captured with a beach seine in July, and data for 2003 represent fish >age 4 captured with a mid-water trawl in Pybus Bay and Frederick .Sound in May. See Figure 2 for site locations. frequently occurring food item and were present in 77%f of stomachs. Discussion Pacific sandfish larvae and juveniles use shallow, near- shore habitats, whereas most adults occupy deeper waters farther from shore. Larvae were similar in size and anatomical features to larvae (0-29 day-old) described by Marliave (1980). Thus, we estimate that the larvae we captured probably hatched between early February and early March — a hatching period similar to that reported by Clemens and Wilby (1967) for British Columbia, Canada, but 1-2 months earlier than hatch- ing dates estimated by Bailey et al. (1983) for northern southeastern Alaska. Pacific sandfish spawn on rocky intertidal shorelines, eggs incubate for about one year, and larvae develop in shallow nearshore areas (Marliave, 1980). Because of the relative high abundance of Pacific 636 Fishery Bulletin 104(4) sandfish larvae in the shallow nearshore waters of The Brothers Islands, this area appears to be an important spawning and nursery area for Pacific sandfish. The length of time that larvae remain in shallow nearshore areas, however, appears to be limited because we cap- tured no age-0 fish in July. Pacific sandfish have been caught incidentally in a variety of habitats in Alaska. In southeastern Alaska, Orsi and Landingham (19851 caught Pacific sandfish in low-gradient beaches composed of sand, gravel, cobble, or a combination of these substrates, and Orsi et al. (2000) caught Pacific sandfish with a rope trawl in deeper waters. In the Aleutian Islands, Isakson et al. (1971) reported that Pacific sandfish were a prominent species in the inshore sand and gravel community. Pa- cific sandfish are known to burrow into soft substrates, but most of the juveniles that we captured were close to shore and adjacent to bedrock outcroppings — area and substrate similar to those observed by Bailey et al. (1983). Although Mecklenburg (2003) reported that Pacific sandfish generally burrow in sand during the day and are active at night, we observed large schools of juvenile Pacific sandfish actively feeding near the surface during the day. Age of Pacific sandfish varied with time of capture. We captured mostly age-1 Pacific sandfish in July; mean size of age-1 fish in our study (75 mm FL) was similar to the size of age-1 fish captured in the northern Gulf of Alaska (71 mm FL; Paul et al., 1997). In spring we cap- tured subadult and adult Pacific sandfish, and in winter we caught only larval Pacific sandfish. Notably absent in all of our catches were age-2 and age-3 fish. Paul et al. (1997) captured age-2 and age-3 Pacific sandfish by a variety of methods in the northern Gulf of Alaska. Most of their sampling, however, occurred in August and in deeper waters (29-55 m). Pacific sandfish may segregate into different habitat types depending on age and size, which may explain the absence of some age classes in our catches. Diets consisting of crustaceans and fish have been reported for Pacific sandfish similar in size to those in our study (Paul et al., 1997; Sturdevant et al.-). Percent FO of fish in the diet of Pacific sandfish was 66% in the Gulf of Alaska (Paul et al., 1997), 100% in Prince Wil- liam Sound (Sturdevant et al.-), and 56% in our study. For crustaceans, the highest %■ FO was for shrimp (19%) in the Gulf of Alaska (Paul et al., 1997), gam- marids (46%) in Prince William Sound (Sturdevant et al.2), and crab larvae (83%o) in our study. Near Kodiak Island, Alaska, stomach contents of 26 Pacific sandfish in July were dominated by fish (70% by number and 97% by weight; Rogers et al.^). ■' Rogers, D. E., D. J. Rabin, B. J. Rogers, K. Garrison, and M. Wangerin. 1979. Seasonal composition and food web relationships of marine organisms in the nearshore zone of Kodiak Island including ichthyoplankton, meroplankton (shellfish), zooplankton. and fish. Environmental Assess- ment of the Alaskan Continental Shelf, p. 529-662, vol. IV. Receptors, Fish, Littoral, Benthos. Outer Continental Shelf Environmental Assessment Program, Boulder, CO. Percent FO of fish in Pacific sandfish stomachs was nearly four times greater in 2001 than in 2002, which probably reflects the abundance of prey. Based on seine catches at The Brothers Islands, the relative abundance of young-of-the-year (YOYi gadids, the dominant fish identified in Pacific sandfish stomachs, was over six times greater in 2001 than in 2002 (Thedinga et al., 2006). In addition, mean length of Pacific sandfish was about 5 mm longer in 2001 than in 2002, which may reflect the increased presence of fish in the diet of Pacific sandfish, or could be due to increased growth time because we sampled 10 days later in 2001 than m 2002. Diet differed by fish size. Larger Pacific sandfish (mean FL 150 mm) ate mostly fish, whereas smaller Pacific sandfish (mean FL 86 mm) ate mostly decapods. Paul et al. (1997) also observed a change in diet based on fish size; % FO of fish in the diet of Pacific sandfish greater than 115 mm FL was 90% compared to 14% for fish less than 99 mm. Also in their study, mean % FO of non-fish food items was 5% for larger fish compared to 24% for smaller fish. Bailey et al. (1983) reported observing juvenile Pacific sandfish in mixed schools with pink salmon iOncorhyn- chiis gorbuscha) fry, but we caught Pacific sandfish with YOY walleye pollock {Theragra chalcograinma), YOY Pacific cod (Gadus macrocephalus), YOY Pacific herring (Chipea pallasii). and juvenile chum salmon (Oncorhyii- chus keta). Apparently, Pacific sandfish exhibit school- ing behavior as larvae and juveniles and co-occur with a variety of forage species, sometimes preying on those that are of consumable size. High predation rates of juvenile walleye pollock by Pacific sandfish have been reported for the Bering Sea and Gulf of Alaska (Gue- nette, 2005). Shallow nearshore waters provide important nurs- ery and spawning habitat for Pacific sandfish. Pa- cific sandfish diet varied between size classes and years which was probably dependent on abundance of YOY walleye pollock. Pacific sandfish are a nutri- tious forage fish with moderately high oil, protein, and caloric content (Anthony et al., 2000; Logerwell and Schaufler, 2005) and could therefore be impor- tant to some predators at certain times of the year. For example. Pacific sandfish occurred in up to 64% of Steller sea lion {Eumetopias jubatus) scats in the Aleutian Islands (Sinclair and Zeppelin, 2002). The dependence of Pacific sandfish upon nearshore ar- eas for spawning, egg incubation, and larval rearing, coupled with the greater sensitivity to pollutants of early life stages than adults (Carls et al., 1999), war- rant the protection of nearshore areas from shoreline development and pollutants to maintain healthy forage fish populations. Acknowledgments We thank D. Csepp, E. Flynn, and M. Johnson for help with fish collections and otolith removal. We also thank NOTE Thedinga et a\ Habitat, age, and diet of Tiichodon tiichodon 637 K. Munk, M. Sturdevant, and M. Cook for help with age and diet analyses. D. Neff assisted with the manuscript figures. The NOAA ship John N. Cobb and the fishing vessel Patriot provided logistical support. J. Rice. A. Moles, N. Hillgruber, and K. Munk provided helpful reviews on earlier drafts of this manuscript. Literature cited Allen, J., and G.B. Smith. 1988. Atlas and zoogeography of common fishes in the Bering Sea and northeastern Pacific. U.S. Dep. Commer., NOAA Tech. Rep. NMFS-66, 151 p. Anthony. J. A., D. D. Roby. and K. R. Turco. 2000. Lipid content and energy density of forage fishes from the northern Gulf of Alaska. J. Exp. Mar. Biol. Ecol. 248:53-78. Bailey, J. E., B. L. Wing, and J. H. Landingham. 1983. Juvenile Pacific sandfish, Trichodon trichodon, associated with pink salmon. Oncorhynchus gorbuscha, fry in the nearshore area, southeastern Alaska. Copeia 2:549-551. Brodeur, R. D., and P. A. Livingston. 1988. Food habits and diet overlap of various eastern Bering Sea fishes. U.S. Dep. Commer.. NOAA Tech. Memo. NMFS/NWC-127, 76 p. Carls, M. G., J. E. Hose, R. E. Thomas, and S. D. Rice. 1999. Sensitivity offish embryos to weathered crude oil: Part \. Low-level exposure during incubation causes malformations, genetic damage, and mortality in larval Pacific herring iClupea pallasii). Environ. Toxicol. Chem. 19(6): 1649-1659. Clemens, W. A., and G. V. Wilby. 1967. Fishesof the Pacific coast of Canada, 433 p. Fish. Res. Board Can.. Bull. 68. Guenette, S. 2005. Models of Southeast Alaska. /;; Food web models and data for studying fisheries and environmental impacts on Eastern Pacific ecosystems (S. Guenette, and V. Christensen, eds.i, p. 106-178. Fisheries Centre Research Report 13. Isakson, J. S.. C. A. Simenstad, and R. L. Burgner. 1971. Fish communities and food chains in the Amchitka area. BioScience 21:666-670. Logerwell, E. A., and L. E. Schaufier. 2005. New data on proximate composition and energy density of Steller sea lion (Eumetopias jubatus) prey fills seasonal and geographic gaps in existing infor- mation. Aquat. Mamm. 31:62-82. Marliave, J. B. 1980. Spawn and larvae of the Pacific sandfish, Trichodon trichodon. Fish. Bull. 78:959-964. Mecklenburg, C. W., T. A. Mecklenburg, and L. K. Thorsteinson. 2002. Fishesof Alaska, 1037 p. Am. Fish. Soc, Bethesda, MD. Mecklenburg, C. W. 2003. Family Trichodontidae Bleeker 1859— sand- fishes. Calif Acad. Sci., Annotated checklists of fishes no. 15, 4 p. Okiyama, M. 1990. Contrast in reproductive style between two spe- cies of sandfishes (family Trichodontidae). Fish. Bull. 88:543-549. Orsi, J. A., and J. H. Landingham. 1985. Numbers, species, and maturity stages of fish captured with beach seines during spring 1981 and 1982 in some nearshore marine waters of southeast- ern Alaska. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-F/NWC-86, 34 p. Orsi, J. A., M. V. Sturdevant, J. M. Murphy, D. G. Morten.sen, and B. L. Wing. 2000. Seasonal habitat use and early marine ecology of juvenile Pacific salmon in southeastern Alaska. NPAFC Bulletin 2:111-122. Paul, J. M., A. J. Paul, T. J. Vogeler, and J. P. Doyle. 1997. Biological investigations on Pacific sandfish {Trichodon trichodon ) in the northern Gulf of Alaska. In Proceedings of the international symposium on the role of forage fishes in marine ecosystems, p. 87-94. Alaska Sea Grant College Program Report no. 97-01. Secor, D. H., J. M. Dean, and E. H. Laban. 1991. Manual for otolith removal and preparation for microstructural examination. Technical Publication 1991-01, 85 p. Belle Baruch Institute for Marine Biol- ogy and Coastal Research, Columbia, SC. Sinclair, E. H., and T. K. Zeppelin. 2002. Seasonal and spatial differences in diet in the west- ern stock of Steller sea lions iEumetopias jubatus). J. Mamm. 83:973-990. Thedinga, J. F., S. J. Johnson, and D. J. Csepp. 2006. Nearshore fish assemblages in the vicinity of two Steller sea lion haul-outs in southeastern Alaska. In Proceedings of the 22"'' Wakefield Fisheries Symposium Sea lions of the world: conservation and research in the 21st century, p. 269-284. Alaska Sea Grant College Program, Univ. Alaska, Fairbanks, AK. Yang, M-S., and M. W. Nelson. 2000. Food habits of the commercially important ground- fishes in the Gulf of Alaska in 1990, 1993, and 1996. U. S. Dep. Commer., NOAA Tech. Memo. NMFS-AFSC- 112, 174 p. 638 Genetic diversity of yellow grouper iEpinephelus awoara) determined by random amplified polymorphic DNA (RAPD) analysis Satyendra K. Upadhyay (contact author) Wang Jun Su Yong-Quan Ding Shao-Xiong Sonal Chaturvedi Laboratory of Fish Genetics Diseases and Breeding College of Oceanograpfiy and Environmental Science Xiamen University Xiamen 361005 Fujian, Peoples Republic of China Present address (for S K Upadhyay): A-37, Satya Vi|aya Society J, M. Road Bhandup (W) Mumbai 400078 (Maharashtra), India Email address for S K. Upadhyay satyaxmu yahoo.com The development of molecular tech- niques has enhanced our ability to identify fish species. A few molecular markers, such as mitochondrial DNA and ribosomal DNA, have been used to assist in the identification of fish species. One such technique now rou- tinely used is the random amplified polymorphic DNA polymerase chain reaction (RAPD-PCR; Williams et al.. 1990). Genetic analysis with RAPD markers is relatively easy, fast, and efficient. RAPD analysis, however, may not be practical for identifing species that interbreed (Martinez et al., 1997). Although the major limi- tation of RAPD technique for iden- tification of intraspecific specimens is its repeatability, adherence to protocol and standardized reactions can improve the method (Jones et al., 1997). This technique has been used for identification and detection of genetic diversity in various fish species (tilapia; Naish et al., 1995; striped bass; Bielawski and Pumo, 1997; grouper; Asensio et al., 2002). Most other DNA-based methods are more laborious and time consuming than RAPD, making them less suit- able for large population or genetic diversity studies. Suitable loci with high reproducibility allow the identi- fication of unequivocally distinct spe- cies (i.e., where there is large genetic differentiation between species, com- pared to within species; Greig et al., 2005). Groupers (Epinephelinae: Serrani- dae) are among the most important and highly valued demersal species of tropical and subtropical coastal ar- eas worldwide. In general, grouper species lack distinct morphological specializations, but color patterns and geographic location are used in the field. Some confusion exists over the names of some important Indo- Pacific grouper species (Heemstra and Randall, 1993; Sadovy, 1997) and little work has been done to es- tablish species identities and their genetic diversity. One alternative to gathering information on these spe- cies is the use of molecular genetic markers. In the present study, RAPD analysis has been used to investigate the genetic variation in two popula- tions of yellow grouper iEpinephelus awoara) from the South China Sea. Materials and methods Sample preparation and DNA extraction Yellow grouper were obtained from two populations (30 individuals from Xiamen and 26 individuals from Guangdong) of the South China Sea. Muscle samples from fish were taken, placed in 95% ethanol, trans- ported to the laboratory, and stored at -20"C until analysis. Genomic DNA was extracted according to the DNA extraction method of DeSalle et al. (1993). Primer selection and RAPD reaction profile Twenty primers (Table 1) were selected on the basis of presence of intense, well-distinguished, and reproducible bands for further analysis. PCRs for each population were performed in 2-,i(L volumes containing 100 mM Tris-HCl (pH 8.3), 500 niM KCl, 15 mM MgCU, 1 f(L dNTPs. 0.4 unit rTaq polymerase and approximately 20 ng of template DNA and 1 ^L of each primer. Thermocycling condi- tions were as follows; initial dena- turation step of 95°C for 5 min., 30 cycles of 45 s at 93°C, annealing tem- perature of 45 s, primary extension of 72'C for 2 min., and a final extension of 72°C for 5 minutes. Detection of amplified DNA Electrophoresis of a lO-j/L portion of the amplification reaction was performed for 45 minutes at 100 V in a 1.2 % agarose gel contain- ing ethidium bromide (1 f(g/mL) in Tris-acetate buffer (0.004 M Tris acetate, 0.001 M EDTA, pH 8.0). DNA fragments were viewed with UV transillumination and analyzed with Genescan (Gene-Geneius Bio- imaging System, Cambridge, Eng- land). Their sizes were estimated by comparison with a commercial 1-kb plus DNA ladder. Manuscript submitted 10 August 2004 to the Scientific Editor's Office. Manuscript approved for publication 19 December 2006 by the Scientific Editor. Fish. Bull. 104:638-642(2006). NOTE Upadhyay et a\ Genetic diversity of Epinephelus awoara determined by RAPD analysis 639 Table 1 Primer details and SN=serial number. patt XM = ;rn of polymorphism in twi Xiamen; GD = Guangdong. sampled popul ations (Xiamen and Guangdong) of Epinephelus awoara. SN Primers DNA sequence No. of bands Total no. of polymorphic loci Polymorphic loci in the two populations Percentage of polymorphic loci XM GD 1 SI 17 CACTCTCCTC 8 7 7 4 87.5 2 S465 CCCCGGTAAC 9 8 8 2 88.9 3 S513 GGACGACAAG 6 5 0 5 83.4 4 S514 CAGGATTCCC 9 5 5 4 55.6 5 S515 GGACAACGAG 7 5 5 3 71.4 6 S517 CCGTACGTAG 10 10 10 5 100.0 7 S520 ACGGCAAGGA 7 6 6 6 85.71 8 S1021 GGCATCGGCT 10 7 7 3 70.0 9 S1023 GGGTCCAAAG 8 7 2 7 87.5 10 S1026 TGCCGCACTT 8 8 8 1 100.0 11 S1036 AAGGCACGAG 10 6 6 5 60.0 12 S1037 CCTCACGTCC 7 5 5 1 71.4 13 S1040 CCTGTTCCCT 10 7 7 2 70.0 14 S1042 TCGCACAGTC 6 2 2 2 33.3 15 S1043 AGCACCTCGT 11 11 11 6 100.0 16 S1044 GAATGCGACC 7 5 5 1 71.4 17 S1045 CAGCGTTGCC 10 7 5 7 70.0 18 S1049 ACGGCACGCA 5 3 3 3 60.0 19 S1053 CAGCCGTTCC 6 4 1 4 66.7 20 S1055 GAATCCGGCA 5 3 3 3 60.0 Total 159 121 106 74 Percentage of polymorphic loci 76.10 66.66 46.54 Data analysis The presence (1) or absence (Ol of RAPD products was scored for all loci based on a minimum of two replicates. Only those characters that could be reliably scored from replicate RAPD reactions were included in the analysis. Loci that failed to generate RAPD amplification products were excluded from the analysis. Probability (Fischer's exact test) was carried out to evaluate differences in homogeneity of gene frequencies across populations for each loci. Test of heterozygosity (H-p), expected hetero- zygosity (Hj^), population differentiation (G^i<), and gene flow (A'',,,) were also conducted between two populations oi Epinephelus awoara. Results Two populations of Epinephelus awoara were analyzed to characterize the RAPD marker for genetic variation. Useful RAPD primers and markers for genetic variation The first group of 20 primers (Table 1) produced a total of 159 bands, among which 121 polymorphic bands (76.10%) were observed. The 38 monomorphic bands (23.90%) could be considered, on a preliminary basis, as population diagnostic bands that allow clear differentia- tion among populations of Epinephelus awoara because they were present in all individuals analyzed. Primers S514, S1021, S1036, S1040 and S1042 generated more diagnostic bands than did other primers. The absence of these markers in related populations must, however, still be confirmed. Primers that showed high levels of reproducibility also produced more polymorphic bands. Over five poly- morphic bands (5.76) per primer were produced from 20 primers on average. All 20 primers (Table 1) yielded satisfactory amplification products for all specimens tested. Each primer produced a unique band pattern of amplified DNA. For a given primer and DNA template, a number of bands appeared intense and reproducible in all the replicates, and bands of weaker stain appeared occasionally. All primers were able to distinguish two populations of E. awoara, reproducing different and well-characterized banding patterns. Banding pattern variations observed with some primers were not consis- tent for all individuals within or between populations and therefore these variations should be attributed to intraspecific polymorphism. 640 Fishery Bulletin 104(4) High rates of RAPD polymorphism Polymorphic rates of RAPD bands were much higher between two populations of Epi?iephehis aivoara. The number of reproducible and well-resolved bands analyzed in the two populations ranged from 5 to 11. Monomor- phic bands, constantly present in all individuals, varied between primers, whereas total polymorphic bands were observed 76.10% of the time for the two populations of E. au'oara. The Xiamen population of £. aivoara exhib- ited 66.66% polymorphism, whereas polymorphism was 46.54% in the Guangdong population (Table 1). RAPD fragments of fewer than 1000 base pairs were found to be more polymorphic than larger fragments. Genetic diversity Genetic diversity between populations was clearly illus- trated in Table 2 and Table 3. The genetic diversity statistics iHj, Hg, G^^, A^,„) estimated between the two populations of Epinephelus awoara for each primer are listed in Table 3. Nei's (1973) unbiased measures of genetic identity and genetic distance between two populations were 0.6216 and 0.4755, respectively. Within the populations, total heterozygosity varied from 0.2189 to 0.4616 and expected heterozygosity ranged between 0.1232 and 0.2920. Probability (Fishers exact test) of homogeneity was significant at P>0.05 for all primers except S465, S513, S1026, S1044, and S1055. Discussion This study showed a considerable amount of genetic variation present in Epinephehis aivoara. RAPD analysis generated a large number of polymorphic DNA bands (Table 1) between the populations of E. aivoara, thus making it one of the most efficient systems for generat- ing DNA markers. However, low levels of interspecific variation were found in RAPD profiles among strains of grouper. Therefore, it was an inefficient system for generating molecular markers for gene mapping (Liu et al., 1999). Several authors (Welsh and McClelland, 1990; Bar- dakci and Skibinski, 1994; Naish et al., 1995) have demonstrated that the RAPD PCR method is a power- ful tool for the assessment of genetic markers that are capable of discriminating between species or subspecies in a wide range of organisms, including fishes. This ca- pacity was confirmed by the results of the present study because, for all screened primers, different RAPD band- ing patterns were observed between the two populations (Table 1). A limitation arising with the application of the RAPD technique is the homology between comigrat- ing bands produced by the same primer in different individuals (Hadrys et al., 1992). Nevertheless, in the present study, homology constitutes a valid assumption because all individuals belong to the same species. Despite the fact that no specific markers were found to identify Epinephelus awoara species, data analysis of the observed and effective number of alleles revealed a degree of divergence (Table 2). The Nei (1973) gene di- versity was also illustrated in Table 2. The two popula- tions of E. aivoara (Table 3) showed high level of genetic variation (i.e.. Shannon's information index; Shannon- Weaver, 1949, average genetic index [Hp,,!,] and genetic diversity index between populations [H^p]). Similarly a high level of genetic variation and genetic distance (0.617-0.949) was observed in dinoflagellates (Baillie et al., 2000). This finding supports a high genetic distance for E. awoara (0.4755) in the present study. This may be due to the selection pressure of pollutants (Nadig et al., 1998) or to overexploitation of groupers in the South China Sea. Total heterozygosity (Hy.) values ranged from 0.2189 to 0.4616, expected heterozygosity (Hg) from 0.1232 to 0.3157, and an estimate of gene flow (iV„,) from 0.3714 to 4.1722. The proportion of total genetic variation within species due to population differentiation (G^.-,) ranged from 0.1070 to 0.5344 (Table 3). Carvalho et al. (1991) observed a very high differentiation (Gc;y,= 0.648) in the guppy (Poecilia reticulata) in northern Trinidad. Similarly, Ward et al. (1994) also observed high G^j values (>0.2000). In any case, the divergence in G^j values indicates that, on average, marine subpopula- tions exchange between one and two orders of magni- tude more migrants per generation. Genetic divergence between areas originates when populations are formed or through the restriction of gene flow (Lage et al., 2004). Higher H^ in marine fish is at least consistent with these fish having, on average, larger population sizes. Homogeneity of gene frequencies was found to be significant (P>0.05) across populations of Epinephelus awoara (Fig. 1) for most of the primers except S465, S513, S1026, S1044,and S1055 (P>0.001). The factor that could influence the genetic variation of the grouper population geographically is the movement of adults, which can be extensive and cover considerable distances (Harding et al., 1997). The otherwise shelter- habituated adults travel long distances necessary to populate remote oceanic islands. The Bermuda fauna include more than 75% of the known west Indian grou- pers even though Bermuda lies more than 800 miles from other coral reefs. Such distributional patterns are best explained as the result of passive transport of larvae by oceanic currents. Similar patterns are also expected in the South China Sea. Adult movement of marine fish is relatively unfettered by physical barri- ers. In addition to gene flow through adult migration, marine fish frequently have a planktonic larval stage of several months duration, which can be expected to further enhance gene flow. Estimation of genetic diversity of Epinephelus awoara by RAPD analysis (i.e., by using the mean of observed number of alleles [Na]. effective number of alleles [Ne] and Nei's gene diversity [//]) was found to be 1.6563, 1.4169, and 0.2915, respectively for Xiamen popula- tions, whereas they were 1.4824, 1.2713, 0.2167, respec- tively for Guangdong populations. Our results provide evidence for a loss (10.5%, 11.3%, and 25.7%, respec- NOTE Upadhyay et al Genetic diversity of Epinephelus awooia determined by RAPD analysis 641 Table 2 Genetic variation in two sampled populations (Xiamen and Guangdong) of Epinephelus awoara. Na = Observed number of alleles; Ne = Effective number of alleles; H = Nei's (1973) gene diversity; Ho (I) = observed heterozygosity (Shannon's information index). Loci Na Ne H Hnil) Xiamen Guangdong Xiamen Guangdong Xiamen Guangdong Xiamen Guangdong S117 1.8750 1..5000 1.5318 1.2272 0.3145 0.1449 0.4675 0.2254 S465 1.8889 1.1111 1.6659 1.0927 0.3673 0.0505 0.5324 0.0719 S513 1.0000 1.8333 1.0000 1.5851 0.0000 0.3298 0.0000 0.4819 S514 1.5556 1.4444 1.3886 1.1905 0.2231 0.1310 0.3269 0.2074 S515 1.7142 1.4285 1.2984 1.1368 0.1924 0.1019 0.3076 0.1565 S517 2.0000 1.5000 1.6602 1.2678 0.3886 1.1696 0.5808 0.2593 S520 1.7142 1.8571 1.5315 1.5871 0.2877 0.3580 0.4162 0.5048 S1021 1.8000 1.3000 1.3246 1.1497 0.1959 0.0946 0.3077 0.1463 S1023 1.2500 1.8750 1.1780 1.4712 0.1004 0.3229 0.1469 0.4806 S1026 2.0000 1.1250 1.6661 1.0575 0.3859 0.0394 0.5691 0.0618 S1036 1.7000 1.4000 1.3534 1.3771 0.3038 0.1940 0.2334 0.2712 S1037 1.7142 1.1428 1.4307 1.0318 0.2254 0.0260 0.3769 0.1039 S1040 1.7000 1.2000 1.5307 1.0605 0.3011 0.0464 0.4348 0.0786 S1042 1.3333 1.3333 1.2429 1.1814 0.1358 0.1107 0.1945 0.1689 S1043 2.0000 1.4545 1.7466 1.1649 0.4081 0.1044 0.5911 0.1694 S1044 1.7142 1.1428 1.4483 1.0057 0.2689 0.0054 0.3997 0.0136 S1045 1.5000 1.8000 1.4107 1.5349 0.2189 0.3146 0.3126 0.4642 S1049 1.6000 1.6000 1.5417 1.3333 0.2839 0.2124 0.3994 0.3228 S1053 1.6666 2.0000 1.1519 1.4657 0.0795 0.3059 0.1116 0.4780 S1055 1.4000 1.6000 1.2359 1.5058 1.1478 0.2722 0.2223 0.3871 Mean 1.6.563 1.4824 1.4169 1.2713 0.2915 0.2167 0.3466 0.2527 o tively for Na. Ne. and H) of genetic diversity; this loss may be due to high fishing pressure of the grouper population in Guangdong. Conclusion These results indicate that commer- cial fishing may result not only in selective genetic changes in exploited stocks but also in reduced genetic diversity due to genetic drift. The loss of genetic diversity may be par- ticularly pronounced during the initial stages of exploitation, and investigations into the advanced stages of exploitation may be less likely to detect significant changes in allelic diversity. This loss can also be tested for other commer- cially exploited species of grouper in genetic and demographic context. In the present study, the RAPD method proved useful and technically convenient for the study of genetic diversity of yellow grouper [Epinephelus awoara): therefore, it can Loci Figure 1 Probabilities (P>0.05) that show the significance of homogeneity of gene frequencies across populations of yellow grouper [Epinephelus awoara I. Arrows show the range of probability. also be applied to other species of grouper and enable the analysis of many samples in a short time. The pres- ent work also indicates that molecular markers used 642 Fishery Bulletin 104(4) in present study are a practical approach for studies of genetic diversity Acknowledgments This work was supported by the Project NSFC 4030602. NFC 40576064 and 2002n6o9 from The Science & Tech- nology Department of Fujian Province. The first author is grateful to the Chinese Government Scholarship Council and to the Ministry of Human Resource and Development, Govt, of India, for providing financial assistance for this study. Literature cited Ansenio L., I. Gonzalez, A. Fernandez, M. A. Rodriguez, E. Lobo, P. E. Hernandez, T. Gracia, and R. Martin. 2002. Application of random amplified polymorphic DNA (RAPDl analysis for identification of grouper tEpinephelus guaza), wreck fish iPolyprion america- nus). and Nile perch iLates niloticus). Food Protection 65(2);432-435. Baillie, B. K., C. A. Baillie, V. Silvestre, M. Si.son, E. D. Gomez, and V. Monje. 2000. Genetic variation in Symbindiuiuin isolates from giant clams based on random amplified polymorphic DNA (RAPD) patterns. Mar. Biol. 136:829-836. Bardakci. F., and D. O. Skibinski. 1994. Application of the RAPD technique in tilapia fish: species and subspecies identification. Heredity 73:117-123. Bielawski, J. P., and D. E. Pumo. 1997. Randomly amplified polymorphic DNA (RAPD) analysis of Atlantic Coast striped bass. Heredity. 78:117-123. Carvalho, G. R., P. R. Shaw, A. E. Magurran, and B. H. Seghers. 1991. Marked genetic divergence revealed by allozymes among populations of the guppy Poecilia reticulata (Poeciliidiael. Biol. J. Linnean See. 42:389-405. DeSalle, R. A., A. K. Williams, and M. George. 1993. Isolations and characterization of animal mito- chondrial DNA. Methods Enzymol. 224:176-202. Greig, T. W., M. K. Moore, C. M. Woodley, and J. M. Quattro. 2005. Mitochondrial gene sequences useful for spe- cies identification of western North Atlantic Ocean sharks. Fish, Bull. 103:516-523 Hadrys, H., M. Balick, and B. Schierwater. 1992. Application of random amplified polymorphic DNA (RAPD) in molecular ecology. Mol. Ecol. 1:55-64. Harding, G. C., E. L. Kenchington, C. J. Bird, D. S. Pezzack, and D. C. Landry. 1997. Genetic relationship among subpopulations of the American lobster iHoniarus americanus) as revealed by random amplified polymorphic DNA. Can. J. Fish. Aquat. Sci. 54:1762-1771. Heemstra P. L., and J. E. Randall. 1993. FAG species catalogue, vol 19: groupers of the world family Serranidae, subfamily Epinephelinae. An annotated and illustrated catalogue of the grouper, rocked, hind coral grouper and lyretail species known to date. FAO Fish Synopsis 16:1-382. Jones, C. J., K. J. Edwards, S. Castaglione, M. O. Winfiled, F. Sala, C. Van de Weil, G. Bredemeijer, B. Vosman, M. Matthes, A. Daly, R. Brettschneiden P. Bettini, M. Buiatti, E. Maestri, A. Malcevschi, N. Marmiroli, R. Aert, R. J. Volckaert, R. Linacero, A. Vazquez, and A. Karp. 1997. Reproducibility testing of RAPD. AFLP and SSR markers in plants by a network of European laboratories. Mol. Breed. 3:381-390. Lage, C, K. Kuhn, and I. Kornfield. 2004. Genetic differentiation among Atlantic cod from Browns Bank, Georges Bank, and Nantucket Shoals. Fish. Bull. 102:289-297. Liu, Z. J., P. Li, B. J. Argue, and R. A. Dunham. 1999. Random amplified polymorphic DNA markers: usefulness for gene mapping and analysis of genetic variation of catfish. Aquacult. 174:59-68. Martinez, L, E. O. Elvevoll. and T. Haug. 1997. RAPD typing of northeast Atlantic minke whale [Balaenoptera acutorostrata). ICES J. Mar. Sci. 54:478-484. Nadig, S. G., K. L. Lee, and S. M. Adams. 1998. Evaluating alterations of genetic diversity in sun- fish populations e.xposed to contaminants using RAPD assay. Aquatic toxicology 43:163-178. Naish, K. A., M. Warren, F. Bardakci, D. O. F. Skibinski, G. R. Carvalho, and G. C. Mair 1995. Multilocus DNA fingerprinting and RAPD reveal similar genetic relationships between strains of Oreo- chromis niloticus (Pisces, Chichlidae). Mol. Ecol. 4:271-274. Nei, M. 1973. Analysis of genetic diversity in subdivided populations. Proc. Nat. Acad. Sci. U.S.A. 70:3321- 3323. Sadovy, Y. 1997. Problems of sustainability in grouper fisheries. /;; Proceedings of the fourth Asian Fisheries Forum, p. 321-324. China Ocean Press. Beijing. Shannon. C. E., and W. Weaver. 1949. The mathematical theory of communication. Univ. of Illinois Press, Urbana, IL. Ward, R. D., M. Woodwark, and D. O .F. Skibinski. 1994. A comparison of genetic diversity levels in marine, freshwater, and anadromous fishes. J. Fish. Biol. 44:213-232. Welsh. J., and M. McClelland. 1990. Fingerprinting genomes using PCR with arbitrary primers. Nucl. Acids Res. 18:7213-7218. Williams, -J. G. K., A. R. Kubelik, and K. J. Livak. 1990. DNA polymorphisms amplified by arbitary prim- ers are useful as genetic markers. Nucl. Acids Res. 18:6531-6533. 643 Improving the precision of otolith-based age estimates for Greenland halibut {Reinhardtius hippoglossoides) with preparation methods adapted for fragile sagittae Jacob L. Gregg Western Fisheries Research Center Marrowstone Marine Field Station United States Geological Survey 616 Marrowstone Pt Road Nordland, Washington 98358-9633 Delsa M. Anderl (contact author) Daniel K. Kimura National Oceanic and Atmosphenc Administration National Marine Fishenes Service Alaska Fisheries Science Center 7600 Sand Point Way NE Seattle, Washington 98115-6349 Email address for D, M. Anderl; Delsa. Anderla noaagov Otolith-based age estimates for Greenland halibut {Reinhardtius hip- poglossoides) have low precision, and there is general uncertainty about their accuracy in older fish (Anon.'; Alpoim et al.-). Low precision can result from inadequate training of age readers, poor aging criteria, or peculiarities of the structure being aged (Kimura and Lyons, 1991). The latter is the primary cause of low pre- cision with Greenland halibut, and it confounds attempts to improve the former two. Sagittae of Greenland halibut are irregular in shape and exhibit marked bilateral asymmetry ' Anonymous. 1997. Report of the ICES/NAFO workshop on Greenland halibut age determination, Reykjavik, Iceland, 26-29 November 1996. ICES CM 1997/0:1, 53 p. Palaegrade 2-4 DK-1261 Copenhagen K Denmark. 2 Alpoim, R., E. Roman, B. Greene, R. Burry, and W. R. Bowering. 2002. Re- sults of the Greenland halibut iRein- hardtiiis hippoglosoides ) otolith exchange between Spain, Canada, and Portugal. /;; NAFO Scientific Council Meeting, June 2002. NAFO SCR Doc. 02/141, 14 p. P.O. Box 638, Dartmouth, Nova Scotia, Canada B2Y 3Y9. (Fig. 1). Much of this irregularity is due to finger-like projections, which begin as small, marginal tubercles in 4- to 6-year-old fish, and can develop into convoluted, fragile structures in older fish. The variable deposition rate of aragonite and protein that produces these structures makes interpretation of growth patterns difficult and results in age estimates that vary depending on which region of the otolith is examined. The amphiboreal distribution of Greenland halibut has led to their exploitation by the industrial fisher- ies of more than ten nations in the North Atlantic and North Pacific Oceans and by several aboriginal fisheries in the near shore regions of Greenland and northern Canada (Al- ton et al., 1988; Witherell^; Anon.^; Treble''). Age determination (aging) and age structure analysis have been undertaken primarily for North Atlantic and Barents Sea stocks. Methods vary between laboratories but the majority of aging is accom- plished by examining the surface patterns of whole sagittae. For the purposes of this note, "surface" (or "surface aging") will refer to the surface pattern of the whole sagitta. Generally, only the left (i.e., blind side offish) sagitta is aged because it has a more centric nucleus, resulting in more evenly spaced annuli (Lear and Pitt, 1975; Bowering, 1978. 1982; Haug and Gulliksen, 1982; Anon.'; Bowering and Nedreaas, 2001; Al- poim et al.'-'). Attempts to improve the resolution of growth patterns have included baking both sagittae, clearing them with oil, grinding the distal surface of the left sagitta, and breaking and burning the left sagitta (Anon.'; Kuznetsova et al., 2001; Al- poim et al.-). To date these processes have had equivocal effects on the pre- cision of age estimates. Internation- al exchanges of Greenland halibut otoliths have yielded mixed results; reported between-reader agreement (±0 year) has ranged from 1% to 69% (Anon.') and from 37% to 51% (CVs ranging from 5.81% to 9.58%) (Alpo- im et al.-). Despite these exchanges, concern about precision still exists and a consensus on preferred aging methods for Greenland halibut has not been reached. The Alaska Fisheries Science Cen- ter (AFSC) has collected Greenland halibut otoliths from the Bering Sea and Aleutian Islands for over 20 years, but little aging was attempted prior to 2003. Initial examination of the otoliths left AFSC age readers 3 Witherell, D. 2000. Groundfish of the Bering Sea and Aleutian Islands area: species profiles 2001. North Pacific Fisheries Managements Council. 60.5 W. 4"' Ave. Ste. 306 Anchorage, AK 99501-2252. ^ Anonymous. 2004. Annual quota table for 2005. In 2004 Annual report of the Northwest Atlantic Fisheries Organiza- tion, P.O. Box 638, Dartmouth, Nova Scotia Canada B2Y 3Y9. ^ Treble, M. 2005 Personal, commun. Arctic stock assessment biologist. Fish- eries and Oceans Canada, 501 University Crescent, Winnipeg, Manitoba Canada R3T 2N6 Manuscript submitted 12 May 2005 to the Scientific Editor's Office. Manuscript approved for publication 6 January 2006 by the Scientific Editor. Fish. Bull. 104:643-648 12006). 644 Fishery Bulletin 104(4) Left Right PP^^^V^"^ "^^^^^^^^^^^^Ht^^^^ posterior ^^^^^^^^^^^^^^kHB|| ventral ^HBH 10 mm Figure 1 Micrograph of sagittae from a 92-cm Greenland halibut iRelnhardtius hippoglos- soides). View is of the distal surface. Dashed line indicates the plane of the cross section used in this study. with a general lack of confidence in surface age esti- mates. This uncertainty, coupled with the observation that otoliths of larger, presumably older, fish tend to grow in thickness rather than in sagittal diameter, led to pilot work for processing and production (large-scale) aging of Greenland halibut sagittae. Various methods reported in the literature and several new techniques were explored. This pilot work converged on a method that involved cutting the left sagitta in the transverse plane and staining the two restulting cross sections. This method is similar to the break-and-burn method (Chilton and Beamish, 1982) but is more amenable to fragile Greenland halibut sagittae. The goal of the present study was to determine whether the precision of Greenland halibut age estimates could be improved by examining the stained cross sections of their sagittae rather than the surfaces of the whole sagittae, and to determine whether there was a significant difference in age estimates made with each of the two methods. Materials and methods Otoliths were collected from Greenland halibut in July 1998 and June-August 1994 as part of the AFSC Bering Sea and Aleutian Islands trawl surveys. Sagittae were removed at sea and stored in a glycerol-thymol solution until the time of our study. Fish length (TL) was mea- sured to the nearest centimeter for each specimen. Surface aging was accomplished by submerging sag- ittae in water over a black background and counting probable annuli (i.e., translucent zones) with a dissect- ing microscope (6-25x) with reflected light. Probable annuli were counted along several vectors on both sides of the right and left sagittae. A count that was repeat- able, with highest reader confidence, was adopted as the surface age estimate. The left sagitta (i.e., on blind side offish) was embed- ded in clear polyester resin and cut into two pieces with a low-speed saw. The cut was made slightly obliquely to the transverse plane and was adjusted for each oto- lith to ensure that the saw blade bisected the nucleus, passed through a thick section of the perisulcular re- gion (i.e., a portion with a large mediolateral dimen- sion), and extended out the center of a prominent dorsal finger (Fig. 1). The two exposed cross sections were then polished with 800-grit wet-dry sandpaper on a lapidary wheel to remove saw marks. Staining techniques were adapted from Richter and McDermott (1990). Polyester blocks containing cut oto- liths were submerged in a solution of 1% Aniline Blue WS (no. B362-03. Mallinckrodt Baker Inc., Phillipsburg, NJ) in 17f acetic acid. Staining times varied from 10 to 15 minutes initially and were consolidated to 13 min- utes as the experiment progressed. Stain solution tem- perature was maintained between 20° and 23°C. Upon removal from the stain, otoliths were rinsed with fresh water and wiped clean to ensure that residual acid and stain were removed. The two cross sections were covered with mineral oil to eliminate glare, and were examined under a dissecting microscope at 12x to 50x magnification with reflected light. Blue stained translu- cent zones (Fig. 2) were counted, and this number was adopted as the cross-section age estimate. Three trials were conducted to examine the possible benefits of cutting and staining Greenland halibut sagit- NOTE Gregg et al : Otolith preparation methods adapted for fragile sagittae 645 Figure 2 Stained cross section of the left sagitta from an 84-cm Greenland halibut tReinhardtiiis hippoglossoides). Arrows point to presumed annual marks in the perisulcular tuberosity where annuli were most readily resolved. Fish age was estimated at 13 years. tae. In trial 1, a training trial, sagittae of 93 Greenland halibut were examined to compare precision between the two aging methods and to calibrate age readers with respect to the first few annuli on the stained cross sec- tions. This sample included fish that ranged from 12 to 84 cm TL, but it was dominated by smaller fish (mean TL=40 cm, median=31 cm). Two readers independently aged sagittal surfaces and stained cross sections. Sur- face aging necessarily preceded cutting and staining, but no consultation occurred between readers until the end of the trial. Readers were aware of fish length during aging. At the end of trial 1 the independently determined age estimates were compared and readers re-examined otoliths that had resulted in age discrepancies. In trial 2, 226 otolith pairs were examined. This sample contained sagittae from many larger specimens (mean TL=75 cm, median=72 cm, and range=57 to 98 cm). Ages were determined in the same manner as in trial 1. However, after surface aging, readers re-exam- ined discrepancies together and assigned, by mutual agreement, a definitive surface age to each sagitta prior to cutting and staining. Similarly, the cross sections were aged independently and then assigned a definitive cross-section age by mutual agreement. This process is similar to that used for production aging of other species at the AFSC (Kimura and Lyons, 1991) and allowed not only a comparison of precision between methods but also a comparison of the final age estimates that resulted from the two methods. In trial 3, sagittae were exam- ined from 76 Greenland halibut with a size range of 12 to 63 cm TL (mean=37 cm, median=39 cm). This trial was conducted in the same manner as trial 2, with the exception that fish length was not provided to the read- ers. We felt that criticism could arise if length data were known because of the potential for reader bias when ag- ing small fish that fall into distinct size classes. Two age readers performed each trial. Reader 1 was relatively inexperienced with six months of experience aging larval otoliths and one month of experience ag- ing adult otoliths. Reader 2 had 14 years of experience aging several species, including other Bering Sea flat- fishes. Neither reader had previously aged Greenland halibut. Between-reader agreement and coefficient of varia- tion (CV) were calculated for each aging method from each trial. CV was used as the measure of preci- sion (Chang, 1982). Percent agreement is not a good measure of precision because it is highly dependent on the age structure of the sample. Bowker's test of symmetry (Hoenig et al., 1995) was used to assess be- tween-reader bias. Definitive ages from trials 2 and 3 were compared by using a two-tailed matched pairs t-test (Snedecor and Cochran, 1967). Von Bertalanffy growth parameters were estimated from surface and cross-section ages combined from trials 2 and 3. An f -test based on the residuals of nonlinear least-squares fit was used to test for difference between the resulting models (Quinn and Deriso, 1999). 646 Fishery Bulletin 104(4) Results Stained cross sections improved the precision of Green- land halibut age estimates for the larger, presumably older, specimens in trial 2 but did not improve precision of estimates for specimens in trials 1 and 3. Percent CVs were 11.33, 16.31, and 8.11 for surface ages and 19.68, 9.64, and 9.96 for cross-section ages from trials 1, 2, and 3, respectively (Table 1). A similar pattern occurred in the symmetry of age estimates. Bowker's test of symmetry indicated that surface age estimates in trial 2 were significantly biased between age readers (P<0.0342) and that cross-section estimates were not {P<0.2159), whereas in trials 1 and 3 significant bias occurred in the cross-section estimates (P<0.0001 and P<0.0012, respectively) (Table 1). These equivocal results were primarily caused by difficulty interpreting the second annuli on cross sec- tions. Reader 1 tended to count a small diameter mark close to the nucleus as the second year whereas reader 2 considered it a check. A post hoc correction of this bias in trial 1 (i.e., adding 1 year to each of reader- 2's cross-section estimates) yielded better precision (CV=7.68) and no significant bias (P<0.2440) (Table 1). This problem in interpretation occurred in all trials but the resulting bias was most noticeable in trials 1 and 3 where fish age estimates were younger. Definitive cross-section ages were significantly greater (older) than definitive surface ages for trial 2 Trial 2 D surface D cross-section n n n ri nl _ n n_ 6 7 8 9 t) n e B u 5 -B 17 B B 20 21 22 23 34 25 26 27 28 29 30 31 32 33 34 35 36 B Trial 3 LteL D surface D cross-section rfi n Age estimate (yr) Figure 3 Age frequencies for age estimates from sagittal surfaces and from sagittal cross sections of Greenland halibut (Reinhardtius hippo- glossoides) in trial 2 (A) and trial 3 (B) of this study. (^=17.32, df = 225, P<0.0001). Mean cross-section age was 17.1 years and had a range from 9 to 36 years, whereas mean surface age was 12.4 years and had a range from 7 to 28 years (Fig. 3A). Differences be- tween definitive cross-section ages and definitive sur- face ages in trial 3 were not significant (/=1.74, df=74, P<0.0858). Mean stained age was 4.29 years and had a range from 1 to 7 years, and mean surface age was 4.15 years and had a range from 1 to 8 years (Fig. 3B). Von Bertalanffy growth parameters calculated from the definitive surface ages (trial 2 and 3 combined) were L, =103.7, iC=0.104, and t„=-0.333. Parameters from definitive cross-section ages were L_^ = 86.2, /<'=0.125, and /|| = -0.233. The models varied significantly from each other (P=40.58, P<0.0001) (Fig. 4). Discussion In larger Greenland halibut (i.e., in trial 2), the precision of age estimates can be improved by aging stained cross sections of sagittae rather than aging sagittal surfaces (Table 1). The sagittae of larger, older Greenland halibut are very difficult to interpret from the surface. Marginal growth increments are very small and are interrupted by the fingerlike projections on the otoliths. Age readers in our study were more confident in the age estimates they made from stained cross sections. The clearest annuli were encountered in the perisulcular region of left sagittae. This region appears to grow more consistently than other areas of the otolith. Staining allowed resolution of very narrow increments in this region that were not visible on the surface of sagittae. Precision did not increase in trials 1 and 3 (Table 1) because these trials contained many smaller specimens. The benefits of cross-sectioning and staining are not as great in otoliths that are still growing rap- idly in sagittal diameter. The magnitude of the difference in age estimates from whole surfaces and stained cross sections did not exceed 1 year in fish less than 46 cm and did not exceed 2 years in fish less than 57 cm. A second confounding factor was interpreta- tion of the second annuli in cross sections. This consistent one-year bias between read- ers outweighed any improvements that may have resulted from cross-sectioning otoliths in smaller specimens. We feel that more in- terreader calibration and validation of cross- sectioned annuli by the Peterson method (Ricker, 1975) can resolve this problem. The increase in precision in trial 2 was accompanied by age estimates that were significantly greater (older) (Fig. 3A). In 24 cases, the cross-section estimate was 10 or more years greater than the surface age es- timate, and in two cases the cross-section estimate was 22 years greater. The mean -JLa NOTE Gregg et a\ Otolith preparation methods adapted lor fragile sagitlae 647 E 60 surface cross-section - V b surface - V b cross-section 0 5 10 15 20 25 30 35 Age (yr) Figure 4 A comparison of plots of length at age for surface ages and cross-section ages of Greenland halibut iReinharcltius hippoglosso/des). Von Bertalanffy (vbl growth curve fits for surface aging (solid line) and cross-section aging (dashed line) are shown. Table 1 Precision of age estimates made from the surfaces and cross-sections of Greenland halibut iRcinhardtius tae. Data from three trials with mean fish length and collection location indicated. BS = Bering Sea. Al = uppoglossoides) sagit- Aleutian Islands. n Collection year, location Mean fish length (cm) Ageing method Percent CV Between reader % agreement Bowker's test of symmetry- ±0 yr ±1 yr ±2 yr /■ df P< Trial 1 93 1998, BS 40 surface 11.33 44.1 81.7 90.3 30.4 22 0.1091 cross-section 19.68 20.4 66.7 81.7 74.0 28 0.0001 cross-section corrected' 7.68 53.8 81.7 93.5 27.3 23 0.2440 Trial 2 226 1994, AI 75 surface 16.31 16.8 42.0 58.4 94.2 71 0.0342 cross-section 9.46 16.3 44.3 68.1 76.9 68 0.2159 Trial 3 75 1994, AI 37 surface 8.11 65.3 89.3 93.3 19.5 11 0.0532 cross-section 9.96 49.3 90.7 98.7 32.3 12 0.0012 ' Percent CV = coefficient of variation x 100; CV calculated following Chang (1982). ^ P<0.05 from Bowker's test indicates asymmetry of discrepancies between readers (i.e., reader bias). ■' Corrected cross-section statistics from trial 1 are the result of apos( hoc adjustment of reader 2's ages (see text). age estimate increased by 4.7 years for 226 fish with mean length of 75 cm. The oldest surface age estimate was 28 years and the oldest cross-section estimate was 36 years (Fig. 3A). Maximum ages of Greenland halibut reported in the literature rarely exceed 20 years (Alton et al.. 1988). These older ages result in smaller size-at- age and have the effect of decreasing estimates of von Bertalanffy "s L^ (Fig. 4). The older cross-section age estimates for Greenland halibut are consistent with natural mortality estimated by the gonadosomatic index method (M=0.112; Cooper et al., in press). Our maximum cross-sectioned age of 36 years indicates M = 0.115, as opposed to an M = 0.149 as indicated by the maximum surface age of 28 years (Hoenig, 1983). Current natural mortality parameters used in management are 0.18 in the Bering Sea and 648 Fishery Bulletin 104(4) Aleutian Islands (lanelli et al.'') and 0.20 in the North Atlantic (Darby et al.~). These values are more consis- tent with surface age estimates. We feel that age estimates made from stained cross sections are an improvement over surface age esti- mates for Greenland halibut. However, validation of the age estimates produced by these methods is nec- essary. Given the large discrepancies that we encoun- tered in some specimens (10 to 20 years) these ages can be roughly tested with tag-recovery or radiometric methods. The methods used in the present study may have application in other hard-to-age species. They are a practical alternative to serial thin sections because the preparation time is shorter and allows the method to be adapted for production aging. The embedding process also preserves the structure of fragile otoliths which can be damaged during the cutting and break-and-burn processes. Aknowledgments This publication was funded by the Joint Institute for the Study of the Atmosphere and Ocean (JISAO) under NOAA Cooperative Agreement no. NA17RJ1232, con- tribution no. 1143. The project was administered by Don Gunderson. We thank Mark Blaisdell, Lyle Britt, Charles Hutchinson, Chris Johnston, Craig Kastelle, and Jon Short for technical advice on this project. Dan Cooper and Margaret Treble provided valuable Green- land halibut life history information and advice on the manuscript. Literature cited Alton. M. S., R. G. Bakkala, G. E. Walters, and P. T. Munro. 1988. Greenland halibut Reinhardtiua hippoglossoi- des of the eastern Bering Sea and Aleutian Islands region. NOAA Tech. Rep. NMFS 71. 31 p. 6 lanelli, J. N.,T.K. Wilderbuer, and T.M. Sample. 2004. As- sessment of Greenland turbot in the eastern Bering Sea and Aleutian Islands. In Stock assessment and fishery evaluation report for the groundfish resources of the Bering Sea/Aleutian Islands region, p. 427-459. North Pacific Fishery Management Council, 605 W. 4'^ Ave., Ste. 306, Anchorage, AK 99501-2252. ' Darby, C., B. Healey, J. Mahe, and W. R. Bowering. 2004. Greenland halibut {Reinhardtius hippoglossoides) in subarea 2 and divisions 3KLMN0: an assessment of stock status based upon e.xtended survivors analysis, ADAPT, and ASPIC analyses, with stochastic projections of potential stock dynamics. NAFO SCR Doc. 04/55, 53 p. Bowering, W. R. 1978. Age and growth of Greenland halibut, Reinhard- tills hippoglossoides (Walbaum), in ICNAF subareas 2-4. ICNAF Res. Bull. 13:5-10. 1982. Population dynamics of Greenland halibut in the Gulf of St. Lawrence. J. Northw. Atla. Fish. Sci. 3:141-147. Bowering, W. R., and K. H. Nedreaas. 2001. Age validation and growth of Greenland halibut (Reinhardtnis luppoglossoides (Walbaum)): a compari- son of populations in the Northwest and Northeast Atlantic. Sarsia 86:53-68. Chang, W. Y. B. 1982. A statistical method for evaluating the reproduc- ibility of age determination. Can. J. Fish. Aquat. Sci. 39:1208-1210. Chilton, D. E., and R. J. Beamish. 1982. Age determination methods for fishes studied by the Groundfish Program at the Pacific Biological Station. Can. Spec. Publ. Fish. Aquat. Sci., 60 p. Cooper, D. W., K. E. Pearson, and D. R. Gunderson. In press. Annual fecundity and an indirect estimate of natural mortality rate for Reinhardtjus hippoglossoides (Greenland turbot or Greenland halibut) from the Bering Sea and Aleutians Islands. Fish. Bull. Haug, T., and B. Gulliksen. 1982. Size, age, occurrence, growth, and food of Green- land halibut Reinhardtius hippoglossoides (Walbaum) in coastal waters of western Spitzbergen. Sarsia 67:293-297. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mor- tality rates. Fish. Bull. 81:898-903. Hoenig, J. M., M. J. Morgan, and C. A. Brown. 1995. Analyzing differences between two age determi- nations methods by testing symmetry. Can. J. Fish. Aquat. Sci. 52:364-368. Kimura, D. K., and J. J. Lyons. 1991. Between-reader bias and variability in the age- determination process. Fish. Bull. 89:53-60. Kuznetsova, E. N.. M. V. Bondarenko. and A.D. Morozov. 2001. Age composition and growth rate of black halibut Reinhardtius hippoglossoides of the Norwegian-Barents Sea stock. J. Ichthyol. 41(2):143-149. Lear, W. H., and T. K. Pitt. 1975. Otolith age validation of Greenland halibut {Rein- hardtius hippoglossoides). J. Fish. Res. Board Can. 32:289-292. Quinn. T. J., and R. B. Deriso. 1999. Quantitative fish dynamics, 542 p. Oxford LTniv. Press, New York, NY. Richter, H., and J. G. McDermott. 1990. The staining of fish otoliths for age determina- tion. J. Fish. Biol. 36:773-779. Ricker, W. E. 1975. Computation and interpretation of biological sta- tistics offish populations. Fish. Res. Board Can. Bull. 191, 382 p. Snedecor, G. W., and W. G. Cochran. 1967. Statistical methods, 6"^ ed., 593 p. Iowa State University Press, Ames, lA. Acknowledgment of reviewers The editorial staff of Fishery Bulletin would like to acknowledge the scientists who reviewed articles published in 2005-2006. Their contributions have helped ensure the publication of quality science. 649 Dr. Peter B. Adams Mr. Robert J. Allman Dr. Freddy Arocha Dr. Peter J. Auster Mr. Keith Bigelow Dr. Reginald B. Blaylock Dr. James A. Bohnsack Mr. Pablo Bordino Dr. Caren Brady Dr. H. Jane Brockmann Ms. Patience Browne Dr. Jeffrey A. Buckel Dr. Steven X. Cadrin Dr. Gregor M. Cailliet Dr. Steven E. Campana Dr. Ruth Carmichael Dr. Martin Castonguay Mr. Gerald J. Chaput Dr. Enric Cortes Dr. Robert Cowen Dr. Maria M. Criales Dr. Michael W. Davis Ms. Michelle L. Davis Ms. Jennifer L. DeBose Dr. J. Brian Dempson Ms. Kim Dietrich Dr. P. J. Doherty Dr. Martin W. Dorn Dr. C. A. J. Duffy Dr. Charles E. Epifanio Dr. Frank Fish Dr. Kevin D. Friedland Dr. Michael G. Frisk Dr. Francisco Javier Garcia-Rodriguez Dr. William F Gilly Dr. C. Phillip Goodyear Dr. Thomas W. Greig Dr. Thomas M. Grothues Mr. John Gunn Dr. Steven L. Haeseker Dr. Euan S. Harvey Dr. James T. Harvey Mr. David Hata Dr. Emma M. C. Hatfield Dr. Pingguo He Dr. Alexander Hesp Dr. John M. Hoenig Dr. Eric R. Hoffmayer Mr. Ingvar Huse Dr. Todd Kellison Dr. Robert D. Kenney Dr. Donald R. Kobayashi Dr. Michael Koopman Dr. Yutaka Kurita Dr. Jeff Laake Dr. Geoffrey Lang Dr. Vladimir V. Laptikhovsky Dr. Kuo-Tien Lee Ms. Susan A. Little Olcott Dr. Svein Lokkeborg Dr. Robert Gregory Lough Dr. Milton S. Love Dr. Christopher Lowe Mr. Mark S. Lowry Dr. Gustavo J. Macchi Ms. Nancy E. Maloney Dr. Stephen Mayfield Dr. Richard S. McBride Ms. Judy McDonald Dr. Jeremy Mendoza-Hill Dr. David A. Methven Dr. Russell B. Millar Dr. Robert J. Miller Dr. Adam Moles Dr. Lance Morgan Dr. Stephan B. Munch Dr. Craig Mundy Dr. Thomas A. Munroe Mr. Michael D. Murphy Dr. John A. Musick Dr. Lisa J. Natanson Dr. Nathaniel Newlands Dr. Michael O'Connell Dr. JeffOlsen Mr. Torstein Pedersen Dr. Jerome J. Pella Dr. Beth Phelan Dr. Clay E. Porch Dr. Steven M. Porter Ms. Jennifer C. Potts Dr. Michael H. Prager Dr. Thomas P. Quinn Dr. Victor R. Restrepo Dr. Harald Rosenthal Dr. Peter C. Rothlisberg Dr. Eric Schultz Dr. Scoresby A. Shepherd Dr. Peter F Sheridan Dr. Robin L. Sherman Dr. Carl N. Shuster, Jr. Dr. Ned Smith Dr. Derke Snodgrass Dr. Sean Y. Sol Dr. Alan M. Springer Mr. Brian P. Steves Dr. Brent S. Stewart Dr. Ian Stewart Dr. Allan W. Stoner Mr. Ben SulHvan Dr. Douglas P. Swain Dr. Emily Beth Szalai Dr. Barbara L. Taylor Mr. Steven Teo Dr. Verena M. Trenkel Dr. Costas Triantaphyllidis Dr. Thomas Trnski Dr. Robert J, Trumble Dr. Fred M. Utter Dr. Teodoro Vaske Mr. William Watson Dr. Daniel Weihs Dr. Tom Weingartner Dr. Steven G. Wilson Dr. Bruce Wing Mr. Arliss J. Winship Dr. Yimin Ye Dr. Lou Zeidberg 650 Fishery Bulletin 104(4) Fishery Bulletin Index Volume 104(1-4), 2006 List ot titles 104(1) 1 Precision and accuracy offish length measurements obtained with two visual underwater methods, by Marie-Joelle Rochet, Jean-Frangois Cadiou, and Verena M. Ti-enkel 10 The effect of predation (current and historical) by humpback whales (Megaptera novaeangUae) on fish abundance near Kodiak Island, Alaska, by Briana H. Witteveen, Robert J. Foy, and Kate M. Wynne 21 Variation in trawl geometry due to unequal warp length, by Kenneth L. Weinberg and David A. Somerton 35 The effect of autotrawl systems on the performance of a survey trawl, by Stan Kotwicki, Kenneth L. Weinberg, and David A. Somerton 46 The fishery for California market squid (Loligo opal- esce;7S) (Cephalopoda; Myopsida), from 1981 through 2003, by Louis D. Zeidberg, William M. Hamner, Nikolay P. Nezlin, and Annette Henry 60 Variability in supply and cross-shelf transport of pink shrimp iFarfcnitepenaeus duorariim) postlar- vae into western Florida Bay, by Maria M. Criales, John D. Wang, Joan A. Browder, Michael B. Robblee, Thomas L. Jackson, and Clinton Hittle 75 Biology and assessment of the painted sweetlips iDiagramma pictum (Thunberg, 1792)) and the spangled emperor {Lethnnus nebulosus (Forsskal, 1775)) in the southern Arabian Gulf, by Edwin M. Grandcourt, Thabit Z. Al Abdessalaam, Ahmed T. Al Shamsi, and Franklin Francis 89 A catch-free stoack assessment model with appli- cation to goliath grouper (Epinephelus itajara) off southern Florida, by Clay E. Porch, Anne-Marie Eklund, and Gerald P. Scott 102 Fish assemblages found in tidal-creek and seagrass habitats in the Suwannee River estuary, by Troy D. Tuckey and Mark Dehaven 118 Prevalence, intensity, and effect of a nematode iPhilometra saltatrix) in the ovaries of bluefish (Pomatoinus saltatrix), by Lora M. Clarke, Alistair D. M. Dove, and David O. Conover 125 Duration of unassisted swimming activity for spot- ted dolphin iSte?ielIa attenuata ) calves; implications for mother-calf separation during tuna purse-seine sets, by Elizabeth F. Edwards 136 Population structure and variance effective size of red snapper (Lutjanus campechanus) in the north- ern Gulf of Mexico, by Eric Saillant and John R. Gold 149 The relationship between smolt and postsmolt growth for Atlantic salmon iSalmo salar) in the Gulf of St. Lawrence, by Kevin D. Friedland, Lora M. Clarke, Jean-Denis Dutil, and Matti Salminen 104(2) 159 Spawning season, maturity sizes, and fecundity in blacktail comber iSerranus atricauda) (Serranidae) from the eastern-central Atlantic, by Mercedes Garcia-Diaz, Jose A. Gonzalez, Maria J. Lorente, and Victor M. Tuset 167 Benthic invertebrates that form habitat on deep banks off southern California, with special reference to deep sea coral, by Brian N. Tissot, Mary M. Yokla- vich, Milton S. Love, Keri York, and Mark Amend 182 A key to selected rockfishes iSebastes spp. ) based on mitochondrial DNA restriction fragment analysis, by Zhuozhuo Li, Andrew K. Gray, Milton S. Love, Akira Goto, Takashi Asahida, and Anthony J. Gharrett 197 Variation in yellowfin tuna (Thunnus albacares) catches related to El Nirio-Southern Oscillation events at the entrance to the Gulf of California, by Ernesto Torres-Orozco, Arturo Muhlia-Melo, Armando Trasvina, and Sofia Ortega-Garci'a 204 Estimating the encounter rate of Atlantic capelin (Mallotus villosus) with fish eggs, based on stomach content analysis, by Pierre Pepin 215 A production modeling approach to the assessment of the horseshoe crab (Limulus polyphemus) popu- lation in Delaware Bay, by Michelle L. Davis, Jim Berkson, and Marcella Kelly 226 Biodiversity as an index of regime shift in the east- ern Bering Sea, by Gerald R. Hoff 238 Triglops dorothy. a new species of sculpin (Teleostei: Scorpaeniformes; Cottidae) from the southern Sea of Okhotsk, by Theodore W. Pietsch and James W. Orr 247 A comparison of two fishery independent survey programs used to define the population structure of American lobster (Homarus americanus) in the Gulf of Maine, by Yong Chen, Sally Sherman, Carl Wilson, John Sowles, and Minora Kanaiwa List of titles 651 256 Juvenile fish assemblages collected on uncon- solidated sediments of the southeast United States continental shelf, by Harvey J. Walsh, Katrin E. Marancik, and Jonathan A. Hare 278 Growth and maturity of salmon sharks (Lamna dit- ropis) in the eastern and western North Pacific, and comments on back-calculation methods, by Kenneth J. Goldman and John A. Musick 293 Potential causes of mortality for horseshoe crabs (Limuhis polyphemvs) during the biomedical bleed- ing process, by Lenka Hurton and Jim Berkson 299 A study of tagging methods for the sea cucumber Cucumaria frondosa in the waters off Maine, by Sheril Kirshenbaum, Scott Feindel, and Yong Chen 303 Parameterizing probabilities for estimating age- composition distributions for mixture models, by Daniel K. Kimura and Martin W. Dorn 306 Morphometric differentiation in small juveniles of the pink spotted shrimp (Farfantepeneaus brasilien- sis) and the southern pink shrimp {F. notialis) in the Yucatan Peninsula, Mexico, by Marco A. May-Kii, Uriel Ordoiiez-Lopez, and Omar Defeo 311 Exploring intraspecific life history patterns in sharks, by Jason M. Cope 104(3) 323 Standard and routine metabolic rates of juvenile sandbar sharks iCcifcharhimis plumbeiis), includ- ing the effects of body mass and acute temperature change, by W. Wesley Dowd, Richard W Brill, Peter G. Bushnell, and John A. Musick 332 Estimating consumption rates of juvenile sandbar sharks iCarcharhiniis plumbeus) in Chesapeake Bay, Virginia, using a bioenergetics model, W. Wesley Dowd, Richard W. Brill, Peter G. Bushnell, and John A. Musick 343 Depredation of catch by bottlenose dolphins ( Tursiops truncatus) in the Florida king mackerel {Scomb- eromorus cavalla) troll fishery, by Erika A. Zollett and Andrew J. Read 350 Reproductive biology, spawning season, and growth of female rex sole [Glyptocephalus zachinis) in the Gulf of Alaska, by Alisa A. Abookire 360 How much does the fishery at Apo Island benefit from spillover of adult fish from the adjacent marine reserve?, by Rene A. Abesamis, Angel C. Alcala, and Garry R. Russ 376 Comparing the identification of southern California juvenile rockfishes (genus Sebastes spp. ) by restric- tion site analysis of the mitochondrial ND3/ND4 region and by morphological characteristics, by Zhuozhuo Li, Mary M. Nishimoto, Milton S. Love, and Anthony J. Gharrett 383 Potential use of offshore marine structures in rebuilding an overfished rockfish species, bocaccio (Sebastes paucispinis), by Milton S. Love, Donna M. Schroeder, William Lenarz, Alec MacCall, Ann Scarborough Bull, and Lyman Thorsteinson 391 Do oil and gas platforms off California reduce recruitment of bocaccio iSebastes paucispinis) to natural habitat? An analysis based on trajectories derived from high-frequency radar, by Brian M. Emery, Libe Washburn, Milton S. Love, Mary M. Nishimoto, and J. Carter Ohlmann 401 Behavioral ontogeny in larvae and early juveniles of the giant trevally iCarafix ignobilis) (Pisces: Caran- gidae), by Jeffrey M. Leis, Amanda C. Hay, Domine L. Clark, LShiung Chen, and Kwang-Tsao Shao 415 Diurnal and nocturnal temperatures for Atlantic salmon postsmolts iSalmo salar L. ) during their early marine life, by David G. Reddin, Peter Down- ton, and Kevin D. Friedland 428 A prior for steepness in stock-recruitment relation- ships, based on an evolutionary persistence prin- ciple, by Xi He, Marc Mangel, and Alec MacCall 434 Survival of white marlin iTetraptiirus cdbidus) released from commercial pelagic longline gear in the western North Atlantic, by David W. Kerstetter and John E. Graves 445 Application of two methods for determining diet of northern fur seals {CaUorhinus ursinus). by Caro- lyn J. Gudmundson, Tonya K. Zeppelin, and Rolf R. Ream 456 Abundance of adult horseshoe crabs (Limuhis poly- phemus) in Delaware Bay estimated from a bay-wide mark-recapture study, by David R. Smith, Michael J. Millard, and Sheila Eyler 465 Estimates of commercial longline selectivity for Pacific halibut (Hippoglossus stenolepis) from mul- tiple marking experiments, by William G. Clark and Stephen M. Kaimmer 468 Hatching date, nursery grounds, and early growth of juvenile walleye pollock (Theragra chalco- gramma) off northern Japan, by Tsutomu Hat- tori, Akira Nishimura, Yoji Narimatsu, and Daiji Kitagawa 652 Fishery Bulletin 104(4) 476 Development of larval and early juvenile penpoint gunnel (Apodichthys flavidus) (family: Pholidae). by Lisa G. De Forest and Morgan S. Busby 482 Spatial and temporal patterns in the bycatch of sea- birds in the Argentinian longline fishery, by Patricia Gandini and Esteban Frere 104(4) 571 Evaluating light-based geolocation for estimat- ing demersal fish movements in high latitudes, by Andrew C. Seitz, Brenda L. Norcross, Derek Wilson, and Jennifer L. Nielsen 579 Abundance and distribution of whale sharks iRhi- nocodon typus) in the northern Gulf of Mexico, by Carolyn M. Burks, William B. Driggers III, and Keith D. Mullin 489 Geographical differences in the feeding patterns of 585 red rockfish (Sebasfes capensis) along South Ameri- can coasts, by Claudio A. Barrientos, M. Teresa Gonzalez, and Carlos A. Moreno Effects of a simulated fishing moratorium on the stock assessment of red porgy iPagriis pagrusK by Michelle L. Davis, and Jim Berkson 498 The use of otolith morphology to indicate the stock structure of common coral trout (Plectropomus leopardus) on the Great Barrier Reef, Australia, by MikaelaA. J. Bergenius, Gavin A. Begg, and Bruce D. Mapstone 605 512 Biological characteristics and mortality of western butterfish (Pentapodus vitta), an abundant bycatch species of prawn trawling and recreational fishing in 616 a large subtropical embayment, by Jason C. Mant, Michael J. Moran, Stephen J. Newman, S. Alex Hesp, Norman G. Hall, and Ian C. Potter 521 The use of multibeam sonar mapping techniques to 623 refine population estimates fo the endangered white abalone (Hallotis sorenseni ), by John Butler, Melissa Neuman, Deanna Pinkard, Rikk Kvitek, and Guy Cochrane 626 533 Experiments in gear conficutation to reduce by- catch in an estuarine squid-trawl fishery, by James P. Scandol, Tony J. Underwood, and Matt K. 631 Broadhurst 542 The relationships between fish assemblages and the amount of bottom horizontal beam exposed at California oil platforms: fish habitat preferences at 638 man-made platforms and (by inference) at natural reefs, by Milton S. Love, and Anne York 550 A Bayesian method for identification of stock mix- tures from molecular marker data, by Jukka Cor- ander, Pekka Marttinen, and Samu Mantyniemi 643 559 The migration patterns of bluefish (Potnatomus saltatrix) along the Atlantic coast determined from tag recoveries, by Gary R. Shepherd, Joshua Moser, David Deuel, and Pam Carlsen 593 Habitat use of the inner continental shelf off south- ern New Jersey by summer-spawned bluefish (Poma- tomus saltatrix). by David L. Taylor, Peter M. Rowe, and Kenneth W. Able Seasonal and size-based predation on two species of squid by four fish predators on the Northwest Atlan- tic continental shelf by Michelle D. Staudinger Evidence for resource partitioning and competition in nursery estuaries by juvenile flatfish in Oregon and Washington, by Christopher N. Rooper, Donald R. Gunderson, and David A. Armstrong Length-specific brood size and winter parturition in pink seaperch (Zalemhius rosaceus) (Perciformes: Embiotocidae), by Lori H. LaPlante Aging fish otoliths recovered from Pacific harbor seal (Phoca vitulina) fecal samples, by Susan D. Riemer and Robert Mikus Habitat, age, and diet of a forage fish in southeast- ern Alaska; Pacific sandfish (Trichodon trichodon), by John F. Thedinga, Scott W. Johnson, and Donald G. Mortensen Genetic diversity of yellow grouper (Epinephelus aivoara ) determined by random amplified poly- morphic DNA (RAPD) analysis, by Satyendra K. Upadhyay, Wang Jun, Su Yong-Quan. Ding Shao- Xiong, Sonal Chaturvedi Improving the precision of otolith-based age esti- mates for Greenland halibut (Reinhardtius hippo- glossoides) with preparation methods adapted for fragile sagittae, by Jacob L. Gregg, Delsa M. Anderl, and Daniel K. Kimura 653 Fishery Bulletin index Volume 104(1-4), 2006 List ot authors LaPlante, Lori H. 623 Leis, Jeffrey M. 401 Lenarz, William 383 Li, Zhuozhuo 182, 376 Lorente, Maria J. 159 Love, Milton S. 167, 182, 376, 383, 391, 542 Abesamis, Rene A. 360 Able, Kenneth W. 593 Abookire, Alisa A. 350 Al Abdessalaam, Thabit Z. 75 Al Shamsi, Ahmed T. 75 Alcala, Angel C. 360 Amend, Mark 167 Anderl, Delsa M. 643 Armstrong, David A. 616 Asahida, Takashi 182 Barrientos, Claudio A. 489 Begg, GavinA. 498 Bergenias, Mikaela A. J. 498 Berkson, Jim 215,293 Berkson, Jim 585 Brill, Richard W. 323, 332 Broadhurst, Matt K. 533 Browder, Joan A. 60 Bull, Ann Scarborough 383 Burks, Carolyn M. 579 Busby, Morgan S. 476 Bushnell, Peter G. 323, 332 Butler, John 521 Cadiou, Jean-Francois 1 Carlsen, Pam 559 Chaturvedi, Sonal 638 Chen, I-Shiung 401 Chen,Yong 247,299 Clark, Domuie L. 401 Clark, William G. 465 Clarke, Lora M. 118, 149 Cochrane, Guy 521 Conover, David O. 118 Cope, Jason M. 311 Corrander, Jukka 550 Criales, Maria M. 60 Davis, Michelle L. 215, 585 De Forest, Lisa G. 476 Defeo, Omar 306 Dehaven, Mark 102 Deuel, David 559 Dorn, Martin W. 303 Dove, Alistair D. M. 118 Dowd, W. Wesley 323,332 Downton, Peter 415 Driggers, William B. Ill 579 Dutil, Jean-Denis 149 Edwards, Elizabeth F. 125 Eklund, Anne-Marie 89 Emery, Brian M. 391 Eyler, Sheila 456 Feindel, Scott 299 Foy, Robert J. 10 Francis, Franklin 75 Frere, Esteban 482 Friedland, Kevin D. 149, 415 Gandini, Patricia 482 Garcia-Diaz, Mercedes 159 Gharrett, Anthony J. 182, 376 Gold, John R. 136 Goldman, Kenneth J. 278 Gonzalez, Jose A. 159 Gonzalez, M. Teresa 489 Goto,Akira 182 Grandcourt, Edwin M. 75 Graves, John E. 434 Gray, Andrew K. 182 Gregg, Jacob L. 643 Gudmundson, Carolyn J. 445 Gunderson, Donald R. 616 Hall, Norman G. 512 Hamner, William M. 46 Hare, Jonathan A. 256 Hattori, Tsutomu 468 Hay, Amanda C. 401 He.Xi 428 Henry, Annette 46 Hesp, S.Alex 512 Hittle, Clinton 60 Hoff, Gerald R. 226 Hurton, Lenka 293 Jackson, Thomas L. 60 Johnson, Scott W. 631 Kaimmer, Stephen M. 465 Kanaiwa, Minoru 247 Kelly, Marcella 215 Kerstetter, David W. 434 Kimura, Daniel K. 303, 643 Kirshenbaum, Sheril 299 Kitagawa, Daiji 468 Kotwicki, Stan 35 Kvitek, Rikk 521 MacCallAlec 383,428 Mangel, Marc 428 Mant, Jason C. 512 Mantyniemi, Samu 550 Mapstone, Bruce D. 498 Marancik, Katrin E. 256 Marttinen, Pekka 550 May-Kii, Marco A. 306 Mikus, Robert 626 Millard, Michael J. 456 Moran, Michael J. 512 Moreno, Carlos A. 489 Mortensen, Donald G. Moser, Joshua 559 Muhlia-Melo, Arturo 197 Mullin, Keith D. 579 Musick, John A, 278, 323, 332 Narimatsu, Yoji 468 Neuman, Melissa 521 Newman, Stephen J. 512 Nezlin, Nikolay P. 46 Nielsen, Jennifer L. 571 Nishimoto, Mary M. 376, 391 Nishimura, Akira 468 Norcross, Brenda 571 Ohlmann, J. Carter 391 Ordofiez-Lopez, Uriel 306 Orr, James W. 238 Ortega-Garcia, Sofia 197 Pepin, Pierre 204 Pietsch, Theodore W. 238 Pinkard, Deanna 521 Porch, Clay E. 89 Potter, Ian C. 512 Read, Andrew J. 343 Ream, Rolf R. 445 Reddin, David G. 415 Riemer, Susan D. 626 Robblee, Michael B. 60 Rochet, Marie-Joelle 1 Rooper, Christopher N. 616 Rowe, Peter M. 593 Russ, Garry R. 360 Saillant, Eric 136 Salminen, Matti 149 654 Fishery Bulletin 104(4) Scandol, James P. 533 Schroeder, Donna M. 383 Scott, Gerald P. 89 Seitz, Andrew C. 571 Shao, Kwang-Tsao 401 Shao-Xiong, Ding 638 Shepherd, Gary R. 559 Sherman, Sally 247 Smith, David R. 456 Somerton, David A. 21. 35 Sowles, John 247 Staudinger, Michelle D. 605 Taylor, David L. 593 Thedinga, John F. 631 Thorsteinson, Lyman 383 Tissot, Brian N. 167 Torres-Orozco, Ernesto 197 Ti-asvina, Armando 197 Trenkel, Verena M. 1 Tuckey, Troy D. 102 Tuset, Victor M. 159 Underwood, Tony J. 533 Upadhyay, Satyendra K. 638 Walsh, Harvey J. 256 Wang, John D. 60 Wang, Jun 638 Washburn, Libe 391 Weinberg, Kenneth L. 21, 35 Wilson, Carl 247 Wilson, Derek 571 Witteveen, Briana H. 10 Wynne, Kate M. 10 Yoklavich, Mary M. 167 Yong-Quan, Su 638 York, Anne 542 York, Keri 167 Zeidberg, Louis D. 46 Zeppelin. Tonya K. 445 Zollett, Erika A. 343 655 Fishery Bulletin Index Volume 103(1-4), 2005 List ot subjects Abundance horseshoe crab 456 flatfish 616 Aerial survey 579 Age 631 and growth painted sweethps 75 salmon shark 278 western butterfish 512 at maturity 350 structure 643 Age-composition distributions 303 Aging 626 Alaska 445, 631 Albatrosses 482 Allelic variation 136 American lobster 247 Apo Island 360 Apodichthys flavidus 476 Argentina 482 Artificial habitat 383 Atlantic capelin 214 Atlantic salmon 149, 415 Australia 533 Autofocus video camera 1 Autotrawl systems 35 Back-calculation methods 278 Bayesian modeling 550 Bayesian posterior distribution 428 Behavioral ontogeny 401 Benthic invertebrates 167 Biodiversity 226 Bioenergetics model 332 Blacktail comber 159 Blood extraction 293 Blood volume 293 Bluefish 118, 559, 593 Bocaccio 383,391 Body size 623 Bottle squid 533 Bottlenose dolphin 343 Bottom trawl 21, 35 Broad squid 533 Brood size 623 Bycatch 482,512,533 Bycatch reduction devices (BRDs) 533 California 542 California market squid 46 Callorhinus ursinus 445 Canary Islands 159 Canonical correspondence analysis 256 Carangidae 401 Caranx ignobiUs 401 Carcharhinus plumbeus 323, 332 Catch-free stock assessment model 89 Cedar Key 102 Chesapeake Bay, Virginia 332 Circle hook 434 Circuli spacing 149 CODAR ocean sensors Common coral trout 489 Competition 616 Consumption 10 Contuiental shelf 167, 605 Continental slope 167 Coral reef fish 489 Coral reefs 360 Correspondence analysis 256 Cottidae 238 Cross validation 311 Cucumaria frondosa 299 Data storage tag(DST) 415 Data-limited situations 89 Deep sea corals 167 Delaware Bay 215 Demography 75 Depredation 343 Development 476 Diagramma pictiim 75 Diet 445, 631 Dispersal 401 Distribution 616 Diurnal movements 415 Dover sole 626 Drafting 125 Dry Tortugas 60 Eastern Bering Sea 226 Eastern tropical Pacific Ocean 125 Ecosystem effects 332 Ecosystems 167 Effective population size 136 ElNiiio 46 El Niiio-Southern Oscillation 197 Elasmobranch 278, 311, 323, 332, 579 Electronic tag 571 Encounter rates 214 Endangered species 521 Energetic constraints 521 Energetics 125 Energy consumption rates 332 Epinephelus awoara 638 Epinephehis itajara 89 Essential fish habitat (EFH) 256 Estuaries 102, 616 False jacopever 489 Farfantepenaeus brasillensis 306 Farfantepenaeus duorarum 60 Farfantepenaeus notialis 306 Fecundity 118, 159 Feeding habits 489 Feeding strategy 489 Fish 1 Fish assemblages 102 Fish eggs 214 Fisheries data 585 Fisheries management 311, 489 Fishery interaction 343 Fishery yields 360 Flatfish' 21,616 Flatfish capture probability 35 Florida 102 Food habits 445,626 Footrope 35 Forage fish 631 Foraging 10, 626 Fourier analysis 489 Generalized linear models 311 Genetic diversity 638 Genetic stock assessment 136 Genetics 182, 376 Geolocation 571 Giant trevally 401 Glyptocephalus zachirus 350 Goliath grouper 89 Gray's Reef National Marine Sanctuary 256 Great Barrier Reef, Australia 489 Greenland halibut 643 Greenland turbot 643 Groundfish 226 Growth 468,559 Growth rate 278,350 Gulf of Alaska 350 of California 197 of Maine 247 of Mexico 136,579 Gulf red snapper 136 Habitat 167,616,631 Habitat complexity 542 Habitat mapping 521 Habitat preferences 542 Habitat use 256,593 Haliotis soreiiseni 521 656 Fishery Bulletin 104(4) Harvest rate 456 Heart rate 323 Herding 21 Heteroscedastic error models 1 High latitude 571 High-frequency radar 391 Hippoglossus stenolepis 465, 571 Homarus americanus 247 Horseshoe crab 215,293,456 Humpback whale 10 Identification 182. 376 Illex iJlecebrosus 605 Interannual variability 197 Intraspecific life history patterns 311 Japan 468 Juvenile shrimp 306 Juvenile stages 616 Juveniles 401, 476 King mackerel 343 Kodiak Island, Alaska 10, 350 Lamna ditropis 278 Larvae 401,476 Larval behavior 60 Larval transport 60 Laser beam 1 Length at maturity 350 Length measurement 1 Lethriiuis nebulosi/s 75 Life history 278 Limulus amebocyte lysate (LAL) 293 Limulus polyphemiis 215, 293, 456 Loligo opalescens 46 Loligo pealeii 605 Loliolus noctiluca 533 Longfin inshore squid 605 Longline fisheries 482 Longline selectivity 465 Lutjanus campechanus 136 Maine 299 Mallotus villosiis 214 Management 46 Marine protected areas (MPAs) 256 Marine reserves 360 Markovian chain Monte Carlo (MCMC) 550 Mark-recapture 456, 465, 559 Megafaunal invertebrates 167 Megaptera novaeangliae 10 Microsatellites 136 Microstomus pacificus 626 Migrant shorebirds 456 Migration 415, 559 Mitochondrial DNA 182, 376 Mixed effects models 1 Mixture models 303 Modeling 10 Molecular markers 550 Moratorium 585 Morphometric differentiation 306 Mortality rate 75 Mother-calf separation 125 Multibeam mapping 521 Natural mortality 512 Nematode 118 Nemipteridae 512 New Jersey 593 New species 238 Newfoundland 214 Northern fur seal 445 Northern shortfin squid 605 Northwest Atlantic Ocean 214.605 Nursery 332 Nursery gi'ound 468 Oceanography 46 Oil and gas platforms 391, 542 Optimal trawl geometry 21 Optimization 303 Orectolobiformes 579 Orientation 401 Otolith 468,626 morphology 489 preparation 643 sectioning 643 staining 643 Overfishing 383 Oxygen consumption 323 Pacific halibut 465,571 Pacific harbor seal 626 Pacific sandfish 631 Pag/us pagrus 585 Painted sweetlips 75 Pancuronium bromide 323 Parameterizing probabilities 303 Parasite 118 Parturition 623 Pathology 118 Pelagic longline 434 Penpoint gunnel 476 Pentapodus vitta 512 Petrels 482 Philippines 360 Philometra saltatrix 118 Phoca vituUna 626 Pholidae 476 PhotoloUgo etheridgei 533 Pink seaperch 623 Pink shrimp 60 Pink shrimp recruitment 60 Pink spotted shrimp 306 Pinnipeds 445, 626 Platform (oil and gas) decommissioning 542 Plectropomus leopardus 489 Pleuronectidae 350 Pomatomus saltatrix 118, 559, 593 Population estimation 456 Population structure 247 Pop-up archival transmitting satellite tags 571 Pop-up satellite archival tags (PSATs) 434 Postlarval influx 60 Postrelease survival 434 Postsmolt growth 149 Postsmolts 415 Predation 214 Prey biomass removal rate 10 Priors 428 QIO (increase in metabolic rate with temperature) 323 Random amplified polymorphic DNA (RAPD) analysis 638 Recruitment 118, 468 Red porgy 585 Redrockfish 489 Reef fish 585 Regime shift 226 Reinhardtius hippoglossoides 643 Remotely operated vehicles (ROVs) 521 Reproduction 75. 278, 512, 623 Reproductive biology 350 Reproductive potential 118 Resource partitioning 616 Restriction fragment analysis 182 Rex sole 350 Rhincodon typiis 579 Rhincodontidae 579 Rockfish 182, 376, 383, 391, 542 Routine metabolic rate (RMR) 323 Salmo salar 149, 415 Salmon shark 278 Sampling effects 623 Sandbar shark 323,332 Scat 445 Scomberomorus cavalla 343 Sculpin 238 Sea cucumber 299 Sea of Okhotsk 238 Seabirds 482 Seagrass habitats 102 Seasonal distribution 559 Seasonal predation 605 Sebastes capensis 489 Sebastes paiicispiuis 383,391 Sebastes spp. 182, 376 Selectivity 533 Serranidae 159 Sen-anus atricauda 159 List of subiects 657 Sharks 311 Size at maturity 159 -based predation 605 -dependent distribution 247 Smolt gi-owth 149 Smolt size 149 South America 489 Southern pink shrimp 306 Spangled emperor 75 Spawning season 159, 350 Spews 445 Spillover offish from marine reserves into fishery areas 360 Spotted dolphin 125 Squid 533 Standard metabolic rate (SMR) 323 Steepness parameter 428 Sfenella attemiata 125 Stochastic optimization 550 Stock assessment 89,215,428, 585 Stock mixtures 550 Stock rebuilding 383 Stock structure 489 Stock-recruitment relationship 428 Stomach content analysis 214 Stratified random survey design 247 Structure-forming 167 Summer-spawn cohort 593 Surplus production model 215 Survival 149 Survival rate 293 Suwannee River estuary 102 Swimming 323 activity 125 speed 401 Tagging 559 Tagging methods 299 Tetrapturus albidus 434 Theragra chalcogramma 468 Thermal habitat 415 Thunnus albacares 197 Tidal-creek habitats 102 Total mortality 512 Trajectories 391 Trap fisheries 75 Trawl bridles 21 catch efficiency 2 1 fishery 533 geometry 35 survey 35, 247 warp 21 Trichodon trichodon 631 Triglops Dorothy 238 Troll fishery 343 Trophic behavior 489 Tursiops truncata 343 Underwater observation 1 Vertical distribution Viviparous 623 401 Walleye pollock 468 Western butterfish 512 Whale shark 579 White abalone 521 White marlin 434 Yellow grouper 638 Yellowfin tuna 197 Yucatan Peninsula, Mexico 306 Zalernbius rosaceus 623 658 Fishery Bulletin Guidelines for authors Content of manuscripts Contributions published in Fishery Bulletin describe original research in marine fishery science, fishery engi- neering and economics, as well as the areas of marine environmental and ecological sciences (including model- ing). Although all contributions are subject to peer review, responsibility for the contents of papers rests upon the authors and not upon the editor or publisher. Submission of an article implies that the article is original and is not being considered for publication elsewhere. Manuscripts must be written in English. Authors whose native lan- guage is not English are strongly advised to have their manuscripts checked by English-speaking colleagues prior to submission. Articles may range from relatively short contributions (10-15 typed, double-spaced pages, tables and figures not included) to extensive contributions (20-30 typed pages). Notes are reports of 5 to 10 pages without an abstract and describe methods or results not supported by a large body of data. Manuscript preparation Title page should include authors" full names and mailing addresses and the senior authors telephone, fax number, and e-mail address, and a list of key words to describe the contents of the manuscript. Abstract should be limited to 150 words (one-half typed page), state the main scope of the research, and emphasize the author's conclusions and relevant findings. Do not review the methods of the study or list the contents of the paper. Because abstracts are cir- culated by abstracting agencies, it is important that they represent the research clearly and concisely. Text must be typed in 12 point Times New Roman font throughout. A brief introduction should convey the broad significance of the paper; the remainder of the paper should be divided into the following sections: Materials and methods, Results, Discussion (or Conclusions), and Acknowl- edgments. 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, abbre- viations, and technical terms should be written out in full the first time they are used. Include FAO common names for species in the list of keywords and in the introduction. Regional common names may be used throughout the rest of the text if they are different from FAO common names which can be found at http://www.fishbase.org/search. html. Follow the U.S. Government Printing Office Style Manual ( 1984 ed. ) and the CBE Style Manual (6th ed.) for editorial style; for fish nomenclature follow the most cur- rent issue of the American Fisheries Society's Common and Scientific Names of Fishes from the United States and Canada. Dates should be written as follows: 11 Novem- ber 2000. Measurements should be expressed in metric units, e.g., 58 metric tons (t); if other units of measure- ment are used, please make this fact explicit to the reader. Write out the numbers zero through nine unless they form part of measurement units (e.g., nine fish but 9 mm). Text footnotes should be inserted in 9-point font at the bottom of the page that displays the first citation of the footnote. Footnotes should be formatted in the same manner as citations. Footnote all personal communica- tions, unpublished data, and unpublished manuscripts with full address of the communicator or author, or, as in the case of unpublished data, where the data are on file. Authors are advised to avoid references to nonstandard (gray) literature (such as internal, project, processed, or administrative reports, ICES Council Minutes, IWC Minutes or Working Papers, any "research" or "work- ing" documents, laboratory or contract reports, Man- agement Council reports, and manuscripts in review) wherever possible. If these references are used, present them as footnotes and list whether they are available at NTIS (National Technical Information Service) or at some other public depository. Cite all software and special equipment or chemical solutions used in the study, not in a footnote but within parentheses in the text (e.g., SAS, vers. 6.03, SAS Inst., Inc., Cary, NO. Literature cited comprises published works and those accepted for publication in peer-reviewed litera- ture (in press). Follow the name and year system for cita- tion format in the "Literature cited" section. If there is a sequence of citations in the text, list chronologically; (Smith, 1932;Green, 1947; Smith and Jones, 1985). Abbre- viations of serials should conform to abbreviations given in the Serial Sources for the BIOSIS Previews Database. Authors are responsible for the accuracy and complete- ness of all citations. Literature citation format; Author (last name, followed by first-name initials). Year. Title of report or manuscript. Abbreviated title of the series to which it belongs. Always include number of pages. If the authorship for a sequence of citations is identical, list works chronologically. Tables and figures — general format • Zeros should precede all decimal points for values less than one. • Sample size, ;?, should be italicized. • Capitalize the first letter of the first word in all labels within figures. • Do not use overly large font sizes in maps and for units of measurements along axes in figures. • Do not use bold fonts or bold lines in figures. • Do not place outline rules around graphs. • Do not use horizontal lines in graphs to indicate mea- surement units on axes. • Use a comma in numbers of five digits or more (e.g. 13,000 but 3000). • Maps should have a North arrow (or compass sign) and degrees latitude-longitude (e.g., 170°E) Fishery Bulletin 104(4) 659 Tables are often overused in scientific papers; it is seldom necessary or even desirable to present all the data associated with a study. Tables should not be excessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All unusual symbols must be explained in the table legend. Other incidental comments may be footnoted with italic footnote markers. Use asterisks to indicate probability in statistical data. Do not type table legends on a separate page; place them on the same page as the table data. Do not submit tables iu photo mode. Figures include line illustrations, photographs (or slides), and computer-generated graphs and must be cited in numerical order in the text. Graphics will aid in the comprehension of the text, but they should be limited to presenting patterns rather than raw data. Figures are costly to print and should be limited to six. Figures must be labeled with authors name and number of figure. Avoid placing labels vertically (except on y-axis). Figure legends should explain all symbols and abbreviations and should be double-spaced on a separate page at the end of the manuscript. Please note that we do not print graphs in color. FAILURE TO FOLLOW THESE GUIDELINES WILL DELAY PUBLICATION OF A MANUSCRIPT Copyright law does not apply to Fishery Bulletin, which falls within the public domain. However, if an author reproduces any part of an article from Fishery Bulletin in his or her work, reference to source is consid- ered correct form (e.g.. Source: Fish. Bull 97:105). Reprints are available free of charge to the senior author (50 copies) and to his or her laboratory (50 copies). Submission The Scientific Editorial Office encourages authors to submit their manuscripts as a single PDF (pre- ferred), or Word (zipped) document by e-mail to Fish- ery.Bulletin(a!noaa.gov. Please use the subject heading "Fishery Bulletin manuscript submission." Do not send encrypted files. For further details on electronic submis- sion, please contact the Scientific Editorial Office directly (see address below). Or you may send your manuscript on compact disc in one of the above formats along with four printed copies (one original plus three copies [stapled]) to the Scientific Editor, at the address shown below. Dr. Adam Moles Scientific Editor, Fishery Bulletin 11305 Glacier Hwy Juneau. AK 99801-8626 Once the manuscript has been accepted for publication, you will be asked to submit a final software copy of your manuscript. When requested, the text and tables should be submitted in Word or Word Rich Text Format. Figures should be sent as PDF files. Windows metafiles, tiff files, or as EPS files. Send a copy of figures in original software if conversion to any of these formats yields a degraded version. Superintendent of Documents Publications Order Form *5178 I I YES, please send me the following publications: Subscriptions to Fishery Bulletin for $55.00 per year ($68.75 foreign) The total cost of my order is $ . Prices include regular domestic postage and handhng and are subject to change. (Company or Personal Name) I Please type or print l (Additional address/attention linel (Street address i (City, State, ZIP Code) (Daytime phone including area code) 'Purchase Order No.) Charge your order. ITS EASY! ^^^^^^™ Please Choose Method of Payment: I I Check Payable to the Superintendent of Documents n GPO Deposit Account | | | | | | | |— fl I 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! This statement is required by the Act of August 12. 1970. Section 3685, Title 39, U.S. Code, showing ownership, management, and circulation of the Fishery Bulletin, publication number 366-370. and was filed on 31 August. 2006. The Bulletin is published quarterly (four issues annually) with an annual subscription price of $55.00 (sold by the Superintendent of Documents, U.S. Gov- ernment Printing Office, Washington. DC 20402). The complete mailing address of the office of publica- tion is NMFS Scientific Publications Office, NOAA, 7600 Sand Point Way NE, BIN C 15700, Seattle. WA 98115. The complete mailing address of the head- quarters of the publishing agency is National Marine Fisheries Service. NOAA, Department of Commerce, 1335 East-West Highway, Silver Spring. MD 20910. The name of the publisher is Willis Hobart and the managing editor is Sharyn Matriotti; their mailing address is: NMFS Scientific Publications Office, 7600 Sand Point Way NE, BIN 15700, Seattle, WA98115. The owner is the U.S. Department of Commerce. 14th St. N.W. Washington, DC 20230; there are no bondholders, mortgages, or other security hold- ers. The purpose, function, and nonprofit status of the organization (agency) and the exempt status for Federal income tax purposes has not changed during the preceding 12 months. The extent and nature of circulation is as follows: total number of copies (A) (average number of copies of each issue during the preceding 12 months) was 1107 and the actual number of copies of the single issue published nearest to the filing dates was 1101. Paid circulation I B i is handled by the U.S. Govern- ment Printing Office. Washington, DC 20402. and the total number printed for sales (mail subscrip- tions and indiridual sales) was 225 for the average number of copies each issue during the preceding 12 months and 225 the actual number of copies of the single issue published nearest to the filing date (C). Free distribution (D) by mail; samples, compli- mentary and other free copies (average number of copies each issue during the preceding 12 months) was 817 and the actual number of copies of the single issue published nearest to the filing date was 811. Free distribution outside the mail (E) by carriers or other means was 0 for both average number of copies and actual number of copies. 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 D and B ) i average number of copies each issue during the preceding 12 months I was 1042 and the actual number of copies of the single issue published nearest to the filing date was 1036. There were 65 copies lavg. annual) not distributed )H). The total (I: sum of G and H ) is equal to the net press run figures shown in Item A: 1107 and 1101 copies, respectively. I certify that the statements made by me above are correct and complete: (Signed) Willis Hobart. Publisher.