U.S. Department of Commerce Volume 111 Number 4 October 2013 ■I / Fishery Bulletin U.S. Department of Commerce Penelope S. Pritzker Secretary of Commerce National Oceanic and Atmospheric Administration Kathryn D. Sullivan Acting Administrator National Marine Fisheries Service Samuel D. Rauch III Acting Assistant Administrator for Fisheries ★ SrATES of Scientific Editor Bruce C. Mundy Associate Editor Kathryn Dennis National Marine Fisheries Service Pacific Islands Fisheries Science Center Aiea Heights Research Facility 99-193 Aiea Heights Drive, Suite 417 Aiea, Hawaii 96701-3911 Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE Seattle, Washington 98115-0070 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, Seattle, WA 98115-0070. 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: $32.00 domestic and $44.80 foreign. Cost per single issue: $19.00 domestic and $26.60 foreign. See back for order form. Editorial Committee Richard Brodeur John Carlson Kevin Craig Jeff Leis Rich McBride Rick Methot Adam Moles Frank Parrish Dave Somerton Ed Trippel Maty Yoklavich National Marine Fisheries Service, Newport, Oregon National Marine Fisheries Service, Panama City, Florida National Marine Fisheries Service, Beaufort, North Carolina Australian Museum, Sydney, New South Wales, Australia National Marine Fisheries Service, Woods Hole, Massachusetts National Marine Fisheries Service, Seattle, Washington National Marine Fisheries Service, Auke Bay, Alaska National Marine Fisheries Service, Honolulu, Hawaii National Marine Fisheries Service, Seattle, Washington Department of Fisheries and Oceans, St. Andrews, New Brunswick, Canada National Marine Fisheries Service, Santa Cruz, California Fishery Bulletin web site: www.fisherybulletin.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 111 Number 4 October 2013 Fishery Bulletin Contents Articles 293-308 Misa, William F. X. E., Jeffrey C. Drazen, Christopher D. Kelley, and Virginia N. Moriwake Establishing species-habitat associations for 4 eteline snappers with the use of a baited stereo-video camera system 309-324 Cappo, Mike, Ross J. Marriott, and Stephen J. Newman James's rule and causes and consequences of a latitudinal cline in the de- mography of John's Snapper ( Lutjanus johmi) in coastal waters of Australia The National Marine Fisheries Service (NMFS) 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, rec- ommends, or endorses any propri- etary product or proprietary mate- rial mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased be- cause of this NMFS publication. The NMFS Scientific Publications Office is not responsible for the contents of the articles or for the standard of English used in them. 325-336 Macchi, Gustavo J., Ezequiel Leonarduzzi, Marina V. Diaz, Marta Renzi, and Karina Rodrigues Maternal effects on fecundity and egg quality of the Patagonian stock of Argentine Hake (Merlucaus hubbsi) 337-351 Porter, Steven M., and Kevin M. Bailey Using measurements of muscle cell nuclear RNA with flow cytometry to improve assessment of larval condition of Walleye Pollock (Godus chalcogrammus ) 352-369 Wuenschel, Mark J., Kenneth W. Able, James M. Vasslides, and Donald M. Byrne Habitat and diet overlap of 4 piscivorous fishes: variation on the inner continental shelf off New Jersey 370-380 Drymon, J. Marcus, Laure Carassou, Sean P. Powers, Mark Grace, John Dindo, and Brian Dzwonkowski Multiscale analysis of factors that affect the distribution of sharks through- out the northern Gulf of Mexico 381-389 Wenzel, Frederick W., Pamela T. Polloni, James E. Craddock, Damon P. Gannon, John R. Nicolas, Andrew J. Read, and Patricia E. Rosel Food habits of Sowerby's beaked whales (Mesoptodon bidens) taken in the pelagic drift gillnet fishery of the western North Atlantic 390-401 Colmenero, Ana 1., Victor M. Tuset, Laura Recasens, and Pilar Sanchez Reproductive biology of Black Anglerfish ( Lophius budegassa) in the northwestern Mediterranean Sea 402 Errata 403 Acknowledgment of reviewers 404-406 Guidelines for authors 293 Establishing species-habitat associations for 4 eteline snappers with the use of a baited stereo-video camera system Email address for contact author: wfmisa@hawaii.edu 1 Department of Oceanography School of Ocean and Earth Science and Technology University of Hawaii at Manoa 1000 Pope Rd., MSB 205 Honolulu, Hawaii 96822 Present address for contact author: Fisheries Research and Monitoring Division Pacific Islands Fisheries Science Center National Marine Fisheries Service, NOAA 2570 Dole St. Honolulu, HI 96822 2 Hawaii Undersea Research Laboratory School of Ocean and Earth Science and Technology University of Hawaii at Manoa 1000 Pope Rd., MSB 303 Honolulu, Hawaii 96822 Abstract— With the use of a baited stereo-video camera system, this study semiquantitatively defined the habitat associations of 4 species of Lutjanidae: Opakapaka ( Pristipo - moides filamentosus), Kalekale ( P . sieboldii), Onaga ( Etelis coruscans ), and Ehu (E. carbunculus ). Fish abundance and length data from 6 locations in the main Hawaiian Islands were evaluated for species- specific and size-specific differences between regions and habitat types. Multibeam bathymetry and back- scatter were used to classify habi- tats into 4 types on the basis of sub- strate (hard or soft) and slope (high or low). Depth was a major influence on bottomfish distributions. Opak- apaka occurred at depths shallower than the depths at which other spe- cies were observed, and this spe- cies showed an ontogenetic shift to deeper water with increasing size. Opakapaka and Ehu had an overall preference for hard substrate with low slope (hard-low), and Onaga was found over both hard-low and hard- high habitats. No significant habi- tat preferences were recorded for Kalekale. Opakapaka, Kalekale, and Onaga exhibited size-related shifts with habitat type. A move into hard- high environments with increasing size was evident for Opakapaka and Kalekale. Onaga was seen pre- dominantly in hard-low habitats at smaller sizes and in either hard-low or hard-high at larger sizes. These ontogenetic habitat shifts could be driven by reproductive triggers be- cause they roughly coincided with the length at sexual maturity of each species. However, further stud- ies are required to determine causal- ity. No ontogenetic shifts were seen for Ehu, but only a limited number of juveniles were observed. Regional variations in abundance and length were also found and could be related to fishing pressure or large-scale habitat features. Manuscript submitted 11 August 2012. Manuscript accepted 9 July 2013. Fish. Bull. 111:293-308 (2013). doi: 10. 7755/FB. 111.4.1 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necesarily reflect the position of the National Marine Fisheries Service, NOAA. William F. X. E. Misa (contact author)’ Jeffrey C. Drazen1 Christopher D. Kelley2 Virginia N. Moriwake1 The catch of deepwater fisheries comprises a multitude of species that live on continental slopes and deep topographic oceanic structures, such as seamounts, ridges, and banks to depths below 2000 m. In the Indo- Pacific region, deepwater snappers (Lutjanidae), groupers (Serranidae), and jacks (Carangidae) that inhabit deep slopes and seamounts at depths of 100-400 m make up a major com- ponent of this fishery. The deepwater handline or “bottomfish” fishery of Hawaii also targets these groups of fishes (Haight et al., 1993a). Some of the commercially important bot- tomfish species can live in excess of 35 years (Andrews et al., 2011; An- drews et al., 2012) — a longevity that indicates low rates of natural mortal- ity and susceptibility to overfishing (Haight et al., 1993a). Four of these key bottomfish species are the focus of this study: Crimson Jobfish ( Pris - tipomoides filamentosus). Lavender Jobfish ( Pristipomoides sieboldii). Flame Snapper ( Etelis coruscans), and Ruby Snapper (Etelis carbun- culus). In Hawaii, these species are known by a different set of common names, and these names will be used for simplicity throughout this article. Pristipomoides filamentosus is com- monly called Opakapaka, P. sieboldii is called Kalekale, E. coruscans is called Onaga, and E. carbunculus is called Ehu. Opakapaka and Onaga rank first and second in total landed weight and value in the Hawaiian Archipelago, and the smaller spe- cies, Ehu and Kalekale, are abun- dant but lower in value and landings (WPRFMC1). From the late 1980s to early 2000s, the Division of Aquatic Re- sources (DAR) of the Hawaii Depart- ment of Land and Natural Resources (DLNR) and the Western Pacific Re- 1 WPRFMC ( Western Pacific Regional Fish- ery Management Council). 2006. Bot- tomfish and seamount groundfish fisher- ies of the western Pacific region, 2005 annual report, 113 p. [Available from Western Pacific Regional Fishery Man- agement Council, 1164 Bishop Street, Suite 1400, Honolulu, HI 96813.] 294 Fishery Bulletin 111(4) 160WW 159WW 158°0'0"W 157”0'0"W 156WW 155°0'0"W 2 o b z b b o CM Figure 1 Map of the current bottomfish restricted fishing areas (BRFAs) in the main Hawai- ian Islands. Highlighted letters indicate the 6 BRFAs — (B) Niihau, (D) Kaena, (E) Makapuu, (F) Penguin Bank, (H) Pailolo Channel, and (L) Hilo — that were sampled from May 2007 to June 2009 with the use of a baited stereo-video camera system for the study of the habitat associations of 4 snapper species. gional Fishery Management Council (WPRFMC) as- sessed bottomfish stocks in the main Hawaiian Islands (MHI) by calculating their estimated spawning poten- tial ratios (SPRs) from annual commercial catch data and established the critical threshold for designation of a stock in a state of recruitment overfishing at a SPR of 20%. Two bottomfish species, the Onaga and Ehu, had SPRs well below 20% for most of this period (DAR2) and were, therefore, considered to be in a state of recruitment overfishing. In 1996, the Magnuson-Stevens Fishery Conservation and Management Act imposed a mandate on regional fishery councils to restore the stocks of overfished spe- cies to healthy levels (i.e. , SPR >20%) within a 10-year time period. To address this problem, the WPRFMC turned to the DAR, which created 19 bottomfish restrict- ed fishing areas (BRFAs) and prohibited bottomfishing in them (Div. Aquatic Resources, Department of Land 2 DAR (Division of Aquatic Resources). 2006. Hawaii’s bot- tomfish fishery, Land Board briefing paper, 17 p. [Available from Division of Aquatic Resources, Hawaii Department of Land and Natural Resources, 1151 Punchbowl St., Rm. 330, Honolulu, HI 96813.] and Natural Resources, Chapter 13-94, Bottomfish Man- agement, Hawaii Administrative Rules). These BRFAs, which took effect on June 1, 1998, were designed to pro- tect 20% of deepwater areas in the depth range of 100- 400 m, where most Onaga and Ehu are found (Parke, 2007). However, identification of suitable geographic ar- eas for closure was difficult at that time because of a lack of adequate habitat data — a common problem for most deepwater fisheries given the logistical challenges involved in sampling the deep sea. In 2007, the DAR revised the BRFA system with data from surveys conducted with a multibeam sonar system, fishing surveys, and analysis of video collected during surveys with a submersible — all of which pro- vided a great deal of new information on bottomfish habitats. The original BRFAs established in 1998 were retained, expanded, relocated, or opened to fishing, and the 12 BRFAs established in 2007 (Fig. 1) contained sig- nificantly more of the hard, steep habitat believed to be preferred by most bottomfish species (Parke, 2007). This belief was formed on the basis of results from submers- ible and fishing surveys that found some species in the water column adjacent to areas of high relief, such as underwater headlands, ledges, outcrops, and pinnacles Misa et al : Establishing species-habitat associations for 4 eteline snappers 295 (Ralston et al., 1986; Haight et al., 1993a). More recent submersible surveys have supported those studies and have indicated that substrate type may be an impor- tant factor that influences distributions of adult bot- tomfishes (Kelley et ah, 2006). However, information on species-specific and age-specific habitat associations for bottomfishes remains limited. Although the preferred habitat of juvenile Opakapaka has been observed to be soft substrates with little to no relief (Moffitt and Par- rish, 1996; Parrish et al., 1997), variations in habitats between adults and juveniles, if any, have yet to be iden- tified for other species of deepwater bottomfishes. Information that can identify fish-habitat associa- tions is fundamental to fisheries science. In addition to the requirement to improve overfished stocks, the Mag- nuson-Stevens Act required federal fishery manage- ment plans to identify the essential fish habitat (EFH) for their managed species (Rosenberg et al., 2000). The EFH for the bottomfish fishery in Hawaii currently is designated as depths from 0 to 400 m without species- specific habitat requirements, despite the notion that habitat requirements probably differ between bottom- fish species and ontogenetic stage of these species. To guide management decisions on the protection and sustainable use of bottomfish resources in Hawaii, this EFH designation should be as complete and as specific as possible (Kelley et al., 2006). New data are needed to obtain a greater under- standing of the habitat associations of bottomfish species. Common shallow-water sampling techniques, such as diver transects, however, are not logistically feasible at depths below 100 m, and fishing surveys can be destructive to local populations. The need for a different survey method has led to the emergence of baited camera systems as cost-effective, nonextractive tools for the estimation of relative abundances of fish species at depths >100 m (Merritt et al., 2011; Moore et al., 2013). With the use of a baited stereo-video camera sys- tem, we aimed to improve our understanding of the habitat associations of 4 species of bottomfishes, within different size classes, in the MHI. Data specific to each species can be used to assess the amount of suitable habitat present in management areas and to relate catch per unit of effort (CPUE) to habitat type. Most important, through expansion of our understanding of the ecology of bottomfishes, more specific and refined EFH designations can be forged and ecosystem-based management strategies can be further developed. Materials and methods The Bottom Camera Bait Station (BotCam) developed by the Coral Reef Ecosystem Division of the NOAA Pa- cific Islands Fisheries Science Center is a remote, fully automated, baited system with stereo-video cameras; it was designed specifically for nonextractive, fishery- independent sampling of deepwater bottomfish species in their habitat and depth range (Merritt, 2005; Mer- ritt et al., 2011). The method for sampling fish popu- lations with a baited stereo-video camera system has been found to generate more consistent data than have comparable unbaited systems (Harvey et al., 2007), has the ability to detect mobile fish species (Harvey et al., 2007; Watson et al., 2010), and has been determined to be effective in sampling bottomfishes in Hawaii (Ellis and DeMartini, 1995; Merritt et al., 2011). The BotCam is a means by which bottomfish abundance estimates can be made within actual bottomfish habitats and fish lengths can be accurately measured. Upon deployment, the BotCam sits about 3 m off the bottom of the seafloor, and, depending on the depth of deployment, amount of light, and water clarity, the field of view may expand or contract. Moore et al. (2013) es- timated that the visual area sampled by the BotCam was between 4 and 400 m2. The BotCam makes use of ambient light, which allows for an operating depth of up to 300 m and is operational on multiple bottom types, including steep slopes and high relief. In our study, the BotCam recorded 30 to 45 min of continu- ous video at each of the 6 deployment locations. Depth data were taken from a conductivity, temperature, and depth profiler attached to the system. The bait canis- ter attached to the BotCam was filled with -800 g of ground anchovy and squid, a mix that is similar to the bait used by bottomfish fishermen (Merritt et ah, 2011). Bottomfish habitat types in the MHI were charac- terized with multibeam bathymetry and backscatter data that originated from a variety of mapping sur- veys conducted with multibeam sonar systems in and around the MHI since the late 1990s. The U.S. Geo- logical Survey in collaboration with the Monterey Bay Aquarium Research Institute carried out the first sur- vey in the MHI in 1998 (U.S. Geological Survey Digital Data Series DDS-55, http://pubs.usgs.gov/dds/dds-55/ index.html; MBARI Hawaii Multibeam Survey, http:// www.mbari.org/data/mapping/hawaii/index.htm) with a 30-kHz Simrad3 EM 300 multibeam sonar system (Kongsberg Maritime AS, Kongsberg, Norway). Both the bathymetry and backscatter data from this survey were processed at a grid resolution of 20 m. The majority of the remaining data came from subsequent surveys con- ducted from 2002 to 2006 by researchers at the Hawaii Undersea Research Laboratory, University of Hawaii at Manoa, with a 95-kHz Simrad EM 1002 multibeam so- nar system. The editing and processing of raw data were carried out by the Hawaii Mapping Research Group of the University of Hawaii at Manoa using the SABER multibeam editing program (SAIC, Inc., McLean, VA) and other proprietary software. Bathymetry data were processed at a 20-m grid resolution, and backscatter data were processed at either a 10-m resolution or a 20-m resolution, depending on the survey. The processed 3 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA 296 Fishery Bulletin 111(4) data from these cruises have not been made publicly available, with the exception of the bathymetry data that have been incorporated into a 50-m-resolution syn- thesis of the entire MHI that is available from the Ha- waii Mapping Research Group (http://www.soest.hawaii. edu/hmrg/multibeam/index.php). Multibeam backscatter data in grids with a 20-m resolution cannot be used effectively to identify specific substrate types, such as mud, sand, pebbles, cobbles, boulders, and bedrock, because more than one of these substrate types often can be found on the seafloor in an area of 20x20 m. Similarly, more than one type of slope can be found in areas of that size because of the presence of small carbonate ledges, large boulders and blocks, sand dunes, and other small-scale topographic features common to seafloors in the Hawaiian Archipel- ago. Multibeam data values for each grid cell (20x20 m) are typically derived through calculation of either the Gaussian weighted means (bathymetry) or the medians (backscatter) of the sonar footprints within each cell. For these reasons, only 4 general habitat types were de- rived from these multibeam data: hard substrate with high slope (hard-high), hard substrate with low slope (hard-low), soft substrate with high slope (soft-high), and soft substrate with low slope (soft-low). Bathymetry data from the different sonar systems generally were consistent. After a number of slope analyses were conducted in ArcGIS 9.1 (Esri, Redlands, CA), a value of 20° was de- termined to be a reasonable boundary between the high and low slopes that appeared in the bathymetry images. Backscatter data, however, are often inconsistent be- tween systems with different frequencies. Furthermore, the backscatter data used in this study were processed in different ways by different technicians. As a result, boundary values between hard and soft substrates had to be determined on a basis of per system and per cruise. A value of 187 was used as the boundary between hard and soft substrates for the EM 300 data and was vali- dated through examination of video from submersible surveys. Boundary values for the EM 1002 data ranged from -41 to 150 and were established through compari- son of areas of overlap with EM 300 data and analysis of video from submersible surveys. Habitat was classified at a resolution of 200x200 m for areas in and around BRFAs. Polygons for high and low slopes and hard and soft substrates were generated with the Raster calculator in ArcGIS 9.1. Intersects of slope and hardness resulted in polygons for the 4 hab- itat types. A grid cell (200x200 m) was superimposed over these polygons, and the areas of the habitat types within each grid cell were calculated. Each grid cell was assigned a habitat type on the basis of which habitat type was observed in the greatest proportion in that area. A stratified-random sampling approach was used to select locations for BotCam sampling. Although the pur- pose of our study was to evaluate species-habitat as- sociations, another goal of this project was to evaluate population changes inside and outside of BRFAs. This objective affected our sampling design. We used data from 625 deployments of the BotCam conducted inside and outside of 6 of the 12 current BRFAs (Fig. 1) be- tween May 2007 and June 2009. The 6 BRFAs that were sampled are located off Niihau (BRFA B), Kaena (BRFA D), Makapuu (BRFA E), and Penguin Bank (BRFA F), in Pailolo Channel (BRFA H), and outside of Hilo (BRFA L). The Niihau and Hilo BRFAs were areas of contin- ued closure from the initial implementation of BRFAs in 1998. The Makapuu and Penguin Bank BRFAs were expanded versions of smaller preexisting BRFAs from 1998, and the BRFAs off Kaena and in Pailolo Channel were areas newly closed in 2007. The BotCam was lowered to depths of 100-300 m. Although the EFH for deep bottomfishes in Hawaii ex- tends to 400 m, the video cameras work under ambient light to only 300 m, thus limiting the depth range of our sampling. Sampling effort was weighted toward known preferred bottomfish habitats to ensure greater replica- tion where fish densities were expected to be higher. Because previous studies have found bottomfishes asso- ciated with hard substrates, high slopes, or a combina- tion of both (Polovina et al., 1985; Ralston et al., 1986; Haight et al., 1993a; Parke, 2007), for our study, hard- high habitats were considered the most suitable and soft-low habitats the least suitable. To sample a BRFA, 32 BotCam deployments inside and 32 outside but ad- jacent to a BRFA were completed over grids of each habitat type with the following replication: 12 hard- high, 8 hard-low, 8 soft-high, and 4 soft-low. BotCam deployments targeted centroids of randomly selected grid cells (200x200 m) and were kept a minimum of 400 m apart to reduce the likelihood of sampling overlap. In regions where a given habitat type was not pres- ent, sampling intensity was increased in the next most suitable habitat. This approach led to skewed sampling across habitat types in Pailolo Channel because only low-slope habitats were identified at a resolution of 200x200 m. When BotCam deployments did not yield usable video (e.g., no recordings or extremely dark im- agery), the BotCam was redeployed at that location on another day. As often happens during sampling efforts in the field, not all targeted grids were sampled because of weather and equipment issues. In the 2-year sam- pling period covered by this study (2007-09), 4 of the 6 BRFAs (Niihau, Makapuu, Penguin Bank, and Pailolo Channel) were sampled twice and the Kaena and Hilo BRFAs were sampled only once. BotCam video footage was reviewed in the labora- tory to estimate the relative abundance, recorded as the maximum number of a particular species observed in a single frame of video (MaxNo), of Opakapaka, Kalekale, Onaga, and Ehu with VF Deep Portal (Deep Develop- ment Corp., Sumas, WA) and Adobe Premiere Pro CS4 (Adobe Systems, Inc., San Jose, CA) software programs. Fishes were identified to the most specific taxonomic classification possible with a species identification ref- erence (Randall, 2007). MaxNo is a conservative abun- Misa et al. : Establishing species-habitat associations for 4 eteiine snappers 297 dance estimate that avoids the potential problem of counting the same fish multiple times as it re-enters a camera’s field of view. Many studies have determined that MaxNo is positively correlated with fish density (Ellis and DeMartini, 1995; Priede and Merrett, 1996; Willis et ah, 2000; Willis and Babcock, 2000; Yau et al., 2001; Cappo et al., 2003). This parameter also has been found to be highly correlated with the traditional parameter of CPUE used in fishing surveys (Ellis and DeMartini, 1995). MaxNo was recorded for all fishes present in the BotCam video footage, but only data for the 4 species of interest were analyzed. Permutational analysis of variance (PERMANOVA) of the data was performed in Primer 6 (PRIMER-E Ltd., Ivybridge, UK) with PERMANOVA+ (Anderson et al., 2008). With PERMANOVA, the data are not assumed to be normally distributed; therefore, this technique was deemed appropriate for analysis of our data, which in- cluded a highly skewed (overdispersed) relative abun- dance distribution due to an unbalanced experimental design and frequent zero counts. The 4 species consid- ered in our study do not all occupy the entire depth range sampled (Polovina et al., 1985; Haight, 1989; Everson et al., 1989; Merritt et al., 2011). To constrain the data to an appropriate range for each species, the depths at which each species had its greatest MaxNo had to be identified. For the initial analysis, depth was divided into 30-m bins from 90 to 300 m. Relative abundance values were square-root transformed to compensate for numerous zero counts and occasional large numbers. A Euclidean distance matrix was used in the statisti- cal test with a type-III sum of squares. If a significant difference (P<0.05) was observed across depth bins, a subsequent pair-wise PERMANOVA was performed to determine the preferred depths of each species. Subse- quent analyses (MaxNo and fork length [FL] ) were then constrained to the depth preferences identified for each of the 4 species studied. Through identification of habitat preferences, the influence of BRFA location (i.e., combined area inside and outside a BRFA) and protection (i.e., area inside versus outside a BRFA) could not be overlooked. PER- MANOVA in a 3-way crossed design was used to deter- mine how BRFA location (BR, 6 levels, fixed), protec- tion (PR, 2 levels, fixed), habitat type (HA, 4 levels, fixed), and the interaction of these factors affected the relative distribution of each species. MaxNo values were square-root transformed, and the PERMANOVA was run on a Euclidean distance matrix with type-III sum of squares. Where significant results (P<0.05) oc- curred, pair-wise testing was performed to identify spe- cific differences. For individual fish visible in both BotCam cam- eras, FL was measured with stereo-photogrammetric measurement software: Visual Measurement System 7.5 (Geometric Software Pty. Ltd., Coburg, Victoria, Depth (m) Figure 2 Mean relative abundance (MaxNo) with standard error (SE) across 7 depth bins for Opakapaka ( Pristi - pomoides filamentosus), Kalekale (P. sieboldii), Onaga (Etelis coruscans), and Ehu ( E . carbunculus) from surveys of these species conducted in the main Hawaiian Islands from May 2007 to June 2009 with the use of a baited stereo-video camera system. Columns with the same letter are not significantly different from each other (P>0.05, post hoc permutational analysis of variance [PERMANOVA] testing). Error bars indicate ±1 SE of the mean. 298 Fishery Bulletin 1 1 1 (4) Table 1 Results of permutational analysis of variance (PERMANOVA) with relative abundance (MaxNo) data from our sur- veys of 4 species — Opakapaka t Pristipomoides filamentosus), Kalekale (P. sieboldii), Onaga ( Etelis coruscans) , and Ehu (E. carbunculus) — in the main Hawaiian Islands between May 2007 and June 2009. The following factors were tested within the preferred depths of each species: bottomfish restricted fishing area location (BR), protection (PR), and habitat type (HA). Preferred depths are noted in the column head for each species. df=degrees of freedom; P=PERMANOVA P-statistic; P=PERMANOVA P-value. Asterisks indicate statistical significance at P<0.05. Factor Opakapaka (90-210 m) Kalekale (180-270 m) Onaga (210-300 m) Ehu (210-300 m) df F P df F P df F P df F P BR 5 2.86 0.02* 5 2.07 0.09 5 1.54 0.17 5 4.78 0.00* PR 1 0.00 1.00 1 0.07 0.79 1 0.07 0.78 1 0.31 0.58 HA 3 8.28 0.00* 3 1.68 0.18 3 3.87 0.02* 3 2.83 0.04* BRxPR 5 0.63 0.66 5 0.55 0.72 5 0.56 0.70 5 0.81 0.54 BRxHA 13 0.64 0.80 12 1.89 0.06 13 0.69 0.71 13 2.33 0.01* PRxHA 3 0.62 0.59 3 0.87 0.45 3 0.56 0.62 3 0.93 0.42 BRxPRxHA 12 1.02 0.42 10 0.44 0.91 9 0.59 0.76 9 0.58 0.79 Residual 247 282 295 295 Australia) and PhotoMeasure 1.74 (SeaGIS Pty. Ltd., Bacchus Marsh, Victoria, Australia). Measurements of individual fish were taken at the point of MaxNo or at the point in the video where the most fish could be measured to ensure that individuals were not repeat- edly measured at various times during video analy- sis. Replicate measurements were taken for individual fish to increase the accuracy of the measurement. An LED device was used to ensure synchronicity of the video footage from the left and right cameras. A root- mean-square error or residual parallax >10 mm and a precision-to-FL ratio >10% were indicative of inac- curate measurements. To ensure the quality of fish length data, these measurements were removed from the analyses in this study. The same 3-way crossed design from the PERMANOVA of relative abundance (BR, PR, HA) was used to test FLs for each species. Transformation of FLs, however, was not necessary be- cause these data typically were normally distributed. Because only variations in mean length were evalu- ated with the previously described approach, additional analyses were undertaken to investigate size-related changes in habitat association. A linear regression was used to evaluate the relationship between depth and FL for each of the 4 species studied to identify ontoge- netic shifts with depth. As part of our examination of ontogenetic shifts across habitat types, a contingency table (tested with Pearson’s chi-square test) was used to determine whether the size-class distribution of each species was independent of habitat type. Fork lengths were grouped into 10-cm bins. This size interval was chosen to maximize the number of observations in each size bin. Merritt et al. (2011) tested and found mea- surements from BotCam video to be accurate to within 0.3-0. 9 cm, making such a grouping very robust. Results For all 4 species studied, significant differences in rel- ative abundance were found across depth bins (PER- MANOVA, P<0.05). Pair-wise comparisons of MaxNo from the 7 depth bins highlighted the depth preference of each species (Fig. 2). MaxNo was highest from 90 to 210 m for Opakapaka (post hoc PERMANOVA, P<0.05). The preferred depths of Kalekale were 180-270 m, and both Onaga and Ehu had the deepest range among spe- cies at 210-300 m (post hoc PERMANOVA, P<0.05). Within the preferred depths of a species, either BRFA location, habitat type, or the interaction of these 2 factors had an effect on the relative abundance of 3 of the 4 species studied (Table 1). Protection and the interaction of all other factors with protection, how- ever, did not have an effect (PERMANOVA, P>0.05). BRFA location and habitat type were each significant factors for Opakapaka. Hilo had the highest relative abundance of this species among sampled locations, and hard-low habitats yielded greater abundance esti- mates for Opakapaka than other habitat types (Fig. 3; post hoc PERMANOVA, P<0.05). Although no signifi- cant location or habitat effects were observed for Kale- kale, the interaction of BRFA location and habitat type was marginal (P=0.06; Table 1); 2 of the largest counts of this species (100 and 85 individuals) occurred on hard-high habitats at Niihau and led to a high mean MaxNo (Fig. 3). Habitat type was the only factor that affected the relative abundance of Onaga. Hard substrate habitats, with either high or low slope, had greater mean MaxNo for Onaga than soft substrate habitats (Fig. 3; post hoc PERMANOVA, P<0.05). BRFA location, habitat type, and the interaction of these 2 factors were significant Misa et a!.: Establishing species-habitat associations for 4 eteline snappers 299 LLJ CO + 15 12 9 - 6 - 3 0 130 25 20 15 10 5 0 i hard-high a hard-low □ soft-high □ soft-low ®BRFA Opakapaka (90-210 m) 30 7 24 2 63 8 12 5 6 31 25 21 25 7 78 9 15 14 5 43 0 25 0 8 33 8 14 13 7 42 80 94 81 35 Kalekale (180-270 m) 22 17 14 10 63 5 14 7 3 29 19 10 15 7 51 31 12 18 12 73 0 77 0 5 82 16 0 7 1 24 93 130 61 38 Habitat Mean relative abundance (MaxNo) with standard error (SE) by location of bottomfish restricted fishing area (BRFA; i.e., combined area inside and outside a BRFA), by habitat type combined for all BRFA locations (habitat), and by habitat type in each BRFA location (BRFAxhabitat) for Opakapaka (Pristipomoides filamentosus), Kalekale (P. sieboldii), Onaga (Etelis coruscans), and Ehu (E. carbunculus) within the preferred depths of each species in the main Hawaiian Islands. A baited stereo-video camera system (BotCam) was used to collect data from May 2007 to June 2009. Columns with the same letter (uppercase type for BRFA; lowercase, bold, italic type for hab- itat; lowercase type for BFRAxhabitat) are not significantly different from each other (P>0.05, post hoc permutational analysis of variance [PERMANOVA]). The number below each column is the number of BotCam deployments. The 4 habitat classifications used in our study were derived from data collected with multitaeam sonar systems: hard substrate with high slope (hard-high), hard substrate with low slope (hard-low), soft substrate with high slope (soft-high), and soft sub- strate with low slope (soft-low). The 8 sampled BRFAs were (B) Niihau, (D) Kaena, (E) Makapuu, (F) Penguin Bank, (H) Pailolo Channel, and (L) Hilo. Error bars indicate ±1 SE of the mean. 300 Fishery Bulletin 111(4) Table 2 Results of permutational analysis of variance (PERMANOVA) with fork length data from our surveys of 4 species — Opakapaka (Pristipomoides filamentosus ), Kalekale (P. sieboldii), Onaga ( Etelis coruscans), and Ehu (E. carbuncu- lus ) — in the main Hawaiian Islands between May 2007 and June 2009. The following factors were tested within the preferred depths of each species: bottomfish restricted fishing area location (BR), protection (PR), and habitat type (HA). Preferred depths are noted in the column head for each species. df=degrees of freedom; P=PERMANOVA F- statistic; P=PERMANOVA P-value. Asterisks indicate statistical significance at P<0.05. Factor Opakapaka (90-210 m) Kalekale (180-270 m) Onaga (210-300 m) Ehu (210-300 m) df F P df F P df F P df F P BR 5 36.04 0.00* 5 28.20 0.00* 4 11.05 0.00* 4 4.90 0.00* PR 1 14.24 0.00* 1 1.43 0.23 0 No test 0 No test HA 3 11.39 0.00* 3 18.38 0.00* 1 0.48 0.49 2 1.77 0.17 BRxPR 5 2.02 0.08 3 16.57 0.00* 3 4.82 0.00* 4 0.84 0.52 BRxHA 9 7.66 0.00* 4 1.16 0.33 1 23.69 0.00* 5 1.62 0.16 PRxHA 2 0.45 0.64 2 0.21 0.82 2 0.48 0.61 2 1.31 0.27 BRxPRxHA 5 3.42 0.01* 2 0.21 0.81 1 13.26 0.00* ONo test Residual 419 446 242 274 for Ehu. The highest relative abundance for this spe- cies was in Pailolo Channel, and the lowest levels were seen at Niihau, Kaena, and Makapuu (Fig. 3; post hoc PERMANOVA, P<0.05). Overall, hard-low habitats had significantly greater numbers of Ehu than did other habitat types. By BRFA location and habitat type, the mean MaxNo of Ehu in Pailolo Channel was higher for hard-low than for soft-low habitats, and similar abundance estimates were found for hard-high, hard- low, and soft-high habitats on Penguin Bank. Niihau and Kaena differed from the other sampled locations in that hard-high habitats had a greater relative abun- dance of Ehu than did hard-low habitats. In our evaluation of mean lengths, BRFA location, protection, and habitat type were all important factors, and the interactions between them were sometimes significant (Table 2). BRFA location, protection, habitat type, the interaction of BRFA location and habitat type, and the interaction of all 3 factors were significant for Opakapaka. Niihau had the largest Opakapaka on average (65.29 cm FL) among sampled locations, and the smallest Opakapaka (28.35 cm FL; Fig. 4; post hoc PERMANOVA, P<0.05) were seen at Hilo. The smallest individual at Hilo measured ~16 cm FL, and the largest individual at Niihau was ~79 cm FL. Opakapaka from outside protected areas had a mean length of 42.89 cm FL and were larger than those fish observed inside the sampled BRFAs (40.53 cm FL; PERMANOVA, P<0.05). The smallest mean lengths of this species were found over hard-low habitats compared with other habitat types overall, other habitats at each BRFA location, and other habitats either inside or outside a particular BRFA (Fig. 4; Table 3; post hoc PERMANOVA, P<0.05). BRFA location, habitat type, and the interaction of BRFA location and protection were significant for Kale- kale. Pair-wise comparisons showed that this species had its smallest mean length (23.64 cm FL) at Kaena, was largest in hard-high habitats (31.46 cm FL) and smallest in soft-low habitats (8.64 cm FL, n= 2), and was larger inside the Penguin Bank and Pailolo Chan- nel BRFAs and outside the Hilo BRFA than in other sampled areas (Fig. 4; Table 3; post hoc PERMANOVA, P<0.05). The smallest individual Kalekale, however, measured 7.63 cm FL at Niihau. BRFA location, the interaction of BRFA location with protection, the in- teraction of BRFA location with habitat type, and the interaction of all 3 of these factors were significant for Onaga. Mean length for Onaga was smallest in Pailolo Channel (42.80 cm FL) than at other locations (Fig. 4) but larger inside the Pailolo Channel BRFA than out- side this protected area (Table 3; post hoc PERMANO- VA, P<0.05). The smallest individual Onaga measured 15.05 cm FL. Although the interaction of BRFA loca- tion and habitat type and the interaction of BRFA location, protection, and habitat type had significant results for Onaga, no clear trends were seen. BRFA location was the only factor that had an influence on mean length for Ehu (Table 2; PERMANOVA, P< 0.05). Overall, mean sizes were very similar for this species but were smallest at Makapuu and Hilo (Fig. 4). For all sampled locations combined, size-related shifts in species-habitat associations were evident. The linear regressions of FL against depth for each species showed that size increased with depth for Opakapaka (coefficient of determination [r2] =0.438, PcO.Ol) but did not for the other 3 species (Fig. 5). In our evaluation of the proportion of fish measured in each habitat type by size class, habitat associations clearly varied by size for Opakapaka, Kalekale, and Onaga (Fig. 6). Ehu had very similar habitat associa- ,Vi Isa et al.: Establishing species-habitat associations for 4 eteline snappers 301 T' a c/3 + E o o> c 0) o i hard-high ■ hard-low □ soft-high □ soft-low sBRFA , Opakapaka (90-210 m) , 37 1 2 0 40 13 19 2 0 34 32 62 7 0 101 22 50 10 6 88 0 28 0 0 28 24 114 18 3 159128 274 39 9 Kalekale {180-270 m) acd; ! C ' 13116 11 2 ISO 21 0 8 0 29 0 1 0 0 1 58 13 74 0 145 0 85 0 0 85 40 0 7 0 47 250115100 2 Onaga (210-300 m) , 23 39 0 0 62 0 3 0 0 3 1 7 0 0 8 43 14 2 0 59 0 1 17 0 0 1 17 10 0 0 0 10 77 180 2 0 Habitat Figure 4 Mean fork length with standard deviation (SD) by location of bottomfish restricted fishing area (BRFA; i.e., combined area inside and outside a BRFA), by habitat type combined for all BRFA locations (habitat), and by habitat type in each BRFA location (BRFAxhabitat) for Opakapaka (. Pristipomoides filamentosus), Kalekale (P. sieboldii), Onaga (Etelis coruscans ), and Ehu (E. carbunculus ) within the preferred depths of each species in the main Hawaiian Islands. A baited stereo-video camera system (BotCam) was used to collect data from May 2007 to June 2009. Columns that have the same letter (uppercase type for BRFA; lowercase, bold, italic type for habitat; lowercase type for BRFAxhabitat) are not significantly different from each other (P>0.05, post hoc permutational analysis of variance [PERMANOVA] testing). Number below each column is the number of fish measured. For protection effects, refer to Table 3. The 4 habi- tat classifications used in our study were hard substrate with high slope (hard-high), hard sub- strate with low slope (hard-low), soft substrate with high slope (soft-high), and soft substrate with low slope (soft-low). The 8 sampled BRFAs were (B) Niihau, (D) Kaena, (E) Makapuu, (F) Penguin Bank, (H) Pailolo Channel, and (L) Hilo. Error bars indicate ±1 SD of the mean. 302 Fishery Bulletin 111(4) Table 3 Summary of significant comparisons from post hoc permutational analysis of variance (PERMANOVA) of fork lengths for bottomfish restricted fishing area location (BR), protection (PR), habitat type (HA), and the interaction of these factors for Opakapaka ( PristipomoicLes filamentosus), Kalekale (P. sieboldii), Onaga ( Etelis coruscans), and Ehu (E. carbunculus ) within the preferred depths of each species from our study of these species in the main Hawaiian Islands between May 2007 and June 2009. Locations of the 6 BRFAs where sampling was conducted are the following: Niihau (B), Kaena (D), Makapuu (E), Penguin Bank (F), Pailolo Channel (H), and Hilo (L). Protection is designated as inside (in) or outside (out) a BRFA. Habitat types are hard-high (HH), hard-low (HL), soft-high (SH), soft-low (SL). NS=nonsignificant comparisons. Preferred depths are noted under the species name in the first column. BR PR HA BRxPR BRxHA PRxHA BRxPRxHA Opakapaka (90-210 m) Largest in B Smallest in L Larger outside Smallest in HL NS (D) largest in SH, smallest in HL (E) largest in SH, smallest in HL (F) largest in high slope, smallest in low slope (L) largest in SL, smallest in HL NS (D in) SH>HL (E in) HH,SH>HL (E out) SH>HH>HL (F in) HH,SH>HL (F out) HH,SH>SL (L in) HH>HL (L out) SL>HH>SH>HL Kalekale (180-270 m) Smallest in D NS Largest in HH (F) larger inside (H) larger inside (L) larger outside NS NS NS Onaga (210-300 m) Smallest in H No test NS (H) larger inside (B) larger in HL than HH (F) similar mean size NS (B in) HL>HH (F in) HH>HL Ehu (210-300 m) Similar mean size No test NS NS NS NS No test tions in all size classes and did not show any habitat shifts with size (Pearson’s chi-square, P>0.05). Opak- apaka had a shift from hard-low habitats to hard-high habitats with an increase in size. There was a greater proportion of sexually mature individuals (>43 cm FL; Kikkawa, 1984) for this species over hard-high habi- tats, and individuals <43 cm FL were seen mostly in hard-low habitats. Although less evident than the habitat shift by Opakapaka, a habitat shift by Kale- kale to hard-high from other habitat types was ob- served within the size class of 25-35 cm. Onaga and Ehu were recorded mostly in hard-low habitats in all size classes. For Onaga, however, the smallest individ- uals (<55 cm FL) were found only in hard-low habi- tats, and, as size increased, hard-high habitats were equally dominant for this species. Discussion Depth has a significant influence on the distribution of bottomfishes in Hawaii. Two distinct depth group- ings were seen within the sampling range of this study. Opakapaka was dominant in the shallower end of the sampling depths (<200 m), and Kalekale, Onaga, and Ehu were observed more frequently toward the deep- er end (>200 m). This finding is consistent with that of previous studies in Hawaii (Haight, 1989; Everson et al., 1989; Merritt et al., 2011) and in the Mariana Archipelago (Polovina et al., 1985). When establishing species-specific differences in distribution, depth must be the first factor evaluated. Although the limitations of our sampling methods have been discussed in previous studies (e.g., Mer- ritt et al., 2011; Moore et al., 2013), it is important to review them here before further discussion of our results. The absence of a quantifiable sampling area, variability in the field of view of the BotCam, and the scale at which habitats were classified are confounding factors that limit the interpretation of the results of this study to a semiquantitative nature. Because the BotCam makes use of ambient light and because envi- ronmental conditions, such as water clarity can differ from site to site, variability in the visual area sampled was unavoidable. However, unlike other visual survey methods, where quadrats or transect lines are used, this approach reduces, but does not eliminate, the ef- fect of visual area because it relies on attracting fishes close to the cameras. What may be more important is the effect of the attracting bait-odor plume. It was our working assumption that any fish seen on BotCam video was from the targeted grid area /Vi j 3 3 et a!.: Establishing species-habitat associations for 4 eteline snappers 303 Opakapaka (90-210 m) o. CD Q 300 250 200 150 100 -j 50 Kalekaie (180-270 m) VL. . ^=0.011 0 20 40 60 80 100 Fork length (cm) Q. 0.05) d=0 d=5 d=43 d=83 d=45 d=9 d=0 d=0 d=0 n-0 n-5 n=82 n-135 n=65 n=9 n-0 n-0 n-0 05-15 15-25 25-35 35-45 45-55 55-65 65-75 75-85 85-95 Size class (cm) Figure 6 Proportion of fish found in each habitat type by size class tested with Pearson’s chi-square (%2) test for Opakapaka (Pristipomoid.es filamentosus), Kalekale (P. sieboldii), Onaga (Etelis coruscans), and Ehu ( E . carbunculus) within the preferred depths of each species in the main Hawaiian Islands. A baited stereo-video camera system (BotCam) was used to collect data from May 2007 to June 2009. The 4 habitat classifications used in our study were hard substrate with high slope (hard-high), hard substrate with low slope (hard-low), soft substrate with high slope (soft-high), and soft substrate with low slope (soft-low). }} (4) where Lf = mean FL (in millimeters) of fish of age t (in years); L„ = asymptotic mean length; K = is a rate constant that determines the rate at which Lf approaches LTC; t = age of a fish; and ty = the hypothetical age at which the mean length is zero. The fit of the VBGF to different data sets was com- pared by using the likelihood ratio test for coincident curves (Cerrato, 1990) across comparable age ranges (Haddon, 2001) and analysis of covariance (ANCOVA) with type-III sums of squares and with loge-trans- formed age (year) as the linear covariate. This ANCO- VA allowed for testing of an interaction of sexxregion and accounted for type-I errors. 312 Fishery Bulletin 1 1 1 (4) Otolith weight at age and gonad weight at length Nonlinear least squares estimation also was used to fit second-order polynomial functions to regional esti- mates of otolith weight at age t (in years) and an ex- ponential function to gonad weight ( Wq ) at FL (Lp in millimeters). The functions were defined with relevant starting parameters (a,b,c) by the following equations: Otolith weight = a + b it) + c ( t 2); (5) Gonad weight = e(a + b ^p'1. (6) Quantile-quantile normal plots and Cook’s Distance were used to identify outliers for exclusion, and plots of residuals were used to test for lack of homogeneity in variances. Comparisons in the 2 responses by sex or region were restricted to the range of explanatory data (sizes or ages) common to each level in the comparison with likelihood ratio tests for coincident curves and ANCOVA with type-III sums of squares. Loge transfor- mations were used to linearize the gonad weight and otolith weight covariates for the ANCOVA. Data from the Arafura Sea were too few for use in these tests and were compared visually with the other regions. Australian and international fishing records An Internet search for record sizes of John’s Snapper and other large lutjanids landed by line and spearfishing was conducted for countries in the Indo-West Pacific, but only world and Australian records were available. The 2011 records maintained by the International Game Fishing Association (IGFA), Australian National Sportfishing As- sociation (ANSA), Australian Angler’s Association (AAA), and Australian Underwater Federation (AUF), were used 60 50 40 30 20 10 0 25 20 15 10 5 0 North Queensland n=216 B Cape York n= 63 Kimberley n= 568 Gin. 60 -i 50 40 30 20 10 ^ ^ ^ V ^ ^ t?N ^ ^ <£>S eft' Fork length (mm) nni /?= 258 Urn Number of annuli Figure 2 The length- and age-frequency distributions of John’s Snapper (Lutjanus johnii ) sampled dur- ing the period of February 1989-April 2002 in 4 regions of Australia: (A) north Queensland, (B) Cape York, (C) Kimberley, and (D) Arafura Sea. The y-axes represent the numbers of fish. Lengths are fork lengths measured in millimeters, and age is measured as the number of an- nuli (annual growth rings) observed in sectioned otoliths. n=sample size. Cappo et al. : Causes and consequences of a latitudinal cline in the demography Lutjanus johnu 313 Table 1 Fitted parameter estimates for the von Bertalanffy growth models for males (M), females (F), and all John’s Snapper ( Lutjanus johnii) sampled from the Kimberley, north Queensland, and Cape York regions in Australia during the period of February 1989-April 2002 (see Fig. 3, for graphs of growth curves by sex and region). L^=mean fork length (mm) of fish of age t (years), L0<)=asymptotic mean length (mm), tQ=the hypothetical age at which the mean length is zero, K is the growth coefficient at which approaches L^. Standard errors of fitted parameter estimates are reported in parentheses, and the coefficient of multiple determination (R2) and sample size (n) are also given. Region n Loo K to R2 Kimberley (M) 255 677.1 (20.53) 0.17 (0.02) -0.81 (0.33) 0.76 Kimberley (F) 294 739.6 (27.05) 0.16 (0.02) -0.62(0.29) 0.79 North Queensland (M) 80 819.7 (25.27) 0.18 (0.02) -0.32 (0.36) 0.85 North Queensland (F) 74 851.4 (37.37) 0.21 (0.03) 0.34 (0.29) 0.83 Kimberley 568 698.0 (14.48) 0.18 (0.01) -0.51 (0.20) 0.79 North Queensland 216 843.5 (15.66) 0.19 (0.01) 0.02 (0.16) 0.87 Cape York 63 685.1 (50.04) 0.18 (0.05) -0.82(0.80) 0.74 to plot the maximum weights of John’s Snapper and 6 other large lutjanids with latitude. These other spe- cies were Mangrove Jack (Lutjanus argentimaculatus), Twospot Snapper (L. bohar), Malabar Snapper (L. mal- abaricus), Emperor Snapper (L. sebae), Chinaman Fish ( Symphorus nema top horns), and Green Jobfish ( Aprion virescens ). Results Length and age distributions Samples from north Queensland had the smallest (55 mm FL) and largest (990 mm FL) fish, a higher modal FL (401-500 mm ) than the samples from the Kim- berley and Cape York regions (301-400 mm FL), and a much higher proportion of very large John’s Snap- per than samples from the other 3 regions (Fig. 2). The overall proportion of samples that were older than 8 years in the Kimberley region (11.6%) was small- er than the proportion of such samples in the north Queensland region (18.2%) but similar to the propor- tion in the Cape York region (11%). The modal age (4 years) for fish from north Queensland was higher than the modal age of fish from Kimberley (3 years), and the 6-year age class was most prevalent in the sam- ples from Cape York. The oldest year classes were from north Queensland (28 year), Kimberley (23 year), and Cape York (18 year). Mortality The maximum observed age (£max) of 28.6 years (a male) was used to produce an estimate of Z=0.146 year-1 with the Hoenig (1983) equation for the north Queensland region. Estimates of Z were not calculated for the other regions because of concerns that the £max values might be skewed by a lack of older year classes that represented undersampling and not truncation of the true age distribution. Growth The von Bertalanffy growth curves fitted to length-at- age data revealed a relatively moderate growth trend (Fig. 3). This trend was reflected in estimates for the K curvature parameter, which ranged from 0.17 to 0.21 year-1 (Table 1). Relatively rapid growth from 0.5 year to 5-7 years was followed by a slower phase after ap- proximately 7-10 years. Asymptotic growth began at approximately 18-20 years (Fig. 3). Females were estimated to grow to a larger aver- age asymptotic length (LJ) in the Kimberley and north Queensland regions (Fig. 3; Table 1). However, sex-spe- cific differences in growth curvature, K, were inconsis- tent between regions; a higher K value was evident for males than for females in the Kimberley region and a lower K value for males than for females in north Queensland (Table 2). The north Queensland was 17.3% (145.5 mm FL) and 18.8% (158.4 mm FL) higher than the for samples from the Kimberley and Cape York regions (Table 1; Fig. 3). Apparent differences in growth trends between sexes and regions were observed to be statistically significant in both the likelihood ratio and ANCOVA tests (Table 2). Significant differences between sex-specific curves were detected within each of the Kimberley and north Queensland regions (Table 2). The lack of significant interaction for sex x region indicates that the larger Loo for females than for males was consistent among regions. Likelihood ratio tests revealed significant differ- ences in growth curves between samples from the 314 Fishery Bulletin 111(4) Kimberley: Females B Kimberley: Males North Queensland: Females D North Queensland: Males Cape York North Queensland and Kimberley Figure 3 Fits of the von Bertalanffy growth function by sex and region for John’s Snapper (Lutja- nus johnii) sampled during the period of February 1989-April 2002 in northern Australia. Females (gray triangles) and males (open circles) are distinguished in separate curves for the (A, B) Kimberley and (C, D) north Queensland regions. A single curve describes the small sample size (n) for (E) Cape York, where a large proportion of fish were of unknown sex (shaded squares). Significantly different growth curves describe eastern and western fish (F) when sexes were pooled. For more information on growth estimates and compari- sons from these models, see Tables 1 and 2. Kimberley and north Queensland regions for each sex, and a significant main effect of region on the slope of transformed length at age (Table 2). The larger L „ and K for samples from north Queensland, therefore, were significantly different from those values for samples from the Kimberley and Cape York regions, but there was no significant difference between values for Cape York and Kimberley fish. Otolith weight at age The sagittae of John’s Snapper are exceptionally large, and the largest otolith weighed in this study exceeded 5 g. Second-order polynomials provided the best fits for otolith weight at age with R2 values of 0.77-0.95 (Table 3; Fig. 4). Very small values for the slope parameter c indicate that these relationships were close to linear Cappo et al.: Causes and consequences of a latitudinal cline in the demography Lut/anus johnii 315 Table 2 Comparison of regional growth data for John’s Snapper (Lutjanus johnii) from our study of a latitudinal cline in the demography of this species in Australia. Likelihood ratio tests, each with 3 degrees of freedom (df), were performed for coincidence of curves (Curves). Tests of the probability (P) of differences among slopes ( p) and intercepts (a) were made for sex and region with 2-way analysis of covariance (ANCOVA) of fork length (mm) against loge (age) for samples from the Kimberley and north Queensland regions (df numerator:df denominator=l:672). Tests of slopes and intercepts were made only for region for Cape York and Kimberley with 1-way ANCOVA (df=l:616) and for Cape York and north Queensland (df= 1:256). P=probability of null hypothesis being true. If the chi-squared goodness of fit statistic (x2) is large, the null model is a poor fit to the curve. The F statistic is the ratio of between-group mean square values to the within-group mean square values for slopes and intercepts. Factor Parameter Test statistic P Excluding Cape York data Kimberley Sex Curves X2=14.42 0.00 North Queensland Sex Curves X:=8.70 0.03 Males Region Curves X2=80.68 <0.00 Females Region Curves X2=99.58 <0.00 Kimberley, north Queensland Sex P F=9.47 0.00 Kimberley, north Queensland Region P F=35.89 <0.00 Kimberley, north Queensland SexxRegion P F=1.07 0.30 Kimberley, north Queensland Sex a F=4.51 0.03 Kimberley, north Queensland Region a F=3.00 0.08 Kimberley, north Queensland SexxRegion a F=0.67 0.41 Including Cape York data: sexes pooled Cape York, Kimberley Region Curves X2=2.06 0.56 Cape York, Kimberley Region P F= 1.45 0.23 Cape York, Kimberley Region a F=1.95 0.16 Cape York, north Queensland Region Curves X:=83.44 <0.00 Cape York, north Queensland Region P F=19.43 <0.00 Cape York, north Queensland Region a F= 3.98 0.05 (Table 3). Departures from a linear relationship were evident for older fish (>10 years), where the rate of otolith weight accretion appeared to decline with age (Fig. 4). Tests that spanned the widest age range, for the north Queensland region, showed no significant effects of sex on accretion rate, but there were significant dif- ferences in the coincidence of curves, slopes, and inter- cepts for pairwise comparisons of age ranges common between north Queensland and other regions (Table 4). Evidence of a difference in otolith weight at age between results for Kimberley and Cape York fish was equivocal because the fitted curves were not coincident, but the slopes and intercepts were not significantly different (Table 4). Cape York and Kimberley curves were inter- mediate between those from the north Queensland and Arafura Sea regions, showing a cline for an increasing rate of accretion of otolith weight with age and distance from the equator by about 0.2 g per degree of latitude for the oldest fish. Visual comparison of the regional fits with data from the Arafura Sea region showed an approximate two-fold difference in otolith weight at age beyond 10 years for north Queensland fish (Fig. 5). Gonad weight at length and maturity Exponential models provided the best fits for gonad weight at length (Table 5; Fig. 4), but the lack of data on gonad weights elsewhere severely restricted the tests in pairwise comparisons. There were no signifi- cant effects of sex for samples from north Queensland (Table 6), a finding that is not coincident with the Kimberley and Cape York fits. The tests on slopes and intercepts, however, were not significant. For fish with gonads that weighed >20 g, maturity was evident from macroscopic classification (Fig. 4). Minimum lengths and ages for the females of these mature fish were 690 mm FL and 9.83 years for north Queensland, 549 mm FL and 7.75 years for Kimberley, and 640 mm FL and 6.3 years for Cape York. Mini- mum lengths and ages for males that had gonads that weighed >20 g were 590 mm FL and 6.16 years for north Queensland, 590 mm FL and 9.75 years for Kimberley, and 620 mm FL and 9.33 years for Cape York (Fig. 4). In contrast, the Arafura Sea fish ap- peared to develop earlier. Fish sampled in that region 316 Fishery Bulletin 111(4) had ovaries >20 g at 448 mm FL (5.0 years) and testes >20 g at 472 mm FL (10.91 years) (Fig. 5). These preliminary data indicate the difference in length at maturity may be up to 24% between the northernmost and southernmost samples, but age at maturity (6-10 years) may be similar among regions, depending on sex. The north Queensland females and males matured at -81% and -72% of L„, respectively, and for samples from the Kimberley region they ma- tured at -74% and -87%, respectively. Latitudinal spread of catch records Catch records for Indo-West Pacific lutjanids indicate that the latitudes farthest from the equator produced the largest individuals for 7 species, in a steeply con- cave, “U-shaped” relationship, but we could not locate any records for other equatorial countries to fit statis- tical relationships (Fig. 6). Most records were obtained from landings in the southern hemisphere, but several world records came from Japan. The largest records for John’s Snapper show a steep rise over about 8°S from 7.2 kg in Darwin (Australian all-tackle [AAA] re- cord) to weights of 10.5 kg (97 cm total length, world all-tackle [IGFA] record), 12.420 kg (6-kg line-class [ANSA] record), and 12.0 kg (spearfishing [AUF] re- cord) near Cairns. The maximum published weight from scientific samples is 4.7 kg for a 71-cm John’s Snapper from the Andaman Sea (Druzhinin and Hla- ing, 1972). Discussion Our detection of latitudinal dines in L,„ of John’s Snapper with distance from the equator is explained Cappo et al. : Causes and consequences of a latitudinal cline in the demography Lut/anus johnii 317 Table 3 Parameters of the functions that relate otolith weight (grams) to age (f years) for males (M), females (F), and all John’s Snapper ( Lutjanus johnii) sampled in northern Australia over the period of February 1989-April 2002. Data were pooled from 3 regions with nonlinear least-squares estimation of otolith weight = a + b (t) + c (t 2). Standard errors of parameters are shown in parentheses, and the range in ages (years), sample size (n), and coefficient of multiple determination (R2) are also given. t range (years) n a b C R2 North Queensland (M) 2.16-28.58 84 -0.31 (0.06) 0.25 (0.02) -0.00 (0.00) 0.95 North Queensland (F) 1.41-25.16 80 -0.34 (0.08) 0.27 (0.02) -0.01 (0.00) 0.90 North Queensland 1.41-28.58 244 -0.34 (0.05) 0.26 (0.01) -0.00 0.91 Kimberley 1.5-23.5 563 -0.16 (0.02) 0.18 (0.01) -0.00 0.89 Cape York 2.16-18.33 99 -0.07 (0.09) 0.18 (0.02) -0.00 (0.00) 0.77 in theories of variation in body size of ectotherms in relation to geography (James’s rule; Blackburn et al., 1999) and temperature (temperature-size rule; Atkinson, 1994). However, our estimation of faster growth rates in cooler water farther from the equa- tor contradicts the temperature-size rule. Here, we discuss the major explanations for the patterns we observed through the use of prevailing ecological the- ories, and we consider the effects of our revised esti- mates of parameters for the VBGF on studies of life history. Explanations of James's rule for tropical fishes Similar dines in maximum size and growth have been documented for some sedentary coral reef species at much larger and smaller latitudinal scales, but, in the case of some lutjanids, the same trends have always been interpreted as an effect of fishing, regional dif- ferences in productivity of habitats, or the evolution of separate stocks. For example, Allman and Goetz (2009) and Burton (2001) reported that mean size at age in Gray Snapper ( Lutjanus griseus ) in Florida in- Table 4 Summary of regional comparisons of otolith weight at age for John’s Snapper ( Lutjanus johnii ) from this study of a latitudinal cline in the demography of this species in Australia. Likelihood ratio tests, each with 3 degrees of freedom, were performed for coincidence of curves (Curves). Tests of differences among slopes ((1) and intercepts (a) were made by using 1-way analysis of covariance of otolith weight against loge (age). All tests were conducted over age ranges present at both levels of the pairwise comparisons. The number of samples (n, otolith weights) is shown in parentheses for each member of the pairs, with the common age range in years (yr). P=probability of null hypothesis being true. If the chi-squared goodness of fit statistic (%2) is large, then the null model is a poor fit to the curve. The F statistic is the ratio of between-group mean square values to the within-group mean square values for slopes and intercepts. Inf=infinity. Region; age range Parameter Test statistic P North Queensland Male (80), Female (77); 2.16-25.16 yr Curves X2=Inf 1 P F=0.59 0.44 a F=0.33 0.57 North Queensland (227), Kimberley (563); 1.5-23.5 yr Curves X2=349.19 <0.00 P F=168.48 <0.00 a F=69.30 <0.00 North Queensland (221), Cape York (98); 2.16-18.3 yr Curves 5C2=86.19 <0.00 P F=168.48 <0.00 a F=69.30 <0.00 Cape York (96), Kimberley (550); 2.16-18.3 yr Curves X-23.21 <0.00 P F=0.31 0.58 a F=1.59 0.21 318 Fishery Bulletin 111(4) 880 770 - E" 660 ■ E £ 550 - a> 0) 440 - £ 330 - 220 - 110 - B 3 60 3.15 270 2.25 1.80 1.35 0.90 0.45 I 1 1 1 1 1 1 1 1 1 1 3 6 9 12 15 18 21 24 27 30 Age (years) 6 8 10 12 14 16 18 20 Age (years) Fork length (mm) Figure 5 Comparisons of parameters by region, and fits of Arafura Sea data points, for John’s Snap- per (Lutjanus johnii) sampled during the pe- riod of February 1989-April 2002 in 4 regions of Australia — north Queensland, Kimberley, Cape York, and Arafura Sea: (A) fork length (mm) at age (years) with von Bertalanffy growth function (sexes pooled; see Table 2); (B) otolith weight at age with first-order poly- nomials (see Table 3); and (C) gonad weight at fork length with exponential relationships (see Table 5). creased with distance from the equator, but they in- voked major regional differences in exploitation rate as an explanation. At a smaller latitudinal scale, Saari (2011) concluded that Red Snapper (L. campeclianus ) from northern Texas and Alabama reach significantly larger L„ than do Red Snapper from southern Texas and northwestern Florida. Saari (2011) discussed se- vere overfishing as the primary cause of the difference, as well as differences in environmental factors, fish- ing behavior between sectors, habitat-preference, and management regimes. In eastern Indonesia, the Crim- son Snapper ( L . erythropterus) and Malabar Snapper grow faster than their conspecifics in northern Austra- lia, but Fry and Milton (2009) interpreted this pattern in relation to the genetic evidence for separate stocks. Mangrove Jack at the southern end of their Austra- lian range have faster juvenile growth and are larger at a given age, but Russell et al.2 were concerned that sample sizes were too small for any inferences to be made from such observations. There is no doubt that heavy fishing can affect body sizes of fishes across latitudes. Throughout the 1970s, there was a ten-fold increase in mean body size of 326 fish species from low to high latitudes in the North Atlantic. However, this trend began to weaken under heavy fishing pressure in the early 1980s, and, by 1991, mean body sizes had declined steeply to the extent that a gradient was no longer detectable (Fisher et al., 2010). This homogenization of community size struc- tures was a breakdown of Bergmann’s rule that Fisher et al. (2010) predicted will lead to declining stability in populations, communities, and ecosystems. The earliest explanations for James’s rule concerned a quandary posed by the temperature-size rule (Atkin- son, 1994); for most ectotherms, decreased nutrition and decreased temperature both reduce growth rates, but each affects maturity differently. Decreased nutri- tion results in delayed maturity at a smaller size, yet decreased temperature usually results in delayed ma- turity at a larger size. This puzzle led Berrigan and Charnov (1994) to propose that the effects of tempera- ture on maturity are associated with the existence of a negative correlation between and the growth coef- ficient, K, in the VBGF. In contrast, the latitudinal studies of tropical fish growth at the largest scales, over 56° of latitude for Ocean Surgeon ( Acanthurus bahianus) and 14° of lati- tude for Stoplight Parrotfish ( Sparisoma viride), have shown that growth rate is faster in cooler waters (22.6- 28.1°C), not slower. Maximum age, adult survivorship, terminal size, and absolute growth rate are inversely related to temperature in populations of Ocean Sur- 2 Russell, D. J., A. J. McDougall, A. S. Reicher, J. R. Ovenden, and R. Street. 2003. Biology, management and genetic stock structure of mangrove jack (Lutjanus argentimacula- tus) in Australia, 198 p. Queensland Department of Prima- ry Industries, Brisbane, Australia. [Available from http;// era.deedi.qld.gov.au/3119/l/BiologyManGeneticStock_report_ final%5Bl%5D-sec.pdf.] Cappo et at: Causes and consequences of a latitudinal cline in the demography Lut/anus /ohnii 319 Table 5 Parameters of the functions relating gonad weight (grams) to fork length (FL) for John’s Snapper ( Lut - janus johnii ) sampled in 3 regions of Australia over the period of February 1989-April 2002. Nonlinear least-squares estimation was calculated with the following equation: gonad weight = e(a + b Standard errors of parameters are shown in parentheses, and the range in FL (Lp), sample size ( n ), and coefficient of multiple determination ( R 2) are also given. Region; sex Lp range (mm) n a b R2 North Queensland (Male) 222-860 67 -5.24 (0.99) 0.01 (0.00) 0.75 North Queensland (Female) 173-830 60 -2.27 (1.08) 0.01 (0.00) 0.57 North Queensland 173-860 127 -2.98 (0.75) 0.01 (0.00) 0.60 Kimberley 473-696 60 -6.49 (0.69) 0.02 (0.00) 0.68 Cape York 290-652 56 -31.64 (2.19) 0.06 (0.00) 0.96 geon (Robertson et al., 2005a) and there are no indi- cations of consistent effects of fishing on longevity or adult survivorship (Robertson et al., 2005b). In relation to Berrigan and Charnov’s (1994) commentary, it is im- portant to note that there was no sign of a negative relationship between locality-specific and K in fig. 5 of Robertson et al. (2005a) for Ocean Surgeon, but such a negative relationship was noted for Stoplight Parrotfish in fig. 4 of Choat et al. (2003). Counter-gradient variation in growth rates has been proposed to explain this contradiction of the temper- ature-size rule. This counter-gradient variation occurs where genetic and environmental influences on pheno- types oppose one another to produce metabolic com- pensation (Conover et al., 1997). Various physiological rates and processes are now known to be elevated to counteract the dampening effect of reduced tempera- ture and diminished length of the optimal season for Table 6 Comparison of regional summaries of gonad weight at length for John’s Snapper ( Lutjanus joh- nii) from our study of a latitudinal cline in the demography of this species in Australia. Likeli- hood ratio tests, each with 2 degrees of freedom ( df) were performed for coincidence of curves. Tests of differences among slopes ( P) and intercepts (a) were made by using 1-way analysis of covariance of gonad weight against loge (age). All tests were conducted over age ranges in fork length (FL) present in both levels of the pairwise comparisons. The number of samples in, shown in parentheses) and gonad weights is shown for each member of a pair, and the common length range in millimetres is also given. P=probability of null hypothesis being true. If the chi-squared goodness-of-fit statistic (%2) is large, the null model is a poor fit to the curve. The F statistic is the ratio of between-group mean square values to the within-group mean square values for slopes and intercepts. Inf=infinity. Region; fork length range Parameter Test statistic P North Queensland Male (66), Female (57); 222-830 mm Curves X2=lnf 1 P F= 0.08 0.77 a .F=0.60 0.44 Regions North Queensland (60), Kimberley (60); 473-696 mm Curves X2=74.17 <0.00 P F=1.05 0.31 a F=3.24 0.07 North Queensland (89), Cape York (56); 290-652 mm Curves X2=281.74 <0.00 P F= 2.24 0.14 a F= 0.55 0.46 Kimberley (55), Cape York (22); 473-652 mm Curves X2=157.12 <0.00 P F=25.74 <0.00 a F= 28.25 <0.00 320 Fishery Bulletin 111(4) 20 - A.vir S.nem S.nem + L.seb .c 1C CT) 3 'd> g Q> O 10 L.arg + L.seb L.arg A.vir ,, + S.nem L.seb ++ L.boh L.boh+ + S.nem AvirL+mal L.mal® l'+ +L.seb o A.vir L.boji L.boh L.mal o L. johnii L. johnii L. johnii 97 cm } Cairns L.arg + L.arg L.mal o L. johnii — Darwin • L. johnii 71 cm Andaman Sea; Druzhinin & Hlaing (1972) ~1~ 20° -30° -20° -10° — r 10° 30° Latitude Figure 6 Records of maximum weight of John’s Snapper (. Lutjanus johnii ) and 6 other large lutjanids landed until 2011, shown by latitude. The data (whole weight in kilograms) for each species was provided by the following sources: the International Game Fishing Association, Australian National Sportfish- ing Association, Australian Angler’s Association, and Australian Underwater Federation. The cluster of data for John’s Snapper is bracketed for comparison with the data from Darwin and the largest weight reported in the scientific literature, by Druzhinin and Hlaing (1972) (bottom right). The other 6 lutjanids were Mangrove Jack (L. argentimaculatus : L.arg), Twospot Snapper (L. bohar ; L.boh), Mala- bar Snapper (L. malabaricus: L.mal), Emperor Snapper (L. sebae: L.seb), Chinaman Fish ( Symphorus nematophorus: S.nem), and Green Jobfish (Aprion virescens : A.vir). feeding and growth at increasing distance from the equator (see Conover et al., 2009, for review). In fact, the crisper clarity of opaque and translucent zones in otoliths of tropical fishes from latitudes where water temperatures are 5-10° Celsius cooler may be a physi- ological product of this counter-gradient variation in growth (see photomicrographs in Choat et al., 2003, 2009; Marriott and Mapstone, 2006; Robertson et al., 2005a). Portner and Knust (2007) proposed a “thermal limitation hypothesis” that natural selection favors individuals that maximize growth and energy effi- ciency at the expense of ranges of thermal tolerance (see also Portner et al., 2008). The underlying con- cept of oxygen- and capacity-limited thermal toler- ance (OCLT) implies that oxygen supply to tissues is optimal between lower and upper temperature limits. Between these limits (termed pejus tempera- tures), oxygen supply also can be increased to exceed maintenance demand and fuel aerobic metabolism for the performance of growth, foraging, migration, and reproduction. These “performances” support the fitness of species, and the excess in oxygen availability that supports them is reflected in a species-specific aerobic scope. The aerobic scope is the difference between the low- est and highest rates of aerobic respiration, with an optimum close to the upper pejus temperature. Beyond upper pejus limits, oxygen supply decreases, mainte- nance demand rises, and aerobic scope begins to de- crease (for review, see Portner 2012). At suboptimal high temperatures, fish cannot consume enough food to meet increasing metabolic needs because aerobic scope is insufficient to satisfy the increase in oxygen demand from exercise and digestion (Portner and Peck, 2010). Populations of Atlantic Cod ( Gadus morhua) also follow James’s rule in the Atlantic, where the growth, spawning, and recruitment of this species are well Cappo et al. : Causes and consequences of a latitudinal cline in the demography Lut/anus johnii 321 known along a latitudinal cline (Portner et al., 2008). Permanent physiological differences induced by tem- perature and climate have been identified in Atlantic Cod populations along that cline, resulting in popu- lation-specific patterns of OCLT. The Hb-I(l/1) allele displays an increasing frequency toward the (warmer) south, leading to a higher oxygen affinity at higher temperatures, and this feature is considered to be a microevolutionary adaptation to optimize oxygen trans- port (Portner et al., 2008). The OCLT concept does not imply strict positive or negative correlations between longevity, maximum size, or growth rate along latitudinal dines. It offers a prom- ising new way forward to use physiological challenges under controlled conditions (see Clark et al., 2012) to disentangle true mechanistic causes (and contradic- tions) of the temperature-size rule from effects of fish- ing and unknown environmental differences between regions. This approach may explain why responses in growth rate and maximum size along long latitudinal gradients are inconsistent in statistical correlations used in intensive field studies of tropical fishes. For example, Robertson et al. (2005a) concluded that varia- tion in growth and terminal size is related strongly to both habitat and temperature, yet Trip et al. (2008) proposed that growth and adult size are most respon- sive to local environmental features unrelated to lati- tudinal (temperature) effects. Growth trajectories and length at maturity Despite vast differences in the local environments sampled, the basic patterns in the growth curves of John’s Snapper are conserved. This snapper species has a relatively gradual growth trajectory through early life, maturing at 6-10 years and at 70-80% of and reaching an asymptotic length at -18-20 years. The consistent, sex-specific differences in growth rates are consistent with functional gonochorism for John’s Snapper, for which there is a higher selective pressure for females to grow to a larger size and have a higher fecundity (Roff, 1983). Longevities >20 years are known for many small and large lutjanids (e.g., Heupel et al., 2010; Martinez-Andrade, 2003) and are considered to be beneficial by ensuring a long reproductive life. This life history minimizes the risk that unfavorable events at large scales will result in the loss of a metapopula- tion. In life history terms, John’s Snapper is an “inter- mediate strategist” falling in the center of a continuum between large species that mature at later ages and have large eggs and those that are long-lived, slow- growing, and highly fecund species (King and MacFar- lane, 2003). Our demonstration of a longevity that is nearly 3 times that reported from early studies is not surpris- ing, or novel, but it is nonetheless very important to improve meta-analyses, such as analyses with Ecopath and stock reduction models. Compared with the pa- rameters derived by Khan (1986), which appear in the online database FishBase (Froese, 2011), the param- eters we have shown for John’s Snapper give evidence of a higher longevity (28.6 versus 10 years derived by Khan), lower K (0.16-0.21 versus 0.28), about 4% of the total instantaneous mortality Z (0.146 versus 2.700), and, consequently, about 25% of the rate of natural mortality M (<0.146 versus <0.590). The estimates of length at maturity of John’s Snap- per are coarse and preliminary but indicate differences as high as 24% between the latitudinal limits sampled. Therefore, it is interesting to note that the minimum esti- mates exceeded predictions from regressions on the basis of L„ and Lmax from published meta-analyses. With our Lmax °f 990 min FL and LOT of 843.5 mm FL for the north Queensland region, we calculated an estimated Lm of 495 ±10 mm FL using the Binohlan and Froese (2009) method and estimated Lm of 448 ±10 mm FL in the Froese and Binohlan (2000) equation. Martinez-Andrade (2003) gen- eralized that Lm occurred at a length about half (0.52) of the L„ for lutjanids, producing an even smaller estimate (Lm=438 mm FL) for John’s Snapper. Our estimates were considerably larger at 590 mm FL for males and 690 mm FL for females in north Queensland, representing from 59.6% (males) to 69.7% (females) of Lmax and from 69.9% to 81.8% of Ltc. The legal limits to fish size at first capture of John’s Snapper in Western Australia (300 mm total length) and Queensland (350 mm total length) do not approach any of the estimates discussed above. The Northern Territory has no size limit. The northernmost (Arafura Sea) samples were at the smallest extremes of length and otolith weight at age and of gonad weight at length, when compared with samples from the other regions. However, the fishery on the coastal reefs of the Northern Territory, inshore of the Arafura Sea trawl grounds, recorded John’s Snapper up to 820 mm FL and 23 years of age (Hay et al.3). Of these coastal females, 50% reached sexual maturity (Lm5o) at a much larger size of 630 mm FL (8-10 years) than did Arafura Sea females, although males reached maturity at a similar size (Lm 5q=470 mm FL). There is clearly a need to accurately measure regional length at maturity and establish fecundity-size curves to fully understand the nested hierarchy within growth curves. In general terms, larger adults of tropical fish populations farther from the equator might be expected to have much larger ovaries (and hence batch fecundity) and a longer spawning life in comparison with their smaller counterparts close to the equator. However, there is no evidence that recruit- ment rates are higher for populations at these margins. Instead, Portner et al. (2008) proposed that recruitment rates should show a dome-shaped distribution about an optimal temperature range. 3 Hay T., I. Knuckey, C. Calogeras, and C. Errity. 2005. NT coastal reef fish: population and biology of the golden snap- per. Fishnote No: 21 Department of Primary Industry, Fish- eries and Mines, Darwin, Northern Territory, Australia, 4 p. [Available from http://www.nt. gov.au/d/Content/File/p/ Fishnote/FN21.pdf.[ 322 Fishery Bulletin 1 1 1 (4) The appearance of a U-shaped relationship between record sizes of lutjanids and latitude among 3 genera highlights the chronic lack of basic length and weight information on equatorial and Asian populations of lut- janids in the Indo-West Pacific. These records also sug- gest that James’s rule may apply in age-based studies when such studies are eventually undertaken in those countries. Conclusions As with some studies of site-attached coral reef fish- es, our findings of larger terminal size, faster growth, and larger size at maturity for John’s Snapper far- thest from the equator agree with James’s rule but do not agree with the presumption of “slower growth in colder water” of the temperature-size rule for ec- totherms. Further, age-based studies alone cannot re- solve the variability in the growth response reported in some tropical studies. More powerful insights can be obtained through the use of the concept of OCLT and measurement of physiological response to exercise and thermal challenges in populations along latitudi- nal dines. The existence of older, larger John’s Snapper in the southern portion of the range of this species has raised some compelling questions concerning the lifetime re- productive output and subsequent recruitment rates of tropical fish populations at the warmer core and cooler limits of their ranges. If recruitment also is marginal at thermal limits, then is the development of larger gonads each year over a longer life in cooler waters a wasted investment for John’s Snapper or is it an adap- tation to episodic recruitment success? Such questions can be investigated only if age-based studies of fish demography are accompanied by information on size- fecundity curves and egg size and quality, along with some relative indices of recruitment. Acknowledgments We wish to thank all the tackle store proprietors, nu- merous anglers, and spearfishing individuals who pro- vided specimens for use in this study or assisted with fieldwork. From the fishing community, we would like to thank, in particular, A. J. McDougall, D. Donald, A. Mead, E. Riddle, M. Kenway, S. Boyle, and R Haz- ard. Archived otoliths and fish frames were supplied also by D. Milton, G. McPherson, M. Sheaves, and A. Coleman. Field and laboratory support was provided by R. Steckis, J. Jenke, C. Skepper, and B. Robertson. We especially appreciate the critical and constructive advice of 3 reviewers including J. H. Choat and A. J. McDougall. Literature cited Allen, G. R. 1985. FAO species catalogue. Vol. 6. Snappers of the world. An annotated and illustrated catalogue of lutja- nid species known to date. FAO Fish. Synop. 125, 208 p. FAO, Rome. Allman, R. J., and L. A. Goetz. 2009. Regional variation in the population structure of gray snapper, Lutjanus griseus, along the west Florida shelf. Bull. Mar. Sci. 84:315-330. Allman, R. J., G. R. Fitzhugh, K. J. Starzinger, and R. A. Farsky. 2005. Precision of age estimation in red snapper (Lutja- nus campechanus). Fish. Res. 73:123-133. Arendt, J. D. 2011. Size-fecundity relationships, growth trajectories, and the temperature-size rule for ectotherms. Evolu- tion 65:43-51. Atkinson, D. 1994. Temperature and organism size — a biological law for ectotherms. Adv. Ecol. Res. 25:1-58. 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. Berrigan, D., and E. L. Charnov. 1994. Reaction norms for age and size at maturity in response to temperature: a puzzle for life historians. Oikos 70:474-478. Berumen, M. L., E. D. L. Trip, M. S. Pratchett, and J. H. Choat. 2012. Differences in demographic traits of four butterfly- fish species between two reefs of the Great Barrier Reef separated by 1,200 km. Coral Reefs 31:169-177. Binohlan, C., and R. Froese. 2009. Empirical equations for estimating maximum length from length at first maturity. J. Appl. Ichthyol. 25:611-613. Blackburn, T. M., K. J. Gaston, and N. Loder. 1999. Geographic gradients in body size: a clarification of Bergmann’s rule. Divers. Distrib. 5:165-174. Burton, M. L. 2001. Age, growth, and mortality of gray snapper, Lutja- nus griseus, from the east coast of Florida. Fish. Bull. 99:254-265. Campana, S. E., M. C. Annand, and J. I. McMillan. 1995. Graphical and statistical methods for determin- ing the consistency of age determinations. Trans. Am. Fish. Soc. 124:131-138. Cappo, M., P. Eden, S. J. Newman, and S. Robertson. 2000. A new approach to tetracycline validation of the periodicity and timing of opaque zone formation in the otoliths of 11 species of Lutjanus from the central Great Barrier Reef. Fish. Bull. 98:474-488. Cerrato, R. M. 1990. Interpretable statistical tests for growth compari- sons using parameters in the von Bertalanffy equation. Can. J. Fish. Aquat. Sci. 47:1416-1426. Choat, J. H., and D. R. Robertson. 2002. Age-based studies. In Coral reef fishes: dynamics and diversity in a complex ecosystem (P. F. Sale, ed.), p. 57-80. Academic Press, San Diego, CA. Cappo et al.: Causes and consequences of a latitudinal cline in the demography Lutjanus johnii 323 Choat, J. H., D. R. Robertson, J. L. Ackerman, and J. M. Posada. 2003. An age-based demographic analysis of the Carib- bean stoplight parrotfish Sparisoma viride. Mar. Ecol. Prog. Ser. 246:265-277. Choat, J. H., J. P. Kritzer, and D. R. Robertson. 2009. Ageing in coral reef fishes: do we need to validate the periodicity of increment formation for every species of fish for which we collect age-based demographic data? In Tropical fish otoliths: information for assessment, management and ecology (B. S. Green, B. D. Mapstone, G. Carlos, and G. A. Begg, eds.), p. 23-45. Springer, New York. Clark, T. D., M. R. Donaldson, S. Pieperhoff, S. M. Drenner, A. Lotto, S. J. Cooke, S. G. Hinch, D. A. Patterson, and A. P. Farrell. 2012. Physiological benefits of being small in a changing world: responses of coho salmon ( Oncorhynchus kisutch) to an acute thermal challenge and a simulated capture event. PLoS ONE 7(6):e39079. Conover, D. O., J. J. Brown, and A. Ehtisham. 1997. Countergradient variation in growth of young striped bass ( Morone saxatilis) from different latitudes. Can. J. Fish. Aquat. Sci. 54:2401-2409. Conover, D. O., T. A. Duffy, and L. A. Hice. 2009. The covariance between genetic and environmen- tal influences across ecological gradients: reassessing the evolutionary significance of countergradient and co- gradient variation. Ann. N.Y. Acad. Sci. 1168:100-129. Druzhinin, A. D. 1970. The range and biology of snappers (Fam. Lutjani- dae). J. Ichthyol. 10:717-736. Druzhinin, A. D., and U. P. Hlaing. 1972. Observations on the trawl fishery of southern Burma. In Proc. Indo-Pac. Fish. Coun, Session 13, Sect. Ill Tech. Pap., Symposium on Demersal Fisheries; Brisbane, Australia, 14-25 October 1968, p. 151-209. IPFC (Indo-Pacific Fisheries Council) Secretariat, FAO Regional Office for Asia and the Far East, Bangkok, Thailand. Fisher, J. A. D., K. T. Frank, and W. C. Leggett. 2010. Breaking Bergmann’s rule: truncation of North- west Atlantic marine fish body sizes. Ecology 91:2499-2505. Froese, R. 2011. The science in FishBase. In Ecosystem approach- es to fisheries: a global perspective (V. Christensen and J. Maclean, eds.), p. 47-54. Cambridge Univ. Press, Cambridge, UK. Froese, R., and C. Binohlan. 2000. Empirical relationships to estimate asymptotic length, length at first maturity and length at maximum yield per recruit in fishes, with a simple method to eval- uate length frequency data. J. Fish Biol. 56:758-773. Fry, G. C., and D. A. Milton. 2009. Age, growth and mortality estimates for popula- tions of red snappers Lutjanus erythropterus and L. malabaricus from northern Australia and eastern Indo- nesia. Fish. Sci. 75:1219-1229. Haddon, M. 2001. Modelling and quantitative methods in fisheries, 424 p. Chapman and Hall/CRC Press, Boca Raton, FL. Heupel, M. R., A. Penny, A. J. Williams, J. P. Kritzer, D. J. Welch, R. J. Marriott, C. R. Davies, and B. D. Mapstone. 2010. Demographic characteristics of exploited tropi- cal lutjanids: a comparative analysis. Fish. Bull. 108:420-432. Hoenig, J. M. 1983. Empirical use of longevity data to estimate mor- tality rates. Fish. Bull. 82: 898-902. Khan, M. G. 1986. Age, growth and mortality rates of the red snap- per (Lutjanus johnii Bloch) of the Bay of Bengal, Ban- gladesh. Bangladesh J. Agric. 1 1(2):49— 61. King, J. R., and G. A. McFarlane. 2003. Marine fish life history strategies: applications to fishery management. Fish. Manage. Ecol. 10:249-264. Kiso, K., and M. I. Mahyam. 2003. Distribution and feeding habits of juvenile and young John’s snapper Lutjanus johnii in the Matang mangrove estuary, west coast of Peninsular Malaysia. Fish. Sci. 69:563-568. Marriott, R. J., and M. Cappo. 2000. Comparative precision and bias of five different ageing methods for the large tropical snapper Lutjanus johnii. Asian Fish. Sci. 13:149-160. Marriott, R. J., and B. D. Mapstone. 2006. Geographic influences on and the accuracy and precision of age estimates for the red bass, Lutjanus bohar (Forsskal 1775): a large tropical reef fish. Fish. Res. 80:322-328. Martinez-Andrade, F. 2003. A comparison of life histories and ecological as- pects among snappers (Pisces: Lutjanidae). Ph.D. diss., 194 p. Louisiana State Univ., Baton Rouge, LA. Newman, S., M. Cappo, and D. McB. Williams. 2000. Age, growth, mortality rates and corresponding yield estimates using otoliths of the tropical red snap- pers, Lutjanus erythropterus, L. malabaricus and L. sebae, from the central Great Barrier Reef. Fish. Res. 48:1-14. Pauly, D., S. Samiblya Jr., and S. Opitz. 1993. Estimates of relative food consumption by fish and invertebrate populations, required for modelling the Bo- linao Reef Ecosystem, Philippines. In Trophic models of aquatic ecosystems (V. Christensen and D. Pauly, eds.), p. 236-251. ICLARM Conference Proceedings 26, International Center for Living Aquatic Resources Management, Metro Manila, Philippines. Portner, H. O. 2012. Integrating climate-related stressor effects on marine organisms: unifying principles linking mole- cule to ecosystem-level changes. Mar. Ecol. Prog. Ser. 470:273-290. Portner, H. O., C. Bock, R. Knust, G. Lannig, M. Lucassen, F. C. Mark, and F. J. Sartoris. 2008. Cod and climate in a latitudinal cline: physiologi- cal analyses of climate effects in marine fishes. Clim. Res. 37:253-270. Portner, H. O., and R. Knust. 2007. Climate change affects marine fishes through the oxygen limitation of thermal tolerance. Science 315:95-97. Portner, H. O., and M. A. Peck. 2010. Climate change effects on fishes and fisheries: to- wards a cause-and-effect understanding. J. Fish Biol. 77:1745-1779. 324 Fishery Bulletin 111(4) Robertson, D, R., J. L. Ackerman, J. H. Choat, J. M. Posada, and J. Pitt. 2005a. Ocean surgeonfish Acanthurus bahianus. I. The geography of demography. Mar. Ecol. Prog. Ser. 295:229-244. Robertson, D. R., J. H. Choat, J. M. Posada, J. Pitt, and J. L. Ackerman. 2005b. Ocean surgeonfish Acanthurus bahianus. II. Fish- ing effects on longevity, size and abundance? Mar. Ecol. Prog. Ser. 295:245-256. Roff, D. A. 1983. An allocation model of growth and reproduction in fish. Can. J. Fish. Aquat. Sci. 40:1395-1404. Saari, C. R. 2011. Comparison of the age and growth of red snapper (. Lutjanus campechanus ) amongst habitats and regions in the Gulf of Mexico. M.S. diss., 134 p. Louisiana State Univ., Baton Rouge, LA. Tanaka, K., Y. Hanamura, V. C. Chong, S. Watanabe, A. Man, F. M. Kassim, M. Kodama, and T. Ichikawa. 2011. Stable isotope analysis reveals ontogenetic migra- tion and the importance of a large mangrove estuary as a feeding ground for juvenile John’s snapper Lutjanus johnii. Fish. Sci. 77:809-816. Travers, M. J., I. C. Potter, K. R. Clarke, S. J. Newman, and J. B. Hutchins. 2009. The inshore fish faunas over soft substrates and reefs on the tropical west coast of Australia differ and change with latitude and bioregion. J. Biogeogr. 37:148-169. Trip, E. L., J. H. Choat, , D. T. Wilson, and D. R. Robert- son. 2008. Inter-oceanic analysis of demographic variation in a widely distributed Indo-Pacific coral reef fish. Mar. Ecol. Prog. Ser. 373:97-109. 325 Abstract— The influences of age, size, and condition of spawning fe- males on fecundity and oocyte qual- ity were analyzed for the Patagonian stock of Argentine Hake (. Merluccius hubbsi). Samples of mature females were collected in the spawning area as part of 2 research surveys con- ducted in January 2010 and 2011, during the peak of the reproductive season. Batch fecundity (BF) ranged between 40,500 (29 cm total length [TL] ) and 2,550,000 (95 cm TL) hy- drated oocytes, and was positively correlated with TL, gutted weight, age, hepatosomatic index (HSI), and the relative condition factor (Kn). Relative fecundity ranged between 85 and 1040 hydrated oocytes g_1 and showed significant positive re- lationships with gutted weight, HSI, and Kn; however, coefficients of de- termination were low for all regres- sions. Dry weights of samples of 100 hydrated oocytes ranged between 1.8 and 3.95 mg and were positively cor- related with all variables analyzed, including batch and relative fecun- dity. Multiple regression models cre- ated with data of the morphophysi- ological characteristics of females supported maternal influences on fecundity and egg weights. Within the studied size range (29-95 cm TL), larger individuals had better somatic and egg condition, mainly revealed by higher HSI and hydrat- ed oocytes with larger oil droplets (275.71pm [standard error 1.49]). These results were associated with the higher feeding activity of larger females during the spawning season in comparison with the feeding ac- tivity of young individuals (<5 years old); the better nutritional state of larger females, assumed to result from more feeding, was conducive to greater production of high-quality eggs. Manuscript submitted 11 January 2013. Manuscript accepted 7 August 2013. Fish. Bull. 111:325-336. doi: 10. 7755/FB. 111.4.3 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necesarily reflect the position of the National Marine Fisheries Service, NOAA. Maternal effects on fecundity and egg quality of the Patagonian stock of Argentine Hake ( Merluccius hubbsi ) Gustavo J. Macchi (contact author)1- 2 Ezequiel Leonarduzzi2 Marina V. Diaz2 Marta Renzi2 Karina Rodrigues 12 Email address for contact author; gmacchi@inidep.edu.ar 1 Consejo Nacional de Investigaciones Cientfficas y Tecnicas (CON ICED Av. Rivadavia 1917 C1033AAJ Buenos Aires, Argentina 2 Instituto Nacional de Investigacion y Desarrollo Pesquero (INIDEP) Paseo Victoria Ocampo Nro. 1 B7602HSA Mar del Plata, Argentina The Argentine Hake (Merluccius hubbsi), with a biomass of around 700,000 metric tons (t) estimated from virtual population analysis in 2009, is one of the most important fish resources for the Argentine fleet (Irusta and Datri1; Villarino and Santos2). Two main stocks have been identified in the Argentine Sea: the northern group (between 34° and 41°S) and the southern group (be- tween 41° and 55°S). The latter group, known also as the Patagonian stock, is the most abundant population of this species, accounting for about 85% of the total biomass estimated for this 1 Irusta, C. G., and L. D’Atri. 2010. Evaluacion del estado del efectivo norte de 41° S de la merluza ( Merluccius hubbsi) y estimacion de la Captura Bio- logicamente Aceptable para el ano 2011. INIDEP Informe Tecnico Oficial No. 42, 28 p. [Available from INIDEP, Paseo Victoria Ocampo Nro. 1, B7602HSA Mar del Plata, Argentina. 2 Villarino, M. F., and B. Santos. 2010. Evaluacion del estado de explotacion del efectivo sur de 41° S de la merlu- za ( Merluccius hubbsi ) y estimacion de las capturas biologicamente aceptables correspondiente al ano 2011. INIDEP Informe Tecnico Oficial No. 43, 27 p. [Available from INIDEP, Paseo Victoria Ocampo Nro. 1, B7602HSA Mar del Pla- ta, Argentina.] fish resource in Argentina (Aubone et al., 2000). This fishery historically has registered the most important commercial landings for Argentina, with annual catches between 170,000 and 370,000 t since 2000 (Sanchez et al., 2012). Nevertheless, during the 1990s, both stocks of Argentine Hake suffered overexploitation and their spawning biomass decreased drasti- cally, and changes also were observed in the parent-stock structure (Aubone et al., 2000). It has been shown that the repro- ductive potential of a fish stock is strongly influenced by the size and age of spawning females. During the reproductive season, older and larg- er individuals produce more oocytes than younger and smaller spawners because fecundity and spawning fre- quency are higher in larger females and, in general, their reproductive period is longer than the period of smaller spawners (Marshall et al., 1998; Marteinsdottir and Thora- rinsson, 1998; Marteinsdottir and Begg, 2002; Macchi et al, 2004; Me- hault et al., 2010). Therefore, varia- tion in length composition of a fish stock caused by fishing activity or by other sources of mortality would af- fect the number of eggs spawned at 326 Fishery Bulletin 1 1 1 (4) a population level. In this sense, fecundity estimates by length class give information on the number of po- tential offspring that can be produced. Nevertheless, viability during the early life stages depends mainly on egg quality, which is associated with the quantity of nutrients stored in the oocytes (Brooks et al., 1997). Bromage et al. (1994) defined egg quality as “those characteristics of the egg that determine its capacity to survive.” There are different methods and indices that can be used to estimate the quality of fish eggs. Some of them are very simple, such as those techniques where morphometrical attributes (e.g., egg diameter or weight and size of the oil droplet) are used, and others are more complex, such as those where the biochemical composition (e.g., lipid or protein concentration) of the oocytes are used (Nocillado et al., 2000). Most studies have concluded that larger eggs are of better quality because they can produce larger larvae with higher survival rates (Hinckley, 1990; Rijnsdorp and Vingerhoed, 1994; Nissling et al., 1998; Trippel, 1998). It is a general assumption that larger larvae may have advantages in finding food and in actively swimming through the water column that increase their probability of survival (Brooks et al., 1997; Al- varez Colombo et al., 2011). In contrast, in some stud- ies, no relationship between egg diameter and larval size was observed (Landaeta and Castro, 2012), or it was suggested that larger offspring could be more eas- ily detected by predators, and, therefore, large larval size would increase the risk of mortality (Kjesbu et al., 1996). However, in general, it is assumed that in- creased egg size leads to higher larval survival. Fish condition can be assessed through the use of different indices, on the basis of length :body-weight ra- tio (K) or liver-weight:body- weight ratio (hepatosomatic index [HSI] ) or on the basis of the biochemical com- position (e.g., lipids, proteins, or glycogen) of different tissues (Dominguez-Petit et al., 2010). It was observed that lipids play an important role as stored energy that can affect fecundity, oocyte quality, and larval viabil- ity (Wiegand, 1996; Lambert et al., 2003; Aristizabal, 2007). Therefore, because the liver is generally the or- gan with the highest levels of lipid storage for gadoids, the HSI may be considered a good descriptor of female condition, and a strong influence on reproduction (Mar- shall et al., 1999; Marteinsdottir and Begg, 2002; Alon- so-Fernandez, 2011). Spawning of the Patagonian stock of the Argentine Hake occurs in waters off the province of Chubut in southern Argentina during spring and summer, and a main peak occurs between December and January (Macchi et al., 2007). This species is a batch spawner with indeterminate annual fecundity, meaning that un- yolked oocytes continuously mature and are spawned throughout the reproductive season (Macchi et al., 2004). It has been reported that the size and age com- positions of Argentine Hake females in Patagonian wa- ters change during the spawning season, affecting egg production of this stock during this period (Macchi et al., 2004). Moreover, Macchi et al. (2006) found posi- tive relationships between oocyte dry weight (DW) and maternal variables, such as total length (TL) and age, relationships that could indicate a possible effect of fe- male size on egg quality. Although different aspects of reproductive biology of Argentine Hake have been studied in recent years (Macchi et al., 2004, 2005, 2006, 2007), information about the nutritional condition of spawning females in relation to reproductive potential is scarce. This infor- mation is essential for understanding the strategy of energy allocation during reproduction and for estimat- ing whether variations in size composition and feeding condition of spawning females could affect egg quality and, therefore, larval survival. Such information could improve the understanding of the recruitment varia- tions that have been observed in the Argentine Hake fishing stocks. The objective of this study was to determine which maternal features — length, weight, age, HSI, and con- dition— were correlated with which measures of egg quality and quantity — DW, oil globule size, and fecun- dity— in Argentine Hake from Patagonian waters. We also examined stomach fullness as a function of fish length to establish the relationship between feeding activity and the size of spawning females. Materials and methods Sample collection Samples of Argentine Hake were obtained from bottom trawls during 2 research surveys carried out in the north Patagonian area during January 2010 and 2011 by the Instituto Nacional de Investigacion y Desarrol- lo Pesquero (INIDEP) (Fig. 1). Argentine Hake speci- mens were captured at depths between 50 and 110 m through employment of a bottom trawl with a mouth width of about 20 m, a height of about 4 m, and with a 20-mm mesh size at the inner lining of the codend. For histological analysis of ovaries, 192 mature fe- males near spawning condition (with hydrated oocytes after germinal vesicle breakdown) were sampled; ova- ries were removed and preserved (fixed) in 10% neu- trally buffered formalin. From each specimen, the fol- lowing measurements were recorded: TL in centimeters and total weight (TW), gutted weight (GW), and liver weight (LW) in grams. In addition, the sagittae oto- liths were collected for age determination. Ages were obtained through the use of the methods of Renzi and Perez (1992). Female condition and fecundity estimation Because fish weight may be greatly influenced by the stomach content and fullness of the gut, GW was used to estimate the condition of Argentine Hake females. jViacchi et a!.: Maternal effects on fecundity and egg quality of Merluccius hubbsi 327 Figure t Map of locations where samples of Argentine Hake ( Merluccius hubbsi) were collected during the peak spawning season (Janu- ary) of 2010 (black dots) and 2011 (squares) in waters off the province of Chubut in Argentina to examine maternal effects on the fecundity and egg quality. Nutritional status was determined by means of 2 indices: 1) the relative condition factor (Kn), expressed as a proportion between the observed GW and the GW determined by the relationship of TL versus GW, de- scribed by the following equation: GW = 0.0148 x TL2-7634 (1) (coefficient of determination [r2]=0.98, n=181, P<0.01) 2) the HSI, which provides an indication of the status of energy reserves in the liver, defined by the follow- ing equation: HSI = ( LW l GW) * 100. (2) Ovaries collected were weighed to the nearest 0.1 g after fixation, and a portion of each sample (about 2.0 g) was removed from each gonad, dehydrated in ethanol, cleared in xylol, and embedded in paraffin. Sections of ovaries that were 5 pm thick were mount- ed and stained with Harris’s hematoxylin followed by eosin counterstain. Histological diagnosis was used to discard ovaries with evidence of recent spawning. Batch fecundity (BF), or the number of oocytes re- leased per spawning, and relative fecundity (RF), or the number of hydrated oocytes per gram of body weight, were estimated with the hydrated oocyte method on fixed ovarian samples (Hunter et al., 1985). Only ova- ries without evidence of recent spawning (no postovula- tory follicles) were selected (102 ovaries from samples collected in 2010 and 79 ovaries from samples collected in 2011). Three pieces of ovary, approximately 0. 1-0.2 g each, were removed from the anterior, middle, and posterior parts of one gonad and weighed to the near- est 0.1 mg, and hydrated oocytes were counted. Batch fecundity for each female was the product of the mean number of hydrated oocytes per unit of weight and the total weight of the ovaries. Relative fecundity was determined as the batch fecundity divided by female weight (without ovary). 328 Fishery Bulletin 111(4) The relationships of BF to TL obtained in 2010 and 2011 were compared on the basis of overlapping length ranges of the females (32-83 cm TL) with analysis of covariance on the log-transformed data (Draper and Smith, 1981). After comparison, the relationships of BF and RF to TL, gutted weight, age, HSI, and Kn ob- tained from samples and data collected in both years were described through the use of simple regression analysis. Egg quality A sample of 100 hydrated oocytes was removed from one ovary of each female selected for estimation of fe- cundity, rinsed in distilled water to remove formalde- hyde remnants, dried for 24 h at 60°C, and weighed to the nearest 0.1 mg. The DW of each of these sam- ples was considered an index of egg quality for the spawning females of Argentine Hake collected dur- ing the 2 research cruises because, in general, dry mass is associated with the quantity of yolk reserves stored in oocytes (Macchi et ah, 2006; Mehault et al., 2010). The relationships between DW and the differ- ent morphophysiological variables (TL, GW, age, HSI, and Kn), on the basis of length, weight, or organ-body weight ratios, were evaluated with simple regression analysis as were the fecundity data. Moreover, the re- lationships between DW and the number of oocytes spawned by batch (BF) and by unit of female weight (RF) were analyzed. To determine the effect of each maternal attribute on fecundity and egg quality and to establish which of these predictor variables had more influence on the number and quality of the oocytes, multiple regres- sion analyses of BF and DW in relation to morpho- physiological variables were carried out with the for- ward stepwise method. In the case of the regression model for BF, the data showed heteroscedasticity and were log-transformed. Colinearity between variables was analyzed to include in the model only those vari- ables that were uncorrelated. All statistical analyses were conducted with R software (R Development Core Team, 2010). Another variable possibly associated with egg qual- ity is the size of the oil droplet in mature eggs, be- cause it functions as an energy source for larvae (Wiegand, 1996; Nocillado et al., 2000; Rodgveller et al., 2012). For this reason, the diameters of oil drop- lets in hydrated oocytes were measured in samples of Argentine Hake females collected in January 2011. Samples collected in 2010 were not used, because hydrated oocytes from this year were in bad condi- tion as a consequence of long-term preservation in formalin solution. To analyze possible differences be- tween old and young females, 2 length groups were considered for this analysis: individuals <40 cm TL (n=12) and >70 cm TL (n=17). Most adult female Ar- gentine Hake <40 cm TL are 2-3 years old and con- sidered first-time spawners for the Patagonian stock of this species (Otero et al., 1986). First-time spawn- ers are thought to produce eggs of lower quality than the eggs of older females, as Trippel (1998) found for Atlantic Cod (Gadus morhua). Hydrated oocytes from each of the 2 length groups of females (n= 641 and 1000 oocytes) were sampled and the diameters of the oil droplets were measured to the nearest 0.01 pm with a Carl Zeiss3 stereomicroscope equipped with AxioVision 4.6 software (Carl Zeiss Microscopy, Jena, Germany). The mean diameters obtained for each size group were compared with a t-test. Because the size of hydrated oocytes in the ovaries may be influenced by the degree of hydration, we used spawned eggs to analyze the relationship between egg diameter and size of the oil droplet. Therefore, egg di- ameters and sizes of oil droplets were measured from Argentine Hake eggs collected in plankton samples, from oblique tows conducted with a bongo net (mesh sizes: 300 and 500 pm) in the research survey of Janu- ary 2010. Samples were fixed in 5% formalin, and lat- er Argentine Hake eggs in early developmental stages were sorted and rinsed in distilled water. Argentine Hake eggs were identified according to Ehrlich (1998). Egg diameter and size of the oil droplet (n=102) were measured to the nearest 0.01 pm, as was done for hy- drated oocytes, and the relationship between them was evaluated with simple regression analysis. Feeding activity during spawning Stomach fullness of Argentine Hake females during spawning was analyzed to find a possible relationship between feeding activity and the size of spawners. Information on stomach content obtained from all adult females sampled in the reproductive area during the 2 research cruises in January 2010 and 2011 (n=2091) was considered in this study. Individuals with everted stomachs as a consequence of pressure changes during capture were not used in this analysis. A gross scale of 4 stages was employed to classify the degree of stomach fullness: 0=empty, l=few contents (<25% full), 2=moderate contents (25-75% full) and 3=full stomach (Prenski and Angelescu4). The frequency distribution of each stomach stage by length class was considered in this analysis, but only stage 0 was em- ployed to determine differences in feeding activity. The relationship between the proportion of empty stomachs and female size was described with simple regression analysis. 3 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. 4 Prenski, L. B., and V. Angelescu. 1993. Ecologia trofica de la merluza comun ( Merluccius hubbsi ) del Mar Argentino. Parte 3. Consumo anual de alimento a nivel poblacional y su relacion con la explotacion de las pesquerfas multiespeci'ficas. INIDEP Documento Cienti'fico No. 1, 118 p. [Available from INIDEP, Paseo Victoria Ocampo Nro. 1, B76G2HSA Mar del Plata, Argentina.! Macchi et al : Maternal effects on fecundity and egg quality of Merluccius hubbsi 329 Results Female condition and fecundity estimation Batch fecundity estimates obtained for Argentine Hake females collected during January 2010 and 2011 ranged between 40,500 hydrated oocytes (29 cm TL) and 2,550,000 hydrated oocytes (95 cm TL). The anal- ysis of covariance applied to compare the coefficients of the relationship of BF to TL obtained for each year did not show statistical differences (P>0.01); therefore, data from 2010 and 2011 were pooled to estimate all the regression analyses. The relationship between BF and TL was curvilinear, fitting better to a power model (Fig. 2A, Table 1). The relationship of BF to GW was linear, and, in the case of BF to age, a power model was fitted (Figs. 2B and C, Table 1). A positive power relationship was observed be- tween BF and the HSI (Fig. 2D, Table 1). In the case of BF and Kn, the relationship was linear and sig- nificant (P<0.01), but a low r2 was obtained (Fig. 2E, Table 1). 330 Fishery Bulletin 111(4) Table 1 Results of the regression analyses between batch fecundity (BF), relative fecundity (RF), and oocyte dry weight (DW) and the variables of total length (TL), gutted weight (GW), age (A), hepatosomatic index (HSI), and relative condition index (Kn) for female Argentine Hake ( Merluccius hubbsi) of the Patagonian stock obtained from bottom trawls during research surveys in January 2010 and 2011. r2=coefficient of determination; a and b=parameters of the equation, n=sample size, P=P-value of the relationship. The sample size for regressions with age (n = 173) was smaller than the sample size for regressions with other variables («=181) because some otoliths were broken or otherwise unusable for age determination. Relationship r 2 a b n P TL Power 0.75 10.07 2.74 181 <0.01 GW Linear 0.81 -64235.60 675.48 181 <0.01 BF A Power 0.68 38103.41 1.67 173 <0.01 HSI Power 0.47 126319.30 1.17 181 <0.01 Kn Linear 0.04 -528543.19 1183337.78 181 <0.01 TL Linear 0.03 396.99 2.37 181 >0.01 GW Linear 0.04 471.76 0.05 181 <0.01 RF A Linear 0.03 448.49 15.12 173 >0.01 HSI Linear 0.11 388.49 33.56 181 <0.01 Kn Linear 0.07 21.68 500.84 181 <0.01 TL Logarithmic 0.18 2.62 0.0001 181 <0.01 GW Logarithmic 0.17 1.48 0.20 181 <0.01 DW A Logarithmic 0.14 2.31 0.34 173 <0.01 HSI Linear 0.18 2.49 0.08 181 <0.01 Kn Linear 0.05 2.01 0.82 181 <0.01 Relative fecundity ranged from 85 to 1040 hydrated oocytes g_1(female weighed without ovaries), with a mean value of 526 hydrated oocytes g^1 (standard er- ror [SE] 183). Positive linear significant relationships (P<0.01) between RF and the morphophysiological vari- ables of GW, HSI, and Kn were observed, but r2 values were low for all variables, explaining in some cases 3% or 4% of the variance (Fig. 3, Table 1). The strongest correlation, with an r2 of 0.11, was obtained with the relationship of RF to the HSI (Table 1). In the multiple regression analysis of the influence of maternal characteristics on BF, 87% of the variabil- ity in BF was explained by GW, DW, and the HSI. and described by the following equation: BF = 5.08 + 0.96LnGW + 1.08LnZ)W (3) + 0.27LnHS7 (r2=0.87, ?i=178, [P<0.01]) Most of this variability was explained exclusively by GW (95%). The forward stepwise method gave evidence that inclusion of other variables did not improve BF predictions significantly. Egg quality For 100 hydrated oocytes, DW estimated from Argen- tine Hake females collected during the spawning peak in January 2010 and 2011 ranged from 1.8 to 3.95 mg, and had a mean value of 2.83 mg (SE 0.36). There were significant positive relationships between DW and all morphophysiological variables considered dur- ing this study: TL, GW, age, HSI, and Kn (Fig. 4, Table 1). Moreover, positive significant relationships (P<0.01) also were obtained between DW and fecundity (BF and RF), described by the following equations: DW = 2.62 + 0.0003BP (4) (r2=0.25, n=181) DW = 2.42 + 0.0008RF. (5) (r2=0.16, n=181) In general, r2 values were low for all relationships estimated and BF was the variable best correlated with DW, although it is possible that size may have some in- fluence on this relationship. Regardless, RF, which was uncorrelated with TL, also showed a significant posi- tive relationship to DW. In the multiple regression analysis carried out with DW as the dependent variable and morphophysiologi- cal variables, only the HSI and GW were included be- cause the other variables did not improve DW predic- tions significantly. This model explained 20% of the total variability, and 90% of this percentage was con- tributed by the HSI: Macchi et al.: Maternal effects on fecundity and egg quality of Merluccius hubbsi 331 DW = 2.49 + 0.061 HSI + 0.00009GW. (6) (r2=0.20, n=178, [P<0.03]) The mean diameters of the oil droplets in hydrated oocytes that were estimated for young (<40 cm TL) and old (>70 cm TL) Argentine Hake females were 265.82 pm (SE 20.33) and 275.71 pm (SE 1.49), respective- ly. Comparison of this variable between both length groups showed highly significant differences (P<0.001), indicating more lipid accumulated in eggs from larger females than in eggs from first-time spawners. The regression analysis between the sizes of the oil droplets (OD) and egg diameters (ED) of Argentine Hake eggs collected during plankton sampling showed a significant positive relationship (P<0.01), described by the following equation: OD = 0.4168PD - 146.46. (7) (r2=0.34, n=102) This result corroborates the notion that larger eggs of Argentine Hake have more lipid reserves for larval growth and, therefore, have a higher probability that they will produce larger larvae. Feeding activity during spawning The regression analysis between the proportion of emp- ty stomachs (PE) and TL for data obtained in January 2010 and 2011 showed a significant negative logarith- mic relationship (P<0.01), indicating that intensity of feeding during spawning increases with female size. This tendency was more evident in Argentine Hake fe- males <55 cm TL because individuals of this species in larger length classes showed high variability in feeding activity (Fig. 5). This relationship was described by the following equation: PE = 1.6824 - 0.2937Ln TL. (8) (r2=0.39, n=45) Discussion Batch fecundity values for the Patagonian stock of Argentine Hake, estimated with samples collected in January 2010 and 2011, did not differ statistically and showed a positive relationship with size, weight, and age of the spawning females. The range of fecundity data (40,500-2,550,000 hydrated oocytes) was simi- lar to estimated values from previous studies of Ar- gentine Hake during the same month in other years (Macchi et al., 2004), but the range was much higher than values obtained for other species of Merluccius : Peruvian Hake (M. gayi peruanus ) (see Canal, 1989), New Zealand Hake (M. australis) (see Balbontin and Bravo, 1993), European Hake (M. merluccius) (see Murua et al., 2006; Recasens et al., 2008; El Habouz et al., 2011), Cape Hake (M. capensis), and Deepwater Hake (M. paradoxus) (see Osborne et al., 1999). Batch fecundity was influenced also by female condition, ex- pressed mainly by liver weight, as was evident from the positive relationship between the number of eggs produced by batch and the HSI. In addition, relative fecundity, a variable that did not show a significant re- lationship with female size in Argentine Hake (Macchi et al., 2004), also was influenced positively by the HSI. 332 Fishery Bulletin 111(4) HSI Figure 4 Dry weight of 100 hydrated oocytes as a function of (A) total length, (B) gutted weight, (C) age, (D) hepa- tosomatic index (HSI), and (E) relative condition factor (Kn) for the Patagonian stock of Argentine Hake (Merluccius hubbsi), which was sampled in waters off the province of Chubut in Argentina during the peak spawning season in 2010 and 2011. The hepatosomatic index was the maternal condi- tion measure that best correlated with fecundity in Argentine Hake in this study. Different authors have reported that the liver is the main deposit of lipids for gadoids and have considered this organ as the main source of energy for reproduction (Lambert and Du- til, 1997; Alonso-Fernandez, 2011). In addition, certain egg constituents, such as yolk proteins and egg coat substances, are synthesized in the liver (Brooks et al., 1997). It has been suggested that lipids play a key role during the early life history of many species (Nocillado et al., 2000), serving as structural components of mem- branes and providing physical protection of organs, buoyancy, or nutrients for embryos (Wiegand, 1996). Lipids of oocytes are stored mainly as lipoproteins, which are the major components of oocyte yolk, or as oil droplets (Wiegand, 1996). In Atlantic Cod, it has been suggested that the HSI is a good index of repro- Macchi et ai.: Maternal effects on fecundity and egg quality of Merluccius hubbsi 333 ( / ) -C 0 8 o E o 0.7 - to > 0 6 a E 70 cm TL) also reflected the maternal contribution to reproductive success. Oocytes of old individuals were characterized by oil globules that were larger than the oil globules of young spawners, indicating that more lipids were stored in the eggs of larger females. In other fish species that have been studied (e.g., Sebastes spp.), older females produced eggs with larger oil globules, which result in larvae with better chance of sur- vival (Berkeley et al., 2004; Sogard et al., 2008). Similar conclusions indicating that fatty acids are essential for egg composition and reproductive success were drawn with species maintained in captivity and fed with different concentrations of lipids (Watanabe et al., 1984; Fernandez-Palacios et al., 1995; Sewall and Rodgveller, 2009). Fur- thermore, the positive relationship observed in Argentine Hake in this study between the sizes of eggs from plankton samples and diameters of their oil droplets indicates that larger eggs contain a higher amount of lipids than do smaller eggs, sup- porting the idea that bigger oocytes are of better quality. The decrease in frequency of empty stomachs with the size of the Argentine Hake females is evidence that feeding activity is higher in larger individuals. In addition, this result supports the hypothesis of continuous feeding during breed- ing season for other hake species (Dominguez-Petit and Saborido-Rey, 2010). The relationship between the percentage of individuals with empty stomachs and TL was more evident in females <55 cm TL; at this size range, the slope of the equation was more pronounced. This break in the regression model is similar to the break previously found for the relationship of oocyte weight to female size in Argentine Hake (Macchi et al., 2006). Those authors reported that differences between egg mass and female size can be observed mainly in young spawners (<55 cm TL) because, in females >55 cm TL, the slope of the model decreased, as a conse- quence of the high variability in oocyte weight. Female condition and the quality of eggs produced by Argentine Hake may be associated with the feeding range of this species. It is possible that larger individu- als would be capable of swimming longer distances as they search for food in deeper waters, where it is com- mon to find great concentrations of squid ( Ilex argenti- nus ) in the austral summer. This squid species is con- sidered the most important prey for larger individuals of Argentine Hake in the Patagonian area (Angelescu and Prenski, 1987). A recent study described an off- shore spawning group of Argentine Hake in Patagonian waters in January that was characterized by a high proportion of females >55 cm TL and older than 5 years (Macchi et al., 2010). The selection of squid for food by larger females may be advantageous for spawning, if the high concentration of fatty acids detected in this mollusk is taken into account (Watanabe et al., 1984). Therefore, studies on Argentine Hake diet and the ef- fect of diet composition on egg quality are needed. In particular, further information is required about lipid, fatty acids, proteins, and ascorbic acid content, all rel- 334 Fishery Bulletin 111(4) evant components that affect egg and embryo survival (Brooks et al., 1997; Sewall and Rodgveller, 2009). Conclusions This study provides further evidence of maternal influ- ence on the reproductive potential and spawning qual- ity of the Patagonian stock of Argentine Hake. Within the analyzed size range (29-95 cm TL), larger females were characterized not only by higher fecundities but also by heavier eggs with larger oil droplets. Larger individuals were in better condition than smaller ones, as shown by higher HSI values, supporting the assump- tion that this index is a good predictor of reproductive potential. Moreover, feeding activity during the spawn- ing peak in January was higher for larger females than for smaller and younger individuals, indicating that Argentine Hake may incorporate energy during the re- productive season. This strategy of energy allocation is typical of an income breeder and is similar to the strategy suggested for the European Hake by Domin- guez-Petit and Saborido-Rey (2010). However, because Argentine Hake may be an opportunistic feeder dur- ing reproduction, it may be advantageous for spawning females to feed on higher quality food, such as squid, found offshore. This benefit may be restricted to larger individuals because they are more likely than smaller females to have the ability to find food offshore. Acknowledgments We wish to thank M. Estrada and H. Brachetta for preparation of the histological sections. Special thanks to M. Iorio and C. Dato for their help during sample collection and the technical staff of the Hake Assess- ment Group of the INIDEP for age determination. We also would like to thank the anonymous reviewers for making suggestions to improve this manuscript. This work was supported by the INIDEP and the CONICET (PIP 112 200801 00815). This is INIDEP contribution no. 1822. Literature cited Alonso-Fernandez, A. 2011. Bioenergetics approach to fish reproductive poten- tial: case of Trisopterus luscus (Teleostei) on the Gali- cian Shelf (NW Iberian Peninsula). Ph.D. diss., 266 p. Univ. Vigo, Vigo, Spain. Alvarez-Colombo, G., C. Dato, G. J. Macchi, E. Palma, L. Machinandiarena, H. E. Christiansen, P. Betti, P. Martos, F. Castro-Machado, D. Brown, M. Ehrlich, H. Mianzan, and M. Acha. 2011. Distribution and behavior of Argentine hake lar- vae: evidences of a biophysical mechanism for self-re- cruitment at the North Patagonian shelf waters. Cienc. Mar 37:633-657. Angelescu, V., and L. B. Prenski. 1987. Ecologia trofica de la merluza comun del mar argen- tine (Merluccidae, Merluccius hubbsi). Parte 2. Dinamica de la alimentacion analizada sobre la base de las condi- ciones ambientales, la estructura y las evaluaciones de los efectivos en su area de distribution. Contrib. Inst. Nac. Invest. Desarr. Pesq. 561:1-205. Aristizabal, E. O. 2007. Energy investment in the annual reproduc- tive cycle of red porgy ( Pagrus pagrus). Mar. Biol. 152:713-724. Aubone, A., S. Bezzi, R. Castrucci, C. Dato, P. Ibanez, G. Irusta, M. Perez, M. Renzi, B. Santos, N. Scarlato, M. Simonazzi, L. Tringali, and F. Villarino. 2000. Merluza ( Merluccius hubbsi). In Sintesis del estado de las pesquerias maritimas argentinas y de la Cuenca del Plata. Anos 1997-1998, con una actual- ization de 1999 (S. Bezzi, R. Akselman, and E. Boschi, eds.), p. 29-40. Publicaciones Especiales INIDEP (In- stitute Nacional de Investigation y Desarrollo Pesque- ro), Mar del Plata, Argentina. Balbontin, F., and R. Bravo. 1993. Fecundidad, talla de la primera madurez sexual y datos biometricos en la merluza del sur Merluccius australis. Rev. Biol. Mar. 28:111-132. Berkeley S. A, C. Chapman, and S. M. Sogard. 2004. Maternal age as a determinant of larval growth and survival in a marine fish, Sebastes melanops. Ecol- ogy 85:1258-1264. Bromage, N., M. Bruce, N. Basavaraja, K. Rana, R. Shields, C. Young, J. Dye, P. Smith, M. Gillespie, and J. Gamble. 1994. Egg quality determinants in finfish: the role of overripening with special reference to the timing of strip- ping in the Atlantic halibut Hippoglossus hippoglossus. J. World Aquacult. Soc. 25:13-21. Brooks, S., C. R. Tyler, and J. P. Sumpter. 1997. Egg quality in fish: what makes a good egg? Rev. Fish Biol. Fish. 7:387-416. Canal, R. L. 1989. Evaluation de la capacidad y condition reproducti- va de la merluza (Merluccius gayi peruanus). Bol. Inst. Mar Peru 13:40-70. Dominguez-Petit, R., and F. Saborido-Rey. 2010. New bioenergetic perspective of European hake ( Merluccius merluccius L.) reproductive ecology. Fish. Res. 104:83-88. Dominguez-Petit, R., F. Saborido-Rey, and I. Medina. 2010. Changes of proximate composition, energy storage and condition of European hake (Merluccius merluc- cius L. 1758) through the spawning season. Fish. Res. 104:73-82. Draper, N. R., and H. Smith. 1981. Applied regression analysis, 2nd ed., 709 p. Wi- ley, New York. El Habouz, H., L. Recasens, S. Kifani, A. Moukrim, A. Bou- haimi, and S. El Ayoubi. 2011. Maturity and batch fecundity of the European hake (Merluccius merluccius, Linnaeus, 1758) in the eastern central Atlantic. Sci. Mar. 75:447-454. Ehrlich, M. D. 1998. Los primeros estadios de vida de la merluza Mer- luccius hubbsi Marini 1933, como aporte al conocimien- to de su reclutamiento y estructura poblacional. Ph.D. diss., 318 p. Facultad de Ciencias Exactas y Naturales, Univ. Buenos Aires, Buenos Aires, Argentina. Macchi et al.: Maternal effects on fecundity and egg quality of Merluccius hubbsi 335 Fernandez-Palacios, H., M. S. Izquierdo, L. Robaina, A. Valen- cia, M. Salhi, and J. M. Vergara. 1995. Effect of n-3 HUFA level in broodstock diets on egg quality of gilthead seabream ( Sparus aurata L.). Aquaculture 132:325-337. Hinckley, S. 1990. Variation of egg size of walleye pollack Theragra chalcogramma with a preliminary examination of the effect of egg size on larval size. Fish. Bull. 88:471-483. Hunter, J. R., N. C. H. Lo, and R. J. H. Leong. 1985. Batch fecundity in multiple spawning fishes. In An egg production method for estimating spawning bio- mass of pelagic fish: application to the northern ancho- vy, Engraulis mordax (R. Lasker, ed.), p. 67-77. NOAA Tech. Rep. NMFS 36. Kjesbu, O. S., P. Solemdal, P. Bratlan, and M. Fonn. 1996. Variation in annual egg production in individual captive Atlantic cod (Gadus morhua ). Can. J. Fish. Aquat. Sci. 53:610-620. Lambert, Y., and J. D. Dutil. 1997. Can simple condition indices be used to monitor and quantify seasonal changes in the energy reserves of Atlantic cod ( Gadus morhua )? Can. J. Fish. Aquat. Sci. 54:104-112. Lambert Y., N. A. Yagarina, G. Kraus, G. Marteinsdottir, and P. J. Wright. 2003. Using environmental and biological indices as proxies for egg and larval production of marine fish. J. Northwest Atl. Fish. Sci. 33:115-159. Landaeta, M. F., and L. R. Castro. 2012. Seasonal and annual variation in Chilean hake Merluccius gayi spawning locations and egg size off Central Chile. Prog. Oceanogr. 92-95:166-177. Leonarduzzi, E. , A. Massa, and E. Manca. 2012. Variacion de la composicion bioqufmica en hem- bras de merluza comun ( Merluccius hubbsi) durante el ciclo reproductive. Rev. Invest. Desarr. Pesq. 21:5-19. Macchi, G. J., P. Martos, R. Reta, and C. Dato. 2010. Offshore spawning of the Argentine hake (Merluc- cius hubbsi) Patagonian stock. PanamJAS 5:22-35. Macchi, G. J., M. Pajaro, and C. Dato. 2007. Spatial variations of the Argentine hake ( Merluc- cius hubbsi) spawning shoals in the Patagonian area during a reproductive season. Rev. Biol. Mar. Oceanog. 42:345-356. Macchi, G. J., M. Pajaro, and M. Ehrlich. 2004. Seasonal egg production pattern of the Patagonian stock of Argentine hake (Merluccius hubbsi). Fish. Res. 67:25-38. Macchi, G. J., M. Pajaro, and A. Madirolas. 2005. Can a change in the spawning pattern of Argen- tine hake ( Merluccius hubbsi) affect its recruitment? Fish. Bull. 103:445-452. Macchi, G. J., M. Pajaro, M. I. Militelli, N. Radovani, and L. Rivas. 2006. I nftuence of size, age and maternal condition on the oocyte dry weight of Argentine hake (Merluccius hubbsi). Fish. Res. 80:345-349. Marshall, T., O. S. Kjesbu, N. A. Yaragina, P. Solemdal, and O. Ulltang. 1998. Is spawner biomass a sensitive measure of the re- productive and recruitment potential of Northeast Arc- tic cod? Can. J. Fish. Aquat. Sci. 55:1766-1783. Marshall, C. T., N. A. Yaragina, Y. Lambert, and O. S. Kjesbu. 1999. Total lipid energy as a proxy for total egg produc- tion by fish stocks. Nature 402:288-290. Marteinsdottir, G., and G. A. Begg. 2002. Essential relationships incorporating the influence of age, size and condition on variables required for esti- mation of reproductive potential in Atlantic cod Gadus morhua. Mar. Ecol. Prog. Ser. 235:235-256. Marteinsdottir G., and K. Thorarinsson. 1998. Improving the stock-recruitment relationship in Icelandic cod (Gadus morhua L.) by including age diversity of spawners. Can. J. Fish. Aquat. Sci. 55:1372-1377. Mehault, S., R. Dominguez-Petit, S. Cervino, and F. Saborido-Rey. 2010. Total egg production and implications for manage- ment of the Southern stock of European hake. Fish. Res. 104:111-122. Murua, H., P. Lucio, M. Santurtun, and L. Motos. 2006. Seasonal variation in egg production and batch fe- cundity of European hake Merluccius merluccius (L.) in the Bay of Biscay. J. Fish Biol. 69:1304-1316. Nissling, A., R. Larsson, L. Vallin, and K. Frohlund. 1998. Assesment of egg and larval viability in cod, Gadus morhua : methods and results from an experimental study. Fish. Res. 38:169-186. Nocillado, J. N., V. D. Penaflorida, and I. G. Borlonga. 2000. Measures of egg quality in induced spawns of the Asian sea bass, Lates calcarifer Bloch. Fish Physiol. Biochem. 22:1-9. Osborne, R. F., Y. C. Melo, M. D. Hofmeyer, and D.W. Japp. 1999. Serial spawning and batch fecundity of Merluc- cius capensis and M. paradoxus. S. Afr. J. Mar. Sci. 21:211-216. Otero, H. O., M. S. Giangiobe, and M. A. Renzi. 1986. Aspectos de la estructura de poblacion de la mer- luza ( Merluccius hubbsi). II Distribucion de tallas y eda- des. Estadios sexuales. Variaciones estacionales. Publ. Com. Tec. Mix. Fr. Mar. 1( 1):147— 179. R Development Core Team. 2010. R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. [Available from: http://R-project.org, accessed December 2010.] Recasens, L., V. Chiericone, and P. Belcari. 2008. Spawning pattern and batch fecundity of the Eu- ropean hake (Merluccius merluccius (Linnaeus, 1758)) in the western Mediterranean. Sci. Mar. 72:721-732. Renzi M. A., and M. A. Perez. 1992. Un criterio para la determination de la edad en juveniles de merluza (Merluccius hubbsi) mediante la lectura de otolitos. Frente Maritimo 1 1( A): 15—3 1 . Rijnsdorp, A. D., and B. Vingerhoed. 1994. The ecological significance of geographical and seasonal differences in egg size in sole Solea solea (L.). Neth. J. Sea Res. 32:255-270. Rodgveller, C. J., C. R. Lunsford., and J. T. Fujioka. 2012. Effects of maternal age and size on embryonic energy reserves, developmental timing, and fecundity in quillback rockfish (Sebastes malliger). Fish. Bull. 110:35-45. Sanchez, R. P, G. Navarro, and V. Rozycki. 2012. Estadfsticas de la Pesca Marina en la Argen- 336 Fishery Bulletin 111(4) tina. Evolucion de los desembarques 1898-2010, 528 p. Ministerio de Agricultura, Ganaderia y Pesca de la Nacion, Buenos Aires, Argentina. Sewall F., and C. J. Rodgveller. 2009. Changes in body composition and fatty acid pro- file during embryogenesis of quillback rockfish ( Sebastes maliger). Fish. Bull. 107:207-220. Sogard, S. ML, S. A. Berkeley, and R. Fisher. 2008. Maternal effects in rockfishes Sebastes spp.: a comparison among species. Mar. Ecol. Prog. Ser. 360:227-236. Trippel, E. A. 1998. Egg size and viability and seasonal offspring pro- duction of young Atlantic cod. Trans. Am. Fish. Soc. 127:339-359. Watanabe, T., T. Arakawa, C. Kitajima, and S. Fujita. 1984. Effect of nutritional quality of broodstock diets on reproduction of Red Sea bream. Bull. Jap. Soc. Sci. Fish. 50:495-501. Wiegand, M. D. 1996. Composition, accumulation and utilization of yolk lipids in teleost fish. Rev. Fish Biol. Fish. 6:259-286. 337 Using measurements of muscle cell nuclear RNA with flow cytometry to improve assessment of larval condition of Walleye Pollock IGadus chalcogrammus ) Email address for contact author: steve.porter@noaa.gov Resource Assessment and Conservation Engineering Division Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 7600 Sand Point Way, NE Seattle, Washington 98115-6349 Abstract— Nuclear RNA and DNA in muscle cell nuclei of laboratory- reared larvae of Walleye Pollock (Gadus chalcogrammus) were si- multaneously measured through the use of flow cytometry for cell-cycle analysis during 2009-11. The addi- tion of nuclear RNA as a covariate increased by 4% the classification accuracy of a discriminant analysis model that used cell-cycle, tempera- ture, and standard length to mea- sure larval condition, compared with a model without it. The greatest improvement, a 7% increase in accu- racy, was observed for small larvae (<6.00 mm). Nuclear RNA content varied with rearing temperature, in- creasing as temperature decreased. There was a loss of DNA when lar- vae were frozen and thawed because the percentage of cells in the DNA synthesis cell-cycle phase decreased, but DNA content was stable during storage of frozen tissue. Manuscript submitted 30 November 2012. Manuscript accepted 12 August 2013. Fish. Bull. 111:337-351. doi: 10.7755/FB.111.4.4 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necesarily reflect the position of the National Marine Fisheries Service, NOAA. Steven M. Porter (contact author) Kevin M. Bailey Marine fish larvae are most vulner- able to starvation during the first few weeks after they begin to feed (O’Connell, 1980; Theilacker, 1986; Theilacker and Porter, 1995). Star- vation may contribute directly to mortality, or it may do so indirect- ly by making larvae more vulner- able to predation (Bailey and Yen, 1983; Folkvord and Hunter, 1986). Many methods, such as the RNA- DNA ratio (Buckley et al., 1999), lipid composition (Lochmann et al., 1995), histological condition of tis- sues (Theilacker, 1978), morphologi- cal measurements (Theilacker, 1978), and flow cytometry (Theilacker and Shen, 1993a), have been developed to measure the physiological condi- tion of fish larvae. Mortality rate has been shown to correlate with nutri- tional condition of fish larvae (Thei- lacker et al., 1996). Hence, accurate assessment of condition can improve understanding of environmental pro- cesses that affect early life survival and recruitment. Previous studies have assessed the condition and growth of fish lar- vae with flow-cytometric cell-cycle analysis, which is a technique that measures the DNA content of indi- vidual cells in a population to deter- mine the proportion of cells in differ- ent phases of the cell cycle (Theilack- er and Shen, 1993b, 2001; Bromhead et al., 2000; Gonzalez-Quiros et al., 2007; Porter and Bailey, 2011; Do- mingos et al., 2012). Cells that are in the process of dividing can have up to twice the amount of DNA as cells that are not dividing. The cell cycle consists of discrete phases: gap 1 (Gl), DNA synthesis (S), gap 2 (G2), and mitosis (M) (Murray and Hunt, 1993). Cell growth occurs during the Gl phase before cell division begins, and cells may enter a GO resting state from this phase in response to starvation or other unfavorable environmental conditions (Murray and Hunt, 1993). For cells to divide, they must first replicate their DNA (S phase) and then grow and produce the structures (G2 phase) necessary for mitosis. Cell division occurs dur- ing mitosis. The proportion of cells in the S, and G2 and M (G2/M) phases are in- dicative of cells that may divide, and the use of flow-cytometric cell-cycle analysis to assess physiological con- dition is founded on the premise that cell proliferation is related to condi- tion. Specific tissue types (e.g., Thei- lacker and Shen, 1993b; Gonzalez- Quiros et al., 2007; Porter and Bai- ley, 2011) or individual whole larval homogenates (Domingos et al., 2012) have been used with this method, 338 Fishery Bulletin 111(4) and flow-cytometric cell-cycle analysis is as responsive to starvation as the RNA:DNA ratio in that it is able to detect starvation within a few days of no feeding (Porter and Bailey, 2011). Flow cytometry has been used to measure the rela- tive amount of cellular RNA and DNA in the brain cells of fish larvae (Theilacker and Shen, 1993b, 2001). Brain cells in the GO and G1 phases were separated through the use of ribonuclease (RNAse): GO were quiescent cells with low RNA content, and G1 were growing cells with high RNA content (Theilacker and Shen, 1993b). Theilacker and Shen (1993b) showed that the propor- tion of high-RNA-content G1 cells differed between fed and starved larvae, and it was suggested that those cells gave an indication of short-term changes in past feeding history. Nuclear RNA (nRNA) may react faster to metabolic changes than does cellular RNA, making it more sensitive to environmental change (Piwnicka et al., 1983) and potentially useful for assessment of the condition of fish larvae. Most nRNA is contained in the nucleolus (Li et al., 2006), the site of ribosome biogenesis (Sirri et al., 2008), and ribosome production correlates with cellu- lar growth (Caldarola et al., 2009). For many cell types grown in culture, nRNA content of G1 cells is highly variable, and a specific amount may be required for entry into the S phase (Darzynkiewicz et al., 1980; Piwnicka et al., 1983; Staiano-Coico et al., 1989). The threshold amount needed for G1 cells to progress into the S phase has been defined as the minimum nRNA content of S-phase nuclei determined from plots of nRNA and DNA fluorescence measured by flow cytom- etry (Darzynkiewicz et al., 1980). In a previous study with flow cytometry (Porter and Bailey, 2011), cell-cycle information (fraction of nuclei in the S and G2/M phases), larval standard length (SL), and temperature were used as covariates in a labora- tory-developed model for measurement of the condition of larvae of Walleye Pollock ( Gadus chalcogrammus). In the study that we describe here, we found that an additional covariate based on nRNA improved the clas- sification accuracy of a similar condition model by more clearly defining healthy (i.e., feeding, growing) and un- healthy (i.e., starving) larval conditions. We also exam- ined the effect of rearing temperature on nRNA mea- surements and the effect of storage on frozen tissue used for measurements with flow cytometry. Materials and methods Larval rearing Adult Walleye Pollock were collected by trawl in She- likof Strait, Gulf of Alaska, during the spawning sea- son in March 2009 and 2010 by the Alaska Fisheries Science Center (AFSC). For our experiments in 2009 and 2010, eggs from a single fish pairing (one female and one male) were fertilized and maintained aboard ship in the dark at 3°C before they were transported to the AFSC in Seattle, Washington. Eggs used for 2011 experiments came from a brood stock of adults kept at the AFSC laboratory at the Hatfield Marine Science Center in Newport, Oregon, and were also the result of a single fish pairing. In 2009 and 2010, larvae were reared in 2 feeding treatments: an always-fed treatment, in which larvae were considered healthy, and an unfed (starved) treat- ment in which larvae were considered unhealthy. Only the always-fed treatment was used in 2011. Rearing methods are described in Porter and Bailey (2011). Two replicate tanks were used for each feeding treatment. The size of the tanks varied with each experiment: 120 L in 2009, 20 L in 2010, and 60 L in 2011. The diet of larvae in the always-fed treatment consisted of laboratory-cultured rotifers ( Brachionus plicatilis ) that were fed an algal diet (Isochrysis galbana and Pavlova lutheri ) and a commercial rotifer supplement. Rotifers were maintained in the rearing tanks at a concentra- tion of 10 mL-1. Natural zooplankton, which included primarily copepod nauplii ( Acartia spp.) and gastro- pod veligers, were collected from a local lagoon and screened through 202-pm mesh; they also were main- tained in the always-fed rearing tanks, at a concentra- tion of 3 mL-1. A 16-h daylight cycle with a light level of 2.5 pmol photon mr2 s-1 from overhead, full-spectrum fluorescent lights was used, and larvae were sampled 4-6 h after the lights turned on at 0600 h. To avoid sampling larvae that were not actively feeding and pos- sibly unhealthy, only larvae that had prey in their gut were sampled from the always-fed treatment. Rearing temperatures varied: 6.0°C in 2009, 2.9°C, 5.9°C, and 8.7°C in 2010, and 6.5°C in 2011. 2009 experiments: nRNA staining protocol and covariate Three methods for preservation of larvae were tested to determine which of them was optimal for simultane- ous staining of DNA and nRNA in muscle cell nuclei of larvae of Walleye Pollock: 1) storage at -80°C, 2) a methanol treatment, and 3) storage at -80°C followed by a methanol treatment for 15 min before tissue pro- cessing. To stain DNA for cell-cycle analysis with flow cytometry, 4',6-diamidino-2-phenyIindole (DAPI), a flu- orescent DNA stain, was used at a concentration of 10 pg mL-1, and nRNA was stained with Invitrogen Syto RNASelect1 green fluorecent cell stain (S32703, Life Technologies Corp., Carlsbad, CA), hereafter referred to as Syto RNASelect stain. Preservation of tissue by freezing works well for DAPI-DNA staining (Theilacker and Shen, 2001), and methanol-preserved tissue has been stained suc- cessfully with Syto RNASelect stain (Life Technolo- 1 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA Porter and Bailey: Using measurements of muscle cell nuclear RNA to assess larval condition of Gadus chalcogrammus 339 gies Corp.2; Molecular Probes, Inc.3). Syto RNASelect stain is known to also bind to DNA and emit an ex- tremely weak fluorescent signal; therefore, we exam- ined through the use of a staining study the effect that varying concentrations of Syto RNASelect stain had on the DAPI-DNA interaction to confirm that the stain concentration we used did not interfere with DAPI staining. Three concentrations of Syto RNASelect stain were mixed in DAPI and tested: 500 nM (concentra- tion recommended by Invitrogen), 1000 nM (maximum concentration recommended by Invitrogen), and 2000 nM. The optimal preservative and stain concentration was defined as that which provided the highest nRNA fluorescence that did not affect DNA fluorescence as measured by flow cytometry. Each preservative was tested with the 3 concentrations of RNA stain previ- ously described and a DAPI-only negative control that was used to determine background RNA fluorescence. For preservative and stain testing, larvae were re- moved from a rearing tank, anesthetized in a 1% solu- tion of tricaine methanesulfonate (MS222), measured for their SL with an ocular micrometer, and placed in a 1.5-mL microcentrifuge tube that was either filled with methanol or inserted into a gel-filled microcentrifuge tube holder that was frozen at -80°C. Tubes were kept in the holder until they were transferred to a -80°C freezer. The nuclei of 29 always-fed larvae were pooled to have enough material to simultaneously test 3 stain concentrations and the control group for each type of preservative. Six replicate aliquots were taken for each stain concentration and control. A specific volume of Syto RNASelect stain was then added to each aliquot to equal the desired final concentration. Larvae were processed and analyzed as described in the section Materials and methods, Flow cytometry. To determine optimal RNA stain concentration, analysis of variance (ANOVA), Tukey’s tests, and Dunnett’s tests were used to compare nRNA and DNA fluorescence associated with each RNA stain concentration and the control. SYSTAT, vers. 13 (Systat Software, Inc., Chicago), was used for all statistical testing. After the optimal preservation method and stain concentration were determined, nRNA staining was confirmed through comparison of untreated muscle cell nuclei to those nuclei treated with RNAse. Nuclei were treated in a mixture of DAPI and RNAse A (50 pg mL_1 RNAse A, Sigma R6513, DNAse free) at room temperature for 25 min before adding the Syto RNAS- elect stain. The nuclei of 12 always-fed larvae were pooled to have enough nuclei to simultaneously test 2 Life Technologies Corp. 2013. Life Technologies: Syto RNASelect green fluorescent cell stain — 5 mM solution in DMSO. Life Technologies Corp., Carlsbad CA. [Product information webpage; Available from http://products. invitro- gen.com/ivgn/product/S32703, accessed August 2013.] 3 Molecular Probes, Inc. 2004. Syto RNASelect green fluo- rescent cell stain (S32703). Molecular Probes, Inc., Eugene, OR. [Product information document; Available from http:// tools, invitrogen.com/content/sfs/manuals/mp32703. pdf.] 3 treatments: a positive control (DAPI+RNA stain), a negative control (DAPI only), and a treatment (RNAse A+DAPI+RNA stain). Four replicate aliquots were tak- en for each treatment, and ANOVA and Tukey’s tests were used to compare RNAse A-treated nuclei to both the positive and negative controls. After the staining protocol, an nRNA covariate to test in a larval condition model was determined with larvae from the always-fed and unfed treatments. Ten individuals were taken from each replicate tank of both treatments at first feeding (defined as the day when 50% of the larvae had prey in their gut) and then at 4, 8, 11, and 14 days after first feeding. Five potential nRNA covariates were investigated: geometric mean fluorescence of nuclei in the GO and G1 (G0/G1), S, and G2/M phases of the cell cycle pooled, mean fluorescence of each cell-cycle phase (G0/G1, S, and G2/M) sepa- rately, and the ratio of the number of S-phase nuclei to the number of Gl-phase nuclei with high RNA con- tent (hereafter, this ratio will be referred to as RSG1; see Materials and methods, Flow cytometry section). RSG1 is a measure of potential cell division based on progression of nuclei from the G1 to S phase. An unhealthy larva would not be expected to grow or it would grow more slowly than a healthy individual. Therefore, an unhealthy larva would have fewer di- viding cells and potentially less S-phase nuclei, and its ratio of S-phase nuclei to Gl-phase nuclei with high RNA content would be smaller than the ratio for healthy individuals. Nuclear RNA fluorescence between treatments was compared with the 2-sample /-test. The Mann-Whitney U test was used to compare RSG1 be- tween treatments. 2010 experiment: nRNA, temperature, and condition Five larvae were sampled from each always-fed tank on 3 separate days after feeding began. A degree-day model (degree-day=temperature*fish age in days) was used so that larvae were sampled on the day of simi- lar developmental stage after first feeding, not on the same calendar day after first feeding (Table 1). Eight larvae from the unfed treatment were sampled from each tank daily, beginning at first feeding and ending at yolk exhaustion. All larvae were frozen at -80°C. Geometric mean nRNA fluorescence for nuclei in the G0/G1, S, and G2/M phases of the cell cycle pooled, and RSG1 for each larva was determined (see Materi- als and methods, Flow cytometry section). ANOVA and Tukey’s tests were used to examine differences in fluo- rescence and RSG1 between feeding treatments and among temperatures. 2011 experiment: effect of freezing and storage on nRNA and DNA After 4 days of feeding, 20 control larvae (not frozen) were processed directly from the rearing tanks (10 from each replicate tank) and analyzed by flow cytome- 340 Fishery Bulletin 1 1 1(4) Table 1 Days after first feeding when laboratory-reared larvae of Walleye Pollock ( Gadus chalcogrammus ) were sam- pled to determine the effect of temperature on nuclear RNA. Rearing tanks were maintained at 3 different temperatures. Because temperature can affect the growth rate of fishes, a “degree-day” model (degree- day=temperature*fish age in days) was used to ensure that that fish were taken at similar developmental stages. Rearing temperature PC) Sampling days after first feeding (determined by degree-day model) Sampling days (calendar days) after first feeding 2.9 17, 35, 61 6, 12, 21 5.9 18,35,65 3,6, 11 8.7 17,33,58 2,4,7 try (see Materials and methods, Flow cytometry section). On the same day, 60 larvae were frozen at -80°C in in- dividual tubes (30 from each tank) for flow-cytometric cell-cycle analysis at later dates. At 4 weeks after they were frozen and at additional intervals of 3, 6, and 10 months, 15 of these frozen larvae were randomly selected and analyzed with flow cytometry. Geometric mean nRNA fluorescence for nuclei in the G0/G1, S, and G2/M phases of the cell cycle pooled and RSG1 was determined for each larva (see Materials and methods, Flow cytometry section). The fraction of nuclei in the G0/G1, S, and G2/M phases of the cell cycle also was determined for each larva (see Materials and methods, Flow cytometry section). ANOVA, Dunnett’s tests, and Tukey’s tests were used to compare nRNA fluorescence, DNA fluorescence, and fractions of nuclei in the GO/ Gl, S, and G2/M phases between the control and frozen samples and among the frozen samples over time. Flow cytometry The tissue preparation protocol described in Porter and Bailey (2011), modified from Theilacker and Shen (2001), was followed. A frozen larva was placed on ice to thaw just before it was processed. A methanol-pre- served larva was processed directly from the preserva- tive. The larva was then placed on a glass depression slide into an approximately 100-pL mixture of DAPI and Syto RNASelect stains. The head and gut were dis- sected away from the trunk musculature, and the mus- cle tissue was sliced into 4 or 5 pieces with 2 scalpels. Only the pieces of muscle tissue were transferred into a microcentrifuge tube that contained a 230-pL mixture of DAPI and Syto RNASelect stains and the mixture was triturated 6 times with a 1-mL syringe with a 25-gauge needle to release the nuclei from the muscle cells. The solution was filtered through a 48- pm filter into another microcentrifuge tube to separate the stained nuclei from large cellular debris. Prepared samples were kept on ice until they were analyzed with a BD Biosciences Influx flow cytometer (BD Bio- sciences, San Jose, CA) typically within 4-5 h of prepa- ration. DAPI was excited with a 350-nm UV laser, and Syto RNASelect stain was excited with a 488-nm laser. The DAPI/DNA detector filter was 450/40, and the de- tector filter for Syto RNASelect stain/RNA was 525/30. Flourochrome compensation for overlapping emissions spectra of RNA and DNA stains is unnecessary when exciting DAPI and Syto RNASelect stain with the BD Biosciences Influx flow cytometer. The beams of the 350-nm and 488-nm lasers intercept the stream at spatially separate points. Each beam excites only the stain for DNA (350 nm) or RNA (488 nm). The emis- sion light is focused on separate mirror pinholes and is detected in separate modular detection blocks. Chicken and trout erythrocyte nuclei (Biosure, Inc., Grass Val- ley, CA) stained with the same mixture of DAPI and Syto RNASelect stains used for the muscle cell nuclei were used as controls. At the beginning of each flow cytometry session, each control type was run and necessary adjustments were made to the flow cytometer to keep control fluo- rescence values similar to previous sessions. DNA and RNA fluorescence values for larvae within the same ex- periment but measured during different flow cytometry sessions were made comparable by standardizing to a common control value. Samples that had <5000 nuclei analyzed or a coefficient of variation >9.00 for the cell- cycle phase of G0/G1 were not used in further analy- ses, and the use of this criteria resulted in rejection of about a third of all samples. For each larva, the fraction of nuclei in the G0/G1, S, and G2/M phases was calculated with MultiCycle A V software, vers. 4.0 (Phoenix Flow Systems, San Diego, CA). FCS Express flow cytometry analysis software, vers. 3.0 (De Novo Software, Los Angeles, CA) was used to calculate RSG1 and geometric mean fluorescence values for nRNA and DNA. DAPI area (DNA content, linear scale) in relation to DAPI height (DNA content, linear scale) was plotted for each larva, and a gate (a boundary used to enclose specific data points) was made to exclude debris, doublets, triplets, and large ag- gregates from nuclei in the G0/G1, S, and G2/M phases (Fig. 1A). Nuclei in those phases were located within the gate (Fig. IB). To calculate geometric mean values of nRNA and DNA fluorescence for each larva, a scatter plot of nRNA fluorescence (log scale) values in relation to DNA fluo- rescence (linear scale) values was made from the data enclosed within the gate for nuclei in the G0/G1, S, and G2/M phases. Each phase was gated separately for mean fluorescence values (Fig. 2), and a single gate that enclosed nuclei of all phases was used for the pooled mean fluorescence value. RSG1 was calculated with the same nRNA and DNA fluorescence scatter plot. Porter and Bailey: Using measurements of muscle cell nuclear RNA to assess larval condition of Gadus chalcogrammus 341 Figure t (A) Scatter plot of 4',6-diamidino-2-phenylindole (DAPI) area (DNA content, linear scale) and DAPI height (DNA content, linear scale) for a laboratory-reared larva of Walleye Pollock (Gadus chalcogrammus ) sampled in 2009, showing the gating strat- egy used to exclude debris, doublets, triplets, and large aggregates from muscle cell nuclei in the GO and G1 (G0/G1), S, and G2 and M (G2/M) phases of the cell cycle. Those nuclei are enclosed within the gate, which is represented by a rectangle. (B) Histogram showing the distribution pattern of the nuclei within the gate. Phases of the cell cycle: Gl=gap 1 or cell growth before cell division, G0=resting state from Gl, S=DNA synthesis, G2=gap 2 or cell growth before mitosis, and M=mitosis. A gate that divided Gl-phase nuclei into fractions of low and high nRNA content on the basis of nRNA fluorescence of S-phase nuclei was made by follow- ing Staiano-Coico et al. (1989). The smallest value of S-phase nRNA fluorescence was the minimum nRNA content needed for Gl nuclei to enter the S phase, and it was assumed that any nuclei with greater fluores- cence also could enter the S phase (Staiano-Coico et al., 1989). A gate that enclosed Gl nuclei was made, begin- ning at the smallest nRNA fluorescence value of the S-phase nuclei and extending to enclose the Gl nuclei with the highest nRNA fluorescence values (designated GIB; Fig. 3, A and B). The GIB group was defined as the number of nuclei with the potential to progress from the Gl to S phase. The number of S-phase nuclei was determined through the use of a gate that enclosed those nuclei (Fig. 3, A and B), and RSG1 was calculated by dividing that number by the number of GIB nuclei. 342 Fishery Bulletin 1 1 1 (4) 30,000 24.500 0 o c 0 o 0 0 | 19,000 < z Q 13.500 8,000 10° io1 io2 103 Nuclear RNA fluorescence Figure 2 Scatter plot of nuclear RNA (log scale) and DNA fluorescence (linear scale) of muscle cell nuclei in the GO and G1 (G0/G1), S, and G2 and M (G2/M) phases of the cell cycle for a laboratory-reared larva of Walleye Pollock (Gadus chalcogrammus ) sampled in 2009. The rectangles outlined in black indicate the gates used to calculate geo- metric mean values of nuclear RNA and DNA fluorescence. Fluorescence values are arbitrary units. Phases of the cell cycle: Gl=gap 1 or cell growth before cell division, G0=resting state from Gl, S=DNA synthesis, G2=gap 2 or cell growth before mitosis, and M=mitosis. Model testing Data from all 3 years were pooled for model testing. One-third of the data were randomly removed from that data set and used for independent cross-valida- tion testing. The remaining data were used to formu- late discriminant analysis models similar to the model described in Porter and Bailey (2011) to classify larvae as healthy (feeding and growing) or unhealthy (starv- ing). Models had SL, temperature, fraction of cells in the S phase, and fraction of cells in the G2/M phases as covariates, and were tested with and without the nRNA covariate included. The arcsin 4% transforma- tion was used to normalize the fraction of cells in the S phase, fraction of cells in the G2/M phases, and RSG1. Models were compared on the basis of the accuracy of their classification of the cross-validation data set and Akaike’s information criterion values (Burnham and Anderson, 2002). Results Nuclear RNA staining protocol Flow-cytometric cell-cycle analysis showed that the tis- sue from frozen samples had a small amount of debris and distinct peaks in G0/G1 and G2/M phases. Fixing larvae in methanol for either a short period of time or long-term storage did not work as well; samples con- tained a large amount of cellular debris and aggregates, and the nuclei did not disassociate from the tissue eas- ily. Samples from larvae preserved long term in metha- nol contained too much debris to be usable for further analysis. Debris was not as great in the samples that were frozen and then received a short-term methanol treatment as it was in the samples stored long term in methanol. The latter samples were usable, but they were not as clean as the samples that were frozen and not treated with methanol. Comparison of the RNA fluorescence between frozen tissues and tissues that were frozen and then treated with methanol indicated that short-term methanol preservation did not improve RNA staining; therefore, frozen tissue (stored at -80°C) worked best for preservation of muscle tissue from lar- vae of Walleye Pollock for nRNA and DNA staining and was used for all further tests and experiments. Syto RNASelect stain concentration affected both nRNA and DNA fluorescence. The fluorescence of nRNA for all stain concentrations was significantly higher than the values observed for the DAPI-only control (ANOVA, F(3;19)=165.94, P<0.001; Dunnett’s test, P<0.001 for each concentration; Fig. 4A). The nRNA fluorescence values of the DAPI+1000-nM Syto Porter and Bailey: Using measurements of muscle cell nuclear RNA to assess larval condition of Gadus chalcogrammus 343 30,000 24,500 19,000 13,500- 8,000 10 10 Nuclear RNA fluorescence 30,000 10 10 Nuclear RNA fluorescence Figure 3 Scatter plots of cell-cycle analysis by flow cytometry of muscle cell nuclei of (A) an always-fed larva (fed 4 days) and (B) an unfed larva (starved 14 days) of Walleye Pollock (Gadus chalcogrammus) reared in 2009. The rectangles outlined in black indicate the gates used to determine the number of Gl-phase nuclei with elevated nuclear RNA content needed to enter the S phase (GIB), and the total number of nuclei in the S phase of the cell cycle (S). The G1 phase of the cell cycle is when cell growth occurs before cell division, and the S phase is when DNA replicates. The always-fed larva had a distinct group of S-phase nuclei, and, for the unfed larva, S-phase nuclei were fewer and dispersed. Fluorescence values are arbitrary units. RNASelect stain concentration and DAPI+2000-nM Syto RNASelect stain concentration were significant- ly higher than the values for the DAPI+500-nM Syto RNASelect stain concentration (ANOVA, Fo igp 165.94, P <0.001; Tukey’s test, P<0.01 for both stain concentra- tions; Fig. 4A), but there was no significant difference in fluorescence between them (ANOVA, F(3 ig)=165.94, P<0.001; Tukey’s test, P=0.44; Fig. 4A). There was no significant difference in DNA fluores- cence between the control and the 500-nM and 1000- nM Syto RNASelect stain concentrations (ANOVA, F(3,19)=10.90, P<0.001; Dunnett’s test, P= 0.57 and 344 Fishery Bulletin 111 (4) 600 -1 0.05). Error bars indicate ±1 standard error of the geometric mean. Fluorescence values are arbitrary units. 0.44 respectively; Fig. 4B). The 2000-nM concentration caused a significant reduction of DNA fluorescence com- pared with the DNA fluorescence of the control (ANO- VA, F(3 i9)= 10.90, P<0.001; Dunnett’s test, P-0.002; Fig. 4B), indicating that a Syto RNASelect stain concentra- tion >1000 nM negatively affected DNA staining. Addi- tionally, no significant difference was observed in DNA fluorescence between the 500-nM and 1000-nM concen- trations (ANOVA, F(3 i9)=10.90, P<0.001; Tukey’s test, P=0.10; Fig. 4B). Therefore, the 1000-nM concentration was optimal for nRNA staining because it produced the highest nRNA fluores- cence and had no effect on DNA staining. The nRNA fluorescence of the RNAse- treated nuclei was significantly less than the values seen for the positive control (DAPI+ 1000-nM Syto RNASelect stain), indicating that nRNA was being stained (ANOVA, F(2,9)=386.26, P<0.0001; Tukey’s test, P<0.0001; Fig. 5A); however, the treated nuclei had a higher fluores- cence than the negative control (DAPI only; ANOVA, F(29)=386.26, P<0.0001; Tukey’s test, P<0.0001; Fig. 5A), indicat- ing fluorescence signal from stained DNA or the incomplete removal of nRNA from the samples. There was no significant dif- ference in DNA fluorescence among the 3 treatments (ANOVA, P(2 g)=0.009, P=0.99; Fig. 5B); therefore, it is most likely that the treatment with RNAse A did not completely remove all of the nRNA. Additionally, this experiment in- dependently confirmed that the 1000- nM concentration of the Syto RNASelect stain does not affect DNA staining be- cause there was no significant difference in DNA fluorescence between the group stained with DAPI only and the group stained with DAPI+ 1000-nM Syto RNAS- elect stain. Nuclear RNA covariate For each treatment, larval SL was not significantly different between replicate tanks (always-fed treatment, 2-sample f-test, £gi=0.196, P= 0.85; unfed treat- ment, 2-sample Utest, £43=0. 45, P=0.65), indicating that larvae in those tanks responded similarly to the same treat- ment. Therefore, replicate measurements of nRNA fluorescence from each tank for each treatment were pooled. The growth rate of larvae in the always-fed treat- ment from hatching to 19 days after hatching was 0.11 mm d_1, a rate that is typical for larvae of Walleye Pollock reared in a 6°C laboratory (Porter and Theilacker, 1999). To formulate the nRNA covariate, we used 113 larvae, 63 always-fed (healthy) and 50 un- fed (unhealthy). There was no significant difference in nRNA fluorescence between feeding treatments when all phases of the cell cycle were pooled or when phases were examined separately (Table 2), indicating that nRNA fluorescence was not a useful indicator of condi- tion. Unlike nRNA fluorescence, RSG1 was responsive to feeding conditions; therefore, it was chosen as the nRNA covariate for model testing. Porter and Bailey: Using measurements of muscle cell nuclear RNA to assess larval condition of Gadus chalcogrammus 345 DAPI DAP1+RNA stain RNAse Treatment 13,400 13,300 c 13,200 - 0) Q CO 0.05). Error bars indicate ±1 stan- dard error of the geometric mean. Fluorescence values are arbitrary units. Plots of nRNA and DNA fluorescence showed that always-fed larvae had a dis- tinct group of aggregated S-phase nuclei that joined the G0/G1 and G2/M phases, and S-phase nuclei were fewer and dis- persed for unfed larvae (Fig. 3, A and B). Overall RSG1 was significantly larger for always-fed larvae than for unfed larvae (0.23 [standard error of the mean 0.01] for always-fed and 0.16 [SE 0.01] for unfed larvae; Mann-Whitney U test, (7=2299.5, Ni=63, N2=50, P<0.001), clearly distin- guishing larvae between the 2 feeding treatments. For always-fed larvae, there was an initial increase in RSG1 and then a gradual decline to its initial value after 2 weeks of feeding (Fig. 6A). For unfed in- dividuals, RSG1 declined throughout the time period to less than half its initial value after 2 weeks of feeding (Fig. 6A). Nuclear RNA fluorescence did not show a distinct difference between feeding treat- ments until after 2 weeks of feeding (Fig. 6B). Nuclear RNA, temperature, and condition Growth in the always-fed treatment at the warmest temperature (8.7°C) was poor; therefore, larvae from this treatment were not used for further analyses. A growth rate of about 0.15 mm d-1 would be ex- pected (senior author, unpubl. data), but larvae grew 0.08 mm d-1. Both mean per- centage of nuclei in the S-phase (8.90) and mean RSG1 (0.17) were small for a typical healthy, feeding larva, supporting the ob- servation that those larvae were not grow- ing well. Growth rates at the temperatures of 5.9°C (0.10 mm d-1) and 2.9°C (0.06 mm d-1) were typical for larvae reared at those temperatures. There was no significant difference in overall nRNA fluorescence between the always-fed and unfed treatments (ANO- VA, F(ii61)=1.25, P=0.27; Table 3). Rearing temperature significantly affected over- all nRNA fluorescence. Larvae reared at 2.9°C had a higher fluorescence than lar- vae reared at 5.9°C (ANOVA, F(i,6i)=4-47, P=0.04; Table 3), indicating more RNA in muscle nuclei from larvae reared at the colder temper- ature. RSG1 was significantly higher for the always-fed treatment than for the unfed treatment at both tem- peratures (ANOVA, F(i 87)=59.65, P<0.01, Tukey’s test, P<0.01; Table 3), similar to the result for the experi- ment conducted in 2009. The effect of temperature on RSG1 was dependent on feeding treatment. RSG1 was smaller for always- fed larvae reared at 5.9°C compared with RSG1 of lar- vae reared at 2.9°C (ANOVA, F(i,87)=18.56, P<0.001, Tukey’s test, P87)=59.65, P<0.001, Tukey’s test, P=0.05). Model testing Data from experiments in 2009, in 2010 (temperatures: 2.9°C and 5.9°C), and in 2011 were pooled for model testing (n= 237). The 8.7°C experiment in 2010 was ex- cluded because larvae grew poorly in the always-fed treatment. A control model used temperature, SL, arcsin \fx -transformed fraction of cells in the S phase, and arcsin \fx -transformed fraction of cells in the G2/M phases; and a test model added arcsin V* -trans- formed RSG1 to the covariates used in the control model (n=158). For independent cross-validation test- ing, 49 always-fed larvae ranging from 5.36 to 8.64 mm SL and 30 unfed larvae from 5.20 to 5.92 mm SL were used (n= 79). Both models significantly dis- criminated between the always-fed and unfed treat- ment groups (control model, Wilks’s lambda=0.46, F(4 i53)=44.19, P<0.001; test model, Wilks’s lambda=0.45, ^(5A52)=37.13, P<0.001). The test model improved overall classification ac- curacy by 4%, and accuracy for both always-fed and unfed larvae increased (Table 4). The improvement in the always-fed treatment larvae was due to the correct classification of additional small, feeding larvae (<6.00 mm SL). The classification accuracy in the test model for those larvae improved 14%, increasing from 53% (8/15) to 67% (10/15) when RSG1 was used. Classifica- tion accuracy of unfed treatment larvae improved 3%, and the test model correctly classified all the unfed larvae tested (Table 4). For larvae <6.00 mm SL from both the always-fed and unfed treatments, classifica- tion accuracy improved 7% (an increase from 37/45 to 40/45). There was no difference in classification accu- racy between models for larvae >6.00 mm SL, and both models correctly classified all of those larvae. Akaike’s information criterion value for the test model was less than that value for the control model (-1130.53 and -1001.38, respectively), indicating that the addition of RSG1 improved model fit. Effect of freezing and storage on nRNA and DNA There was a loss of DNA when larvae were frozen and thawed, but that process did not affect nRNA mea- surements. The DNA fluorescence of fresh tissue was significantly greater than DNA fluorescence of frozen Porter and Bailey: Using measurements of muscle cell nuclear RNA to assess larval condition of Gadus chalcogrammus 347 Figure 6 For always-fed and unfed treatments of larvae of Walleye Pollock ( Gadus chalcogrammus) reared in 2009, (A) mean ratios of the number of S-phase nuclei to the number of Gl-phase nuclei with high nuclear RNA content (RSG1) and (B) geometric mean fluo- rescence of nuclear RNA of all cell-cycle phases pooled. The G1 phase of the cell cycle is when cell growth occurs before cell divi- sion, and the S phase is when DNA replicates. Error bars indicate ±1 standard error of the mean. Fluorescence values are arbitrary units. tissue (ANOVA, F4 48=36. 67, P<0.001, Dunnett’s tests, P<0.05 for all frozen groups; Table 5), indicating that DNA was lost when tissue was frozen and thawed. The percentage of nuclei in the G2/M phases was not sig- nificantly different between the fresh and frozen tissue treatments (ANOVA, 2*4 48= 2.00, P=0.11; Table 5), but freezing affected the percentage of nuclei in the GO/ G1 and S phases. There was a significant increase in the percentage of nuclei in the G0/G1 phases between fresh and frozen tissue (ANOVA, P4 48=19.90, P<0.001, Dunnett’s tests, P<0.01; Table 5), and freezing had the opposite effect on S-phase nuclei, namely a decrease in the percentage of nuclei in that phase (ANOVA, P4 4g=19.14, P<0.001, Dunnett’s tests, P<0.01; Table 5). The increase in the percentage of nuclei in the G0/G1 phases in the freezing treatment may be due to a loss of DNA from S-phase nuclei that caused them to be identified as nuclei in the G0/G1 phases on the basis of their fluorescence signal. There was no signif- icant change in the percentage of G0/G1- or S-phase nuclei among the 4 frozen groups of larvae tested (Tukey’s tests, P >0.35; Table 5), indicating that DNA was stable during stor- age of frozen tissue. This result indicates that DNA was lost by S-phase nuclei either during the freezing process or subsequent thawing. There was no significant difference in nRNA fluorescence between fresh and frozen tissues (ANOVA, P4 48=2.38, P=0.06; Table 5), indi- cating no loss of nRNA. RSG1 showed results similar to those as DNA in that RSG1 of fro- zen tissue was significantly less than RSG1 of fresh tissue, and it was stable during the 10 months of storage of frozen tissue (ANO- VA, F 4 16.70, P<0.001, Dunnett’s tests, P<0.01; Tukey’s tests >0.90; Table 5). Discussion We developed a protocol for staining nRNA in muscle cell nuclei of larvae of Walleye Pollock for use in flow cytometry, and we showed that the inclusion of an nRNA covariate in a mod- el resulted in more accurate measurement of physiological condition than did cell-cycle analysis alone. Accurate assessment of condi- tion of fish larvae is essential because small changes in mortality rate over a long period of time can strongly influence future recruit- ment (Houde, 1987). RSG1, based on nRNA fluorescence, proved to be an indicator of po- tential growth that was responsive to feeding conditions, and it contributed meaningful im- provement to the discriminant analysis model for assessment of larval condition, as shown by an increase in classification accuracy and a decrease in Akaike’s information criterion value. Data from 3 years were used for model testing; there- fore, results are not unique to a single group of larvae. The classification accuracy of small larvae (<6.00 mm SL) was most improved (7%). This result is important because that size class includes larvae that have just started to feed, and the S- and G2/M-phase fractions can be highly variable in first-feeding larvae and may not always distinctly indicate condition. Additionally, 348 Fishery Bulletin 111(4) Table 3 The ratio of the number of S-phase nuclei to the number of Gl-phase nuclei with high nuclear RNA content (RSG1) and values of nuclear RNA (nRNA) fluorescence for larvae of Walleye Pollock ( Gadus chalcogrammus ) sampled from always-fed and unfed treat- ments reared at different temperatures in 2010. Fluorescence values are arbitrary units and were adjusted on the basis of controls to make samples comparable among sessions of flow cytometry. Standard errors of geometric means of nRNA fluorescence (all cell- cycle phases pooled) and of RSG1 means are reported in parentheses. Cell-cycle phases: Gl=gap 1 or cell growth before cell division, G0=resting state from G1 phase, S=DNA synthesis, G2=gap 2 or cell growth before mitosis, and M=mitosis. Rearing temperature (°C) Treatment n RSG1 nRNA fluorescence 2.9 Unfed 28 0.14 (0.01) 32.3 (1.5) 2.9 Always-fed 26 0.26(0.02) 32.5 (2. OF 5.9 Unfed 13 0.09(0.01) 26.7 (0.8) 5.9 Always-fed 24 0.19(0.01) 30.5 (1.9)2 8.7 Unfed 17 0.13(0.02) 26.5 (1.1) 2n= 14. 2n= 10. the sizes of healthy and unhealthy larvae overlap in that size class. Nuclear RNA varied with rearing temperature, in- creasing as temperature decreased, a result similar to the findings of other studies on the effect of temper- ature on RNA content of larval fishes (Canino, 1994; Malzahn et al., 2003). This result also indicates that our nRNA staining protocol worked as intended. Temperature affected RSG1, and, therefore, it needs to be accounted for when our method is used to mea- sure the condition of larvae sampled from the field. RSG1 for healthy larvae was smaller at warmer tem- peratures (5.9°C in our study), and this observation may indicate that nuclei were cycling faster through the cell cycle than nuclei at colder temperatures. Oth- er studies have shown that increasing temperature decreases cell-cycle duration in other species, such as yeast ( Saccharomyces cereuisiae) (Jagadish and Carter, 1978) and Magellan Plunderfish (Harpagifer bispinis) (Brodeur et al., 2003). There was no difference in the percentage of S-phase brain cells of fed larvae of At- lantic Cod (Gadus morhua) reared at 6°C and 10°C — a finding that was explained by an increased rate of pro- gression by cells through the cell cycle at the higher temperature (Gonzalez-Quiros et al., 2007). A similar result was also found for S-phase nuclei from muscle cells of larvae of Walleye Pollock reared at 3.2°C and 5.9°C, temperatures comparable to those used in our study (Porter and Bailey, 2011). RSG1 of unhealthy larvae was not affected by tem- perature, probably as a result of the slow or ceased growth of these larvae. The brain cells of starved At- lantic Cod larvae reared at 10°C had a smaller per- centage of S-phase cells than the brain cells of larvae starved at 6°C (Gonzalez-Quiros et al., 2007), a differ- ence that may be due to the length of time that the larvae were starved. Fish larvae in general starve faster at higher temperatures, and, for larvae of Wall- eye Pollock in a previous study, the percentage of S- phase nuclei decreased the longer larvae were starved (Porter and Bailey, 2011). Atlantic Cod larvae at both temperatures were starved for 5 days; therefore, the percentage of S-phase cells of the larvae starved at the higher temperature (10°C) would be expected to be less than the percentage for the larvae starved at the lower temperature. The loss of DNA during freezing and thawing has been documented for human blood, where about 25% of the DNA was lost (Ross et al., 1990). Differences in nuclear membrane permeability among cell-cycle phases may account for the loss of DNA when larvae of Walleye Pollock were frozen and thawed, and they may explain why there was a decrease in the percent- age of muscle cell nuclei in the S phase. S-phase nuclei may be more permeable than G2-phase nuclei (Coverly et al., 1993; Leno and Munshi, 1994), resulting in diffu- sion of DNA out of the nucleus, and they may rupture because they may be more fragile than nuclei at other phases — an outcome that also would contribute to loss of nuclei. Crytoprotectant has been used to stabilize DNA in brain cells of Walleye Pollock larvae during freezing (Theilacker and Shen, 1993a), and it could possibly be used to prevent the loss of DNA from mus- cle cell nuclei as well, but the use of cryptoprotectant needs further investigation, particularly for between- laboratory comparisons where standardized protocols are used (Caldarone et al., 2006). In our study, the loss of DNA was unchanged for up to 10 months when the Porter and Bailey: Using measurements of muscle cell nuclear RNA to assess larval condition of Gadus chalcogrammus 349 Table 4 Results of larval condition for laboratory-reared larvae of Walleye Pollock ( Gadus chalcogrammus) sampled in 2009, 2010, and 2011. Independent cross-validation testing of models without (control) and with (test) the ratio of the number of S-phase nuclei to the number of Gl-phase nuclei with high nuclear RNA content (RSG1) in- cluded as a covariate. The arcsin Vx transformation was used to normalize the fraction of cells in the S phase of the cell cycle, fraction of cells in the G2/M phases of the cell cycle, and RSG1. G2/M=fraction of nuclei in the G2 and M phases combined. Cell-cycle phases: S=DNA synthesis, G2=gap 2 or cell growth before mitosis, and M=mitosis. Control model Covariates: standard length, temperature, arcsin -y/S- phase fraction , arcsin J G2/M- phase fraction Treatment Healthy Classification Unhealthy Percentage correct Always-fed 42 7 86 Unfed 1 29 97 Overall correct 90 Test model Covariates: standard length, temperature, arcsin ^/S- phase fraction , arcsin ^/G2/M- phase fraction , arcsin V RSG1 Treatment Healthy Classification Unhealthy Percentage correct Always-fed 44 5 90 Unfed 0 30 100 Overall correct 94 Table 5 Percentages of nuclei in the G0/G1, S, and G2/M phases of the cell cycle, the ratio of the number of S-phase nuclei to the number of Gl-phase nuclei with high nuclear RNA content (RSG1), nuclear RNA fluorescence, and DNA fluorescence for muscle cell nuclei from fresh and frozen larvae of Walleye Pollock (Gadus chalcogrammus) sampled in 2011. Fluorescence values are arbitrary units and were adjusted on the basis of controls to make samples comparable among sessions of flow cytometry. Standard errors of geometric means of nRNA and DNA fluorescence (all cell-cycle phases pooled), of means for RSG1 and percentages of nuclei in cell-cycle phases are reported in parentheses. G0/Gl=percentage of nuclei in GO and G1 phases combined, G2/M= percentage of nuclei in G2 and M phases combined. Cell-cycle phases: Gl=gap 1 or cell growth be- fore cell division, G0=resting state from G1 phase, S=DNA synthesis, G2=gap 2 or cell growth before mitosis, and M=mitosis. Treatment n Percentage of nuclei in G0/G1 Percentage of nuclei in S phase Percentage of nuclei in G2/M phases RSG1 nRNA fluorescence DNA fluorescence Fresh tissue 18 76.7 (1.2) 20.64 (0.97) 2.63 (0.42) 0.37 (0.03) 22.22 (0.43) 14,691.03 (194.93) Frozen 4 weeks 11 85.3 (0.8) 11.78 (1.27) 2.96 (0.73) 0.18 (0.01) 23.72 (0.68) 12,082.43 (137.11) Frozen 3 months 10 87.4(1.0) 8.22 (1.22) 4.36 (0.59) 0.18 (0.02) 21.30 (0.59) 12,866.44(187.72) Frozen 6 months 6 87.8 (1.4) 8.50 (1.85) 3.71 (1.05) 0.20 (0.02) 23.69 (0.68) 14,280.30 (144.20) Frozen 10 months 8 88.2 (1.4) 10.01 (2.00) 1.79 (0.84) 0.19 (0.03) 22.83 (0.82) 14,104.69 (221.88) 350 Fishery Bulletin 111(4) [ tissue was frozen at -80°C, indicating that DNA was stable during that time and under that condition. Neither freezing and thawing nor storage of frozen tissue affected RNA in muscle cell nuclei of larvae of Walleye Pollock in our study. For frozen human tissue, no significant RNA degradation (as measured by gene expression and electropherograms) was detected after 16 h on ice (Micke et ah, 2006), and another study showed that significant RNA degradation did not be- gin until 30 min after thawing at room temperature (Botling et al., 2009). Although those studies did not measure the amount of RNA present, they do support our assertion that the protocol used in our study was adequate to preserve RNA because larvae were frozen quickly, thawed tissue was kept cool on ice, and the time from tissue thawing to analysis with flow cytom- etry was typically not longer than 5 h. Our results dif- fer from the findings of Theilacker and Shen (1993b) in that their study indicated that a cryoprotectant and acid were needed before freezing to stabilize RNA in brain cells of larvae of Walleye Pollock. This difference in results may be due to the difference in type of tissue used (muscle cell nuclei versus whole brain cells) (Oli- var et al., 2009) or in the method used for preparation of tissue for flow cytometry. In Theilacker and Shen (1993b), tissue dissection occurred before freezing and whole cells were analyzed; however, in our study, tis- sue preparation occurred after freezing, and only nuclei were used. Our results indicate that only frozen tissue should be analyzed when condition is measured with the method described here. Results would be inaccurate if fresh tissue were used because of its higher fraction of S-phase nuclei. Cell-cycle measurements (fraction of S- and G2/M-phase nuclei pooled) of independent groups of larvae of Walleye Pollock reared under similar condi- tions were not significantly different (Porter and Bailey, 2011), indicating that the effect of freezing and thaw- ing was consistent. Freezing and thawing of tissue was also part of the method used in another study where standardized protocols were used for spectrofluoromet- ric analysis of RNA and DNA for assessing larval fish physiological condition (Caldarone et al., 2006). Conclusions An nRNA covariate improved larval condition measure- ments. For minimal cost (the cost of the Syto RNAS- elect stain), model accuracy increased, and the greatest improvement was for small larvae. The assay that we developed in our study quickly determines condition, and, therefore, many larvae can be analyzed in a short period of time. In addition, larvae can be kept frozen for at least 10 months without affecting condition measurements. There is no diel pattern in RNA con- tent or measurements of the cell-cycle phases for lar- vae of Walleye Pollock (Bailey et al., 1995; Theilacker and Shen, 2001); therefore, our method can be applied in the field, where sampling can occur anytime during a 24-h period. Because we found that temperature af- fected RSG1, future studies should include measure- ments at additional temperatures to formulate a model for field-sampled larvae. Acknowledgments We would like to thank A. Dougherty for collection of Walleye Pollock eggs and her assistance in the labo- ratory. D. Prunkard at the University of Washington, Department of Pathology, Cytometry Core Facility as- sisted with flow cytometry. F. Morado, M. Paquin and 3 anonymous reviewers provided helpful comments on earlier drafts of the manuscript. This research was funded by the North Pacific Research Board (NPRB grant no. 926, publication 432) and the Alaska Fisher- ies Science Center. It is contribution EcoFOCI-0800 to NOAA’s Ecosystems and Fisheries-Oceanography Coor- dinated Investigations. Literature cited Bailey, K. M., and J. Yen. 1983. Predation by a carnivorous marine copepod, Eu- chaeta elongata Esterly, on eggs and larvae of the Pa- cific hake, Merluccius productus. J. Plankton Res. 5:71-82. Bailey, K. M., M. F. Canino, J. M. Napp, S. M. Spring, and A. L. Brown. 1995. Contrasting years of prey levels, feeding condi- tions and mortality of larval walleye pollock Theragra chalcogramma in the western Gulf of Alaska. Mar. Ecol. Prog. Ser. 119:11-23. Botling, J., K. Edlund, U. Segersten, S. Tahmasebpoor, M. Eng- strom, M. Sundstrom, P. Malmstrom, and P. Micke. 2009. Impact of thawing on RNA integrity and gene ex- pression analysis in fresh frozen tissue. Diagn. Mol. Pathol. 18:44-52. Brodeur, J. C., J. Calvo, A. Clarke, and I. A. Johnston. 2003. Myogenic cell cycle duration in Harpagifer species with sub-Antarctic and Antarctic distributions: evidence for cold compensation. J. Exp. Biol. 206:1011-1016. Bromhead, D., J. Kalish, and P. Waring. 2000. Application of flow cytometric cell cycle analysis to the assessment of condition and growth in larvae of a freshwater teleost Galaxias olidus. Can. J. Fish. Aquat. Sci. 57:732-741. Buckley, L., E. Caldarone, and T. L. Ong. 1999. RNA-DNA ratio and other nucleic acid-based indi- cators for growth and condition of marine fishes. Hy- drobiologia 401:265-277. Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a prac- tical information-theoretic approach, 2nd ed., 497 p. Springer- Verlag New York, Inc., New York. Caldarola, S., M. C. De Stefano, F. Amaldi, and F. Loreni. 2009. Synthesis and function of ribosomal proteins — fading models and new perspectives. FEBS Journal 276:3199-3210. Porter and Bailey: Using measurements of muscle cell nuclear RNA to assess larval condition of Gadus chalcogrammus 351 Caldarone, E. M., C. M. Clemmesen, E. Berdalet, T. J. Miller, A. Folkvard, G. J. Holt, M. P. Olivar, and I. M. Suthers. 2006. Intercalibration of four spectroflourometric proto- cols for measuring RNA/DNA ratios in larval and juve- nile fish. Limnol. Oceanogr. Methods 4:153-163. Canino, M. F. 1994. Effects of temperature and food availability on growth and RNA/DNA ratios of walleye pollock Ther- agra chalcogramma (Pallas) eggs and larvae. J. Exp. Mar. Biol. Ecol. 175:1-16. Coverly, D., C. S. Downes, P. Romanowski, and R. A. Laskey. 1993. Reversible effects of nuclear membrane permea- bilization on DNA replication: evidence for a positive licensing factor. J. Cell Biol. 122:985-992. Darzynkiewicz, Z., F. Traganos, and M. R. Melamed. 1980. New cell cycle compartments identified by multi- parameter flow cytometry. Cytometry 1:98-108. Domingos, J. A., P. Fromm, C. Smith-Keune, and D. R. Jerry. 2012. A robust flow-cytometric protocol for assessing growth rate of hatchery-reared barramundi Lates cal- carifer larvae. J. Fish Biol. 80:2253-2266. Folkvord, A., and J. R. Hunter. 1986. Size-specific vulnerability of northern anchovy, Engraulis mordax, larvae to predation by fishes. Fish. Bull. 84:859-869. Gonzalez-Quiros, R., I. Munuera, and A. Folkvord. 2007. Cell cycle analysis of brain cells as a growth index in larval cod at different feeding conditions and tem- peratures. Sci. Mar. 71:485-497. Houde, E. D. 1987. Fish early life dynamics and recruitment variabil- ity. Am. Fish. Soc. Symp. 2:17-29. Jagadish, M. N., and B. L. A. Carter. 1978. Effects of temperature and nutritional conditions on the mitotic cell cycle of Saccharomyces cerevisiae. J. Cell Sci. 31:71-78. Leno, G. H., and R. Munshi. 1994. Initiation of DNA replication in nuclei from qui- escent cells requires permeabilization of the nuclear membrane. J. Cell Biol. 127:5-14. Li, Q., K. Yunkyung, J. Namm, A. Kulkarni, G. R. Rosania, Y-H. Ahn, and Y. T. Chang. 2006. RNA-selective, live cell imaging probes for study- ing nuclear structure and function. Chem. Biol. 13:615-623. Lochmann, S. E., G. L. Maillet, K. T. Frank, and C. T. Taggart. 1995. Lipid class composition as a measure of nutrition- al condition in individual larval Atlantic cod (Gadus morhua). Can. J. Fish. Aquat. Sci. 52:1294-1306. Malzahn, A. M., C. Clemmesen, and H. Rosenthal. 2003. Temperature effects on growth and nucleic acids in laboratory-reared larval coregonid fish. Mar. Ecol. Prog. Ser. 259:285-293. Micke, P, M. Ohshima, S. Tahmasebpoor, Z.-P. Ren, A. Ost- man, F. Ponten, and J. Botling. 2006. Biobanking of fresh frozen tissue: RNA is stable in nonfixed surgical specimens. Lab. Invest. 86:202-211. Murray, A., and T. Hunt. 1993. The cell cycle: an introduction, 251 p. Oxford Univ. Press, Inc., New York. O’Connell, C. P. 1980. Percentage of starving northern anchovy, Engrau- lis mordax, larvae in the sea as estimated by histologi- cal methods. Fish. Bull. 78:475-489. Olivar, M. P., M. V. Diaz, and M. A. Chicharo. 2009. Tissue effect on RNA:DNA ratios of marine fish larvae. Sci. Mar. 73S1: 171-182. Piwnicka, M., Z. Darzynkiewicz, and M. R. Melamed. 1983. RNA and DNA content of isolated cell nuclei mea- sured by multiparameter flow cytometry. Cytometry 3:269-275. Porter, S. M., and K. M. Bailey 2011. Assessing the condition of walleye pollock Ther- agra chalcogramma (Pallas) larvae using muscle-based flow cytometric cell cycle analysis. J. Exp. Mar. Biol. Ecol. 399:101-109. Porter, S. M., and G. H. Thei lacker. 1999. The development of the digestive tract and eye in larval walleye pollock, Theragra chalcogramma. Fish. Bull. 97:722-729. Ross, K. S., N. E. Haites, and K. F Kelly. 1990. Repeated freezing and thawing of peripheral blood and DNA in suspension: effects on DNA yield and integ- rity. J. Med. Genet. 27:569-570. Sirri, V., S. Urcuqui-Inchima, P. Roussel, and D. Hernandez- Verdun. 2008. Nucleolus: the fascinating body. Histochem. Cell Biol. 129:13-31. Staiano-Coico, L., Z. Darzynkiewicz, and C. K. McMahon. 1989. Cultured human keratinocytes: discrimination of different cell cycle compartments based upon measure- ment of nuclear RNA or total cellular RNA content. Cell Tissue Kinet. 22:235-243. Theilacker, G. H. 1978. Effect of starvation on the histological and mor- phological characteristics of jack mackerel, Trachurus symmetricus , larvae. Fish. Bull. 76:403-414. 1986. Starvation-induced mortality of young sea-caught jack mackerel, Trachurus symmetricus, determined with histological and morphological methods. Fish. Bull. 84:1-17. Theilacker, G. H., and S. M. Porter. 1995. Condition of larva! walleye pollock, Theragra chalcogramma, in the western Gulf of Alaska assessed with histological and shrinkage indices. Fish. Bull. 93:333-344. Theilacker, G. H., and W. Shen. 1993a. Fish larval condition analyzed using flow cytom- etry. In Physiological and biochemical aspects of fish development (B. T. Walther, and H. J. Fyhn, eds.), p. 346-355. Univ. Bergen, Bergen, Norway. 1993b. Calibrating starvation-induced stress in lar- val fish using flow cytometry. Am. Fish. Soc. Symp. 14:85-94. 2001. Evaluating growth of larval walleye pollock, Ther- agra chalcogramma, using cell cycle analysis. Mar. Biol. 135:897-907. Theilacker G. H., K. M. Bailey, M. F. Canino, and S. M. Porter. 1996. Variations in larval walleye pollock feeding and condition: a synthesis. Fish. Oceanogr. 5(suppl. 1 >: 1 12—123. 352 Habitat and diet overlap of 4 piscivorous fishes: variation on the inner continental shelf off New Jersey Email address for contact author: mark.wuenschel@noaa.gov 1 Marine Field Station Rutgers University 800 do 132 Great Bay Boulevard Tuckerton, New Jersey 08087-2004 Present address for contact author: Northeast Fisheries Science Center National Marine Fisheries Service, NOAA 166 Water St. Woods Hole, Massachusetts 02543 2 New Jersey Department of Environmental Protection Nacote Creek Research Station P.O. Box 418 Port Republic, New Jersey 08241 Deceased Abstract— Piscivorous fishes, many of which are economically valuable, play an important role in marine ecosystems and have the potential to affect fish and invertebrate popula- tions at lower trophic levels. There- fore, a quantitative understanding of the foraging ecology of piscivores is needed for ecosystem-based fishery management plans to be successful. Abundance and stomach contents of seasonally co-occurring piscivores were examined to determine overlap in resource use for Summer Floun- der (Paralichthys dentatus; 206-670 mm total length [TL] ), Weakfish (Cynoscion regalis\ 80-565 mm TL), Bluefish (Pomatomus saltatrix\ 55-732 mm fork length [FL] ), and Striped Bass ( Morone saxatilis; 422- 920 mm FL). We collected samples from monthly, fishery-independent trawl surveys conducted on the in- ner continental shelf (5-27 m) off New Jersey from June to October 2005. Fish abundances and overlaps in diet and habitat varied over this study period. A wide range of fish and invertebrate prey was consumed by each species. Diet composition (determined from 1997 stomachs with identifiable contents) varied with ontogeny (size) and indicated limited overlap between most of the species size classes examined. Al- though many prey categories were shared by the piscivores examined, different temporal and spatial pat- terns in habitat use seemed to alle- viate potential competition for prey. Nevertheless, the degree of overlap in both fish distributions and diets increased severalfold in the fall as species left estuaries and migrated across and along the study area. Therefore, the transitional period of fall migration, when fish densities are higher than at other times of the year, may be critical for unraveling resource overlap for these seasonally migrant predators. Manuscript submitted 24 October 2012. Manuscript accepted 26 August 2013. Fish. Bull. 111:352-369. doi: 10.7755/FB.111.4.5 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necesarily reflect the position of the National Marine Fisheries Service, NOAA. Mark J. Wuenschel (contact author)1 Kenneth W. Able1 James M. Vasslides1 Donald M. Byrne2* Predator species and their interac- tions with prey and other predator species play an important role in determination of the structure and function of ecosystems (Schmitz, 2007; Braga et al., 2012) — an espe- cially important concern because populations of many predators have declined in abundance (Myers and Worm, 2003; Heithaus et ah, 2008). In marine ecosystems, piscivorous fishes have the potential to affect fish and invertebrate populations at lower trophic levels. In some cases, direct removals of prey resources by piscivorous fishes have been shown to rival or even exceed the remov- als by commercial fisheries (Buckel et al., 1999c; Overholtz et ah, 2000; Overholtz and Link, 2007). There- fore, fish trophic ecology is relevant to several aspects of fisheries man- agement (Link, 2002). With the gen- eral move toward multispecies and ecosystem-based approaches to fish- eries management, there is a need for more comprehensive information on food web structure, interspecific trophic interactions, and predator movements (Andrews and Harvey, 2013). In temperate zones, many coastal marine fishes undergo large-scale seasonal and ontogenetic shifts in their spatial distribution. Examples from the east coast of the United States include species that migrate north to New England in summer and south to the Carolinas in winter (e.g., Striped Bass [Morone saxatilis ] and Bluefish [Pomatomus saltatrix ]) and species that move inshore in summer and offshore in winter (e.g., Summer Flounder [Paralichthys den- tatus] and Weakfish [Cynoscion rega- Z/s] ) (Able and Fahay, 2010). Further, many temperate, estuarine-depen- dent species in the Middle Atlantic Bight leave estuaries in the fall of their first year to avoid cold winter temperatures (Able and Fahay, 1998; 2010). Many of these young-of-the- year (YOY) fishes are piscivorous and their egress from estuaries to coastal waters acts to concentrate them in time and space with other Wuenschel et al.: Habitat and diet overlap of 4 piscivorous fishes 353 species or size classes, increasing the potential for both interspecific and intraspecific interactions. Interspe- cific competition and resource partitioning have been well documented for fish species in freshwater systems (Persson et ah, 1999; Sutton and Ney, 2002; Bellgraph et ah, 2008), where the potential for interactions may be greater than it is in marine systems given the closed nature of freshwater systems and fish populations. This interspecific competition and resource partition- ing may apply to some degree in estuaries as well, as has been reviewed for European estuaries (Elliot and Hemingway, 2002). In contrast, because of the openness of marine populations and the ability of individuals to move great distances, interspecific competition in most marine systems likely is highly variable in time and space, making it more difficult to document and study interspecific competition in marine systems than in freshwater populations. Summer Flounder, Weakfish, Bluefish, and Striped Bass are important commercial and recreational spe- cies in New Jersey and elsewhere on the east coast of the United States. These species co-occur seasonally and feed on similar prey, indicating potential for com- petitive interactions. However, studies of food habits for these species generally have focused on estuarine collections (Gartland et ah, 2006; Latour et ah, 2008) or have been limited to seasonal, offshore (at depths of 5-400 m) collections aggregated over multiple years (Buckel et al., 1999b, Garrison and Link, 2000a, 2000b; Link et ah, 2002; Overton et ah, 2008; Woodland et al., 2011). Further, most prior studies on these spe- cies typically have focused on a single (Gartland et ah, 2006; Latour et ah, 2008) or a pair of species (Buckel and McKown, 2002; Buckel et ah, 2009). Because of the spatial and temporal variability in competitive interactions between migratory fishes, studies span- Figure t (A) Study area where Summer Flounder (Paralichthys dentatus), Weakfish (Cynoscion regalis), Bluefish (Poma- tomus saltatrix), and Striped Bass ( Morone saxatilis ) were sampled in 2005 along the northeastern coast of the United States for a 5-month study of habitat and diet overlap of these 4 piscivorous fishes and (B) sample col- lection area off New Jersey with the 15 strata outlined (strata were defined on the basis of latitudinal boundar- ies and depth contours of 9, 18, and 27 m). In June, August, and October, all strata were sampled. In July and September, only strata indicated with diagonal lines were sampled. 354 Fishery Bulletin 1 1 1 (4) ning multiple years, such as many of the ones listed above, potentially blur or miss finer-scale interactions that occur over shorter intervals. Therefore, to evalu- ate potential interactions between mobile, seasonally migratory species, information on spatial distributions and food habits is needed at finer spatial and temporal scales (Rudershausen et ah, 2010). Although resource overlap among Summer Flounder, Weakfish, Bluefish, and Striped Bass has been indicated by or inferred in prior studies (Garrison and Link, 2000b), often over broad areas or time periods, a rigorous evaluation of these interactions at a more relevant ecological scale is lacking. The objectives for this study were to compare habi- tat use and food habits of 4 common predators of the Middle Atlantic Bight over the course of a 5-month period of co-occurrence in nearshore (inner continen- tal shelf) waters. The degree of overlap in both habi- tat and diet between different size classes of these 4 predators was quantified to determine resource overlap at a fine spatial and temporal scale. Materials and methods The distribution, abundance, and diet of Summer Flounder, Weakfish, Bluefish, and Striped Bass were evaluated from 20-min bottom trawls (30-m headrope, 6-mm codend liner [Byrne, 1994; Wuenschel et ah. 2012]) conducted in inner continental shelf waters of New Jersey in collaboration with the Bureau of Ma- rine Fisheries of the New Jersey Department of En- vironmental Protection. Through the use of a depth- stratified random sampling design, samples were taken during daylight hours at depths of 5 to 27 m along the New Jersey coast from the entrance of New York Harbor to the entrance of Delaware Bay (Fig. 1). The survey area was divided into 15 strata (Fig. 1) on the basis of latitudinal boundaries and depth contours (9, 18, and 27 m; Byrne, 1994). In June, August, and Oc- tober 2005, all depths and strata were sampled with 2 tows per strata, plus 1 additional tow in each of the 9 largest strata (39 tows, Table 1; see Byrne 1994 for details). In the intervening months (July and Septem- ber), sampling was undertaken with a bottom trawl net, bridles, and towing cables that were identical to the ones used in June, August, and October and were fished from the same vessel (RV Seawolf, SUNY Sto- nybrook) used during the other months, but because of constraints on vessel time, sampling was limited to 2 or 3 tows in nearshore and mid-shore depths for all but the northernmost and southernmost strata (18 tows, July; 12 tows, September; Table 1, Fig. 1). Representa- tive subsamples, with a mean of 8.4 (9.6 standard de- viation [SD]) per species per tow, of Summer Flounder, Weakfish, Bluefish, and Striped Bass were selected to cover the range of lengths of fish collected in a given tow for analysis of gut contents. Stomachs were re- Table 1 Summary of samples collected in 2005 off the coast of New Jersey for this study of habitat and diet overlap of 4 piscivorous fishes. Collection month and sampling effort ( numbers of tows in parentheses), numbers of fish collected, stomachs analyzed in the laboratory, and stomachs with prey in the gut for small, medium, and large Summer Flounder (Paralichthys dentatus ), PdS, PdM, PdL; small, medium, and large Weakfish ( Cynoscion regalis), CrS, CrM, CrL; small and large Bluefish ( Pomatomus saltatrix), PsS, PsL; and Striped Bass ( Morone saxatilis), Ms. PdS PdM PdL CrS CrM CrL PsS PsL Ms June (39) Fish collected 85 170 80 1 774 1 9 4 7 Stomachs analyzed 47 63 35 0 46 1 0 6 0 Stomachs with prey 30 42 26 0 42 1 0 4 0 July (18) Fish collected 134 296 61 64 6319 0 48 3 0 Stomachs analyzed 72 153 51 0 125 0 21 3 0 Stomachs with prey 39 74 12 0 116 0 18 3 0 August (39) Fish collected 302 950 148 2587 4232 31 890 23 214 Stomachs analyzed 71 245 108 100 109 7 150 20 27 Stomachs with prey 64 178 62 99 84 5 105 17 11 September (12) Fish collected 29 783 116 17,836 11,752 189 2370 6 0 Stomachs analyzed 4 117 22 134 58 5 185 5 0 Stomachs with prey 3 74 8 96 45 5 147 2 0 October (39) Fish collected 12 192 31 9746 10,275 1786 616 70 123 Stomachs analyzed 6 134 26 172 93 86 156 67 89 Stomachs with prey 3 79 13 130 75 66 130 52 37 All months ( 147) Fish collected 562 2391 436 30,234 33,352 2007 3927 112 344 Stomachs analyzed 200 712 242 406 431 99 512 101 116 Stomachs with prey 139 447 121 325 362 77 400 78 48 Wuenschel et al.: Habitat and diet overlap of 4 piscivorous fishes 355 moved immediately after capture and preserved in for- malin for laboratory analysis. In some months, tagging and releasing Striped Bass was a higher priority than determining stomach contents; therefore, all fishes cap- tured were not available for diet analysis. To account for size-related changes in habitat use (Able and Fahay, 2010), diet composition within species (Garrison and Link, 2000b), and interactions across species (Buckel and McKown, 2002), species were split into multiple size classes when data permitted: small (Summer Flounder: 200-300 mm total length [TL] ; Weakfish: 80-200 mm TL; Bluefish: 55-300 mm fork length [FL]), medium (Summer Flounder: 301-400 mm TL; Weakfish: 201-350 mm TL), and large (Summer Flounder: 401-670 mm TL; Weakfish: 351-565 mm TL; Bluefish: 301-732 mm FL). For Striped Bass, a single size class was used because of limited sample sizes, the absence of prior evidence for ontogenetic shifts beyond the YOY stage (Walter et al., 2003), and the relatively large sizes of our specimens (422-920 mm FL). Diet analysis In the laboratory, preserved stomachs were carefully opened and the contents transferred to a solution of rose bengal stain and 95% ethyl alcohol. Prey items were identified to the lowest practical taxonomic level by using available keys and guides for the Mid-Atlantic region (Weiss, 1995; Able and Fahay, 1998) and enu- merated. For each stomach, abundant or large prey types were sorted and placed on preweighed filter pa- pers or aluminum weighing pans and dried to a con- stant weight (+0.0001 g) in a drying oven (70°C). Dry weights were chosen because they are more representa- tive of nutritional value and have less weighing error than wet weights (Hyslop, 1980), especially for small or partial prey (Carr and Adams, 1972). For small and, therefore, hard-to-separate prey items (e.g., copepods and mysids), an aggregate sample was dried and the percent contribution by volume of different prey types was recorded and later converted to weights. Through the use of this protocol, prey-specific dry weights were obtained directly for larger prey or estimated from ag- gregate samples of smaller, mixed prey items for each stomach analyzed. Trawl collections yielded “clusters” of individuals within species and size classes per location; therefore, the percent contribution by weight of prey items was calculated with the following cluster sampling estima- tor (Buckel et al., 1999a; 1999b; Gartland et al., 2006). For a given size class of a predator, the percent contri- bution by weight of each prey type k (%W^) to the diet was calculated with the following equation: £ MMik %wk=^ 100, (1) I Mi i=l where q., = — — ; lk w; n = the number of trawls; Mx = the number of species size class sampled per tow i; W{ = the total dry weight of all prey in stomachs for that species size class in tow /; and Wfe = the total dry weight of prey type k in all stomachs for that species size class collected in tow i. To facilitate analysis, prey items were grouped into the following general categories: squids (predominantly Loligo spp.), decapod crustaceans (including swimming crabs, sand crabs, rock crabs [Cancer borealis and C. ir- roratus], spider crabs, hermit crabs, decapod zoea, and shrimps [predominately Crangon septemspinosa and Palaemonetes sp.]), nondecapod crustaceans (including amphipods, isopods, cumaceans, mysids, and mantis shrimp), bivalves (clams and periwinkles), fishes (44 species identified), worms and wormlike organisms (nematodes, polychaetes, annelids, and leeches), and other unidentified (UID) items (inorganic matter, or- ganic matter, eggs, and insects). In addition, prey items (species or higher taxa) that contributed on average >5% by weight to the overall mean diet of a species size class were included as additional prey categories. Therefore, if Bay Anchovy (. Anchoa mitchilli) composed >5% of the diet for a given species size class in any month, it was included as a prey category and the cat- egory “fishes” represented all remaining fish prey that contributed <5% to the diet of that species size. The cumulative trophic diversity was calculated for each species size class to determine whether the sample sizes that were analyzed were sufficient to describe the diet of a given species size class in each month (Ferry and Cailliet, 1996; Cortes, 1997; Braccini et al., 2005; Belleggia et al., 2008). The Shannon-Wiener index ( H '), which describes entropy on the basis of information theory, was calculated as each stomach that contained prey was added to the analysis for 100 randomizations of the data for each species size class: H' = -'L(pi)-(Xogcpi), (2) i=l where S - the number of prey categories; and Pi = the proportion of the cumulative (total) sam- ple (gut contents) represented by the ith prey category. Following Jost (2006), H' was converted to effective number of species (exp(iL')), a true diversity. Only groups (monthly species size classes) with mean tro- phic diversity curves that appeared asymptotic or with >40 sampled guts were included in the similarity anal- ysis (described in the next paragraph). To evaluate the degree of similarity in diets between species and size classes, nonmetric multidimensional scaling (nMDS) and hierarchical clustering were used 356 Fishery Bulletin 111(4) within each month and across all months (PRIMER-E, Ltd., Plymouth, UK1). The nMDS data were calculated as the percentage of diet by weight for each month and each species-size-class combination and were log-trans- formed before use in the Bray-Curtis index to construct the sample similarity matrix. Group-average hierarchi- cal clustering was then used to identify those predators that had dietary similarities at the 60% level following Jaworski and Ragnarsson (2006) and Clarke and War- wick (2001). Habitat and diet overlap Habitat and diet overlap between pairs of species size classes were determined through the use of Schoener’s index (Schoener 1970). This index was calculated with this equation: a = l-0.5(x|py-pi^|)» (3) i=l which shows the overlap (a), where pij and pik are the proportions of the ith resource (trawl station or prey proportion) used by species j and k, respectively. Index values range from 0 to 1, with values >0.6 representing biologically important overlap in resource use (Wallace, 1981; Buckel and McKown, 2002; Bethea et al., 2004). Results Spatial distribution, abundance, and sizes The spatial distribution, abundance, and size distribu- tion for each of the 4 predators were variable over the course of our 5-month study (Table 1, Figs. 2 and 3). Summer Flounder were the most consistently collected species throughout this study. They were distributed throughout our study area, with a slight shift in abun- dance from offshore to inshore in summer followed by the reverse in fall. The size distribution of Sum- mer Flounder was relatively constant, with individu- als of 206-670 mm TL representing a broad range of age classes (YOY to 4+) collected from June to Octo- ber. Weakfish (80-565 mm TL) also were collected con- sistently, with greater abundances occurring inshore. Catches of Weakfish in June and July were dominated by larger size classes (>200 mm TL), with smaller size classes (YOY or 1+, <200 mm TL) becoming abundant from August to October. Similarly, Bluefish (55-732 mm FL) were most abundant inshore. They were dominat- ed by larger size classes (>300 mm FL) in June, with smaller size classes (<300 mm FL) becoming abundant inshore from July to October. Striped Bass, which were typically larger (422-920 mm FL) than the other 3 species (Fig. 3), were less abundant and highly variable in time and space. In 1 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. June and August, collections of Striped Bass were lim- ited to the northernmost strata of the sample area. No Striped Bass were collected in July and September, likely because the northern strata were not sampled during those months. In October, Striped Bass were distributed throughout our study area. In all months in which they were collected, they were typically inshore. Diet analysis Because of limited or zero abundance of some species size classes in some months (described previously), sampling limitations, and the occurrence of empty stomachs, adequate food habit information was not ob- tained over all species size classes or months. The rela- tive contributions of prey types to the diets for species size classes in each month are summarized in Table 2 (species size classes with insufficient samples sizes to be included in subsequent analyses are presented). For the groups that were considered adequately sampled for diet description (Fig. 4) and, therefore, included in the cluster analysis, the cumulative trophic diversity curves indicated that sample sizes of 30-40 guts were sufficient to characterize the diet in most cases. How- ever, for some monthly species size classes (e.g., me- dium Weakfish in August and medium Summer Floun- der in October) trophic diversity continued to increase beyond 40 guts analyzed. Size-specific patterns in trophic diversity differed across species, with Summer Flounder showing de- creased diversity with size in June. In contrast, larg- er size classes of Weakfish in August and Bluefish in October had more diverse diets than did smaller size classes of these species. Overall, diversity of prey items increased throughout time in Summer Flounder and Striped Bass, and it remained low for small Blue- fish. Weakfish diet diversity also was relatively stable through the 5-month period of this study, with the ex- ception of the high trophic diversity of the medium size class of this species in August. Cluster analysis separated the size classes for each of the 4 species into 3 groups in June at the 60% simi- larity level (Fig. 5). The first group consisted of large (>400 mm TL) Summer Flounder, which preyed pre- dominantly on squids (88.1%) (Table 2). Small and medium Summer Flounder formed the second group, and medium (201-350 mm TL) Weakfish the third. Al- though the second and third groups consumed mostly fishes (0-73.3% sand lances [Ammodytes spp.j and 6.1- 52.9% UID and other fishes), they were separated by the amounts of decapod crustaceans (2.1-20.9%) and squids (8.3-21.5%) in the former and mysids (33.4%) in the latter. In July, the cluster analysis identified 3 groups at the 60% similarity level from among the 4 species size classes for which enough data were available. Small and medium Summer Flounder were grouped together, with diets consisting of both pelagic and benthic prey, including Butterfish ( Peprilus triacanthus ; 15.8-18.7%), Wuenschel et al.: Habitat and diet overlap of 4 piscivorous fishes 357 Summer Flounder Weakfish Bluefish Striped Bass Figure 2 Distributions of Summer Flounder ( Paralichthys dentatus), Weakfish ( Cynoscion regalis), Bluefish (Pomatomus saltatrix), and Striped Bass ( Morone saxatilis ) sampled in June, July, August, September, and October of 2005 along the coast of New Jersey for this study. Circles are shaded by species and represent abundance (log transformed), with the same scale in all frames, and crosses indicate zero catches. See Table 1 for a summary of numbers caught for each species. 358 Fishery Bulletin 111(4) Summer Flounder Weakfish Bluefish Striped Bass Total length (mm) Total length (mm) Fork length (mm) 15 0 SO r L T 25 0 .) — — r — 20 r Hi _ 55 r , Fork length (mm) Figure 3 Length distributions (percent frequency) for Summer Flounder ( Paralichthys dentatus ), Weakfish (Cynoscion regalis), Blue- fish (Pomatomus saltatrix), and Striped Bass (Morone saxatilis) collected from June (top row) to October (bottom row) in 2005 off the coast of New Jersey for this study. Bars are shaded by species (as in Fig. 2). Vertical, dashed gray lines indicate breakpoints between size classes. Note that the j:-axis scales are different across species and the y-axis scales are different across months within species. Wuenschel et at: Habitat and diet overlap of 4 piscivorous fishes 359 Table 2 Stomach contents of small, medium, and large Summer Flounder (Paralichthys dentatus), PdS, PdM, PdL; small, medium, and large Weakfish ( Cynoscion regalis), CrS, CrM, CrL; small and large Bluefish ( Pomatomus saltatrix ), PsS, PsL; Striped Bass ( Mo- rone saxatilis), Ms, for this study of habitat and diet overlap of these 4 piscivorous fishes. All samples were collected in 2005 off the coast of New Jersey. Diet is summarized by month and for all months. %W ^ is the proportion of identifiable prey to the diet by weight. Note that proportions from cluster estimators do not add up to 100% in all cases (see main text for calculation). An asterisk (*) indicates groups with insufficient sample sizes for inclusion in the analysis with nonmetric multidimensional scaling. UID=unidentified. CO o 8 CO # £ .co 09 Co * 0); however, a was below the biologically important level of 0.6 in most months (Fig. 6). Some groups had substantial overlap ( >0.6) in habitat but not diet (e.g., large, medium, and small Summer Flounder). There also were instances of overlap ( >0.6) in diet but not habitat (e.g., small Weakfish versus medium Weakfish and small Sum- mer Flounder). There were no cases where substantial overlap ( >0.6) in both habitat and diet occurred at the same time. For many of the pairwise comparisons, over- laps in both habitat and diet increased severalfold dur- ing the fall in this 5-month study, from levels of ~0.2 in Wuenschel et al.: Habitat and diet overlap of 4 piscivorous fishes 363 PdL 0 "D "O c cn _q CQ CL 03 s _c CO Ms J J ASO J JASOJ JASOJ JASOJ J A S O J JASOJ J ASOJ JA Month of year Figure 6 Pairwise monthly overlap (a), from Schoener’s index, in habitat (open circles) and diet (triangles) between species size classes sampled from June to October in 2005 off the coast of New Jersey for this study. Habitat overlap is shown for all cases when both groups were captured in the same month. Diet overlap is shown only for groups with sufficient sample sizes (see main text). Hori- zontal dashed lines indicate biologically important overlap threshold (0.6). Species size classes are indicated in this manner: small, medium, and large Summer Flounder ( Paralichthys dentatus ), PdS, PdM, PdL; small, medium, and large Weakfish ( Cynoscion regalis), CrS, CrM, CrL; small and large Bluefish (Pomatomus saltatrix), PsS, PsL; and Striped Bass ( Morone saxatilis), Ms. June and July to levels >0.4 in September and October. Small and medium Weakfish had relatively high over- lap in diet, but lower overlap in habitat, with small and medium Summer Flounder from August to Octo- ber. The limited spatial distribution of Striped Bass, in addition to their absence during some months, resulted in a low degree of overlap in habitat with groups of other species, but Striped Bass were most abundant in the fall and, therefore, overlapped with the other predators at that time. Discussion The degree of habitat and diet overlap among Summer Flounder, Weakfish, Bluefish, and Striped Bass varied with size and season. The size classes of these 4 preda- tors examined in our study exhibited different patterns in spatial distribution, depth, and habitat use. Sum- mer Flounder were most evenly distributed throughout our study area — a pattern also observed in a longer (1982-2003) time series of sampling at generally deep- 364 Fishery Bulletin 111(4) er depths (from 27 to >193 m) (Able and Fahay, 2010), with similar sizes present during our 5-month study on the inner continental shelf. Weakfish in our study occurred primarily inshore in shallow strata. Their abundance increased through the summer because of the appearance of YOY fish, and they dominated the catch by October — a trend also observed in composite collections over time on the deeper continental shelf (Able and Fahay, 2010). The size structure of Bluefish was similarly variable with multiple YOY cohorts appearing during summer and eventually dominating the proportion of catch by October. Bluefish were concentrated at inshore, shallow stations, as had been documented previously (Able et ah, 2003; Wiedenmann and Essington, 2006; Wuenschel et al., 2012). In October, the distribution of Bluefish was more uniform with depth, and they appeared to utilize a greater portion of the inner continental shelf during the period of southward migration, as was evi- dent from other composite sampling at similar and deeper depths (Able and Fahay, 2010). Bluefish and Weakfish distributions indicated a high degree of similarity in their use of the inshore shelf during the summer months. In contrast, the distribu- tion of Striped Bass was limited to the northernmost stations during summer. Like Bluefish, Striped Bass were more abundant throughout the study area during their fall migration (October). The limited collections of Striped Bass, coupled with a priority to tag and release them during certain months, restricted our ability to describe Striped Bass diets. Nevertheless, Striped Bass showed little overlap in habitat use with the other predators studied. The species-level distributions and monthly size fre- quencies tended to overestimate overlap because dif- ferent sizes may have occupied different locations in a given month. When viewed at the size-class level, spatial (habitat) overlap was rarely above 0.6. Simi- larly, assessment of diet overlap at the species level (i.e. , combining size classes and ignoring ontogenetic shifts in diet) would likely increase perceived overlap. For example, large and medium Summer Flounder diet overlap was moderate (a -0.2-0. 5) and overlap between medium and small sizes was similarly vari- able (a -0.3-0. 6), but overlap between large and small Summer Flounder was much lower (a -0.2-0. 3). This result supports the interpretation of gradual onto- genetic changes in diet and some similarity between adjacent size classes. Use of a single size group for Striped Bass because of limited sample sizes prevented the exploration of ontogenetic shifts for this species, which were significant in one study (Smith and Link, 2010) but not another (Walter et al., 2003). Regardless of potential ontogenetic differences, their extensive use of estuaries in New Jersey and throughout their range during many seasons (Able and Fahay, 2010) probably also contributes to the limited co-occurrence of this species with the other species size classes studied. Unidentified fish remains were large components of the diet for many of the species size classes in our study — a common problem encountered in the analysis of stomach contents (Garrison and Link, 2000a; 2000b). The presence of large portions of UID fishes serves to increase the overlap in diet and decrease the num- ber of distinct clusters from the groups analyzed. As- suming that identifiable and unidentifiable prey oc- cur in similar relative proportions, we consider the separation of distinct groupings in our analysis to be conservative. Where diets between species size classes showed little overlap, we can conclude diets were in- deed different. The diets of Summer Flounder, Weakfish, Bluefish, and Striped Bass on the inner continental shelf of New Jersey during summer varied month to month — an observation that could not be obtained from previ- ous studies conducted in the region across a greater portion of the shelf region over multiple years (Gar- rison and Link, 2000b; Link et ah, 2002; Walter et al., 2003; Buckel et al., 2009; Smith and Link, 2010). The finer scale in our study accounts for the slightly dif- ferent patterns of similarity among Summer Flounder, Weakfish, Bluefish, and Striped Bass diets in compari- son with results that had been reported previously for the shelf ecosystem of the northeastern United States (Garrison and Link, 2000b). Although the size classes differed slightly, this previous study reported Summer Flounder (21-70 cm) and Bluefish (>31 cm) in a pisci- vore guild distinct from small Bluefish (10-30 cm) and Weakfish (10-50 cm). On a much finer spatial scale, we observed greater overlap in diets, which varied through time, between Summer Flounder and Weakfish, particularly for the small and medium size classes. This similarity was driven by large amounts of mysids and other crusta- ceans in the diet. Both Summer Flounder and Weakfish incorporated more fishes and squids in their diet as they increased in size. The early onset of piscivory by Bluefish separated them from the smaller size classes of Summer Flounder and Weakfish. Together, these dif- ferences relative to earlier studies point out the ad- vantages of the finer temporal, spatial, and size-class scales used in this study. Of the 4 predators analyzed from the Mid-Atlantic region in our study, only Weakfish were consumed in appreciable numbers by the other species. Weakfish were consumed by small Summer Flounder (16.5%) in July and small Bluefish (9.5%) in August, and an appre- ciable amount of cannibalism by medium-size Weakfish (23.5%) was observed in August. In October, large Blue- fish and Summer Flounder fed extensively on Weakfish (25.2-26.7%), with medium and small Summer Floun- der and Striped Bass also consuming Weakfish but in lesser proportions. Similarly, in Long Island bays, YOY Bluefish and Summer Flounder consumed Weakfish (7.5% and 3.0% by weight, respectively); however, data were reported across sizes and seasons (spring, sum- mer, and fall), and Weakfish diet was not investigated Wuenschel et al : Habitat and diet overlap of 4 piscivorous fishes 365 (Sagarese et al., 2011). Predation on juvenile Weak- fish has also been observed in Delaware Bay by adult Weakfish and Summer Flounder (Taylor, 1987). A recent synopsis of diets documented across the continental shelf in the region indicated that Weakfish occurred at relatively low levels (~5%) in diets of Blue- fish, Summer Flounder, and Weakfish (Smith and Link, 2010). Although direct predation on Weakfish by the 4 piscivores investigated was high at different times during our study, large numbers of Weakfish were also collected in our surveys, indicating high availability of this prey type. However, the possibility that the high degree of predation may influence the continuing low- population levels of Weakfish (NEFSC2) needs further study, especially because some of these predators, such as Summer Flounder (Able et al., 2011) and Striped Bass, have reached high population levels in recent years (ASMFC3). Another prey species, Atlantic Menhaden, has re- ceived increased attention because of its historical im- portance as the main prey of Striped Bass at other lo- cations during other times (Nelson et al., 2003; Walter et al., 2003; Overton et al., 2008). Uphoff (2003) sug- gested that a shortage of Atlantic Menhaden as prey for Striped Bass in the Chesapeake Bay may have re- duced the nutritional condition of those Striped Bass in the 1990s, making them susceptible to disease. In our study, there was little evidence of Atlantic Menhaden in the diets of any of these predators except in October when Striped Bass, Weakfish, and Bluefish consumed them to some degree, although Atlantic Menhaden are typically present in the area during the other months (Ahrenholz, 1991; Smith, 1999). Adult Atlantic Menha- den occur along the coast in summer (Smith, 1999), but YOY Atlantic Menhaden reside in estuaries in summer and are not plentiful in the ocean until October (Able and Fahay, 2010), when they appeared in the diets of the piscivores examined in our study. Adult Atlantic Menhaden exceed the gape limitation for most of the species size classes examined, except that for Striped Bass and large Bluefish; therefore, it is not surprising that little consumption of this species was documented in summer. Diet analyses for larger predator species and larger individuals of those species considered in our study are needed to fully evaluate consumption of large Atlantic Menhaden in the Middle Atlantic Bight. As with our study, Woodland and Secor (2011) re- ported no clupeids in the diet of YOY Bluefish collected 2 NEFSC (Northeast Fisheries Science Center). 2009. 48th Northeast Regional Stock Assessment Workshop (48th SAW) Assessment Report. U.S. Dept. Commer., Northeast Fish. Sci. Cent. Ref. Doc. 09-15, 834 p. [Available from National Marine Fisheries Service, 166 Water Street, Woods Hole, MA 02543-1026 or http://www.nefsc.noaa.gov/nefsc/publications/.] 3 ASMFC (Atlantic States Marine Fisheries Commission). 2011. Striped Bass Stock Assessment Update 2011, 207 p. Prepared by the Striped Bass Stock Assessment Subcom- mittee and Striped Bass Tagging Subcommittee. [Available from http://www.asmfc.org/speciesDocuments/stripedBass/re- ports/stockassmts/20 HStripedBassAssmtUpdate.pdf] on the inner continental shelf off Maryland in August, although clupeids made up 9% of the diet of YOY Blue- fish collected in Chesapeake Bay in the same month. However, Atlantic Menhaden contributed a large por- tion to Bluefish diets, but less so for Summer Floun- der in Long Island bays (Sagarese et al., 2011), and large proportions to the diets of Weakfish, Bluefish, and Sandbar Shark ( Carcharhinus plumbeus) in Delaware Bay (Taylor, 1987), underscoring the importance of At- lantic Menhaden as prey in estuaries during summer. Additionally, we acknowledge that the degree of con- sumption and overlap in diets may vary with the popu- lation size of the predators. At the time of our study, the populations of Striped Bass (ASMFC3) and Sum- mer Flounder (Able et al., 2011; Terceiro, 2011) were relatively high and the populations of Bluefish (Shep- herd and Nieland4) and Weakfish (NEFSC2) were rela- tively low. More detailed estimates of population-level consumption with more diet information from larger size classes would help to determine the direct effect of the predation of these piscivores on Weakfish and Atlantic Menhaden populations. One behavioral attribute that may reduce resource overlap between these predators is their use of the wa- ter column. Recent studies have indicated that a pre- sumed benthic species, Summer Flounder, may spend considerable time in the water column (Yergey, 2011; Henderson, 2012). This observation is consistent with the surprisingly large proportion of pelagic prey in their diets and their ability to feed in the water column in the laboratory (Olla et al., 1972). In addition, aggre- gation of diet data across the 24 h cycle may obscure some interactions with prey that undergo diel migra- tions (e.g., mysids and Summer Flounder; Buchheister and Latour, 2011). Although sampling was limited to daylight hours, given that gut evacuation rate gener- ally decreases with fish size (Wuenschel and Werner, 2004) and with the relatively slow passage of food in carnivores (Smith, 1989; Adams and Breck, 1990), the results of our study capture much of the nighttime feeding habits of the predators examined. Another caveat to consider when evaluating overlap in habitat between these species is that overlap may be confounded by the length of a tow with a relatively large net. Although the duration of tows was short in our study, the distance covered (-1.85 km) may have included multiple discrete habitats, inflating assess- ments of overlap. However, the overlap estimates were not systematically high and the area covered per tow was very small relative to the overall study area and to the distances that the species examined are known to move; therefore, analysis at the tow level is appropriate. This study has described in fine spatial and tem- poral detail the degree of overlap in habitat and diet 4 Shepherd, G. R., and J. Nieland. 2010. Bluefish 2010 stock assessment update. U.S. Dept. Commer., Northeast Fish. Sci. Cent. Ref. Doc. 10-15, 33 p. [Available from National Marine Fisheries Service, 166 Water Street, Woods Hole, MA 02543-1026 or http://www.nefsc.noaa.gov/nefsc/publications/.! 366 Fishery Bulletin 111(4) of 4 piscivores on the inner continental shelf off New Jersey from summer through fall. It does not consider these parameters on adjacent estuaries where these species can also be abundant (Able and Fahay, 2010). For example, Striped Bass are known to use estuar- ies in the region during the summer (Tupper and Able, 2000; Able and Grothues, 2007; Ferry and Mather, 2012). The same is true for Bluefish (Grothues and Able, 2007; Sagarese et al., 2011), Summer Flounder (Sackett et al., 2007; 2008; Sagarese et al., 2011), and Weakfish (Taylor, 1987; Turnure 2010). Therefore, our findings from the inner shelf should be viewed with this qualification in mind. As an example, the diets de- scribed here for inshore waters revealed differences in diets, compared with diets observed in studies carried out within estuaries (Gartland et al., 2006; Latour et al., 2008; Sagarese et al., 2011). Although not necessarily unexpected given differ- ences in prey availability between estuarine and ocean habitats, these differences in diets underscore the need to incorporate into ecosystem models both distribution and diet information from the inner shelf and adja- cent estuaries because of connectivity between them, especially over seasonal scales (Able, 2005). These data have typically been lacking. Many prey categories were shared by the piscivores examined in our study on the inner continental shelf, but different patterns in habitat use (in time and space) seemed to allevi- ate overlap between these predators for prey. However, overlaps in fish distributions and diets increased dur- ing fall as species left estuaries and underwent coastal migrations. Therefore, the information on spatial dis- tributions and food habits reported here at finer spatial and temporal scales, compared to results from previous studies, complements and provides a critical link be- tween what is known about these mobile, seasonally migratory predators within estuaries and about these same piscivores across larger geographic scales (e.g., the northeastern shelf of the United States). It is rea- sonable to ask whether diet overlap may continue to increase during winter when many of these predators are concentrated farther south and likely co-occur in thermal refuges in the South Atlantic Bight. Conclusions Results from this study, at finer temporal, spatial, and size-class scales than have been attempted typically in other studies, indicate that overlap in distribution on the inner continental shelf for the 4 piscivores ex- amined was not uniform through time and that they had moderate levels of overlap in diet. The exception was observed in the fall, when many of these species became concentrated as both larger individuals and smaller YOY fishes left estuaries, gathered with other individuals on the inner shelf, and began their south- ward migration as temperatures cooled. Given the high seasonal variability in water temperature and produc- tivity in the system studied and the migratory nature of the species investigated, it is not surprising that habitat use and species interactions were variable. Our understanding of species interactions on the inner continental shelf has been limited, in the past, by the gap between sampling programs within estuar- ies and bays and programs occurring farther offshore. The intensive monthly sampling of this study on the inner continental shelf revealed the dynamic nature of habitat and diet overlap for Summer Flounder, Weak- fish, Bluefish, and Striped Bass. The use of inner shelf resources by these 4 important species had previously been poorly defined. The limited degree of resource overlap in summer and the increasing overlap in fall for the 4 piscivores indicate that this period of change- able overlap may be important for the population dy- namics of these species and that information about it should be incorporated into not only species-specific models of population dynamics but also broader eco- system-level models. | Acknowledgments This study was funded through a grant from the col- laborative Bluefish/Striped Bass Dynamics Research Program of Rutgers University and the National Ma- rine Fisheries Service. The authors thank New Jersey Department of Environmental Protection personnel (L. Barry, A. Mazzarella and S. Reap), the staff at the Rut- gers University Marine Field Station and Captain S. Cluett and crew of the RV Seawolf for assistance. The following individuals provided field and laboratory as- sistance: R. Nichols, M. Greaney, J. Conwell, J. Lamo- naca, J. Eppenstiener, and J. Bunkiewicz. B. Smith, S. Rowe, and anonymous reviewers provided constructive comments on the manuscript. We are grateful to all of the above. This article is contribution no. 2013-3 from the Institute of Marine and Coastal Sciences, Rutgers University. Literature cited Able, K. W. 2005. A reexamination of fish estuarine dependence: evidence for connectivity between estuarine and ocean habitats. Estuar. Coast. Shelf Sci. 64:5-17 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. 2010. Ecology of estuarine fishes: temperate waters of the western North Atlantic, 584 p. Johns Hopkins Univ. Press, Baltimore. Able, K. W., and T. M. Grothues. 2007. Diversity of estuarine movements of striped bass ( Morone saxatilis ): a synoptic examination of an es- tuarine system in southern New Jersey. Fish. Bull. 105:426-435. Wuenschel et al.: Habitat and diet overlap of 4 piscivorous fishes 367 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 York Bight. Fish. Bull. 101:201-214. Able, K. W., M. C. Sullivan, J. A. Hare, G. Bath-Martin, J. C. Taylor, and R. Hagan. 2011. Larval abundance of summer flounder ( Paralich - thys dentatus) as a measure of recruitment and stock status. Fish. Bull. 109:68-78. Adams, S. M., and J. E. Breck. 1990. Bioenergetics. In Methods for fish biology (C. B. Schreck, and P. B. Moyle, eds.), p. 389-415. Am. Fish. Soc., Bethesda, MD. Ahrenholz, D. W. 1991. Population biology and life history of the North American menhadens, Brevoortia spp. Mar. Fish. Rev. 53:3-19. Andrews, K. S., and C. J. Harvey. 2013. Ecosystem-level consequences of movement: sea- sonal variation in the trophic impact of a top predator. Mar. Ecol. Prog. Ser. 473:247-260. Belleggia, M., E. Mabragana, D. E. Figueroa, L. B. Scenna, S. A. Barbini, and J. M. Diaz de Astarloa. 2008. Food habits of the broadnose skate, Bathyraja brachyurops (Chondrichthyes, Rajidae), in the south- west Atlantic. Sci. Mar. 72:701-710. Bellgraph, B. J., C. S. Guy, W. M. Gardner, and S. A. Leathe. 2008. Competition between saugers and walleyes in non- native sympatry. Trans. Am. Fish. Soc. 137:790-800. Bethea, D. M., J. A. Buckel, and J. K. Carlson. 2004. Foraging ecology of the early life stages of four sympatric shark species. Mar. Ecol. Prog. Ser. 268:245-264. Braga, R. R.. H. Bornatowski, and J. R. S. Vitule. 2012. Feeding ecology of fishes: an overview of world- wide publications. Rev. Fish Biol. Fish. 22:915-929. Braccini, J. M., B. M. Gillanders, and T. I. Walker. 2005. Sources of variation in the feeding ecology of the piked spurdog (Squalus megalops ): implications for in- ferring predator-prey interactions from overall dietary composition. ICES J. Mar. Sci. 62:1076-1094. Buchheister, A., and R. J. Latour. 2011. Trophic ecology of summer flounder in lower Chesa- peake Bay inferred from stomach content and stable isotope analyses. Trans. Am. Fish. Soc. 140:1240-1254. Buckel, J. A., and K. A. McKown. 2002. Competition between juvenile striped bass and bluefish: resource partitioning and growth rate. Mar. Ecol. Prog. Ser. 234:191-204. Buckel, J. A., D. O. Conover, N. D. Steinberg, and K. A. McKown. 1999a. Impact of age-0 bluefish (Pomatomus saltatrix) predation on age-0 fishes in the Hudson River estu- ary: evidence for density-dependent loss of juvenile striped bass (Morone saxatilis). Can. J. Fish. Aquat. Sci. 56:275-287. Buckel, J. A., M. J. Fogarty, and D. O. Conover. 1999b. Foraging habits of bluefish, Pomatomus saltatrix, on the U.S. east coast continental shelf. Fish. Bull. 97:758-775. 1999c. Mutual prey of fish and humans: a comparison of biomass consumed by bluefish, Pomatomus salta- trix, with that harvested by fisheries. Fish. Bull. 97:776-785. Buckel, J. A., J. P. Pessutti, J. E. Rosendale, and J. S. Link. 2009. Interactions between bluefish and striped bass: behavior of bluefish under size- and number-impaired conditions and overlap in resource use. J. Exp. Mar. Biol. Ecol. 368:129-137. Byrne, D. 1994. Stock assessment of New Jersey’s nearshore recre- ational fisheries resources. In Proceedings of the work- shop on the collection and use of trawl survey data for fisheries management; Charleston, SC, 15-17 August (T. Berger, ed.), p. 36-42. Atlantic States Mar. Fish. Comm., Washington, D.C. Carr, W. E. S., and C. A. Adams. 1972. Food habits of juvenile marine fishes: evidence of the cleaning habit in the leatherjacket, Oligoplites sau- rus, and the spottail pinfish, Diplodus holbrooki. Fish. Bull. 70:1111-1120. Clarke, K. R., and R. M. Warwick. 2001. Change in marine communities: an approach to statistical analysis and interpretation, 2nd ed., 172 p. PRIMER-E Ltd., Plymouth, UK. Cortes, E. 1997. A critical review of methods of studying fish feed- ing based on analysis of stomach contents: applica- tion to elasmobranch fishes. Can. J. Fish. Aquat. Sci. 54:726-738. Elliot, M., and K. L. Hemingway (eds.). 2002. Fishes in estuaries, 636 p. Blackwell Science Ltd., Oxford, UK. Ferry, L. A., and G. M. Cailliet. 1996. Sample size and data analysis: are we charac- terizing and comparing diet properly? In “Gutshop ’96:” feeding ecology and nutrition in fish symposium proceedings. International Congress on the Biology of Fishes; San Francisco, 14-18 July (D. MacKinley and K. Shearer, eds.), p. 71-80. Am. Fish. Soc., Bethesda, MD. Ferry, K. H., and M. E. Mather. 2012. Spatial and temporal diet patterns of subadult and small adult striped bass in Massachusetts estuar- ies: data, a synthesis, and trends across scales. Mar. Coast. Fish. 4:30-45. Garrison, L. P, and J. S. Link. 2000a. Diets of five hake species in the northeast United States continental shelf ecosystem. Mar. Ecol. Prog. Ser. 204:243-255. 2000b. Dietary guild structure of the fish community in the Northeast United States continental shelf ecosys- tem. Mar. Ecol. Prog. Ser. 202:231-240. Gartland, J., R. J. Latour, A. D. Halvorson, and H. M. Austin. 2006. Diet composition of young-of-the-year bluefish in the lower Chesapeake Bay and the coastal ocean of Vir- ginia. Trans. Am. Fish. Soc. 135:371-378. Grothues, T. M., and K. W. Able. 2007. Scaling acoustic telemetry of bluefish in an estua- rine observatory: detection and habitat use patterns. Trans. Am. Fish. Soc. 136:1511-1519. Heithaus, M. R., A. Frid, A. J. Wirsing, and B. Worm. 2008. Predicting ecological consequences of marine top predator declines. Trends Ecol. Evol. 23:202-210. Henderson, M. J. 2012. Movements, growth, and mortality of Chesapeake Bay summer flounder based on multiple tagging tech- nologies. Ph.D. diss., 262 p. College of William & Mary, Williamsburg, VA. 368 Fishery Bulletin 111(4) Hyslop, E. J. 1980. Stomach contents analysis — a review of methods and their application. J. Fish Biol. 17:411-429. Jaworski, A., and S. A. Ragnarsson 2006. Feeding habits of demersal fish in Icelandic wa- ters: a multivariate approach. ICES J. Mar. Sci. 63:1682-1694. Jost, L. 2006. Entropy and diversity. Oikos 113:363-375. Latour, R. J., J. Gartland, C. F. Bonzek, and R. A. Johnson. 2008. The trophic dynamics of summer flounder ( Para - lichthys dentatus ) in Chesapeake Bay. Fish. Bull. 106:47-57. Link, J. S. 2002. Ecological considerations in fisheries management: when does it matter? Fisheries 27:10-17. Link, J. S., K. Bolles, and C. G. Milliken. 2002. The feeding ecology of flatfish in the northwest At- lantic. J. Northwest Atl. Fish. Sci. 30:1-17. Myers, R. A., and B. Worm. 2003. Rapid worldwide depletion of predatory fish com- munities. Nature 423:280-283. Nelson, G. A., B. C. Chase, and J. Stockwell. 2003. Food habits of striped bass (Morone saxatilis) in coastal water of Massachusetts. J. Northwest Atl. Fish. Sci. 32:1-25. Olla, B. L., C. E. Samet, and A. L. Studholme. 1972. Activity and feeding behavior of the summer floun- der ( Paralichthys dentatus) under controlled laboratory conditions. Fish. Bull. 70:1127-1136. Overholtz, W. J., and J. S. Link. 2007. Consumption impacts by marine mammals, fish, and seabirds on the Gulf of Maine-Georges Bank Atlan- tic herring ( Clupea harengus) complex during the years 1977-2002. ICES J. Mar. Sci. 64:83-96. Overholtz, W. J., J. S. Link, and L. E. Suslowicz. 2000. Consumption of important pelagic fish and squid by predatory fish in the northeastern USA shelf ecosys- tem with some fishery comparisons. ICES J. Mar. Sci. 57:1147-1159. Overton, A. S., C. S. I. Manooch, J. W. Smith, and K. Brennan. 2008. Interactions between adult migratory striped bass {Morone saxatilis) and their prey during winter off the Virginia and North Carolina Atlantic coast from 1994- 2007. Fish. Bull. 106:174-182. Persson, L., R Bystrom, E. Wahlstrom, J. Andersson, and J. Hjelm. 1999. Interactions among size-structured populations in a whole-lake experiment: size- and scale-dependent pro- cesses. Oikos 87:139-156. Rudershausen, P. J., J. A. Buckel, J. Edwards, D. R Gannon, C. M. Butler, and T. W. Averett 2010. Feeding ecology of blue marlins, dolphinfish, yel- lowfin tuna, and wahoos from the North Atlantic Ocean and comparison with other oceans. Trans. Am. Fish. Soc. 139:1335-1359. Sackett, D. K., K. W. Able, and T. M. Grothues. 2007. Dynamics of summer flounder, Paralichthys denta- tus, seasonal migrations based on ultrasonic telemetry. Estuar. Coast. Shelf Sci. 74:119-130 2008. Habitat dynamics of summer flounder, Paralich- thys dentatus, within a shallow USA estuary, based on multiple approaches using acoustic telemetry. Mar. Ecol. Prog. Ser. 364:199-212. Sagarese, S. R., R. M. Cerrato, and M. G. Frisk. 2011. Diet composition and feeding habits of common fishes in Long Island bays, New York. Northeast. Nat. 18:291-314. Schoener, T. W. 1970. Nonsynchronous spatial overlap of lizards in patchy habitats. Ecology 51:408-418. Schmitz, O. J. 2007. Predator diversity and trophic interactions. Ecol- ogy 88:2415-2426. Smith, B. E., and J. S. Link. 2010. The trophic dynamics of 50 finfish and 2 squid species on the northeast US continental shelf. NOAA Tech. Memo. NMFS-NE-216, 640 p. Smith, J. E. 1999. Distribution of Atlantic menhaden, Brevoortia tyrannus, purse-seine sets and catches from southern New England to North Carolina, 1985-96. NOAA Tech. Rep. NMFS 144, 22 p. Smith, L. S. 1989. Digestive functions in teleost fishes. In Fish nu- trition, 2nd ed. (J. E. Halver, ed.), p. 332-423. Academic Press, San Diego, CA. Sutton, T. M., and J. J. Ney. 2002. Trophic resource overlap between age-0 striped bass and largemouth bass in Smith Mountain Lake, Virginia. N. Am. J. Fish. Manage. 22:1250-1259. Taylor, E. T. 1987. Food habits of dominant piscivorous fishes in Delaware Bay, with special reference to predation on juvenile weakfish. M.S. thesis, 82 p. Univ. Delaware, Graduate College of Marine Studies, Lewes, DE. Terceiro, M. 2011. The summer flounder chronicles II: new science, new controversy, 2001-2010. Rev. Fish Biol. Fish. 21:681-712. Tupper, M., and K. W. Able. 2000. Movements and food habits of striped bass ( Mo- rone saxatilis) in Delaware Bay (USA) salt marshes: comparison of a restored and a reference marsh. Mar. Biol. 137: 1049-1058. Turnure, J. T. 2010. Estuarine habitat ecology of adult weakfish ( Cy - noscion regalis ): a multi-scale approach. M.S. thesis, 125 p. Rutgers Univ., New Brunswick, NJ. Uphoff, J. H„ Jr. 2003. Predator-prey analysis of striped bass and Atlan- tic menhaden in upper Chesapeake Bay. Fish. Manage. Ecol. 10:313-322. Wallace, R. K„ Jr. 1981. An assessment of diet-overlap indexes. Trans. Am. Fish. Soc. 110:72-76. Walter, J. F., Ill, A. S. Overton, K. G. Ferry, and M. E. Mather. 2003. Atlantic coast feeding habits of striped bass: a synthesis supporting a coast-wide understanding of tro- phic biology. Fish. Manage. Ecol. 10:349-360. Weiss, H. M. 1995. Marine animals of southern New England and New York: identification keys to common nearshore and shallow water macrofauna. State Geological and Natural History Survey of Connecticut, Department of Environmental Protection. Bulletin 115, 344 p. Wiedenmann, J., and T. E. Essington. 2006. Density-dependent overwinter survival in young- of-year bluefish (Pomatomus saltatrix )? A new approach for assessing stage-structured survival. Can. J. Fish. Aquat. Sci. 63:1934-1943. Wuenschel et al : Habitat and diet overlap of 4 piscivorous fishes 369 Woodland, R. J., and D. H. Secor. 2011. Differences in juvenile trophic niche for two coast- al fish species that use marine and estuarine nursery habitats. Mar. Ecol. Prog. Ser. 439:241-254. Woodland, R. J., D. H. Secor, and M. A. Wedge. 2011. Trophic resource overlap between small elas- mobranchs and sympatric teleosts in Mid-Atlantic Bight nearshore habitats. Estuaries Coasts 34: 391-04. Wuenschel, M. J., K. W. Able, J. A. Buckel, J. W. Morley, T. Lankford, A. C. Branson, D. O. Conover, D. Drisco, A. Jordaan, K. Dunton, D. H. Secor, R. J. Woodland, F. Juanes, and D. Stormer. 2012. Recruitment patterns and habitat use of young-of- the-year bluefish along the United States east coast: in- sights from coordinated coastwide sampling. Rev. Fish. Sci. 20:80-102. Wuenschel, M. J., and R. G. Werner. 2004. Consumption and gut evacuation rate of laborato- ry-reared spotted seatrout (Sciaenidae) larvae and juve- niles. J. Fish Biol. 65:723-743. Yergey, M. S. 2011. Movement of summer flounder (Paralichthys den- tatus ): application of telemetry to understanding ecol- ogy and bycatch discard mortality. M.S. thesis, 69 p. Rutgers Univ., New Brunswick, NJ. 370 Multiscale analysis of factors that affect the distribution of sharks throughout the northern Gulf of Mexico J. Marcus Drymon (contact author)’-2 Laure Carassou3 Sean P. Powers1- 2 Mark Grace4 John Dindo2 Brian Dzwonkowski2 Email address for contact author: mdrymon@disl.org 1 Department of Marine Sciences University of South Alabama, LSCB-25 Mobile, Alabama 36688 2 Dauphin Island Sea Lab 101 Bienville Boulevard Dauphin Island, Alabama 36528 3 Department of Zoology and Entomology Rhodes University P.O. Box 94 Grahamstown 6140, South Africa 4 Mississippi Laboratories Southeast Fisheries Science Center National Marine Fisheries Service, NOAA 3209 Frederic Street Pascagoula, Mississippi 39567 Abstract— Identification of the spa- tial scale at which marine com- munities are organized is critical to proper management, yet this is particularly difficult to determine for highly migratory species like sharks. We used shark catch data collected during 2006-09 from fish- ery-independent bottom-longline surveys, as well as biotic and abiotic explanatory data to identify the fac- tors that affect the distribution of coastal sharks at 2 spatial scales in the northern Gulf of Mexico. Cen- tered principal component analyses (PCAs) were used to visualize the patterns that characterize shark distributions at small (Alabama and Mississippi coast) and large (north- ern Gulf of Mexico) spatial scales. Environmental data on tempera- ture, salinity, dissolved oxygen (DO), depth, fish and crustacean biomass, and chlorophyll-a (chl-a) concentra- tion were analyzed with normed PCAs at both spatial scales. The re- lationships between values of shark catch per unit of effort (CPUE) and environmental factors were then an- alyzed at each scale with co-inertia analysis (COIA). Results from COIA indicated that the degree of agree- ment between the structure of the environmental and shark data sets was relatively higher at the small spatial scale than at the large one. CPUE of Blacktip Shark (Carcha- rhinus limbatus) was related posi- tively with crustacean biomass at both spatial scales. Similarly, CPUE of Atlantic Sharpnose Shark (Rhizo- prionodon terraenovae) was related positively with chl-a concentration and negatively with DO at both spa- tial scales. Conversely, distribution of Blacknose Shark (C. acronotus ) displayed a contrasting relationship with depth at the 2 scales consid- ered. Our results indicate that the factors influencing the distribution of sharks in the northern Gulf of Mexico are species specific but gen- erally transcend the spatial bound- aries used in our analyses. Manuscript submitted 25 October 2012. Manuscript accepted 29 August 2013. Fish. Bull. 111:370-380. doi: 10.7755/FB. 11 1.4.6 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necesarily reflect the position of the National Marine Fisheries Service, NOAA. Paramount to the conservation of marine resources and ecosystems is the identification of proper spatial scales for management plans. Al- though long recognized as a central, if not universal, concept in ecology, the notion of scale more recently has begun a transition from qualitative description to quantitative assess- ment (Schneider, 2001). For marine systems, this transition is particu- larly important because choice of spatial scale directly affects the iden- tification of patterns (Perry and Om- mer, 2003). As fisheries management plans transition to an ecosystem- based approach, the identification of suitable spatial scales becomes even more important (Hughes et ah, 2005; Francis et al., 2007). For sharks, many of which are considered top predators and play a central role in regulation of marine ecosystems (Heithaus et al., 2008), the identification of appropriate spa- tial scales for management is made more difficult than the identification of spatial scales for bony fishes be- cause of their highly migratory na- ture and relative paucity. Traditional mark-and-recapture methods allow for examination of gross spatial- scale patterns in sharks, but these methods are limited by low recap- ture rates. Pop-up satellite archival tags circumvent this problem by ex- ponentially increasing the odds of retrieving data from tagged sharks. Unfortunately, their use is often cost prohibitive, and the algorithms presently employed to estimate geo- graphic locations are too coarse to provide reliable spatial pattern data on small scales (i.e., tens of kilome- ters) (Sims, 2010; Hammerschlag et al., 2011). Consequently, information Drymon et al.: Factors that affect the distribution of sharks throughout the northern Gulf of Mexico 371 Figure 1 Spatial extent of the area used in our small-scale analysis of shark distribution in the northern Gulf of Mexico during 2006-09. (A) Eight blocks (1-8, west to east), which spanned depths from 1 to ~20 m, where the shark bottom-longline survey was conducted during all months by the Dauphin Island Sea Laboratory. (B) Sample locations for the small- scale bottom-longline data during 2006-09 (filled circles) and trawl data during 2007-09 from the Southeast Area Monitoring and Assessment Program database (http://seamap.gsmfc.org) (open circles). concerning spatial patterns and distributions of shark communities in coastal marine systems is still needed before resource managers can successfully incorporate sharks into sustainable ecosystem management plans (Heithaus et ah, 2007). Long-term, fishery-independent monitoring pro- grams are one of the most common ways to assess spatial patterns for marine vertebrates. The NOAA Southeast Fisheries Science Center (SEFSC) Mississip- pi Laboratories have been conducting annual bottom- longline surveys to assess patterns of shark distribu- tions across the entire northern Gulf of Mexico since 1995, and the data from these surveys are incorporated into stock assessments that ultimately shape fishery management plans for these animals. Given the im- portance of merging biological scales with the scales of fisheries management, we sought to examine spatial patterns in assemblages of shark species on the scale of the northern Gulf of Mexico and to investigate to what extent those patterns in shark communities were present regionally, along the coasts of Mississippi and Alabama. The goal of this investigation was to characterize the spatial distribution of shark communities in coastal waters of the northern Gulf of Mexico. Previous studies have examined the distributions of coastal sharks in the northern Gulf of Mexico (Drymon et al., 2010), and we sought to further the approach in these studies by relating spatial trends in shark species assemblages to abiotic and biotic data, including the degree to which these patterns were driven by the availability of po- tential prey items. Ultimately, we wanted to determine whether patterns in the structure of shark communi- ties and the factors that drive them are independent of scale. We predict this multifaceted approach will allow for a more precise understanding of the determinants of the spatial distributions of these predators in na- ture and for a definition of appropriate management measures. Materials and methods Small-scale study site A bottom-longline survey was initiated in May 2006 by the National Marine Fisheries Service (NMFS) and the Dauphin Island Sea Lab (DISL). During this survey sharks were sampled from waters at depths of 1-20 m along the Alabama and Mississippi coastlines (Fig.l). Sampling occurred during all months (January-Decem- ber) on NMFS research vessels (all 20-30 m in length), such as the RV HST, RV Gandy, and RV Caretta. A stratified random block design was used and 8 blocks were established along the combined coast of Missis- sippi and Alabama. Each block was ~10 km east-west and extended from the shoreline to approximately the 20-m isobath. Blocks 1-4 were located west of 88°00 W (western blocks), and blocks 5-8 were located east of 88°00 W (eastern blocks) (Fig. 1A). Sampling was allo- cated evenly and replicated within each block. For this study, we analyzed data collected in 2006-09 as part of this survey. Small-scale sampling methods Between 12 and 16 stations were randomly selected and sampled each month using a stratified random 372 Fishery Bulletin 111(4) Figure 2 Spatial extent of the area used in the large-scale analysis of shark distribution in the northern Gulf of Mexico dur- ing 2006-09. (A) Seventeen National Marine Fisheries Service statistical zones (4-21 east to west, excluding zone 12), which spanned depths from 1 to -250 m, where the bottom-longline sets were conducted by the NOAA Southeast Fisheries Science Center Mississippi Laboratories during the months of August and September. (B) Sampling locations for the large-scale bottom-longline data for 2006-09 are indicated by filled circles; trawl data for 2007-09 from the Southeast Area Monitoring and Assessment Program database (http://seamap.gsmfc.orgl are indicated by open circles. survey design that ensured equal effort across blocks 1-8 and the range of depths sampled (Fig. 1A). At each station, a single bottom-longline was set and soaked for 1 h. The main line consisted of 1.85 km (1 nmi) of 4-mm monofilament (545-kg test) that was set with 100 gangions. Gangions consisted of a longline snap and a 15/0 circle hook baited with Atlantic Mackerel ( Scomber scombrus). Each gangion was made of 3.66 m of 3-mm monofilament (320-kg test). Sharks that could be boated safely were removed from the main line, unhooked, and identified to species following Castro (2011). For each individual, sex, length (precaudal, fork, natural, and stretch total in centime- ters), weight (in kilograms), and maturity stage (when possible) were recorded. All length measurements originated at the tip of the rostrum and terminated at the origin of the precaudal pit, the noticeable fork in the tail, the upper lobe of the caudal fin in a natu- ral position, and the upper lobe of the caudal fin in a stretched position for precaudal, fork, natural, and stretch total lengths, respectively. Maturity in males was assessed according to Clark and von Schmidt (1965). Sharks were tagged either on the anterior dor- sal fin with a plastic Rototag1 (Dalton ID, Henley-on- Thames, UK) or just below the first dorsal fin with a metal dart tag. Which tag type was used depended both 1 Mention of tradenames or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. Drymon et al.: Factors that affect the distribution of sharks throughout the northern Gulf of Mexico 373 on species and size of a shark at capture (Kohler and Turner, 2001). Additional data sets To determine whether the patterns that characterize the shark community assemblage in our study region (coasts of Alabama and Mississippi, hereafter referred to as small scale ) were applicable across the northern Gulf of Mexico (hereafter called large scale), we ob- tained bottom-longline data from the SEFSC Mississip- pi Laboratories. This information included catch, fork length, and environmental data collected across Gulf of Mexico statistical zones 4-21 (Fig. 2A) in 2006-09. Dur- ing that period bottom-longline sets were conducted by the Mississippi Laboratories in August and September. The methods used for the bottom-longline survey were identical at the small and large scales, and a complete description of these methods is provided in Driggers et al. (2008). To examine factors that potentially influence the distribution of sharks on both small and large scales, we analyzed the relationships between longline shark data and a set of environmental factors, including trawl data and abiotic parameters. Biotic trawl data were obtained from the Southeast Area Monitoring and As- sessment Program (SEAMAP) database (http://seamap. gsmfc.org, November 2010) of the Gulf States Marine Fisheries Commission. We restricted our analysis of SEAMAP data to those years for which trawling was conducted across the entire northern Gulf of Mexico (2007-09). The data from those years that were used in our analysis originated from both state (Louisiana, Alabama, Mississippi, and Florida) and federal (NOAA Fisheries) regulatory agencies. All data archived in the SEAMAP trawl database were collected according to standard SEAMAP trawl protocols (Rester, 2012). Two biotic variables, representative of the availability of potential prey for sharks, were selected for inclusion in our analysis of SEAMAP trawl data: fish biomass and crustacean biomass per station in kilograms. All biomass data from the trawl data set were standard- ized to kilogram per minute. The abiotic variables tem- perature (degrees Celsius), salinity (practical salinity unit), dissolved oxygen (milligrams per liter) and depth (meters) were collected with conductivity, temperature, and depth (CTD) instruments (SBE 911p/zzs and SBE 25 plus Sealogger, Sea-Bird Electronics, Inc., Bellevue, WA) during bottom-longline sampling at both the large and small scales. To include a proxy for primary production in our analysis, we used data on chlorophyll-a (chl-a) concen- tration as a measure of phytoplankton biomass (Can- ion, 2008; Martinez-Lopez and Zavala-Hidalgo, 2009). The satellite-based ocean color data used in this study were derived from the moderate resolution imaging spectroradiometer (MODIS) on the Aqua satellite (for a detailed sensor description go to the MODIS mis- sion website at http://modis.gsfc.nasa.gov). The data on chl-a concentration used for analyses in our study were downloaded from the Ocean Color website (http:// oceancolor.gsfc.nasa.gov, accessed April 2012). For this study, annual binned level-3 chl-a data (Campbell et al., 1995) at a spatial resolution of 4 km were used from 2006 to 2009. The annual composites are produced by averaging all valid, cloud-free acquisitions for each ocean pixel. The valid pixels are determined by using an extensive quality control process that tests for nu- merous factors known to degrade data accuracy. Addi- tional details for the level-3 chl-a data can be found at http://modis.gsfc.nasa.gov/data/atbd/index.php. Despite that extensive quality control process, the optically complex nature of the coastal zone can still present dif- ficulties for ocean color algorithms. In the case of data on chl-a concentration, algorithms are known to over- estimate concentrations in coastal zones, particularly in regions that are influenced by a river, because of estuarine materials, such as suspended sediment and concentrations of dissolved organic material. However, this phenomenon occurs primarily at depths <10 m (Martinez-Lopez and Zavala-Hidalgo, 2009); therefore, we obtained data on chl-a concentration from the 25-m isobath to limit the effect of these degrading influences. Data analyses Bottom-longline data sets were limited to those spe- cies observed in both the small- and large-scale bot- tom-longline surveys. Data of catch per unit of effort (CPUE), measured as sharks 100 hooks-1 h-1, were iog(x+l)-transformed to reduce the influence of the most common species and to standardize the data (Leg- endre and Legendre, 1998). All sets, including those with zero catches, were included in our analyses. Mean CPUE data were then analyzed as a function of block (blocks 1-8) across the small scale (Fig. 1A) and as a function of statistical zone (zones 4-21, minus zone 12) across the large scale (Fig. 2A). A centered principal component analysis (PCA) was performed on mean transformed shark CPUE data for both small- and large-scale data. The data collected with the CTD instruments and the MODIS satellite data (collectively hereafter referred to as environmen- tal data) at both small and large scales were analyzed with a normed PCA. At each spatial scale, centered (for shark CPUE data) and normed (for environmental data) PCAs allowed for the identification of the major spatial patterns that characterize shark assemblages and environmental conditions and for the visualiza- tion of covariances between shark species and of cor- relations between environmental factors (Legendre and Legendre, 1998). The relationships between values of shark CPUE and environmental factors were then analyzed at each spatial scale with co-inertia analyses (COIA). Co-in- ertia analysis is a flexible, multivariate method that couples environmental and faunal data and measures the degree of agreement between them (Doledec and 374 Fishery Bulletin 111(4) Chessel, 1994; Dray et al., 2003). This method has been used successfully on diverse ecological data sets and organisms, including fishes (e.g., Mellin et ah, 2007; Carassou et al., 2011; Lecchini et ah, 2012), zooplank- ton (e.g., Carassou et ah, 2010), benthic invertebrates (e.g., Bremner et ah, 2003), and bacteria (e.g., Jardillier et ah, 2004). In our study, each COIA was based on the matching of a normed PCA of environmental data and a centered PCA of shark abundance data (PCA- PCA-COIA, Dray et ah, 2003). Monte Carlo tests with 10,000 permutations between observations were used to confirm the significance of COIA results (fixed-D test; Dray et ah, 2003), with significance assessed at P<0.05. For each COIA, the vectorial correlation (RV) coefficient, a multivariate generalization of the squared Pearson’s correlation coefficient, provided a quantita- tive measure of the co-structure between explanative (environmental) and explained (shark CPUE) vari- ables, with a value of 1 indicating a perfect match be- tween the 2 data sets (Doledec and Chessel, 1994; Dray et ah, 2003). The criterion of total inertia was used to compare the amount of agreement between environ- mental and shark data for the 2 spatial scales consid- ered (Dray et ah, 2003). All multivariate analyses were performed with the ADE-4 software (Thioulouse et ah, 1997, 1995-2000). Results Small-scale sampling During small-scale sampling, 353 stations were sur- veyed, spanning the months from March to November during 2006-09 (Fig. IB). Winter months (December, January, and February) were excluded from subsequent analyses because of the complete absence of sharks in the small-scale survey area during this time (2100 hooks with no sharks). Over the course of this survey, 2417 individuals representing 12 shark species were encountered. Of these 12 species, 5 species met our criteria for inclusion in subsequent analyses (i.e., they also were abundant in the large-scale data set): At- lantic Sharpnose Shark ( Rhizoprionodon terraenovae), Blacktip Shark ( Carcliarhinus limbatus), Blacknose Shark (C. acronotus ), Spinner Shark (C. brevipinna), and Bull Shark (C. leucas). Mean CPUE (±standard er- ror [SE] ) ranged from 2.88 [0.28] sharks 100 hooks-1 h-1 for Atlantic Sharpnose Shark to 0.11 (0.02) sharks 100 hooks-1 h-1 for Bull Shark (Table 1). Wide size ranges, with size measured as fork length (FL) in centimeters, were found for Atlantic Sharpnose (36.0-96.3 cm FL), Blacktip (59.8-164.0 cm FL), Blacknose (40.9-136.0 cm FL), and Spinner (49.9-165.9 cm FL) Sharks. A small- er size range was seen for Bull Sharks (73.0-155.5 cm FL), the least commonly encountered of the 5 species (Table 1). The centered PCA conducted with small-scale data on shark abundance explained 91.88% of the variabil- ity between observations (across blocks 1-8) on the first 2 principal components (PCI and PC2) (Fig. 3A). Variation along PCI was explained primarily by data for Atlantic Sharpnose Shark, which was most abun- dant in blocks 2, 3, and 4 (western blocks), less com- mon in block 1, and relatively rare in blocks 5, 6, 7, and 8 (eastern blocks). Spinner Shark showed a similar but less marked spatial pattern (Fig. 3A). Variation along PC2 was explained primarily by data for Blacktip Shark, which was more abundant in block 1 (western block), and relatively rare in block 5 and 6 (eastern blocks) (Fig. 3A). Patterns were less clear for Blacknose and Bull Shark. The normed PCA on small-scale environmental data explained 74.18% of the variability between observa- tions (blocks) on the first 2 principle components (PCI and PC2) (Fig. 4A). Temperature and crustacean bio- mass were positively correlated with each other and both of those variables had a high negative correlation with salinity. These 3 variables explained most of the variability along PCI. Fish biomass was negatively cor- related with depth. Chl-a concentration and dissolved oxygen were negatively correlated, together explain- ing most of the variability along PC2. Blocks 7 and 8 (eastern blocks) were characterized by high dissolved oxygen and low concentration of chl-a, and the inverse was true for block 3 (a western block) (Fig. 4A). The COIA that coupled small-scale shark abundance and environmental data was characterized by a total inertia of 0.22 and an RV coefficient of 0.65, indicat- ing a relatively high degree of agreement between the structures of the 2 data sets. Axes 1 and 2 supported 99.17% of this common structure (Fig. 5A). Atlantic Sharpnose Shark abundance was positively related with chl-a concentration and negatively related with dissolved oxygen and salinity. Abundance of Blacktip Shark was more positively associated with crustacean biomass than wfith other environmental variables. Blacknose and Spinner Sharks had high negative associations with dissolved oxygen, and Blacknose Shark had a strong positive association with depth (Fig. 5A). Large-scale sampling Across the large-scale survey area, shark abundance data were obtained from 551 stations sampled during the months of August and September during 2006-09 (Fig. 2B). Over the course of this survey, 4493 sharks, comprising 26 species, were captured. Mean catch per unit of effort (±SE) ranged from 4.74 (0.41) sharks 100 hooks-1 h-1 for Atlantic Sharpnose Shark to 0.06 (0.01) sharks 100 hooks-1 h-1 for Bull Shark (Table 1). Wide size ranges were observed for Atlantic Sharpnose (33.0-115.5 cm FL), Blacktip (38.2-157.0 cm FL), Blac- knose (40.0-104.9 cm FL), and Spinner (54.0-169.0 cm FL) Sharks. The smallest size range was seen in Bull Shark (131.4-176.0 cm FL), the least commonly en- countered of the 5 species (Table 1). Drymon et al. : Factors that affect the distribution of sharks throughout the northern Gulf of Mexico 375 Table 1 Data that we used in our analyses of shark distribution in the northern Gulf of Mexico during 2006-09. Number, mean size (measured as fork length [FL] in centimeters and standard error of the mean [SE] ), size range, and mean catch per unit of effort (CPUE), measured as sharks 100 hooks-1 h-1, are shown for the 5 shark species common to both of the 2 data sets: small (Alabama and Mississippi coasts) and large (across the northern Gulf of Mexico). The 5 species were Atlantic Sharpnose Shark ( Rhizoprionodon terraenouae), Blacktip Shark ( Carcharhi - nus limbatus), Blacknose Shark (C. acronotus ), Spinner Shark (C. breuipinna), and Bull Shark (C. leucas). n= no. of sharks sampled. Species n Mean size ±SE (cm FL) Range (cm FL) Mean CPUE ±SE (sharks 100 hooks-1 h-1) Small scale Atlantic Sharpnose Shark 1016 68.8 (0.43) 36.0-96.3 2.88 (0.28) Blacktip Shark 474 102.7 (0.93) 59.8-164.0 1.34 (0.14) Blacknose Shark 600 91.1 (0.42) 40.9-136.0 1.70 (0.19) Spinner Shark 147 70.1 (1.86) 49.9-165.9 0.42 (0.06) Bull Shark 40 102.8 (6.09) 73.0-155.5 0.11 (0.02) Large scale Atlantic Sharpnose Shark 2596 73.9 (0.20) 33.0-115.5 4.74 (0.41) Blacktip Shark 254 111.7 (1.19) 38.2-157.0 0.51 (0.08) Blacknose Shark 530 85.0 (0.52) 40.0-104.9 0.98 (0.11) Spinner Shark 158 104.7 (2.17) 54.0-169.0 0.29 (0.08) Bull Shark 21 155.0 (2.77) 131.4-176.0 0.06 (0.01) Figure 3 Results of the centered principal components analysis (PCA) on shark data from (A) small-scale and (B) large- scale bottom-longline surveys conducted in the northern Gulf of Mexico during 2006-09. Numbers within the panels correspond to the sampling blocks (1-8) and statistical zones (4-21, minus 12) of the small- and large- scale surveys, respectively, used in our analysis (blocks and zones are defined in Figs. 1 and 2). Filled circles represent shark species. The sum of the variation explained by the first (PCI) and second (PC2) principal com- ponents is 91.88% for small-scale survey and 87.30% for large-scale survey. The scale is shown in ovals at top of each panel. 376 Fishery Bulletin 111(4) Figure 4 Results of the normed principal components analysis (PCA) on (A) small- and (B) large-scale environmental data from CTD casts conducted during the bottom-longline survey in the northern Gulf of Mexico during 2006-09, trawl data from the Southeast Area Monitoring and Assessment Program database (http://seamap.gsmfc.org) for 2007- 2009, and from the moderate resolution imaging spectroradiometer on the Aqua satellite (http://modis.gsfc.nasa.gov) for 2006-2009. Numbers within the panel correspond to the sampling blocks (1-8) and statistical zones (4-21, minus 12) of the small- and large-scale surveys, respectively, used in our analysis (blocks and zones are defined in Figures 1 and 2). Arrows represent abiotic variables, and dashed-line circles represent correlation circles with a unit of 1. Variation explained by the first (PCI) and second (PC2) principal components is 74.18% for the small-scale survey and 65.88% for the large-scale survey. The scale is shown in ovals at top of each panel. Cbio=crustacean biomass, Chl-a=chlorophyll-a, DO=dissolved oxygen, Fbio=fish biomass, Sal=salinity, Temp=temperature. The centered PCA conducted with large-scale data on shark abundance explained 87.30% of the vari- ability between observations (across NMFS statistical zones) on the first 2 principal components (Fig. 3B). Variation along PCI was mainly explained by data for Atlantic Sharpnose Shark, which was more abundant in zones 11, 14, and 16 (western zones), and variation along PC2 was mainly explained by data for Blacknose Shark, which was more abundant in zones 3 and 5 (eastern zones) (Fig. 3B). Compared with other species, Bull Shark displayed a weaker pattern because of their lower abundances (Fig. 3B). The normed PCA on large-scale environmental data explained 65.88% of the variability between observa- tions (NMFS statistical zones) on the first 2 principle components (Fig. 4B). Fish biomass and temperature were correlated, and both of these variables were negatively correlated with depth. These 3 variables explained most of the variability along PCI (Fig. 4B). Chl-a concentration and crustacean biomass were positively correlated, and concentration of chl-a had a strong negative correlation with dissolved oxygen. To- gether, these 3 variables explained the majority of vari- ability along PC2. NMFS statistical zones 11, 14, 15, and 16 were characterized by high chl-a concentration, and zones 18 and 19 were characterized by high fish biomass. Conversely, eastern zones 4-6 were character- ized by low fish and crustacean biomass (Fig. 4B). The COIA that coupled large-scale shark abundance and environmental data was characterized by a total in- ertia of 0.20 and a RV coefficient of 0.42, indicating good agreement between the 2 data sets. Axes 1 and 2 support- ed 97.32% of this common structure (Fig. 5B). Abundance of Atlantic Sharpnose Shark was strongly related to chl-a concentration and had a strong negative relation to dis- solved oxygen. Spinner Shark showed a similar pattern. Blacktip Shark abundance was related to crustacean biomass and had a strong negative relation to salinity. Abundance of Blacknose Shark was strongly related to temperature and inversely related to depth (Fig. 5B) Discussion For comparison of the factors that affect the distribu- tion of sharks across spatial scales, COIA provides a Drymon et al Factors that affect the distribution of sharks throughout the northern Gulf of Mexico 377 Figure 5 Results of the co-inertia analyses on (A) small- and (B) large-scale shark and environmental data from the north- ern Gulf of Mexico during 2006-09. Small-scale total inertia is 0.22, and axes 1 and 2 supported 99.17% of this structure. Large-scale total inertia is 0.20, and axes 1 and 2 supported 97.32% of this structure. The scale is shown in ovals at top of each panel. Arrows and dotted lines represent environmental variables. Filled circles and full lines represent shark species. Cbio=crustacean biomass, Chl-a=chlorophyll-a, DO=dissolved oxygen, Fbio=fish bio- mass, Sal=salinity, Temp=temperature. robust tool. Examination of the results for total inertia indicates that analyses at both small and large scales were equally useful for identification of patterns be- tween sharks and explanatory variables. However, the RV coefficients indicate that explanatory variables were better correlated with shark abundances at the small scale (RV=0.65) than at the large scale (RV=0.42). Giv- en 1) the unique coupling of bottom-longline data sets collected through the use of identical methods across the same temporal scale and 2) the similarity in shark size and catch between the surveys at 2 spatial scales, our data are particularly well suited to COIA. Our re- sults indicate that the factors affecting the distribution of sharks in the Gulf of Mexico are species specific but relatively well conserved across spatial scales. The factors that affect the distribution of Blacktip Shark were similar at small and large scales, and the distribution of this species was best explained by crus- tacean biomass at both scales. However, it is unlikely that Blacktip Shark responded to crustaceans as po- tential prey. Although previous studies of feeding hab- its of Blacktip Shark in the northern Gulf of Mexico (Hoffmayer and Parsons, 2003; Barry et al., 2008) and Florida (Heupel and Hueter, 2002) have identified crus- tacean components, these same studies have revealed that Blacktip Sharks prey predominately on teleosts. That Blacktip Shark distributions may not be influ- enced by the distribution of their preferred prey is not surprising. In an acoustic telemetry study in Terra Ceia Bay, Florida, no correlation was found between juvenile Blacktip Shark and their prey (Heupel and Hueter, 2002). After examination of the influence of prey abundance on the distribution of sharks (includ- ing Blacktip Shark) at 2 spatial scales, Torres et al. (2006) showed no correlation between shark catch and teleost abundance at individual sampling locations, al- though a correlation was shown between shark catch and teleost abundance within a region. The strong rela- tionship observed in our study between Blacktip Shark and crustacean biomass at both spatial scales indicates that perhaps the underlying mechanism that most in- fluences the distribution of this species correlates with crustacean biomass. Distribution of Atlantic Sharpnose Shark was best explained by chl-a concentration, a pattern that, like the one seen for Blacktip Shark, was independent of scale. However, although Blacktip Shark may have been influenced by factors other than prey, Atlantic Sharpnose Shark may have been indirectly responding to available prey as indicated by the observed relation- ship with concentration of chl-a. The contrast between Blacktip Shark and Atlantic Sharpnose Shark may il- lustrate basic differences in the ecology of these 2 spe- cies. As adults, Blacktip Sharks are a larger, more mo- 378 Fishery Bulletin 111(4) bile species than Atlantic Sharpnose Sharks, and they are capable of moving hundreds of kilometers on short time scales, as illustrated by traditional (Kohler et al., 1998) and pop-up satellite archival (senior author and S. Powers, unpubl. data) tagging data. In contrast, At- lantic Sharpnose Sharks have a relatively small home range (Carlson et al., 2008). Blacktip Sharks, compared with Atlantic Sharpnose Sharks, may show higher va- gility when faced with a patchy prey environment. For example, Atlantic Sharpnose Sharks sampled in the small-scale survey showed relative trophic plasticity. Portunid crabs and shrimp contribute more to the diet of Atlantic Sharpnose Sharks sampled west (blocks 1-4 in the current study) than to the diet of this shark spe- cies east (blocks 5-8) of Mobile Bay, and therefore may reflect differences in the prey base between these 2 ar- eas (Drymon et al., 2012). These findings indicate that the Atlantic Sharpnose Shark may have a wider di- etary breadth than the Blacktip Shark and may, there- fore, be responding to gradients in overall production as opposed to fish or crustacean biomass, specifically. Distributions of Atlantic Sharpnose and Spinner Sharks at both large and small scales were negative- ly related to dissolved oxygen. This relationship has been previously identified for other species of juvenile sharks. In Chesapeake Bay, Virginia, tree-based regres- sion models indicated the importance of dissolved oxy- gen as a factor that influences the distribution of juve- nile Sandbar Shark ( Carcharhinus plumbeus) (Grubbs and Musick, 2007). Similarly, researchers have noted that, although dissolved oxygen is not as widely consid- ered as temperature or salinity, it may play an impor- tant role as an environmental influence that affects the distribution of top predators in coastal environments, as has been demonstrated for juvenile Bull Shark in Florida waters (Heithaus et al., 2009). In our study, a wide size range of Spinner Shark was documented across both the small- and large-scale surveys. On the basis of age and growth data (Carl- son and Baremore, 2005), the mean sizes of Spinner Shark captured in small- and large-scale surveys corre- sponded to the ages of approximately 1 and 4 years old, respectively. Conversely, across surveys at both spatial scales, the mean size of Atlantic Sharpnose Shark was indicative of mature, adult animals (Carlson and Bare- more, 2003). Our findings, therefore, support previous work that indicated the importance of dissolved oxygen as an influence on the distribution of juvenile sharks (Grubbs and Musick, 2007; Heithaus et al., 2009) and indicates that dissolved oxygen may influence the dis- tribution of adult sharks as well. Distributions of Blacknose Shark were best ex- plained by depth, the direction of which varied as a function of scale. On the small scale, Blacknose Shark distribution was strongly and positively associated with water depth (i.e., deeper water resulted in higher Blacknose Shark CPUE). Conversely, at the large scale, distribution of Blacknose Shark were strongly and neg- atively associated with deep water (i.e., the shallower the depth, the lower the observed CPUE Blacknose Shark). This apparent dichotomy highlights differenc- es in the range of depths associated with each spatial scale and likely reflects a preferred depth range for this species. Small-scale sampling occurred at depths up to ~20 m, and large-scale sampling occurred pri- marily at depths >20 m. Discrete depth preferences for Blacknose Shark have previously been documented. Analyzing the same 2 bottom-longline data sets used in our analyses, Drymon et al. (2010) showed a discrete depth preference of 10-30 m for Blacknose Shark. Our data support these findings yet provide no additional insight into why Blacknose Shark occupy these depths. Although our analyses identified factors that may influence the distribution of selected shark species at 2 different spatial scales, our approach has certain limitations. For instance, the faunal component of our analyses was based on catch data (CPUE). Bait loss can affect CPUE calculations (Torres et al., 2006). In areas where (or during times when) bait loss is high, CPUE may be artificially low. Recording the status of individual gangions (i.e., fish caught, bait present, bait absent) allows for hook-specific CPUE to be calculated, resulting in more accurate determination of CPUE and, hence, improving the power of this approach. In addi- tion, the analyses we used are sensitive to the tem- poral alignment of the data sets used. Restriction of analyses to data collected with the same methods and during the same time period will facilitate the iden- tification of reliable relationships between faunal and explanatory data. Conclusions Identification of the factors that affect the distribution of large predators is challenging. Our analysis encom- passes physical parameters (salinity, temperature, dis- solved oxygen, and depth), proxies for primary (chl-a concentration) and secondary (trawl) productivity, and predatory data sets across 2 spatial scales. Our results indicate that the factors that affect the distribution of sharks in the Gulf of Mexico are species dependent but may transcend the spatial boundaries that we exam- ined. As physical and biological characteristics of eco- systems in the Gulf of Mexico change, species-specific knowledge of how these factors influence the distribu- tions of top predators will be critical for the implemen- tation of proactive management measures. Acknowledgments The authors wish to thank all members of the Fisher- ies Ecology Laboratory at Dauphin Island Sea Labora- tory (DISL), as well as members of the NOAA South- east Fisheries Science Center Mississippi Laboratories shark team for the countless hours they spent at sea collecting valuable data. Data from DISL’s Fisheries Drymon et a!.: Factors that affect the distribution of sharks throughout the northern Gulf of Mexico 379 Oceanography of Coastal Alabama research program, which is funded by the Alabama Department of Con- servation and Natural Resources, Marine Resources Di- vision, were used to provide ground truth for remotely sensed data from NASA’s Sea-viewing Wide Field-of- view Sensor (SeaWifs) Project. We wish to thank T. Henwood and E. Hoffmayer from the National Marine Fisheries Service for constructive comments that im- proved this manuscript. Literature cited Barry, K. P., R. E. Condrey, W. B. Driggers III, and C. M. Jones. 2008. Feeding ecology and growth of neonate and juve- nile blacktip sharks Carcharhinus limbatus in the Tim- balier-Terrebone Bay complex, LA, U.S.A. J. Fish Biol. 73:650-662. Bremner, J., S. I. Rogers, and C. L. J. Frid. 2003. Assessing functional diversity in marine benthic ecosystems: a comparison of approaches. Mar. Ecol. Prog. Ser. 254:11-25. Campbell, J. W., J. M. Blaisdell, and M. Darzi. 1995. Level-3 SeaWiFS data products: spatial and tem- poral binning algorithms. NASA Tech. Memo. 104566, vol. 32, 73 p. Canion, A.K. 2008. Variability in productivity model inputs on mul- tiple timescales: implications for productivity monitor- ing in Weeks Bay, Alabama. M.S. thesis, 99 p. Univ. South Alabama, Mobile, AL. Carassou, L., R. LeBorgne, E. Rolland, and D. Ponton. 2010. Spatial and temporal distribution of zooplankton related to the environmental conditions in the coral reef lagoon of New Caledonia, Southwest Pacific. Mar. Pol- lut. Bull. 61:367-374. Carassou, L., B. Dzwonkowski, F. J. Hernandez, S. P. Powers, K. Park, W. M. Graham, and J. Mareska. 2011. Environmental influences on juvenile fish abun- dances in a river-dominated coastal system. Mar. Coast. Fish. 3:411-427. Carlson, J. K., and I. E. Baremore. 2003. Changes in biological parameters of Atlantic sharpnose shark Rhizoprionodon terraenovae in the Gulf of Mexico: evidence for density-dependent growth and maturity? Mar. Freshw. Res. 54:227-234. Carlson, J. K., and I. E. Baremore. 2005. Growth dynamics of the spinner shark (Carcha- rhinus brevipinna ) off the LInited States southeast and Gulf of Mexico coasts: a comparison of methods. Fish. Bull. 103:280-291. Carlson, J. K., M. R. Heupel, D. M. Bethea, and L. D. Hollensead. 2008. Coastal habitat use and residency of juvenile At- lantic sharpnose sharks ( Rhizoprionodon terraenovae). Estuaries Coasts 31:931-940. Castro, J. I. 2011. The sharks of North America, 640 p. Oxford Univ. Press, Inc., New York. Clark, E., and K. von Schmidt. 1965. Sharks of the central Gulf coast of Florida. Bull. Mar. Sci. 15:13-83. Doledec, S., and D. Chessel. 1994. Co-inertia analysis: an alternative method for studying species-environment relationships. Freshw. Biol. 31:277-294. Dray, S. D., D. Chessel, and J. Thioulouse. 2003. Co-inertia analysis and the linking of ecological data tables. Ecology 84(ll):3078-3089. Driggers, W. B., G. W. Ingram, M. A. Grace, C. T. Gledhill, T. A. Henwood, C. N. Horton, and C. M. Jones. 2008. Pupping areas and mortality rates of young tiger sharks Galeocerdo cuvier in the western North Atlantic Ocean. Aquat. Biol. 2:161-170. Drymon, J. M., S. P. Powers, J. Dindo, B. Dzwonkowski, and T. A. Henwood. 2010. Distributions of sharks across a continental shelf in the northern Gulf of Mexico. Mar. Coast. Fish. 2:440-250. Drymon, J. M., S. P. Powers, and R. H. Carmichael. 2012. Trophic plasticity in the Atlantic sharpnose shark (Rhizoprionodon terraenovae ) from the north central Gulf of Mexico. Environ. Biol. Fishes 95:21-35. Francis, R. C., M. A. Hixon, M. E. Clarke, S. A. Murawski, and S. Ralston. 2007. Ten commandments for ecosystem-based fisheries scientists. Fisheries 32:217-233. Grubbs, R. D., and J. A. Musick. 2007. Spatial delineation of summer nursery areas for juvenile sandbar sharks in Chesapeake Bay, Virginia. Am. Fish. Soc. Symp. 50:87-108. Hammerschlag, N., A. J. Gallagher, and D. M. Lazarre. 2011. A review of shark satellite tagging studies. J. Exp. Mar. Biol. Ecol. 398(1-21:1-8. Heithaus, M. R., D. Burkholder, R. E. Heuter, L. I. Heithaus, H. R. Pratt Jr., and J. C. Carrier. 2007. Spatial and temporal variation in shark commu- nities of the lower Florida Keys and evidence for his- torical population declines. Can. J. Fish. Aquat. Sci. 64:1302-1313. Heithaus, M. R., A. Frid, A.J. Wirsing, and B. Worm. 2008. Predicting ecological consequences of marine top predator declines. Trends Ecol. Evol. 23:202-210. Heithaus, M. R., B. K. Delius, A. J. Wirsing, and M. M. Dunphy-Daly. 2009. Physical factors influencing the distribution of a top predator in a subtropical oligotrophic estuary. Limnol. Oceanogr. 54:472-482. Heupel, M. R., and R. E. Hueter. 2002. Importance of prey density in relation to the movement patterns of juvenile blacktip sharks (Carcha- rhinus limbatus ) within a coastal nursery area. Mar. Freshw. Res. 53:543-550. Hoffmayer, E. R., and G. R. Parsons. 2003. Food habits of three shark species from the Mis- sissippi Sound in the northern Gulf of Mexico. South- east. Nat. 2:271-280. Hughes, T. P, D. R. Bellwood, C. Folke, R. S. Steneck, and J. Wilson. 2005. New paradigms for supporting the resilience of marine ecosystems. Trends Ecol. Evol. 20:380-386. Jardillier, L., M. Basset, I. Domaizon, A. Belan, C. Amblard, M. Richardot, and D. Debroas. 2004. Bottom-up and top-down control of bacterial com- munity composition in the euphotic zone of a reservoir. Aquat. Microb. Ecol. 35:259-273. 380 Fishery Bulletin 111(4) Kohler, N. E., J. G. Casey, and P. A. Turner. 1998. NMFS cooperative shark tagging program, 1962- 1993: an atlas of shark tag and recapture data. Mar. Fish. Rev. 60(2): 1-87. Kohler, N. E., and P. A. Turner. 2001. Shark tagging: a review of conventional methods and studies. Environ. Biol. Fishes. 60:191-223. Lecchini, D., L. Carassou, B. Frederich, Y. Nakamura, S. C. Mills, and R. Galzin. 2012. Effects of alternate reef states on coral reef fish habitat associations. Environ. Biol. Fishes 94:421-429. Legendre, P, and L. Legendre. 1998. Numerical ecology, 2nd ed., 853 p. Elsevier Sci- ence, Amsterdam. Martinez-Lopez, B., and J. Zavala-Hidalgo. 2009. Seasonal and interannual variability of cross-shelf transport of chlorophyll in the Gulf of Mexico. J. Mar. Syst. 77:1-20. Mellin, C., M. Kulbicki, and D. Ponton. 2007. Seasonal and ontogenetic patterns of habitat use in coral reef fish juveniles. Estuar. Coast. Shelf Sci. 75:481-491. Perry, R. I., and R. E. Ommer. 2003. Scale issues in marine ecosystems and human in- teractions. Fish. Oceanogr. 12:513-522. Schneider, D. C. 2001. The rise of the concept of scale in ecology. Biosci- ence 51:545-553. Sims, D. W. 2010. Tracking and analysis techniques for understand- ing free-ranging shark movements and behavior. In Sharks and their relatives II: biodiversity, adaptive physiology, and conservation (J. C. Carrier, J. A. Mu- sick, and M. R. Heithaus, eds.), p. 351-392. CRC Press, Boca Raton, FL. Rester, J. K. (ed.). 2012. SEAMAP environmental and biological atlas of the Gulf of Mexico, 2010, 84 p. Gulf States Marine Fisheries Commission, Ocean Springs, MS. Thioulouse, J., D. Chessel, S. Doledec, and J-M. Olivier. 1997. ADE-4: a multivariate analysis and graphical dis- play software. Stat. Comput. 7:75-83. Thioulouse, J., D. Chessel, S. Doledec, J-M. Oliver, F. Goreaud, and R. Pelissier. 1995-2000. ADE-4. Ecological data analysis: explorato- ry and euclidean methods in environmental sciences. Version 2001, CNRS, Lyon, France. [Available from http://pbil.univ-lyonl.fr/ADE-4-old/ADE-4.html.] Torres, L. G., M. R. Heithaus, and B. Delius. 2006. Influence of teleost abundance on the distribution and abundance of sharks in Florida Bay, USA. Hydro- biologia 569:449-455. 381 Abstract-— We describe the food hab- its of the Sowerby’s beaked whale (Mesoplodon bidens) from observa- tions of 10 individuals taken as bycatch in the pelagic drift gillnet fishery for Swordfish ( Xiphias gla- dius ) in the western North Atlan- tic and 1 stranded individual from Kennebunk, Maine. The stomachs of 8 bycaught whales were intact and contained prey. The diet of these 8 whales was dominated by meso- and benthopelagic fishes that composed 98.5% of the prey items found in their stomachs and cepha- lopods that accounted for only 1.5% of the number of prey. Otoliths and jaws representing at least 31 fish taxa from 15 families were pres- ent in the stomach contents. Fishes, primarily from the families Mori- dae (37.9% of prey), Myctophidae (22.9%), Macrouridae (11.2%), and Phycidae (7.2%), were present in all 8 stomachs. Most prey were from 5 fish taxa: Shortbeard Codling ( Lae - monema barbatulum) accounted for 35.3% of otoliths, Cocco’s Lantern- fish (Lobianchia gemellarii) contrib- uted 12.9%, Marlin-spike (Nezumia bairdii) composed 10.8%, lantern- fishes ( Lampanyctus spp.) accounted for 8.4%; and Longfin Hake ( Phycis chesteri) contributed 6.7%. The mean number of otoliths per stomach was 1196 (range: 327-3452). Most of the fish prey found in the stomachs was quite small, ranging in length from 4.0 to 27.7 cm. We conclude that the Sowerby’s beaked whales that we examined in this study fed on large numbers of relatively small meso- and benthopelagic fishes that are abundant along the slope and shelf break of the western North Atlantic. Manuscript submitted 4 February 2013. Manuscript accepted 30 August 2013. Fish. Bull. 111:381-389. doi: 10.7755/FB.111.4.7 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necesarily reflect the position of the National Marine Fisheries Service, NOAA. Food habits of Sowerby's beaked whales C Mesoplodon bidens ) taken in the pelagic drift gillnet fishery of the western North Atlantic Frederick W. Wenzel (contact author)1 Pamela T. Polloni2 James E. Craddock2 (deceased) Damon P. Gannon3 John R. Nicolas1 (deceased) Andrew J. Read4 Patricia E. Rosel5 Email address for contact author: frederick.wenzel@noaa.gov 1 Protected Species Branch Northeast Fisheries Science Center National Marine Fisheries Service, NOAA 166 Water Street Woods Hole, Massachusetts 02543 2 Biology Department Woods Hole Oceanographic Institution Woods Hole, Massachusetts 02543 3 Department of Biology Bowdoin College 6500 College Station Brunswick, Maine 04011 4 Division of Marine Science and Conservation Nicholas School of the Environment Duke University Beaufort, North Carolina 28516 5 Protected Resources and Biodiversity Division Southeast Fisheries Science Center National Marine Fisheries Service, NOAA 646 Cajundome Blvd Lafayette, Louisiana 70506 The Sowerby’s beaked whale ( Meso- plodon bidens) is 1 of 4 species of the genus Mesoplodon (Family Ziphiidae) in the western North Atlantic. The Sowerby’s beaked whale is restricted to the North Atlantic and the most boreal species in its genus, with ob- servations recorded as far north as 71°N (Carlstrom et al., 1997; Hook- er and Baird, 1999; McAlpine and Rae, 1999; Lucas and Hooker, 2000; Waring et al., 2010). There is also a single record of a stranded Sowerby’s beaked whale from the Gulf of Mexi- co (Bonde and O’Shea, 1989). Most information on the distribu- tion and abundance of beaked whales off the northeastern coast of the United States has been derived from vessel surveys conducted by NOAA Fisheries. It is difficult to identify Mesoplodon beaked whales to species level at sea; therefore estimates of abundance are often reported at the generic level in stock assessments (e.g., Waring et al., 2010). Waring et al. (2001) reported that off the north- eastern coast of the United States, Mesoplodon beaked whales were encountered most frequently along the shelf break and north wall of the Gulf Stream. The habitat prefer- ences of these animals overlap with the habitat preferences of the sperm whale (Physeter macrocephalus), but Sowerby’s beaked whales were con- centrated on the colder shelf edge (Griffin, 1999; Waring et al., 2001). MacLeod et al. (2003) reviewed available information on the diet of beaked whales and concluded that fishes are important prey of 5 of the 10 (Family Ziphiidae) species for which diet information was avail- able. This conclusion stands in con- trast to earlier reviews of the diet of beaked whales where the impor- tance of squids was emphasized (e.g., Clarke, 1986). Beaked whales are cryptic, deep-diving odontocetes, and. 382 Fishery Bulletin 111(4) as a result, direct observation of foraging is impossi- ble. Most insight into their feeding behavior has come from digital acoustic tags, which record the 3-D move- ment and acoustic environment of tagged individuals (e.g., Madsen et al., 2005). Application of these tags to individuals of the Blainville’s beaked whale ( Mesoplo - don densirostris) indicates that this species forages at depths of more than 1000 in in dives that may last for almost 1 h (Arranz et ah, 2011). To date, however, no Sowerby’s beaked whales have been studied with digi- tal acoustic tags. Given the challenges of studying live whales, all published information on the food habits of the Sow- erby’s beaked whale has been acquired from stranded specimens (Dix et al., 1986; Lien and Barry, 1990; Lien et al. 1990; Ostrom et al., 1993; Pereira et ah, 2011; Spitz et al., 2011; Santos et al.1’2). Recent analysis of the stomach contents of 10 stranded Sowerby’s beaked whales from the Azores in the eastern North Atlan- tic (Pereira et al., 2011) provided evidence that small meso- and bathypelagic fishes constitute an important part of the diet of this species in this area. One largely untapped source of information on the biology of the Sowerby’s beaked whale comes from a sample of animals taken as bycatch in a pelagic drift gillnet fishery for Swordfish ( Xiphias gladius) that op- erated in the western North Atlantic between 1989 and 1998. The pelagic drift gillnet fishery was monitored by observers from the Northeast Fisheries Observer Pro- gram (NEFOP); these observers documented bycatch consisting of more than 1100 individuals of 14 marine mammal species (Waring et al., 2000). This bycatch included 46 beaked whales taken in the “northern or summer stratum” of the fishery that operated along the continental shelf break along the southern side of Georges Bank (Waring et al., 2009). Pelagic drift gill- nets were prohibited after 1998 because of the large number of cetaceans taken during fishing operations that used them (Waring et al., 2000; 2002). Here, we describe the stomach contents of Sowerby’s beaked whales taken in this pelagic drift net fishery, and we provide the first detailed account of the food habits of Sowerby’s beaked whales from the western North Atlantic. Materials and methods We examined the stomach contents of 10 Sowerby’s beaked whales taken incidentally in the pelagic drift gillnet fishery for Swordfish in the Atlantic between August 1989 and July 1996 and a single dead stranded 1 Santos, M. B., G. J. Pierce, H. M. Ross, R. J. Reid, and B. Wilson. 1994. Diets of small cetaceans from the Scottish coast. ICES Council Meeting (C.M.) document, 1994/N:11. [Presented as a poster.] 2 Santos, M. B., G. J. Pierce, G. Wijnsma, H. M. Ross, and R. J. Reid. 1995. Diets of small cetaceans stranded in Scot- land 1993-1995. ICES Council Meeting (C.M.) document, 1995/N:6. individual from Kennebunk, Maine (Table 1 and Fig. 1). We obtained skin tissue from each bycaught speci- men and conducted DNA analysis at the NOAA South- east Fisheries Science Center to confirm that each ani- mal was in fact a Sowerby’s beaked whale. DNA was extracted from the tissue through the use of standard proteinase K digestion followed by organic extraction (Rosel and Block, 1996). The quality of the DNA was assessed through agarose gel electrophoresis, and DNA quantity was measured with a fluorometer (Amersham Biosciences3, now GE Healthcare Life Sciences, Little Chalfont, UK). To confirm field identifications on the basis of mor- phology, the 5’-end of the mitochondrial DNA control region was amplified and sequenced as described in Sellas et al. (2005). Resultant DNA sequences were identified to species through phylogenetic reconstruc- tion with an alignment that contained the new con- trol region sequences and the sequences obtained from the 5 species of beaked whales present in the western North Atlantic. Mesoplodont whales form strongly sup- ported clades in phylogenetic analyses of control region sequences; therefore, this method is well suited to spe- cies identification of unknown samples (Henshaw et al., 1997; Dalebout et al., 2004). The unusual stomach anatomy of beaked whales has been described in detail by Mead (1989, 1993, 2007). We examined the contents of the esophagus and upper digestive tract, including the fore stomach, main stom- ach, connecting chambers, and pyloric stomach. We fol- lowed a standard protocol for analysis of stomach con- tents (see Craddock et al., 2009), separating hard parts from the remaining digesta by elutriation and then decanting them through a sieve with a 0.5-mm mesh. We then sorted, dried, and identified all hard parts to the lowest possible taxonomic level. Certain diagnostic bones of fishes (e.g., otoliths, dentaries, premaxillar- ies, and maxillaries) were stored separately from other hard parts. Squid beaks and all parasites were counted and preserved in 70% ethanol. We archived the con- tents of each stomach separately. We identified the hard parts of prey items through the use of published guides (Roper et al., 1984; Clarke, 1986; Harkonen, 1986; Vecchione et ah, 1989; Campa- na, 2004) and the otolith and skeletal bone reference collection prepared by J. E. Craddock at the Woods Hole Oceanographic Institution (WHOI). This collec- tion is now part of the ichthyology collection of the Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts (http://www.mcz. harvard. edu/Departments/Ichthyology/researchcoIl.html, ac- cessed May 2013). We estimated the number of fish prey using half the number of otoliths when more than 50 otoliths were present. When fewer than 50 otoliths were present, we 3 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. Wenzel et al.: Food habits of Mesopiodon bidens in the western North Atlantic 383 Table I Origin and description of Sowerby’s beaked whales (Mesopiodon bidens) obtained in the western North Atlantic between Au- gust 1989 and October 2003. Most whales were retained as bycatch in the pelagic drift gillnet fishery for Swordfish (Xiphias gladius). One whale was collected stranded in Kennebunk, Maine. Latitudes and longitudes are given in decimal degrees. NA=not available; M=male; F=female. Wt. of stomach Whale contents identification Latitude Longitude Depth (m) Year Month Day Source Sex Length (cm) (g) D00253 40.23 67.90 1050 1989 10 10 Drift gillnet M 491 1816 D00341 40.02 68.80 1200 1995 6 24 Drift gillnet F 485 5830 D01369 40.87 66.42 1600 1994 6 10 Drift gillnet M 462 5897 D01380 40.97 66.32 1900 1994 6 3 Drift gillnet F 460 4082 D03070 40.03 68.77 1350 1996 7 4 Drift gillnet F 476 2700 D03202 40.35 67.35 1350 1996 7 6 Drift gillnet M 470 2650 D03458 40.03 68.63 1600 1996 7 4 Drift gillnet F 471 NA D03486 40.97 66.40 750 1994 7 10 Drift gillnet F 495 4082 D06061 40.78 66.57 1250 1990 8 9 Drift gillnet F 274 Empty F00120 40.87 66.48 1150 1989 8 26 Drift gillnet M 473 Partial MH03-604 43.33 70.52 NA 2003 10 2 Stranding M 442 Empty counted the maximum number of either right or left otoliths for each fish species. We assessed relative prey importance by frequency of occurrence (FO) and pro- portion of numerical abundance (%N). FO is the pro- portion of stomachs that contained a particular type of prey, %N is the proportion that each prey type rep- resented of the total number of prey items recovered, and the minimum number of fish, is determined by the number of paired otoliths found in each stomach, with any odd numbered otolith raising the minimum num- ber of fish by one (Table 2). We measured whole undi- gested otoliths from abundant fish prey with a stage micrometer or vernier calipers and estimated prey sizes with linear regressions derived from the WHOI reference collection (Table 3). We estimated the num- ber of individual cephalopod prey from the maximum number of either upper or lower beaks (Table 4). Results One of the 10 stomachs which we examined was from a calf and contained only mucous or milk, and the stom- ach of the stranded individual was completely empty. In addition to the 10 stomachs examined for this study, another stomach was dissected and examined at sea by a NEFOP observer who retained only 14 otoliths: 9 Marlin-spike ( Nezumia bairdii) and 5 Cocco’s Lantern- fish ( Lobianchia gemellarii). Because of the incomplete examination of the stomach contents of this individual, we did not include it in the quantitative analysis of food habits. The remaining 8 stomachs were intact and contained prey; therefore we used the contents of these stomachs in our quantitative analysis of the frequency, numerical abundance, and size of prey. Genetic analy- sis confirmed that these stomachs were all from Sow- erby’s beaked whales. Fishes dominated the diet of these whales; in to- tal, we recovered 9451 otoliths of fishes and jaws of 18 Sloan’s Viperfish (Chauliodus sloani ). The only prey represented by more jawbones than otoliths was Sloan’s Viperfish, The mean number of otoliths per stomach was 1196 (range: 327-3452). The recovered otoliths represented at least 31 species from 15 fami- lies of deepwater fishes (Table 2). Fishes from the fami- lies Moridae (%N=37.9% of prey), Myctophidae (22.9%), Macrouridae (11.2%), and Phycidae (7.2%) were pres- ent in all 8 stomachs. Most (74.1%) prey were from the following 5 taxa, ordered by proportion of numerical abundance: 1) Shortbeard Codling ( Laemonema bar- batulum), Moridae, 35.3%; 2) Cocco’s Lanternfish, Myc- tophidae, 12.9%; 3) Marlin-spike, Macrouridae, 10.8%; 4) lanternfishes ( Lampanyctus spp. ), Myctophidae, 8.4%; and 5) Longfin Hake ( Phycis ehesteri), Phycidae, 6.7%. In each stomach, 12-19 different fish taxa were present, with a mean of 15. The estimated standard lengths of fish prey ranged from 4.0 cm to 27.7 cm (Ta- ble 3). In its esophagus, whale DO 1380 had 13 whole Cocco’s Lanternfish, ranging in length from 8.0 to 10.0 cm (mean: 9.2 cm), similar to lengths of fish estimated from otoliths of this species found in the other 7 stom- achs examined (Table 3). Squid remains were found in 7 of the 8 analyzed stomachs, but they were represented by only 123 beaks (minimum 73 individuals) of 3 identified taxa (Table 4); cephalopods accounted for only 1.5% of total prey. The mean number of squid beaks per stomach was 15.4 (range: 0.0-35.0). 384 Fishery Bulletin 111(4) 72°W 71 "W 70°W 69"W 68" W 67" W 66”W Figure 1 Map of locations where 10 specimens of the Sowerby’s beaked whale ( Mesoplodon bidens ) were taken in the pelagic drift gillnet fishery for Swordfish ( Xiphias gladius ) in the western North Atlantic between August 1989 and July 1996. The stomach contents of 8 of these whales were examined to determine the food habits of this species. Discussion The Sowerby’s beaked whales that we examined had been feeding primarily on large numbers of relative- ly small meso- and benthopelagic fishes before their death; cephalopod prey constituted a very minor part of the diet of these animals. Our findings are simi- lar to those of Pereira et al. (2011), who examined the stomach contents of 10 stranded Sowerby’s beaked whales from the Azores and found a predominance of small fish prey. It is important to note that all but one of the whales that we examined were killed at sea, and, with the exception of the single stranded animal, they were apparently healthy at the time of their death. The presence of intact prey in the esopha- gus of one specimen and the large numbers of prey items in the stomachs that we examined indicate that these animals had been foraging before death. The av- erage minimum number of prey in the stomachs that we examined was more than 600 (4789 fishes, plus 73 squids, in all 8 stomachs combined), compared with 85 prey in the stranded specimens examined by Pereira et al. (2011). We believe the stomach contents of the whales that we examined are representative of the summer diet of Sowerby’s beaked whales along the continental shelf break off the northeastern coast of the United States. Nevertheless, biases from several sources could affect our conclusions. For example, our analysis of the stom- ach contents of these bycaught cetaceans could have been biased if these whales had been feeding in or around a fishing gear. Such behavior has not been re- ported for beaked whales, however, and the Sowerby’s beaked whales that we examined were taken in pelagic drift gillnets that targeted large Swordfish and tunas in the top 10 m of the water column. These large-mesh gillnets could not have captured the prey we identi- fied in stomachs of Sowerby’s beaked whales; the only whole fish we recovered were 13 small (<10.0 cm) Coc- co’s Lanternfish found in one esophagus. Therefore, the Sowerby’s beaked whales that we examined were not actively feeding in nets when captured. Wenzel et al.: Food habits of Mesoplodon Dicers in the western North Atlantic 385 Table 2 Analysis of fish prey identified in stomachs of Sowerby’s beaked whales (Mesoplodon bidens ) taken in the pelagic drift gillnet fishery for Swordfish ( Xiphias gladius ) in the western North Atlantic between August 1989 and July 1996. %FO= percentage of frequency of occurrence; %N=percentage of number of otoliths. Unidentified means that the structure of the otoliths was distinct for identification but the otoliths were not identified. Unidentifiable otoliths were worn or digested and not identifiable. Family Species Common name Occurrence (no. of stomachs) %FO Number of otoliths %N Minimum number of fish Alepocephalidae Alepocepkalus cf. agassizii Agassiz’s Smoothhead 4 50 10 0.1 6 Diretmidae Diretmus argenteus Spinyfin 1 13 2 0.0 1 Gonostomatidae Gonostoma elongatum Longtooth Anglemouth 2 25 6 0.1 4 Macrouridae Coelorinchus sp. Grenadier 1 13 4 0.0 2 Macrouridae Coryphaenoides sp. Grenadier 3 38 39 0.4 21 Macrouridae Nezumia bairdii Marlin-spike 8 100 1019 10.8 515 Melamphaidae Poromitra capita Ridgehead 1 13 3 0.0 2 Melamphaidae Scopelogadus beanii Bean’s Bigscale 7 88 344 3.6 178 Merlucciidae Merluccius albidus Offshore Hake 1 13 1 0.0 1 Merlucciidae Merluccius bilinearis Silver Hake 2 25 311 3.3 156 Moridae Gadella imberbis Beardless Codling 3 38 65 0.7 33 Moridae Laemonema barbatulum Shortbeard Codling 8 100 3332 35.3 1672 Moridae Unidentified morid Codling 1 13 178 1.9 89 Myctophidae Bentkosema glaciale Glacier Lanternfish 3 38 5 0.1 3 Myctophidae Bolinichthys supralateralis Stubby Lanternfish 1 13 2 0.0 1 Myctophidae Ceratoscopelus maderensis Horned Lanternfish 7 88 75 0.8 40 Myctophidae Hygophum hygomii Bermuda Lanternfish 4 50 6 0.1 4 Myctophidae Lampadena speculigera Mirror Lanternfish 7 88 36 0.4 19 Myctophidae Lampanyctus spp. Lanternfishes 8 100 797 8.4 403 Myctophidae Lobianchia gemellarii Cocco’s Lanternfish 8 100 1222 12.9 613 Myctophidae Nannobrachium cf. atrum Dusky Lanternfish 5 63 15 0.2 8 Paralepididae Arclozenus risso White Barracudina 8 100 90 1.0 49 Paralepididae Unidentified paralepidid 2 25 4 0.0 2 Paraliehthyidae Paralichthys oblongus Fourspot Flounder 1 13 1 0.0 1 Phycidae Phycis chesteri Longfin Hake 7 88 634 6.7 321 Phyeidae Urophycis chuss Red Hake 2 25 30 0.3 15 Phycidae Urophycis tenuis White Hake 2 25 8 0.1 4 Scorpaenidae Helicolenus dactylopterus Blackbelly Rosefish 7 88 350 3.7 179 Serrivomeridae Serrivomer beanii Stout Sawpalate 1 13 3 0.0 2 Sternoptychidae Polyipnus clarus Slope Hatchetfish 1 13 2 0.0 1 Stomiidae Chauliodus cf. sloani Sloan’s Viperfish 4 50 15 0.2 9 Unidentified otoliths 8 100 215 2.3 114 Unidentifiable otoliths 8 100 627 6.6 318 Total otoliths 9451 Stomiidae Ckauliodus spp. Viperfish jaws 2 25 18 3 Total fishes 4789 A second potential bias arises because hard parts of different prey may pass through the gastrointesti- nal tract at different rates. For example, squid beaks are resistant to digestion and often accumulate in stomachs of marine mammals, but the soft tissue and bones of fishes are more readily digested (Bigg and Fawcett, 1985). The complex structure of beaked whale stomachs (Mead 1989, 1993, 2007) makes it likely that relatively indigestible squid beaks are re- tained for prolonged periods. Therefore, the results reported here may overestimate the already low im- portance of cephalopods in the diet of the Sowerby’s beaked whale. A third possible bias arises from the secondary in- gestion of prey, in which recovered hard parts enter whales in the stomachs of prey and are not consumed directly by whales themselves. It is difficult to evalu- ate this potential source of bias. It is possible, for ex- ample, that the Horned Lanternfish ( Ceratoscopelus maderensis) we recovered could have been secondarily introduced into the stomachs of the Sowerby’s beaked whales that we examined, given the small size of the 386 Fishery Bulletin 1 1 1 (4) Table 3 Habitat and size of the most abundant prey species found in stomachs of Sowerby’s beaked whales ( Mesoplodon bidens) taken in the pelagic drift gillnet fishery for Swordfish ( Xiphias gladius) in the western North Atlantic between August 1989 and July 1996. Habitat/depths are taken from Fishbase (http://www.fishbase.org/search.php, accessed June 2013). Standard length was used to measure fish lengths. i?2=coefficient of multiple determination; NA=not available Prey species Habitat Diurnal migrant Depth range (m) Number measured Mean otolith length (cm) Otolith length range (cm) Otolith length (OL) - fish length regression Mean length of prey (cm) with range Laemonema barbatulum Benthopelagic No 50-1620 136 0.4 0.2-0. 6 NA NA Lobianchia gemellarii Mesopelagic Yes 200-800 140 0.7 0.6-0. 8 Fish length = 0.06430L + 1.0482 R2= 0.9799 9.7 (8-11) Nezumia bairdii Benthopelagic No 90-700 198 0.8 0 1 o Fish length = 0.0350L- 0.0961 i?2=0.9348 22 (13-28) Lampanyctus spp. Mesopelagic Yes 40-1000 127 0.3 0.2-0. 4 NA NA Phycis chesteri Benthopelagic No 90-1400 74 1 0.6-1. 3 NA NA Helicolenus daetylopterus Bathydemersal No 50-1100 40 0.3 0.2-0. 6 NA NA Scopelogadus beanii Meso- to bathypelagic Yes 400-1000 38 0.3 0.2-0. 4 Fish length = 0.01690L + 1.6267 i?2=0.5816 8.1 (5-11) Arctozenus risso Mesopelagic No 200-1000 43 0.4 0.3-0. 4 Fish length = 0.00860L + 1.6552 R2= 0.8359 21.5 (10-27) Ceratoseopelus Mesopelagic Yes 330-600 24 0.3 0.2-0. 3 Fish length = 5.4 (4-6) maderensis 0.05110L - 0.1393 fl2=0.9454 fish we recovered (4. 0-6. 3 cm) and their low numbers. This fish species is generally abundant and found in schools in the deep scattering layer (DSL) along the shelf break on the southern side of Georges Bank (Backus et al., 1968) where the Sowerby’s beaked whales that we examined were taken. The mechanisms by which the Sowerby’s beaked whale locates and captures prey are largely unknown. All whales in the genus Mesoplodon have relatively small mouths and few teeth, and they are believed to employ suction while feeding (Mead et al. 1982; Heyn- ing and Mead, 1996). The Sowerby’s beaked whale has 2 teeth that erupt only in sexually mature males (Mead, 1989; Heyning and Mead, 1996). The relatively small mouth and 2 teeth of this species may explain why Sowerby’s beaked whales typically are found to have only small prey items in their stomachs. Studies that employed digital acoustic tags on other beaked whales in this genus have provided brief but exceptionally rich glimpses into the foraging behav- ior of these animals. For example, tagged Blainville’s beaked whales in the Canary Islands have been report- ed to feed on prey in the lower part of the DSL and within the benthopelagic zone (Arranz et al., 2011). Almost half of the attempts at prey capture made by these whales in the Canary Islands occurred in the benthic boundary layer, reinforcing the importance of benthopelagic prey for them. In addition, these Blain- ville’s beaked whales appeared to focus on the oxygen minimum layer just below the DSL in areas of steep topography. Johnson et al. (2008) described the behav- ior of a tagged Blainville’s beaked whale in the Baha- mas that appeared to provoke a schooling reaction in mesopelagic prey that resulted in a school of prey up to 4 m in diameter and created an opportunity for the whale to more easily capture those prey. Until a tag is deployed for the deep-diving Sowerby’s beaked whale, we can only speculate about the foraging behavior of Wenzel et al.: Food habits of Mesoplodon bidens in the western North Atlantic 387 Table 4 Cephalopod prey from the stomachs of Sowerby’s beaked whales ( Mesoplodon bidens) taken in the pelagic drift gillnet fish- ery for Swordfish ( Xiphias gladius) in the western North Atlantic between August 1989 and July 1996. %N=percentage of number of total beaks. %FO=percentage of frequency of occurrence, on the basis of the number of stomachs studied. Prey item D00253 D00341 D01369 D01380 D03070 D03202 D03458 D03486 Total %N %FO Unidentified upper beaks 0 2 2 4 10 19 12 49 39.8 87.5 Unidentified lower beaks 0 5 8 16 29 23.6 50.0 Histioteuthis spp. 4 13 7 9 33 26.8 50.0 Taonius pavo 9 2 11 8.9 25.0 Chiroteuthis veranyi 1 1 0.8 12.5 Total beaks 14 0 15 9 9 18 35 23 123 100.0 Total cephalopods 73 this species, but the results presented here indicate that their hunting strategies may be similar to those of their better-studied congener. Therefore, on the basis of knowledge of the habi- tat preferences of prey recovered from the stomachs of Sowerby’s beaked whales, we conclude that these animals feed in the meso- and benthopelagic environ- ments along the shelf break, foraging in the water col- umn and near the seafloor. Mesopelagic fishes in this region are important prey for several other cetacean species. Horned Lanternfish, in particular, is consumed by the Atlantic white-sided dolphin ( Lagenorhynchus acutus ) (Craddock et al., 2009) and by the common dolphin ( Delpliinus delphis), both of which are also caught incidentally in the pelagic drift gillnet fishery for Swordfish in the Atlantic (Craddock and Polloni4). The stomach of a harbor porpoise ( Phocoena phocoena), captured in a pelagic drift net fishery off North Caro- lina was found to contain more than 1900 otoliths of Horned Lanternfish (Read et ah, 1996). Many marine organisms are concentrated in oceano- graphic frontal zones, as a result of increased produc- tion and advection (Jahn and Backus, 1976; Backus et al., 1977; Olson and Backus, 1985). As a consequence of these aggregations, predators (including swordfish) and fishermen exploit fronts. The mosaic of oceanic fronts associated with the Gulf Stream and its warm- and cold-core rings have long been targeted by fishermen of Swordfish, particularly along the shelf break (Smith et al., 1996). Swordfish have been reported to feed on some of the same prey items that we recovered from 4 Craddock J. E., and P. T. Polloni. 2005. Food habits of small marine mammals from the Gulf of Maine and from slope water off the northeast US coast. Year 3, Final Re- port, revised, 31 p. Request no. EA 133F-02-RQ-0081. Req- uisition no. NFFM7320-2-15375. [Available from Northeast Fisheries Science Center, National Marine Fisheries Service, NOAA, 166 Water St., Woods Hole, MA 02543 http://www. nefsc.noaa.gov/publications/reports/EA 133F02RQ0081.pdf] Sowerby’s beaked whales (Scott and Tibbo, 1968; Still- well and Kohler, 1985). For example, barracudinas (Paralepididae) are im- portant food items for Swordfish in the northwestern Atlantic (Scott and Tibbo, 1968) and were common prey of the Sowerby’s beaked whales that we exam- ined; White Barracudina ( Arctozenus risso ) is the most common barracudina in this region (Moore et al., 2003). Lanternfishes (Myctophidae) also are consumed by Swordfish in large numbers, but, because of their rela- tively small size, they do not contribute significantly to the mass ingested by these predators (Scott and Tibbo, 1968). The pelagic drift gillnet fishery in the Atlantic targeted Swordfish and tunas, and the fishing effort focused on thermal fronts along the shelf break, as de- scribed by Podesta et al. (1993). Therefore, the com- mon prey fields and habitats of Swordfish and Sow- erby’s beaked whales may help to explain the relatively high bycatch rates of Sowerby’s beaked whales in this fishery. Conclusions The diet of Sowerby’s beaked whales in the western North Atlantic is dominated by meso- and benthope- lagic fishes (98.5%), and cephalopods accounted for only 1.5% of their prey. Future research with digital acoustic tags would be helpful to examine the diving and echolocation behavior of Sowerby’s beaked whales in relation to the vertical and horizontal distribution of prey. A study that combines both the tagging methods used by Arranz et al. (2011) and survey data of the prey field documented with the use of scientific echo- sounders and by direct capture of voucher specimens would be particularly profitable. The regular occur- rence of Sowerby’s beaked whales in and near the can- yons on the southern margin of Georges Bank, where the whale specimens we studied were captured, offers a promising field opportunity for such research. 388 Fishery Bulletin 1 1 1 (4) Acknowledgments We dedicate this paper to the memory of coauthors James Craddock and John Nicolas, who were instru- mental in this study. We thank J. Galbraith, T. Sutton, and the Northeast Fisheries Science Center for provid- ing fish specimens to add to the reference collection; E. Josephson and H. J. Foley for providing maps; B. Hay- ward and M. Moore for assistance with sorting stomach contents; K. Hartel and C. Kenaley of the Museum of Comparative Zoology, Harvard University, for refer- ence material and additional otolith measurements. We also recognize K. Hartel, D. Waples, F. Serchuk, M. Simpkins, G. Waring, and T. Fenster and 3 anonymous reviewers for the useful comments that helped to im- prove this manuscript. Our work was made possible by the dedication of observers from NEFOP and the co- operation of pelagic drift net fishermen. We thank our colleagues aboard the Abel-J for their assistance in the field. This work was funded by the Northeast Fisheries Science Center, Woods Hole, Massachusetts. Literature cited Arranz, P., N. Aguilar de Soto, P. T. Madsen, A. Brito, F. Bordes, and M. P. Johnson. 2011. Following a foraging fish-finder: diel habitat use of Blainville’s beaked whales revealed by echoloca- tion. PLoS ONE 6(12):e28353. doi:10. 1371/journal, pone. 0028353. Backus, R. H., J. E. Craddock, R. L. Haedrich, and B. H. Robison. 1977. Atlantic mesopelagic zoogeography. In Fishes of the Western North Atlantic, part 7. (R. H. Gibbs, ed.), p. 266-287. Foundation Mar. Res., Yale Univ., New Ha- ven, CT. Backus, R. H., J. E. Craddock, R. L. Haedrich, D. L. Shores, J. M. Teal, A. S. Wing, G. W. Mead, and W. D. Clarke. 1968. Ceratoscopelus maderensis: peculiar sound-scatter- ing layer identified with this myctophid fish. Science 160:991-993. 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. Beddington, R. J. H. Beverton, and D. M. Lavigne, eds), p. 284-291. George Allen & Unwin, Ltd., London. Bonde, R. K., and T. J. O’Shea. 1989. Sowerby’s beaked whale ( Mesoplodon bidens ) in the Gulf of Mexico. J. Mammal. 70:447-449. Campana, S. 2004. Photographic atlas of fish otoliths of the northwest Atlantic Ocean, 284 p. Can. Spec. Publ. Fish. Aquat. Sci. 133. Carlstrom, J., J. Denkinger, P. Feddersen, and N. Oien. 1997. Record of a new northern range of Sowerby’s beaked whale ( Mesoplodon bidens ). Polar Biol. 17:459-461. Clarke, M. R. 1986. A handbook for the identification of cephalopod beaks, 273 p. Oxford Univ. Press, Oxford. Craddock, J. E., P. T. Polloni, B. Hayward, and F. Wenzel. 2009. Fishes and cephalopods in the diet of Atlantic white-sided dolphins ( Lagenorhynchus acutus) off the coast of New England. Fish. Bull. 107:384-394. Dalebout, M. L., C. S. Baker, J. G. Mead, V. G. Cockcroft, and T. K. Yamada. 2004. A comprehensive and validated molecular tax- onomy of beaked whales, family Ziphiidae. J. Hered. 95:459-473. Dix, L., J. Lien, and D. E. Sergeant. 1986. A North Sea beaked whale, Mesoplodon bidens, in Conception Bay, Newfoundland. Can. Field-Nat. 100:389-391. Griffin, R. B. 1999. Sperm whale distributions and community ecol- ogy associated with warm-core ring off Georges Bank. Mar. Mamm. Sci. 15:33-51. Harkonen, T. 1986. Guide to the otoliths of the bony fishes of the Northeast Atlantic, 256 p. Danbiu ApS, Biological Con- sultants, Hellerup, Denmark. Henshaw, M. D., R. G. LeDuc, S. J. Chivers, and A. E. Dizon. 1997. Identifying beaked whales (Family Ziphiidae) using mtDNA sequences. Mar. Mamm. Sci. 13:487-495. Heyning, J. E., and J. G. Mead. 1996. Suction feeding in beaked whales: morphological and observational evidence. Nat. Hist. Mus. Los Ang. Cty. Contrib. Sci. 464, 12 p. Hooker, S. K„ and R. W. Baird. 1999. Observations of Sowerby’s beaked whale, Mesoplo- don bidens in the Gully, Nova Scotia. Can. Field-Nat. 113:273-277. Jahn, A. E., and R. H. Backus. 1976. On the mesopelagic faunas of the Slope Water, Gulf Stream, and northern Sargasso Sea. Deep Sea Res. 23:223-234. Johnson, M., L. S. Hickmott, N. Aguilar Soto, and P. T. Madsen. 2008. Echolocation behaviour adapted to prey in forag- ing Blainville’s beaked whale (Mesoplodon densirostris). Proc. R. Soc. Lond., Ser. B: Biol. Sci. 275:133-139. Lien, J., and F. Barry. 1990. Status of Sowerby’s Beaked Whale, Mesoplodon bidens, in Canada. Can Field-Nat. 104:125-130. Lien, J., F. Barry, K. Breeck, and U. Zuschlag. 1990. Multiple strandings of Sowerby’s Beaked Whale, Mesoplodon bidens, in Newfoundland. Can. Field-Nat. 104:414-420. Lucas, Z. N., and S. K. Hooker. 2000. Cetacean strandings on Sable Island, Nova Scotia, 1970-1998. Can. Field-Nat. 114:45-61. MacLeod, C. D., M. B. Santos, and G. J. Pierce. 2003. Review of data on diets of beaked whales: evidence of niche separation and geographic separation. J. Mar. Biol. Assoc. U.K. 83:651-665. Madsen, P. T., M. Johnson, N. Aguilar de Soto, W. M. X. Zim- mer, and P. Tyack. 2005. Biosonar performance of foraging beaked whales (. Mesoplodon densirostris). J. Exp. Biol. 208:181-194. McAlpine, D. F., and M. Rae. 1999. First confirmed reports of beaked whales, cf. Meso- plodon bidens and M. densirostris (Ziphiidae), from New Brunswick. Can. Field-Nat. 113:293-295. Wenzel et al.: Food habits of Mesoplodon bidens in the western North Atlantic 389 Mead, J. G. 1989. Beaked whales of the genus Mesoplodon. In Handbook of marine mammals, vol. 4: river dolphins and the larger toothed whales (S. H. Ridgway and R. Harrison, eds.), p. 349-430. Academic Press, San Di- ego, CA. 1993. The systematic importance of stomach anatomy in beaked whales. Int. Mar. Biol. Res. Inst. [Japan] IBI Reports 4:75-86. 2007. Stomach anatomy and use in defining systemic relationships of the Cetacean Family Ziphiidae (beaked whales). Anat. Rec. 290:581-595. Mead, J. G„ W. A. Walker and W. J. Houck. 1982. Biological observations on Mesoplodon carlhubbsi (Cetacea: Ziphiidae). Smithsonian Contributions to Zo- ology 344:iii + 1-25. Moore, J. A., K. E. Hartel, J. E. Craddock, and J. K. Galbraith. 2003. An annotated list of deepwater fishes from off the New England Region with new area records. North- east. Nat. 10:159-248. Olson, D. B., and R. H. Backus. 1985. The concentrating of organisms at fronts: a cold- water fish and a warm-core Gulf Stream ring. J. Mar. Res. 43:113-137. Ostrom, P. H., J. Lien, and S. A. Macko. 1993. Evaluation of the diet of the Sowerby’s beaked whale, Mesoplodon bidens, based on isotopic compari- sons among northwestern Atlantic cetaceans. Can. J. Zool. 71:858-861. Pereira, J. N., V. C. Neves, R. Prieto, M. A. Silva, I. Cascao, C. Oliveira, M. J. Cruz, J. V. Medeiros, J. P. Barreiros, F. M. Porteiro, and D. Clarke. 2011. Diet of mid-Atlantic Sowerby’s beaked whales, Me- soplodon bidens. Deep Sea Res. (I Oceanogr. Res. Pap.) 58:1084-1090. Podesta, G. P., J. A. Browder, and J. J. Hoey. 1993. Exploring the association between swordfish catch rates and thermal fronts on U.S. longline grounds in the western North Atlantic. Cont. Shelf Res. 13:253-277. Read, A. J., J. R. Nicolas, and J. E. Craddock. 1996. Winter capture of a harbor porpoise in a pelagic drift net off North Carolina. Fish. Bull. 94:381-383. Roper, C. F. E., M. J. Sweeney, and C. E. Nauen. 1984. FAO species catalogue. Vol. 3. Cephalopods of the world. An annotated and illustrated catalogue of species of interest to fisheries. FAO Fish. Synop. 125, vol. 3, 277 p. FAO, Rome. Rosel, P. E., and B. A. Block. 1996. Mitochondrial control region variability and global population structure in the swordfish, Xiphias gladius. Mar. Biol. 125:11-22. Scott, W. B., and S. N. Tibbo. 1968. Food and feeding habits of swordfish, Xiphias gladius, in the western North Atlantic. J. Fish. Res. Board Can. 25:903-919. Sellas, A. B., R. S. Wells, and P. E. Rosel. 2005. Mitochondrial and nuclear DNA analyses reveal fine scale geographic structure in bottlenose dolphins ( Tursiops truncatus) in the Gulf of Mexico. Conserv. Genet. 6:715-728. Smith, T. D., R. B. Griffin, G. T. Waring, and J. G. Casey. 1996. Multispecies approaches to management of large marine predators. In The Northeast Shelf ecosystem: assessment, sustainability, and management (K. Sher- man, N. A. Jaworski and T. J. Smayda, eds.), p. 467-490. Blackwell Science, Cambridge, MA. Spitz, J., Y. Cherel, S. Bertin, J. Kiszka, A. Dewez, and V. Ridoux. 2011. Prey preferences among the community of deep- diving odontocetes from the Bay of Biscay, North- east Atlantic. Deep Sea Res. (I Oceanogr. Res. Pap.) 58:273-282. Stillwell, C. E„ and N. E. Kohler. 1985. Food and feeding ecology of the swordfish Xiphi- as gladius in the western North Atlantic Ocean with estimates of daily ration. Mar. Ecol. Prog. Ser. 22:239-241. Vecchione, M., C. F. E. Roper, and M. J. Sweeney. 1989. Marine flora and fauna of the eastern United States: Mollusca: Cephalopoda. NOAA Tech. Rep. NMFS 73, 23 p. Waring, G. T., T. Hamazaki, D. Sheehan, G. Wood, and S. Baker. 2001. Characterization of beaked whale (Ziphiidae) and sperm whale ( Physeter macrocephalus) summer habitat in shelf-edge and deeper waters off the Northeast U.S. Mar. Mamm. Sci. 17:703-717. Waring, G. T., E. Josephson, K. Maze-Foley, and P.E. Rosel (eds.). 2009. U.S. Atlantic and Gulf of Mexico marine mammal stock assessments — 2009. NOAA Tech. Memo. NMFS- NE-213, 528 p. 2010. U.S. Atlantic and Gulf of Mexico marine mammal stock assessments — 2010. NOAA Tech. Memo. NMFS- NE-219, 595 p. Waring, G. T., J. M. Quintal, and C. P. Fairfield (eds.). 2002. U.S. Atlantic and Gulf of Mexico marine mammal stock assessments — 2002. NOAA Tech. Memo. NMFS- NE-169, 318 p. Waring G. T., J. M. Quintal, and S. L. Swartz (eds.). 2000. U.S. Atlantic and Gulf of Mexico marine mammal stock assessments — 2000. NOAA Tech. Memo. NMFS- NE-162, 298p. 390 Abstract— Gonadal morphology and reproductive biology of the Black Anglerfish ( Lophius budegassa ) were studied by examining 4410 speci- mens collected between June 2007 and December 2010 in the north- western Mediterranean Sea. Ovaries and testes presented traits common among fishes of the order Lophi- iformes. Spawning occurred between November and March. Size at first maturity (L50) was 33.4 cm in total length (TL) for males and 48.2 cm TL for females. Black Anglerfish is a total spawner with group-synchro- nous oocyte development and de- terminate fecundity. Fecundity val- ues ranged from 87,569 to 398,986 oocytes, and mean potential fecun- dity was estimated at 78,929 (stan- dard error of the mean [SE] 13,648) oocytes per kilogram of mature fe- male. This study provides the first description of the presence of 2-3 eggs sharing the same chamber and a semicystic type of spermatogen- esis for Black Anglerfish. This new information allows for a better un- derstanding of Black Anglerfish re- production— knowledge that will be useful for the assessment and man- agement of this species. Manuscript submitted 19 October 2012. Manuscript accepted 19 September 2013. Fish. Bull. 111:390-401 (2013). doi: 10.7755/FB. 11 1.4.8 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necesarily reflect the position of the National Marine Fisheries Service, NOAA. Reproductive biology of Black Anglerfish C Lophius budegassa) in the northwestern Mediterranean Sea Ana I. Colmenero (contact author) Victor M. Tuset Laura Recasens Pilar Sanchez Email address for contact author: colmenero@icm.csic.es Institut de Ciencies del Mar (ICM) Consejo Superior de Investigaciones Cientificas (CSIC) Passeig Maritim de la Barceloneta 37-49 08003 Barcelona, Spain Lophius, a genus commonly known as anglerfishes or monkfishes, includes 7 species broadly distributed and ex- ploited worldwide. Most of these spe- cies inhabit the northwestern Atlan- tic, as do Goosefish ( Lophius ameri- canus ) and Blackfin Goosefish (L. gastropliysus ), or the northeastern Atlantic, as do Cape Monk (L. vome- rinus), Shortspine African Angler (L. vaillanti), Black Anglerfish (L. bude- gassa), and White Anglerfish (L. pis- catorius), although Black Anglerfish and White Anglerfish also live in the Mediterranean Sea and Yellow Goosefish (L. litulon ) can be found only in the northwestern Pacific (Farina et al., 2008). In the past, spe- cies of Lophius have been captured as bycatch in mixed fisheries, but an increase in their economic value, together with the overexploitation of other groundfish species, has led to the development of targeted an- glerfish fisheries (Hislop et al., 2001). In the northwestern Mediterranean Sea, landings of Black Anglerfish and White Anglerfish have accumu- lated to just over 6000 metric tons during the last 10 years.1 The scarce reproductive information available for these species does not allow for a 1 Tudo Vila, P. 2012. Unpubl. data. Directorate of Fishing and Maritime Af- faires, Government of Catalonia, Avin- guda Diagonal, 523-525, 08029 Barce- lona, Spain. proper assessment or informed man- agement of anglerfish fisheries. This study focuses on Black An- glerfish, a demersal fish distributed along the Mediterranean Sea, as well as the northeastern Atlantic from the British Isles to Senegal (Caruso, 1986). This species is found over the continental shelf and upper slope at depths of up to 800 m and inhabits sandy, muddy, and rocky bottoms (Carlucci et al., 2009). They occupy the water column as eggs and larvae, and then they shift to a benthic exis- tence as juveniles and adults (Farina et al., 2008). This species co-occurs with White Anglerfish over all its bathymetric range, although White Anglerfish has a deeper distribu- tion that reaches to depths >1000 m (Afonso-Dias and Hislop, 1996). De- spite the overlapping distributions of these species, Colmenero et al. (2010) concluded that no ecological compe- tition exists between these species because of a temporal segregation in their biorhythms; Black Angler- fish is more active at nighttime, and White Anglerfish is more active dur- ing daytime. Most studies of these species have been undertaken in northeast- ern Atlantic waters, and they often have dealt with age and growth (Dupouy et al., 1984; Landa et al., 2008a; Woodroffe et al., 2003), feed- ing habits (Crozier, 1985; Laurenson Colmertero et al.: Reproductive biology of Lophius budegassa in the northwestern Mediterranean Sea 391 Figure 1 Map of study area where specimens of Black Anglerfish (Lophius budegassa) were collected in the northwestern Mediterranean Sea between June 2007 and December 2010 to examine the reproductive biology of this species. The zone shaded in light gray closest to the Catalan coast indicates the sampling area. The cities shown on the coast were the ports of commercial trawl fishery vessels from which specimens were collected. and Priede, 2005; Preciado et al., 2006), geographical and depth distribution (Caruso, 1985; Landa et al., 2008b; Velasco et al., 2008), and reproduction (Afonso- Dias and Hislop, 1996; Duarte et al., 2001; Laurenson, 2006). In the Mediterranean Sea, studies have been less numerous and for the most part have focused on biological aspects similar to the ones examined in the studies just described (Carlucci et al., 2009; Colme- nero et al., 2010; Garcfa-Rodriguez et al., 2005; La Mesa and De Rossi, 2008; Maravelias and Papaconstantinou, 2003; Negzaoui-Garali and Ben Salem, 2008; Negzaoui- Garali et al., 2008; Tsimenidis, 1984; Tsimenidis and Ondrias, 1980; Ungaro et al., 2002). However, only Tsi- menidis (1980) and Carbonara et al. (2005) focused on reproductive traits of Black Anglerfish. Information about the duration of the spawning sea- son, fish size at first maturity, reproductive strategy, maturation of oocytes, and fecundity are very relevant for studies of the biology and population dynamics used in stock assessments for management of fishery resources. Of all these reproductive features, fecundi- ty is the most difficult biological parameter to obtain, although it is critical for accurate stock assessments (Trippel et al., 1997). In the peculiar case of species of Lophius, studies on fecundity are scarce because of the difficulty of 1) acquisition of suitably mature individu- als in the maturity phases of spawning capable or ac- tively spawning and 2) the use of a proper method for fecundity estimation, which is especially complicated because of the morphological features of the gonads of these species. During reproduction, a gelatinous mate- rial is secreted into the lumen, and enumeration and measurement of eggs embedded in this mucus matrix is extremely difficult. Although species of Lophius have similar traits throughout the world, some biological aspects and catch trends of fisheries present interspecific and spa- tial variations (Farina et al., 2008). For that reason, reproductive parameters of the Atlantic stock of Black Anglerfish cannot be applied to the stock in the Medi- terranean. In addition, knowledge of the reproductive biology of this species in the northwestern Mediterra- nean Sea is very limited. Therefore, our study is the first one to take a detailed approach to the examina- tion of reproductive traits of Black Anglerfish in the Mediterranean Sea, and the results of this study can contribute to improvement of stock assessment and ef- fective management of this species in this region. Materials and methods Sample collection Monthly samples of Black Anglerfish were obtained from 467 sampling stations situated in the fishing grounds off the Catalan coast in the northwestern Mediterranean Sea from 40°5.980'N to 43°39.310'N 392 Fishery Bulletin 1 1 1 (4) Table 1 Macroscopic and microscopic description of the 5 maturity phases in the reproductive cycle of male and female Black An- glerfish (Lophius budegassa ) collected from the northwestern Mediterranean Sea between June 2007 and December 2010; adapted from Afonso-Dias and Hislop (1996) and Brown-Peterson et al. (2011). Phase Males Females Immature (I) Testes are long, narrow, and tubular shaped. They are translucent with no visible vascularization. The medial semi- niferous duct is distinct. Only spermato- gonia and primary spermatocytes are present. Ovaries are very narrow, thin, and flattened-tube shaped. They are translucent; no oocyte clusters visible and minimal vascularization. Only oogonia and primary growth oocytes are present. Developing/ Regenerating (II) Testes are small with visible blood ves- sels around the seminal duct. Spermato- gonias, primary and secondary spermato- cytes are predominant. Spermatids are scarce. Ovaries are small. Still no noticeable individual oo- cyte clusters. They acquire a cream color and vas- cularization is visible. Only oogonia and primary growth oocytes are present. Spawning capable (III) Testes increase in length and width. They have a firm texture and cream color. Sem- inal duct is highly vascularized. Germ cells at all stages of spermatogenesis are present. Spermatids are predominant with a lot of spermatozoa in the lumen of the sperm duct. Ovaries increase in width and length. They have a light orange color, and blood vessels are prominent. The edges of the ovaries start to curl and they oc- cupy a larger proportion of the body cavity. A mucus matrix starts to develop. Primary growth, cortical alveolar, and primary and secondary vitellogenic oo- cytes are present. Actively spawning (IV) Testes are large and firm and have a creamy coloration. Large amounts of sperm produced when testes are dissect- ed. Abundant quantities of spermatozoa are present in the seminiferous tubules. Ovaries are extremely long and wide and occupy most of the body cavity. The color of the oocytes is or- ange, and they are visible macroscopically. Ovaries are characterized by the presence of large hyaline oocyte clusters enclosed in a transparent gelatinous matrix that is completely developed. High vascular- ization is present. Oocytes are in tertiary vitellogen- esis, migratory nucleus and hydration. Regressing (V) Testes are small, flaccid, and have brown or red areas in their beige surface. They are still highly vascularized. Sperm and residual spermatozoa can be found in the lumina of the sperm duct. Spermatogonia are present in the testes cortex. Ovaries are flaccid and highly vascularized and of- ten have longitudinal striations. Their color is dark pink or red. Atresia and postovulatory follicles, together with primary growth stages, are present. Cortical alveolar, primary and secondary vitellogen- esis can be found. and from 0°32.922'E to 3°35.718'E between June 2007 and December 2010. Specimens were collected onboard commercial trawl fishery vessels at depths of 20-600 m and identified according to Caruso (1986). The trawl fleet belonged to the ports of Roses, Blanes, Arenys de Mar, Vilanova i la Geltru, and Sant Carles de la Rapita (Fig. 1). For this study, 4410 specimens were measured to the nearest centimeter in total length (TL), weighed to the nearest gram in total weight (TW) and gutted weight (GW), and measured with an accuracy of 0.01 g in gonad weight (GNW) and liver weight (LW). Macroscopic and histological description of gonads Of the total number of specimens, 3562 fish had gonads removed and their sex was determined, and they were assigned macroscopically to a gonadal stage on the basis of a scale with 5 maturity phases that were described in previous studies: immature (phase I), developing or regenerating (phase II), spawning capable (phase III), actively spawning (phase IV), and regressing (phase V) (Afonso-Dias and Hislop, 1996; Brown-Peterson et ah, 2011) (Table 1). Sex was easily assessed macroscopical- ly in mature individuals. However, gonads from small individuals (approximately <20 cm TL) were indistin- guishable macroscopically because ovaries and testes were small, translucent, and string-like. Fish that were too small to determine their sex or assign to a gonadal phase were classified as indeterminate. To corroborate the macroscopic classification of some unclear and undetermined gonads, 372 speci- mens were histologically examined. They were fixed in 10% buffered formalin solution before they were dehy- drated and embedded in a methacrylate polymer resin. Colmenero et al.: Reproductive biology of Lophius budegassa in the northwestern Mediterranean Sea 393 Cross sections, each 3-4 pm thick, were made with a manual microtome Leica Reichert-Jung 2040 (Leica Microsystems,2 Wetzlar, Germany), stained with Lee’s stain (methylene blue and basic fuchsin), and mounted in a synthetic resin of dibutyl phthalate xylene (DPX) on microscope slides. Gonads were classified according to the morphological features and the presence of spe- cific inclusions (oil droplets, yolk granules, yolk vesi- cles, or postovulatory follicles) (Wallace and Selman, 1981). The ovarian and testicular phases were defined by the developmental stage of the most advanced cell within the gonad (Yoneda et al., 1998b). Spawning season and size at first maturity The spawning season was established from the analy- sis of the monthly variation of the maturity phases and the changes in gonadosomatic (GSI) and hepatosomat- ic (HSI) indices for each sex (Afonso-Dias and Hislop, 1996). Because immature specimens were not consid- ered, 1437 males and 1167 females were used to deter- mine both indices, which were calculated according to Yoneda et al. (2001) with the following equations: GSI = (GNW / GW) x 100. ( 1) HSI = ( LW / GW) x 100. (2) Size at first maturity (L50) was determined through the examination of males and females in mature phases (phase III, phase IV, or phase V) and immature indi- viduals collected during the spawning period (Duarte et al., 2001). Total length of all individuals was used to estimate L50, defined as the size at which 50% of all fish sampled were at sexually mature phases. Maturity curves were determined with a logistic curve (Pope et al., 1975): P = 100 / (1 + exp [a + 6TL]), (3) where P = the percentage of mature individuals as a function of size class (TL); and a and b = specific parameters that can change during the life cycle. A logarithmic transformation was applied to this equa- tion to calculate the parameters 0 and b by means of linear regression. Reproductive strategy and fecundity Patterns of ovarian organization and fecundity were tested by oocyte size-frequency distributions (West, 1990). For our analysis, 36 fish, with lengths between 20.0 and 72.5 cm TL, were randomly selected from all maturity phases. From these fish, 4428 oocytes — with more than 300 oocytes from each maturity phase (1=961; 11=1106; 111=1046; IV=381; V=934) — were mea- 2 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. sured for their diameter with an image analysis pro- gram (Image-Pro Plus, vers. 5.0, Media Cybernetics, Inc., Rockville, MD) in combination with an Axioskop 2 Plus microscope (Carl Zeiss Microscopy, LLC, Thorn- wood, NY) and a ProgRes C14 digital microscope cam- era (Jenoptik AG, Jena, Germany). Diameters were measured to the nearest 0.01 pm. The mean oocyte diameter by developmental stage was determined by calculating the diameter of all oocytes encountered in each subsample. Measurements were taken only of oo- cytes that were sectioned through the nucleus (Afonso- Dias and Hislop, 1996). Before fecundity was estimated, the gonads of 7 in- dividuals were divided into 3 sections (anterior, middle, and posterior) to test differences in mean oocyte den- sity within the ovary by using a one-way analysis of variance (ANOVA). This use of 3 sections ensured that the analyzed subsample represented the entire ovary (Murua et al., 2003). Batch fecundity (BF), the total number of mature eggs produced in a single spawn- ing batch by an individual female, was estimated by using the gravimetric method on the basis of the rela- tion between ovary weight and the density of oocytes in the ovary (Hunter and Goldberg, 1980). Three ovarian tissue samples of known weight, representing 10% of the total ovarian weight, were extracted from different areas of the same ovary (anterior, middle, and posterior ovarian lobe). These subsamples were collected from 15 specimens with ovaries in phases III and IV with nei- ther postovulatory follicles nor atretic oocytes present. Because the oocytes could not be extracted from their mucogelatinous matrix without destroying them, whole tissue subsamples were mounted on several slides for analysis and covered with a cover slip. Images of each ovarian tissue sample were taken with a Canon Powershot SD870 IS digital camera (Canon USA, Melville, NY), and oocytes were counted manually with Image-Pro Plus software. Fecundity val- ues were obtained by examining Black Anglerfish with total lengths of 46-65 cm, TW of 1096-5592 g, GW of 986-3600 g, and GNW of 88.70-2300 g. Batch fecun- dity for each female was calculated as a product of the number of secondary vitellogenic oocytes per unit of weight multiplied by the total ovarian weight (Yoneda et al., 2001). Relative batch fecundity (RBF), the total number of mature eggs released by a female during the spawning batch per gram of body weight of gutted fish, was calculated as BF divided by GW (Pavlov et al., 2009): BF = ( oocyte number / sampled GNW) x total GNW. (4) RBF = BF / GW. (5) Linear regression analysis was used to examine the relationships between BF and fish TL, TWT, and GW (Armstrong et al., 1992). Linear regression analy- sis also was applied to analyze the relationship be- tween RBF and TL. Mean potential fecundity was 394 Fishery Bulletin 1 1 1 (4) Figure 2 Micrographs of transverse sections of ovaries, oocytes inside the gelatinous matrix, and testes of Black Anglerfish ( Lophius budegassa) collected in the northwestern Mediterranean Sea between June 2007 and December 2010. The final 4 stages of ovary development are shown in the left column: (A) Phase II: developing or regenerating; (B) Phase III: spawning capable; (C) Phase IV: actively spawning; (D) Phase V: regressing. Oocytes are featured in the middle column: (E) 1 oocyte in a chamber (486 pm in diam- eter), (F) 2 oocytes floating in a chamber (427 pm and 359 pm in diameter), (G) 3 oocytes in a chamber (341 pm, 330 pm, and 322 pm in diameter), and (H) closed-up division between chambers. In the right column, transverse sections of testes show (I) its lobular organization, (J) empty lobules, (K) seminal lobule during spermatogenesis, and (L) spermatozoa in the lumen of the seminal lobule. Ov=ovigerous membrane; n=nucleus; Od=oil droplet; Yv=yolk vesicle; m=mucus matrix; Cn=chromatin nucleolar; Pn=perinucleolar; Pof=postovulatory follicle; l=seminal lobule; Sg=spermatogonia; Sc=spermatocyte; St=spermatid; Sz=spermatozoa. Scale bars=100 pm. Colmenero et al.: Reproductive biology of Lophius budegassa in the northwestern Mediterranean Sea 395 also calculated as the number of vitellogenic oocytes divided by kilogram of mature female (Murua et al., 2003). Results Gonadal morphology Ovarian structure consists of a flattened band with 2 distinctive lobes that are folded up and connected to each other at their posterior end. The lobes form a sin- gle organ attached to the abdominal cavity by a black mesenteric tissue called the mesovarium. One side of the ovarian wall is made of an ovigerous membrane and connective tissue. The nonovigerous side is made of epithelial cells. A single layer of oocyte clusters proj- ects from the ovigerous membrane to the lumen (Fig. 2A). Inside each gonad, the clusters can be in different development stages. Only the oocytes situated closest to the tip of the clusters have progressed through all maturity stages, and the other oocytes are only oogonia or in the primary growth stage (Fig. 2B). A gelatinous material is secreted into the lumen during the late phases of gonad maturation, producing the mucus matrix characteristic of the reproduction of Lophius species (Fig. 2C). Hydration of the oocytes oc- curs just before spawning, and postovulatory follicles (Fig. 2D) are found during the regression phase of the reproductive cycle. Ripe eggs, which are usually situat- ed on the tip of the oocyte cluster, rupture the follicles and are pressed into the layer of mucus. In this study, every chamber examined contained at least 1 egg in the gelatinous matrix (Fig. 2E), although the presence of 2 (Fig. 2F) or 3 eggs (Fig. 2G) floating in separate chambers was also noted (Fig. 2H). Oocyte diameter appeared to differ depending on the quantity of oocytes present in each chamber; the oo- cytes that were isolated in their chambers were found to be larger in size than other oocytes. A diameter of 486 pm was obtained for the oocyte that was the single oocyte in its chamber. Diameters of 427 pm and 359 pm were found for the 2 oocytes floating together in a chamber, and diameters of 341 pm, 330 pm, and 322 pm were observed for the 3 oocytes that shared the same chamber. Measurements of more oocyte diameters are needed to confirm these preliminary observations. The testes are a pair of elongated and tubular struc- tures located in the dorsal portion of the abdominal cavity, and they are bean-shaped in transverse section. The organization of the testes is lobular: the connec- tive tissue extends from the testicular capsule to form lobules that have their blind ends on the surface of the gonads, converging ventrally towards the sperm duct (Fig. 21). These lobules are fused to the posterior end of each testicular lobe to form a common sperm duct that leads to a genital pore (Fig. 2J). Spermatogenesis takes place in a capsule-like sac called a cyst, but it is not completed within the cyst. Each cyst contains spermatogonia or developing spermatocytes (Fig. 2K). Before the end of the spermatogenesis, the cyst breaks up and spermatids are released into the lumina of the lobules, where spermatogenesis is then completed and spermatids transform into spermatozoa (Fig. 2L). The cysts appear to be arranged in order of maturation, with a gradient of germ cells of increasing maturation from the cortex to the sperm duct. The shape of the spermatozoa head seems to be elongated. Spawning season and size at first maturity Monthly distribution of macroscopic classification of the maturity phases (Fig. 3, A and B) revealed that the period of maximum occurrence of females in the spawn- ing capable phase (III) was from November to January. The presence of females in the actively spawning phase (IV) was observed from November to March, with a maximum peak in January. Females in the immature, regressing, and developing and regenerating phases (I, V, and II, respectively) were found throughout the year, with the highest percentage of immature individuals seen in May. A slight increase in phase-III females was observed in August, and that increase would likely re- sult in spawning activity in September, indicating the possibility of a secondary breeding season. Males in all maturity phases were observed throughout the year, with 2 maxima of mature males occurring in Decem- ber and July. For mature males and females, GSI and HSI indi- ces were calculated. In males, GSI was fairly constant throughout the year and a maximum index value of 1.06 was reached in January (Fig. 3C). The mean GSI for females was highest from December to March, with a peak of maximum activity in January (4.94) and Feb- ruary (2.43) (Fig. 3D). The mean HSI for females and males followed the same pattern. The highest value for males was found in September (2.50), and the lowest value in February (1.65) (Fig. 3C). In females, HSI val- ues ranged from 3.19 in January to 1.86 in March (Fig. 3D). The highest HSI values were found just at the beginning of the main spawning season. GSI and HSI results, together with observations of maturity phases throughout the year, indicate that there is one main spawning season from November to March. Comparison of L50 curves showed a clear difference between males and females. The size at 50% sexual maturity was 33.4 cm TL for males (Fig. 4A) and 48.2 cm TL for females (Fig. 4B). Reproductive strategy and fecundity The size-frequency distributions of oocyte diameters in each of the 5 maturity phases indicate that oocytes in different stages of development were found in each ma- turity phase (Table 2; Fig. 5). During phase I, only oo- cytes in the primary growth stage (chromatin nucleolar and perinucleolar) with a narrow range of diameters were present (Fig. 5A). In phase II, cortical alveolar 396 Fishery Bulletin 111(4) Gonad maturity stages: □ I oil 0 III □ IV b V A n = 103 123139 125 57 135 129 141 99 132130124 ^ ^ ^ V* ^ o* ^