Fishery Bulletin 'ATES O^ ^ Vol. 86, No. 1 Marine Biological Laboratory ' LIBRARY JUL 6 1000 Woods Hole, m^IS^^'^^^^QS oplank- 1 25 LOEB, VALERIE J., and OMAR ROJAS. Interannual variation of ichthyopl ton composition and abundance relations off northern Chile, 1964-85 VETTER, E. F. Estimation of natural mortality in fish stocks: A review ... NOTARBARTOLO-DI-SCIARA, GUISEPPE. Natural history of the rays on the genus Mobula in the Gulf of California 45 WILLIAMS, AUSTIN B. Notes on decapod and euphausiid crustaceans, continen- tal margin, western Atlantic, Georges Bank to western Florida, USA 67 WARLEN, STANLEY M. Age and growth of larval gulf menhaden, Brevoortia patronus , in the northern Gulf of Mexico 77 O'BRIEN, LORETTA, and RALPH K. MAYO. Sources of variation in catch per unit effort of yellowtail flounder, Limanda ferruginea (Storer), harvested off the coast of New England 91 McEACHRON, LAWRENCE W., JEFF F. DOERZBACHER, GARY C. MATLOCK, ALBERT W. GREEN, and GARY E. SAUL. Reducing the bycatch in a commer- cial trotline fishery 109 BRUCE, B. D. Larval development of blue grenadier, Macruronus novaezelandiae (Hector), in Tasmanian waters 119 COWAN, JAMES H., JR., and RICHARD F. SHAW. The distribution abundance, and transport of larval sciaenids collected during winter and early spring from the continental shelf waters off west Louisiana 129 HAMNER, WILLIAM M., GREGORY S. STONE, and BRYAN S. OBST. Behavior of southern right whales, Eubalaena australis, feeding on the Antarctic krill, Euphausia superba 143 Notes PENSON, JOHN B., JR., ERNEST 0. TETTY, and WADE L. GRIFFIN. An econo- metric analysis of net investment in Gulf shrimp fishing vessels 151 SHIRLEY, SUSAN M., and THOMAS C. SHIRLEY. Appendage injury in Dunge- ness crabs. Cancer magister, in southeastern Alaska 156 (Continued on back cover) Seattle, Washington U.S. DEPARTMENT OF COMMERCE C. William Verity, Jr., Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION William E. Evans, Under Secretary for Oceans and Atmosphere NATIONAL MARINE FISHERIES SERVICE Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. 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 instead of being issued individually. 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. SCIENTIFIC EDITOR, Fishery Bulletin Dr. Andrew E. Dizon Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service, NOAA P.O. Box 271 La Jolla, CA 92038 Editorial Committee Dr. Jay Barlow National Marine Fisheries Service Dr. William H. Bayliff Inter-American Tropical Tuna Commission Dr. George W. Boehlert National Marine Fisheries Service Dr. Robert C. Francis University of Washington Dr. James R. Kitchell University of Wisconsin Dr. William J. Richards National Marine Fisheries Service Dr. Bruce B. CoUette National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service Mary S. Fukuyama, Managing Editor The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications GfTice, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115. Second class postage is paid in Seattle, Wash., and additional offices. POSTMASTER send address changes for subscriptions to Fishery Bulletin, Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Although the contents have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has determined that the publication of this periodical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing this periodical has been approved by the Director of the Office of Management and Budget. Fishery Bulletin Marine Biolomlil^Z^ CONTENTS / LIBRARY JUL 6 ma Vol. 86, No. 1 I January 1988 LOEB, VALERIE J., and OMAR ROJAS. Interannual variation oPrc'Sfhj^^k^k-Mass. ton composition and abundance relations off northern Chile, 1964-85 VETTER, E. F. Estimation of natural mortality in fish stocks: A review 25 NOTARBARTOLO-DI-SCIARA, GUISEPPE. Natural history of the rays on the genus Mobula in the Gulf of California 45 WILLIAMS, AUSTIN B. Notes on decapod and euphausiid crustaceans, continen- tal margin, western Atlantic, Georges Bank to western Florida, USA 67 WARLEN, STANLEY M. Age and growth of larval gulf menhaden, Brevoortia patronus , in the northern Gulf of Mexico 77 O'BRIEN, LORETTA, and RALPH K. MAYO. Sources of variation in catch per unit effort of yellowtail flounder, Limanda ferruginea (Storer), harvested off the coast of New England 91 McEACHRON, LAWRENCE W., JEFF F. DOERZBACHER, GARY C. MATLOCK, ALBERT W. GREEN, and GARY E. SAUL. Reducing the bycatch in a commer- cial trotline fishery 109 BRUCE, B. D. Larval development of blue grenadier, Macruronus nouaezelandiae (Hector), in Tasmanian waters 119 COWAN, JAMES H., JR., and RICHARD F. SHAW. The distribution abundance, and transport of larval sciaenids collected during winter and early spring from the continental shelf waters off west Louisiana 129 HAMNER, WILLIAM M., GREGORY S. STONE, and BRYAN S. OBST. Behavior of southern right whales, Eubalaena australis, feeding on the Antarctic krill, Euphausia superba 143 Notes PENSON, JOHN B., JR., ERNEST O. TETTY, and WADE L. GRIFFIN. An econo- metric analysis of net investment in Gulf shrimp fishing vessels 151 SHIRLEY, SUSAN M., and THOMAS C. SHIRLEY. Appendage injury in Dunge- ness crabs, Cancer magister, in southeastern Alaska 156 iContinued on next page) Seattle, Washington 1988 For sale by the Superintendent of Documents, U.S. Government Printing Office, Washing- ton DC 20402— Subscription price per year: $16.00 domestic and $20.00 foreign. Cost per single issue: $9.00 domestic and $11.25 foreign. Contents — Continued CURRNES, KENNETH P., CARL B. SCHRECK, and HIRAM W. LI. Reexamina- tion of the use of otolith nuclear dimensions to identify juvenile anadromous and nonanadromous rainbow trout, Salmo gairdneri 160 BUTLER, JOHN L., and DARLENE PICKETT. Age-specific vulnerability of Pacific sardine, Sardinops sagax, larvae to predation by northern anchovy, En- graulis mordax 163 EPIFANIO, CHARLES E., DAVID GOSHORN, and TIMOTHY E. TARGETT. Induction of spawning in the weakfish, Cynoscion regalis 168 The National Marine Fisheries Service (NMFS) does not approve, recommend or endorse any proprietary product or proprietary material mentioned in this publi- cation. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. AWARDS The Publications Advisory Committee of the National Marine Fisheries Service has announced the best bublications authored by the NMFS scientists and published in the Fishery Bulletin for 1986 and the Marine Fisheries Review for 1985. only effective and interpretive articles which significantly contribute to the understanding and knowledge of NMFS mission-related studies are eligible, and the following papers were judged as the best In meeting this requirement. Fishery Bulletin 1986 — Starvation-induced mortalitv of young sea-caught jack mackerel, Trachurussymmetrlcus, determined with histological and morphological methods, by Gail M. Theilacker. Fish. Bull., U.S. 84: 1-17. Gail M. Theilacker is with the Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service, NOAA, La Jolla, CA. Marine Fisheries Review 1985 — Biology of the red sea urchin, Strongylocentrotus franclscanus, and its fishery in California, by Susumu Kato and Stephen c. schroeter. Mar. Fish. Rev 47(3):i-20. Susumu Kato is with the Southwest Fisheries center Tiburon Laboratory National Marine Fisheries Service, NOAA, Tiburon, CA, and Stephen C. schroeter is with the Department of Biological Sciences, university of California, university Park, CA. INTERANNUAL VARIATION OF ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE RELATIONS OFF NORTHERN CHILE, 1964-83 VaLKRIK J LOEB' AND OMAR ROJAS^ ABSTRACT Larval fishes were collected off of northern Chile during winter (July-September) ichthyoplankton surveys undertaken in 1964-70, 1972-73, and 1983. The 19-year timespan included a wide variety of hydrographic conditions in the Humboldt Current area (cold years. El Nino events, and intervening transition years); it also included the decline and collapse of the anchoveta fisheries and increases of sardine, mackerel, and jack mackerel stocks off northern Chile and Peru. The ichthyoplankton data are examined in relation to ambient hydrographic conditions as well as to possible chronological changes in environmental conditions which led to the increased Chilean sardine stocks and anchoveta fishery collapse. More coherent patterns come from considerations of larval fish species composition in 1964-69 and 1970-73 data sets than from years of "similar" hydrographic conditions. A marked shift in relative abundances of nonfished mesopelagic species in 1969-70 is associated with changes within long-term physical data bases from Chile and Peru suggesting a large-scale environmental change. Sardine stock growth began with successful larval survival of 1968-69 and later year classes. Anchoveta stock decline began in 1972 probably due to poor larval survival. Affiliation of anchoveta and coastal species larval abundance implies that they are similarly influenced by coastal processes. An atmos- pherically driven oceanic circulation change beginning in the late 1960's and possibly involving onshore presence of subtropical and or oceanic waters and altered coastal processes may have been responsible for the changes in the northern Chilean fish assemblages. The Humboldt Current region, like the other major eastern boundary current systems (Califor- nia, Benguela, and Canary Currents) is domi- nated by pelagic schooling fish stocks including anchoveta (Engraulis), sardine iSardinops), hake iMerluccius), mackerel (Scomber), jack mackerel (Trachurus), and bonita iSarda) (Par- rish et al. 1983). These fish stocks, like those in the other eastern boundary current areas, exhibit extreme population fiuctuations. Most notable in the past 30 years are the collapses of Peruvian and Chilean anchoveta stocks in the mid-1970's and their succession by sardine and, to a lesser extent, mackerel and jack mackerel stocks (San- tander and Flores 1983; Serra 1983). Hydrographic complexity and variability are characteristic of eastern boundary current sys- tems. Included in the Humboldt Current region are equatorial, subequatorial, subantarctic, and antarctic oceanic water masses; northward flow- ing currents and opposing countercurrents; and wind driven, seasonally variable coastal up- welling (Wyrtki 1967). Additionally, the region is iMoss Landing Marine Laboratories, P.O. Box 450, Moss Landing, CA 95039. 2Instituto de Fomento Pesquero, Avenida Pedro de Valdivia 2633, Casilla 1287, Santiago, Chile. subject to 1) large seasonal and longer period fluctuations in advection of water masses of markedly different properties and 2) large in- terannual differences in the timing and intensity of seasonal upwelling processes (Bakun 1987; Bernal et al. 1983; Parrish et al. 1983; Robles et al. 1976). The clearest and generally considered most important of the nonseasonal processes in- fluencing the biology of the current system is the El Nino phenomenon (Bernal et al. 1983; Guillen 1983). El Nino events off Peru and Chile are marked by large-scale atmospherically driven southward and coastward advection of warm, high-salinity equatorial and subequatorial sur- face waters, weakening of coastal upwelling (or upwelling of warm nutrient-poor waters), and weakening of subsequent phytoplankton blooms. These El Nino or warm-water periods are vari- able in their intensity and duration (Guillen 1983; Santander and Flores 1983). In contrast to these periods are more "normal" cold-water events resulting from atmospherically driven in- tensification of northward flowing cold, low- salinity subantarctic waters and seasonal up- welling of cold, nutrient-rich water. Major El Nino events occurred in 1891, 1925-26, 1940-41, 1957-58, 1965, 1972-73, 1976, and 1982-83; Manu.script accepted September 1987. FISHERY BULLETIN: VOL. 86, NO. 1, 1988. FISHERY HULLKTIN: VOL 8(i, NO 1 major cold events over the past 20 years occurred in 1964, 1967-68, 1970-71, and 1974-75 (Guillen 1983). The decline and ultimate collapse of the an- choveta fisheries of Peru began in 1970 and was finalized by the intense 1972-73 El Nino; the northern Chilean stock decline started in 1972 and was finalized by 1977. Factors facilitating these declines are generally believed to include overfishing and the devastating effects of the El Nino on anchoveta spawning behavior and in- tensity as well as on subsequent recruitment. Competition and/or predation pressure result- ing from increasing abundances and distributions of sardine and mackerel have also been hy- pothesized (Santander and Flores 1983; Serra 1983). Because of the great socioeconomic value of the dominant pelagic fish species of the Peru-Chile ecosystem, their population fluctuations have re- ceived a great deal of attention over the past 20 years. However, coincidental changes in the com- position, abundance, or spawning intensities of other commercially less important and non- harvested species have not been examined. Infor- mation on the changes of these unfished species in relation to hydrographic conditions and fluctu- ations of the dominant pelagic fish stocks provide additional insight into the ecology of the Hum- boldt Current and may elucidate possible causes for the dramatic changes which occurred during the 1970's. In the present work we examine the abundance and composition of total ichthyoplankton assem- blages collected off of northern Chile (lat. 18°- 24°S) during 1964-73 and 1983 in relation to am- bient hydrographic conditions. "Normal" cold water as well as warm-water and El Nino events occurred during the 19-yr sampling span. We also examine our results with respect to possible chronological change in environmental condi- tions which led to the 1977 anchoveta fishery col- lapse off northern Chile. Our results may be ap- plicable for interpreting coincidental changes in the Peruvian ecosystem and may also be broadly applicable for studies of similar changes in the other eastern boundary current ecosys- tems. METHODS Samples were collected during 1964-73 and 1983 ichthyoplankton surveys conducted by the Instituto de Fomento Pesquero. The area most intensively surveyed was a narrow coastal strip extending between Arica and Antofagasta (lat. 18°-24°S, long. 70°-72°W; Fig. 1). This area in- cludes one of two major anchoveta (Engraulis rin- gens ) spawning grounds off Chile and the pri- mary sardine iSardinops sagax) spawning area off Chile prior to 1973 (Fig. 2A, B). All samples used for interannual comparisons were collected during late July-September following peak win- ter anchoveta and sardine spawning periods. Be- tween 21 and 87 samples from the 18°-24°S area were analyzed for each of 11 cruises (Table 1). In one case data from two cruises (August and Sep- tember 1968) were pooled to provide adequate coverage. Sampling was done annually from 1964 to 1970 and in 1972 and 1973. There was a 10-yr hiatus before regular sampling was resumed in 1983. The 1964-73 samples were collected with Hensen nets (0.28 m^ mouth opening; 300 ixm mesh). Prior to 1973 the vertical net hauls were 50-0 m; in 1973 haul depth was increased to 100 m. The 1983 100-0 m vertical hauls were made with WP2 nets (0.25 m^ mouth opening; UNESCO 1968) of 300 ^JLm mesh. Samples were preserved using buffered 5% formalin solution. Sea surface temperature and salinity data were collected at most sampling stations for all but two winter cruises; these data are lacking for 1970 and salinity data are minimal for 1967. All fish eggs and larvae were removed from samples, and invertebrate zooplankton biomass was measured. Wet weight displacement volume was measured for 1964-73 samples; in 1983 the Yashnov (1959) technique modified by Robertson (1970) was used. A calculated correction factor of 1.44 (±3.34) was applied to the 1983 biomass values to permit comparison with the earlier data. All fish larvae were identified to lowest taxon possible and counted. We herein treat the larvae of six commercially important species (anchoveta [Engraulis ringens], Pacific sardine [Sardinops sagax], jack mackerel [Trachurus murphyi; also known as T. symmetricus in U.S.A.], chub mack- erel [Scomber japonicus]. South Pacific men- haden [Ethmidium maculatum], and hake [Mer- luccius gayi]) separately from the other 35 identified taxa. These six species are referred to as the "PL" (larvae of pelagic schooling species). The other larval taxa considered together are the "OL". The PL and OL categories are treated sepa- rately because abundances of the PL (especially of anchoveta and sardine) mask abundance rela- LOEB and ROJAS: ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE 90° 75° 60° 45< z < UJ u o < a. 15 - 30' 45' 60' Figure 1. — Ichthyoplankton study area between Arica and Antofagasta, Chile (18°-24°S), 1964-83. tions of the OL. The OL are further separated into myctophid, "other mesopelagic" and "coastal" fractions. Ichthyoplankton abundances are expressed as numbers per 10 m^ sea surface. Flow meters were not used with the vertical Hensen net hauls; numbers of larvae caught in each tow are multi- plied by 30 to provide numbers per 10 m^ esti- mates. A flow meter used with WP2 nets provided water volume measurements and more accurate abundance estimates. Based on these measure- ments the conversion factor applied to Hensen net catches appears to be reasonable: water volumes filtered by 100-0 m vertical WP2 net hauls aver- 76* 74° "I T" COQUIMBO VALPARAISO FALC AHUANO VALDIVIA PTO MONTT _ai aJl_ 18 20 22 24' 26 ° 28 ° 30° 32 34' 36' 38' - 40' 42< 44' FISHKKY lUlI.I.KTIN: VOl. Hfi. NO 1 72° 70° I 8 ° S - 20 ° 22 o -24 26 -28 - 30 - 32 - 34 - 36 -38 - 40' - 42 44 Figure 2. — Spawning areas of (A) anchoveta iEngraiilis ringens) and (Bi sardine tSardinops sagax) off Chile based on egg abundances during July-September ichthyoplankton surveys, 1964-73. LOEB and ROJAS: ICHTHYOPLANKTON COiMPOSITION AND ABUNDANCE Table 1. — Cruises yielding samples used for examination of ichthyoplankton abundance and composition variations oft northern Chile Only samples from 1 8 -24 S are used for interannual comparisons. Data from two 1 968 cruises are combined; data from cruise 71(4)69CD are used for analysis of sampling depth-related catch differ- ences. N = number of samples used in ichthyoplankton analyses Tow types: H = Hensen net; WP2 = WP2 net; V = vertical. Year Cruise Dates Location N Tow Type Depth (m) 1964 06(3)64GE 16 08-23 09 18 20', , 23 38'S 70 ir. , 71 50'W 68 H V 0-50 1965 13(3)65CD 15 08-09 09 18 20' , 23 50'S 70 00', , 72 08'W 76 H V 0-50 1966 25{3)66CD 21 08-31 09 18 28' , 23 52'S 7016' , 72 16'W 72 H V 0-50 1967 37(3)67CD 17 08-1009 18 25', , 23 45'S 70 05' , 71 38'W 59 H V 0-50 1968 47(3)68NO 25 08 18 28' , 23 OO'S 70 05' , 70 58'W 37 H V 0-50 49(3)68NO 29 09 18=26' , 23 01'S 70 11' , 70 59'W 41 H V 0-50 1969 70(3)69NO 23 08-25 08 18^29' , 23 01'S 70 09' , 71 10'W 35 H V 0-50 71(4)69CD 71(4)69CD 03 12-17 12 03 12-17 12 28 29' 28 29' , 38 OO'S , 38 OO'S 71 22' 71 22' , 73 55'W , 73 55'W 43 39 H H V V 0-100 0-50 1970 86(3)70NO 25 09-26 09 19°27' , 21 58'S 70 14' , 71 03'W 21 H V 0-50 1972 109(3)72NO 04 09-15 09 18 '29' , 22^58 'S 70 '10' , 71 25'W 87 H V 0-50 1973 130(3)73CP 28 07-08 08 18 17' , 23 OS'S 7005' , 72 20'W 42 H V 0-100 1983 277(3)83CP 07 08-15 09 18°33' . 23 48'S 70 09' , 71 38'W 38 WP2 V 0-100 aged 28.3 m'^ yielding a raw count to numbers per 10 m'- conversion factor of 35; this 179f increase in conversion factor is associated with a 12*7^ de- crease in mouth opening of WP2 vs. Hensen nets. Larval fish diversity is expressed as total num- bers of taxa per sampling period and mean num- bers of taxa per tow. Various parametric and nonparametric tests were used for statistical analyses. Differences in mean abundances are tested with 2-tailed Z tests and Mann Whitney U tests (Dixon and Massey 1969; Conover 19711. Similarity of abundance ranks within data sets are tested with Kendall's concordance iW) test (Tate and Clelland 1969) and Spearman's rho (p) correlation test (Conover 1971). Significant values resulting from these tests are indicated, but due to multiple testing these values should be used only as indicators of the relative strengths of relationships. Percent similarity indices (PSI; Whittaker 1975) are used for comparisons of species percentage composi- tion. Because PSIs are strongly influenced by abundant species, we apply these tests to the OL fraction as well as to total larvae. We define as "high" all PSI values >80, as "moderate" PSI's >65 and <80, and as "low" values <65. SAMPLING CONSIDERATIONS Sampling Depth Differences The 100 m sampling depths in 1973 and 1983 potentially effect direct comparisons of abun- dance estimates and species composition in these vs. earlier data sets owing to individual species' depth distributions. Evaluation of depth-related sampling differences is possible through a com- parison of data obtained from coincidental 50-0 m and 100-0 m Hensen net samples taken during December 1969 at 28°-38"S (Table 1). These data indicate that abundance estimates and percent- age contribution of the PL to total larvae were higher in 50 m tows while those of the OL were higher in 100 m tows (Table 2). Four of five PL species caught were more abundant and frequent in the 50 m samples. The greatest differences were for Trachurus murphyi, which was 5x more abundant and 3.6 x more frequent, and Merluc- cius gayi , which was present only in 50 m sam- ples. Abundance estimates were higher for 11 of 16 OL taxa in 100 m tows although, with a few exceptions, catch frequencies were similar. The greatest difference was for Triphoturus mexi- FISHERY BULLETIN: VOL 86, NO. 1 Table 2. — Catch comparisons of 50-0 m and 100-0 m Hensen net tows taken at lat. 28 -38 S, long. 71 -74'W during 3-17 December 1969. Abundances expressed as mean and standard deviations of numbers per 10 m2, F = percent frequency of occurrence in samples Depth-related catch differences are shown as ratios of 50:100 m abundance estimates, species percentage contribution to total identified larvae, PL and OL percentage contribution to total larvae, and taxonomic diversity (mean numbers per tow and total numbers of taxa). N = number of samples. PL = larvae of pelagic species; OL = other larval taxa. 50- -0 m 100- ■0 m (N = 43) (/V = 39) Taxon X (S) F % X (S) F % Ratio Engraulis ringens 34.2 (122.8) 23.2 13.42 22.3 69.4) 15.4 7.23 1.53 Clupea bentincki 6.3 ( 28 2) 9.3 2.47 1.5 6.7) 5.1 0.49 4.20 Ethmidium maculatum 0.7 ( 4.6) 2.3 0.27 0.8 4.8) 2.6 0.26 0.88 Merluccius gayi 1.4 ( 6.4) 46 0.55 Trachurus murphyi 4.2 ( 14.0) 9.3 1.65 0.8 4.8) 2.6 0.26 5.25 Total PL 46.8 (131.1) 34.9 1836 25.4 71.0) 23 1 8.24 1.84 Bathylagus nlgngenys 0.7 ( 4.6) 2.3 0.27 0.8 4.8) 2.6 0.26 0.88 Vinciguerria lucetia 10.5 ( 33.9) 14.0 4.12 16.9 75.6) 7.7 5.48 0.62 Diogenlchthys spp. 31.4 (107.0) 20.9 12.32 37.7 79.9) 28.2 12.22 0.83 Hygophum bruuni 121.4 (238.2) 58.1 47.64 166.2 291.0) 53.8 53.86 0.73 Protomyctophum sp. 3.1 9.2) 10.3 1.00 Diaphus sp. 77 ( 18.6) 18.6 3.02 9.2 24.0) 17.9 2.98 0.84 Lampanyctus parvicauda 7.7 ( 18.6) 163 3.02 6.9 21.2) 15.4 2.24 1.12 Lampanyctus sp. 1.4 { 6.4) 4.6 0.55 1.5 6.7) 5.1 0.49 0.93 Triphoturus mexicanus 7.7 ( 21.8) 14.0 3.02 18.5 57.4) 20.5 5.99 0.42 Scopelosaurus sp. 0.8 4.8) 2.6 0.26 Normanichthys crockeri 15.3 ( 38.9) 18.6 6.00 20.0 66.5) 10.3 6.48 0.76 Sebastes sp. 1.4 ( 9.1) 2.3 0.55 0.8 4.8) 2.6 0.26 1.75 Blennild A 0.7 ( 4.6) 2.3 0.27 Blenniid D 0.7 ( 4.6) 2.3 0.27 0.8 4.8) 2.6 0.26 0.88 Bothid 0.7 ( 4.6) 2.3 0.27 Unid. 2 0.7 ( 4.6) 2.3 0.27 Small damaged myctophids 4.8 9.2 Other unidentified 13.2 7.0 Total OL 226.0 (316.0) 86.0 81.59 299.4 405.9) 82.0 91.78 0.84 Total larvae 272.8 (340.7) 324.8 400.0) 0.84 No. taxa/tow 2.5 ( 2.0) 2.2 2.0) 1.13 Total no. taxa 18 17 0.94 canus which was 2.4 x more abundant and 1.5 x more frequent in 100 m tows. Because of the large catch variability, none of the species abundance differences nor the abundance differences of the PL, OL, and total larvae are significant (Z tests, P's all >0.10). Additionally, species abundances within all positive tows from the two sampling depths are not significantly different (Mann Whitney U tests, P > 0.10 in all cases). The overall species composition of 50 and 100 m tows was similar. Despite greater proportions of PL in 50 m tows, the PSI value from comparisons of total species lists was high (87.7). Species per- centage contribution within the OL fraction of the two tow types was also quite similar (94.5). Spe- cies abundance rankings within the two total lar- val data sets are significantly correlated (p = +0.80; P < 0.01 ). Species diversity estimates (total numbers of taxa and mean numbers of taxa per tow) are also similar. From these comparisons it is apparent that the PL predominantly occur within the upper 50 m. Similar shallow (e.g., <50 m ) distributions have been described for dominant PL species off of Peru (anchoveta, sardine, and hake; Sameoto 1982). The generally lower 100 m abundance estimates of these species is puzzling, but suggests possi- bly shorter sampling time and/or less efficient sampling within the upper 50 m of these tows. Higher catch frequency and abundance of T. mexicanus in 100 m tows suggest that large pro- portions (e.g., 30-60%) of these larvae are at 50- 100 m. As a result of these catch differences we suggest caution in making direct numerical comparisons between the 1973 and 1983 vs. earlier data sets. Although the overall compositions and abun- dance relations should not be markedly altered, some accommodation should be allowed for the percentage contributions and across-year abun- dance ranks of PL species (especially Trachurus) and Triphoturus mexicanus. LOEB and ROJAS: ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE Net Type Differences There are no data available for a direct evalua- tion of catch differences between the vertical Hensen and WP2 net tows. However, comparisons of data from the 1983 WP2 net hauls (Table 3) do not indicate that this net is more or less efficient than the Hensen net. Mean abundances and spe- cies diversity (total numbers of taxa and mean numbers of taxa per tow) are within the ranges of values from Hensen net tows. Day-Night Catchi Considerations Day and night sample data have been combined for each cruise. Day samples (0600 -1800 hours) outnumbered (generally 55-64%) night samples during all cruises. In nine cases there were non- significant day-night larval catch differences (Z tests, P >0.05) and overall similarity in night: day catch ratios (0.74-1.8:1; mode = 1.2:1). One cruise had a significantly higher night vs. day catch (2.5 X; P < 0.01); this cruise and one with Table 3. — Abundance estimates and diversity of larval fishes and abundance estimates of zooplankton collected in July-September samples off northern Chile (18 -24-S), 1964-83^ Larval fish abundances as mean numbers per 10 m^. PL = larvae of commercially important pelagic species; OL = other larval taxa (myctophid, other mesopelagic, and other categories). Number of larvae is total raw count of identified larvae for each sampling period. Larval fish diversity expressed as total number of identified taxa and mean number of taxa per tow. Zooplankton abundance is mean displacement volume (cc/10 m2). N.A. = data not available. Mean abundance for Taxon 1964 1965 1966 1967 1968 1969 1970 1972 1973 1983 Engraulis ringens 302.0 36.7 3,478.3 224.7 72.3 181.7 620.0 1,634.1 1,816.4 — Sardinops sagax 5.3 5.1 12.5 4.6 21.1 24.9 1.4 327.2 52.1 594.9 Ethmidium maculatum — — — — — 2.6 — 0.7 — — Trachurus murphyi 2.6 7.1 10.8 3.1 0.4 6.8 5.7 100.3 2.1 — Scomber japonicus — 0.4 — — — — — 25.9 — — Merlucaus gayi — — — — — — — — 0.7 2.6 Total PL 309.9 49.3 3,501.6 232.4 93.8 216.0 627.1 2,088.2 1,871.3 597.5 Bathylagus nigrigenys 4.0 5.1 6.7 9.7 6.2 0.9 7.2 56.6 35.0 31.5 Vinciguema lucetia 3.1 0.4 2.9 6.1 3.1 2.6 21.4 54.8 18.6 3.7 Stemoptyx diaphana — — — — — — — 0.7 0.7 1.0 Chaultodus sp. — — — — — — — 0.3 — — Stomias spp. — — — — — — — — — 1.9 Lestidiops pacificum — — — — — — — 0.3 0.7 — Melamphaes sp. — — — 0.5 0.4 — — 0.3 2.1 — Beryciform — — — — — — 1.4 — — — Other mesopelagics 7.1 5.5 9.6 16.3 9.7 3.5 30.0 113.0 57.1 38.1 Diogenichthys spp. 6.2 18.9 14.6 30.5 7.3 14.6 21.4 59.3 39.3 114.2 Hygophum bruuni — 2.8 1.2 3.6 — — — — — 0.9 Hygophum atratum — — — — — — — — — 6.7 Metelectrona ventralis — — — — — — — — — 0.9 Myctophum nitidulum 0.4 — 0.4 — 0.4 0.9 — 4.8 4.3 4.6 Diaphus sp. — 0.4 0.4 — — — 1.4 0.3 — 4.5 Lampanyctus parvicauda 41.0 32.0 65.8 63.0 10.8 38.6 5.7 29.7 12.1 15.5 Lampanyctus spp. — — 0.8 — 1.1 — — — 0.7 — Triphoturus mexicanus 31.8 15.8 20.8 29.0 4.6 31.7 11.4 19.7 23.6 33.6 fvlyctophids 79.4 69.9 104.0 126.1 24.2 85.8 39.9 113.8 80.0 180.9 Normanichtys crockeri 17.2 1.5 281.2 5.6 2.3 26.6 10.0 2.4 31.4 — Sebastes sp. 0.4 — 2.5 1.0 0.8 — — — 0.7 — Gadiform D — — 0.8 — — — — — — — Macrourid A — — — 0.5 — — — — 0.7 — Macrourid C — — — 0.5 — — — — — — Blenniid A 0.4 — 0.8 — 1.2 — — — 1.4 — Blenniid B — — 0.4 — — — — — — — Blenniid C — — — 1.5 — — — 0.3 — — Blenniid D 5.3 0.4 16.7 7.1 4.2 2.6 — 2.8 8.6 1.0 Gobiesocid A 0.4 — 0.8 — — — — 1.0 0.7 — Gobiesocid B — — — — 0.4 — — — — — Unid. 1 — — — — 0.8 ^ — — 5.7 8.6 Unid. 2 — — 0.4 — — — — 0.3 12.1 — Unid. 3 — — — — 0.4 1.7 — 0.3 0.7 — Unid 4 — — — — — — — 0.3 0.7 — Ophidiid — — — 1.0 0.8 — — — — — Hippoglossina sp. — — — — — — — 0.3 — — Other larvae 23.7 1.9 303.6 17.2 10.9 30.9 10.0 7.7 62.7 9.6 Total OL 110.2 77.3 417.2 159.6 44.8 120.2 79.9 234.5 199.8 228.4 FISHERY BULLETIN: VOL. 86, NO 1 Table 3. — Continued. Mean abundance for Taxon 1964 1965 1966 1967 1968 1969 1970 1972 1973 1983 Larval fish abundance and di iversity Total ID PL 309.9 493 3,501 6 2324 938 216.0 627 1 2,0882 1,871 3 597.5 Total ID OL 110.2 77,3 417.2 159 6 448 120.2 799 234 5 1998 2286 Total ID larvae 420.1 126.6 3,9188 3920 138.6 336.2 707.0 2,322.7 2,071.1 826 1 Unid missing OL 10.2 4.4 170 220 3.4 11.9 7.2 104.5 57.2 76.6 Total larvae 430.3 131.0 3.9358 4140 142.0 348.1 714.2 2,427.2 2,1283 9027 Number of larvae 953 321 9,406 771 360 650 495 6,738 2,900 893 Number of taxa 15 14 20 19 20 13 12 26 25 19 Number of taxa tow 22 16 32 25 1.4 1.9 23 5.3 4 1 3.7 Number of samples 68 76 72 59 78 35 21 87 42 38 Zooplankton abundance 333.9 2422 406.5 370.8 279.2 141.0 N.A 301 7 100.6 168.7 Number of samples 85 124 72 59 78 35 110 42 30 a 1.8:1 night:day catch ratio were represented by fairly equal day {55-5T/( ) and night sam- ples. TAXONOMIC PROBLEMS The 576 samples used for interannual compari- sons yielded a total of 41 taxa including 19 spe- cies, 7 genera, and 11 higher taxa (Table 4). The PL and most mesopelagic forms were identified to species. During several cruises there were large proportions of small Diogenichthys spp. (Myc- tophidae) larvae which could not be identified to species. As a consequence, data on the two spe- cies, D. atlanticus and D. laternatiis, were lumped to permit reasonable between-year taxonomic composition comparisons. In all but one cruise, D. laternatiis dominated (77-1007f ) the identifiable Diogenichthys larvae. Total within-year Dio- genichthys spp. abundances were multiplied by proportions of identified D. laternatus and D. at- lanticus larvae to provide between-year abun- dance rankings for each species. Species identifications of coastal forms are lim- ited by inadequate taxonomic information and by the presence of generally early larval develop- mental stages in samples. These larvae are pri- marily classified at familial and ordinal levels. Because the classifications include few multispe- cies groupings and those were numerically rare the taxonomic limitations offer no severe analyti- cal problems. Largest taxonomic problems occurred in cruises when large numbers of small unidentifi- able larvae were caught (e.g., 1972 and 1983; Table 3). Additionally, most cruises had "miss- ing" larvae (e.g., "other larvae" enumerated when the samples were first processed but not ac- counted for during later species identification work). With the exception of 1983 the uniden- tified and missing larvae made up <5% of the total larval abundance for each sampling pe- riod. RESULTS Overall Ichthyoplankton Composition The 576 July-September samples used for in- terannual comparisons yielded a total of 23,487 identified larvae. These larvae were dominated (85.17r) by PL species (Table 4). Overall domi- nants were anchoveta (Engraulis ringens; 74.37c) and sardine (Sardinops sagax; 9.3%). The other PL species were relatively rare: Trachurus mur- phyi contributed 1.2% and Scomber japonicus, Merluccius gayi , and Ethmidium maculatum to- gether formed 0.3% of the total. The larval abun- dances of these species off of Chile are strongly influenced by sampling time and location. Mer- luccius gayi occurs primarily to the south of the study area (24°-43°S) and Scomber japonicus and Trachurus murphyi have later summer (Novem- ber-February) spawning peaks. The OL were dominated by mesopelagic fishes (18 taxa, 10.6% of total larvae). Myctophids were most abundant (8.0%) primarily because of the large numbers of Diogenichthys spp., Lampanyc- tus parvicouda , and Triphoturus mexicanus, which together made up 7.7% of the total. One bathylagid (Bathylagus nigrigenys) and one gonostomatid (Vinciguerria lucetia ) were also rel- atively abundant (together 2.5%). Coastal fish larvae (14 taxa) made up 4.2% of the total; a scorpaeniform (Normanichthys crockeri; 3.4%) and blenniid (Blenniid D; 0.4%) dominated this group. Eighteen taxa were relatively frequent (e.g., in 8 LOEB and ROJAS: ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE >50% of cruises) and/or abundant across the 10 sampling periods (Table 4). These taxa (four PL species, seven myctophid taxa, two other meso- pelagic species, and five coastal forms) made up 99.69; of the identified larvae; they also con- tributed 97-1009r of the identified larvae (91.2- 99.6% of total larvae) and included the top 9-12 ranked taxa within each sampling period (Table 5). Table 4— Ichthyoplankton species collected in July-September samples oft northiern Chile (18 -24 S, 70 -72 W), 1964-83. Rela- tive abundance (ROA) and percentage composition based on summed cruise mean abundances (no. 10 m2) of all identified forms. Frequency is number of total 10 sampling periods when taxon was caught. Categories are PL (commercially important pelagic species) and other taxonomic components (M = myc- tophids: OM = other mesopelagic taxa; C = coastal forms). Fre- Cate- Taxon ROA % quency gory Engraulis ringens 1 74.29 9 PL Sardinops sagax 2 9.33 10 PL Normanichthys crockerl 3 3.36 9 C Diogenichthys latema- tus + D. atlanticus 4 2.90 10 M Lampanyctus parvicauda 5 2.79 10 M Tnphoturus mexicanus 6 1.97 10 M Bathylagus nigngenys 7 1.44 10 OM Trachurus murphyi 8 1.23 9 PL Vinciguerna lucetia 9 1.04 10 OM Blenniid D 10 0.43 9 C Scomber japonicus 11 023 2 PL Myctophum nttidulum 12 14 7 M Unid. 1 13 0.13 3 C Unid. 2 14 0.11 3 C Hygophum bruuni 15 0.075 4 M Diaphus sp. 16 0.062 5 M Hygophum atratum 17 0.060 1 M Sebastes sp. 18 0050 5 C Blenniid A 19 0.034 4 C Ethmidium macu latum 20.5 0030 1 PL Merluccius gayi 20.5 0.030 2 PL Unid. 3 22 0028 4 C Melamphaes sp 235 0.026 3 OM Gobiesocid A 23.5 0026 4 C Lampanyctus spp. 25 0024 3 M Sternoptyx diaphana 26 0,020 3 OM Stomias spp. 27 0.017 1 OM Blenniid C 285 0016 2 C Ophidiid 285 0016 2 C Beryciform 30 0.012 1 OM Macrourid A 31 0011 2 Lestidiops pacificum 32.5 0.009 2 OM Unid. 4 32.5 0.009 2 C Metelectrona ventralis 34 0.008 1 M Gadiform D 35 0.007 1 Macround C 36 0,004 1 Blenniid B 37.5 0.0036 1 C Gobiesocid B 37.5 0.0036 1 C Chaultodus sp. 39.5 0.0030 1 OM Htppoglossina sp. 39.5 0.0030 1 C Total number of identified larvae: 23,487 Total number of samples: 576 Total number of taxa: 41 INTERANNUAL VARIATIONS IN ABUNDANCE AND COMPOSITION The ichthyoplankton demonstrated extreme in- terannual variations in abundance and composi- tion (Tables 3, 5; Fig. 3A, B). Most obvious are the 1) total larval and PL abundance peaks of 1966, 1972, and 1973 and 2) shift from anchoveta to sardine dominance between 1973 and 1983. The maximum total larval abundance values in 1966, 1972, and 1973 were 2.4-30 x higher than those of other years; the PL had 3-71 x higher abun- dances during these vs. other years. Interannual abundance variations during the seven years of lower abundance were also large for total larvae (to 6.9 X ) and the PL (to 12.7 x ). Anchoveta domi- nated the ichthyoplankton through 1973 (29.0- 88.8% ) and was directly responsible for the ex- treme abundance variations; anchoveta were absent from the 1983 samples, and sardine larvae (72.0%) contributed to the moderately high total larval and PL abundances (Table 5). The OL fraction had less extreme abundance variations than the PL: maximum 1966, 1972, 1973, and 1983 mean abundance values ranged from 1.4 to 9x those of the other years; mean values within the six years of lower abundance varied to 3.8 x. Unlike the PL, OL abundance fluctuations were not attributed to any one taxo- nomic component (Fig. 3B): the 1966 peak was largely due to coastal taxa (73% of OL); myc- tophids and other mesopelagic taxa equally domi- nated the 1972 peak (48%); myctophids domi- nated (79%) in 1983; and all three components were relatively abundant (29-40%) in 1973. Myc- tophids were generally the dominant component (50-90% ) during the six years of lower OL abun- dance. Abundance fluctuations (as range of mean abundance values) across the 10 years were less extreme for myctophids (8.1 x ) than for the other mesopelagic taxa (32.3 x) and coastal forms (160x). In contrast to the ichthyoplankton, July-Sep- tember zooplankton biomass values were rela- tively constant between years and exhibited only a 4x range in values (Table 3). Despite large between-year variability in rela- tive proportions of the PL and OL, there is a sig- nificant agreement of their ranked mean abun- dances across the 10 years (p = +0.81, P < 0.01). There is also a general agreement of the ranked mean abundances of PL and the three OL compo- nents across the 10 years (W = 0.44; P = 0.05). These categories were generally more abundant 9 FISHERY BULLETIN: VOL. 86, NO 1 Table 5. — Comparisons of relative abundances of dominant larval fish taxa collected off of northern Chile (18- 24°S, 70°-72 W) during July-September sampling periods, 1964-83. Relative abundances within each year are presented as (A) percentage contribution to total identified larvae and (B) ranked abundance. Taxa are listed in order of total summed 10-yr mean abundances Taxon 1964 1965 1966 1967 1968 1969 1970 1972 1973 1983 A. Percentage contribution to total identified larvae Engraulis nngens 71.89 28.99 88.76 57.32 52 20 54.05 87.69 70 35 87.70 — Sardinops sagax 1.26 4.03 0.32 1.17 15 23 7.41 0.20 14.09 2.52 72.01 Normanichthys crocken 409 1.18 718 1.43 1.66 7.91 1.41 0.10 1.52 — Diogenichthys spp. 1.48 14.93 0.37 7.78 5.27 4.34 303 2.55 1.90 13.82 Lampanyctus pan/icauda 9.76 25.28 1.68 16.07 7.80 11.48 0.81 1.28 058 1 88 Tnphoturus mexicanus 7.57 12.48 0.53 7.40 332 9.43 1 61 085 1.14 4.07 Bathylagus nigrigenys 095 4.03 0.17 2.47 4.48 027 1.02 244 1.69 381 Trachurus murphyi 0.62 5.61 0.28 0.79 0.29 2.02 0.81 4.32 0.10 — Vinciguerna lucetia 0.74 0.32 0.07 1.56 2.24 0.77 3.03 2.36 0.90 0.45 Blenniid D 1.26 0.32 0.43 1.81 3.03 0.77 — 0.12 0.42 012 Scomber japonicus — 0.32 — — — — — 1.12 — — Myctophum nitidulum 0.10 — 0.01 — 0.29 0.27 — 0.21 0.21 0.56 Unid. 1 — — — — 0.56 — — — 0.27 0.96 Unid. 2 — — 0.01 — • — — — 0.01 0.57 — Hygophum bruuni — 2.21 0.03 0.92 — — — — — 0.11 Diaphus sp. — 0.32 0.01 — — — 0.20 0.01 — 0.54 Hygophum atratum — — — — — — — — — 0.81 Sebastes sp. 0.10 — 0.06 0.26 0.58 — — — 0.03 — Other taxa 0.19 — 0.09 1.02 3.03 1.28 0.20 0.19 0.43 0.78 B. Ranked within-year abundance Engraulis nngens 1 1 1 1 1 1 1 1 1 — Sardinops sagax 6.5 6.5 7 9 2 5 10 2 2 1 Normanichthys crocken 4 9 2 8 9 4 5 12 5 — Diogenichthys spp. 5 3 6 3 4 6 2.5 4 3 2 Lampanyctus parvicauda 2 2 3 2 3 2 7.5 7 8.5 5 Triphoturus mexicanus 3 4 4 4 6 3 4 9 6 3 Bathylagus nigrigenys 8 6.5 9 5 5 12.5 6 5 4 4 Trachurus murphyi 10 5 8 11 17 7 7.5 3 13.5 — Vinciguerria lucetia 9 11.5 10 7 8 9 2.5 6 7 10 Blenniid D 6.5 11.5 5 6 7 9 — 11 10 13.5 Scomber japonicus — 11.5 — — — — — 8 — — Myctophum nitidulum 12.5 — 18.5 — 17 12.5 — 10 12 8 Unid. 1 — — — ^ 12.5 — — — 11 6 Unid. 2 — — 18.5 — — — — 20.5 8.5 — Hygophum bruuni — 8 12 10 — — — — — 15.5 Diaphus sp. — 11.5 18.5 — — — 10 20.5 — 9 Hygophum atratum — — — — — — — — — 7 Sebastes sp. 12.5 — 11 13.5 14 — — — 21 — in 1966, 1972, and 1973 and relatively rare in 1965 and 1968. The ranked abundance patterns of each of the components differ from one another (e.g., all pairwise correlation coefficients [p = —0.21 to +0.61] are nonsignificant). Larval diversity is strongly correlated with total larval abundance (p = +0.88, P < 0.01). There are no significant correlations between abundance ranks of invertebrate zooplankton biomass and total larvae (p = +0.27) or any of the larval categories (p = -0.03 to +0.22; P > 0.05 in all cases). Species Abundance Variations and Relations The top 10 ranking larval fish taxa were caught during at least 9 of the 10 sampling periods (Table 4). All of these taxa exhibited large interannual abundance fluctuations (Table 3). Most marked were the abundance changes of anchoveta, sar- dine, and coastal species Normanichthys crockeri. This latter species (rank 3 in overall abundance) was frequently abundant prior to 1983; like an- choveta it was absent from 1983 samples. Among the 10 taxa only Triphoturus mexicanus had <10x changes in mean abundance values; <20x changes occurred for Lampanyctus parvicauda (11.5X) and Diogenichthys spp. (18.4x); all other taxa had >20x mean abundance changes over the 10 years. The abundance fluctuations of these 10 taxa are primarily responsible for the interan- nual abundance and composition variations (Table 5; Fig. 3A, B). 10 LOEB and ROJAS: ICHTHYOPLANKTON COMI'OSITION AND AHUNDANCE A. TOTAL ICHTHYOPLANKTON E O a. t :1 Anchoveta Other PL I I Sordine OL B. OL CATEGORIES II Other mesopeiogic toxo M yctophids Coastal toxa * j > \ i j 83 Figure 3. — Mean abundance (numbers per 10 m2) of (A) total ichthyoplankton and major PL components and (Bl total OL and major OL components collected off northern Chile (18'-24°S) during July-September sampling periods, 1964-83. Between-year comparisons of the species per- centage compositions of total larvae give a wide range of PSI values (3.3-95.0; Table 6A) which primarily reflect similarity in percentages of an- choveta; 51% of these values are moderate to high (e.g., >65). Highest values (91.5-95.0) come from comparisons between 1966, 1970, and 1973 when anchoveta contributed >87% of the larvae. High values (80.7-82.5) also result from comparisons between 1964, 1967, and 1969 and result from moderate anchoveta abundance (54.0-71.9%) and relatively similar proportions of other taxa. Low- est values (3.3-31.0) result from comparisons of 1983 vs. all other years and reflect the absence of anchoveta larvae in 1983 samples. The 1973 and 1983 PSI values are little affected (e.g., <2.3) by 11 FISHERY BULLETIN: VOL. 86, NO 1 Table 6. — Between-year percent similarity index (PS!) values from compari- sons of (A) total larvae and (B) \he OL (ottier larval taxa) fraction collected during July-September sampling periods off northern Cfiile (18-24 S), 1964- 83. PS! values for 1965 1966 1967 1968 1969 1970 1972 1973 1983 A. Total larvae 1964 52.44 79.94 8222 71.25 80.70 79.70 77.89 80.30 10.30 1965 X 33.95 67.41 55.54 62.38 38.16 45.32 38.74 28.47 1966 X 62.69 57.83 65.08 91.55 73.65 91 88 3.32 1967 X 77.96 8247 67.75 68.21 66.82 18.05 1968 X 79.39 62.81 76.39 63.68 30.95 1969 X 62.95 69.67 63.50 18.80 1970 X 79.06 94.95 7.31 1972 X 79 42 22.02 1973 X 8.95 B. OL fraction 1964 69.88 47.01 75.76 57.56 85.16 45.99 36.08 51.29 33.15 1965 X 29 22 88.07 60 36 68.42 55.43 54.30 47.13 54.39 1966 X 34.96 3678 50.09 30 53 26 15 37 50 18.48 1967 X 70.21 71.01 53.92 52.55 55.68 49.12 1968 X 58.46 54.85 61.57 67 92 51.92 1969 X 48.98 39.22 51 92 37.18 1970 X 74.35 68.40 60.21 1972 X 66 23 58.74 1973 X 58.61 accommodations for possible sampling depth- catch differences of Trachurus gayi and Triphotu- rus mexicanus. When the PL are excluded, comparisons be- tween the OL taxa yield generally lower PSI val- ues than those of the total larvae; only 2¥7( of the 18.5-88.1 values are moderate to high (Table 6B). Moderate to high values (69.9-88.1) come from comparisons between 1964, 1965, 1967, and 1969, and in part result from similar proportions of Lampanyctus parvicauda (32.1-41.4%) and Tri- photurus mexicanus (18.2—28.99^) during those years. Moderate values (66.2-74.4) also come from comparisons of 1967 vs. 1968 and 1969 (sim- ilar proportions of L. parvicauda, T. mexicanus, and Diogenichthys [12.2-39.5%!); 1970 vs. 1972 (similar proportions of Vinciguerria lucetia and Diogenichthys [23.4-26.8%!); and 1972 vs. 1973 (similar proportions of Bathylagus nigrigenys, V. lucetia, and Diogenichthys [9.3-25.3%!). Lowest PSI values (<30) result from comparisons of 1966 vs. 1965, 1970, 1972, and 1983, and are due largely to extreme dominance by Normanichthys crockeri (67.4% of OL) in 1966. Recalculations to accommodate for possible depth-related increased catches of T. mexicanus in most cases decrease 1973 and 1983 PSI values (e.g., by 2.6-7.2) and in two cases (1968 and 1970 vs. 1973) change the value characterization from moderate (67.9 and 68.4) to low (63.8 and 64.5). With one exception (1972 vs. 1973, PSI = 67.0) justed values are low. all of the other ad- Species Across-Year Ranked Abundance Patterns Individual species across-year abundance rank- ings demonstrate a variety of patterns. Three pat- terns are shared by nine of the more frequently occurring taxa (Table 7). These involve 1) a group formed by anchoveta and three coastal forms; 2) a group formed by one myctophid and two other mesopelagic species; and 3) a species pair consisting of sardine and a myctophid. An- other species pair (two myctophids) can be formed if the 1973 and 1983 abundances of T. mexicanus are adjusted. Group I includes anchoveta, Normanichthys crockeri, Blenniid D, and Sebastes sp. (Table 7). There is a significant concordance among these species as to years of highest (1966 and 1973) and lowest (1965 and 1983) abundance (W = 0.69, P < 0.01). The abundance rankings of anchoveta and A'^. crockeri (p = +0.79) and of Blenniid D and Sebastes (p = +0.88) are significantly correlated (P < 0.01). None of the correlations between spe- cies of the two pairs are significant due to differ- ences in 1967-68 vs. 1970-72 relative abun- dances. The three Group II species, Diogenichthys later- 12 LOEB and RO.I.XS ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE notus. Bathylagus nigrigenya , and Vinciguerna lucetia, have a concordance of higher abundances in 1972, 1973, and 1983 vs. other years (W = 0.81, P<0.01). The abundance rankings of B. ni- grigenys are strongly correlated with those of D. laternatus and V. lucetia (p = +0.82, P < 0.01 in both cases); the correlation between D. laternatus and V . lucetia is not significant. The abundance rankings of sardine and Mycto- phum nitidulum (Pair I species) are significantly correlated (p = +0.96, P<0.01); both species were rare or absent in 1970 and most abundant in 1972, 1973, and 1983. With adjustment to 1973 and 1983 abundances of T. mexicanus, its abundance rankings are strongly correlated with those of Lampanyctus parvicauda (Pair II species; p = +0.84, P < 0.01). Highest relative abundances of both species were in 1964, 1966, 1967, and 1969. Three relatively frequent species iTrachurus murphyi, Diogenichthys atlanticus , and Diaphus sp.) do not conform to any of the above patterns. If the 1973 Trachurus abundance is adjusted to accommodate for possible undersampling, its abundance pattern is similar to that of anchoveta (p = +0.68, P = 0.05) but not to any of the other Group I species (p = +0.03 to +0.46). VARIATIONS IN ABUNDANCE AND COMPOSITION RELATIVE TO HYDROGRAPHIC CONDITIONS Ichthyoplankton abundances in the 10 years sampled show no consistent patterns relative to warm water-cold water events (Table 8). High PL and OL abundances occurred during strong El Nino events (1972 and 1983) and during cold or transition years immediately following El Nihos (1966 and 1973). Lowest abundances of both frac- tions were associated with the 1965 El Nino and warm 1968. Neither the PL nor OL have signifi- cant correlations with ranked (high to low) July- Table 7. — Larval fish taxa grouped according to similar across-year (1964-83) ranked abundance patterns. Significant agreement of group rankings indicated by Kendall's concordance {W) values. Correlations between species pair rankings indicated by Spearman's rhio (p) values. Significant values at P £ 0.05 are indicated, but note use of multiple testing. " = abundance ranks adjusted to accommodate for possible sampling deptti related catch differences (Table 2). Abundance rank foi r Species 1964 196£ . 1966 1967 1968 1969 1970 1972 1973 1983 Group 1 Engraulis nngens Normanichthys crocken Blenniid D 5 4 4 9 9 9 1 1 1 6 6 3 8 8 5 7 3 7 4 5 10 3 7 6 2 2 2 10 10 8 Sebastes sp. 5 8 1 2 3 8 8 8 4 8 E nngens-N- crocken Blenniid D-Sebastes sp, E nngens- Blenniid D E- nngens-Sebastes sp. N. croc/(en -Blenniid D W = P = P = P = P = P = 0.69 (P + 0.79 (P + 0.88 (P +0.59 +0.44 + 0.61 o o o V II V N. crocken-Sebastes sp. P = + 0.48 Group II Diogenichthys laternatus Bathylagus nigngenys Vinciguerna lucetia 10 9 6.5 5 8 10 7.5 6 8 4 4 4 9 7 6.5 7.5 10 9 6 5 2 2 1 1 3 2 3 1 3 5 B nigngenys-D. laternatus B nigrigenys-V. lucetia D. laternatus-V lucetia W = P = P = P = 0.81 (P + 0.82 (P + 0.82 (P + 0.52 <0.01) <0.01) < 0.01) Pair 1 Sardinops sagax Myctophum nitidulum 7 6 8 9 6 6 9 9 5 6 4 4 10 9 2 1 3 3 1 2 S sagax-M. nitidulum P = + 0.96 (P<0.01) Pair II Lanpanyctus parvicauda Tnphoturus mexicanus ' ' 3 1 5 6 1 4 2 3 9 10 4 2 10 8 6 5 8 9 7 7 L. parvicauda-T. mexicanus" P = + 0.84 (P <0.01) 13 FISHERY BULLETIN VOL H6, NO 1 Table 8. — Range, mean, standard deviation and ranked (high to low) values of sea ichthyoplankton sampling periods, 1964-83. N = number of observations. N.A. ^ data not (1976), Bernal et al. (1982), and Kelly and Blanco (1983). 1964 1965 1966 1967 1968 Temperature (°C) N 85 128 72 55 81 Range 13.5-17.9 13.6-18.1 13.7-17.1 13.3-16.0 14.3-18.6 X 15.7 16.5 15.3 14.8 16.5 (S) (1.0) (1.0) (0.8) (0.8) (0.9) Rank 7 3.5 8 9 3.5 Salinity (%o) N 84 124 72 4 81 Range 34.51-34.99 34.80-35.40 34.53-35.03 34.74-34.82 34.69-35.23 X 34.74 35.06 34.83 34.78 34.92 (S) (0.12) (0.12) (0.11) (0.04) (0.12) Rank 9 2 6 8 4 Hydrographic condition: Cold El Nino Transition Cold Warm September mean temperature and salinity val- ues; larval diversity (mean number of taxa/tow) also shows no correlation with these values (Table 9). Within the PL, anchoveta were most abundant during years immediately following El Nihos (cold 1966, transition 1973), the 1972 El Nirio, and cold 1970; lowest abundances were dur- ing the 1965 and 1983 El Nihos (Table 8). There are no significant correlations between ranked anchoveta abundances and ranked values of tem- perature or salinity (Table 10). Sardine larvae were most abundant during and after the 1972 El Niiio; prior to this moderate abundances and rela- tively large percentage contributions to the ichthyoplankton occurred only during the warm 1968-69 period (Tables 5, 7). Lowest sardine abundances were during cold years 1964, 1967, and 1970 and the 1965 El Nino. Despite low abun- dances during the 1965 El Nino, there is a signif- icant positive correlation between ranked sardine abundance and temperature (p = +0.69, P < 0.05). Ranked larval anchoveta and sardine abundances are not correlated (p = -0.07). The only apparent warm-cold year abundance pattern among the OL categories is that of the coastal taxa; this group had lowest abundances during the 1965, 1972, and 1983 El Nihos and highest abundances in subsequent 1966 and 1973 transition years. The ranked abundance pattern of this category has negative correlations (P < 0.05) with ranked temperature (p = -0.69) and salinity (p = -0.68) values (Table 9). Both the myctophid and other mesopelagic categories appear to have abundance patterns unrelated to warm-cold hydrographic conditions (Table 9). Table 9. — Correlations of across-year abundance ranks of zooplankton and ichthyoplankton categories with ranked (high to low) mean temperature and salinity values from nine July-Septem- ber sampling periods off northern Chile, 1964-83. Correlations based on Spearman's rho tests. Significant values at P s 0.05 are indicated, but note use of multiple testing. PL = larvae of pelagic species; OL = other larval taxa. Temperature Salinity (^■C) (%o) Zooplankton -0.48 -0.37 PL -0.10 -0.17 OL -0.02 + 0.05 Larval diversity: (mean no. taxa/tow) + 0.12 0.00 Myctophids + 0.02 +0.12 Other mesopelagic taxa + 0.28 + 0.12 Coastal taxa -0.69 P<0.05 -0.68 P = 0.05 Zooplankton biomass values show negative but nonsignificant correlations with temperature and salinity (Table 9). This is in agreement with the time-series analysis results of Bernal et al. (1983) which demonstrated no consistent relations of zooplankton biomass with cold- or warm-water events. Species Groups and Hydrographic Conditions The species groups formed by similarity of between-year abundance ranks demonstrate both positive and negative correlations with cold- and warm-year conditions (Table 10). Group I and Pair II and their member species have negative correlations with ranked temperature and salinity values indicating a tendency for higher 14 LOEB and ROJAS: ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE surface temperature (°C) and salinity (%o) values during July-September available Hydrographic "condition" descriptions for these penods are from Robles et al. 1969 1970 1972 1973 1983 Temperature ("C) N 35 N A 110 43 24 Range 14.6-17.5 15.5-18.5 14.3-17.4 15.1-19.0 X 16.0 17.2 15.8 17.2 (S) (1.0) (0.7) (0.9) (0.9) Rank 5 1.5 6 1.5 Salinity (%o) N 35 N.A. 108 43 20 Range 34.60-35.19 34.75-35.96 34.60-35.13 34 81-35.39 X 34.87 35.05 34.82 35.09 (S) (0.16) (0.18) (0.13) (0.19) Rank 5 3 7 1 Hydrographic condition: Warm Cold El Nirio Transition El Nino Table 10. — Correlations of across-year abundance ranks of larval fish species groups and member species with ranked (high to low) mean temperature and salinity values from nine July-September sampling periods, 1964-83. Correlations based on Spearman's rho tests. Group correlations based on ranks of summed within- year ranks of member species. Significant values at P < 0.05 are indicated but note use of multiple testing. " = abundance ranks adjusted to accommodate for apparent large sampling depth- related catch differences. Temperature Salinity Species (C) (%o) Group 1 -0.78 P < 0.05 -0.78 P<0.05 Engraulis nngens -0.52 -0.62 Normanichthys crocken -0.69 P < 0.05 -0.72 P<0.05 Blenniid D -0.74 P < 0.05 -0.77 P<0.05 Sebastes sp. -0.67 -0.55 Group II + 0.27 + 0.20 Diogenichthys laternatus + 0.45 + 0.50 Bathylagus nigngenys + 0.27 + 0.20 Vinciguerna lucetia + 0.12 -0.14 Pair 1 + 0.70 P < 0.05 + 0,49 Sardinops sagax + 0.69 P < 0.05 + 0.50 Myctophum nitidulum + 0.61 + 0.35 Pair II -0.62 -0.57 Lampanyctus parvicauda -0.68 P = 0.05 -0.48 Tnphoturus mexicanus" -0.50 -0.53 abundances during colder, lower salinity periods. The rankings of Group I (based on ranks of summed within-year member species ranks) are significantly correlated (P < 0.05) with both temperature and salinity (p = -0.78 in both cases). Within this group the rankings of Normanichthys crockeri and Blenniid D are cor- related (P < 0.05) with temperature and salinity (p = -0.69 to -0.77); the correlation of Sebastes with temperature is also relatively strong (p = -0.67). Within Species Pair II, Lampanyctus parvicauda abundance has a relatively strong negative correlation with temperature (p = -0.68). Group II and Pair I and their member species have positive correlations of ranked abundance with temperature and (with one exception) salin- ity values (Table 10) suggesting a tendency for higher abundances during warmer, higher salin- ity conditions. These correlations are all non- significant and generally weak for Group II and its member species (Bathylagus nigrigenys , Vin- ciguerria lucetia, and Diogenichthys laternatus). Pair I has a positive correlation (P < 0.05) with temperature (p = +0.70) primarily due to sardine abundance ranks. Species Percentage Composition Relative to Hydrographic Conditions Ichthyoplankton percentage composition shows no striking warm year vs. cold year related patterns. Total ichthyoplankton composition comparisons between years of "similar" hydro- graphic conditions do not give overall higher PSI values than do comparisons between years of dif- ferent conditions (Table llA). PSI values (ranges, means, and proportions of high and moderate val- ues) from comparisons of cold, transition, and warm years are similar. However, highest values (91.6-94.9) come from cold vs. transition year (1970 vs. 1966 and 1973) and between-transition year (1966 vs. 1973) comparisons. Additionally, intercomparisons of the transition, warm, and El Niilo years give relatively lower values than do cold-year comparisons. Comparisons between El Nino years give generally PSI low values. 15 FISHERY BULLETIN: VOL. 86, NO. 1 Table 11. — Within- and between-hydrographic period ichthyoplankton composition comparisons presented as range, mean, and standard errors of percent similarity index (PSI) values and numbers (N) out of total comparisons hiaving moderate to high (e.g., -65) values. A. Total larvae. B. OL (other larval taxa) fraction. PSI values for Cold Transition Warm El Nino A. Total larvae Cold years Range 67 8-822 62.7-94.9 62.8-82.5 7.3-79.1 (1964, 1967, 1970) X 76.6 79.4 73.0 46.5 (SE) ( 4.5) ( 5.2) ( 3.6) ( 9 6) N 33 5/6 4 6 49 Transition years Range 57.8-65.1 3.3-79.4 (1966, 1973) X 91.9 625 39.7 (SE) ( 1-6) (12.9) N 1/1 1/4 2/6 Warm years Range 188-76.4 (1968, 1969) X (SE) 79.4 52.3 ( 9.2) N 1/1 2/6 El Nmo years Range 220-45.3 (1965, 1972, 1983) X (SE) N 31.9 ( 7.0) 03 B. OL Cold years Range 46 0-758 30 5-68 3 47.0-85,2 332-88 1 (1964, 1967, 1970) X 58.6 48.0 64.6 57.6 (SE) ( 8,9) ( 5.7) ( 5.4) ( 5.9) N 1,3 1/6 3/6 3/9 Transition years Range 36.8-67.9 18 5-66.2 (1966, 1973) X 37.5 51.7 41.0 (SE) ( 6.4) ( 7.8) N 0/1 1 4 1 6 Warm years Range 37,2-68,4 (1968, 1969) X (SE) 58.5 53,1 ( 5.2) N 0/1 1/6 El Nino years Range 54.3-58.7 (1965, 1972. 1983) X (SE) N 55.8 ( 1.5) 03 The OL percentage composition similarly does not demonstrate clear hydrographically related patterns (Table IIB). As with the total larvae, comparisons of cold vs. other years yield most of the moderate to high PSI values. Highest values (88.1 and 85.2) come from comparisons of cold vs. El Nino (1965 vs. 1967) and cold vs. warm (1964 vs. 1969) years. Comparisons within and between transition, warm and El Nino years give primar- ily low values. Recalculation of PSIs to accommo- date for Triphoturus mexicanus lowers mean val- ues for comparisons with 1973 and 1983 by only 0.4-2.3 and does not affect the overall results. Chronological Considerations of Species Composition When the total larval and OL PSI data are con- sidered in terms of chronological rather than hy- drographic periods, various patterns become ap- parent (Table 12). For the total ichthyoplankton, comparisons within the 1964-69 data set and be- tween this and the 1970-73 data set give similar means, ranges, and proportions of moderate to high values. In contrast, comparisons within the 1970-73 data set provide more similar values and a significantly higher mean value than results from comparisons within the 1964-69 set (Z test, P < 0.01). This suggests that, despite the varied hydrographic conditions represented during the 1970-73 period, conditions were favorable for a repeated fairly similar anchoveta-dominated ichthyoplankton assemblage during July- September months. Chronologically grouped comparisons of the OL fraction provide somewhat different patterns from those of the total ichthyoplankton (Table 12B) and indicate a marked change in species 16 LOEB and ROJAS: ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE proportions between 1964-69 and later years. Comparisons within the 1964-69 OL data set yield all of the high and most of the moderate PSI values. The mean PSI value from these compari- sons is significantly higher (P < 0.05) than that from 1964-69 vs. 1970-73 comparisons. As with the total larvae, the 1970-73 PSI values are sim- ilar and moderately high and the 1983 compari- son values are relatively low compared with other Table 12. — Within- and between-time period ichthyoplank- ton composition comparisons presented as ranges, means, and standard errors of percent similarity index (PSI) values and numbers (A/) out of total comparisons having moderate to high (e.g.. -65) values. A. Total larvae. B. OL {other larval taxa) fraction. PSI values for 1964-69 1970-73 1983 A. Total larvae 1964-69 Range 33 9-825 38 2-91.9 3 3-31.0 X 67.4 67.7 18.2 (SE) { 3.6) ( 3.6) ( 4.3) N 9 15 11 18 06 1970-73 Range 79.1-95.0 7.3-22.0 X 84.4 12.8 (SE) ( 5.2) ( 4.7) N 3/3 03 B. OL 1964-69 Range 29.2-88.1 26.2-67.9 18.5-54.4 X 60.2 48.4 40.7 (SE) ( 4.6) ( 2.5) ( 5.6) N 715 1 18 06 1970-73 Range 66.2-74.4 58.6-60.2 X 69.7 592 (SE) ( 2.4) ( 0.5) N 33 0/3 years. Accommodation for T. mexicanus abun- dance reduces slightly (1.1-3.1) the mean values from comparisons with 1970-73 and 1983 data sets and strengthens the significance of difference (P<0.01) between 1964-69 and 1964-69 vs. 1970-73 mean values. The difference in OL PSI values between 1964- 69 and 1970-83 is, to a great extent, due to abun- dance shifts of Group II and Pair II species. The abundance ranks of all three Group II species (Table 7) indicate significantly higher abun- dances during 1970-83 than in earlier years (Man Whitney U tests, all P's < 0.05). Addition- ally, the averaged abundance estimates from these four years are significantly higher than from earlier years (Z tests; P < 0.01 for Bathyla- gus nigrigenys and Vinciguerria lucetia,P < 0.05 for Diogenichthys spp.). Together B. nigrigenys , V. lucetia, and Diogenichthys spp. (primarily D. laternatus ) contributed 46.5-72.8^7^ of the OL col- lected during 1970-83 compared with 5.8-37.1% during 1964-69 (Table 13). In contrast. Pair II species Lampanyctus parvi- cauda and Triphoturus mexicanus were rela- tively less abundant during 1970-83 than in pre- vious years. These two species contributed 21-66% of the OL during 1964-69 compared with <22% during 1970-83 (Table 13). With adjust- ments to T. mexicanus abundance these species proportions in 1973 and 1983 decrease to 11.8% and 13.2%, and those of the Group II species in- crease to 49.9% and 71.4%, respectively. The rela- tive abundance decrease of Pair II was primarily Table 13.— Percentage contribution by dominant OL (other larval taxa) species collected sampling penods, 1 964-83. Species arranged according to group affiliations based on across patterns. during July-September ■year ranked abundance Percentage contribution for 1964 1965 1966 1967 1968 1969 1970 1972 1973 1983 Group 1 Normanichthys crocken Sebasles sp. Blenniid D 1561 0.36 4.81 1.94 0.52 67.40 060 4.00 3.51 0.63 4.45 5.15 1.79 940 22.13 2.16 12.52 1.02 1.19 15.72 0.35 4.31 0.44 Group total 20.78 2.46 72.00 8.59 16.34 24.29 12.52 2.21 20.38 0.44 Group II Bathylagus nigrigenys Vinciguerria lucetia Diogenichthys spp. 3.63 2.81 5.63 6.60 0.52 24.45 1.61 0.70 350 608 3.82 19.11 13.87 6.94 16.33 0.75 2.16 12.15 9.01 26.78 26.78 24.14 23.37 25.29 17.53 9.31 19.68 13.78 1.62 49.96 Group total 12.07 31.57 5.81 2901 37.14 15.06 62.57 72.80 46.52 65.36 Pair 1 Myctophum nitidulum 0.36 — 0.10 — 0.89 0.75 — 2.05 2.15 2.01 Pair II Lampanyctus parvicauda Triphoturus mexicanus 37.21 28.86 41.40 20.44 15.77 4.99 3947 18 17 24.16 10.29 32.11 26.37 7.13 14.27 12.67 8.40 6.06 11.82 6.78 14.70 Pair total 66.07 61.84 20,76 57.64 34.45 58 48 21.40 21.07 17.88 21.48 17 FISHERY BULLETIN: VOL. 86. NO 1 due to the increased numbers of Group II species and decreased numbers of L. parvicauda . The 1970-83 averaged abundance of L. parvicauda is significantly lower (Z test; P < 0.01) than that of 1964-69. Triphoturus mexicanus averaged abun- dance (both adjusted and unadjusted values) is similar (P > 0.05) between the two time peri- ods. DISCUSSION The northern Chilean ichthyoplankton data set is obviously weakened by lack of information from the 1974-82 period; this missing informa- tion is critical for an appreciation of the temporal extent and relative constancy of the apparent ichthyoplankton composition change in 1970-73 vs. earlier years. This data set also suffers from limited seasonal coverage which prohibits exami- nation of between-year variations in spawning time and intensity as a cause of interannual abundance fluctuations and apparent composi- tion change. However, the existing data set does provide coherent coverage over varied hydro- graphic conditions between 1964 and 1973 and is sufficient to test for correlations with short-term (e.g., year to year) fluctuations in hydrographic conditions. The large interannual changes in abundance and composition of the northern Chilean ichthy- oplankton can to a certain extent be related to interannual changes of hydrographic conditions in the Humboldt Current. This has been demon- strated through correlations of ranked tempera- ture and salinity values and abundances of coastal species, sardine, and Lampanyctus parvi- cauda (Tables 9, 10). The temperature and salin- ity values used in these correlation tests represent ambient conditions during the July- September spawning period and therefore possibly reflect only conditions affecting egg and early larval (e.g., to stages capable of substantial net avoidance) survival. These values do not nec- essarily reflect longer term conditions affecting abundance, distributions, and fecundities of adult populations or later larval survival and recruit- ment. However, there is a generally good corre- spondence between these values and reported longer term hydrographic conditions in the Hum- boldt Current over the 19-yr timespan (e.g., Table 8; Robles et al. 1976; Bernal et al. 1983; Guillen 1983; Bakun 1987). Despite significant correlations between abun- dances of some ichthyoplankton components and temperature and salinity values, there is no ap- parent consistency of total larval or OL species percentage compositions during years of "similar" hydrographic conditions (Table 11). More coherent patterns emerge from consider- ations of the 1964-69 and 1970-73 data sets (Table 12). This chronological separation is also supported by the ranked abundance patterns of the various species groups and pairs {Table 7). Among the least confusing across-year abun- dance patterns demonstrated by the ichthy- oplankton are 1) generally greater abundance of Group II species after 1969, 2) greatest abun- dance of Pair I species after 1970, and 3) pre- dominantly higher abundances of one of the Pair II species prior to 1970 (Table 7). Associated with the Group II and Pair II abundance patterns are large shifts in their relative proportions (Table 13). The shift from relatively large percentage con- tributions by Lampanyctus parvicauda and Triphoturus mexicanus to larger proportions of Diogenichthys spp., Bathylagus nigrigenys , and Vinciguerria lucetia after 1969 is notable. The abundances of these mesopelagic species, unlike those of anchoveta and sardine, are not directly influenced by man's fishing activities and so may be interpreted as indicators of environmental change. Furthermore, the timing of these species absolute and relative abundance changes pre- ceded by several years the dramatic changes in anchoveta and sardine stocks off of northern Chile (Fig. 4) and so cannot be directly related to biological consequences of change in the domi- nant pelagic schooling fish stocks. Although fragmentary, there is evidence for a change in zooplankton biomass values off north- ern Chile (18°-24°S) occurring in 1969 (Fig. 5) which, like OL percentage composition, suggests a possible environmental change. Time series analysis of quarterly zooplankton biomass values during 1964-73 indicate generally lower biomass during 1969-73 relative to the 1964-68 period. As with total larval abundance (Table 9), these zooplankton biomass variations do not appear to be related to warm year-cold year events (Bernal et al. 1983). The changes in OL composition and zooplank- ton biomass suggest that there was subtle but large-scale (low-frequency) environmental tran- sition occurring in the 1969—70 period. Various indications of environmental change occurring about this time are present in long-term physical data bases from Chile and Peru. Predominantly 18 LOEB and ROJAS: ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE 3,0 1955 83 84 Figure 4.— Total catch and catch of dominant species taken in the northern Chilean pelagic fishery, 1955-84. negative sea surface atmospheric pressure anomalies occurred off of Arica from 1960 to 1972 with a strong negative anomaly occurring in 1969; predominantly positive anomalies occurred there after 1972 (Fig. 6A). A similar but less ex- treme change from negative or neutral anomalies to predominantly positive anomalies occurred off Iquique (20"S) in 1970 (Fig. 6B; Kelly and Blanco 1983). Off Peru (5°-15°S) the wind driven turbulent mixing index of surface waters shows a general increase during and after the 1972 El Nino event (Fig. 7A; Bakun 1987). A probable result of this increased turbulence is an increase in standard deviations associated with monthly temperature values; standard deviations above the 30-yr mean generally persisted throughout the year from 1972 to 1984 (Fig. 7B) and suggest increased physical variability and heterogeneity in this later period. Comparable data sets from northern Chile are not available to determine if these lat- ter two features were also characteristic of the Chilean area. How these observed atmospherically related changes could be related to changes in the marine environment off northern Chile is uncertain. It is possible that the observed changes in atmos- pheric pressure off Arica and Iquique have associ- ated changes in advection of water mass and fau- nal sources. Bernal et al. (1983) discussed El Nino related changes in water mass distribution off Chile in 1972 and 1973 relative to cold-year 1967. These changes involved southerly extensions of oceanic subtropical and equatorial subsurface waters, strengthening of the spring-summer ther- mocline, and cessation of coastal upwelling. These authors did not examine water mass distri- butions in the 1968-70 period. However, lowered zooplankton biomass starting in 1969 and the OL composition change starting around 1970 suggest that the hydrographic conditions attributed to the 1972 El Nino may have been an intensification of 19 w ZOOPLANKTON BIOMASS W FISHERY BULLETIN: VOl- 86, NO 1 W 2 1 .0 o ■o c o to I - - 2 — — 1 1 1 1 1 - 19 64 9 S5 1966 1967 1966 969 1970 1971 972 1973 Year Figure 5. — Time series estimates of quarterly zooplankton biomass values from northern Chile (18°-24°S), 1964-73, standardized accordmg to the long term standard deviation. W = warm years; C = cold years. From Bernal et al. 1983. SEA SURFACE ATMOSPHERIC PRESSURE ANOMALIES 3 - 3 - 3 - 2 - B - 1 - /^ 1 V^ \/ L-A /V r^ A r\ f V A \ \ \ r\ \ '' yt \' V 1 - J ^ V H V^ /•M -J V \H 'V 1 ^ yl V \ I960 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 Year Figure 6. — Sea surface atmospheric pressure anomalies off of (A) Arica (18'Sl and (B) Iquique (20''Sl, 1960-82. From Kelly and Blanco 1983. 20 LOEB and ROJAS: ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE TURBULENT MIXING INDEX 350 - 300 - 250 - 200 - 55 65 70 Year STANDARD DEVIATIONS 4 0- 3 5 3 2 5 2.0 I 5 1.0 B Std. dev 12- mo- running meon of std. dev. .— Long - Itrtn meon of tid. dev. 85 Yeor Figure 7.— (A) Surface layer turbulent mixing index values and (B) standard deviations and 12-mo running mean of standard deviations associated with mean monthly sea surface temperature values off of Peru (5°-15°S), 1953-84. Dotted lines represent long-term mean values of each index. From Bakun 1987. conditions initiated during the warm 1968-69 pe- riod; these could possibly have persisted and in- tensified again during the 1976 and 1983 El Nino events. It is also likely that the 1970 change in OL composition is related to onshore advection of northern or oceanic water masses and associated faunal assemblages, but this cannot be confirmed. All involved species are relatively abundant in coastal Peruvian and south eastern tropical Pacific waters (Ahlstrom 1971, 1972; de Castillo 1979; Santander and de Castillo 1979) and in more southern coastal Chilean waters (Table 2), but their wintertime relative abundances in these areas have not been documented. It is also possi- ble that the changes in species composition are related to locally lowered zooplankton abundance (e.g., that the Group II species are relatively more successful than Lampanyctus parvicauda and Triphoturus mexicanus during periods of lowered secondary productivity levels). Alternatively, the observed change could be due to altered seasonal spawning activity which is not treated in the present study. However, whatever the cause, there is evidence for an environmental change in the study area, and this may be also implicated in changes occurring within the PL ichthyoplank- ton fraction. Increased abundances of sardine and Mycto- phum nitidulum (Pair I species) during and after 1972 may be further evidence for a changed marine environment off of northern Chile. Addi- tionally, the significant correlation between lar- val sardine abundance and temperature (Table 10) suggests that elevated temperatures may have been important for increased spawning activity and/or increased success of hatching and early larval survival. Given this observation one may speculate that the increased sardine catches after 1973 (Fig. 4) are related to increased fre- quency of warm-water events in the 1964-84 pe- riod relative to earlier years. Increased sardine 21 FISHERY BULLETIN: VOL. 86, NO. 1 catches in 1973 resulted from apparently ex- tremely good survival and recruitment of individ- uals spawned during the warm 1968-69 period (Serra"^). Similarly, good survival of the large 1972 El Nino spawn could explain the huge catch increases in 1976 and later years (Fig. 4). How- ever, mean biomass estimates of age groups con- tributing to the Chilean fisheries catch from 1974 to 1981 (Serra 1983) indicate increasing contribu- tions after the 1967 year class with marked in- creases beginning with the 1970 year class; this suggests that factors other than temperature (e.g., environmental change starting in 1969-70) may also be responsible for increased larval sur- vival and recruitment. A possible cause is in- creased nearshore influence of equatorial and subtropical waters (Santander and Flores 1983). Because the sardine abundance increase was ini- tiated prior to the 1972 anchoveta decline off northern Chile (Fig. 4), it is difficult to implicate reduced anchoveta competition as the cause of the early sardine population growth in this area. The grouping of anchoveta with three coastal species (Group I), and significant correlation of anchoveta and Normanichthys crockeri larval abundances are extremely interesting and imply that the spawning intensity and/or early stage survival of these four species are influenced in similar ways by interannual changes off northern Chile. Unfortunately, little is known about the natural histories or population abundances of the coastal species. Because of the group composition, it is logical to suspect that coastal processes are important factors influencing their larval abun- dance. The significant negative correlation of the group as a whole, and of two of the coastal species, with ranked temperature and salinity values (Table 10), suggests that coastal upwelling and/or increased coastal influence by subantarctic waters, and theoretically enhanced food supplies, are important factors. Given the present data set and information from recent publications, a case can be made for a low-frequency environmental change influencing the abundances of anchoveta and sardine larvae as well as the larvae of coastal and mesopelagic species during the 1964-84 period. The Chilean OL composition suggests an environmental change (e.g., an atmospherically related oceanic circulation change) starting with the 1968-69 warm-water event. This coincided with apparent successful survival of sardine larvae and markedly increased recruitment by 1968 and later year classes despite varied warm water-cold water events between 1968 and 1973. Physiologi- cal anomalies of Peruvian anchoveta stocks in 1971 suggest that these fishes may have experi- enced environmental change at that time. Unusu- ally low proportions (e.g., 40% vs. typically 90%) of potential spawning-sized fish were sexually mature during the 1971 spawning season and fat content of the 1971-72 catch was anomolously high, indicating unusually low transfer of body fat to gonadal products (Sharp 1980). Starting with the 1972 El Nino was 1) an obvious in- creased incidence of penetration of subtropical surface waters toward the Peruvian coast, 2) co- incidental onshore and southward expansion of sardine spawning activity off both Peru and Chile, 3) southward expansion of Peruvian an- choveta spawning activity into new spawning areas between 14°S and 18°S (e.g., to northern Chile), and 4) a succession of years of poor an- choveta larval survival off Peru and Chile (San- tander and Flores 1983; Serra 1983). Environ- mental conditions favorable for growth of sardine populations, as well as of mackerel and jack mackerel populations, off both Chile and Peru have persisted since the early to mid-1970's (San- tander and Flores 1983; Serra 1983). The lack of Chilean ichthyoplankton data from the 1974-82 period precludes evaluation of the constancy of altered species composition during that time. However, there are indications that change is once more occurring off northern Chile. Preliminary analysis of ichthyoplankton samples collected between Arica and Antofagasta during 4-14 August 1985 indicates a clear dominance by anchoveta larvae at a markedly higher mean abundance level than encountered in the 1964- 83 samples; sardine and Trachurus larval abun- dances are comparable to those in the 1973 sam- ples (Table 14). The other species have not yet been analyzed, but Normanichthys crockeri is Table 14. — Mean abundance estimates and standard errors (num- bers per 10 m2) and percent frequency of occurrence (F) of PL taxa, OL and total larvae collected in 81 1 00-0 m WP2 net samples off northiern Chile (18°-24°S) during 4-24 August 1985. Species (SE) (F) 3Serra, R. Unpubl. manuscr. Subsecretaria de Pesca, Teatinos 120, Piso 11, Of. 44, Santiago, Chile. Engraulls ringens 5,535.3 (1,844.1) (90.4) Sardinops sagax 63.8 ( 24.9) (26.6) Trachurus murphyi 1.2 ( 0.9) (2.1) Otfier species 233.7 ( 28.4) (92.5) Total 5,834.0 (1,838.2) 22 LOEB and ROJAS: ICHTHYOPLANKTON COMPOSITION AND ABUNDANCE once again noted to be among the abundant OL taxa. It will be of great interest to see the recruit- ment resulting from this 1985 anchoveta spawn. If the 1985 anchoveta year class is relatively suc- cessful during this period of continued large sar- dine stocks, it will lend support to the idea that the anchoveta-sardine population fluctuations have been primarily regulated by low-frequency hydrographic events. In a long-term context such events may not be unusual to the Humboldt Cur- rent area. Fish scales present in sediment records from coastal Peruvian waters indicate that sar- dine replacements of typically dominant an- choveta stocks have occurred at infrequent inter- vals over the past 11,000 years (De Vries and Pearcy 1982). In light of this, the ecological events of the past 20 years may be naturally oc- curring, physically mediated, but probably fish- eries enhanced, fluctuations in this hydrographi- cally complex region. Because of our limited time reference, we have not previously acknowledged such fluctuations as being "normal". ACKNOWLEDGMENTS We extend our appreciation to Dick Parrish and Andy Bakun, Pacific Fisheries Environmental Group (SWFC/NMFS/NOAA), whose interest in eastern boundary current comparison studies has made this work possible. We also thank Rodolfo Serra and Gary Sharp for their valuable com- ments and discussions of the manuscript. Appre- ciation is extended to numerous other people in- cluding ship captains and crew and laboratory assistants, especially Hernan Miles, who have helped in sample collection and processing over the 20-yr period. This work was funded by NMFS/NOAA Solici- tation WASC-84-00075. LITERATURE CITED Ahlstrom, E. H. 1971. Kinds and abundance of fish larvae in the eastern tropical Pacific, based on collections made on EASTROPAC I. Fish. Bull., U.S. 69:3-77. 1972. Kinds and abundance of fish larvae in the eastern tropical Pacific on the second multivessel EASTROPAC survey, and observations on the annual cycle of larval abundance. Fish. Bull., U.S. 70:1153-1242. Bakun, A. 1987. Monthly variability in the ocean habitat off Peru as deduced from maritime observations, 1953-84. In D. Pauly and I. Tsukayama (editors), the anchoveta and its ecosystem, p. 46-74. International Center for Living Aquatic Resources Management (ICLARM), Manila. BERNAL, p. a., F. L. RoBLES, AND O. RoJAS. 1983. Variabilidad fisica y biologica en la region merid- ional del sistema de corrientes Chile-Peru. FAO Fish. Rep. 291:683-711. CONOVER, W. J. 1971. Practical nonparametric statistics. John Wiley and Sons, N.Y., 462 p. DE Castillo, O. S. 1979. Distribucion y variacion estacional de larvas de peces en la costa Peruana. Inf. Inst. Mar Peru 63:1- 32. De Vries. T J . and W G. Pearcy. 1982. Fish debris in sediments of the upwelling zone off central Peru: a late quarternary record. Deep-Sea Res. 28(1A):87-109. DiXON. W J., AND F. M. Massey, Jr. 1969. Introduction to statistical analysis. McGraw-Hill, N.Y., 638 p. Guillen, O 1983. Condiciones oceanograficas y sus fluctuaciones el el Pacifico sur oriental. FAO Fish. Rep. 291:607-658. Kelly, R , and J. L. Blanco. 1983. Fluctuaciones ambientales y su relacion con la abundancia de recursos pelagicos en la zona norte-centro de Chile. Inst. Fom. Pesq. 830040, 22 p. Parrish. R H , A Bakun. D M Husby, and C S Nelson. 1983. Comparative climatology of selected environmental processes in relation to eastern boundary current pelagic fish reproduction. FAO Fish. Rep. 291:731-777. Robertson. A 1970. An improved apparatus for determining plankton volume. Fish. Bull., S. Afr. 6:23-26. ROBLES, F L , E ALARCON. AND A ULLOA. 1980. Water masses in the northern Chilean zone and their variations in the cold period (1967) and warm periods ( 1969, 1971-73). Proceedings of the workshop on the phenomenon known as "En Nino", p. 83-174. UNESCO. Sameoto. D 1980. Distribution and abundance of six species of fish larvae in Peruvian waters and their relationship with the physical and biological environment. Bol. Inst. Mar Peru Callao 5:164-170. Santander, H , and O S de Castillo 1979. El ictioplancton de la costa Peruana. Bol. Inst. Mar Peru Callao 4:69-112. Santander. H . and R Flores 1983. Los desoves y distribucion larval de cuatro especies pelagicas y sus relaciones con las variaciones del ambi- ente marino frente al Peru. FAO Fish. Rep. 291:835- 867. Serra. J R 1983. Changes in the abundance of pelagic resources along the Chilean coast. FAO Fish. Rep. 291:255-284. Sharp. G D 1980. Report of the workshop on effects of environmental variation on survival of larval pelagic fishes. In G. D. Sharp (editor). Workshop on the effects of environmental variation on the survival of larval pelagic fishes, Lima, Peru, April-May 1980. Workshop Report No. 28, p. 15- 59. Intergovernmental Oceanographic Commission, UNESCO, Paris. Tate. M W , and R C Clelland 1957. Nonparametric and shortcut statistics in the social, biological and medical sciences. Interstate Printers and Publishers, Danville, IL, 171 p. 23 FISIIKRY BUI.LKTIN VOL HH. NO 1 UNESCO Wyktki, K 1968. Zooplankton sampling. Monographs on oceano- 1967. Circulation and water masses in the eastern equa- graphic methodology 2. Imprimeries Fopulaires, torial Pacific Ocean. Int. J. Oceanol. Limnol 1:1 17-147. Geneva. 174 p. Yashnov, V A WmiTAKKK. K H 1959. A new model of a volume meter for rapid and pre- 1975. Communities and ecosystems MacMillan Pub- cise plankton evaluation undt-r field conditions. Zool. lishing Co., N.Y., 385 p. Zh. (Moscowl 38:1741-1744. 24 ESTIMATION OF NATURAL MORTALITY IN FISH STOCKS: A REVIEW E. F VetterI ABSTRACT The instantaneous rate of natural mortality (M) is an important but poorly quantified parameter m most mathematical models of fish stock dynamics. This report reviews methods used commonly to estimate M for fish stocks, sensitivity of some common fishery models to values chosen for M, and evidence refuting the common assumption that a constant value can be an adequate approximation of A/ within single stocks. With the exception of simple surplus production models (e.g., Schaefer 1954; Pella and Tomlinson 1969) all mathematical models offish stock dy- namics include as a parameter the instantaneous rate of natural mortality (M). The models do not require explicitly any particular form for M; it can be constant or can vary in any imaginable form. But because natural mortality has proved extremely difficult to measure directly, M is as- sumed almost universally to be some constant specific to whatever stock is being modeled. This is particularly true for analyses of commercial fish stocks, which often require estimates of M only for the postrecruit ages. Decreases in natural mortality with increasing age during egg and postlarval stages are so dramatic compared to ap- parent changes during postrecruitment ages (e.g., Gushing 1975) or compared to differences be- tween different sexes, collection sites, seasons, years, cohorts, or stocks within species, that vari- ations in M during these later (postrecruitment) ages are often assumed negligible. Whether this assumption is in fact acceptable is the subject of this report. The answer is no, it is probably not acceptable in most cases. That an- swer follows from the information presented in Sections II through V, with the following conclu- sions: Section II: Current methods for estimating nat- ural mortality: a review of methods used cur- rently to estimate M in fish populations. All of these methods have strong limitations or disad- vantages. ^Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. Manuscript accepted October 1987. FISHERY BULLETIN: VOL. 86, NO. 1. 1988. Section III: Sensitivity of fishery models to choices for M: a review of the sensitivity of some standard fishery models to different choices for input value(s) of M. Existing studies show that the models are sensitive and that sensitivity is affected not only by the values chosen for M, but also by interactions between M and the values chosen for other parameters in the models. Section IV: Evidence for nonconstant M; factors influencing death rate: a review of factors as- sumed or shown to affect M in fish stocks. Avail- able information implies that many such factors exist, acting alone or in concert. Section V: Evidence for nonconstant M; vari- ability within and between groups: a review of existing quantitative evidence for the extent of variability in M between but especially within stocks. Because almost all fishery models focus on single stocks, variability within stocks (as op- posed to between stocks) is the most important question. Some studies show strong differences between mortality rates of various groups offish; some do not. Those which do not have tended to assume there would be none, and have often used catch curve regression analysis to derive a single estimate from data combined over many groups (usually years) of data. The few studies from which it is possible to determine ranges of esti- mates show differences of at least 50 to 100% be- tween minimum and maximum estimates for sin- gle groups (e.g., stocks) offish. The report's major conlcusions are that natural mortality is far from constant for many fish stocks, and that this variability is extensive enough that it should not be ignored. Analyses of 25 FISHERY BULLETIN: VOL. 86, NO. 1 fish stock dynamics need much more rigorous es- timates of within-stock variability (both trends and variance) in M for exploited fish stocks. II. CURRENT METHODS FOR ESTIMATING NATURAL MORTALITY Three methods are used currently or have been proposed to estimate M in fish populations: 1) analysis of catch data, usually from commer- cial fisheries but also from sampling programs specifically conducted for stock assessment (this includes mark-recapture studies), 2) correlations of M with other life history parameters, and 3) estimation of deaths due to predation. I de- scribe below each method in turn, listing both advantages and disadvantages of each. Catch-Analysis Methods Methods for deriving estimates of natural mor- tality from catch data are based on measuring decreases in abundance, either relative or abso- lute, in groups of fish during two or more succes- sive periods of time. Groups may be distinguished on the basis of any identifiable characteristic, e.g., size (length or weight), age, sex, location and time of capture, or some identifiable tag or mark. The most common grouping is by age, for two reasons. First, age has been considered histori- cally the most important factor potentially affect- ing estimates of mortality rate and subsequent results from the most commonly used fishery models (e.g., Heincke 1913; Baranov 1918). This is probably because the methods were developed for temperate water fisheries which tend to have obvious annual reproductive cycles, so that indi- vidual year classes are often relatively easy to distinguish. Second, the earliest method of esti- mating M (catch curve analysis, discussed below) depends on determining the rate during succes- sive ages. Regardless of the grouping criterion, methods for estimating M use generally one of two types of data. The first type is simply subsamples of un- marked catch. These subsamples contain fish se- lected randomly and classified into groups on the basis of size (length or weight). The second type is mark-recaptures, in which previously marked in- dividual fish can be identified and classified after recapture into groups on the basis of this positive identification. Estimates of mortality are usually derived from samples of unmarked fish by analy- sis of resulting "catch curves" (Ricker 1975). Be- cause it has been used so frequently, catch curve analysis is discussed below in some detail. With marking it is possible to follow the history of individual fish, so many different types of esti- mation procedures exist for deriving estimates of mortality from mark-recapture data (e.g., Ricker 1975; Jones 1979; Brownie et al. 1985). Because so many variations are possible, marking experi- ments are discussed only generally, stressing the basic advantages and disadvantages of mark- ing data relative to data from unmarked samples, in deriving estimates of mortality from these data. Size-frequency distributions from unmarked subsamples of catch (the first type of data) are converted usually to age-frequency distributions, on the basis of previously determined relation- ships between age and length or age and weight. Subsequent analyses concentrate on analyzing this resulting curve of age-composition (e.g., Ricker 1975). Abundance usually decreases expo- nentially with size (or age) in this type of sample. Converting the abundances to their logarithmic values often results in a relatively linear decrease during most exploited ages (or sizes), after some initial increase in vulnerability. Graphs of these logged-frequency distributions are usually called "catch curves", and their analysis, "catch curve analysis". "Catch curve analysis" generally con- sists of determining the best-fit straight line through the decreasing portion of the logged- frequency distribution, because if the decrease in abundance is truly exponential, the slope of this line through the log-transformed data is the in- stantaneous rate of decrease in abundance (e.g., Ricker 1975). There are two basic t3T)es of catch curves, dis- tinguished on the basis of when the data were collected and how many groups are represented in the curves. The first, horizontal catch curves, in- cludes data from several groups (e.g., size or age classes) collected at a single point in time (or com- bined from two or more points in time). Thus, horizontal catch curves reflect "ancient history". The individuals contributing to the frequency dis- tribution were not originally all members of the same group. To use this type of catch curve, one must assume that for each successive age, risk of mortality has been historically the same for all individuals achieving that age. If this has not been the case, the catch curves may show various types of curvature in the descending leg, but ab- sence of curvature is no guarantee that the rates have in fact been constant. 26 VETTER: NATURAL MORTALITY IN FISH STOCKS Thus, horizontal curves are subject to the ex- tremely restrictive assumption that the groups from which the data were collected must be in steady state relative to each other, i.e., their rela- tive abundances must be constant through time. If this is true, then a graph of data collected at a single point in time, which may include, for ex- ample, individuals from 5 consecutive year classes displayed as frequencies at 5 consecutive ages, will look the same as the 5 graphs of data that will result from collecting samples during 5 consecutive years (ages) from each of the 5 year classes. If these conditions are not met, simple linear fitting to determine a single estimate for mortality will be inappropriate. The second type of catch curve, longitudinal, includes data collected from a single identifiable group over a protracted period of time. Most often, this will be a single cohort of fish such as single year class, sampled during successive years. Lon- gitudinal curves are not subject to the assumption of steady state, but do share with horizontal catch curves several other severe disadvantages. These include 1) groups must be adequately identifi- able; 2) groups must be closed to migration, so that changes in abundance are due only to fishing or natural mortality, or if migration does occur, it must occur in proportion to the age distributions in the local groups; 3) samples must represent adequately the true composition of the groups in nature; 4) rate(s) of mortality must be relatively constant between groups over time, so that the log-transformed frequency distributions are truly linear (e.g., Jensen 1984); 5) compensatory rela- tionships between stock levels and natural mor- tality, or fishing mortality and natural mortality, must not be present. Methods for estimating M, which assume to greater or lesser degrees that the conditions listed above are met, have been described repeatedly. The methods tend to fall into two categories. Methods in the first category estimate M from catch records of unexploited or lightly exploited groups of fish. In these groups, F equals or ap- proximates zero. Therefore, the observed rate of decrease (Z ) equals or approximates M, because Z equals the sum of F and M (e.g., Heincke 1913; Baranov 1918; Ricker 1947; Beverton and Holt 1957; Robson and Chapman 1961; Pauly 1982; Munro 1982; and among others). Methods in the second category estimate M by determining Z at various levels of fishing effort, then using the observed relationship between Z and effort to predict, via regression analysis or manipulation of various ratios, the value of Z at zero effort (e.g., Silliman 1943; Beverton and Holt 1957; Paloheimo 1961; Lander 1962; Chap- man and Murphy 1965; Paulik and Robson 1969; Gulland 1983; Butler and MacDonald 1979; Fournier and Archibald 1982; Caddy 1984; and others). These methods are most appropriate for analyz- ing catches of unmarked fish. Accurate results depend strongly on accurate measures of catch per unit effort (CPUE) and constant catchability (q) because if these conditions (in addition to those listed above) are not met, observed relation- ships between abundances in different sample groups may not reflect true differences between groups in situ. Marked fish present fewer problems. Advan- tages include 1) concentration on measuring rela- tive rather than absolute differences between abundances of different groups, 2) immigration need not be considered, as entire original groups are known to carry marks, and 3) with suffi- ciently large samples, it becomes possible to test for differences in mortality rate between different groups (e.g., between ages, between sexes, or be- tween sampling sites), rather than having to as- sume that such effects are negligible. Reviews and descriptions of various mark- recapture methods appear in papers by Seber (1973), Ricker (1975), Jones (1979), and Brownie et al. (1985). Some of the newer types of marking analyses can solve many of the most vexing prob- lems associated with traditional catch curve anal- ysis (e.g.. Reed and Davies 1980; Hochbaum and Walters 1984; Burnham and Andersen 1984; Burnham et al. 1984; Brownie et al. 1985). Several disadvantages unique to marking oper- ations counteract these advantages, however, even with the newer methods. These disadvan- tages include various types of mark-induced effects on mortality rates, behavior, and vulnera- bility to capture, as well as mark loss, unrepre- sentative mixing of marked fish with their origi- nal groups prior to recapture (e.g., Ricker 1975), and especially in commercial fisheries, under- reporting or incorrect reporting of recaptures. Both analysis of catch curves from unmarked fish and analysis of mark-recapture data have the advantage of requiring only catch (and usually effort) data, and these data can generally be col- lected by sampling catches from commercial fish- eries. However, in addition to problems specific to each method, they have in common one or more other major disadvantages: 1) inability to distin- 27 FISHERY BULLETIN; VOL 86, NO 1 guish between losses (or gains) from migration or recruitment versus losses due to fishing or natu- ral causes, 2) imprecision in the estimates of M obtained (e.g., Beverton and Holt 1957; Taylor 1958; Bishop 1959; Paloheimo and Dickie 1966 Ricker 1975, 1977; Doubleday 1976; Pauly 1980 Larkin and Gazey 1982; Paloheimo 1980, 1982 Myers and Doyle 1983; Roff 1984), 3) sensitivity to size-specific mortality affecting the estimated age-structure of the group (Ricker 1969), 4) errors in estimates of age, such that abundances-at-age derived from age-length conversions are unrepre- sentative, 5) where analyses are conducted on data combined over two or more cohorts, the un- likely condition that mortality rates were in fact similar for all cohorts, and 6) problems inherent in the analyses themselves (e.g.. Barlow 1984). Disadvantages 1, 4, and 5 may not apply to marked fish. Disadvantage 5 does not apply to single cohorts. But collections from marked groups and single cohorts are still vulnerable to the other problems. Further, although in principle it would be pos- sible to estimate M for different ages, times, or places, most commonly in practice a single, fishery-wide constant M is estimated by pooling data from throughout the fishery. By implication, the analyst is assuming that the exploited stock was more or less in steady-state over all times and areas of catch so that M was relatively constant while that data set was collected and while (his- torically) the observed age-distributions were being created. In fact, substantial evidence exists that M is not constant, either within a single stock over time (age) or between stocks of a given species in different areas (Sections IV and V). A final disadvantage is that catch-curve analy- ses are fundamentally unmechanistic, generated simply by charting changes in abundance. Catch- curve analyses cannot predict the effect of changes in factors that control M; thus there is little hope of predicting M in the future should conditions change. Life History Methods A second approach to estimating the instanta- neous rate of natural mortality in fish stocks is based on the observation that M often correlates strongly with life history parameters, such as growth rate, age at sexual maturity, costs of re- production, and maximum age (Table 1). Typically in such studies, analytical formulas are derived from theoretical relationships be- tween the various parameters (e.g., Beverton 1964), or empirical formulas are derived from re- gression of M against one or more of the parame- ters (e.g., Hoenig 1983). These models have two significant advantages: 1) they require minimal amounts of data, and 2) they are useful in demon- strating broad trends across species and in devel- oping ecological theory. But because they produce only a single and often very imprecise estimate of M for any given group offish, they are not partic- ularly effective for generating precise estimates of natural mortality or for determining the exis- tence or extent of trends and variability in M for given stocks. They will also be no better than the methods used to estimate the values of M used in the regressions. Table 1. — Studies relating instantaneous rate of natural mortality to life history traits in fish. Traits Species Source ^^max. ^'f' ~''-inf' nietabolic rate, Various Beverton and Holt 1959 reproduction ^max. !<• '-inf. "'-asm. f'Shing clupeids, engraulids Beverton 1963 ^^V,n, general Ursin 1967 'max' 'maxbiomass, '*• general Alverson and Carney 1975 growth rate young fish Ware 1975 '-asm. gonad size, condition factor gadoids Jones and Johnston 1977 'max general Blinov 1977 gonad body weight index. Ms/w, r^3„ /.,„ general Gunderson 1980 tV|n(, L|n), k. water temperature 175 stocks Pauly 1980 energy cost of reproduction general Myers and Doyle 1983 max various Hoenig 1983 weight various Peterson and Wroblewski 1983 '<. '-inf. '-asm various Roff 1986 ^Maximum age, 2Von Bertalanffy growth parameter. 3Maximum length. ''Length at age of sexual maturity. ^Maximum weight. 6Age at occurrence of cohort's max biomass. ''Age of sexual maturity. 28 VETTER: NATURAL MORTALITY IN FISH STOCKS Predation Methods A third class of estimators extends single spe- cies cohort analysis to a multispecies assemblage incorporating the major predators and alterna- tive prey of the stock in question. Single species cohort analysis is used to estimate population abundances and annual values for the instanta- neous rate of fishing mortality (F) for single groups, usually year classes, of fish (e.g., Pope 1972; Ricker 1975; Gulland 1983). The multispe- cies extension simply combines cohort analyses for several species (e.g., Anderson and Ursin 1977). The methods all generate estimates of M as the sum of some constant rate of nonpredatory, nonfishing mortality plus the total estimated flux of prey (stock) to each of the major predators. This feeding flux to predators is estimated by first using cohort analyses to reconstruct population sizes of the various groups of predator and prey, then combining these population sizes with ob- served growth rates for the predators and with estimated preferences for various prey. Thus it becomes possible to estimate the predatory com- ponent of M. Versions of the method have been described by Anderson and Ursin (1977), Majkowski (1981), and Pope and Knights (1982). Applications in a marine system (North Sea) have been described by Anderson and Ursin (1977), in an ecosystem context, by Laevastu et al. (1982) and in lake systems by Forney (1977) and Stein et al. (1981). The predation method has been developed pri- marily from analyses of marine systems, espe- cially the North Sea, and much of the literature exists only as "mimeos" or notes associated with ICES (International Council for the Exploration of the Seas) activities. The most readily available discussion of this approach appeared in Mercer (1982), which includes a critical review and dis- cussion by Ursin (1982) of the various methods. Several other discussions appear in Pauly and Murphy's (1982) volume of collected papers from a symposium on theory and management of trop- ical fisheries. Most of these papers specifically address tropical multispecies systems, but the concepts are broadly applicable. References to other, often less accessible, works can be found in these two general references. The predation method is elegant in concept but often difficult to apply. Studies by Forney (1977) and Stein (1981) had the distinct advantages of limited species numbers in a small system, and direct quantification of stomach contents. Yet even in lake systems, the sampling problems of estimating Z, population abundances, and so forth, remain often as intractable as in large marine systems. The two greatest problems are 1) the difficulty in defining vulnerability and prefer- ence functions for the various prey stocks (e.g., Ursin 1982) and 2) the need to include cohort analyses of all the major interacting species, some or many of which may not be available commer- cially (and for which therefore data will be scarce). Despite these problems the approach can cer- tainly generate, for stocks that suffer heavy predatory mortality from other fished stocks, more realistic estimates of M than approaches that simply generate a globally fixed and invari- ant M. More importantly (and in contrast to the age-frequency or life history methods) the preda- tion method has the advantage of being mecha- nistic. Predation-related causes and conse- quences of age, size, site, stock, geographic, or time trends in M can be investigated via pertur- bation and sensitivity analysis in computer simu- lation studies or, alternatively, investigated through analysis of existing catch data. It be- comes possible (not necessarily feasible) to inves- tigate the implications of varying age or abun- dance structures of interacting fishery resources. Thus the predation approach has considerable conceptual appeal for fairly simple systems in which 1 ) predation is the major force controlling prey abundance, 2) predators have few alterna- tive prey, 3) the possibility can be ignored that predators prefer moribund prey which were about to die anyway, and 4) all major species of predator and prey are sought commercially so that data on abundances and feeding preferences are or can be made available. Unfortunately, the number of systems satisfy- ing these requirements appears to be fairly small, and of course where predation is a relatively small fraction of M, the multispecies predation method will be particularly ineffective. III. SENSITIVITY OF FISHERY MODELS TO CHOICES FOR M Although catch-analysis, life history, and pre- dation methods all exist currently for estimating M in fish stocks, in practice the only method used extensively is the first — direct estimation of M from analysis of catch structure. Thus the discus- sion below of model sensitivity to M is based on this type of estimate. The conclusions reached are 29 FISHERY BULLETIN: VOL. 86, NO. 1 not specific to this one method. Model sensitivity to a given derived value of M will be the same, regardless of the method used to derive the value. General Patterns Sensitivity analyses of M in fishery models have evolved through two phases. Earlier studies noted the influence of M on estimates of maxi- mum yield (Fmax) or maximum yield per recruit {(Y/R)jnax\ and on F^ax (the fishing pressure re- quired to produce maximum yield) in Beverton- Holt yield models (Beverton and Holt 1957; Hen- nemuth 1961; Francis 1974; Parks 1977; Bartoo and Coan 1979; Bulgakova and Efimov 1982). More recently, as cohort analyses have become more popular, more attention has been directed toward assessing the influence of M on age- specific estimates of stock sizes (A^, ) and fishing mortalities (F, ) produced by these models (Mur- phy 1965; Pope 1971; Ricker 1971; Agger et al. 1973; Doubleday 1976; Ulltang 1977; Doubleday and Beacham 1982; Pope and Shepard 1982; Sims 1982a, 1982b, 1984). A few other studies have investigated the effect of M on estimates of maxi- mum sustainable yield (MSY) or total biomass (Francis 1976; Deriso 1982; Beddington and Cooke 1983; Tyler et al. 1985). Most of these studies have used a single, invari- ant value for M . Model sensitivity is then as- sessed by comparing model results using some "best" estimate of M, to results using one (or rarely, more) pair(s) of M values some arbitrary percentage above and below the best estimate. Only a few studies exist of the effects of noncon- stant M, where M varies in different groups of fish within a given stock. These include Beverton and Holt's (1957) example of density -dependent M in plaice, and several investigations of age- specific M (Parks 1977; Ulltang 1977; Bartoo and Coan 1979; Sandland 1982; Bulgakova and Efi- mov 1982; Caddy 1984; Tyler et al. 1985). No study to date has specifically addressed the problems of estimating values of M for a full fish- ery analysis, leading from cohort analyses (using M to estimate F, , Ni , and recruitment R ) to esti- mates of yield or yield-per- recruit using the same M(s) and R subsequently in the Beverton-Holt formulas. Also, no study to date has addressed the possibility and consequences of differing patterns of variability in M , although it has been sug- gested in one case (Ulltang 1977) that random variations will be unimportant if the rate is con- stant (on average) over the fished ages. In general, the earlier analyses with yield mod- els assuming a constant M show that higher esti- mates of M lead to 1) lower estimates of y^ax oi" iY/R)jnax (because fewer survive to be caught), 2) higher estimates of Fj^ax 'yo^J must fish a bit harder to catch a given amount of those left), and 3) lower estimates of age at first capture it^.; be- cause it pays to catch them before they die, rather than waiting for them to grow bigger but less abundant). Including density-dependence tends to exag- gerate these trends, at least for plaice in the North Sea (Beverton and Holt 1957). Including age-structured M in yield models also affects the estimates, but not necessarily in a straightfor- ward manner. As described below in the section on numeric results, change in model output for a given change in M depends not just on the values chosen for M, but also on those chosen for the other parameters. M is not an independent parameter in these models. Analyses with cohort or virtual population models which assume a constant value for M show that in general the effect of increasing M is to increase estimates of N, (because the higher M is, the more fish died in addition to those being caught) and to decrease estimates of F,. The data show only Z , which is the sum of M and F, . As- suming Z has been constant, a decrease in F, requires an increase in M. If Z has been variable, the lower F, may be explained on the basis of higher A^,, a smaller proportion of which (F,) would account for the observed catch. The actual effect, particularly on estimates of A^,, is not necessarily that simple. As with yield models, a given change in M does not always produce the same change in model output. The result depends also on values chosen for other parameters; M is not an independent param- eter. In cohort analysis the results (estimates of A'^, and F, ) are particularly sensitive to the relative sizes of F and M (i.e., to the exploitation ratio E = F/(F + M)). The effect of assuming an incor- rect value (or series of values) of M tends to build up as the analysis proceeds backward in time. This is because with every time step backward the catch (C) is inflated by the factor M in order to estimate at that time the size of the entire stock, not just the size of the catch. That is A^, =A^, + i + C,(F, +M)/F, where F, satisfies the catch equation (1) 30 VETTER: NATURAL MORTALITY IN FISH STOCKS C, =N,,i (FJF, + A/)(e' (2) If M is large relative to F (i.e., the exploitation ratio is low), then errors in A^, can increase pro- gressively and become quite large at the younger ages (e.g., Agger et al. 1973; Murphy 1965; Ull- tang 1977; Sims 1982a, 1982b, 1984). Numeric Results Although general responses of various models can be determined simply by inspection of the analytic models themselves, the quantitative change to expect in the result (output) for a quan- tified change in M (input) is not always immedi- ately obvious. This is because M tends to occur more than once in various formulas. For example, M appears in both the numerator and denomina- tor in the solution to the Beverton-Holt yield equation (Ricker 1975). Y = FA^oe'-^'-WxdAM + F) -3e'-*'"V(M +F + k) + 3e(-2*^V(M +F + 2k) e'-s^'-'/CM +F ^3k)). (3) So, rather than derive analytical expressions (e.g., Sims 1984), I resort below to a simpler ap- proach. Sensitivity of fishery models to changes of given magnitude in M is assessed by comparing percent change reported in model response (out- put) to percent change in M (input). In cases for vector (age or density-dependent) M, I have merely described the shape of the M -vector. For these different vectors, I report the percent change in the result due to switching from a vec- tor of one shape to a vector of another shape. Yield Models At least four studies (Beverton and Holt 1957; Hennemuth 1961; Francis 1974; Bartoo and Coan 1979) have shown that errors in estimates of M propagate into roughly equal errors in esti- mates of (y//? )maxj but with sign reversed (Table 2). For example, a 10% overestimate in M will lead to approximately 10% underestimate of (Y/ R 'max- An equally important result is that the actual magnitude of the effect induced depends strongly not just on the error in M, but on the values chosen for the other parameters in the model. In another study, Beddington and Cooke (1983) used the Beverton-Hol formulation to investi- gate the influence of M (constant; 0.1 to 0.8 year"M, t^ (0 to 4 years), and K (the von Berta- lanffy growth parameter; 0.1 to 0.5 year"M on MSY (maximum sustainable yield), expressing the result as "MSY as a % of Bq," where Bq is the initial or recruited biomass. Higher percentages indicate that more of the original biomass is being taken at MSY. Increasing M by a factor of 8 (0.1 to 0.8 year-i) increased MSY/Bq by a factor of about 4 to 8, depending on the particular values of tf. and K. Again, errors in M produced roughly the same relative error in the result; and again the actual effect of any given change in M de- Table 2— Sensitivity of estimated maximum yield per recruit ((V/Rj^ax) 'o changes in instantaneous rate of natural mortality (M) and other input conditions. Sensitivity of (>^'W)max ^^d of changes in Ware expressed as percentage difference from nominal responses at nominal (best-guess) M. Symbols are: t^ = age-at-first-capture, F = instantaneous rate of fishing mortality, M = nominal value for M. Frances (1974) used an age-structured simulation model. All other citations used standard yield- per-recruit analyses. Input conditions % change % change in in M (^/'^Jmax Species Source fc = constant (3.72) F = variable /W = 0.10 -1-50 -50 -20 -^30 fp = variable F = constant (0.73) M = 0.10 -1-50 -50 -60 -^50 F = constant (0.95) M = 0.8 + 20 -20 -21 + 32 /W = 0.8 -HO -10 -14 -M6 M = 0.60 + 2b -20 plaice plaice Beverton and Holt 1957 Beverton and Holt 1957 yellowfin tuna Hennemuth 1961 yellowfin tuna Francis 1974 yellov\rfin tuna Bartoo and Coan 1979 31 FISHERY BULLETIN: VOL. 86, NO. 1 pended on the values chosen for the other parameters. Pope and Garrod (1973) present another exam- ple of sensitivity in MSY to values chosen for M. They describe briefly the consequences of using an incorrect constant for M of cod stocks when estimating the F required to generate MSY (Fmsy*- Underestimating M by 507r (assumed M = 0.1 year^^; true M = 0.2 year"M leads to a choice of Fmsy that is 67% too high. Overestimat- ing M by 50% (assumed M = 0.3 year ^ true M - 0.2 year"M underestimated Fmsy by 50%. The simulations described above tested the ef- fects of choosing alternative constant values for M. Choosing a vector alternative can also have significant effects; again, the magnitude of the effect depends on the values chosen for other parameters. Beverton and Holt (1957) showed that incorporating density-dependence in M for plaice decreased (y/i?)max by 12%, when holding tc constant at 3.72 years and letting F vary. Con- versely, holding F constant and letting t^, vary decreased {Y/R )max by about 37%. Age-dependent values for M were compared with age-constant values by Bartoo and Coan (1979), Bulgakova and Efimov (1982), and Tyler et al. (1985). In their analysis of Atlantic yel- lowfin tuna stocks, Bartoo and Coan found that replacing an assumed constant M of 0.8 year"^ with an age-structured M increasing from 0.1 year~^ at age to 1.2 year~^ at age 7, increased (Y/R )max by 17% (from 6 to 7 kg). Estimating total yield (7,) rather than {Y/R ) and estimating R as a function of constant versus age-specific M in analysis of catch curves for rela- tively unexploited stocks of Pacific ocean perch and Oregon hake, Bulgakova and Efimov (1982) found that replacing a constant (age-averaged) M with age-variable M tended to increase estimated Yi when fish recruited fairly late to the fishery, but decreased Yf if the fish recruited early. This is because of the interaction between the values as- sumed for M (constant or age-variable) and the value calculated for R from each type of mortality curve. Starting with a given value for recruitment at age 6 years (from Efimov 1976), they calculated R twice for ages 4 and 8 years — once with age- averaged M and once with age-specific M. Be- cause in this set of data the age-averaged M was generally higher than the age-specific M at the tested ages of recruitment (ages 4, 6, or 8 years), back-calculations with age-averaged (i.e., con- stant) M predicted fewer recruits than back- calculations with age-specific M. With fewer re- cruits and generally higher M , potential yield at later ages obviously must drop. Differences in predicted potential yield ranged from about -30% at ^4 (age-specific estimate lower than age- averaged estimate, when fish were assumed to recruit to the fishery at age 4 years) to +15% at ^6 (age-specific estimate higher) and to +60% at Tyler et al. (1985) tested (among other things) the effects of ignoring "true" age-structure in M and using instead a constant value in estimating stock biomass using Deriso's (1980) delay- difference model. They did the tests on catch data generated by Walter's (1969) age-structured sim- ulation model of cod, using three different (input) age structures for M in Walter's model. After gen- erating "catch data" from Walter's model, they analyzed the simulated data set using Deriso's model with constant M (= 0.5 year~^). The age structures tested were 1) mortality increasing and then decreasing with age (Walter's original mortality vector spanning ages 3 to 12 years; age- averaged M = 0.55 year~\ range = 0.33 to 0.70 year~^), 2) mortality increasing with age (ages 7 to 12 years; average M = 0.5 year~^, range 0.3 to 0.7 year"^) and 3) mortality decreasing with age (ages 7 to 12 years; average M = 0.5 year"\ range 0.7 to 0.3 year M. In all three cases Deriso's model with constant M misestimated the "true" biomass generated by Walter's model (with age- structured values for M). The differences were relatively small, however: —13% for the increas- ing and then decreasing series, +19% for the de- creasing series, and +4% for the increasing series. These differences were due to the differ- ences in M, and not the differences in model structure; generating and analyzing biomass with the same constant M in both models led to a discrepancy of only 0.5%. By analogy to life history patterns in other adult animals, M (after recruitment into most fished stocks) is more likely to increase with age than to cycle or decrease. By implication, the sim- ulation results from the increasing series are probably most realistic. If so, the effects of ignor- ing age-structure in favor of using a constant M may be relatively small (5 to 20%), at least for the cod stock simulated in this study. But the results obviously depend again not just on correctly choosing the values for M, but on the values cho- sen for the other parameters. In this case, Tyler et al.'s (1985) results imply that age-structure in M can be relatively unimportant, at least when the 32 VETTER: NATURAL MORTALITY IN FISH STOCKS assumed constant is evenly bracketed by the "true" age-structure in M . Further simulations by Tyler et al. (1985) using a wide range of constant values for M (0.4 to 1.4 year M and the growth rate parameter rho (mean Ford growth coefficient for the fishable stock; 0.46 to 1.6) showed that incorrect guesses of M (and rho) could produce errors up to 1,000^^ in estimated biomass. More realistic ranges for the two parameters (0.4 to 0.8 year" ^ forM, 0.6 to 1:2 for rho), extending about 509r above and below the "true" values for these parameters, induced much lower error in biomass estimates (about the same order of magnitude, 50 to 100% below and above the "true" biomass). As before, changes (errors) of a given amount in M (expressed as fraction or percentage of the original value) ap- pear to produce about the same amount of change (expressed as percent of original value) in simple estimates of yield, depending on the conditions of other parameters in the model. Chatwin (1958) compared estimates of Yj^ax from lingcod populations. Rather than compare constant and age-variable values for M, he as- sumed several different values for an average (constant) M in adults, but assumed that M in- creased from the assumed average for adults to higher values in both juveniles and senescent fish. He reports no quantitative results but states, as found above, that increasing the average M, for a given F, considerably decreased l^max' that decreasing M increased Yrna-x' snd that size at first capture changed relatively little with those changes in M. These comparisons between age-structured versus constant M, or between different constants have demonstrated that effects on results can be large for some combinations of parameters yet small for others. Alternative choices drawn from apparently realistic parameter values lead to rel- atively small differences in estimates of M. Specific amounts of change depend strongly not only on the values chosen for M, but also on the value of M relative to values chosen for the other interacting parameters in the yield models. For most choices of parameter values, sensitivity of output is roughly equal to perturbation of input. Cohort Analyses Effects of interactions between changes in M and values chosen for other parameters is even more obvious in stock reconstruction analyses (e.g., cohort analysis and virtual population anal- ysis (VPA)). These analyses are used to "recon- struct" estimates of stock abundance during pre- vious years, based on catch data and assumptions about the value(s) of M during those previous years. Studies of sensitivity to M in Beverton- Holt types of yield or biomass assessments were usually empirical, based on analyses of catch data from specific fisheries. Studies of sensitivity to M in VPA and cohort analysis include both theoret- ical and empirical studies; i.e., simulations using totally contrived data sets (e.g.. Agger et al. 1973), analyses of specific data sets (e.g.. Pope 1971; Doubleday and Beacham 1982) and combi- nations of analytical evaluations and analysis of specific data sets (e.g., Doubleday 1976; Ulltang 1977; Sims 1982a, 1982b, 1984). Simple analyses of sensitivity to M, in which M is varied but all else is held constant, include 1) Pope's (1972) analysis of Atlantic yellowfin tuna, in which he found that replacing constant M with age-structured M (higher Ms for older fish) produced lower estimates for fishing mortal- ity {F, ) in the later ages, but had little effect on estimates for the younger ages, and 2) Doubleday and Beacham's (1982) statement that I07c error in constant M translated into 9 to 149f error in estimates of i? (at age 3) for cod in the Gulf of St. Lawrence. Somewhat more complicated analyses are pre- sented by Ulltang ( 1977 ) and Sims ( 1982a, 1982b, 1984). Ulltang evaluated the effects on model pre- dictions of F, and A'^, , of several types of variation in M. These included no variation (uniformly con- stant M), M constant within years but varying randomly between years, M varying with age, and M varying with season. Sims evaluated the effects of choosing various constants for M on esti- mates of A'^,, and derived an analytical expression relating variance in M to expected variance in estimates of abundance. In Ulltang's simulations, increasing (decreas- ing) a constant M by 50*^ (from 0.2) decreased (increased) F by about 207c ("true" F's ranging from 0.4 to 0.8). Creating a data set with M vary- ing randomly from one year to the next, then an- alyzing those data with an assumed constant M, Ulltang (1977) found that the Z calculated from the constant-M model was on average the same as the "true" Z from the random-M model. He con- cluded that random fluctuations in M will cancel out during analysis and so can be ignored. Ull- tang assessed the influence of age-dependent M compared with constant M by generating a catch curve with age-variable M (decreasing curvilin- 33 FISHERY BULLETIN: VOL 86, NO 1 early from 0.3 at age 1 to 0.1 at age 10, average about 0.2) and F equal to 0.2, then analyzing the catch with F equal to 0.2 or 0.6, and M equal either to 0.1 or 0.2. Choice of M made little differ- ence in estimates of stock size for the case of high F (0.6), because most of the deaths were due to (observed) fishing. When F was low (0.2), stock- size estimates were much more sensitive to incor- rect choices for M , because most of the deaths were due in this case to M, which was unmea- sured and therefore unobserved. Ulltang (1977) simulated seasonal changes in M by concentrating all deaths in either the first or last quarter of a year. Estimated stock sizes (N, ) changed relatively little; with F = 1.2 and M - 0.4, A'^, was a maximum of 10% higher if all deaths occurred first quarter, 10% lower if all oc- curred in the last quarter. A serious problem with the conclusions reached by Ulltang (1977) is also common to all the other studies discussed above; they are based on rela- tively few combinations of values for the various parameters, and relatively few simulations. For example, the conclusion that random errors in M will tend to even out is intuitively attractive, pro- vided the time scale of variation is short relative to the generation time of the fish. In fact random variation in M did even out in the two sets of simulations he conducted. But the examples he chose included only one set of ages (2 to 10 years), with relatively high values of F (0.5 to 0.8 year"^) compared to the values tested for M (0.1, 0.3 year"M. The gravity of consequences from choos- ing an incorrect M depends very heavily on the size of M relative to the size of F, i.e., on E. Had he chosen different values for his simulations, he might have reached very different conclusions. This is probably the basis for the discrepancy be- tween Ulltang's conclusion that seasonal effects are minor, versus Sims' (1984) conclusion that seasonal effects can be quite large, if M is high. Sims (1984) attempted to overcome this prob- lem (trying to draw general conclusions from the results of simulations based on particular, or rel- atively few, sets of parameters) by analytically deriving formulas for relative error in stock-size estimates, and then testing the formulas with data from actual fisheries. He used this approach twice: once to assess the effects of seasonality (Sims 1982a) and once to consider in general the effects of different choices (errors) for constant M (Sims 1984). But his results (and equations for error) show clearly that error in estimated stock sizes depends on several parameters and that the effects of one can be strongly dependent on the values chosen for the others. Choosing a high M (0.6 year~M and concentrating catch during the first quarter of the year overestimated R by 20%; concentrating catch during the last quarter underestimated R by 23% (compared with the 10% error found by Ulltang). Within the same analysis, reducing M by half (to 0.3 year^ ^) reduced the error in R by half, but the same reduction of error inR was also achieved by leaving M high and reducing F . In assessing specifically the effects of error in M on error in R , Sims (1984) showed very different effects on esti- mates of/? in heavily fished versus lightly fished cohorts of Atlantic bluefin tuna. Changing M by 50% led to changes in estimated R of 60 to 260% in the lightly fished cohort, but only to relatively smaller changes of 35 to 70% in the heavily fished cohort. Again, the magnitude of the error in model predictions depended not just on the mag- nitude of M, but on its relationship to the other parameters in the catch equation, particularly F. Errors (expressed as percentage change in out- put for a given change in input) in model output in the simulations described above, all of which tried to use apparently realistic values for model parameters, rarely exceeded 50%, and were often less than the error introduced into values chosen for M . By implication, the effects of incorrectly guessing M may be relatively unimportant if M is relatively small (e.g., in this situation not more than about 0.5 year"M and relatively invariant, although the actual magnitude of effect due to any given percentage change in M depends on the values chosen for other parameters. So, inaccurate estimates of M might be impor- tant or they might not. It all depends on the mag- nitude and variability of M within a given stock (or group). Although untested, it seems likely that estimates of M for groups in which M varies little and is relatively low, are more likely to be reasonably accurate than estimates of M from groups in which M is large and variable. The fol- lowing section reviews evidence that M does in fact vary both within and between groups of fish, and the succeeding section reviews evidence for the magnitude of that variability in ostensibly similar groups. IV. FACTORS INFLUENCING DEATH RATE Despite the fact that in most fishery models, M is assumed to be constant for all exploited ages in 34 VETTER; NATURAL MORTALITY IN FISH STOCKS any given stock, abundant evidence exists to the contrary. Natural mortality has been shown to vary with age, density, disease, parasites, food supply, predator abundance, water temperature, fishing pressure, sex, and size. Evidence for rela- tionships between these factors and M, and se- lected references for each, are presented below. Changes in mortality rate with age, within sin- gle groups of fish, have been demonstrated and discussed more frequently than changes with any other factor. References include, among others, Baranov (1918, plaice), Sette (1943, Atlantic mackerel), Ricker (1945, 1947, lake fish; 1969, 1975, various species), Beverton and Holt (1959, many species of marine fish), Beverton (1963, en- graulids and clupeids), Boiko (1964, sturgeon). Gushing (1975, plaice), Blinov (1977, fish in gen- eral), Bulgakova and Efimov (1982, Oregon hake and sea perch), Sandland (1982, fish in general). Smith (1985, clupeoids), Roff (1986, fish in gen- eral). Evidence for changes with senescence for fish in general has been discussed or documented by, among others, Woodhead (1979) and Craig (1984). Although specific patterns vary with species (e.g., Woodhead 1979), in general M is extremely high during egg and larval stages (e.g., 2 to 10% per day in plaice and clupeoids (Cushing 1975; Smith 1985)), falls precipitously during the juve- nile period, becomes relatively stable during in- termediate adult ages and increases again with senescence. But even during these relatively sta- ble mid-adult ages, changes in M with age can be substantial, particularly in short-lived fish (e.g., Ricker 1947, stunted versus "normal" whitefish). Changes in natural mortality rate with size (rather than age) within single groups of fish (usually stocks), have been discussed by Baranov (1918, plaice), Ricker (1969, size-selective mortal- ity in general). Ware (1975, larval fish), and Peterson and Wroblewski (1984, many species). Differences in natural mortality rate between populations of the same species in different envi- ronments, or even in different areas of a single environment (e.g., a single lake) are documented by Ricker (1947), Kennedy (1954), and Schupp (1978). Year-to-year differences in natural mor- tality rates of single stocks from a given area are shown by Pope and Knights (1982, plaice) and by Henderson et al. (1983, whitefish). Density- dependent changes in M are discussed by Bever- ton and Holt (1957), Cushing (1967), Tyler and Gallucci (1980), Backiel and LeCren (1978), Jones (1982), and others. Differences in M between sexes have been documented by Beverton and Holt (1957, plaice), Ricker (1947, rock bass), and others. Changes in natural mortality rate related to the cost of reproduction have been discussed by Jones and Johnston (1977), Roff (1984), and others. Other factors that affect M either alone or in combination with other factors include disease and parasitism (reviewed by Lester 1984), starva- tion (Hewitt et al. 1985; Theilacker 1986: larval anchovy), physiological state (Smith 1985), and fishing pressure (Ursin 1982; Munro 1982). Addi- tional examples are cited by Beverton and Holt (1957), Anderson and Ursin (1977), Sissenwine (1984), and Hunter (1984). Most of the factors listed above (e.g., age, size, sex) are indirect influences on M . The most im- portant factor directly affecting natural mortality rate is probably predation; this is implied by a large body of literature describing changes in prey community composition and abundance fol- lowing changes in composition and abundance of predators (e.g., Carpenter et al. 1985). Direct evidence that predators account for most natural mortality in fish stocks is difficult to gather (Section II). To quantify the fraction of M due to predation, one must know, not only rela- tive changes in abundance, but absolute popula- tion density of all predators and prey together with consumption rates and prey preferences of all the predators. Although this is rarely possible, at least two studies from freshwater systems do present quantified estimates of predatory mortal- ity in relation to available prey. Forney (1977) quantified predation mortality in a relatively simple, unmanipulated lake system where there were few species of predator and prey. Combining stomach-content estimates of prey consumed with trawl-sample estimates of predator and prey abundance, he concluded that 30 to 100% of yel- low perch production was consumed by walleye, their principal predator. In a manipulated sys- tem. Stein et al. (1981) assessed predatory mor- tality of young tiger muskellunge after they were stocked in a small pond and lake. During the time of the study, a single predator (largemouth bass) accounted for 25 to 45% of losses to natural mor- tality. In marine systems evidence for the relative im- portance of predation can be gleaned from com- paring total natural mortality with estimated predatory mortality based on abundance of preda- tors and feeding preferences. For example, multi- species cohort analyses reported by Pope and 35 FISHERY BULLETIN: VOL. 86, NO. I Knights ( 1982) show predatory mortality as 80 to 909c of M for age-0 cod, whiting, and haddock in the North Sea (the fraction of M due to predatory mortahty cannot be assessed accurately in the older ages because predators appropriate to these sizes were not included in the analysis). In an- other example, estimating M from energy flow models, Sissenwine (1984) demonstrated that predation in the Georges Bank ecosystem can ac- count for all production by prey fish; nonpreda- tory mortality was negligible. Thus a multitude of factors, acting alone or in concert, can be expected to produce variations in M between individuals within single groups of fish, as well as between groups. Differences can be expected between species, between stocks within species, and from place-to-place and time- to-time within given stocks. In the following sec- tion, I will review more completely existing evi- dence for, and the extent of, this expected variability in M . V. VARIABILITY WITHIN AND BETWEEN GROUPS As discussed above (Section III), simulation studies generally show that effects of choosing a particular value or set of values for M can range from insignificant to considerable, depending in part on the model used, in part on the values chosen for other parameters, and in part on the form chosen for the estimate(s) of M. Authors suggest that in the future, simulations should be conducted with a range of values for M, to bracket probable values (e.g., Beverton and Holt 1957; Tyler et al. 1985). The problem with this advice is identifying the appropriate range and distribution of M for any given group offish. Obviously, wide ranges for M will lead to great discrepancies between model predictions based on one end of the range or the other. It has been shown above, however, that model output can be relatively insensitive to small changes in M . This is particularly true if F is much larger than M (i.e., if the stock is highly exploited so that losses to fishing far exceed losses to natural mortality). The problem is determining whether, for a given stock in situ, changes in M are in fact large or small. Compensatory changes in M, in response to changes in F, will further confound the problem, because variations in M will then be a function of the value(s) of F, in addition to the suite of other factors that may be affecting estimates of M . M does appear to vary considerably between groups offish. Estimates of M compiled by Pauly (1980) (Fig. 1) for 175 stocks and species offish worldwide differ greatly between groups, ranging from a minimum of about 0.1 year"^ to several unusual values as high as 7.0 year ^ Even within a group as ostensibly homogeneous as the tunas, the range of estimated mortality constants spans the majority of the common values (0.2 to 2.0 year"i. Murphy and Sakagawa 1977). Estimates of variability in M within groups of fish are much less common, but are actually more important than the obvious differences between groups with obviously different characteristics such as differing lifespans. Most fishery analyses are directed toward understanding or predicting dynamics of single stocks (single groups offish). The most important considerations for natural mortality parameter values in these single- species analyses are whether and if so over what values M varies for the group of fish in question. But measuring trends or variability in natural mortality rates withingiven groups (e.g., stocks) is difficult and, with the exception of trends with age, rarely attempted. This is primarily because the only extant methods for estimating M depend either directly or indirectly on analysis of catch data (Section II), and catch data are prone to many well known (but largely unsolved) prob- lems. Problems with analysis of catch data fall into two general categories: 1) problems with sam- pling procedure, such that fish are caught or counted out of proportion to their true abundance and 2) problems with fish appearing or disappear- ing from the "unit stock" due to causes other than birth or natural mortality (i.e., migration, fishing mortality, or tagging mortality), again resulting in catch data that do not represent the true struc- ture of the stock. If sampling biases can be over- come, the problem reduces to partitioning total disappearance offish into fractions owing to fish- ing, tag mortality, and migration. The first parti- tion can be eliminated by studying unfished popu- lations, the second by quantifying tag mortality, and the third by studying only closed or tagged populations. Unfortunately, very few sampled populations satisfy completely even one of these criteria. Re- gardless, we still need at least some crude esti- mates of M in order to determine whether M truly varies enough to invalidate the standard assump- tion in fisheries models that M is effectively con- stant during exploited ages. The question here 36 VETTER NATURAL MORTALITY IN FISH STOCKS >- o z UJ 3 O UJ QC U. 28 26- 24- 22- 20- 18- 16- 14- 12 10- 8- 6- 4- 2- NATURAL MORTALITY ESTIMATES: FROM PAULY (1980) (FISH STOCKS) T nni Ji [,i i"i'"('|"i ' i'i I I I I'i'i I I I I I I I ' l'i'i ' i' n I I I i T i I I I I I I I I I I I I I I I I I I i M i I I I'l'i I 0.5 1.0 1.6 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 ESTIMATED NATURAL MORTALITY (M) Figure 1. — Frequency of estimated instantaneous annual rates of natural mortality (M) in 175 different fish stocks, populations, or species. Estimates include both freshwater and marine species. Data from Pauly 1980. concerns variability of M within groupings that would be used commonly to estimate M , such as stocks of single species, rather than general pat- terns across species. If M truly varies relatively little during these ages (so that the log of catch- at-age decreases linearly with age), if the age structure has been relatively constant histori- cally (so that catch curves are actually linear, rather than curvilinear as seen in stocks with inconstant age structure, e.g., chapter 2, Ricker 1975), and if catch curves actually reflect rela- tively accurately this constancy and low variabil- ity, then most estimates of M derived from analy- sis of appropriately processed catch curves cannot help but be relatively close to the true rate. Model predictions, although in theory sensitive, would in practice be fairly robust to any particular value chosen from the true range of values for M . Despite the potential problems with accuracy or precision of existing estimates from single groups of fish, I list in Table 3 most of the esti- mates available for unexploited populations, and some of the few existing estimates from exploited populations. My purpose is to identify the appar- ent range of variability in M within single stocks. The estimates are drawn from references cited by Pauly (1980) and other sources. Only references that reported multiple estimates for M are in- cluded, thus excluding most of the references re- viewed. Because these estimates are derived from catch data, the stated ranges are "apparent", rather than demonstrably the "true" values. Estimated rates of natural mortality are not particularly constant for either unexploited or ex- ploited groups, and are only slightly less variable within stocks than they are within species. Al- though the range of rates within groups may ap- pear relatively small compared to the total range of rates reported for all fish species (e.g., 0.36 to 0.56 for sauger from Lake Nipigon [Table 3] vs. approximately 0.1 to 3.0 for most species listed by Pauly 1980), the maximum and minimum rates reported for single groups differed by at least 50% in 20 of the 22 comparison listed in 37 FISHERY BULLETIN: VOL. 86, NO. 1 Table 3. In at least one case maximum and mini- mum estimates differ by as much as a factor of 7 (i.e., young vs. old whitefish in Shakespeare Is- land Lake, Ricker 1947, Table 3). The range of reported estimates of M for species (rather than single groups or stocks within a spe- cies as compared above) is even greater. Even the least variable estimates differed by a factor of 1.75 (75%, male vs. female plaice, Beverton 1964). In whitefish, the species for which the most esti- mates exist, maximum estimates are 20 times greater than minimum estimates (Table 3). Table 3. — Ranges in estimates of instantaneous rate of natural mortality in unexploited and exploited fisfi populations. 'Wr^ax'''^min is expressed as ffie ratio between tfie maximum (Mmax) and minimum (/Wmm) values reported for M for that species, Values in parenttieses are total range of estimates and ratios for those species where multiple reports exist. Common Age Species name name Body of water Sex (years) M range 'Wmax^'^mir , Source 1) Unexploited populations: Amboplites rupestris rock bass Nebish Lake m 10-12 1.47-2.1 1.49 Ricker 1947 Nebish Lake f 10-14 1.1-1.6 1.45 Ricker 1947 Nebish Lake both 10-14 1.08-1.56 (1.08-2.1) 1.44 (2.01) Ricker 1947 Stizostedion canadensis sauger Lake Nipigon 8-14 0.36-0.56 1.56 Ricker 1947 Coregonus clupeaformls whitefish Lake Opeongo 6-13 0.53 Ricker 1947 Shakespeare Island Lake 11-27 0.08-0.60 7.51 Ricker 1947 Great Slave Lake 17-22 . 0.71-0.99 1.39 Kennedy 1953 Lake Nueltin 13-15 0.84 Kennedy 1963 Lake IVIcDonald 11-14 9-10 1.34 1.66 (0.08-1.66) (20.75) Kennedy 1963 Kennedy 1963 Leuachthys sardinella Ikroavik Lake 6-10 0.2-1.4 7.00 Wohlschlag 1954 Cnstovomer namayacush Great Slave Lake 1-26 0.31-1.61 5.19 'Kennedy 1954 Great Slave Lake 15-23 0.49-0.92 1.88 'Kennedy 1954 Great Slave Lake 15-23 0.52-0.75 (0.31-1.61) 1.44 (5.19) 2Kennedy 1954 Perca fluviatilus perch River Thames m 3-8 0.56-0.98 1.75 Williams 1967 River Thames f 3-7 0.2-0.64 3.20 3Williams 1967 River Thames juv 3-5 0.53-1.69 (0.2-1.69) 3.19 (8.45) 3Williams 1967 Leuciscus leuciscus dace River Thames 5-11 0.36-1.31 3.64 "Williams 1967 Alburnus alburnus bleak River Thames 3-8 0.6-2.4 4.00 ■•Williams 1967 Rutilus rutilus roach River Thames River Stour 2-11 3-12 0.22-1.38 0.44 (0.22-1.38) 6.27 (6.27) "Williams 1967 Mann 1973 Cheilodactylus macropterus tarakihi Chatham Islands Chatham Islands New Zealand 5-35 5-22 0.03 0.08 0.15 (0.03-0.15) (5.00) Vooren 1977 Vooren 1977 Vooren 1977 II) Exploited Populations: Pleuronectes platessa plaice North Sea f 5-13 0.08 Beverton 1964 North Sea m 5-13 0.14 (0.08-0.14) (1.75) Beverton 1964 Brevooiiia petronus gulf menhaden Gulf of fvlexico, 1-3 0.7-1.6 2.29 5Ahrenholz 1981 Central (1969, 1971) Gulf of fvlexico, 1-3 0.88-0.98 1.11 SAhrenholz 1981 Eastern (1969, 1971) Gulf of IVIexico, 1-3 1.17-1.23 1.05 SAhrenholz 1981 Western (1969, 1971) Gulf of IVIexico, 1-3 0.95-1.2 1.26 SAhrenholz 1981 all areas (1969, 1971) (0.7-1.6) (2.29) Gadus morhua cod North Sea 0.5-1 0.59-1.46 2.47 7Pope and Knights 1982 Coregonus clupeaformls whitefish Lake Huron 3.8 0.34-1.67 4.91 SHenderson et al. 1983 'Increasing with age. 2Year to year variation (1946-52); ages 15-23 combined. 3Not consistent with age. "Generally increasing with age. SAssuming 20% tag loss rate, spor tag loss rates from 10 to 30%. 78 different years (1967-75). 823 different year classes (1947-75). 38 VETTER: NATURAL MORTALITY IN FISH STOCKS As discussed previously, these different esti- mates can lead to at least as great a difference in results produced by fishery analyses such as yield models or stock reconstruction analyses (Sec- tion II). Reported differences in estimates of M for whitefish stocks living in Shakespeare Lake com- pared with other relatively small lakes (e.g., Lake McDonald, Table 3) are particularly significant. Because both stocks are of the same species and living in more or less similar environments (small lakes), one might easily (and incorrectly) assume that both have the same rate of natural mortality; but they did not. The lower rates oc- curred in the stock occupying a small lake with no predators. This is a clear example of the effect that environment, particularly the predator envi- ronment, can have on the realized rate of natural mortality in a fish stock. Obviously, choosing a single appropriate constant for this species would be difficult. Choosing an appropriate species- specific constant for some of the other species with multiple estimates might be difficult as well (e.g., rock bass, lake trout, perch, roach, tarakihi, or menhaden. Table 3). None of these studies from either unexploited or exploited stocks support the assumption that M is constant for any given stock or species, nor are these within-stock ranges particularly narrow. In addition, treatment of the original catch data may have in some cases obscured the "true" vari- ability. Ricker (1947) and Kennedy (1953, 1954, 1963), for example, use a 3-yr smoothing tech- nique to reduce the effects of unequal recruit- ment; this also serves to reduce variability that may actually be due to differences in natural mor- tality. Also, single estimates from data collected during only one or two years of sampling (e.g., Wohlschlag 1954; Williams 1967; Mann 1973; Vooren 1977) can be seriously biased by annual changes in either recruitment or mortality rates. If the estimates reported above are even approxi- mately accurate, it is apparent that the range of possible values for M is wide, and that variability can be considerable even within single stocks. A solution to this problem of choosing a reason- able value for M, at least for long-lived fish, is suggested by the possibility that variation in M (not just the mean value) may be related to max- imum lifespan. Fish that live for many years must naturally have lower mortality rates than more short-lived fish. These lower rates may also be less variable in the longer lived stocks, if as in many other biological processes, variability is proportional to the mean. This could account for the ubiquity and apparent effectiveness of the constant 0.2 year"\ used almost universally for the long-lived (20 to 30 years) and well-studied fish stocks from northern European seas (e.g., Beverton 1964). If so, assuming a constant M might be valid for these longer lived stocks. Unfortunately, the few studies cited above do not support this attractive idea. Although in gen- eral, mortality rates decease as lifespan in- creases, the variability in estimates does not ap- pear to follow the same trend. This may be due partially to the relatively similar lifespans (10 to 20 years) for most of the species for which esti- mates exist. But the apparent range in rates for the shortest lived species cited above (Ahrenholz 1981, Brevoortia patronus, ages 1 to 3 years, M range 0.7 to 1.6 year"^) is certainly not greater than ranges reported for the longer lived white- fish (Henderson et al. 1983, Coregonus clu- peaformis, ages 10+ years, M range 0.34 to 1.67 year"^). VI. SUMMARY AND RECOMMENDATIONS Thus it appears that rates of M, or at least rates of M derived by existing estimation methods, do in fact vary widely within many fish stocks. Be- cause the variations appear to be considerable and because the results from fishery models can be sensitive to large variations in M , one must conclude that assuming constancy without proof can have serious consequences for fishery man- agement. A better approach may be to discard the notion that a single "best" estimate of M can be found, and instead try to tailor estimates of M to local groups, based on some combinations of the meth- ods discussed in Section III. Obviously, practical considerations of time and resources will limit the accuracy and precision with which M can be esti- mated. Also, the estimates in the studies re- viewed here are prone to all the artifacts men- tioned in the previous sections. True rates of natural mortality, and their variability, are still very poorly known for even the great stocks of commercial fish in temperature regions that have been subject to continuous exploitation for decades. Careful, repeated tagging experiments probably hold the most promise for determining with any reasonable degree of accuracy, rates of natural mortality in fish stocks. But even these have inherent problems that are not easily 39 FISHERY BULLETIN: VOL 86, NO. 1 solved. There remains a great need both for new methods, and refinements of the old. ACKNOWLEDGMENTS This review was initiated in response to a re- quest for Deus ex machina in mortality estima- tion for fish stocks, elicited from participants at the 35th Annual Tuna Conference held at Lake Arrowhead, CA, 20-22 May 1984, during a discus- sion organized in an attempt to identify the single most important but least well-estimated parame- ter in fishery models. The author apologizes for the lack of divine inspiration herein, but extends grateful thanks to Chris Boggs, Andy Dizon, John Hoenig, Pierre Kleiber, Robert Olsen, and two anonymous reviewers for their insightful discus- sion and helpful comments during preparation of this review. LITERATURE CITED Agger, P , I Boetius, and H Lassen 1973. Errors in the virtual population analysis: The efiects of uncertainties in the natural mortality coefficients. J. Cons. Int. Explor. Mer 35:98. Ahrenholz, D W 1981. Recruitment and exploitation of gulf menhaden, Brevoortia patronus. Fish. Bull., U.S. 79:325-335. Alverson, D L , AND M J Carney 1975. A graphic review of the growth and decay of population cohorts. J. Cons. Int. Explor. Mer 36:133- 143. Anderson, K P , and E Ursin 1977. A multispecies extension to the Beverton and Holt theory of fishing, with accounts of phosphorus circulation and primary production. Medd. Dan. Fisk. Havunders, 7:319-435. BaCKIEL, T , AND E D. LECREN. 1978. Some density relationships for fish population parameters. In S. Gerking (editor), Ecology of freshwater fish production, p. 279-302. Blackwell Scientific, Oxf. Barlow. J 1984. Mortality estimation: biased results from unbiased ages. Can. J. Fish. Aquat. Sci. 41:1843-1847. Baranov. F I 1918. On the question of the biological basis of fisheries. Reports of the Department of Fisheries and Scientific and Industrial Investigations, Vol I, Part 1. (Engl, translat.; Library, Southwest Fisheries Center, P.O. Box 271, La Jolla, CA.) BARTOO, N W , AND A L COAN 1979. Changes in yield recruit of yellowfin tuna Thunnus albacares under the ICCAT minimum size regulation. ICCAT Coll. Vol. Sci. Pap. VIII:120-138, BEDDINGTON, J R . AND J G COOKE. 1983. The potential yield offish stocks. FAO Fish. Tech. Rep. 242, 47 p. Beverton, R J H 1963. Maturation, growth and mortality of clupeid and engraulid stocks in relation to fishing. Rapp. P. -v. Cons, int. Explor. Mer 154-44-6J. 1964. Differential catchability of male and female plaice in the North Sea and its effect on estimates of abundance. Cons. Perm. Int. Explor. Mer Rapp. et P.-v. 155, p. 103-112. Beverton, R J H , and S J Holt 1957. On the dynamics of exploited fish popu- lations. Fish. Invest. Minist. Agric, Fish. Food (G. B.), Ser. II, 19:21-42. 1959. A review of the lifespan and mortality rates of fish in nature and their relation to growth and other physiological characteristics. In CIBA Foundation. Colloquium on aging V. The lifespan of animals, p. 142-180. Churchill, Lond. Bishop, Y. 1959. Errors in estimates of mortality obtained from virtual populations. J. Fish. Res. Board Can. 16:73-90. Blinov, V V 1977. The connection between natural mortality and re- production of food fishes and modeling of the processes involved. J. Ichthyol. 16:1033-1036. BoiKO, E G 1964. Evaluating natural mortality in the Azov Zander iStizostedion lucioperca ). Fish. Res. Board Can. Transl. Ser. No. 541, 15 p. Brownie. C , D. R Anderson, K Burnham, and D. S. Robson. 1985. Statistical inference from band recovery data - a handbook. 2d ed. U.S. Fish Wildl. Serv. Resour. Publ. # 156. BULGAKOVA. T I. AND Y N EFIMOV. 1982. A method for forecasting catch by consideration of the dependence of natural mortality on age. J. Ichthyol. 22:24-32. Burnham, K., and D. R Andersen. 1984. Tests of compensatory vs. additive hypotheses of mortality in mallards. Ecology 65:105-112. BuRHAM, K , G C White, and D R Anderson 1984. Estimating the effect of hunting on annual survival rate of adult mallards. J. Wildl. Manage. 48:350-361. Butler, S A , and L. L MacDonald. 1979. A simulation study of a catch-effort model for esti- mating mortality rates. Trans. Am. Fish. Soc. 108:353- 357. Caddy, J F 1984. Method of solving for natural mortality rate from stocks with different schedules of growth and mortal- ity. Can. J. Fish. Aquat. Sci. 41:1226-1230. Carpenter, S R , J F. Kitchell. and J R Hodgson 1985. Cascading trophic interactions and lake productiv- ity. Bioscience 35:634-639. Chapman, D G , and G I Murphy 1965. Estimates of mortality and population from survey- removal records. Biometrics. 21:921-935. Chatwin, B M 1958. Mortality rates and estimates of theoretical yield in relation to minimum commercial size of ling cod (.Ophid- ian elongatus ) from the Strait of Georgia, British Colum- bia. J. Fish. Res. Board Can. 15:831-849. Craig. J F 1984. Aging in fish. Can. J. Zool. 63:1-8. Gushing, D. H 1967. The possible density-dependence of larval mortality and adult mortality in fishes. In J. H. S. Blaxter (edi- tor), The early life history of fishes, p. 103- 111. Springer- Verlag, N.Y, 40 VETTER: NATURAL MORTALITY IN FISH STOCKS 1975. The natural mortality of the plaice. J. Cons. Int. Explor. Mer 36:150-157. Deriso. R B 1980. Harvesting strategies and parameter estimation for an age-structured model. Can. J. Fish. Aquat. Sci. 37:268-282. 1982. Relationship of fishing mortality to natural mortal- ity and growth at the level of maximum sustainable yield. Can. J. Fish Aquat. Sci. 39:1054-1058. DOUBLEDAY. W G 1976. A least squares approach to analyzing catch at age data. ICNAF Res. Bull. No. 12, p. 69-81. DOUBLEDAY. W G , AND T D BEACHAM 1982. Southern Gulf of St. Lawrence cod: a review of multispecies models and management advice. In M. C. Mercer (editor), Multispecies approaches to fisheries management advice, p. 133-140. Can. Spec. Publ. Fish. Aquat. Sci. No. 59, Efimov, Y N 1976. Assessment of the magnitude of the reserves of Pacific hake by the virtual population method. [In Russ.) In Abstracts of All-Union Conference of Young Scientists, Fisheries Exploitation in the Resources of the Worlds Oceans. Forney. J L 1977. Reconstruction of yellow perch iPerca flavescens) coherts from examination of walleye iStizostedwn vit- reum vitreum ) stomachs. J. Fish Res. Board Can. 34:925-932. FOURNIER, D . AND C P ARCHIBALD 1982. A general theory for analyzing catch at age data. Can. J. Fish. Aquat. Sci. 39:1195-1207. Francis. R C 1974. TUNPOP. a computer simulation model of the yel- lowfin tuna population and the surface tuna fishery of the eastern Pacific Ocean. Inter-Am. Trop. Tuna Comm., Bull. 16:234-258. 1976. Relationship of fishing mortality to natural mortal- ity at the level of maximum sustainable yield under the logistic stock production model. J. Fish. Res. Board Can. 31:1539-1542. GULLAND. J A 1983. Fish stock assessment. A manual of basic meth- ods. FAOAViley series on food agriculture. Vol 1. John Wiley and Sons, N.Y., 223 p. Gunderson, D R 1980. Using R-K selection theory to predict natural mor- tality. Can. J. Fish. Aquat. Sci. 37:2266-2271. Heincke, F 1913. Investigations on the Plaice. General report 1. The plaice fishery and protective regulations. Rapp. Cons. Explor. Mer 17A, p. 1-153. Henderson. B A . J Collins, and J. A Reckahn 1983. Dynamics of an exploited population of whitefish (Coregonus dupeaformis I in Lake Huron. Can. J. Fish. Aquat. Sci. 40:1556-1567. Hennemuth.R. C. 1961. Year class abundance, mortality and yield-per- recruit of yellowfin tuna in the eastern Pacific Ocean, 1954-1959. Inter-Am. Trop. Tuna Comm., Bull. 6:1-51. Hewitt. R . G. H Theilacker. and N C H Lo 1985. Causes of mortality in young jack mackerel. Mar. Ecol. Prog. Ser. 26:1-10. Hochbaum. G S . AND C J Walters. 1984. Components of hunting mortality in ducks: a man- agement analysis. Can. Wildl. Serv, Occas. Pap. # 52. Hoenig, J 1983. Empirical use of longevity data to estimate mortal- ity rates. Fish. Bull., U.S. 82:898-903. Hunter, J R 1984. Inferences regarding predators in the early life stages of cod and other fishes. In E. Dahl, D. S. Danielsen, E. Moksness, and P. Solemdal (editors), The propagation of cod Gadus morhua . L. Flodevigen Rapp. 1:533-562. Jensen, A L 1984. Non-linear catch curves resulting from variation in mortality among subpopulations. J. Cons. Int. Explor. Mer 41:121-124. Jones. R 1979. Materials and methods used in marking experi- ments in fishery research. FAO Fish. Tech. Pap. No. 190, 134 p. 1982. Ecosystems, food chains, and fish yield. In D. Pauly and G. Murphy (editors). Theory and management of tropical fisheries, p. 195-240. ICLARM Conf Proc. 9. Jones. R . and C Johnston 1977. Growth, reproduction, and mortality in gadoid fish species. In J. H. Steele (editor). Fisheries mathemat- ics, p. 37-62. Acad. Press, N.Y. Kennedy, W A 1953. Growth, maturity, fecundity and instability in un- exploited whitefish iCoregonus dupeaformis) of Great Slave Lake. J. Fish. Res. Board Can. 10:413-441. 1954. Growth, maturity, and mortality in the relatively unexploited lake trout, (Cristivomer namaychush), of Great Slave Lake. J. Fish. Res. Board Can. 11:827- 852. 1963. Growth and mortality of whitefish in three unex- ploited lakes in northern Canada. J. Fish. Res. Board Can. 20:265-272. Laevastu. T , F Favorite. H A Larkin. 1982. Resource assessment and evaluation of the dynam- ics of the fisheries resources in the northeastern Pacific in numerical ecosystem models. In M. C. Mercer (edi- tor), Multispecies approaches to fisheries management advice, p. 70-81. Can. Spec. Publ. Fish. Aquat. Sci. 59. Lander. R H 1962. A method of estimating mortality rates from changes in composition. J. Fish. Res. Board Can. 19:159-168. Larkin, P. A , and W Gazey. 1982. Applications of ecological simulation models to management of tropical multispecies fisheries. In D. Pauly and G. Murphy (editors). Theory and management of tropical multispecies fisheries, p. 123-140. ICLARM Conf Proc. 9. Lester. R J G 1984. A review of methods for estimating mortality due to parasites in wild fish populations. Helgol. Meersunters. 37:53-64. MA.JK0WSKI, J. 1981. Application of a multispecies approach for assessing the population abundance and age-structure of fish stocks. Can. J. Fish. Aquat. Sci. 38:421-431. Mann. R H K 1973. Observations on the age, growth, reproduction, and food of the roach, Rutilus rutilus (L.), in two rivers in England. J. Fish. Biol. 5:707-736. Mercer, M C 1982. Multispecies approaches to fisheries management advice. Can. Spec. Publ. Fish. Aquat. Sci. 59, 169 p. 41 FISHERY BULLETIN; VOL. 86, NO. 1 MUNRO, J L 1982. Estimation of biological and fishery parameters in coral reef fisheries. In D. Pauly and G. I. Murphy (edi- tors), Theory and management of tropical multispecies fisheries, p. 71-83. ICLARM Conf Proc. 9. Murphy, G. I. 1965. A solution to the catch curve. J. Fish. Res. Board Can. 22:1965. Murphy. T. C, and G T. Sakagawa 1977. A review and evolution of estimates of natural mor- tality rates of tunas. In ICCAT Coll. Vol. Sci. Pap. VI(SCRS-1976):117-123. Myers, R A., and Doyle, R W. 1983. Predicting natural mortality rates and reproduction-mortality trade offs from life-history data. Can. J. Fish. Aquat. Sci. 40:612-620. Paloheimo, J E 1961. Studies on estimation of mortalities I. Comparison of a method described by Beverton and Holt and new linear formula. J. Fish. Res. Board Can. 18:645-662. 1980. Estimation of mortality rates in fish popula- tions. Trans. Am. Fish. Soc. 109:378-386. 1982. Estimation offish mortality. Trans Am. Fish. Soc. 111:535-537. Paloheimo, J E , and L M Dickie. 1966. Food and growth of fishes III. Relations among food, body size, and growth efficiency. J. Fish. Res. Board. Can. 23:1209-1247. Parks, W W 1977. Cohort and equilibrium yield-per-recruit analyses for the Atlantic bluefin tuna fisheries system, accounting for two system configurations and two natural mortality models. ICCAT Coll. Vol. Sci. Pap. VI:385-401. PAULIK, J , AND D S ROBSON. 1969. Statistical calculations for change-in-ratio estima- tors of population parameters. J. Wildl. Manage. 33:1- 27. Pauly, D 1980. On the interrelationships between natural mortal- ity, growth parameters, and mean environmental tem- perature in 175 fish stocks. J. Cons. Int. Explor. Mer 39(2):175-192. 1982. Studying single-species dynamics in a tropical mul- tispecies context. In D. Pauly and G. Murphy (editors). Theory and management of tropical fisheries, p. 33- 70. ICLARM Conf Proc. 9. Pauly, D., and G. Murphy (editors). 1982. Theory and management of tropical fish- eries. ICLARM Conf Proc. 9, 360 p. International Center for Living Aquatic Resources Management, Manilla, Phillipines. PELLA, J J . AND P K TOMLINSON. 1969. A generalized stock-production model. Inter-Am. Trop. Tuna. Comm., Bull. 13:416-496. Peterson. J , and J S Wroblewski 1984. Mortality rate of fishes in the pelagic ecosys- tem. Can. J. Fish. Aquat. Sci. 41:1117-1120. Pope. J G 1971. Note on cohort analysis and age-specific fishing mortality. ICCAT Tuna Meeting Working Paper, 2 p. 1972. An investigation of the accuracy of virtual popula- tion analysis using cohort analysis. ICNAF Res. Bull. 9, p. 65-74. Pope, J G , and D J Garrod. 1973. A contribution to the discussion of the effect of error on the action of catch quotas and effort quotas. Ann. Meet. Inst. Comm. Northwest Atl. C. Fish. 1973, Res. Doc. 119, Serial #3074. Pope, J. G , and B J Knights. 1982. Simple models of predation in multispecies fisheries for considering the estimation of fishery mortality and its effects. In M. C. Mercer (editor), Multispecies ap- proaches to fisheries management advice, p. 64- 69. Can. Spec. Publ. Fish. Aquat. Sci. 59. Pope, J. G., and J, G Shepard 1982. A single method for the consistent interpretation of catch-at-age data. J. Cons. Int. Explor. Mer 40:176- 184. Reed, R. B., and W D Davies 1980. Estimating rate of exploitation from tag returns and fishing effort. Proc. Conf Southeast. Assoc. Fish. Wildl Agencies 34:151-157. RICKER, W. E 1945. Natural mortality among Indiana bluegill sun- fish. Ecology 26:111-122. 1947. Mortality rates in some little-exploited populations of fresh-water fishes. Trans. Am. Fish. Soc. 71:114-128. 1969. Effects of size-selective mortality and sampling bias on estimates of growth, mortality, production, and yield. J. Fish. Res. Board Can. 26:479-541. 1971. Derzhavin's biostatistical method of population analysis. J. Fish. Res. Board Can. 28:1666-1672. 1975. Computation and interpretation of biological statis- tics offish populations. Bull. Fish Res. Board Can. 191, 382 p. 1977. The historical development, /n J. A. Gulland (edi- tor). Fish population dynamics, p. 1-26. Wiley- Interscience, Lond. RoBSON. D S., AND D G Chapman 1961. Catch curves and mortality rates. Trans. Am. Fish. Soc. 90:181-189. ROFF, D A 1984. The evolution of life history parameters in teleosts. Can. J. Fish. Aquat. Sci. 41:989-1000. 1986. Predicting body size with life history model. Bioscience 6:316-323. Sandland, R. L. 1982. Estimation, inference, and data analysis for log- linear models in tagging studies. J. Cons. Cien. 40:290- 303. Schaefer, M B 1954. Some aspects of the dynamics of populations impor- tant to the management of commercial marine fish- eries. Inter-Am. Trop. Tuna. Comm., Bull. l(2):27-56. Schupp. D 1978. Walleye abundance, growth, movement and yield in disparate environments in a Minnesota lake. Am. Fish. Soc. Spec. Pub. 11:58-65. Seber, G a F. 1973. The estimation of animal abundance and related parameters. Griffin, Lond., 506 p. Sette, O E 1943. Biology of the Atlantic mackerel (Scomber scom- brus) of North America. U.S. Fish. Wildl. Serv., Fish. Bull. 50:149-237. Silliman, R p. 1943. Studies on the Pacific pilchard, or sardine iSardinops caerula ) 5. A method of computing mortali- ties and replacements. U.S. Fish Wildl. Serv., Spec. Sci. Rep. 24, 10 p. Sims. S. E 1982a. The effect of unevenly distributed catches on stock 42 VETTER: NATURAL MORTALITY IN FISH STOCKS size estimates using virtual population analysis (cohort analysis). J. Cons. Int. Explor. Mer 40:47-52. 1982b. Algorithms for solving the catch equation forward and backward in time. Can. J. Fish. Aquat. Sci. 39:197- 202. 1984. An analysis of the effect of errors in the natural mortality rate on stock-size estimates using virtual popu- lation analysis (cohort analysis). J. Cons. Int. Explor. Mer 41:149-153. SISSENWINE. M P 1984. Why do fish populations vary? In R. M. May (edi- tor). Exploitation of marme communities, p. 59-94. Springer-Verlag, N.Y. Smith. P E 1985. Year-class strength and survival of 0-group clu- peoids. Can, J. Fish. Aquat. Sci. 42:69-82. Stein. R A , R F Carline. and R S Hayward. 1981. Largemouth bass (Micropterus salmoides) preda- tion on stocked tiger muskellunge. Trans. Am. Fish. Soc. 110:604-612. Taylor, C C 1958. Natural mortality rate of Georges Bank had- dock. U.S. Fish Wildl. Serv., Fish Bull, 58:1-7, Theilacker. G H 1986. Starvation-induced mortality of young sea-caught jack mackerel, Trachurus synimetricus, determined with histological and morphological methods. Fish, Bull,, U,S, 84:1-17, Tyler, A V . and V Gallucci 1980. Dynamics of fished stocks. In R. T. Lackey and L. A. Nielsen (editors). Fisheries management, p. 111- 149, John Wiley & Sons, NY, Tyler, A V , L L Sebring, M C Murphy, and L F. Murphy. 1985. A sensitivity analysis of Deriso's delay-difference equation using simulation. Can. J. Fish. Aquat. Sci. 42:836-841. Ulltang. O 1977. Sources of errors in and limitations of virtual popu- lation analysis (cohort analysis). J. Cons. Int. Explor. Mer 37:249-260. Ursin, E 1967, A mathematical model of some aspects of fish growth, respiration and mortality, J, Fish, Res, Board Can, 24:2355-2452, 1982, Multispecies fish stock and yield assessment in ICES, In M, C, Mercer (editor), Multispecies approaches to fisheries management advice, p, 39-47, Can, Spec. Publ, Fish, Aquat, Sci, 59, VOOREN.C M 1977, Growth and mortality of tarahiki (Pisces: Cheilo- dactylidae) in lightly exploited populations. N.Z. J. Mar. Freshwater Res. ll(l):l-22. Walters, C J 1969, A generalized computer model for fish population studies. Trans, Am, Fish, Soc, 98:505-512, Ware, D M. 1975. Relation between egg size, growth and natural mor- tality of larval fish. J. Fish. Res. Board Can. 32:2503- 2512, Williams, W P 1967, Growth and mortality of 4 species of fish in the Rivers Thames and Reading. J, Anim, Ecol, 266:695- 720, Wohlschlag, D E 1954, Mortality rates of whitefish in an Arctic lake. Ecology 35:388-396, Woodhead, a D 1979, Senescence in fish. In P, J, Miller (editor), Fish phenology: Anabolic adaptations in teleosts, p, 179- 205, Symp, Zool, Soc, Lond, 44, 43 NATURAL HISTORY OF THE RAYS OF THE GENUS MOBULA IN THE GULF OF CALIFORNIA Giuseppe Notarbartolo-di-Sciara1 ABSTRACT Mobulid rays, which abound during summer in the southern Gulf of California, southern Baja California, Mexico, were monitored for a period of four years during a study of their biology. A total of 262 specimens belonging to four species of Mobula were examined. Mobula thurstoni was the most abundant l58'/f of the catch), followed by M. japanica (30%), M. munkiana (9'5f ), and M. tarapacana (3'^). The study area served as a nursery ground for M. thurstoni, a summer feeding and mating ground for M. thurstoni and M. japanica, and a wintering ground for M. munkiana and young M. thurstoni ; M. tarapacana was rare. Data on size, weight, sex ratio, life history, seasonality, feeding habits, behavior, habitat, and symbionts are presented for each species. Size segregation was a common feature of M. thurstoni , M. japanica , and M. munkiana ; sex segregation was not evident. An extreme degree of feeding specialization was noted: summer prey were almost exclusively the eu- phausiid Nyctiphanes simplex: the mysid Mysidium sp. dominated in the winter. A key to the genus Mobula in the Gulf of California is presented as an aid for species identification. This paper reports on natural history aspects of rays of the genus Mobula (Mobulidae), a poorly known group of elasmobranchs commonly called manta rays or devil rays, frequent in the Gulf of California. A good early overview of the family Mobulidae was given by Gill (1908). Cadenat (1960) described the natural history of the mobu- lids of tropical west African waters, based on specimens which were occasionally captured by the local fishermen. However, with the exception of observations carried out with some regularity by Coles (1910, 1913, 1915, 1916a, 1916b) of Mo6- ula olfersi (= M. hypostoma ) and Manta birostris off North Carolina, most of the available litera- ture is purely anecdotal and deals with occasion- ally encountered or harpooned specimens. Long- term field investigations of devil ray ecology and behavior are wanting. As a result, mobulids are among the least known of the batoid taxa. This was recognized by Bigelow and Schroeder (1953) in their comprehensive review of the knowledge of this family. No major contribution to the understanding of any aspect of mobulid biology has since been published. Regular fisheries for mobulids were not known to exist, because mobulid meat is generally con- sidered of little market value. However, in 1981, during a reconnaissance trip to the southern Gulf iScripps Institution of Oceanogi'aphy, University of Califor- nia, San Diego, La Jolla, CA 92093; present address: Museo Civico di Storia Naturale, corso Venezia 55, 20121 Milano, Italy. Manu.script accepted October 1987. FISHERY BULLETIN: VOL. 86, NO. 1. 1988. of California (Mexico), in the vicinity of La Paz, Baja California Sur, a regular, seasonally impor- tant fishery was discovered. This activity afforded the opportunity to study several aspects of the natural history and the ecology of these batoids. Preliminary oral interviews revealed that local fishermen in the Gulf of California knew of, and routinely captured, four species of devil rays, in addition to the well-known giant manta ray, Manta birostris. This information contrasted with the scientific literature, where only two mobulid species, Manta birostris and Mobula lu- casana, were reported for the area (Beebe and Tee-Van 1941; Fowler 1944; Castro-Aguirre 1965). The confusing state of mobulid taxonomy demanded a revisionary work of the genus Mob- ula (Notarbartolo-di-Sciara 1987), and a discus- sion of the systematics of Manta in the eastern Pacific (Notarbartolo-di-Sciara in press). Such ef- fort permitted designation of names for all species of Mobula found in the Gulf of California: M. thurstoni (Lloyd 1908), of which M. lucasana Beebe and Tee Van (1938) is a junior synonym; M. japanica (Miiller and Henle 1841); M. tara- pacana (Philippi 1892); and M. munkiana Notarbartolo-di-Sciara (1987), which had not been described before. Many of the reports of M. lucasana { = M. thurstoni ) from Central and South America (Beebe and Tee- Van 1941; Fowler 1944; Nichols and Murphy 1944; Barton 1948; Castro-Aguirre 1965; Chirichigno 1974; Pequeno 1983) undoubtedly refer to other species of Mob- ula . A key to the genus Mobula in the Gulf of 45 F1SHP:RY BULLETIN: VOL 86. NO 1 0) 3 a: D O 46 NOTARBARTOLO-DI-SCIARA: NATURAL HISTORY OF MOBULA California is presented as an aid to future studies of mobulids from this region. Working relationships were established with the local fishing communities, and their activities were intermittently monitored between 1981 and 1984. Captured rays were examined and mea- sured before their pectoral fins were filleted; stomach contents and reproductive organs were examined later. Information was gathered on size, weight, sex ratios, life history, seasonality, feeding habits, habitat, behavior, and symbionts of four species of rays belonging to the genus Mob- ula (M. thurstoni, M. japanica, M. munkiana, and M. tarapacana ). Detailed descriptions and morphometries of those species are given by Notarbartolo-di-Sciara (1987). The manta ray, Mania birostris, was also occasionally captured (Notarbartolo-di-Sciara in press), but is not treated in the present study. METHODS Although mobulids are locally said to be abun- dant on both sides of the southern Gulf of Califor- nia, for logistic convenience collecting trips were made only to the peninsular coast (Fig. 1). The fishing cooperative based at Punta Arena de la Ventana was selected as the prime collecting site, because mobulids were caught there more consis- tently than at other localities. Fishing camps on Isla El Pardito, at Cueva de Leon, Ensenada de los Muertos, and Bahia de los Frailes were also sources of study material. Other fishing commu- nities, such as Juncalito, San Evaristo, El Sar- gento. La Ventana, and San Jose del Cabo were occasionally visited, but yielded no data because mobulids were not specifically sought by the fish- ermen. Seven field trips were made. Six were short-term (24 January-8 February 1981, 25 November 1981, 16-21 December 1981, 20-23 December 1982, 19-26 January 1984, and 28 Oc- tober- 1 November 1984); one lasted almost six months (26 January-15 July 1983). Mobulids of all available species and sizes are caught with nets and harpoon; their meat is fil- leted out of the pectoral fins for human consump- tion and used as shark bait. Gill nets are either strung just under the surface or are set on the bottom perpendicular to shore, usually at depths between 10 and 200 m. Fishing vessels were 5-7 m fiberglass launches, locally called "pangas", powered by an outboard engine. Fishing occurred within a radius of about 15 km from a base camp. Nets are checked once a day, early in the morn- ing. Rays weighing up to approximately 100 kg were hauled on board, larger specimens were towed ashore. Rays that were dead in the nets, after several hours (i.e., three unsexed specimens of Mobula thurstoni), were often partially de- stroyed by gammarid amphipods (locally called "plaga"), and were unmarketable. Specimens were weighed and measured before being processed by the fishermen. Weights (WT) were taken with calibrated spring-scales. Rays lighter than 20 kg were weighed to the nearest pound with a 50-lb scale; weights were subse- quently converted to kg. Heavier rays were weighed to the nearest kg with a 150 kg scale. Specimens which exceeded 150 kg (all postnatal M. tarapacana ) were cut in four pieces and weighed separately. Ten percent was then added to the total weight to compensate for body fluid loss. All the specimens could not be weighed, as occasionally a large number of rays were beached simultaneously, and because of the intense heat the fishermen could not delay their processing. A set of 29 measurements was taken for mor- phometric analysis and systematic purposes. Methods and results are presented in Notarbartolo-di-Sciara (1987). Measurements relevant to the present paper were disc width (DW), greatest dimension between outermost tips of pectoral fins, pelvic fin length, from anterior margin of vent to tip of pelvic, and clasper length, from anterior margin of vent to tip of clasper. Most specimens were discarded after measuring and sampling. All preserved specimens were deposited in the Marine Vertebrate Collection of the Scripps Institution of Oceanography. Raw data listing all specimens examined and pre- served can be found in Notarbartolo-di-Sciara (1985). The size and shape of the testes were inspected in male specimens, and the ducti deferentes were cut slightly above the genital papilla. Presence or absence of seminal fluid was determined by running a finger in the caudal direction over the ducts anterior to the cut. Clasper length in thousandths of disc width (DW) was plotted against DW to determine size at maturity, and the presence or absence of seminal fluid was noted. Relative size and contents of uteri and nidamental glands were examined in female spec- imens, and right and left ovaries were compared. The diameter of the largest ovum was plotted against DW to determine size at maturity of fe- male Mobula. Eggs were extracted from the germinative epithelium and their greatest di- 47 KISHKRY [BULLETIN: VOL 86, NO 1 ameter was measured to the nearest 0.1 mm with a steel dial caliper. Stomach contents, if not larger than approxi- mately 200 cc, were sampled whole; otherwise, the bolus was made homogeneous by stirring, and about 200 cc were preserved. Stomach content samples were fixed and preserved in 10% buffered formalin. The stomach content of each sample was thoroughly agitated, separated with a plank- ton strainer (mesh size 0.5 mm), rinsed of forma- lin in deionized water, and blotted for 30 seconds on blotting paper. The lump was then molded into a cylindrical shape, and a portion of one end was separated to make up 1 g of wet weight, measured to the nearest 0.1 g. The subsample was then placed with water in a gridded tray, and exam- ined under a dissecting microscope. Contents of the spiral intestine were discarded, because the small crustacean prey was rapidly digested. Feeding habits were analyzed quantitatively by computing the Index of Relative Importance (%IRI) (Pinkas et al. 1971; Hyslop 1980) for each prey species. The IRI combines percentage by number iN), mass (M), and frequency of occur- rence (F) in the formula: mi = (%N + %M) X %F Prey items were identified, when possible, to lowest taxa or species, then the %N of all prey species within each subsample was calculated. When more than one species was present, all items were individually counted. To obtain the %M term of the equation, mean mass was calcu- lated for each species by measuring the length of each item contained in five randomly selected squares on a tray, calculating the average length of each prey species, and obtaining mass values from Miller's (1966) Plankton Conversion Tables, where mass is related to length for all main planktonic taxa. The %N from all subsamples were summed, and the percent from the new sum was calculated, to calculate %IRI for each prey species. The same procedure was applied to %M and 7cF. The total 7(N was then added to the total 7cM, and that sum was multiplied by the total %F, to obtain total IRI for each prey species, from which the %IRI was calculated. When few items (e.g., copepods) were found among a large amount of partially digested euphausiid or mysid shrimps, the possibility of reconstructing the shrimp number within the subsample by count- ing the digestion-resistant eye pairs was dis- carded to avoid bias in favor of the shrimp frac- tion. The following method was adopted instead: all odd prey items were counted and measured, and their total mass was obtained from Miller's tables; this was subtracted from the total weight (1 g) of the subsample. The remaining weight was divided by the mean weight of each individual item, calculated by averaging the lengths of all available intact specimens, and obtaining the cor- responding weight in Miller's tables. A potential biasing factor existed, when only a few prey re- mains were found (e.g., when a relatively uncom- mon item occurred alone in a stomach, therefore contributing a value of 100 7(N and 7 0.5), with the exception of the November to February period, when females were larger ( T-value = 2.331, df = 12, P < 0.05). A total of 105 specimens (210-1,770 mm DW) were weighed. The WT/DW relationship is best described by the equation: WT - 4.817 X 10-8 ^Y)^^)2.^8 r = 0.99 WT is given in kg, DW in mm. The largest speci- men in the sample was a female; DW was not measured because the fishermen had already started filleting the pectoral fins. Calculated DW, regressed from disc length, cranial width, and upper toothband length, was 1,801 mm (multiple correlation coefficient = 0.99). The second largest specimen was also a female, 1,799 mm DW. The largest male had a DW of 1,770 mm and weighed 53 kg. The smallest freshly caught specimen was 876 mm DW and weighed 6.4 kg. The smallest postnatal specimen was a male, the carcass of which was found drying on the beach in Ensenada de los Muertos. Its calculated DW, regressed from toothbands length, was 864 mm (multiple correla- tion coefficient = 0.99). Overall ratio of males to females caught was 1.18 {N = 148). Catch sex ratios varied with sea- son. Females appeared to be dominant in winter (ratio of males to females 0.27; A'^ = 14). The re- verse was true in March, in favor of males. A significant difference from a 1:1 ratio (chi square testP > 0.05) was not noted. Geographical segre- gation, either of sex or size, was not apparent for M. thurstoni during the warmer months when a wide array of size classes and both sexes were found in the same fishing area. Males and fe- males were occasionally harpooned from the same group basking at the surface. This fact argues against behavioral sex segregation. Winter data, however, were suggestive of size segregation at that time of year. It was common knowledge among local fishermen that during the winter months all M. thurstoni caught are small. The bimodal size-frequency distribution for early 49 FISHERY BULLETIN: VOL. 86, NO. 1 OVERALL Xtot= 13456 ±4022 (N=I48) V,f 13449+5513 {N-79) 30 -] H% - I356.3±59.27(N = 66) 20- 10- z UJ o UJ q: u. lij o V) CD < 30nx,<„: 20- 10- '•%% MARCH -APRIL 1211 1±I52 30 (N = I5) II28.81I6I 34(N = I0) I346.0±370.37(N = 4) £22. rr-irT^rrr-iEin. i .h^. 30 20- 10- JUNE Xtot= 1401 22+39.45 (N=76) X^j.:l38l42±55 22(N^43) Xjj:l42703±55 I0(N = 33) NOVEMBER -FEBRUARY Xtot = 95969±42 75(N-I6) Xj.jt = 86733±48l(N-3) X^j =987 46±52 26(N = II) D TOTAL E3 ?? □73. 1 MAY Xtot= 1435 4 + 69 87 (N = 37) Xj.^ = 1436.61 107.98 (N: 20) Xj^ = l433.9±87.27(N=l7) r^m. , .mm i lA. JULY X,0T = 15I275±341.72(N = 4) X^^ = 1409 67 ± 385.40 ( N = 3) K^ :|80I(N=I) 950 1150 1350 1550 1750 950 1150 1350 1550 1750 DISC WIDTH (mm) Figure 2. — Size-frequency distributions of Mobula Ihurstoiii (means ±2 SE). spring (Fig. 2) suggested that larger rays began to move into the area in the spring from their un- known wintering grounds. Mature testes are large, elongated structures attached by the mesorchia to the upper anterior wall of the pleuroperitoneal cavity, on either side of the vertebral column. A large epigonal organ is associated with each testis. Both testes appeared to be functional. They were usually about the same size, although occasionally the left testis was nearly 25''f larger. The paired ducti defer- entes remain separated from each other through- out their length and open into the cloaca at the tip of the urogenital papilla through two distinct pores, rather than merging in the urogenital sinus, as in most elasmobranchs (Daniel 1934), including Manta ehrcnhcrgi (Gohar and Bayoumi 1959). Clasper length was plotted against DW for 43 M. thurstoni (Fig. 3), to determine the onset of male maturation. Rapid increase in relative size of the claspers, beginning at a DW of about 1,500 mm, was concomitant with the incipient presence of abundant seminal fluid in the lower portion of each ductus deferens. The pelvic fin area, and 50 NOTARBARTOLO-ni-SCIARA NATl'RAl. HISTORY OV MORULA especially the tissue at the bases of claspers, of the larger males appeared swollen and congested in May and June, and some of the skin had as- sumed a pink coloration. Similar observations in carcharhinid and odontaspid sharks have been linked with mating activities (Springer 1960; Gilmore et al. 1983). The ovaries in M. thursfoni are paired, elon- gated organs located inside the pleuroperitoneal cavity, analogously to the testes, and are con- nected to a large epigonal organ. Eggs are pro- duced within the germinative epithelium. The largest eggs were found at the anterior end of the ovary. Only the left ovary develops and is func- tional, whereas the size and aspect of the right ovary remains comparable with those of the im- mature stage. Asymmetry is also present in the oviducts, the left uterus being usually the largest in mature females. It consists of a voluminous, thick-walled expansion of the lower tract of the oviduct; its lumen is lined with a highly devel- oped epithelium consisting of elongated, flattened villi (trophonemata), a well-known mobulid (Gill 1908; Setna and Sarangdhar 1950; Wourms 1977) and rhinopterid (Schwartz 1966; Smith and Mer- riner 1986) feature. In several instances both uteri were found to contain a viscous, whitish or greenish substance. Oviducts open separately into the cloaca. A progression of sexual maturity in female mobulids was evident from the exam- ined ovary's developmental condition. In the im- mature female the germinative epithelium is a narrow, leaf-shaped band, tapering at both ends, located opposite to the mesovarium (facing the center of the cavity) along the ovary's longitudi- nal axis. In mature females the germinative ep- ithelium takes over most of the ovary's ventral side, making room for the mature ova. To deter- mine the size at maturity of female M. thurstoni, the diameter of the largest ovarian egg was plot- ted against DW (19 specimens. Fig. 4). An egg growth plateau was not evident, because data on Figure 3. — Relationship between clasper size and body size in Mobula thurstoni. 50 I 40 g^30 _i -^ o UJ " Q. £ 20 OT ^ < s i 10 • ••mtn«l fluid pr»««nl O ••minal fluid aba«nl y o ca> to 500 1000 1500 DISC WIDTH (mm) 15 E > o (- CO LU o < • •• - •• • • •• • • - • • • • •• • • 1 1 ] Figure 4. — Relationship between size of largest ovum and body size in Mobula thurstoni. 1000 1500 DISC WIDTH (mm) 2000 51 FISHKKY BULI.KTIN VOL 86. NO 1 the larger sizes were insufficient, and therefore the maximum egg diameter was not l^^nown. It appears from the scanty available data that fe- male M. thurstoni began to mature at a DW of about 1,500 mm (the point at which the slope of the curve becomes steeper). Mobulo thurstoni, like all mobulids, is a vivi- parous matrotroph ( Wourms 1981 ), the near-term embryo being three orders of magnitude larger than the mature egg. Uteri and nidamental glands of 68 females were inspected. No adujt female exammed from March through June iN = 55) was pregnant. Embryos were found in July and October. Two embryos from females caught in July were near-term. Four embryos found in October were in an early stage of development. All females (A^ = 4) inspected in October were pregnant, and all embryos were in the same de- velopmental stage, suggesting coordinated breed- ing activity. The largest female (DW 1,801 mm) had a single embryo, which appeared to be in the final stage of fetal development, fully pigmented, the yolk sac completely absorbed, and the umbili- cus a mere scar (Fig. 5A). The embryo's DW was 630 mm; its WT, 3.4 kg. It occupied the left uterus with the rostrum pointed forward. Its pectoral fins were folded dorsally, the right pectoral on top. The cephalic fins were almost totally un- rolled and extended ventrally towards the mid- line of the body. The uterus, with the embryo inside, occupied roughly one third of the female's pleuroperitoneal cavity. Lack of space inside the cavity and the distended skin on the abdomen made it apparent that no other embryo had been recently expelled or aborted. Uniparity appears to be a common pattern within the genus Mobiila (Hill 1862; Gill 1908; Coles 1913, 1916b; Barnard 1925; Setna and Sarangdhar 1950; Cadenat 1960; Wallace 1967; Capape and Zaouali 1976; Figueiredo 1977). Only Risso (1826) asserted that Cephaloptera giorna (= M. mohular) may have one or two young, but his statement was not doc- umented. Since the smallest free-swimming spec- imen noted had a DW of 864 mm, the average size at birth for M. thurstoni is probably between 650 and 850 mm DW, and a WT of approximately 4.5 kg. The second largest female (DW 1,799 mm) had also only one embryo in her left uterus. The embryo was unpigmented, with disc 210 mm wide, and weighed 173 g. Like the term-embryo, its rostrum was pointing forward; unlike it, how- ever, its pectoral fins were folded ventrally. Mating, parturition, and early mobulid life his- tory take place in the shallower portion of a popu- lation's range, not an uncommon elasmobranch feature. McLaughlin and O'Gower (1971) dis- cussed inshore movements in the mating Port Jackson shark Heterodontus portusjacksoni , as did Springer (1960) for the sandbar shark Eu- lamia milberti (= Carcharhinus plumbeus). One- year-old gray reef sharks, Carcharhinus am- bhrhynchos, were observed in French Polynesia in shallower waters than adults by Nelson and Johnson (1980). A similar result was reported for the hammerhead shark, Sphyrna lewini , by Clarke (1971) in Hawaii and by Klimley ( 1983) in the Gulf of California. Bullis (1967) hypothesized an upward movement to shallower depths for newborn marbled cat sharks, Galeus arae. There is likely an advantage for juveniles to remain in relatively protected areas during the earlier stages of their life, when they are most vulnera- ble to predation (Springer 1967). Examination of 139 stomachs indicated that M. thurstoni was extremely specialized in its feeding habits. Eighty one (58.3%) stomachs were empty, or had only traces of food (<1 g wet weight). The remaining 58 stomachs (41.7%) had quantifiable contents. All recognizable prey items were planktonic crustaceans (with the ex- ception of a few fish eggs, one nematod, and a small coleopteran, probably ingested accidentally when it was floating). They were listed, ranked by decreasing ^IRI, in Table 1. Mobula thurstoni Table 1 —Prey species found m 57 stomachs of Mobula thurstoni, ranked by decreasing Index of Relative Importance (IRI). N = percentage of prey species by number; M = percentage of prey species by mass; F = percent frequency of occurrence of prey species. Prey species N %N M °/oM %F IRI , %IRI Nyctiphanes simplex 4,940.8 86 70 4.982.0 87 40 877 15,268 1 97,90 Mysidium sp. 6352 11 10 631.2 11.10 12.3 273 ' 1.75 Copepoda 99.9 1 80 74.5 1.31 15.8 49 ; Megalopa larvae 12.0 0.21 4.1 0.07 14.0 3.95 1 Hyperiid amphipods 6.8 0.12 5.8 0.10 5.3 1.18 ' [ 0,35 Fisfi eggs 1,4 0.03 1,8 0.03 5.3 0.302 \ Nematoscelis difi. 0.5 001 1.1 002 1.8 0.050 1 Stomatopod larvae 0.7 0.01 0.6 0.01 1.8 0.041 'J 52 NOTARBARTOLO-DI-SCIARA: NATURAL HISTORY OF MOBULA B i Figure 5. — A: term-embryo of Mobula thurstoni . The scale in the photograph is in centimeters. B; embryo of M. tarapacana . The scale bar equals 5 cm. 53 KISHEKY BULLETIN: VOL 86, NO. 1 fed mostly on adult and juvenile euphausiids, Nyctiphanes .s///?p/c.v. The niysid Mysidium sp. (underscribed, Thomas E. Bowman^) was second in order of importance. Mysids and euphausiids were never found together in the same stomach. The overall importance of the two food items dif- fered by two orders of magnitude. All other prey species found in the stomachs were rare and prob- ably fortuitously ingested. These included one zoea larva and the following copepod species: Undinula vulgaris, Eucalanus subcrassus, E. subtenuis, Ternora discaudata, Scolecithrix danae, Nannocalanus minor, Euchaeta remana, Euchaeta sp., and Labidocera diandra. Diet varied with season (Fig. 6A), with mysids being dominant from December through March, and eu- phausiids during the warmer months. Diet varied with predator size (Table 2): smaller individuals fed both on euphausiids and mysids; the larger rays fed only on euphausiids. This result probably reflects the predominance of smaller rays during winter, when fewer euphausiids are available, rather than an ontogenetic change in food prefer- ences. Table 2. — Size differences in the diet of Mobula thurstoni. %IRI Prey species DW < 1 ,300 (n = 21) mm DW > 1 ,300 mm (n = 36) Nyctiphanes simplex Mysidium sp. Other 79.04 18.13 2.83 99.91 0.09 Two remoras (Echeneididae) were occasionally found on large M. thurstoni: Remora remora (3 specimens; range: 98-200 mm SL), and R. albescens (3 specimens; 93-100 mm SL). Crus- tacean parasites were encountered: Pupulina minor (Copepoda: Caligidae), Ecthrogaleus den- ticulatus (Copepoda: Pandaridae) sparsely on the skin, and Ecthrogaleus disciarai (Benz and Deets 1987) in large patches on the dorsal surface, En- tepherus laminipes (Copepoda: Cecropidae) from the branchial filter plates, Eudactylina oliveri (Copepoda: Eudactylinae) from the gill lamellae, and Kroeyerina sp. (Copepoda: Kroyeriidae) from the olfactory lamellae. Mobula thurstoni was usually observed at the surface in coastal waters of Bahia de la Ventana, Cueva de Leon, and Bahia de los Muertos, often within a few hundred meters of land and occa- 2Thoma.s E. Bowman, Smithsonian Institution, Washington, D.C., pers. commun. 1984. sionally as far as 6 km. When sighted offshore, it was sometimes found over considerable depths O500 m), although it appeared to be more abun- dant in shallower, neritic waters. Mobula thurstoni was always caught in the shallower part of the nets, usually at a depth of < 100 m. The greatest part of the catch, however, was surface- dwelling rays. Beginning in mid-April, numerous M. thurstoni were consistently seen in the early morning hours cruising slowly at the surface. They would frequently pause, conspicuous on calmer days, with the tips of their pectoral fins protruding out of the water. This behavior is well known in mobulids (Norman and Eraser 1937); it has been observed also in connection with mating activities in M. olfersi ( = M. hypostoma ) by Coles (1910). During such occasions, fishermen could easily approach the rays and harpoon them, be- fore startling them and causing them to dive. Re- peated captures within the same aggregation re- vealed that rays of various sizes and both sexes could be found together. While at the surface, M. thurstoni was usually solitary or in small, nonpolarized groups (2-6), rather than in larger aggregations or schools. The species was fre- quently seen jumping out of the water in spectac- ular, often reiterated somersaults; it was recog- nized by the distinctive ventral markings. It is not known to what extent mobulids make use of the sea bottom. Beebe and Hollister (1935) observed a group of 12 small devilfish (most likely Mobula ) lying on the sandy substrate off Frigate Islet, in the British West Indies. Bigelow and Schroeder (1953) speculated that Manta spends much of its time resting quietly on the seafloor. During an experiment organized in conjunction with Sea World of San Diego, aimed at establish- ing whether M. thurstoni could survive in a con- fined environment, five young specimens were captured with gill nets and kept in a large pen (6 m in diameter) anchored in 2.5 m of water in Ensenada de los Muertos. None of the rays sur- vived 24 hours of captivity; the reasons for their deaths were not clear, although the particularly stressful capturing method appeared as a likely cause. During that experiment the negatively buoyant rays (sinking tail-first as soon as they stopped swimming) spent a great deal of time resting on the bottom, and were able to circulate water through their gills while resting, by a syn- chronized maneuvering of the oral valve and of the gill covers (as judged from the flow made vis- ible by the numerous particles suspended in the water). A frequent method of turning around 54 NOTARBARTOLO-Dl-SCIARA: NATURAL HISTORY OF MOBULA XII -II IV VI VII VIM IX XI Figure 6. — A. seasonal variation of the relative importance ofNyctiphanes simplex and Mysidium sp. in the diet of Mobula thurstoni. B. seasonal variation of the abundance of adult and juvenile A^. simplex in the coastal areas of the southwestern Gulf of California (modified from Brinton and Townsend 19801. C. captures of M. tarapacana . D. captures of M. munkiana . E. mean number of daily captures of M. japanica; bars represent 2 SE on either side of the mean. F. mean number of daily captures of M. thurstoni: error bars as in E. (*): although specimens were also captured between December and February, data are not comparable with spring and summer captures because catch effort was minor and inconsistent in the colder months. 55 FISHKKY BULI.KTIN: VOL 86, NO 1 (e.g., when swimming towards the wall of the pen) was to dive vertically in a tight circle until swimming in the opposite direction in an inverted position, and then spinning around the body axis to brmg the dorsal side up, rather than turning by banking to the right or to the left. An indication of the seasonal abundance of M. thiirstoni in the surface waters of the study area was obtained by the mean number of rays caught daily from March to July 1983 (Fig. 6; Table 3). Mean daily catch should be taken as a rough indication of the relative abundance of M. thurstom rather than as a precise index because the fishing effort was difficult to quantify. Mean effort, however, was roughly constant from March through July because the mean monthly number of working boats (about 20) and the number and size of the nets set then was constant. Further- more, the fishermen would harpoon a ray every time they had the opportunity to do so. Peak of abundance was in June, a result which appears to be consistent with the fishermen's past experi- ence, despite the 1983 abnormally high water temperatures (Cane 1983). In July the number of M. thurstom caught had dropped drastically, and most of the catch consisted of M. japanica . No information was obtainable for the August-Sep- tember period. Eighteen specimens were cap- tured during six fall and winter field trips (24 January-8 February 1981; 25 November 1981; 16-21 December 1981; 20-23 December 1982; 19-26 January 1984; 28 October-1 November 1984), but that figure was not comparable with other data because part of the fishing cooperative migrated south to Los Frailes during the cooler months. It is common knowledge, however, among the local fishermen, that M. thurstom in the colder season is present, but in fewer numbers than during the summer. The study area constitutes a feeding, mating and nursery ground for M. thurstoni. The eu- phausiid Nyctiphanes simplex , the main diet item and the only food of the adults when in the area, is the most abundant and widespread euphausiid in the Gulf of California, and has been observed in dense swarms (Brinton and Townsend 1980). Al- though it is found in the study area year-round, its juvenile and adult stages are most abundant between February and August, peaking in June on the west side of the Gulf of California (Brinton and Townsend 1980). The seasonal abundance of M. thurstoni in the southern Gulf thus seems to be closely related to the seasonal abundance of its main prey. It is impossible to describe the general Table 3— Mean number of daily captures of Mobula thurstom a - total monthly number of captures; b monthily number of days of monitonng; X = mean number of daily captures; SE = standard error of tfie mean; so = standard deviation. Month a 6 b X 2SE SD range March 16 038 054 1.1 0-4 April 9 13 0.69 0.69 1.3 0-4 May 35 7 5.00 1.95 2.6 1-9 June 77 15 5.10 332 6.4 0-23 July 5 9 0.56 0.34 1.0 0-3 movement and life history pattern of M. thurstom in the Gulf of California from the fragmentary information available. The scanty data, however, suggest the following: 1 ) adult male and nonpreg- nant adult female M. thurstoni enter the area in spring to feed and to mate, 2) pregnant females segregate from the rest of the population in spring (as is also suggested by the slight predom- inance of males in spring and early summer), 3) gestation period is one year and females give birth to one young every two or more years, 4) the young are born in the study area or near it in midsummer and remain there throughout their early life, and 5) in late summer, when the num- bers of adult and juvenile Nyctiphanes simplex decline due to intense heating of the water (Brin- ton and Townsend 1980), adult M. thurstom leave the area, whereas the young switch their diet from euphausiids to mysids. Further investiga- tions are needed for additional corroboration of these hypotheses. Mobula japanica (MuUer and Henle 1841) Local name: cubana de lomo bianco A total of 78 specimens, 34 males (DW range 1,316-2,386 mm) and 44 females (1,470-2,302 mm), were caught at three stations (Punta Arena de la Ventana, Cueva de Leon, and Ensenada de los Muertos) and adjacent waters, between 16 De- cember 1981 and 13 July 1983. Overall and sea- sonal size-frequency distributions for M. japanica are shown in Figure 7. With the exception of April, when females were larger (T -value = 4.697, df = 3, P < 0.02), there are no significant size differences between the sexes (T -value - 0.535, df= 76, P > 0.5). Most of the rays in the sample were large; only three were < 1,900 mm DW. Twenty-seven specimens (size range: 1,316- 2,285 mm DW; 18.6-115 kg) were weighed. The WT/DW relationship is described by the equation: 56 NOTARBARTOLO-DI-SCIARA: NATURAL HISTORY OF MOBULA WT = 4.29 X 10-10 (DW)'^^ r = 0.98 where WT is given in kg, DW in mm. The overall male to female ratio was 0.89 (A'^ = 781. Females dominated the June through December period; males were predominant in April and May. Sex ratios, however, never significantly differed from 1 (x" test P > 0.05). Sex segregation, behavioral or geographical, was never observed in M. japan - ica (both sexes were caught together in nets and by harpoon); sampling bias (N = 3) may explain why only males were caught in May. Geographi- cal size segregation, by contrast, was an evident feature of sexually mature specimens in the Gulf of California (Fig. 7). No pregnant females were found, although in some specimens the left uterus had a flabby and dilated appearance, suggesting recent delivery. Tissues at the base of the claspers of most of the larger males were swollen and reddened in June and July when the tips of the claspers were flex- ible and the rhipidion could be easily spread, and in doing so a white, viscous fluid would ooze from the hypophyle; all this suggested mating activity. The clasper length-DW relationship for M Japan - ica (Fig. 8) did not exhibit a clear pattern as in M. 30 -I 20 10 - OVERALL Xtot = 2I25 I±3674 (H-7Q) ^^^ = 21016167 12 (N-34) Xjj -2143.3139 28 (N = 44) frrn r-[7^ r-TH , i .^^ I. i DECEMBER Xtot = I86I4± 385 74 {H-5) X^^ = I3I6(N = I) Xj^:l997.8± 352 22(N = 4) rmrTTi. i ., i .. i ., i .. i ,. i .1 ^. i ,. i > u z LU O UJ o en m < 30 n 20 - 10 - APRIL X,o,^ 2156 166 58 (N^5) X^^ = 2l050i28.84(N-3) H% = 22325 + 53.00(N = 2) mm , MAY Xtot = X^^= 2057 33 1122 56 (N^ 3) .■ I I I I .Fxl— I, 30 20- 10 - JUNE Xtot=2I45 5131.32 (N^26) X^2,150 mm), indi- cating that male sexual maturity in M. Japanica begins at a DW of 2,100 (±50) mm. Lack of infor- mation on smaller specimens prevented a clear understanding of the onset of female sexual ma- turity (Fig. 9). Large eggs were found in speci- mens as small as 2,070 mm DW, possibly indicat- ing that female M. japanica began to mature at that size. Only 19 (24%) of 78 specimens had quantifiable stomach contents (>1 g wet WT). The remaining 59 stomachs (76%) were empty or had only traces of food. All M. japanica fed largely on the eu- phausiid Nyctiphanes simplex (Table 4); no mysids were found. Other species occurring in the stomachs, including copepods, megalopa larvae, stomatopod larvae, hyperiid amphipods, caridean decapods (Crangon sp., Pasiphaea sp., and one alpheid decapod), and one cumacean, had an over- all relative diet importance of only 0.38%>. Mobula japanica were therefore very similar in feeding habits to large M. thurstoni. Information on the Figure 9. — Relationship between size of largest ovum and body size in Mobula japanica . 40 • • ^^ E ^ 30 • \. 5 • < • • Q 2 • 3 20 •- • > • • O • • 1- V) • • LU Sio - • • < _l ' 1 1 2000 2200 DISC WIDTH (mm) 2400 Table 4. — Prey species found in 19 stomachs of Mobula japanica, ranked by decreas- ing Index of Relative Importance (symbols as in Table 2). Prey species N %N M %/W %F IRI %IRI Nyctiphanes simplex Copepoda Other 1,890.64 4.61 4.74 99.51 0.24 0.25 1,869.85 1.15 29.00 9841 0.06 1.53 100.00 31 58 36.84 19,792 9.47 65.58 99 62 0.05 0.33 58 NOTARBAKTOI.O-DI SCIARA: NATl'HAI. HISToHY OF MOBVLA winter diet of M.Japanica was lacking, as four of the five specimens collected then had empty stom- aches; the fifth, a large female, contained a small fragment of a partially digested fish carcass. Since quantifiable stomach contents were found only in large rays between April and July, no size or seasonal differences in the diet of M.japanica could be detected. Mohiila Japanica was often found carrying Remora remora , usually seen clinging to the out- side of body, but found once inside a spiracle. Six specimens of R. remora (range 109-217 mm SL) were collected from M. japanica . Only one speci- men of/?, albescens (97 mm SL) was found, in the mouth cavity of a M.japanica. A pilot fish, Nau- crates ductor (Carangidae) also associated with M. japanica, swam alongside a harpooned ray that was being towed inshore and remained for some time at the water's edge, where the ray was beached. Mobula japanica was parasitized by the following crustaceans: Nerocila acuminata (Isopoda: Cymothoidae), Pupulina brevicauda and P . minor (Copepoda: Caligidae) on the skin; Eudactylina oliveri (Copepoda: Eudactylinae) in the gills; and Kroeyerina sp. (Copepoda: Kroyeri- idae) among the olfactory lamellae. Unidentified trypanorhynch cestodes were occasionally found within the pleuroperitoneal cavity. Habitat preference of M . japanica did not ap- pear to differ from that of M. thurstoni . However, the use of the habitat differed seasonally: in April and May, when M. thurstoni was abundant at the surface, M. japanica was never seen, and few specimens were bottom gillnetted during those months. Conversely, M.japanica, in June and July, was seen in the late morning hours at the surface in groups of several individuals swim- ming parallel to the shore. Occasionally speci- mens were seen in water <1 m deep. Mobula japanica is not known to school, and I never ob- served schooling. Coles (1910) reported that M . olfersi (= M. hy- postoma ) utters a "musical, bell-like bark" when dying. A similar account was given by Risso (1810) of Cephalopterus massena {= M . mobular). This information led subsequent authors (Nor- man and Fraser 1937; Bigelow and Schroeder 1953) to wonder whether mobulid rays are capa- ble of producing sounds while in the water. Sound production is a fairly widespread phenomenon among bony fishes (Fish and Mowbray 1970; Tavolga 1971); however, elasmobranchs lack the traditional structures used by teleosts to generate sound, i.e., the swim bladder and bony skeletal parts (Marshall 1962), and recognizable sounds have not been recorded from these animals (Backus 1963). Sound production among elasmo- branchs has been reported only for the Atlantic cownose ray, Rhinoptera bonasus (Myrberg 1981); in that case clicks and scraping sounds were presumably produced with the dental plates, elicited when strongly prodding three rays which were confined in a tank (Fish and Mowbray 1970). Mobula japanica, when beached alive, often emitted a distinctive noise which could have been the equivalent of Coles' "bark". This noise, however, was apparently caused by the periodic, spasmodic contractions of the mandibular, pha- ryngeal, and hypobranchial musculature of the asphyxiating ray, which forced air from the mouth cavity out of the gill openings through the meshlike branchial filter plates. Although under- water sonic recordings have never been made, it seems unlikely that under normal circumstances any audible sound could be produced in this fash- ion by submerged mobulids. This area served as a spring and summer feed- ing and mating ground for adult M. japanica, rather than as a pupping or nursery ground, as indicated by the lack of small-sized specimens. Seasonal abundance of M .japanica in the surface waters was indicated by the catch data (Table 5) and is comparable to the seasonal abundance of M . thurstoni (Fig. 6). No M . japanica were ob- served in March; in April and May they occurred occasionally. By mid-June large numbers ap- peared in the nearshore surface waters near Punta Arena de la Ventana, and were easily har- pooned. Most of the July mobulid catch consisted of M .japanica, when the numbers of M. thurstoni had declined. Data are lacking for the August- October period, therefore it was impossible to tell whether the peak of abundance occurred in July or later. Fishermen's reports were not clear, al- though there was agreement on an overall decline of mobulid abundance in late summer. Mobula japanica fed exclusively on the euphausiid A^yc- tiphanes simplex, and its numbers apparently de- clined concomitant with the late summer decline Table 5. — Mean number of daily captures of Mob- ula japanica (symbols as in Table 1). Month a b X 2SE SD range March 16 0-0 April 5 13 0.4 0.36 0.7 0-2 May 3 7 0.4 0.40 0.5 0-1 June 26 15 1.7 1.02 2.0 0-6 July 39 9 4.3 2.89 4.3 0-14 59 FISHERY BULLETIN: VOL 86, NO. 1 of abundance oC their prey (Brinton and Townsend 1980). Some M . japanica , however, in- cluding individuals both large and small, were found in this region throughout December. Like M. tharstoni , winter catch data are not compara- ble because of differences in fishing effort, when M . Japanica is apparently caught less frequently. Mobula miinkiatia NotJirbartolo-di-Sciara 1987 Local name: tortilla Twenty-four specimens, 10 males (DW range 686-900 mm) and 14 females (719-1,097 mm), were caught at four stations (Punta Arena de la Ventana, Ensenada de los Muertos, Bahia de los Frailes, and Isla El Pardito) and adjacent waters, between December 1982 and October 1984. Size- frequency distributions for M. munkiana (Fig. 10) revealed that female mean size was greater than male, although not significantly (T-value = 1.724, df = 22,P > 0.1). Seasonal differences in size-frequency distribution could not be examined because M . munkiana were only collected during the fall and winter. All 10 freshly captured speci- mens were weighed (size range: 686-1,097 mm DW; 4.1-11.8 kg). Their WT/DW relationship is described by the following equation: WT= 1.041 X 10-«(DW)2 34 r = 0.95. WT is given in kg, DW in mm. The largest specimen in the sample, a female, was one of the largest "tortillas" ever seen. There are no data on size at birth, as no embryos were found. Lack of knowledge of the size of the young of the year also prevented insight on size segrega- >- o z LlJ o UJ q: u. LlI O CE < 10 5 - p?rn r-p^ Xtot = 886.5 ±33.4 mm (M = 24) X^^ = 853.8 ± 39.5mm (N = IO) X^^ =909.9 ±47. 1 mm (N:|4) D TOTAL m is ?? mm 650 750 850 950 1050 DISC WIDTH (mm) Figure 10. — Size-frequency distribution.s oi' Mubula munki- ana (means ±2 SE). tion. Term-embryos in M . rochebrunei , a closely related, similar sized species from west Africa, were 340-350 mm wide (Cadenat 1960). Since Mobula at birth has a DW of about 1/3 of the adult, size at birth would be about 350 mm DW. This information argues in favor of size segrega- tion in M. munkiana. Male to female ratio was 0.71 (N = 24), insignificantly different from 1 (x" test P > 0.05). Both sexes were caught in the same net sets, indicating that males and females school together, and that there was no sex segre- gation, either geographic or behavioral. A dried, twisted male carcass, for which mor- phometries could not be obtained, with a calcu- lated DW of 895 mm, had long, well-developed claspers, markedly protruding beyond the pelvic fins. Based on other mobulid species, this condi- tion indicates sexual maturity. A second speci- men, with a disc 686 mm wide, had small and pliable claspers, and the ratio between clasper length and pelvic fin length, both measured from tip to anterior margin of vent, was 0.88. At this ratio, both M. thurstoni and M . japanica are im- mature. Two M . munkiana , 871 and 872 mm DW, differed greatly in the relative size of their claspers: one possessed slightly longer claspers than the pelvic fins (ratio = 1.10), and an incipi- ent hardening of the cartilage was apparent; in the other specimen the claspers were much shorter than the pelvics (ratio = 0.84), and still soft. This information suggested that male sexual maturity in M. munkiana began at about 870 mm DW. The largest female specimen was sexually ma- ture, as it had a large, flaccid left uterus, and the enlarged left ovary consisted mainly of about 30 macroscopic ova (size range 2-15.2 mm). Ten stomachs were examined: four were empty, three contained unidentifiable whitish matter, and three were full of planktonic crustaceans. A list of prey species ranked by decreasing 9rIRI is given in Table 6. Mysidium sp. appeared to be the main staple of M. munkiana's diet. Mobula munkiana is thus similar to the wintering young of M. thurstoni. One of the stomachs contained coarse coral fragments and small gastropod shells, perhaps ingested by the ray while foraging on mysids near a sandy substrate. The west African species M . rochebrunei appeared to have similar feeding habits, as Cadenat (1960) found mysids and a few larger postlarval stomatopods in the stomachs of several specimens. I have no record of remoras associating with M. munkiana. The only parasitic crustacean 60 NOTARBARTOLO-Dl-SCIARA: NATURAL HISTORY OF MOBULA found was Pupulina cf. minor (Copepoda: Caligi- dae) from the skin. Among M. miinkiana's distinguishing features are its neritic preferences combined with its so- cial habits. This is the only mobulid species in the Gulf of California that was consistently seen in schools. It is not known whether this is a seasonal behavior, or a permanent ethological feature of the species. Schools appear as a conspicuous dark patch, sometimes a few tens of meters in diame- ter, as they slowly cruise along the coastline in shallow water. The presence of the school is often also highlighted by the frequent, simultaneously leaping individuals, which betray its position from a long distance. Similar behavioral traits (schooling and leaping) have been reported for two closely related species from the Atlantic, M. hypostoma (Bancroft 1829; Coles 1910, 1916a) and M. rochebrunei (Cadenat 1960). During leaps, M. munkiana occasionally reached a height of about two DWs. Two types of leaps were observed: rising vertically head first and landing flat with the belly on the sea surface with a loud clap (breach), and spinning one to three times around the main transverse body axis (somer- sault). A salient feature of M. munkiana's ecology in this area is its winter occurrence when all other mobulids are absent or at their lowest numbers. Mobula munkiana apparently subsists then chiefly on the mysid shrimp, Mysidium sp., which is also the main food for young wintering M. thurstoni . However, M . munkiana frequents the area occasionally in summer: two specimens were caught by surface gill net in Bahia de la Ventana in July 1983. Even during the season in which it is most abundant, M. munkiana is seen in "pulses", as its occurrence at any particular loca- tion is spotty. It may occur in large numbers at one location for a few days, and then be absent for 1 or 2 weeks. This observation suggests the possi- bility that M . munkiana lives in large concentra- tions, perhaps composed by several schools, which travel along the coast. A similar phenomenon was observed off the Senegal coast by Cadenat (1960) in M . rochebrunei, a species which is closely re- lated to M. munkiana both morphologically and ecologically. It is conceivable that mobulids in the northern half of the eastern tropical Pacific mate and give birth in summer, based on the few term and near- term embryos found in summer in M . thurstoni and M . japanica, and from anatomical evidence of mating activity in adult males M . thurstoni and M .japanica . That such a hypothesis can also be extended to M . munkiana is supported by lack of reproductive activity in any of the specimens collected during the fall or winter, and that new- born and young-of-the-year are missing from the sample. This evidence corroborates the hypothe- sis that the local waters are a wintering ground for M. munkiana, which then migrates into an unknown area (perhaps the northern Gulf of Cali- fornia) during the warmer season for mating and pupping. The possible causes of this ecological difference between M. munkiana and the other mobulids are many, and open to speculation. Mysid abun- dance may be declining in summer in the south- ern Gulf of California, and M . munkiana perhaps migrates to areas where this crustacean or re- lated species abound during the warmer season. Alternatively, M. munkiana could be excluded from this region in spring by competition with the incoming, larger M. japanica and adult M. thurstoni. Finally, M. munkiana may be moving during the summer into an area which is more suitable for its reproductive needs. Unfortu- nately, this recently discovered species is very little known, and it has been reported only from the Gulf of California and Ecuador, although its distribution probably extends to other coastal areas of the tropical east Pacific (Notarbartolo-di- Sciara 1987). Table 6^ — Prey species found in three stomachs of Mobula munkiana ranked by decreasing Index or Relative Importance (symbols as m Table 2). Identifiable copepod species included Undinula vulgans, Rhincalanus nasutus, and Scolecithnx danae. The stomatopods found were "erhithrus" larvae. One unidentified food item was a fragment of a larger crustacean, probably an euphausnd. Prey species N °/oN M °/oM %F IRI %IRI Mysidium sp. 293.40 97.80 287.40 95.80 100.00 19.360 Stomatopod larvae 4.65 1.55 9.06 3.02 66.67 304 Copepoda 1.59 0.53 1.44 0.48 66.67 67.3 Other 0.37 0.12 2.10 0.70 66.67 54.7 97.84 1.54 0.34 0.28 61 FISHERY BULLETIN: VOL. 86, NO. 1 Mob III (I tarapacatm (Philippi 1892) Local name: vaquetilla Mobula tarapacann is not a common species in the study area. Seven specimens were collected, one of which, a premature male embryo, was ex- pelled by a large female while she was being landed. Of the postnatal individuals, two were male (DW range 2,476-2,494 mm), and four were female (2,704-3,052 mm). All were caught in Bahia de la Ventana between 9 June and 30 Octo- ber 1983. All but two of the specimens were weighed. The following equation describes the WT/DW relationship for M. tarapacana (where WT is given in kg, DW in mm): WT = 2.378 X 10'*^(DW)2-92 r = 0.998. Although all sampled postnatal M . tarapacana were large, smaller individuals are known from the area, as can be seen in photographs taken at Punta Arena de la Ventana in summer 1981 (courtesy Felipe Galvan Magaha, CICIMAR, La Paz, Mexico; also Greg B. Deets"^). This informa- tion argues against geographical size segregation of M. tarapacana . Data on the embryo provide no indication of size at birth, since it was still far from term. Pale pigmentation was apparent only around the head and pelvic regions, and the ex- ternal yolk sac was present (Fig. 5B). The em- bryo, expelled tail first, was alive at birth. Judg- ing by its size it had filled the left uterus completely and must have been the sole develop- ing embryo. The small size of the sample does not permit any clear inference on size at sexual maturity for M . tarapacana . Some indication, however, can be obtained by comparison with similar species. Of the two postnatal males, the specimen with a disc 2,476 mm wide appeared to be immature: no sem- inal fluid was found in the ducti deferentes, the testes were small and apparently little developed, and the ratio between clasper length and pelvic fin length was 0.94. Conversely, the second post- natal male, with a disc of approximately the same width (2,494 mm), possessed claspers longer than pelvics (ratio = 1.14), and the testes were well de- veloped. Thus, sexual maturity in male M . tara- pacana begins around a DW of 2,400-2,500 mm. The specimen with a DW of 2,704 mm, one of two nonpregnant females, had a bulky left ovary, con- taining numerous large eggs; the largest, 32 mm in diameter, weighed 12 g. Similar features ap- peared to be associated with sexual maturity in female M. thurstoni and M. Japanica. The left ovary of another specimen, DW 2,831 mm, was smaller, and the diameter of the largest ovum was 18.6 mm, indicating that a DW of 2,700- 2,800 mm denotes a transitional stage for female M. tarapacana , in which both mature and non- mature individuals can occur. Twelve echeneidids were recovered from M . tarapacana . Three were Remora remora (size range: 108-229 mm SL), and nine were R. albescens (74-159 mm SL). The following crusta- cean parasites were also found: one cymothoid isopod (still in an unidentifiable aegathoid stage) and Pupulina [lores (Copepoda: Caligidae) on the skin, Entepherus laminipes (Copepoda: Cecropi- dae) on the branchial filter plates, and Eu- dactylina sp. (Copepoda: Eudactylinae) in the gills. Mobula tarapacana is strictly a summer and fall visitor to this region (Fig. 6C). This species is often found farther from the coast than M . thurstoni and M. japanica, and may have more pelagic habits. Four of the five stomachs examined (all from specimens caught in summer) were almost empty. Only traces of food were found among the folds of the stomach epithelium. Prey included four species of copepods (Acartia sp. , Pontella sp., Tetnora discaudata , and Undin- ula vulgaris ), hiperiid amphipods, one brachi- uran (family Calappidae), one euphausiid, two caridean decapods (one of which belonged to the family Alpheidae), megalopa and stomatopod lar- vae, and a fish egg. The fifth stomach, from a late October capture, contained the remains of 27 fishes (probably carangids 15-30 cm long, and a smaller anchovy-like species). Small tetraodon- tids had been found before in the stomach of a M . tarapacana caught in Bahia de la Ventana (Felipe Galvan Magaha'*). On this basis it is im- possible to determine whether M. tarapacana is a specialized ichthyophagous ray, with the few crustacean items accidentally ingested while swimming, or a generalized feeder. The mesh size of this species' branchial filter plates is indeed greater than in other Mobula species (Notarbartolo-di-Sciara 1987). However, filter- •*Greg B. Deets, Long Beach State University, CA, pens, com- mun. 1984. ^Felipe Galvan Magana, CICIMAR, La Paz, Mexico, pers. commun. 1983. 62 NOTARBARTOLO-DI-SCIARA NATURAL HISTORY OK MOBrLA feeding on planktonic Crustacea still appears to be a feasible foraging technique for M. tara- pacana, judging from the size of its branchial sieve as it compares with the average-sized crus- tacean prey. SUMMARY AND CONCLUSIONS Four species ofMobula were found in the south- ern Gulf of California. The most abundant spe- cies, M. thurstoni , was present year-round, but only the smaller individuals were seen during the winter. The bulk of the population, including the adults, appeared in early spring. Numbers began declining in July. Mobula japanica , the second most abundant species, was comprised of only large individuals; numbers progressively in- creased from March throughout July. Large M. japanica were rare in winter, but were occasion- ally caught then. Mobula tarapacana is the rarest mobulid in the area, yet its presence as a summer and fall visitor is well known and predictable; it is believed by the local fishermen to be more abundant farther offshore. All three species share a similar pattern of peak summer seasonal abun- dance. The reverse is true for Mobula munkiana , it being most abundant in winter, and almost to- tally absent during the rest of the year. It is not known where any species goes when not seen in the area. Seasonal migrations within the epipelagic habitat to different areas of the Panamic region are likely, but unverifiable be- cause of the present lack of knowledge of the oc- currence of identified Mobula species south of the Gulf of California. Alternatively, devil rays may spend part of the year in midwater, or near the sea bottom, therefore disappearing from sight and reach. There is a striking similarity between the array of mobulid species found in the Gulf of Cali- fornia (and probably along the Pacific coast of tropical America) and the mobulid fauna from the tropical waters off west Africa. The family is rep- resented in both areas by Manta birostris and by four species oi Mobula : M . thurstoni; M Japanica (reported from west Africa as M. rancureli by Cadenat 1959); M . tarapacana (reported as M. coilloti for African waters by Cadenat and Ran- curel 1960 and Stehmann 1981); and a small gi-e- garious form, represented in the Gulf of Califor- nia by M . munkiana and off west Africa by the closely related M . rochebrunei (Notarbartolo-di- Sciara 1987). Tropical coastal areas off west America and west Africa are known to be among the most productive tropical waters in the world, because of comparable large-scale atmospheric and oceanographic circulation patterns (Sverdrup et al. 1942). It is conceivable that the ecological similarity between these two regions is reflected in similar faunal associations, especially as far as low levels of the tropic chain (e.g., plankton- feeding vertebrates) are concerned. The Gulf of California presents a unique envi- ronment in the eastern Pacific Ocean, with ex- treme annual water temperature ranges, wind- induced mixing and upwellings, and subsequent great productivity (Roden 1964; Brusca 1980). Upwelling is caused along the peninsular coast by the southerly winds prevailing during the warmer months. This environment apparently creates optimal conditions for the existence of the euphausiid shrimp A^yc^/p/?a/?es simplex , which is found in great abundance in the neritic habitat between spring and midsummer, before the in- tense August heat causes a decline in its numbers (Brinton and Townsend 1980). The following data are combined in Figure 6 to provide an overview of the possible relationship between the seasonal- ity of predator and prey in the study area: a) the relative importance of Nyctiphanes simplex and Mysidium sp. in the diet of M. thurstoni; b) the relative abundance of A'^. simplex; the occurrence of M. tarapacana (c) and M . munkiana (d) in the catch; and the relative abundances of M .japanica (e) and M . thurstoni (f) (no data on the biology of Mysidium sp. are available). Young M . thurstoni and all M . munkiana ex- amined in winter appeared to subsist largely on Mysidium sp., whereas adult M. thurstoni and M .japanica caught during the warmer months fed exclusively on N . simplex . An extreme degree of feeding specialization was evident in all mobu- lid species in which quantitative analyses of the stomach contents was possible; most prey forms, other than N . simplex and Mysidium sp., were so rare that they were probably ingested acciden- tally. Stenophagy was linked to feeding special- ization in another myliobatiform species, the mollusk-feeder Rhinoptera bonasus (Schwartz 1966; Smith and Merriner 1985). These results suggest that devil rays are highly efficient in lo- cating and selecting their preferred food. They may be aided during this behavior by their prey's habit of swarming. Competitive interaction is to be expected between sympatric species-pairs which are closely related both taxonomically and ecologically. Food-resource partitioning is known to occur in sympatric species-pairs of skates 63 FISHERY BULLETIN: VOL. 86, NO. 1 (McEachran et al. 1976). This condition, however, is not necessarily true when the sought-after re- sources are not in short supply (Zaret and Rand 1971). This may be the case of M. thurstoni and M . japanica feeding together on A^. simplex when the abundance of euphausiids is at its peak. Com- petition should occur, however, in late summer, when prey numbers decline. It would be interest- ing to determine whether the slight morphologi- cal and behavioral differences between poten- tially competing species pairs (M. thurstoni/ M. munkiana in winter, M. thurstoni IM . japan- ica in spring and summer) influence or reflect partitioning of their habitat when food resources become limiting, as was described for both fresh- water (Werner and Hall 1977) and marine teleosts (Hixon 1980; Larson 1980). This overview of the ecology and natural his- tory of mobulids in the Gulf of California is based on field investigations made chiefly in 1983, a year in which the El Nino perturbation was par- ticularly severe (Cane 1983). Although in terms of fishermen's experience the year 1983 was not unduly different, as far as mobulid relative abun- dance and seasonality are concerned, the abnor- mally high water temperatures resulting from El Nino may have affected the devil rays studied in subtle ways; therefore this investigation should be repeated in a normal year. According to the fishermen, the abundance of sharks (mostly carcharhinids and sphyrnids) on which their activity is based is declining. This decline will probably result in an increase of mob- ulid fishing effort. It is of concern that 12% of the specimens of M. thurstoni caught were immature (DW <1,500 mm). ACKNOWLEDGMENTS I owe deep gratitude to the many persons who assisted me in this investigation: Richard H. Rosenblatt, Theodore H. Bullock, Paul K. Dayton, William E. Evans, and Walter H. Munk, mem- bers of my doctoral committee; Edward Brinton, Robert Cowen, Abraham Fleminger, Nicholas Holland, Margaret Knight, Spencer Luke, William Newman, Mark Grygier, Jeff Schweitzer, George Shor, George Snyder, and Fred White of the Scripps Institution of Oceanog- raphy; Thomas Bowman (United States National Museum, Washington, D.C.); Daniel Brooks (Uni- versity of British Columbia, Vancouver); Greg Deets (Long Beach State University); Dennis Bedford and Robert Lea (California Department of Fish and Game); Fay Wolfson (Hubbs Marine Research Institute); Alexis Fossi (Institut Na- tional des Techniques de la Mer, Cherbourg, France); Felipe Galvan Magaiia (Centro Inter- diciplinario de Ciencias Marinas, La Paz, Mex- ico); Lalo Cuevas, Marcelo Geraldo, Juan Lucero, and their colleagues of the Cooperativa Pesquera de Punta Arena de la Ventana; Steven Kramer (National Marine Fisheries Service, Honolulu, Hawaii); Carl A. Jantsch and Steven D. Kamol- nich (Sea World, Inc., San Diego). The Hubbs Marine Research Institute (San Diego) loaned a sailing vessel, the "Fling", for the field study; Rodney Black helped in the outfitting of the ves- sel and in sailing it to the Gulf of California. This investigation was supported in part by a grant from the Foundation for Ocean Research (San Diego). LITERATURE CITED Backus. R H 1963. Hearing in elasmobranchs. In P. W. Gilbert (editor), Sharks and survival, p. 243-254. D. C. Heath & Co., Lexington, MA. Bancroft, E. N 1829. On the fish known in Jamaica as the sea-devil. Zool. J. 4:444-457. Barnard. K H 1925. A monograph of the marine fishes of South Africa. Part 1. Ann. S. Afr. Mus. 21:1-415. Barton, O 1948. Color notes on Pacific manta, including a new form. Copeia 1948:146-147. BEEBE, W , AND M A HOLLISTER 1935. The fishes of Union Island, Grenadines, British West Indies, with the description of a new species of stargazer. Zoologica, N.Y. 19:209-224. Beebe. W., and J Tee- Van 1938. Eastern Pacific Expeditions of the New York Zo- ological Society. XV. Seven new marine fishes from Lower California. Zoologica, N.Y. 23(15):299-301. 1941. Eastern Pacific expeditions of the New York Zoolog- ical Society. XXVIII. Fishes from the Tropical Eastern Pacific. (From Cedros Island, Lower California, south to the Galapagos Islands and northern Peru). Part 3. Rays, mantas and chimaeras. Zoologica, N.Y. 26(26):245- 280. Benz. G W . AND G B Debts 1987. Echthrogaleus disciarai sp. nov. (Siphonostoma- toida: Pandaridae), a parasitic copepod of the devil ray Mobula lucasana Beebe and Tee- Van, 1938 from the Sea ofCortez. Can. J. Zool. 65:685-690. BiGELOW. H B , AND W C SCHROEDER 1953. Sawfishes, guitarfishes, skates and rays. In J. Tee- Van, C. M. Breder, A. E. Parr, W. C. Schroeder, and L. P. Schultz (editors), Fishes of the western north At- lantic, Memoir 1, part 2, p. 1-514. Sears Foundation for Marine Research, New Haven. BrINTON, E . AND A W TOWNSEND 1980. Euphausiids in the Gulf of California - the 1957 64 NOTAKliAIMOI.O 1)1 SflARA NATl'KAl. HISTORY OF .\f()IUl.A cruises. Calif. Coop Oifaiiic Fi.sh Inve.st. Rt'p. 21:211- •J.'i.''). Bki-.sca. R C 1980. Common intertidal invcrti-brate.-i of thf (iulf of California Univ. Arizona Pres.s, Tust "H ';^ WILMINGTON OCEAN <^ f:i BALTIMORE 36 70 39 % ^ * !) .it WASHINGTON 35 77 V g ^y\ NORFOLK 71 38 / ■\ \ 78 \ ^ ^^X VI 34 72 / \ x'S / -^Xj p. CAPE HATTERAS \ y \ 37 78 77 36 76 35 75 74 34 73 Figure l. — East coast of United States showing continental shelf, slope and submarine canyon areas from which species listed in Table 1 were collected. Base map adapted from Uchupi ( 1965) and Veatch and Smith (1939). Contours m m: dotted = 100, dashed = 200, solid = 1,000. 72 WILLIAMS: DECAPOD AND EUPHAUSIID CRUSTACEANS 83 31 30 81 80 29 79 28 78 V \ ^ A / \ ■\K 78 27 79 y 26 25 80 85 27 84 83 26 82 25 81 Figure 2. — Florida peninsula including continental platform showing slope localities from which species listed in Table 1 were collected. Base map adapted from Uchupi (1965). Contours in m; dotted = 100, dashed = 200, solid = 1.000. Some of the more obvious differences for the Veatch Canyon material are: Rostrum relatively shorter in relation to basal antennal article. Dor- sal spines of telson more distally positioned, ante- rior pair at about midlength of telson but subject to some variation. Antennal scale with distal spine exceeding antennular peduncle. Major chela of mature male missing, but juvenile with fingers bent mesad; dactyl moderately arched in profile and greatly overreaching fixed finger, somewhat twisted, compressed proximally and dorsally producing thin dorsal margin, external surface somewhat concave, occlusive surface lacking plunger but broadened and strongly calci- fied distally, fitted to obliquely flattened occlu- sive surface of fixed finger lacking socket but provided with small stout tooth on mesial sur- face; palm with obsolescent dorsal and ventral notches. In the Florida material, the differences are: Chela relatively stout, fingers stout and thick, dactyl opening and closing in oblique plane, tip rounded, bearing short plunger fitting into shal- low socket on occlusive surface of fixed finger, latter with 2 short spines on mesial surface; palm with shallow notch on dorsal margin and slight offset on ventral margin, outer surface smooth but base of dactyl flanked by distodorsal groove and longer mesial groove. Second pleopod of male with appendix masculina exceeding appendix in- terna. Uropodal exopod with lateral margin end- ing in single sharp tooth and rather long, uncol- ored movable spine; endopod lacking distal spines but bearing subterminal tuft of setae on dorsal surface. Alpheus amblyonyx is distributed from Quin- tana Roo (type locality, Bahia de la Ascen- sion), Yucatan Peninsula, Mexico, to Puerto Rico, Saint Thomas and Dominica; sublittoral (Chace 1972). Family Hippolytidae Three small shrimps, two males and one female from JSL 1673 in the Gulf of Mexico off Florida, represent an undescribed species resembling members of the genus Ligur Sarato, 1885 from the western Mediterranean and Indo-Pacific re- gion (see Holthuis 1947, 1955). The specimens were associated with burrow systems of the blue- line tilefish, Caulolatilus microps Goode and Bean. 73 FISHERY BULLETIN: VOL. 86, NO. 1 SUPERFAMILY PAGUROIDEA Family Lithodidae Lithodes maja (Linnaeus 1758). Southern limit extended from Sandy Hook, NJ (see Williams 1984) to Baltimore Canyon. SUPERFAMILY GALATHEOIDEA Family Galatheidae Munida forceps A. Milne Edwards 1880. Geographic range extended from south of Norfolk Canyon, 36°43.2'N, 74°38.0'W, 252 m (Wenner 1952), to Veatch and Lydonia Canyons off south- ern New England, 103-337 m. The distinctive color pattern of this species was described on 14 October 1981 from specimens pre- served in formalin 25 July 1981. Carapace (Fig. 3), salmon color with lavender submesial spots on gastric region and interrupted U-shaped bands of same color in nested series on mesogastric region, posterior to cephalic groove, and arching across posterior and posterolateral parts. Oblique red lines on lateral wall of carapace below suture, most prominent band along anterior edge, contin- ued dorsally anterior to antennal peduncle and ending on lateral side of supraocular spine. Me- dian band of same intensity on epistome and labium. Paler oblique lateral band on basal an- tennular article. Some flecks of red on merus of chelipeds and cross banding on fingers of some individuals. Munida longipes A. Milne Edwards 1880. Northern limit extended from off Cape Lookout, NC (Williams 1984) to Baltimore Canyon. Munidopsis cf. transtridens Pequegnat and Pe- quegnat 1971. Munidopsis transtridens is known only from the holotype female taken in the southeastern Gulf of Mexico at 1,280 m. The spec- imens reported here from Baltimore, Hendrick- son, and Lydonia Canyons off New Jersey and southern New England, 906-1,425 m, are all males. They resemble M. transtridens but differ from it in rostral characters (both longer and shorter, variably narrower or broader, in degree of lateral convexity) and in having chelae strik- ingly larger than the slender ones of the holotype. Although these differences may be attributable to sexual dimorphism, provisional identification seems best until more material is available for study. Figure 3. — Munida forceps, male. Dorsal view of carapace, dia- grammatic representation of lavender bands on salmon ground color, carapace length 17.7 mm to base of supraocular spine. SUPERFAMILY XANTHOIDEA Family Goneplacidae Chacellus filiformis Guinot 1969. Geographic range extended from the northern Gulf of Mexico east of the Mississippi River Delta and off the east coast of Florida, 328-400 m (Guinot 1969), to Bal- timore and Lydonia Canyons, 160-244 m. Goneplacid crabs were driven out of secondary burrows in walls of larger burrows constructed by tilefish, Lopholatilus chamaeleonticeps Goode and Bean, with rotenone. The poison did not kill the crabs but caused them to emerge from the burrow systems enough that they could be collected by 74 WILLIAMS: DECAPOD AND EUPHAUSIID CRUSTACEANS "slurp gun". These burrow systems in Pleistocene clay, referred to by Warme et al. (1978) and Cooper and Uzmann (1980) as "Pueblo Villages", shelter a number of invertebrate and vertebrate species (Able et al. 1982; Bowman 1986; Grimes et al. 1980a, 1980b, 1986). Goneplacids are rare in collections made from surface vessels probably because trawls or grabs cannot efficiently sample the burrow systems in which these crabs have been observed. Galatheids from shallower bur- rows are more open to capture by conventional means (Churchill B. Grimes^). To the brief color description quoted from Chace by Guinot (1969), the following can be added from notes made 14 October 1981 on ma- ture males, females, and juveniles that were pre- served in formalin 25 July 1981, and personal communication from Churchill Grimes (fn. 2). Carapace dorsally spotted with red on off-white background. Same type of spots on pterygosto- mian, subocular, epistomial, and subbranchial areas, on external maxillipeds, and on merus, car- pus, and propodus of chelipeds (dorsally, later- ally, and mesially). Spots tending to coalesce along front of carapace and on chelae. Red color more diffuse on dorsal or exposed surfaces of walking legs, becoming more distinct and intense with increasing size. Fingers of chelae black. Dactyls of walking legs white except for darkened tips, but setae pinkish. There is some variation in pattern on individual crabs. There is variation also in the length of the male first pleopod, both in the USNM series of speci- mens studied by Guinot (1969) and in the new material reported here. In some specimens of the latter, this appendage exceeds or at least reaches the distal edge of the telson, whereas it is shorter in specimens previously reported from localities further south. In the latter, the third abdominal segment is more angled laterally than in speci- mens from the north. Thus, there seem to be some differences between the northern and southern populations. SUPERFAMILY GRAPSIDOIDEA Family Grapsidae Euchirograpsus americanus A. Milne Edwards 1880. Geographic range extended north from off ^Churchill B. Grimes, Southeast Fisheries Center Panama City Laboratory, National Marine Fisheries Service, NOAA, .3500 Delwood Beach Road, Panama City, FL .32407-7499, pers. commun. February 1982. Oregon Inlet, NC (Williams 1984) to Oceanogra- pher Canyon at the edge of Georges Bank and nearby continental slope at 155-200 m. SUPERFAMILY PINNOTHEROIDEA Family Pinnotheridae Dissodactylus juvenilis Bouvier 1917. The ovigerous female from the Gulf of Mexico off west- ern Florida, though similar in general features to D. juvenilis, is very large for that species. In a recent review of the genus Dissodactylus , Griffith (1987) reported D. juvenilis from north of Yu- catan and the Mississippi Delta. Members of the genus are found in association with clypeastroid echinoids (Schmitt et al. 1973), as was this speci- men in a sample that included Clypeaster ravenel- lii A. Agassizi. ACKNOWLEDGMENTS The following persons brought these records to my attention through requests for identifications: Barbara Hecker and Dennis T. Logan, Lamont- Doherty Geological Laboratory, Columbia Uni- versity, Palisades, NY; Kenneth W. Able, Center for Coastal and Environmental Studies, and Churchill B. Grimes, Department of Horticul- ture and Forestry, Rutgers University, New Brunswick, NJ (CBG now with Southeast Fish- eries Center Panama City Laboratory, National Marine Fisheries Service, NOAA, Panama City, FL). Maureen E. Downey identified the echinoid. Ruth E. Gibbons drafted the maps and Keiko Hi- ratsuka Moore figured the galatheid. The manuscript was critically reviewed by B. B. Col- lette and D. L. Felder. LITERATURE CITED Able, K W , C B Grimes, R A Cooper, and J. R Uzmann 1982. Burrow construction and behavior of tilefish, Lopholatilus chamaeleonticeps, in Hudson Submarine Canyon. Environ. Biol. Fishes 7(3):199-205. Able. K. W., D. C. Twitchell. C. B. Grimes, and R. S. Jones. 1987. Tilefishes of the genus Caulolatilus construct burrows in the sea floor. Bull. Mar. Sci. 40:1-10. Bowman. T H. 1986. Tridentella recava , a new isopod from tilefish burrows in the New York Bight (Flabellifera: Tri- dentellidael. Proc. Biol. Soc. Wash. 99(2):269-273. Chace. F A, Jr 1972. The shrimps of the Smithsonian-Bredin Caribbean Expeditions with a summary of the West Indian shallow-water species (Crustacea: Decapoda: Natantiaj. Smithson. Contrib. Zool. 98:1-179. 75 FISHERY BULLETIN: VOL. 86, NO 1 Cooper. R A , and J R Uzmann. 1980. Ecology of juvenile and adult Homarus. In J. S. Cobb and B. F. Phillips (editors), The biology and man- agement of lobsters. Vol. II, Ecology and management, ch. 3, p. 97-142. Acad. Press, NY. Griffith. H 1987. Taxonomic revision of the genus Dissodactylus (Crustacea: Brachyura: Pinnotheridae). Bull. Mar. Sci. 40:397-422. Grimes, C B , K W Able, and R S Jones. 1986. Tilefish, {Lopholatilus chamaeleonticeps), habitat, behavior and community structure in mid-Atlantic and southern New England waters. Environ. Biol. Fishes 15(41:273-292. Grimes, C B . K W Able, and S C Turner 1980a. A preliminary analysis of the tilefish, Lopholatilus chamaeleonticeps fishery in the Mid-Atlantic Bight. Mar. Fish. Rev. 42(11):13-18. Grimes, C G , K W Able, S C. Turner, and S J Katz 1980b. Tilefish: Its continental shelf habitat. Under- water Nat. 12(41:34-38. Guinot, D 1969. Recherches preliminaires sur les groupements na- turels chez les Crustaces Decapodes des Brachyoures. VII. Les Goneplacidae (suite et fin). Bull. Mus. Natl. Hist. Nat. Ser. 2, 41(3):688-724. HOLTHUIS, L B 1947. The Decapoda of the Siboga-Expedition. Part IX. The Hippolytidae and Rhynchocinetidae collected by the Siboga and Snellius Expeditions with remarks on other species. Siboga-Exped. Monogr. 39aS, 100 p. 1955. The recent genera of caridean and stenopodidean shrimps (Class Crustacea, Order Decapoda, Supersection Natantia) with keys for their determination. Zool. Verb. (Leiden) 26:1-157. PEQUEGNAT, W E , AND L H Pequegnat 1971. New species and new records o{ Munidopsis (Deca- poda: Galatheidae) from the Gulf of Mexico and Caribbean Sea. Texas A & M Univ. Oceanogr. Stud., 1 (suppl.):3-24. Schmitt. W L , J C McCain, and E S Davidson 1973. Decapoda I, Brachyura I, Fam. Pinnotheridae. In H.-E. Gruner and L. B. Holthuis (editors), Crustaceorum Catalogus 3, 160 p. Dr. W. Junk B. V.-Den Haag. Squires, H J 1965. Decapod crustaceans of Newfoundland, Labrador and the Canadian eastern Arctic. Fish. Res. Board Can., Manuscr. Rep. Ser. (Biol.) 810, 212 p. UCHUPI, E. 1965. Map showing relation of land and submarine topog- raphy. Nova Scotia to Florida. U.S. Geol. Surv. Misc. Geol. Invest., Map 1-451 (Sheets 1-3). Veatch, a C, and P a Smith 1939. Atlantic submarine valleys of the United States and the Congo Submarine Valley. Geol. Soc. Am. Spec. Pap. 7, 101 p. Warme. J E , R A Slater, and R A Cooper 1978. Bioerosion in submarine canyons. In D. J. Stanley and G. Kelling (editors). Sedimentation in submarine canyons, fans, and trenches, ch. 6, p. 65-70. Dowden, Hutchinson & Ross, Inc., Stroudsburg, PA. Wenner, E L 1982. Notes on the distribution and biology of Galathei- dae and Chirostylidae (Decapoda: Anomura) from the Middle Atlantic Bight. J. Crustacean Biol. 2(3):360- 377. Wenner, E L , and D F Boesch 1979. Distribution patterns of epibenthic decapod Crus- tacea along the shelf-slope coenocline. Middle Atlantic Bight, USA. Bull. Biol. Soc. Wash. 3:106-133. Wenner, E L , and N. T Windsor 1979. Parasitism of galatheid crustaceans from the Nor- folk Canyon and Middle Atlantic Bight by bopyrid isopods. Crustaceana 37(3):293-303. Williams, A B. 1984. Shrimps, lobsters, and crabs of the Atlantic coast of the eastern United States, Maine to Florida. Smithson. Inst. Press, Wash., DC, 550 p. Williams, A B , and R L Wigley 1977, Distribution of decapod Crustacea off northeastern United States based on specimens at the Northeast Fish- eries Center, Woods Hole, Massachusetts. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 407, 44 p. 76 AGE AND GROWTH OF LARVAL GULF MENHADEN, BREVOORTIA PATRONUS, IN THE NORTHERN GULF OF MEXICO Stanley M. Warlen' ABSTRACT Experiments on laboratory-spawned and -reared larval gulf menhaden, Brevoortia patronus , showed that they formed one otolith growth increment per day and that the increments could be used to estimate their age. Wild larvae from collections in the northern Gulf of Mexico along three transects I Cape San Bias, Florida; Southwest Pass, Louisiana; and Galveston, Texas) were aged. Gompertz growth equations were used to describe the relationship between age and standard length for larvae collected at various locations, and in different seasons and years. MANOVA tests and subsequent pairwise tests were used to test for differences among these growth curves. For the most extensive data set (Southwest Pass, Louisiana), there were significant differences in growth between early season (December) and late season (February) larvae. Early season larvae grew faster than late season larvae. Growth of larvae also differed among December collections and among February collections. The growth model for the pooled data for all wild larvae predicted that they grew from 2.4 mm SL at hatching to 20.4 mm SL at 62 days. Gulf menhaden, Brevoortia patronus , is the most abundant commercial finfish in the Gulf of Mex- ico and, with 883,500 metric tons (t) landed in 1985 (U.S. National Marine Fisheries Service 1986); it constitutes the largest fishery in the United States. Some aspects of the oceanic early life history of this clupeid are known and are re- viewed by Turner (1969), Christmas and Waller (1975), Lewis and Roithmayr (1981), Govoni et al. (1983), and Shaw et al. (1985a). However, virtu- ally nothing is known about the age and growth of the larvae, much less how these parameters vary spatially and temporally. Daily growth in- crements on otoliths of larval fishes can be used as an indicator of their age, and once the use of this technique, first described by Pannella ( 1971 ), is validated for the larvae of an individual spe- cies, their ages can be estimated with confidence and growth rates can be determined. Intraspecific growth may be compared for larvae from different areas and seasons (Lough et al. 1982), and from this it may be possible to ascertain how biotic and abiotic environmental variables affect larval growth and survival. The objectives of this study are to 1) validate the periodicity of increment for- mation in otoliths of larval gulf menhaden, 2) estimate larval growth rates, 3) compare growth rates of larvae from different locations and times, 4) estimate spawning times, and 5) examine pos- •Southeast Fisheries Center Beaufort Laboratorv, National Marine Fisheries Service, NOAA, Beaufort, NC 28516-9722. Manuscript accepted September 1987 FISHERY BULLETIN. VOL. 86, NO. 1, 1988. sible relationships between larval growth and surface water temperature. This work was part of a larger project designed to investigate the early life history of several economically important fishes and the marine planktonic food webs that support their growth and survival in the northern Gulf of Mexico. METHODS Spawning and Larval Rearing Adult gulf menhaden were collected near Gulf Breeze, FL, and transported to the Beaufort Lab- oratory, Beaufort, NC (Hettler 1983). After a pe- riod of acclimation, adults were induced to spawn in the laboratory. The resultant larvae were used in experiments to validate the periodicity of in- crement formation on their otoliths and the age at first increment formation. Beginning February 1983, several thousand newly spawned gulf menhaden eggs were trans- ferred to a tank containing 90 L of filtered sea- water. The static water in this tank, kept at 20.5° ± 0.5°C throughout the experiment, was continuously aerated and the salinity maintained at 31 ± 17cc. Photoperiod was 12 hours light:12 hours dark. A food concentration of 25 rotifers (Brachionus plicatilis) mL"^ was maintained. A green alga, Nanochloris sp., was added periodi- cally as food for the rotifers and to aid in remov- ing toxic metabolites. The otoliths of larvae sam- 77 KISHERY BULLETIN; VOL. 86, NO. 1 pled at 10-, 17-, 24-, and 31-d posthatch were ex- amined. In January 1984 additional larvae were reared to compliment results of the earlier experiment. Smaller tanks with 60 larvae in 10 L of filtered water were used. Experimental conditions were the same as for the first experiment. The otoliths of larvae sampled at 7-, 14-, and 20-d posthatch were examined. Larval Collections Larval gulf menhaden were collected in the northern Gulf of Mexico during six cruises of the RV Oregon II. Sampling stations (Fig. 1) along transects LA (off Louisiana) and FL (off Florida) were occupied during 11-19 December 1979, 5- 15 February 1980, and 2-12 December 1980 and along transects LA, FL, and TX (off Texas) during 9-24 February 1981, 2-13 December 1981, and 4-16 February 1982. Transect LA is near the Mississippi River outflow off Southwest Pass, LA; transect FL is southwest of Cape San Bias, FL; and transect TX is located off Galveston Bay, TX. Sampling stations were in water depths of 18, 91, and 183 m except off Texas where only the 18 and 91 m depths were sampled. A multiple opening-closing net and environ- mental sensing system (MOCNESS) as described by Wiebe et al. ( 1976) were the primary sampling gear used to capture larvae. Additional samples were taken in oblique tows with a 60 cm bongo frame also fitted with 505 fjim mesh nets. Samples were collected day and night and were preserved in 95Vf ethanol (final concentration ^757^ ) within 5 minutes of collection. The ethanol was changed in all samples at least once after initial preserva- tion to prevent dissolution of otoliths in fish from any samples that may have been inadequately preserved. Data from larvae collected at all sta- tions within a transect were combined for that transect. Estimating Age and Growth All gulf menhaden larvae were measured to the nearest 0.1 mm standard length (SL). The largest otolith pair (sagittae) was teased from the sur- rounding tissue, cleaned in distilled water, and then placed on a glass microslide under a thin layer of Flo-Texx-^ mounting medium. Otoliths were viewed with a compound micro- scope fitted with a television camera. Growth in- crements were counted from otolith images on a video monitor at magnifications of at least 400 x. An increment appeared as a light, wide incremen- tal band and a dark, narrow, discontinuous band (Tanaka et al. 1981). Increments were generally clearly discernable and easily counted (Fig. 2). Estimated age was the number of increments counted plus an empirically derived value for the 2Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 96 00' 94° 00' 92° 00' 90° 00' 88° 00' 86° 00' 84° 00' 30° 00' 28° 00' 26° 00' Figure l. — Location of sampling sites from which larval gulf menhaden were collected during crui-ses of the RV Oregon II in December 1979-81 and February 1980-82. 78 WARLEN; AGE AND GROWTH OF LARVAL GULF MENHADEN Figure 2. — Photomicrograph of a saggital otohth with 22 increments from a 17.4 mm SL field collected larval gulf menhaden. Scale bar represents 10 pim. Growth increments appear as pairs of wide incremental and narrow discontinuous bands. number of days from spawning to first increment formation. Results of the laboratory experiments established the periodicity of otolith increment formation. A spawning date was assigned each ageable larva by using the estimated age of the fish in days to back-calculate from the date of capture. It was assumed that there were no differences in either the age at initial increment deposition or the otolith increment deposition rate between lo- cations and seasons and that the rate was not a function of temperature, food, or photoperiod. Average growth of larvae was described by the Laird version (Laird et al. 1965) of the Gompertz growth equation (Zweifel and Lasker 1976) fitted to estimated age and size at time of capture for fish from all cruises and transects. To stabilize the variance of length over the observed age in- terval, length data were log-transformed and model parameters were estimated from the log- transformed version of the growth equation. The model was fit to data for each transect within each cruise and for pooled data from all cruises. Potential differences in the overall growth curves among years and between seasons for lar- vae caught off Louisiana and between years ( 1981 and 1982) for larvae caught off Louisiana and Texas were examined by treating the parameters of the Gompertz equation as dependent variables in two-way multivariate analysis of variance (MANOVA) designs. A one-way MANOVA de- sign was used to test for differences among transects (LA, FL, TX) within one season (February 1982). Following significant MANOVA results, prespecified pairwise Hotelling's T'^ test comparisons (Bernard 1981, as modified by Hoenig and Hanumara 1983) were made using the Bonferroni procedure (Harris 1975) to provide conservative tests of statistical significance. Bonferroni critical values for these individual tests were equal to the overall error rate (significance level = 0.05) divided by the number of possible comparisons in the particular MANOVA design. The emphasis in the compari- sons was to look for overall differences in the growth of larvae using these statistics as a guide and not to look for differences in individual parameters of the growth models. Hotelling's T^ test and MANOVA both require that the data fit a multivariate normal distribu- tion and that the variance-covariance matrices of the populations are not different (Harris 1975). These assumptions are difficult to test and are almost certainly not valid for real data sets (par- 79 FISHERY BULLETIN: VOL 86. NO 1 ticularly field data), but they may be nearly valid for many sets of data (Harris 1975). No direct test of normality in a trivariate, joint probability dis- tribution is available (Bernard 1981), but bias arising from nonnormal, multivariate, joint dis- tributions is minimized with large sample sizes (Bernard 19811. While methods are available to test the hypothesis of equal variance-covariance matrices (e.g.. Box's modification of Bartlett's test), these methods are very sensitive and even minor differences between group dispersions will likely be discovered (Pimentel 1979). In any event, the use of MANOVA in this paper relies on variance-covariance matrices estimated from nonlinear regressions, and these are not amenable to testing. However, both MANOVA and Hotel ling's tests are extremely robust even under violation of the assumptions of homo- scedasticity and multivariate normality (Harris 1975). RESULTS Increment Formation The age of gulf menhaden at formation of the first otolith growth increment was estimated from laboratory-reared larvae. The intercept (2.6 days) of the regression of the number of growth increments on known posthatch age of 36 larval gulf menhaden (Fig. 3) was used to estimate posthatch age at formation of the first increment. This value was added to the time from spawning to hatching which at 20°C is 2 days (Powell'M. This sum (4.6 days) is the estimated time from spawning to formation of the first increment. Hence, it was necessary to add 5 days to each increment count to estimate the age of larval gulf menhaden from spawning. The periodicity of increment formation was as- certained from the regression of the number of growth increments on the known age (Fig. 3). The slope did not differ significantly (/-test, P < 0.05) from 1.0, and thus, on the average, one otolith growth increment was formed per day in laboratory-reared larvae up to 31 days after hatching. Results of a second experiment (Table 1 ) confirmed this periodicity. The age of gulf men- haden larvae estimated from otolith increment counts ( + 5) closely approximated the known ages of 51 laboratory-reared larvae. Mean estimated age of larvae differed by < 1 day from the known 3A. B. Powell, Southeast Fisheries Center Beaufort Labora- tory, National Marine Fisheries Service, NOAA, Beaufort, NC 28516-9722, pers. commun. February 1986. CO H Z 111 liJ DC O z O o o tr LU m Z 35 30 25 20 15 10 No. increments=-2.617+ 0.921(known age) r= 0.928 n= 36 5- 10 15 20 25 KNOWN AGE (days) 30 35 Figure 3. — Regression of the number of growth increments on the known posthatch age of 36 laboratory-reared gulf menhaden. Standard error of the slope IS 0.108. 80 WARLEN; AGE AND GROWTH OF LARVAL GULF MENHADEN Table 1. — Standard length (mm) and estimated age (num- ber of otolith growth increments +5) of laboratory-reared larval gulf menhaden. Values in parentheses are 95% inter- val estimates. Known age Number of Mean estimated (days) fish age Mean SL 7 14 20 13 16 22 7.8 (±0.38) 13.2 (±1.14) 19.3 (±0.77) 4.7 (±0.28) 6.4 (±0.38) 8.0 (±0.42) ages and the 957c confidence intervals included the known age in each of the three groups. Some of the variation in the number of gi'owth incre- ments observed in known age larvae (Fig. 3, Table 1) may have been due to 1) poor growing conditions during rearing that could have re- sulted in reduced growth in underfed larvae, 2) variation in the inception of increment formation as has been observed in other species (Laroche et al. 1983; Fives et al. 1986), and 3) faintness of growth increments in some larvae. In contrast, increments on otoliths of field collected larvae (Fig. 2) were usually very regular and distinct and were more easily observed than those on otoliths of laboratory-reared larvae. I assumed that the gi'owth increment deposition rate was also daily in wild larvae examined in this study. Age and Growth of Larvae Average growth of larval gulf menhaden dur- ing their first two months of life was described by the Gompertz growth model for pooled length at age data for 2,003 fish representing collections from all six RV Oregon II cruises (Table 2, Fig. 4). Larvae ranged in age from 5 to 62 days (x=24A days) and in SL from 3.4 to 28.0 mm (x = 12.6 mm). In the log-transformed model, age ac- counted for 82% of the variation in length. Gulf menhaden were predicted to have grown from 2.4 mm SL at hatching to 20.4 mm at age 62 days. The size at hatching estimated from the Gom- pertz equation was only slightly less than the hatching size, 2.6-3.0 mm SL, observed in the laboratory (Hettler 1984). Age-specific growth rates declined from — 79f/day at age 10 days to <0.4%/day at age 60 days. Maximum absolute growth rate occurred when gulf menhaden larvae were 7.9 mm SL and 13 days old. The asymptotic length of larvae (21.5 mm SL), determined from the variables of the growth equation, is approximately the size when larvae begin to transform into juveniles. This transfor- mation, described by Lewis et al. (1972) for the closely related Atlantic menhaden, B. tyrannus, apparently ends when the fish reach 28-30 mm SL (Suttkus 1956). In all instances except one transect (TX Decem- ber 1981, where there was no convergence in the parameter values in the computer fitting proce- dure and the model would not fit the data), the Gompertz growth model could be used to describe the growth of gulf menhaden larvae from each cruise and transect (Figs. 5-7, Table 2). The growth model for the FL December 1980 larvae approximates an exponential form because of the exceptionally low value for a. This may be due to the preponderance of small, young larvae. GROWTH COMPARISONS Louisiana - Seasons and Years There were statistically significant differences (MANOVA, P < 0.001) in the growth curves for larvae caught off Louisiana for two seasons (De- cember, February) and three years (1979-80, 1980-81, 1981-82). To determine if differences ex- isted between seasons in each year and among any two years within each season, I selected 9 of the possible 15 pairwise comparisons for testing. The Bonferroni critical value in these tests was 0.0033 (0.05/15). The inability to fit a Gompertz growth model to the TX December 1981 data pre- cluded a comparison with the larvae collected off Texas in February 1982. Pairwise comparisons for within years data for Louisiana larvae showed significant differences (P < 0.003) in growth curves between early sea- son (December) and late season (February) for each year. Faster growth of early season larvae is evident if the respective curves (Figs. 5; 6a, c; 7a, c) are compared. For any age, the predicted size is greater for early season than for late season lar- vae. Only for the third year did the length at age 40+ days of February-caught larvae exceed that for December-caught larvae. In similar comparisons for larvae caught off Louisiana in December of all three years, there were significant differences (P< 0.003) in the growth curves (Figs. 5a, 6a, 7a) for any two years. As judged by the predicted size at any age, larvae appeared to grow faster in 1979 than in either 1980 or 1981. While the curves for the 1980 and 1981 larvae overlapped, larvae from 1980 were larger at 30+ days than were the 1981 larvae. Significant differences were also found among the curves (Figs. 5b, 6c, 7c) for larvae caught in 81 F1SHF:KV BUI.l.KTlN VOL Hli, NU 1 Table 2.— Estimates of Gompertz growth model parameters and mean age (days) and mean SL (mm) for larval gulf menhaden collected m the northern Gulf of Mexico during the winters 1979-80, 1980-81, and 1981-82. fl2 is the coefficient of determination for the respective models. Number fish Growth model parameters^ fvlean estimated tVlean Date Transect' aged R2 '-(O) A(0) a age (d) SL (mm) Winter 1979-80 Dec. 1979 LA 42 0863 2 768 (1.270) 1701 (0,0711) 00809 (00190) 28 1 174 Feb. 1980 LA 324 0954 2.131 (0 138) 1592 (00105) 00710 (0 0031) 302 12.7 Winter 1980-81 Dec. 1980 LA 191 0931 2888 (0.188) 1125 (0,0097) 00496 (0.0044) 220 118 Dec. 1980 FL 80 0701 3418 (0.994) 00561 (0 0370) 0.0001 (0,0407) 14 9 79 Feb. 1981 LA 338 0849 2702 (0.231) 1159 (0 0120) 00577 (0,0049) 246 11.7 Feb. 1981 TX 223 0526 5839 (0.579) 00305 (00088) 0,0125 (0,0110) 21,5 10.3 Winter 1981-82 Dec. 1981 LA 370 0921 0384 (0.076) 04780 (0,0433) 1240 (0,0054) 21 7 13.3 Dec. 1981 TX 114 — .1 3 ,1 25,4 11.9 Feb. 1982 LA 191 0,736 0807 (0,337) 02729 (0,6093) 0,0851 (0,0090) 31 2 15.5 Feb. 1982 FL 88 0.624 2278 (0,824) 1067 (0,0454) 0,0394 (0,0205) 17,6 89 Feb 1982 TX 42 0946 1 798 (0,392) 0,1276 (0,0340) 00500 (0,0142) 225 100 All years All data pooled 2,003 0822 2,355 (0.098) 0,1345 (0,0059) 0,0608 (0,0020) 24.4 12.6 iLA=IVIississippi River Delta (Southwest Pass Louisiana); FL = Cape San Bias, FL; TX = Galveston, TX. 2L|0| = length at hatching, A(0) = specific growth rate at hatching, « = exponential decline in A,o)- Values in parentheses are estimated standard errors from the nonlinear regressions. 3Gompertz growth model did not fit the data. E E X »- CD Z lU _l Q < Figure 4.— Growth of gulf menhaden larvae collectccl in the winters of 1979- 80, 1980-81, and 1981-82 m the north- ern Gulf of Mexico. The Gompertz growth model was used to describe the pooled data. Two through nine coincident data points are labelled with their numeral. Coincident points of 10 and above are la- belled A, B, etc. CO 30 25 20 15 10 1 1 1 1 1 3 1 1 1 1 1 1 1 2 2231 1 1111 1 11 1121 1 112 11 13313221331 21 4 222635231 8 3687iaC6K5H28 111 1 2194EaC6546383331 1 55E5BaC665A2A2 1 " ::: _ 31 1 i L59422 -0.0608t n= 2,003 10 20 30 40 50 60 70 ESTIMATED AGE (cJays) 82 WARLEN; AGE AND GROWTH OF LARVAL GULF MENHADEN E E I (- o z LU _l Q cr < Q W a ? 1 20 1 1 > ' 12^ ''^1 1 1 1 1 15 " A 1 1 1 10 - / t ' \ 5 - 1 1 1 1 Louisiana Dec 1979 1 1 10 20 30 40 50 60 20 15 10- 11 11 12 1 2 2232221 1 123 11 114" " ' 1 322 1 1 833122 1 32 212 1 223 121 10 20 30 40 Louisiana Feb 1980 50 60 ESTIMATED AGE (days) Figure 5. — Growth of larval gulf menhaden collected in the winter 1979-80 in the northern Gulf of Mexico. The Gompertz growth model was used to describe the data. Coincident data points are labelled as in Figure 4. February in three years. Larvae caught in 1980 grew faster than larvae caught in 1981 and larvae caught in 1982 up to 25 days, thereafter 1980 and 1982 had very similar size at age esti- mates. Louisiana vs. Florida vs. Texas - February 1982 There were no significant differences (MANOVA, P = 0.212) among the growth curves for larvae caught in February 1982 off Louisiana, Texas, and Florida (Figs. 7c, d, e), and hence no pairwise comparisons were necessary. Louisiana vs. Texas - February 1981-82 Statistically significant differences (MAN- OVA, P < 0.002) in larval growth were observed in the LA and TX transects from February 1981- 82. Pairwise comparisons indicated significant differences (P < 0.008) in the growth of larvae collected off Texas in 1981 (Fig. 6d) and 1982 (Fig. 7e) and in growth between LA 1981 (Fig. 6c) and TX 1981 collections. The earlier pairwise com- parisons had already shown a significant differ- ence in growth of larvae from LA February 1981 and February 1982 collections (Fig. 7c), but none for growth of larvae caught in the LA February 1982 and TX February 1982 collections. Two other potential tests, between transects of differ- ent areas and different years, were not considered to be meaningful. Larvae caught off Louisiana in February 1981 were larger at age 18+ days than were larvae caught off Texas in February 1981, and might be considered to be faster growing fish. There was a 83 FISHERY BULLETIN: VOL 86, NO 1 E E I H o z LU _J Q cr < Q Z < I- 25 20 15 10 5- 25 20 15 10 5- Louisiana Dec 1980 j_ Louisiana Feb 1981 Florida Dec 1980 -L, I 1 1 1 1 2 2 31J3512 1 1 4 244 2 126263M?22 3122ttaft?7311 511 1 __J3323 23 1 2 12131 11112 10 20 30 40 50 60 10 20 30 40 Texas Feb 1981 50 60 ESTIMATED AGE (days) Figure 6.— Growth of larval gulf menhaden collected in the winter 1980-81 in the northern Gulf of Mexico. The Gompertz growth model was used to describe the data. Coincident data points are labelled as in Figure 4. statistical difference in growth of larvae caught off Texas in February 1981 and 1982, and it ap- pears that the 1982 larvae grew at a faster rate. Conclusions from these statistical differences in- volving TX February 1981 larvae collections should be viewed with caution because of the rel- atively poor fit (r^ = 0.526) of the model. Inade- quacies, such as the lack of larvae <13 or >31 days old, in that data set probably resulted in the relatively poor parameter estimates (Table 2). Additional sampling would be necessary to fur- ther test the hypotheses of differences in growth between geographic areas in the northern Gulf of Mexico and between years for Texas larvae. 84 Estimated Spawning Times Gulf menhaden larvae observed in this study were estimated to have been spawned from mid- October to mid-February (Fig. 8). The limited ex- tent of seasonal sampling precluded estimation of the probable total range of the spawning season. Most larvae captured in December and February had been spawned in November and January re- spectively (Fig. 8). The considerable overlap in spawning times of larvae caught the same month in different years is a reflection of the similarity of sampling dates. The relatively narrow distribu- tion of spawning dates for larvae caught off Flor- WARLEN; AGE AND GROWTH OF LARVAL GULF MENHADEN 20 15 10 E 3 20 I (- (0 z lb LU _l D CC 10 < O z < t- b co 20 15- 10- Louisiana Dec 1981 I b - 2 41 1 173 721 1 1252925 1 21435131 1 13262 112 122 611 3 1 1 Texas Dec 1981 1 1 1 1 1 1 1 1 1512 21; 21246212 " 2 1611 12 51 i^21 1 t r4121 1 1 1212 ^\ 14 31 nil 222 21 1 2 1 1 Louisiana Feb 1982 _L Texas Feb 1982 J_ J- 10 20 30 40 ESTIMATED AGE (days) 50 60 Florida Feb 1982 I 10 20 30 40 50 ESTIMATED AGE (days) 60 Figure 7. — Growth of larval gulf menhaden collected in the winter 1981-82 in the northern Gulf of Mexico. The Gompertz model was used to describe the data from all transects except Texas December 1981 where it could not be made to fit the data. Coincident data points are labelled as in Figure 4. ida in both December 1980 (Fig. 6b) and in Febru- ary 1982 (Fig. 7d), off Louisiana in February 1982 (Fig. 7c), and off Texas in February 1981 (Fig. 6d) represent larvae from fewer cohorts. DISCUSSION Laboratory observations indicate that larval gulf menhaden on the average form one growth increment per day on their otoliths and that counts of these increments can be used to esti- mate age. Otoliths of larval gulf menhaden are thin and round, and the increments are generally easily counted and consequently are ideally suited for ageing. The most closely spaced incre- ments, those occurring near the focus, were at least 1.5 |xm wide and were above the 0.2 jxm resolution of the light microscope (Campana and Neilson 1985). First increment formation occurs about 5 days after spawning and probably coin- cides with first exogenous feeding. This is sup- ported by Hettler ( 1984) who found that gulf men- haden eggs hatched at about day 1.7 at 19°-20°C. Four days after hatching larvae had functional 85 FISHERY BULLETIN: VOL. 86, NO. 1 (42) 1979 D ILA (324) I — C 1980 LU < Q LU CC Q. < O (191) (80 )l— {JH FL rn — HLA 1980 (338)h (223)1 — 1981 LA c±i ^L/ -Cp ITX (88)1 IJHFL 1982 DEC I Jan I FEB ' SPAWNING DATE Figure 8. — Schematic plots of the spawning times of larval gulf menhaden collected in the northern Gulf of Mexico during 6 cruises of the RV Oregon II from December 1979 to February 1982. In each distribution the vertical line is the median value and 50% of the data points fall within the block. Lines beyond the boxes represent the range of data points. The value in parentheses to the left of each distribution is the number of fish. mouths and were 4.5 mm SL. However, develop- mental rates are probably temperature depen- dent (Powell and Phonlor 1986), and hence larvae at lower temperatures would be older at first feed- ing. The Gompertz growth model appears to ade- quately describe the growth of larval gulf men- haden in most cases. Except where data are some- what limited (Figs. 6b, d; 7b, d) the fit of the model is relatively good and the r'^ is >0.73 for each transect (Table 2). Gompertz gi'owth models have been used (Zweifel and Lasker 1976; Methot and Kramer 1979; Laroche et al. 1982; Warlen and Chester 1985) to describe growth of larval fishes where the length-age plots are nonlinear and upper asymptotes were apparent. Average growth rate of larval gulf menhaden to day 60 was 0.30 mm/day throughout its oceanic existence. This rate was very similar to that, 0.28 mm/day (estimated from figure 2 of Hettler 1984), for larvae reared in the laboratory at 20° ± 2°C for 60 days. However, wild-caught larvae were from wider extremes in water temperature, with mean early season (December) temperatures from 17.4° to 21.2°C and late season (February) 12.9° to 16.4°C. The growth rate of larval Atlantic herring, Clupea harengus , up to 50 days old was similar and varied between 0.23 and 0.30 mm/day (Lough et al. 1982). However, gulf menhaden lar- vae grew slower than the fast growing but rela- tively short-lived engraulids — bay anchovy, An- choa mitchilli (Fives et al. 1986) and northern anchovy, Engraulis mordax (Methot and Kramer 1979). Only a small number of larvae from all the collections were 2^50 days old. Larvae of this age 86 WARLEN; ACE AND GROWTH OK LARVAL C.rLF MENHADEN were either not in the sampling area or were inac- cessible to the fishing gear used. Although the latter cannot be fully discounted, the former pos- sibility is most likely, since larvae as they g:-ow are known to be transported (Shaw et al. 1985b) toward estuaries. Larvae are about 15-25 mm SL (estimated from Suttkus 1956) when they enter estuaries in Louisiana, and the smallest immi- grating larvae are estimated from the growth model 23 days old caught off Florida in Febru- ary 1982 (Fig. 7d) suggests that comparisons of that data set with the data sets for larvae caught off Louisiana and Texas in February 1982 would be of little value. The estimated spawning period for gulf men- haden extended from mid-October to mid- February (Fig. 8). These results agree with Fore (1970) and Christmas and Waller (1975) who, using the occurrence of eggs and larvae, esti- mated that gulf menhaden spawned from mid- October through March. Gonad weight-body weight ratios of adults (Lewis and Roithmayr 1981) and morphological and physiological fea- tures of ovarian tissue (Combs 1969) also indicate that spawning extends from October to early March. Based on the movement of late larvae into Lake Pontchartrain, Suttkus (1956) presumed that gulf menhaden spawning began in October and ceased in February. He suggested that the beginning and end of the spawning period fluctu- ates from year to year, and that there is no spawn- ing activity during the spring and summer months as Higham and Nicholson (1964) have reported for the closely related Atlantic men- haden. Most larvae caught in December were spawned in November (Fig. 8) regardless of the year. Lar- vae caught in February were spawned mostly in January but estimated spawning dates extended from mid-December to mid-February. For any given cruise, larvae from off Texas and Louisiana were spawned at about the same time. There was also considerable overlap in the spawning dates in any cruise off Florida and the other areas. The distribution of the central 507^ of spawning dates from the Louisiana sample in February 1980 ex- tended over a 29-d period and was wider than for any other data set. This unusually wide distribu- tion may have been due to the presence of two distinct cohorts, one spawned in late December and one in late January, collected on the Febru- ary 1980 cruise. Combs (1969) found that this species had intermittent total spawning. Lewis and Roithmayr (1981) inferred that gulf men- haden were intermittent, or fractional spawners. Christmas and Waller (1975) noted a modal tem- poral distribution of eggs in the region from the Mississippi delta to east of Cape San Bias. Bal- dauf sampled young menhaden in the lower Neches River, TX, from November through April and found two incoming populations from which he suggested that there may have been two spawning peaks. Only in the larval collections of December 1981 did spawning date distribution appear to be bimodal; 7 and 20 November for Lou- isiana and 8 and 19 November for Texas. Future sampling throughout the spawning season will be necessary to determine the seasonal periodicity and peaks of gulf menhaden spawning. Relative numbers of larvae in cohorts within the spawning season could then be compared with measure- ments of environmental conditions as a test of the match-mismatch hypothesis (Cushing 1975) and to further test, as Methot (1983) has done, whether larvae spawned during favorable envi- ronmental periods constitute the greatest per- centage of the year class. ACKNOWLEDGMENTS I thank the following persons of the Beaufort Laboratory: M. Boyd who extracted, mounted, and aged otoliths, W. Hettler who spawned gulf menhaden and furnished eggs for the laboratory experiments, A. Chester who advised on statisti- cal procedures and problems, D. Ahrenholz and A. Powell who reviewed early drafts of the manuscript, and the crew and scientists on the RV Oregon II cruises. P. Ortner of the Atlantic Oceanographic and Meteorological Laboratories, NOAA, Miami, FL, provided raw data from which zooplankton counts were summarized. This re- search was supported by a contract from the Ocean Assessments Division, National Ocean Service, NOAA. LITERATURE CITED Bernard. D R 1981. Multivariate analysis as a means of comparing growth in fish. Can. J. Fish. Aquat. Sci. 38:233-236. Campana, S. E., and J D. Neilson, 1985. Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Christmas. J Y , and R S Waller. 1975. Location and time of menhaden spawning in the 4Baldauf R. J. 1954. Survey and study of surface and sub- surface conditions in and around Beaumont, Texas. Biological survey of the Neches River in the region of Beaumont, Texas. Texas A&M Res. Found., Mimeo. Rep., 184 p. 88 WARLEN; AGE AND GROWTH OF LARVAL GULF MENHADEN Gulf of Mexico. Gulf Coast Res. Lab.. Ocean Springs, MS, 20 p. Combs. R M 1969. Embryogenesis, histology, and organology of the ovary of Breioortia patronus . Gulf Res. Rep. 2:333- 434. CUSHINO. D H 1975. Marine ecology and fisheries. Cambridge Univ. Press. Cambridge, Eng., 278 p. Fives. J M , S M Warlen. and D E Hoss 1986. Aging and growth of larval bay anchovy, Anchoa mitchilli, from the Newport River estuary, North Caro- lina. Estuaries 9(4B):362-367. Fore. P L 1970. Eggs and larvae of the Gulf menhaden. In Report of the Bureau of Commercial Fisheries Biological Labo- ratory, Beaufort. N.C., p. 11-13. U.S. Fish Wildl. Serv. Circ. 341. GovoNi. J J , D E Hoss, AND A J Chester 1983. Comparative feeding of three species of larval fishes in the northern Gulf of Mexico: Brevoortia patronus, Leiostomus xanthurus, and Micropogonias undula- tus. Mar. Ecol. Prog. Ser. 13:189-199. GUILLORY. V . J GEAGHAN. AND J. ROUSSEL 1983. Influence of environmental factors on gulf men- haden recruitment. La. Dep. Wildl. Fish. Tech. Bull. 37, 32 p. Harris. R J 1975. A primer of multivariate statistics. Acad. Press, N.Y., 332 p. Hettler, W F 1983. Transporting adult and larval gulf menhaden and techniques for spawning in the laboratory. Prog. Fish- Cult. 45:45-48. 1984. Description of eggs, larvae, and early juveniles of gulf menhaden, Brevoortia patronus. and comparisons with Atlantic menhaden, B. tyrannus, and yellowfin menhaden, B. smithi . Fish. Bull.. U.S. 82:85-95. HiGHAM, J R . AND W R NICHOLSON 1964. Sexual maturation and spawning of Atlantic men- haden. Fish. Bull., U.S. 63:255-271. Hoenig. N a . AND R G Hanumara 1983. Statistical considerations in fitting seasonal growth models for fishes. Cons. Int. Explor. Mer, CM. 1983/ D:25, 25 p. JONES. C 1985. Within-season differences in growth of larval At- lantic herring, Clupea harengus harengus . Fish. Bull., U.S. 83:289-298. Laird. A K . S A Tyler, and A D Barton 1965. Dynamics of normal growth. Growth 29:233- 248. Laroche. J L . S L Richardson, and A A Rosenberg 1982. Age and growth of a pleuronectid, Parophrys ve- tulus, during the pelagic larval period in Oregon coastal waters. Fish. Bull., U.S. 80:93-104. Laurence. G C 1978. Comparative growth, respiration and delayed feed- ing abilities of larval cod (Gadus morhua ) and haddock {Melanogrammus aeglefinus) as influenced by tempera- ture during laboratory studies. Mar. Biol. (Berl.) 50:1- 7. Laurence, G C . A S Smigielski. T A Halavik. and B R Burns 1981. Implications of direct competition between larval cod {Gadus morhua I and haddock i Melanogrammus ae- glefinus ) in laboratory growth and survival studies at different food densities. Rapp. P. -v. Reun. Cons. int. Ex- plor. Mer 178:304-311. Lewis. R M . and C M Roith.mayr 1981. Spawning and sexual maturity of gulf menhaden, Brevoortia patronus . Fish. Bull., U.S. 78:947-951. Lewis. R M . E P H Wilkins, and H R Gordy 1972. A description of young Atlantic menhaden, Bre- voortia tyrannus, in the White Oak River estuary, North Carolina. Fish. Bull., U.S. 70:115-118. Lough. R G . M Pennington. G R. Bolz. and A A Rosenberg 1982. Age and growth of larval Atlantic herring, Clupea harengus L., in the Gulf of Maine-Georges Bank region based on otolith growth increments. Fish. Bull., U.S. 80:187-199. Methot. R D. JR 1983. Seasonal variation in survival of larval northern anchovy, Engraulis mordax. estimated from the age dis- tribution of juveniles. Fish. Bull., U.S. 81:741-750. Methot, R.. D., Jr , and D Kramer 1979. Growth of northern anchovy, Engraulis mordax, larvae in the sea. Fish. Bull., U.S. 77: 413-423. Pannella. G 1971. Fish otoliths: Daily growth layers and periodical patterns. Science (Wash., D.C.) 173: 1124-1127. Pimentel. R a 1979. Morphometries, the multivariate analysis of biolog- ical data. Kendell/Hunt Publ. Co., Dubuque. lA, 276 P- Powell. A B , and G Phonlor 1986. Early life history of Atlantic menhaden, Brevoortia tyrannus , and gulf menhaden, B . patronus . Fish. Bull., U.S. 84:991-995. Reintjes. J W 1970. The gulf menhaden and our changing estuar- ies. Proc. Gulf Caribb. Fish. Inst. 22:87-90. Shaw. R F . J H Cowan. Jr , and T L Tillman 1985a. Distribution and density of Brevoortia patronus (gulf menhaden) eggs and larvae in the continental shelf waters of western Louisiana. Bull. Mar. Sci. 36:96- 103. Shaw. R F . W J Wiseman, Jr . R E Turner. L J Rouse. Jr . R E Condrey. and F J Kelly. Jr 1985b. Transport of larval gulf menhaden Brevoortia pa- tronus in continental shelf waters of western Louisiana: A hypothesis. Trans. Am. Fish. Soc. 114:452-460. Suttkus, R D 1956. Early life history of the gulf menhaden. Brevoortia patronus. in Louisiana. Trans. N. Am. Wildl. Conf 21:390-407. Tanaka. R. Y Mugiya. and J Yamada 1981. Effects of photoperiod and feeding on daily growth patterns in otoliths of juvenile Tilapia nilotica. Fish. Bull., U.S. 79:459-466. Turner. W R 1969. Life history of menhadens in the eastern Gulf of Mexico. Trans. Am. Fish. Soc. 98:216-224. US National Marine Fisheries Service 1986. Fisheries of the United States, 1985. U.S. Natl. Mar. Fish. Serv. Curr. Fish. Stat. 8380, 121 p. Warlen, S. M.. and A. J Chester. 1985. Age, growth, and distribution of larval spot, Leios- tomus xanthurus, off North Carolina. Fish. Bull., U.S. 83:587-599. 89 FISHERY BULLETIN: VOL. 86, NO. 1 WiEBK. P H . K H Burt. S H Boyd, and A W Morton Zweifei,. J R . and R Lasker 1976. A multiple opening/closing net and environmental 1976. Prehatch and posthatch growth of fishes — a general sensing system for sampling zooplankton. J. Mar. Res. model. Fish. Bull., U.S. 74:609-621. 34:313-326. 90 SOURCES OF VARIATION IN CATCH PER UNIT EFFORT OF YELLOW! AIL FLOUNDER, LIMANDA FERRUGINEA (STORER), HARVESTED OFF THE COAST OF NEW ENGLAND LORETTA O BrIEN AND RALPH K. MaYO^ ABSTRACT Factors affecting variability in commercial catch per unit effort (CPUE) of yellowtail flounder were examined in order to establish a basis for standardizing fishing effort. Analysis of variance ( ANOVA) procedures were employed to test for differences in CPUE among vessel tonnage class, fishing area, and depth zone and the interactions between tonnage class and area, and tonnage class and depth. Vessel tonnage class and fishing area accounted for highly significant IP < 0.01) sources of variation in CPUE whereas depth was not significant (P > 0.05) in most cases. Interactions between tonnage class and stock area were also highly significant in all cases. A series of annual fishing power coefficients was computed for each tonnage class relative to a standard for each stock based on parameter estimates obtained by fitting the CPUE observations to a linear model with tonnage class as the independent variable. Deviations of annual fishing power coefficients from the 20-year mean were found to exhibit significant first order autocorrelations. Consequently, annual coefficients were computed over the entire 1964-83 period by incorporating tonnage class, annual and seasonal effects as independent variables in a three-way linear model. Although the standardized CPUE estimates obtained from this procedure are similar to those obtained by previous methods, the revised proce- dures described in this paper insure adequate representation of all vessel classes engaged in the yellovfcrtail fishery in the CPUE calculations. Fishing effort and resulting catch per unit effort (CPUE) indices are routinely used in assessing the impact of commercial fishing operations on stock abundance. The traditional concept that ag- gregate CPUE indices may be used to measure annual changes in relative stock abundance is based on the principal assumption that the catch- ability coefficient (q) either remains constant over all fleet components, or that nominal effort is adjusted to account for differences in relative effi- ciencies (Pope and Parrish 1964; Kimura 1981). Variation in q may be due to persistent differ- ences in fishing power of various types of gear or to technological innovations which may be gradu- ally introduced over time (Gulland 1964; Sis- senwine 1978). Biological interactions such as changes in availability of a species due to sea- sonal distribution patterns or to annual changes in abundance may also affect the overall catcha- bility of demersal species (Garrod 1964; Pope and Garrod 1975). Variability in catchability coeffi- cients may be taken into account by relating nom- inal fishing effort of each fleet component to some chosen standard category. 'Northeast Fisheries Center Woods Hole Laboratory, Na- tional Marine Fisheries Service. NOAA, Woods Hole, MA 02543. Manuscript accepted October 1987. FISHERY BULLETIN; VOL. 86. NO. 1, 1988. Numerous authors have described the basic procedures for calculating relative fishing power of various fleet components. Beverton and Holt (1957) provided evidence to suggest that the dis- tribution of logarithms of fishing power factor/ vessel tonnage ratios could be described by a nor- mal curve while Gulland (1956) employed an analysis of variance (ANOVA) model of log CPUE. The properties of the ANOVA model were further examined by Robson (1966) who extended the techniques developed by Gulland (1956) and formally specified the analysis of Beverton and Holt (1957) as a two-way multiplicative ANOVA model. Stern and Hennemuth (1975) employed the method of Robson (1966) in their analysis of fishing effort in the U.S. Georges Bank haddock fishery using depth fished and vessel tonnage as classification variables. In a previous study, Rounsefell (1957) computed standardized log CPUE indices to determine relative abundance of several co-occurring species on Georges Bank ac- cording to depth. More recently, Gavaris (1980) and Kimura (1981) have developed modifications of the ANOVA model to estimate annual stan- dardized CPUE indices from time series of catch and effort data by incorporating a year effect in the model. Standardized annual CPUE indices based on 91 I-ISIIKKY lUU.l.KTIN: VOL, 86, NO. 1 criteria established by Lux (1964) have been rou- tinely used to monitor relative abundance of three stocks of yellowtail flounder, Limanda fer- ruginea (Storer), in the commercial fishery off the New England coast (Fig. 1 ). Lux calculated CPUE indices for otter trawlers ranging from <26 gross registered tons (GRT) to 100 GRT based on trips in which yellowtail flounder accounted for bQ^/< or more of the total landed weight between 1942 and 1961. A fishing power coefficient was then com- puted for each of several GRT categories as the ratio of CPUE to a standard GRT category CPUE for the entire timespan. A separate set of fishing power coefficients was computed for each of the three stocks. Lux's (1964) work improved upon an earlier analysis of yellowtail flounder CPUE by Royce et al. ( 1959) which was based only on rela- tively small vessels ranging in size from 5 to 50 GRT that dominated the fishery during the 1940's. Since 1964, numerous technological innova- tions have drastically changed the character of the New England fishing fleet as traditional side trawlers have gradually given way to larger, more efficient stern trawlers equipped with so- phisticated electronic navigation and hydroa- coustic devices. This gradual alteration in the fleet characteristics over time suggests that pre- viously documented relationships among vessel categories may no longer be applicable to the cur- rent fishery, and that use of nominal effort in CPUE calculations will tend to overestimate rela- tive abundance in the more recent years (Westrheim and Foucher 1985). Long-term de- clines in yellowtail flounder abundance on each of the principal fishing grounds (Clark et al. 1984) also indicate that current catchability coefficients may differ from previous values. Accordingly, up- dated fishing power coefficients are required to adequately assess changes in effective fishing ef- fort and CPUE which have occurred during the past two decades. Further, to obtain annual effort and CPUE estimates over such a broad period of years, techniques for computing relative fishing power should incorporate a time element in the analysis. In this paper we examine variation in CPUE with respect to fishing area, depth, vessel tonnage class, season, and year for three stocks of yellow- tail flounder on Georges Bank, Southern New England, and Cape Cod grounds between 1964 and 1983. Before evaluating differences in rela- tive fishing power among vessel classes, we inves- tigate potential interactions between tonnage class and area and tonnage class and depth within each year, and partition the data to mini- mize tonnage class-area interactions. For each stock, fishing power coefficients are examined for 68 67 66 Figure 1.— Yellowtail flounder stocks off the coast of New England (After Lux 1963). 92 O'BRIEN AND MAYO: CPUE OF YELLOWTAIL FLOUNDER annual and seasonal interactions. A three-way linear model, incorporating annual and seasonal components, is employed to compute relative fishing power coefficients over the entire 20-yr period and estimate annual standardized CPUE indices. DISTRIBUTION OF CATCH AND EFFORT Commercial exploitation of yellowtail flounder began in the late 1930's following the decline of the winter flounder fishery (Royce et al. 1959). Nominal catches^ for the three grounds combined rapidly increased to 31,500 metric tons (t) in 1942 but subsequently declined to 5,500 t in 1954. Landings by U.S. vessels gradually increased to a record high of 36,900 t in 1963, but declined again to 10,500 t in 1978 (Fig. 2). Distant water fleet (DWF) catches were also substantial during this 2Nominal catch defined as live weight equivalent of landings, excluding discards. period, peaking at 20,700 t in 1969. Overall catches from the three fishing grounds have re- cently increased to 30,800 t in 1983, although 1984 landings declined to 15,500 t (Clark et al. 1984). The decline in catch during the 1940's was not due to overfishing (Royce et al. 1959) but may have been related to a warming trend in the re- gion which affected recruitment (Sissenwine 1974). The more recent decline between 1969 and 1978, however, has been attributed to increased fishing effort by both domestic and distant water fleets (Brown et al. 1980). In the early 1940's the size of vessels fishing for yellowtail flounder varied from 5 to 75 GRT. The predominant vessels on Southern New England and Cape Cod grounds ranged from 26 to 50 GRT; on Georges Bank, the dominant vessels were in the 51-75 GRT range. By the mid-1960's larger vessels had begun to enter the fishery, increasing the maximum size to 215 GRT. During this period the size range of the dominant vessels on South- ern New England grounds and on Georges Bank had increased to 51-72 and 73-104 GRT, respec- 55 n O CO O O o I O) c C <0 50- 45 'cJj S 40 30- Total DWF nVja-yy 1940 1945 1950 1955 1960 1965 Year Figure 2. — Yellowtail flounder landings (metric tons) by United States and distant water fleet (DWF) vessels from the combined Georges Bank, Southern New England, and Cape Cod grounds, 1940-84. 93 FISHKKY MUI.LKTIN: VOL. 86, NO 1 tively. Vessels fishing the relatively nearshore Cape Cod grounds remained in the 34-50 GRT range. Larger vessels continued to enter the fish- ery during the 1970's and, by 1983, several ves- sels were in the 311-400 GRT range. Since 1964, vessel CRT's have been categorized by tonnage class (TO as given in Table la. A review of Lux's (1964) data from 1942 to 1961 and the distribution of more recent yellowtail flounder landings from 1964 to 1983 (Figs. 3-5) reveal that vessels of similar size have continu- ally fished the same general areas over the past 40 years. The TC 21-24 vessels fish primarily on Southern New England and Cape Cod grounds (Fig. 3 1, although TC 24 vessels occasionally enter the southwest part of Georges Bank. The TC 25-33 vessels fish on both Georges Bank and Southern New England grounds (Fig. 4), while TC 41-43 vessels concentrate on Georges Bank (Fig. 5). Although TC 41 vessels operate at times on the eastern part of the Southern New England grounds, the TC 42 and 43 vessels fish exclusively on Georges Bank. The distribution plots (Figs. 3-5) reveal a grad- ual phaseout of smaller (TC 21-24) vessels on the inshore Southern New England and Cape Cod grounds and a concurrent increase in the activity of large (TC 41-43) vessels on Georges Bank, Southern New England, and, to a lesser extent, on the Cape Cod grounds. In evaluating trends in CPUE we must ask whether these changes in the yellowtail fishery over the past 20 years (i.e., the shift in the predominant vessel size on two of the Table 1. — Definition of vessel tonnage classes (a) and depth ranges and corresponding zones (b) included in analysis of vari- ance of yellowtail flounder CPUE. a Gross registered tonnage Vessel tonnage (range) class 5-10 21 11-15 22 16-22 23 23-33 24 34-50 25 51-72 31 73-104 32 105-150 33 151-215 41 216-310 42 311-400 43 b Depth range Depth (m) zone 1-55 1 56-110 2 111-183 3 three grounds and the addition of larger vessels to the fieet on all three grounds) affect CPUE as calculated by the traditional method (Lux 1964). If the same size range of vessels (5-100 GRT) had fished for yellowtail flounder throughout the years, a shift in the dominant vessel class would not affect CPUE estimates since effort would be standardized against the same class and is, there- fore, relative. However, the maximum vessel size has increased and the predominant TC now repre- sents vessels larger than 100 GRT. Since landings and effort data contributed by these larger vessels were not incorporated into previous CPUE calcu- lations, CPUE estimates will not necessarily rep- resent overall fleet performance in recent years. The following procedures, therefore, were devel- oped to calculate new fishing power coefficients that encompass the entire size range of vessels currently in the fishery. METHODS OF ANALYSIS Catch and effort data recorded by trip were ob- tained from Northeast Fisheries Center (NEFC) detailed commercial landings files. Fishing effort or days fished (df) is defined on a 24-h basis as number of hours of actual fishing time divided by 24. Only trips landing 50% or more of yellowtail flounder were analyzed; trips included within the qualification level generally accounted for 70- 90% of the total yellowtail landings over the en- tire period, except on Cape Cod grounds where qualified trips accounted for 40-60% of the total. Catch per day fished (CPUE) was computed for each trip and transformed to In CPUE since pre- liminary analysis indicated a positive correlation of the arithmetic mean CPUE with the variance. Use of the log transformation, however, stabilized the variance and created a lognormal distribution as noted by Gulland (1956) and Steel and Torrie (1980). Trips landing between 1964 and 1983 are clas- sified in the data base by vessel tonnage class, statistical area, and depth zone fished. Vessels ranging in size from 5 to 310 GRT (Table la) and statistical areas corresponding to the three major stocks were selected for analysis as follows: Georges Bank (areas 522-525), Southern New England (area 526 and 537-539), and Cape Cod (areas 514 and 521) (Fig. 6). Because of their spo- radic representation, TC 21-23 vessels have been excluded from the Georges Bank analyses and have been combined as one category on the South- ern New England grounds. Depth zones 1, 2, and 94 O'BRIEN AND MAYO CPUE OF YELLOWTAIL FLOUNDER O — — — "X^ to o IT) O o o o o f\j m =r u — LJ o — z a: a: «- O — . . . . ^ o \r> o o o — nj m • •»!* ST^'^ CM I 1— ( bo aj c c o cS CO 00 05 T3 c (8 05 O 05 in CO 05 bo C ■3 c (8 c 3 CO c o 3 95 FISHERY BULLETIN: VOL 86, NO. 1 CO CO lO (M w (V CO 1 « U-) ca OJ CD (D LlJ be tx CO C —i c (_) rn o CO ■w Lu cn ^M O — a; cr -■ « z cc z cr O LiJ t— >- u a CO 00 OS 1—1 -a a CD to o 05 tH O t~ 05 r-l U5 CO o> Z o oj o in o in o z — f\j i/l r- IT) _ (_, o — Z (T cr "- —1 UJ O l/i o o o oo O Ul O — Aj in _, O UJ 03 ts^ z a O UJ bo S 1 OJ -a c 3 o c _o "-^ 3 'u in (5 D O 96 O'BRIEN AND MAYO: CPUE OF YELLOWTAIL FLOUNDER CO I CO be ;2 CO 00 Oi c o 09 bo C s ,2 >H a> -a e 3 o t 3 Q P 2 97 FISHERY BULLETIN: VOL. 86. NO. 1 Figure 6. — Yellowtail flounder fishing grounds defined by U.S. statistical area are as follows: Southern New England. 526-539; Georges Bank, 522-525; Cape Cod, 514 and 521. 3 (Table lb) were also selected from six possible zones based on the bathymetric distribution of yellowtail flounder. Trip data were aggi'egated at different levels of spatial resolution to examine variability in CPUE over the entire region and within each of the three established stocks. Two-way ANOVAs with interaction were performed on annual data sets using the BMDP statistical software progi'am P4V (Dixon 1981). Given the large number of ob- servations in each analysis, the more rigorous 99% significance level was chosen to test the null hypothesis (no significant differences) since relatively small differences in mean CPUE can produce statistically significant results. The ANOVA was performed initially to test for differ- ences in CPUE among all tonnage classes and statistical areas and to determine the overall ex- tent of tonnage class-area interactions. Second- ary analyses were performed to examine the ef- fect of tonnage class-area and tonnage class-depth interactions among and within each of the stocks. All subsequent tests for significance of tonnage class, area, and depth main effects were per- formed with the interaction effects absorbed in the error sum of squares. Estimates of annual geometric mean CPUE were obtained by combi- nations of tonnage class, stock area, and depth from the row and column means provided by the P4V software (Tables 2, 3). A standard vessel class was selected for each of the three stocks for use in calculation of fishing power coefficients based on the prevalence of the vessel class in the fishery and its relative contri- bution to the landings over the 20 years. The TC 32 category was chosen as the standard for both Georges Bank and Southern New England stocks, and the TC 25 class was chosen as the standard for the Cape Cod stock. Within each stock annual fishing power coefficients were derived for each tonnage class relative to the standard by fitting In CPUE to a one-way linear model using the GLM procedure of the Statistical Analysis Sys- tem (SAS Institute 1982) as follows: U = CC + ^ [^jXj] + ^. Annual deviations of the coefficients from the 20-yr mean were tested for first order autocorre- lation using the Durbin-Watson test statistic (Neter and Wasserman 1974). A time component was subsequently incorporated in the linear model to account for annual trends in the data; seasonal effects were also included by classifying the data according to calendar quarter. The ini- tial year (1964) and the fourth quarter were se- lected as reference categories. The general model is specified as: 98 O'BRIEN AND MAYO: CPUE OF YELLOWTAIL FLOl'NDER Table 2. — Geometric mean CPUE^ (landings per day fished, metric tons) for Georges Bank, Southern New England, and Cape Cod yellowtail flounder trips by vessel tonnage class, 1964-83. Vessel tonnage class Year Vessel tonnage class Year 21 22 23 24 25 31 32 33 41 42 21 22 23 24 25 31 32 33 41 42 Georges Bank 1964 — — — 3.65 4.17 4.88 5.04 4.58 4.52 2.14 1974 — 1.08 1 62 231 2.14 2.07 2.07 2.64 1.30 — 1965 — — — 2.41 3.06 3.62 3.55 3.39 4.97 — 1975 — 0.57 1.57 1.13 1.29 1.33 1.55 1.55 1.72 — 1966 — — 1.18 1.25 2.27 276 2.78 2.76 2.71 — 1976 — 0.72 0.90 0.53 0.86 1.13 1 32 1.69 2.06 — 1967 — — — — 2.20 2.74 2.96 278 265 — 1977 — 0.93 1.73 1.44 1.77 1.61 1.82 1.98 1.81 — 1968 — — — 3.88 2.99 3.47 3.55 384 3.53 2.66 1978 — — 2.72 1.15 1.58 1.52 2.04 2.26 282 — 1969 — — — 3.28 2.82 3.06 3.30 3.02 2.81 3.88 1979 0.79 1.74 2.28 1.59 2.17 2.32 2.93 2.88 3.64 — 1970 — — — 3.74 2.40 2.88 308 2.81 2.70 — 1980 — 1.02 246 1.91 2.54 2.68 2.81 3.07 3.00 — 1971 — — — 1.66 2.07 2.70 252 2.24 2.35 — 1981 — 1.88 2.50 1.43 2.58 2.76 2.61 3.28 2.23 — 1972 — — — — 1.96 224 239 2.29 1.96 — 1982 249 1.96 2.42 2.29 3 10 3.20 3.72 3.98 3.97 — 1973 — — — 1.38 2.64 2.61 2.81 3.05 2.76 3.20 1983 8.81 1.74 2.51 3.35 3.43 3.27 3.65 3.57 3.79 — 1974 — 2.52 — — 2.79 2.13 2.23 2.19 1.30 3.42 pane Cnd 1975 — — 1.93 1.91 1.64 1.93 2.04 1.72 3.19 *» ^i^l^f u » 1976 1.35 1.69 1.77 1.84 2.16 2.06 1964 — 1.65 1.38 1.94 2.64 2.85 2.76 — — — 1977 1.17 1.66 1.77 226 1.81 1965 — 1.45 1.20 1.59 2.16 3.56 2.12 2.07 — — 1978 1.67 1.50 1.72 2.03 282 1966 1.13 1.07 1.23 1.48 2.55 3.64 1.76 — — — 1979 2.69 1.31 1.80 2.41 2.37 2.82 3.64 1967 — 0.78 0.90 1.56 2.46 3.21 1.95 2.62 — — 1980 0.30 1.09 0.90 2.88 2.62 3.27 3.00 1968 — 1.16 0.99 1.89 2.65 4.11 2.42 — 1.13 — 1981 _ 1.12 1.24 1.98 2.48 2.61 223 1969 — 1.66 1.33 202 2.77 3.64 2.47 — — — 1982 1.88 0.48 2.02 1.71 2.10 2.53 397 2.07 1970 — 1.03 1.05 2.34 2.33 3.64 26.96 234.18 2.91 — 1983 — — — — 0.79 2.46 1.97 2.21 3.79 — 1971 1972 1.71 1.17 1.63 1.89 1.16 2.43 2.02 2.04 1.85 2.24 2.02 1.90 1.46 2.45 1.79 3.36 — 1.93 — Southern New England 1973 1.02 1.32 1.06 1.87 1.99 1.90 1.91 1 63 284 — 1964 234 1.94 6.15 4.88 4.66 4.44 5.04 4.59 4.09 — 1974 0.83 1.15 0.90 1.59 2.08 1.74 1.82 1.87 1.37 — 1965 1.77 1.55 3.87 3.62 3.12 3.56 3.81 4.05 1.50 — 1975 0.77 1.13 1.23 1.42 1.92 1.45 1.64 1.09 1.19 — 1966 1.21 306 3.19 2.93 2.52 261 2.56 2.35 3.03 — 1976 0.34 1.31 1.38 1.69 1.93 1.42 1.54 1.88 2.15 — 1967 1.80 1.14 294 3.38 323 2.67 2.66 2.31 1.80 — 1977 0.28 0.86 1.25 1.50 2.03 1.45 1.42 2.19 2.57 — 1968 2.24 4.15 4.69 4.42 4.02 3.41 3.62 3.63 3.53 — 1978 — 0.84 1.59 1.90 2.11 1.96 1.53 1.92 5.99 — 1969 265 372 5.17 3.81 4 10 3.39 3.43 4.01 2.81 — 1979 0.98 0.91 1.42 1.83 2.10 2.24 1.74 2.23 3.12 — 1970 7.74 3.28 4.76 3.77 3.96 3.08 3.64 2.87 2.70 — 1980 0.26 0.58 1.30 1.73 2.18 2.24 2.07 2.19 3.34 — 1971 11.72 1.60 4.06 3.14 3.32 2.85 2.87 3.16 2.35 — 1981 0.70 0.76 1.17 1.69 2.24 2.40 1.22 1.85 2.99 — 1972 — 1.96 3.19 2.62 3.30 3.01 2.96 3.04 1.96 — 1982 0.54 0.71 1.22 1.79 2.02 2.41 1.57 1.94 2.37 — 1973 — 1.01 1.05 2.26 2.66 2.29 2.22 2.12 2.76 — 1983 0.69 1.51 1.23 1.96 1.91 1.29 1.53 1.46 0.83 — ^Calculated as exp 1 nX In ' landings I effort I 20nly one trip by tonnage classes 32 and 33 in 1970 on Cape Cod grounds. u a + 2! lP#(/] + e where U = \n CPUE, a = intercept estimate, P,Y = model parameter estimates in loga- rithmic units for category 7 for ton- nage class, season, and year, X,j = dummy variable for tonnage class, season, and year ( = 1 when category j occurs; = otherwise), and e — error term. All tests for significance of main effects were based on the above model without interaction. Separate ANOVAs were also performed to exam- ine first order interactions. Parameter estimates obtained from the model without interaction were retransformed following methods described by Bradu and Mundlak (1970) to derive unbiased fishing power, seasonal, and annual coefficients. Annual coefficients corresponding to the 1965-83 period were multiplied by the reference year CPUE to compute annual standardized CPUE es- timates. RESULTS Smaller vessels (TC 21-24) generally exhibited the lowest CPUE indices in all three areas, al- though TC 21-23 vessels were not represented on Georges Bank (Table 2). Catch rates of medium vessels (TC 25; 31-33) were similar to each other, and were generally greater than those for TC 21- 24 vessels. Mean CPUE indices for the largest vessels (TC 41 and 42) were more variable, but generally were equal to or greater than those 99 FISHERY BULLETIN: VOL. 86, NO. 1 Table 3. — Geometric nriean CPUE^ (landings per day fished, metnc tons) for Georges Bank, Southern New England, and Cape Cod yellowtail flounder tnps by depth zone, 1964-83. Georges Bank So ut herr New England pth zone Cape Cod Depth zone De Depth zone Year 1 2 3 Combined 1 2 3 Combined 1 2 3 Combined 1964 3.92 4.91 2.74 4.80 5.05 4.96 4.71 503 2.13 2.36 — 2.25 1965 3.76 3.48 — 3.55 3.61 3.88 2.28 366 1.61 1.80 2.36 1.71 1966 3.02 2.62 1.29 2.71 2.71 2.66 0.71 2.71 1.99 2.13 1.70 2.07 1967 3.19 2.69 2.81 2.79 2.97 2.77 1.14 2.91 1.52 2.16 — 1.80 1968 3.38 3.51 — 3.49 3.70 3.97 3.24 3.76 2.15 2.31 — 2.24 1969 3.36 3.11 — 3.16 3.62 3.93 1.58 3.66 2.16 2.66 — 2.27 1970 2.81 2.99 1.00 2.93 3.44 3.68 1.62 3.49 1.10 2.24 234.18 1.40 1971 2.20 2.30 2.75 2.28 3.26 2.96 — 3.14 1.03 1.94 — 1.46 1972 2.28 2.37 1.34 2.35 3.28 3.25 2.74 3.27 1.56 1.88 — 1.82 1973 2.30 3.06 2.85 2.91 2.41 2.55 4.40 2.47 1.75 1.83 2.27 1.81 1974 2.10 2.32 2.90 2.30 2.31 2.02 — 2.19 1.86 1.78 — 1.82 1975 1.85 1.99 1.29 1.98 1.59 1.63 — 1.60 1.53 1.38 1.32 1.46 1976 1.75 2.06 — 2.03 1.34 1.56 — 1.43 1.54 1.88 2.93 1.70 1977 1.92 2.15 1.67 2.14 1.92 1.97 — 1.94 1,70 1.89 1.27 1.77 1978 1.78 2.09 1.98 2.07 2.45 2.17 — 2.32 1.90 1.71 3.49 1.86 1979 2.47 2.79 2.04 2.74 2.88 2.39 — 2.76 2.03 2.03 2.47 2.03 1980 2.58 3.50 1.53 3.37 2.71 3.51 1.13 3.04 2.06 2.31 2.38 2.16 1981 2.26 2.88 1.54 2.76 2.46 2.78 — 2.58 1.74 1.96 5.67 1.79 1982 2.34 2.66 2.53 2.62 3.20 3.54 4.53 3.31 1.81 1.44 0.91 1.67 1983 1.98 2.28 3.04 2.26 3.08 3.13 5.91 3.10 1.70 1.41 — 1.62 1 Calculated as exp 1 n-:i In ( andings effort )] 20nly one tnp in depth zone 3 in 1970 on Cape Cod grounds. corresponding to medium and small vessels, par- ticularly in the later years (Table 2). The initial ANOVAs performed over all statistical areas re- vealed highly significant iP < 0.01) differences in CPUE for tonnage class and area main effects in each of the 20 years (Table 4). The interaction of tonnage class and area was also highly signifi- cant in all years, suggesting that relative fishing power of the individual vessel classes varies ac- cording to area. ANOVA results for the compari- son of CPUE among stocks were highly signifi- cant for area main effects in 19 out of 20 years, and the tonnage class-stock area interaction term was highly significant in all years (Table 4). Grouping the data according to stock tended to reduce the amount of significant tonnage class- area interaction within each stock, although dif- ferences among tonnage classes remained highly significant. On Georges Bank the differences in CPUE were highly significant for statistical area and tonnage class main effects in 80 and lOO'/r of the years, respectively, while the tonnage class-area inter- action was highly significant in only 40% of the years. Differences due to area on Southern New England grounds were highly significant in all years except 1978, and differences due to tonnage class were highly significant in all years. The in- teraction term was highly significant in 70% of the years. Differences due to area on the Cape Cod grounds were highly significant in all years ex- cept 1975, and differences due to tonnage class were highly significant for all years. The interac- tion was highly significant in only 35% of the years (Table 4). Differences in CPUE by depth zone were gener- ally not significant. Depth main effects yielded Table 4. — Frequency with which highly significant (P < 0.01) re- sults were obtained from analysis of vanance (ANOVA) tests of yellowtail flounder annual CPUE data. (Total number of years tested = 20.) N/A = Not applicable (tests not performed). Main effects Area Tonnage class Depth All areas 2020 20/20 N/A Among stocks 19 20 N/A N'A Within stocks Georges Bank 16 20 20/20 10 20 So. New England 1 9, 20 20/20 7/20 Cape Cod 19/20 20/20 3/20 Interactions Tonnage class Tonnage class area depth All areas 20/20 N/A Among stocks 20/20 N/A Within stocks Georges Bank 8/20 4/20 So. New England 14/20 2/20 Cape Cod 7/20 1/20 100 O'BRIEN AND MAYO: CPUE OF YELLOWTAIL FLOUNDER highly significant differences in only 50, 35, and IdVc of the years for Georges Bank, Southern New England, and Cape Cod grounds, respectively (Table 4), while the tonnage class-depth interac- tion was highly significant in no more than 20% of the years on each of the three grounds. Interaction between tonnage class and statis- tical area throughout the region was highly sig- nificant in all cases. Further analyses yielded highly significant differences in CPUE among the three stocks and highly significant tonnage class- stock interactions which suggests that relative fishing power among vessel classes is not consis- tent from stock to stock, implying a need for com- puting a separate set of fishing power coefficients for each stock. Within each stock, differences in CPUE among statistical areas and tonnage classes were also highly significant in most cases, although the tonnage class-area interaction was not. Standardized CPUE Annual fishing power coefficients, obtained by retransforming linear model parameter esti- mates for each tonnage class, are presented in Table 5 by stock. Cursory examination of the coef- ficients reveals distinct trends throughout the 20- yr series. On Georges Bank and Southern New England grounds, fishing power coefficients for the smaller vessels (TC 23-25) relative to the standard declined over time, whereas coefficients for the larger vessels increased over time. On Cape Cod grounds the coefficients increased for the smaller vessels (TC 23 and 24) although trends were less pronounced. These trends are illustrated graphically by plotting annual devia- tions from the 20-yr average in Figures 7-9. A Durbin-Watson test for first order autocorrela- tion of the annual deviations (Neter and Wasser- man 1974) was significant for most tonnage classes within each of the stocks, suggesting the presence of a substantial tonnage class-time in- teraction. The three-way linear model, modified to in- clude interaction terms, also revealed highly sig- nificant tonnage class-year as well as tonnage class-season and year-season interactions within each of the three stocks (Table 6). When interac- tions are significant, they can be examined in detail or absorbed in the error term when testing for main effects. Since tonnage class effects have already been examined on an annual basis, the interaction terms were excluded from the three- way model used to obtain parameter estimates for tonnage class, season, and year. The model is specified as follows: f/ = a + X [Pi, Xy + ^2j X2j + Paj Xsj] + where ^y, P2/, P3; = model parameter estimates in logarithmic units for category j for tonnage class, season, and year, Xij , X2J, X2,j dummy variables for ton- nage class, season, and year ( = 1 when category 7 occurs; = otherwise). ANOVA results obtained without interaction are presented in Table 6 for each of the three stocks. For Georges Bank and Southern New England stocks, year accounts for the greatest reduction in error sums of squares; on Cape Cod grounds tonnage class accounts for the greatest overall reduction. Coefficients for tonnage class, year, and season, derived from model parameter estimates for the combined 1964-83 period are presented in Table 7. Tonnage class coefficients for Georges Bank and Southern New England are relatively homogeneous, as compared with those obtained for Cape Cod grounds, owing to the narrower range of vessel tonnage classes which have con- sistently exploited these fisheries. Seasonal coef- ficients exhibit the same pattern on Georges Bank and Southern New England with the highest catch rates occurring during the third quarter; on Cape Cod grounds the highest catch rates occur during the second quarter. Trends in annual coefficients are similar on all three grounds. Standardized CPUE indices based on the annual coefficients are illustrated in Figure 10, and traditional indices based on the methods of Lux (1964), as given by Clark et al. (1984), are also presented for comparative purposes. Although each series indicates similar trends, CPUE indices obtained from the linear model for Georges Bank and Southern New England have remained slightly higher than the traditional in- dices since the early 1970's. Prior to this, the tra- ditional CPUE indices were greater than the re- vised indices. On Cape Cod grounds, differences between the two series are considerably greater, particularly in the early years. 101 Georges Bank 1-l TC24 0.5- • -0.5- • c ■^^ o 65 67 89 71 73 75 77 79 81 83 4-» - ^1 TC25 > 0.5- • . £J 0- * , ' — -0.5- (0 Zj ■' 1 I i I I I 1 J- 65 87 89 71 73 75 1 1 I 1 77 79 81 83 C < 1- TC31 0.5- . -0.5- * ' 1 1 1 1 1 I 65 67 89 71 73 75 1 1 1 1 77 79 81 83 Table 5. KIS51EKY BULLETIN: VOL 86, NO 1 -Annual fishing power coeflicients calculated by vessel tonnage Georges Bank, Southern New England, Vessel tonnage class^ Year 21 22 23 24 25 31 32 33 41 42 Georges Bank 1964 073 083 0.97 0.91 0,90 043 1965 068 0.86 1.02 095 1,40 — 1966 0.45 0.81 0.99 099 097 — 1967 — 0.74 0.93 094 0,89 — 1968 1.09 084 098 1 08 088 0,75 1969 0.99 0.85 0.93 0.92 0,95 1.17 1970 1.21 0.78 0.94 0.91 094 0.92 1971 0.66 0.82 1.07 089 1,12 — 1972 — 082 0.94 096 0,93 — 1973 not 0.49 0.94 0.93 1.00 1 09 1,09 1.14 1974 calculated — 1 25 096 098 097 1.53 1975 1.00 0.99 0.85 1.06 1 04 1 66 1976 0.73 0.92 0.96 1.17 1,01 — 1977 — 0.66 0.94 1.28 1,67 — 1978 — 0.97 0.87 1,18 1,54 — 1979 0.55 0.76 1.02 1.19 1.35 — 1980 0.42 0.35 1.10 1.25 1.43 — 1981 0.45 0.50 0.80 1.05 1.16 — 1982 0.23 0.96 0.82 1.21 1.41 099 1983 — 0.40 1.25 1 12 1.26 — Southern New England 1964 - 20.95 — 0.97 0.92 0.88 0.91 0.81 1965 — 0.86 — 0.95 0.82 0.93 1 06 0.39 1966 — 1.12 — 1.14 0.98 1.02 092 1.18 1967 — 0.94 — 1.27 1.22 1.00 0.87 068 1968 — 1.21 — 1.23 1.11 0.94 1,01 098 1969 — 1.32 — 1.11 1.19 0.99 1.17 0.82 1970 — 1.23 — 1.04 1.09 085 0.79 0.74 1971 — 1.06 — 1.10 1 16 0.99 1.10 082 1972 — 0.99 — 088 1.12 1.01 1,03 0.66 1973 — 0.46 — 1.01 1.20 1 .03 1 .00 0,95 1.24 'Standard vessel class on Georges Bank and Southern New England rounds = 32. Standard vessel class on Cape Cod grounds = 25. 1- TC33 0.5- . 0.5- -1- [ I 1 1 1 1 1 1 I 1 0.5 66 67 69 71 73 75 77 79 81 83 TO 41 ~~i I I I I I — I — I — I — I 65 67 69 71 73 75 77 79 81 83 1 0.5 TC42 ~i I I I — I — I — I — I — I — I 66 67 69 71 73 75 77 79 81 83 Figure 7. — Deviations in annual fishing power from the 1964-83 20-yr mean for major vessel tonnage classes fish- ing on Georges Bank. Year 102 O'BRIEN AND MAYO: CPUE OF YELLOWTAIL FLOUNDER class relative to a standard class vessel fishing for yellowtail flounder on and Cape Cod grounds, 1964-83. Vessel tonnage class^ Year 21 22 23 24 25 31 32 33 41 42 Southern New England 1974 — 0.67 — 1.11 1 03 1 00 1.27 0.62 1975 — 072 — 0.73 0.83 085 1.00 1.11 1976 — 0.66 — 0.40 0.66 086 1.28 1.56 1977 — 087 — 0.79 0.97 088 1.09 0.99 1978 — 1 33 — 0.56 0.78 0.75 1.11 1.38 1979 — 069 — 0.54 0.74 079 098 1.24 1980 — 082 — 0.68 0.90 0.95 1.09 1.07 1981 — 0.92 — 0.55 099 1.06 1.26 0.86 1982 — 0.64 — 0.62 0.84 0.86 1.07 1.07 1983 — 0.68 — 092 0.94 0.89 0.98 1.04 Cape Cod 1964 — 062 0.52 0.73 1.08 1.04 — — 1965 — 067 0.55 0.73 1.65 0.98 0.96 — 1966 044 042 048 0.58 1.42 0.69 — — 1967 — 032 0.36 0.64 1.30 0.79 1.07 — 1968 — 044 0.37 0.71 1.55 0.91 0.99 0.43 1969 — 060 048 0.73 1.31 0.89 — — 1970 — 0.44 0.45 1.01 1.56 (3) (3) 1.25 1971 — 0.57 0.93 1.19 1.10 0.93 1.20 1.65 1972 0.93 088 063 1 09 1.09 0.79 0.97 1.05 1973 0.51 066 053 0.94 1.00 0.95 0.96 0.82 1.43 1974 0.40 0.55 0.43 0.76 083 0.87 0.90 0.66 1975 0.40 059 0.64 0.74 0.75 085 0.57 0.62 1976 18 068 072 088 0.73 0.80 0.97 1.12 1977 0.14 042 061 0.74 0.71 0.70 1.08 1.26 1978 — 040 075 0.90 0.93 0.73 0.91 2.84 1979 0.47 043 0.68 0.87 1.06 0.83 1.06 1.48 1980 12 027 0.60 0.80 1.03 095 1 01 1.53 1981 0.31 0.34 052 0.75 1.07 0.54 0.83 1.34 1982 0.27 035 060 088 1.19 0.77 0.96 1.17 1983 0.36 0.79 0.64 1.02 0.68 0.80 0.77 0.43 2Vessel classes 21, 22, and 23 combined on Southern New England grounds 3|nsuf1icient data for these categories. Southern New England c o 1- 05- TC 21-22-23 0.5- -1- 1 1 ; 1 \ 1 1 1 ^ 1 0.5- -0.5 65 67 69 71 73 75 77 79 81 83 TC31 — I 1 1 1 1 1 \ 1 1 I 65 67 69 71 73 75 77 79 81 83 TC33 <0 D C C < Figure 8. — Deviations in annual fishing power from the 1964-83 20-yr mean for major vessel tonnage classes fish- ing on Southern New England grounds. 1 0.5 H -0.5- -I — I — I — I — I — I — I — I r 65 67 69 71 73 75 77 79 81 83 TC25 — I 1 1 1 1 1 1 I I I 65 67 69 71 73 75 77 79 81 83 —I — 1 1 1 1 1 1 1 1 I 66 07 69 71 73 76 77 79 81 83 1- 0.5- TC41 0.5- -1- — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 65 67 69 71 73 75 77 79 81 83 Year 103 FISHERY BULLETIN VOL 86. NO 1 Cape Cod 1- 5 -0 5- TC23 — I — I — I — I — 1 — I — I — I — I — I 65 67 69 71 73 75 77 79 81 83 1 0.5 -0.5 -1 TC32 —1 — I — I — I — I — I — I — I — I — I 65 67 69 71 73 75 77 79 81 83 Figure 9.— Deviations in annual fishing power from the 1964-83 SO-yr mean for major vessel tonnage classes fish- ing on Cape Cod grounds. > Q • CD Q 0} a w c ,o Southern New England n ' I •—] I |— ' I I I — ■— I — I — I 1 I I — I 64 66 68 70 72 74 76 78 80 82 84 Year 4-1 ■D 0) JO. (f) IT >~ CD Q a3 Q. CO c ,o Cape Cod I ' I ' I ' I ' I ' I 64 66 68 70 72 74 76 78 80 82 84 Year Traditional CPUE index Linear Model CPUE Index Figure 10. — Trends in annual yellowtail flounder CPUE (metric tons per day fished) calculated with traditional CPUE based on Lux (1964) and annual coefficients (linear model CPUE). 104 O'BRIEN AND MAYO; CPUE OF YELLOWTAIL FLOUNDER Table 6. — ANOVA results obtained from a three-way linear model incorporating year, quarter, and tonnage class for Georges Bank, Southern New England, and Cape Cod stocks of yellowtail floun- der. Table 7. — Tonnage class, quarter, and year coefficients derived from a three-way linear model for the 1964-83 period for yellowtail flounder stocks on Georges Bank, Southern New England, and Cape Cod grounds. Sum of Mean F Georges Southern Source df squares Georges square Bank value P Tonnage Bank New England Cape Cod With interaction class Model 211 1.432 92 6.79 29.53 <0.01 21 0.92 0.32 Year 19 908.42 47.81 207.88 <0.01 22 0.92 0.50 Qtr 3 107.36 35.79 155.59 <0.01 23 0.92 0.57 TC 9 84.24 9.36 40.70 <0.01 24 057 0.93 0.80 Yr'Tc 103 116.81 1.13 4.93 <0.01 25 0.80 0.96 1.00 Yr*Qtr 57 202.29 3.55 15.43 <0.01 31 96 0.92 1.03 Qtr'Tc 20 13.80 0.69 3.00 <0.01 32 1.00 1.00 0.84 Error 18,437 4.171.68 023 33 1.08 1.09 1.00 Total 18,648 5.604.60 030 41 1.13 1.05 1.24 Without inte iraction 31 1.100.02 35.48 147.83 <0.01 42 1.19 — — Model Quarter Year 19 908.42 47.81 199.21 <'0.01 Qtr 3 107.36 35.79 149.11 <0.01 1 0.96 1.00 1.07 TC 9 84.24 936 39.00 <0.01 2 0.89 1.06 1.15 Error 18,617 4.504.58 0.24 3 4 1.09 1.00 1.18 1.00 0.90 1.00 Total 18,648 5 604 60 n .^n Southern New England Year With interaction 1964 1.00 1.00 1.00 Model 217 2,85995 13.18 33.79 <0.01 1965 0.73 0.72 0.84 Year 19 2,085.95 109.79 281.50 <0.01 1966 0.55 0.54 0.84 Qtr 3 111.50 37.17 95.30 <0.01 1967 0.54 0.60 0.81 TC 6 59.80 9.97 25.56 <0.01 1968 0.68 0.79 0.93 Yr*Tc 114 257.12 2.26 5.78 <0.01 1969 0.61 0.77 1.06 Yr-Qtr 57 304.89 535 13.72 <0.01 1970 0.58 0.75 1.00 Qtr-Tc 18 40.69 2.26 5.80 <0.01 1971 0.49 0.63 0.95 Error 26,879 10.612.21 0.39 1972 0.45 0.65 0.88 Total 27,096 13,472.16 0.50 1973 1974 0.55 0.42 0.49 0.45 0.84 0.77 Without interaction 1975 0.36 0.30 0.72 Model 28 2.257.25 80.62 196.62 <0.01 1976 0.37 0.25 0.76 Year 19 2,085.95 109.79 267.77 <0.01 1977 0.37 0.36 0.74 Qtr 3 111.50 37.17 90.66 <0.01 1978 0.34 0.39 0.83 TC 6 59.80 9.97 24.31 <0.01 1979 0.48 0.52 0.85 Error 27,068 11.214.91 0.41 1980 0.55 0.56 0.78 Total 27,096 13.472,16 0.50 1981 0.46 0.54 0.80 Cape Cod 1982 1983 0.44 0.40 0.71 0.70 0.77 0.78 With interaction Model 247 2.438.73 987 2468 <:0.01 Year 19 166.49 8.76 21.91 <0.01 Qtr TC 3 9 174.44 1,331.92 58.15 147.99 145.37 369.98 <0.01 <0.01 DISCUSSION Yr'Tc 135 526.18 3.90 9.74 <0.01 Yr'Qtr 57 167.62 2.94 7.35 <0.01 The analytical approach adopted in this paper Qtr'Tc 24 72.09 3.00 7.51 <0.01 is based on _ _ the hypothesis that CPUE of yellow- Error 19,097 7,731.24 0.40 tail flounder differs among the various tonnage Total 19.344 10.169.97 0.53 classes of vessels and ge _ _ sographic regions associ- Without interaction Model 31 1,672.84 53.96 122 64 <0.01 ated with the fishery. In all of the analyses, the Year 19 166.49 876 19.92 <0.01 null hypothesis (i.e., no 1 significant differences) Qtr TC 3 9 174.44 1,331.92 58.15 147.99 132.15 336.34 <0.01 <0.01 was rejected only when the probabil iity of obtain- Error 19.313 8.497.13 0.44 ing a greater F statistic was <0.01. Even at this Total 19.344 10.169.97 0.53 probability level, statistically signi ficant results were often obtained when difi'erences among vari- able levels appeared to be minimal due primarily to the large number of observations included in most analyses. The initial series of ANOVAs, based on pooled CPUE data from all statistical areas encompass- 105 FISHEKY BULLETIN; VOL. 86, NO. 1 ing the three stocks, provided sufficient evidence to reject the null hypothesis. In each of the 20 years analyzed, the main effects of tonnage class and statistical area represented highly signifi- cant (P < 0.01) sources of variation. The tonnage class-area interaction term was also highly sig- nificant in all cases, implying that vessels of var- ious tonnage classes exhibit different CPUE trends relative to each other in different areas. The initial results established the basis for fur- ther investigations. In subsequent analyses, the data were grouped to test the null hypothesis that no significant differences in CPUE existed among the three traditionally accepted stock definitions (Georges Bank, Southern New England, and Cape Cod). The highly significant differences ob- tained in 19 out of 20 years indicate that catch rates differ among the three stocks. The resulting highly significant tonnage class-stock area inter- action term obtained from the ANOVAs in all years suggests that standardization of CPUE among tonnage classes should be performed sepa- rately for each stock. Analysis of variance within each stock provided the final basis for performing the standardized CPUE calculations. In these tests, the rejection of the null hypothesis for tonnage class main effects in all years for each stock suggests that separate fishing power coefficients must be calculated for each tonnage class even though the coefficients are similar in many cases. The ANOVA results also indicated that differences in CPUE among statistical areas within each stock were highly significant in SO^f or more of the years implying that, within each stock region, yellowtail abun- dance is not homogeneous. This is not surprising since yellowtail flounder are prevalent only on certain grounds within each geographic region. Further analyses of the data by depth indicated no overall significant differences in CPUE be- tween the two primary depth zones (1-55 m and 56-110 m) where yellowtail flounder are consis- tently caught. The infrequent number of significant interac- tions on Georges Bank and Cape Cod grounds relative to Southern New England (Table 4) sug- gests a greater independence of the tonnage class and area main effects with respect to CPUE, i.e. both large and small vessels exhibited relatively similar changes in mean CPUE among statistical areas within each stock. In choosing data sets for computing fishing power coefficients we sought to minimize the amount of interaction among the vessel tonnage classes and geographic areas in- volved. This criterion was met to a greater extent for the Cape Cod and Georges Bank stocks than for the Southern New England stocks. It appears that yellowtail flounder inhabiting this region are subject to a more complex set of interactions perhaps due to temperature and bottom type. We decided, however, to accept the results for each of the three stocks and proceed with the calculations of fishing power coefficients. Annual fishing power coefficients computed for each vessel tonnage class fishing on Georges Bank, Southern New England, and Cape Cod grounds provided a basis for examining the con- sistency in relative fishing power of individual tonnage classes over time. Annual deviations for Georges Bank and Southern New England grounds indicated a gradual change in relative fishing power of most tonnage classes between 1964 and 1983, and tests for autocorrelation of residuals indicated significant time effects. On these grounds, larger vessels exhibited higher catch rates relative to the standard in the later years as compared with the earlier years. Since many of the larger vessels have been replaced in recent years by newer vessels which are, pre- sumably, equipped with more sophisticated elec- tronics, any attempt to relate CPUE to stock abundance must account for such technological advances. Similarly, changes in seasonal availability are often great enough to mask interannual variation in stock abundance. Thus, the presence of signifi- cant tonnage class-season interactions may be ex- plained by the ability of certain vessel classes to effectively target seasonal concentrations. Since peak spawning of yellowtail flounder occurs dur- ing late spring (Lux 1964), the presence of high seasonal coefficients during the second and third quarters is not surprising. By specifying the model to include tonnage class, annual, and seasonal components, we have attempted to account for technological and sea- sonal availability factors which interact with temporal changes in abundance. Although other factors could be incorporated in the model to ac- count for a larger portion of the variation in CPUE, analyses of historical commercial fishing operations of this type are often limited to those attributes which can be directly linked to land- ings records (Kimura 1981; Westrheim and Foucher 1985). An alternate approach adopted by Stern and Hennemuth (1975) involved the use of a study fleet of selected vessels whose characteris- tics and fishing practices were closely monitored. 106 O'BRIEN AND MAYO CPUE OF YELLOWTAIL FLOUNDER In our study, factors were selected for inclusion in the model based on prior knowledge of fleet characteristics and seasonal and spatial distribu- tion patterns of the species. Despite this, the three attributes incorporated in the final model accounted for 15-25% of the total variation in CPUE, depending on the stock. Undoubtedly, other factors such as experience of the captain, net design and rigging, and variation in local fish abundance contribute substantially to overall variation in catch rates. Differences between the annual CPUE esti- mates based on Lux's original fishing power coef- ficients and the recalculated indices occur in many cases because of shifts in the vessel compo- sition of the fleet over the past 20 years. The in- clusion of larger vessels in the more recent years, particularly on Georges Bank and Southern New England grounds, may account for the consis- tently higher CPUE estimates obtained for these areas since the mid-1970's. On Cape Cod grounds, CPUE estimates differ substantially prior to this time. Lux (1964) has stated that a relatively low proportion of the landings from this area were used in his CPUE computations and, conse- quently, the indices were not considered to be as valid a measure of relative abundance as those obtained for Georges Bank and Southern New England. Our analyses for Cape Cod grounds, based on data for the period since 1964, are sub- ject to the same concerns since a large proportion of the yellowtail flounder landings continues to be taken incidentally. Although the revised standardized CPUE esti- mates presented in this paper are based on a dif- ferent standardization technique, trends are gen- erally similar to those obtained previously. The revised procedure, however, accounts for seasonal and technological influences and insures com- plete representation of all vessel classes engaged in the yellowtail fishery. ACKNOWLEDGMENTS We wish to sincerely thank Stephen H. Clark for his advice throughout this study, and for crit- ically reviewing the manuscript. Michael J. Fo- garty reviewed the final draft and advised on statistical procedures. We are also grateful for the suggestions provided by two anonymous referees. LITERATURE CITED Beverton, R J H . AND S J Holt 1957. On the dynamics of exploited fish popula- tions. Fish. Invest., Lond., Ser. 2, 19:1-533. Bradu, D. andY Mundlak 1970. Estimation in lognormal linear models. J. Am. Stat. Assoc. 65(329):198-211. Brown. B E . M P Slssenwine. and M M McBride. 1980. Implication of yellowtail flounder stock assessment information for management strategies. U.S. Dep. Commer., NMFS, NEFC, Woods Hole Lab. Ref Doc. No. 80-21, 12 p. Clark, S. H., M. M. McBride, and B Wells. 1984. Yellowtail flounder assessment update - 1984. U.S. Dep. Commer., NMFS, NEFC, Woods Hole Lab. Ref. Doc. No. 84-39, 30 p, Dixon, W J (editor), 1981. BMDP statistical software, Univ. Calif. Press, Berkeley, p. 388-412. Garrod. D J 1964. Effective fishing effort and the catchability coeffi- cient q. Rapp. Cons. int. Explor. Mer 155, No. 14, p. 66- 70. Gavaris. S 1980. Use of a multiplicative model to estimate catch rate and effort from commercial data. Can. J. Fish. Aquat. Sci. 37:2272-2275. Gulland. J A 1956. On the fishing effort in English demersal fish- eries. Fish. Invest., Lond., Ser. 2, 20(51:1-41. 1964. Catch per unit effort as a measure of abundance. Rapp. Cons. int. Explor. Mer 155, No. 1, p. 8-14. KiMURA. D K 1981. Standardized measures of relative abundance based on modelling log (c. p. u.e.), and their application to Pacific ocean perch (Sebastes alutus). J. Cons. int. Explor. Mer 39:211-218. Lux. F E 1963. Identification of New England yellowtail flounder groups. Fish. Bull, U.S. 63:1-10. 1964. Landings, fishing effort, and apparent abundance in the yellowtail flounder fishery. Int. Comm. North- west Atl. Fish. Res. Bull. No. 1, p. 5-21. NETER. J . AND W WaSSERMAN 1974. Applied Linear Statistical Models. Richard W. Irwin, Inc., Homewood, IL, 843 p. Pope. J A . and B B Parrish 1964. The importance of fishing power studies in abun- dance estimation. Rapp. Cons. int. Explor. Mer 155, No. 17, p. 81-89. Pope, J G . and D J Garrod 1975. Sources of error in catch and effort regulations with particular reference to variations in the catchability coef- ficient. Int. Comm. Northwest Atl. Fish. Res. Bull. 11, p. 17-30. Robson, D S. 1966. Estimation of the relative fishing power of individ- ual ships. Int. Comm. Northwest Atl. Fish. Res. Bull. 3, p. 5-14. Rounsefell. G a 1957. A method of estimating abundance of groundfish on Georges Bank. Fish. Bull., U.S. 113:264-278. RoYCE, W. F., R J Buller, E. D. Premetz 1959. Decline of the yellowtail flounder (Limanda fer- ruginea) off New England. Fish. Bull., U.S. 59:169- 267. SAS Institute 1982. SAS User's Guide: Statistics. 1982 ed. SAS Insti- tute Inc., Gary, NC. 107 FISHERY BULLETIN: VOL. 86, NO. 1 SissENWiNE. M P Stern. H , ,Jk , and R C Hknnemuth 1974. Variability in recruitment and equilibrium catch of 1975. A two-way model for estimating standardized fish- the Southern New England yellowtail flounder fish- ing crfort applied to the U.S. haddock fleet. Rapp. Cons. ery. J. Cons. int. Explor. Mer 36:15-26. int. Explor. Mer 168:44-49. 1978. Is MSY an adequate foundation for optimum Westrheim, S J , AND R. P. FOUCHER. yield? Fisheries 3(6):22-42. 1985. Relative fishing power for Canadian trawlers land- Steel. R G D, AND J H TORRIE. ing Pacific cod iCiadus macrocephalus) and important 1980. Principles and procedures of statistics. McGraw- shelf cohabitants from major offshore areas of western Hill, N.Y., 633 p. Canada, 1960-1981. Can. J. Fish. Aquat. Sci. 42:1614- 1626. 108 REDUCING THE BYCATCH IN A COMMERCIAL TROTLINE FISHERY Lawrence W McEachron,' Jeff F Doerzbacher,- Gary C Matlock,^ Albert W. Green,^ and Gary E, Saul- ABSTRACT Reducing the bycatch of red drum, Sciaenops ocellatus, and spotted seatrout, Cynoscion nebulosus, in the Texas commercial trotline fishery is desirable. Hook placement within the water column was examined as a means of accomplishing this objective. The commercial trotline fishery was simulated in the Laguna Madre during February 1985 through January 1986. Requiring placement of trotline hooks on bottom will reduce bycatch of red drum, spotted seatrout, and other nonmarketable fishes and improve operational efficiency of commercial fishermen without significantly reducing catch of black drum, Pogonias cromis, a target commercial species. Other than crab and shrimp being more effective baits than oleander leaves, no other generalization could be made concerning baits and seasons. Longlines catch species unwanted or legally non- retainable by fishermen and have been regulated to reduce the bycatch of nontargeted species (South Atlantic Fishery Management Council 1985). Trotlines (Figs. 1, 2) are a specialized long- line used in shallow (<4 m) Texas estuaries to catch fish (Simmons and Breuer 1962; Breuer 1973, 1974, 1975; Matlock 1980). Red drum, Sciaenops ocellatus, and spotted seatrout, Cynoscion nebulosus, were the primary targets until 1981 when their sale was prohibited be- cause of overfishing (Matlock et al. 1979; Anony- mous 1979, 1981, 1983). The effort has since been redirected toward black drum, Pogonias cromis. Regulations requiring the use of circle hooks and placement of the mainline under water were en- acted to reduce the bycatch of red drum and spot- ted seatrout. However, a bycatch still occurs. This study was conducted to determine if the bycatch could be further reduced by additional regulation of where in the water column hooks are fished and bait types. MATERIALS AND METHODS The catch on trotlines with hooks placed on the bottom or in the top of the water column was compared by simulating commercial fishing tech- niques in the Laguna Madre, TX (Fig. 3). Bottom trotlines were set with the mainline on the bot- tom. Top trotlines had the mainline floated with iTexas Parks and Wildlife Department, P.O. Box 1717, Rock- port, TX 78382. 2Texas Parks and Wildlife Department, 4200 Smith School Road, Austin, TX 78744. the hooks suspended in water >0.6 m deep to insure hooks fished in the water column. Texas Parks and Wildlife Department (TPWD) trotlines were set in the same area as commercial trotlines. Commercial fishermen were contacted by tele- phone within 24 hours prior to TPWD sets to de- termine areas of commercial activity. All TPWD trotlines were at least 15 m apart. Trotlines with 100 hooks each were built ac- cording to commercial fishermen specifications (McEachron et al. 1985). The mainline (182.9 m long) consisted of #36 nylon twine, knotted twice every 1.8 m for swivel (1/0 black brass barrel) placement (Figs. 1, 2). Hooks (#8 Mustad 39960ST) were attached by a 610-686 mm long staging (56.7 kg test monofiliament) to the swivel at 1.8 m intervals. Stakes (51 cm x 76 mm) and/or anchors were placed on each end to stretch the mainline. Floats (3.8 L) were attached to the mainline every 15 hooks for navigation identifi- cation. Eighteen trotlines were set overnight each month in both the upper and lower Laguna Madre during 1 February 1985 through 31 January 1986. Six (3 top; 3 bottom) were set during each of two monthly sampling periods (first and last 15 days of the month). Another six sets were made in either the first or last half of each month; the period was randomly selected each month. Each trotline was baited completely with one of three bait types — cut portions of blue crab, Callinectes sapidus; dead shrimp, Penaeus sp.; or oleander, Nerium sp. leaves — so that all bait types were used on both top and bottom trotlines during every period. These baits represented the most Manuscript accepted October 1987. FISHERY BULLETIN; VOL. 86. NO 1, 1987. 109 <5-* 5.1 cm X 7.6 cm WOODEN STAKES Figure 1.— Top trotline. 1 cm X 7.6 cm WOODEN STAKES Figure 2. — Bottom trotline. commonly used baits by commercial trotline fish- ermen (McEachron et al. 1980, 1986). Fishes caught were identified (Hoese and Moore 1977; Robins et al. 1980), counted (Table 1), and total length (TL) was measured to the nearest 1 mm. Data were pooled into fall (Septem- ber-November), winter (December-February), spring (March-May), and summer (June-Au- gust) to examine seasonal variation. A catch rate (No. /line • h) for black drum; red drum; spotted seatrout; hardhead catfish Arius felis ; and total fishes was computed for each trot- line set by dividing the number caught by the number of hours fished. Catch rates were trans- formed to log (catch rate + 1) and analyzed using a four-factor fixed-effects model analysis of vari- ance (AOV). The four factors were 1) hook place- ment, at two levels — top and bottom; 2) bait, at three levels — crab, shimp, and leaves; 3) bay, at two levels — upper Laguna Madre and lower La- guna Madre; 4) season, at four levels — fall, win- ter, spring, and summer. Diff'erences in main effect means were evalu- ated with Ducan's multiple range test. However, when significant first-order interactions were found, comparisons were made within levels of the interacting factors using the mean square error (MSE) from the AOV. Total lengths of each species were analyzed in a nested AOV to investigate differences among the four factors. However, because fish were not caught in all factor level combinations, factors and/or factor levels for each species were elimi- nated from analyses. Spotted seatrout lengths 110 McEACHRON ET. AL: BYCATCH IN TROTLINE FISHERY Figure 3.— Texas coast. Table 1 —Number of fishes caught on trotlines in the upper and lower Laguna Madre during February 1985-January 1986. Upper Lower Laguna Laguna Species Madre Madre Total Arlus felis 977 1,652 2,629 Sciaenops ocellatus 352 658 1,010 Pogonias cromis 67 265 332 Cynoscion nebulosus 29 103 132 Micropogonias undulatus 36 51 87 Opsanus beta 34 1 35 Archosargus probatocephalus 1 24 25 Dasyatis americana 17 17 Dasyatis sabina 6 7 13 Elops saurus 4 4 8 Orthopristis chrysoptera 1 7 8 Bagre marinus 1 4 5 Lagodon rhomboides 5 5 Paralichthys lethostigma 1 3 4 Rhinoptera bonasus 4 4 Chilomycterus schoepfi 3 3 Ophichthus gomesi 3 3 Cynoscion arenanus 2 2 Negapnon brevirostris 1 1 Trachinotus carolinus 1 1 All species 1.512 2,812 4,324 were pooled for both bay systems because an in- sufficient number of spotted seatrout were caught for individual bay analyses. Factor levels elimi- nated from length analyses were leaves and win- ter from hardhead catfish, leaves and crab from spotted seatrout, and fall, spring, and summer from black drum. Spring and winter red drum lengths were pooled. Lower Laguna Madre data only were used for red drum, black drum, and hardhead catfish length analyses. Each measured fish length was an observational unit of a trotline set. Sets were a random factor nested within fixed main effect combinations. The nested set effect mean square was used for testing other effects when the set effect was significant. However, the AOV yields approximate F values because un- equal numbers of fish were caught among sets. SAS procedures (SAS Institute, Inc. 1980, 1982) were used for all analyses. The significance level for each AOV test was a = 0.01 because the AOV used to examine catch rates of each species had 15 potential F tests. This alpha value assured that 111 KKSHEKY BULLETIN: VOL. 86, NO. 1 the family level of significance would not exceed 0.15. All other tests were made with a signifi- cance level of u = 0.05. Mean catch rates and con- fidence intervals computed from transformed data were back-transformed for tabular and graphic presentation (Elliott 1979). RESULTS Fishing trotlines on the bottom reduces bycatch without affecting catches of black drum, the target species. A significant difference could not be detected between black drum catch rates on top and bottom trotlines regardless of bait, sea- son, or bay (Tables 2, 3). Catch rates for hardhead catfish, red drum, spotted seatrout and total fishes were significantly lower on bottom trot- lines than on top trotlines (Tables 2, 3). Differ- ences in catch rates between top and bottom trot- lines for hardhead catfish, red drum, and total fishes did not vary significantly among seasons and baits but did vary between bays based on the first-order interactions (Tables 4, 5). A significant second-order interaction of position x bay x bait for spotted seatrout revealed that differences be- tween top and bottom trotlines were affected by both bait and bay but not by season (Fig. 4). No significant difference was found in red drum catch rates among baits nor in spotted seatrout catch rates among seasons (Tables 2, 3). All other main effects were significant for catch rates of all species and total fishes. Of the possible first-order interactions involving bait, season, and bay, only season x bait for hardhead catfish, red drum, and total fishes, bait x bay for black drum and total fishes, and season x bay for black drum, hard- head catfish, and total fishes were significant (Ta- bles 4, 5). The second-order interaction of bait X bay x season for total fishes was signifi- cant (Fig. 5). No significant differences were found in mean lengths of black drum, hardhead catfish, red drum, and spotted seatrout between top and bot- tom trotlines (Tables 6-8). Significant differences in mean length of hardhead catfish were detected for main effects of bait and season (Table 7). DISCUSSION Management objectives could be better met by requiring placement of trotline hooks on bottom than by allowing hooks to be fished from the sur- face. Red drum and spotted seatrout mortality would be reduced without significantly affecting TS r ro CL o c- o 4.^ (0 m £ -) QJ (0 o C r OJ If) o i^ c Q) nj •o r O O re ^o o^ LO _) 05 () CO <0 0) I/I O ■n OJ (/) c (U (5 3 Q. Ifl (0 > E 55 2 •o Q. CL 3 n (1) "O c . (0 r . o (0 3 ,->" c (0 —i ■D in (» « ^— o o £:- m (TJ (/I 3 (0 JO u 01 t_ LL m Ol .c C 11 3 c ■D — C o 2 o CO CO ~ — 0) (11 en (fl ■D *- c r (D o to (J (0 J2 o 01 *-' 1- (0 JD o >. o i JD (11 _j (11 < 1- 0) E E 3 C/5 Q. C/) (0 LL 0) i_ (0 Q) C S CD (0 S (0-5 01 ^ (tl Q. 3 T3 0.01(0 ^(0 3 (/) (0 6 E o o CD en o (D Q. CD (JD t~- CM CO O) CO in 0° (£) CM CO •^ ■^ '- 0^ T- CO cp CM cp CD o^ d t^ CO -* 1^ d in in d CD CM CO d d rg ~ ? in — in in 00 , — T— t^ ° CO to CM CD r^® CO in cp CD cp r^ ^ CO ,_ tT '^co d c\j d ■t d in 5 in d d d d r? in CO CM £" LO CJ) CO ■* t^ CO "" CO OM cr> CO CO t-" r- (D -"-cfi d in d in d CM d TJ- CO CD d d d d ? en CD ^° 00 CO in "" CO coP ■ in r-~ d d d d CO ST CM G^ CM in ■* CD r~- 0)0 1^ CM r- in a)P CO cp '^ d d •.- °in CM in .,^ yL Pco eg d CO d CO d CD CM d d d d d <5" CD CO — in (D CO in ■^ ■^ 0)° (D in O) CM .^ OlP '^ CO CD CO cp CO °co '«■ CM r^ CM ^O) CO d , — d co d CM CO ■ m C (0 -p 00 CD k_ J3 0) c 3 "D (0 TO c 3 O) TO _l (/5 CO E o o n c TO a. o c o QJ TO sz o TO O "CJ 0) E c TO c TO £ > O < TO ^_ 3 O 3 CD .< O q; O 0) ii ., en o Sf E TO 3 3 C/5 a: o c a> o o TO 3 O TO C/5 > ^i;r;i:'5'~ir'~'~'~oowcD'-r-- ooooininoooocDoooo-^ OOCDOOOCDOcioCDOOOcb »— OCDTj-cooOCXJCOOCDCMCDlfl'l-CO r--coa)inior-.cococDLnoT}--t-coo> C\J lO CO o o r^'^Lrii-ocbcNJcNJo CO -^ 00 OCVJinCMCDCD'— COO'J-CDO'3''-T-r— CO cocO'-cDTrojcocDininocnocococo o ocDLnr-'.-coc\iooa)ooocoa5cot-^r-^ CO h^'-OJCOOOini-Tj-^OOOCMOCO oj "-cvJocDdooocbcbcbczicbocbcri ■^'-'— '— lOt-i-t-i— ,-^C\JCDCOCO oooocjTj-ocoooair-^-'-Ttoo cbc3cbc)C)cbocboc3ooodcD V V V V V V V cococD''-a5- o o ■'-maioo-^'j-oo-^Ln cDcbcbcDcJcbcbcbcodcDcbcridd COOO'-S-CDt^OOCDCMCOOfOitCD OOCDi-CpOOr^CO-^COOOCOCOOOCOCD Lric\icMihc\ic>cb->-T-cbuSc3CDdc) CM -^ •>- CM CD tncOOOOCDOOOCMO-r-OOO-^CDCO-r- co ■^■^CMr^cM-^'^'-in-^LnO'-ocM'- 00 ■^'-oooo-'-oooooooor-- CM ooooooooooooooo--- d cDdcbcDdcDdcDcbcbcbcbcDdcbd '-oo-^T-cDCM-.-co-'-r^ococDLnco p-r-oo'^oooooococD-'j-or- CDOcDcricJdcocricDCDebcbcDCDcb CM''--^cocMa)t~^rgcDcoooocMO i-r--.T3-cpt^T}-r~.-- cDi^CM''--'-'-'.-0'>^-.- dcd"^dcbcb<3r^'r^cricDC3C3c\id ^ CM T- cnCMtDOLnCOCOinOTOOTCOCO^COCM '-inoocoooocot^ooo'— ocD'-'- OCMt~^C0OOOC\|T-LnOOOCMOO OOOOOOOOOOOOOOOh^ cbdcDocbcDdcbdcricDcbddcbd '-CMCO'-CMCO-'-CMCOCOCMCOCOCOCO'V CO oo c o (/) (1) o § ~ 00 "' -5 CO tr TO >. in O TO Q> TO O Q- CD en m Q. c o uo - TO Q) TO E n U) ^ X X X 00 c c c o o o oo E Q) 00 >, O) >. TO XI c o TO 0) en c o U) TO 0) E . CD TO If) E to >. en >. TO E B oo >, TO <^ B B 5 ^ ^ ,_ TO TO TO O CD m CD Q. O TO CL CD O ^ g A JO TO c 'c O) C3) (/> CO c .^^ o >. c TO Q) ^ TO X O) •^^ c TO J3 T3 TO •o OJ C ^ TO (0 >, Q) TO CJ n 0) X Q. 05 c TO o '^ OJ CO ■D O C Q. 3 O g O B TO TO u. C ^ O CD 15 C o Q) ■o (U O E E TO o 03 In 0) c £ TO >. ^ C TO ■D a> a> 2 g o 1 ^ ^ 03 LLI C —i CO < TO 0) 03 < CD a> > CO TO o CO 0) o o _l d d Q. < CD E CO CM c CO in n o o en o o < < CM CO ■•- CO O -r- CD CD ■* O •- CD 00 CD (U QJ ■O JJ TO TO E 22 03 TO TO 03 C C >- 3 3 0) O) 03 TO TO TO _l _l CD k_ k_ 03 03 1=1- g Q. o D -J 03 C TO 0) 2 03 CL Q. 3 II 03 c3 => TO 03 -.^ ? i g^ c o TO • O ^ CL C ^-' 03 >, .£ TO >..y TO ^ X -^ ^ 03 C o 03 S 2? 05 TO T3 C TO 03 C T3 TO 03 03 03 O 03 - TO I -Q 2 ^"S TO - I § m o o3 2 P "1 x: 03 2 ^ _l TO _J O "O TO 03 x: "O 2 _l TO I D c o 03 TO 03 cn < CD m < CO Tj- r^ CD 03 tT ■.- tT OJ -"t CD 00 d d d d < < < < CD ■>- CO in CM O •* O CO O -^ CO c:i d d> d < CD < < o in in CD 00 O CO ■•- o c\j o o cs d d d < < < < CO oo CO o O CM CO O o o o o d d d d < CD < < h- 00 r-- CD ■^ 00 t^ o o ■.- o 1- d d d d < CD < < oo ■^ in TT CD CD CO •* O TJ- o o d d d d < < < < O TJ- CD O o CM r~. o ■•- C\J o ••- d d d d O < O O CM oo CO in •,-■,- r-- ■* CM T- T- o d d d d CD < CD CD O O ■* -r- "- o ■* o CO CM ■^ CM d d d d < < < < ■* CD CD r^ •■- CM o in CD CO O 1^ d d y-^ d 03 03 w g'E = -E i E TO •> CL 3 LL S CO Cn 113 0.12 0.10- O.OB E 3 0.06 < 0.04 a. I u < 0.02 o 0.00- — I — ULM FISHERY BULLETIN; VOL. 86, NO. 1 SHRIMP TOP CRAB LEAVES TOP TOP BOTTOM BOTTOM — I LLM — I ULM — I LLM — I ULM — I LLM BAY SYSTEM Figure 4.— Significant second-order interaction among positions, baits, and bay system for spotted seatrout catch rates. LLM = lower Laguna Madre, ULM = upper Laguna Madre. Figure 5. — Significant second-order interac- tion among bait types, bay system, and sea- sons for total fishes catch rates. LLM = lower Laguna Madre, ULM = upper Laguna Madre. - 1.5 \ • 1.2 E I 0.9 < K I O I- < u uj 0.6 • 0.3 0.0 ,. l.B 1.5 1.2 0.9 0.6 0.3 0.0- FALL CRAB SHRIMP LLM ULM LEAVES LLM ULM I I WINTER SPRINS SEASON SUMMER 114 McEACHRON ET. AL: BYCATCH IN TROTLINE FISHERY Table 6. — Mean length (mm t 1SE) of black drum, hardhead catfish, red drum, and spotted seatrout caught on top and bottom trotlines by bait, bay and season during February 1985- January 1986 Number in parentheses is number of fish measured. Position Bait Species Top Bottom Crab Shrimp Leaves Black drum 539 ± 10 550 ± 13 566 ± 1 1 493 ± 12 588 i 21 (177) (146) (172) (108) (43) Hardhead catfish 338 ± 1 343 ± 2 346 ± 1 328 ± 2 337 ± 3 (1.363) (675) (1,201) (651) (186) Red drum 521 ± 4 504 ±9 540 ± 6 481 ±6 542 ± 7 (657) (173) (260) (336) (234) Spotted seatrout 439 ± 9 420 i 24 471 ± 23 426 ± 10 445 ± 19 (111) (18) (17) (86) (26) Bay Season Upper Lower Laguna Laguna Species fvladre (vladre Fall Winter Spnng Summer Black drum 484 ± 19 560 ±8 464 ± 20 538 ±8 630 ± 21 416 ± 19 (67) (256) (35) (201) (71) (16) Hardhead catfish 327 ± 1 348 ± 1 329 ±2 368 ±2 333 ±2 345 ±2 (824) (1,214) (538) (277) (724) (499) Red drum 506 ± 7 523 ± 5 493 ± 1 1 506 ± 5 546 ±9 551 ± 10 (308) (522) (133) (427) (114) (146) Spotted seatrout 411 ± 15 443 ± 10 410 ±21 404 ± 1 2 482 ± 25 478 ± 13 (29) (100) (25) (51) (16) (37) Table 7. — Summary of results of the AOV s of mean length for black drum (winter season only), hardhead catfish (excludes leaves and winter season), and red drum (winter and spring seasons combined) on top and bottom trotlines in lower Laguna Madre during February 1985-January 1986. NA = not analyzed. Black drum Hardhead catfish Red drum Source of Sum of Sum of Sum of variation df squares F PR > F df squares F PR>F df squares F PR>F Total 176 1,941,401 1,050 2,310,898 521 6,484,793 Position 1 5,978 0.15 0.70 1 80 0.02 0.90 1 6,122 0.26 0.61 Bait 2 15,082 0.19 083 1 33,650 6.60 0.01 2 89,151 1.90 0.15 Season NA 3 89,202 5.83 <0.01 2 133,190 2.85 0.06 Position X bait 2 49,063 0.69 0.55 1 8,242 1.62 0.21 2 10,913 0.23 0.79 Season ^ position NA 3 1,517 0.10 0.96 2 789 0.02 0.98 Season x bait NA 3 6,132 0.40 0.75 4 68,535 0.73 0.57 Season '; position x bait NA 3 10,269 0.67 0.57 4 63,388 0.68 0.61 Set (position x bait) 27 10,906,616 8.26 -O.OI NA NA Set (season '- position bait) NA 89 453,750 3.22 <0.01 91 2,129,491 2.61 <0.01 Error 144 704,611 946 1,499,474 413 3,706,235 Table 8. — Summary of results of the AOV of mean length for spot- ted seatrout (shnmp bait only) on top and bottom trotlines in upper and lower Laguna Madre combined dunng February 1985- January 1986. Source of variation df Sum of squares PR>F Total 85 750,362 Position 1 601 0.04 0.84 Season 3 58,796 1.32 0.29 Position X season 3 58,462 1.31 0.29 Set (position ^ season) 28 416,855 3.42 <0.01 Error 50 217,744 115 FISHERY BULLETIN: VOL 86, NO 1 black drum catches. Operational efficiency of commercial fishermen should improve with less handling of nontarget species. Mortality of non- target fishes would decrease because they would not be caught and subsequently handled. For red drum and spotted seatrout that are caught, sur- vival would be high for those released back into the water. Survival of released red drum caught on trotlines in winter and summer and of spotted seatrout in winter was 100^^ (Martin et al. 1987). About 509^ of the spotted seatrout died in summer cage studies; but few commercial trotlines are fished during this period (TPWD unpubl. data). Thus, the goal of reducing the catch of nontarget species and reducing mortality due to trotlines can be achieved with minimal impact on the com- mercial fishermen. Interactions between bay system and the other three factors for some species probably reflect dif- ferences in relative abundance. Fewer fish were available to be caught in upper Laguna Madre than in lower Laguna Madre (Crowe et al. 1986). The effects of bait and season on trotline catches cannot be determined in bay systems where the fish abundance approaches zero. No spotted seatrout were caught on crab bait on bottom in either bay; but they were caught on all baits on top in the lower Laguna Madre leading to the significant second-order interaction of posi- tion X bay X bait. This condition was not unex- pected because spotted seatrout are predomi- nately sight feeders (Vetter 1977), and might not take baits on bottom as readily as baits suspended in the water column. Crab and shrimp were more effective baits than oleander leaves for all four species. No other gen- eralizations could be made concerning baits and seasons. Selection of crab or shrimp as the bait of choice for reducing bycatch while maximizing black drum catch is unclear because catch rates for black drum and red drum were greater on crab than shrimp, especially in winter. ACKNOWLEDGMENTS We would like to thank all Laguna Madre field personnel who diligently collected the samples. Tom Heffernan, Ed Hegen, Lynn Benefield, Maury Ferguson, and Tony Maciorowski re- viewed the manuscript. LITERATURE CITED Anonymous 1979. Saltwater finfish research and management in Texas. A report to the Governor and the 66th Legislature. Tex. Parks Wildl. Dep., Coastal Fish. Branch, PWD Rep. No. .3000-59, 21 p. 1981. Saltwater finfish research and management in Texas. A report to the Governor and the 67th LegLslature. Tex. Parks Wildl. Dep., Coastal Fi.sh. Branch, PWD Rep. No. ;J000-10K, ,31 p. 1983. Saltwater finfish research and management in Texas. A report to the Governor and the 68th Legislature. Tex. Parks Wildl. Dep., Coastal Fish. Branch, PWD Rep. No. 3000-154, 48 p. Breuer, J P 1973. A survey of the juvenile and adult food and game fish of the Laguna Madre. Tex. Parks Wildl. Dep., Coastal Fish. Branch, Proj. Rep. 173-202. 1974. Juvenile and adult food and game fish of the Laguna Madre. Tex. Parks Wildl. Dep., Coastal Fish. Branch, Proj. Rep. 109-130. 1975. Biological studies in the lower Laguna Madre of Texas, 1975. Tex. Parks Wildl. Dep., Coastal Fish. Branch, Proj. Rep. 158-196. Crowe, A L , L W McEachron, and P C Hammerschmidt. 1986. Trends in relative abundance and size of selected finfish in Texas bays: November 1975-December 1985. Tex. Parks Wildl. Dep., Coastal Fish. Branch, Manage. Data Ser. No. 114, 259 p. Elliott. J. M 1979. Some methods for the statistical analysis of samples of benthic invertebrates. Freshwater Biol. Assoc, Sci. Publ. No. 25, 160 p. HoESE. H D , AND R. H Moore 1977. Fishes of the Gulf of Mexico, Texas, Louisiana, and adjacent waters. Texas A&M Univ. Press, College Station, 327 p. Martin. J H , K W Rice, and L W McEachron 1987. Survival of three fishes caught on trotlines. Tex. Parks Wildl. Dep., Coastal Fish. Branch, Manage. Data Ser. No. Ill, 21 p. Matlock, G. C 1980. History and management of the red drum fishery. In Proceedings Colloquium on red drum and sea- trout, p. 37-54. Gulf States Mar. Fish. Comm. No. 5. Matlock, G C , P L Johansen, and J P Breuer 1979. Management of red drum in a Texas estuary - a case study. Proc. Annu. Conf Southeastern Assoc. Fish Wildl. Agencies 33:442-450. McEachron, L W , A W Green, G. C Matlock, and G E. Saul. 1985. A comparison of trotline catches on two hook types in the Laguna Madre. Tex. Parks Wildl. Dep., Coastal Fish. Branch, Manage. Data Ser. No. 86, 44 p. 1986. Evaluation of the commercial trotline fishery in the Laguna Madre during fall 1984. Tex. Parks Wildl. Dep., Coastal Fish. Branch, Manage. Data Ser. No. 93, 25 p. McEachron. L W . G C Matlock. A R Martinez, and J P. Breuer 1980. Evaluation of natural, leaf, vegetable, worm and cork baits used on trotlines in upper and lower Laguna Madre, Texas (September 1977-October 1978). Tex. Parks Wildl. Dep., Coastal Fish. Branch, Manage. Data Ser. No. 8, 68 p. Robins, C R , R M Bailey, C E Bond. J R Brooker, E A Lachner, R N Lea, and W. B Scott. 1980. A list of common and scientific names of fishes from the United States and Canada. (4th ed. ) Am. 116 McEACHRON ET AL: BYC'AIVH IN TROTLINE FISHERY Fish. Soc. Spec. Pub. No. 12, 174 p. SOUTH ATLANTIC FISHERY MANAGEMENT COUNCIL. SAS Institute. iNC 1985. Source document for the swordfish fishery 1980. SAS supplemental library user's fjuide. SAS management plan. South Atl. Fish. Manage. Counc, Institute Inc. Cary, NC, 202 p. Charleston, SC. 1982. SAS users guide: Statistics. SAS Institute Inc. Vetter, R D Cary, NC, 584 p. 1977. Respiratory metabolism of, and niche separation Simmons, E G . and J P Breuer between two co-occurring congeneric species, Cynoscion 1962. A study of redfish, Sciaenops ocellata Linnaeus nebulosus and Cynoscion arenarius in a south Texas and black drum, Pogonias cromis Linnaeus. Publ. estuary. MA Thesis, Univ. Tex, 113 p. Inst. Mar. Sci., Univ. Tex. 8:184-211. 117 LARVAL DEVELOPMENT OF BLUE GRENADIER, MACRURONUS NOVAEZELANDIAE (HECTOR), IN TASMANIAN WATERS B D BruceI ABSTRACT The development of Macruronus novaezelandiae is described and illustrated from both reared speci- mens and larvae from Tasmanian waters. Eggs of M. novaezelandiae are pelagic, spherical (1.08-1.18 mm diameter), and have a single oil droplet (0.36-0.42 mm diameter). Eggs hatch after 55-60 hours at 14°-19°C. Larvae are 2.2-2.3 mm at hatching. Characteristic pigmentation, a myomere count of 78-80, and the sequence of fin development separate M. novaezelandiae from other known gadiform larvae. Development is direct, with no marked change in body morphology. Fin development proceeds in the sequence: second dorsal, anal, first dorsal, pelvic, caudal, pectoral. However, adult fin comple- ments are reached in the sequence: first dorsal, pelvic, anal, second dorsal, caudal, pectoral. Caudal development is late in Macruronus. Flexion begins at 20 mm and is not complete until 28 mm. The caudal fin is based on two ural centra, four hypurals, two epurals, and a parhypural. X and Y bones are present although they are not readily distinguishable from dorsal and anal pterygio- phores. The genus Macruronus comprises four nominal species, which occur in southern temperate conti- nental shelf and slope regions. Two species, Macruronus novaezelandiae and M. magellanicus support commercial fisheries. The blue grenadier, M. novaezelandiae , forms the basis of fisheries in New Zealand and Australia where total annual catches range up to 97,750 and 1,100 t respec- tively (Patchell 1982; Wilson 1981, 1982). Macruronus magellanicus is fished commercially off South America. The remaining species, M. maderensis and M. capensis , are known only from a limited number of specimens (Svetovidov 1948; Cohen 1986). Despite their economic importance and widespread distribution, very little is known of the early life history of any member of the genus. Patchell (1982) identified winter spawn- ing grounds on the west coast of the South Island for New Zealand populations of M. novaeze- landiae and similarly Wilson (1981, 1982) has suggested a winter spawning, on the west coast of Tasmania, for Australian M. novaezelandiae. This paper presents the first published informa- tion on the larvae of Macruronus . In 1984, the Division of Fisheries Research of the Commonwealth Scientific and Industrial Re- search Organization established a multidisci- ICSIRO Division of Fisheries Research, GPO Box 1538, Ho- bart, Tasmania 7001, Australia; present address: South Aus- tralia Department of Fisheries, GPO Box 1625, Adelaide, South Australia 5001, Australia. Manu.scnpt accepted September 1987. FISHERY BULLETIN: VOL. 86. NO. 1, 1988. plinary program to investigate the biology and ecology of blue grenadier in Tasmanian waters. An integral part of this program was a study of larval ecology. As such, it was first necessary to establish criteria for the identification of blue grenadier larvae. This paper describes the larval development of M. novaezealandiae from Tasma- nian waters. MATERIALS AND METHODS Specimens were obtained from samples col- lected aboard the CSIRO Fisheries Research Ves- sel Soela between April 1984 and September 1985. Details of sampling strategies, locations, and procedures will be described in a subsequent manuscript. Larvae were obtained by sampling with a rectangular midwater trawl (RMT 1+8; Baker et al. 1973), aim diameter ring net (500 [im mesh), and free-fall, vertical drop nets of 64 |jLm and 200 fxm mesh (Heron 1982). Juvenile specimens were obtained with an Engels 352 pelagic trawl fitted with a 10 mm liner. Newly hatched larvae were reared from eggs stripped and fertilized at sea. Eggs and milt stripped from ripe adults trawled from 500 m were mixed in 1 L plastic jars filled with sea- water. Despite the jars being located in a sea- water bath, incubation temperatures varied con- siderably (14°-19°C). On return to the laboratories at Hobart, the eggs were transferred to 2 L glass jars and placed in a constant temper- 119 FISHERY BULLETIN: VOL. 86, NO. 1 ature incubation chamber set at 14.0° ± 0.2°C. In- cubation jars were not aerated, and no attempt was made to feed the larvae. All specimens used for description were fixed in a IQ% formalin-seawater solution buffered with sodium p-glycerophosphate and later transferred to a 59^ solution. This description is based on a series of 74 lar- vae, 2.2-34.2 mm in length, although comments on pigment and meristic variability stem from routine examination of several hundred speci- mens. A representative series of larvae is de- posited with the South Australian Museum, Ade- laide, South Australia. Developmental terminology follows Ahlstrom et al. (1976). Body measurements follow Matarese et al. (1981). Length measurements are reported as notochord length, NL (i.e., from the snout tip to the end of the notochord) in preflexion and flexion larvae, and standard length, SL (i.e., from the snout tip to the posterior margin of the superior hypural elements) in postflexion larvae and juve- niles. Larvae were measured under a dissecting microscope fitted with an ocular micrometer and a camera lucida. Juveniles were measured with vernier calipers. Meristic counts and examination of ossification sequences were made on specimens cleared and stained using Alizarin Red S-KOH-glycerine (Hollister 1934). Caudal osteology follows Inada (1981), Marshall and Cohen (1973), and Monod (1968). Vertebral counts include the first vertebrae, the neural spine of which is fused to the supraoc- cipital crest (Marshall 1966), and both ural cen- tra. Vertebral centra were counted as ossified only when a complete band of stain connected both neural and haemal spines. RESULTS Identification of M. novaezelandiae larvae was based on their typical gadiform morphology (large head, compact gut, tapering body form), myomere count, and the development of confluent dorsal-caudal-anal fins (see section on Distin- guishing Features). Identification of field- collected specimens was confirmed by comparison with reared larvae. Distinguishing Features Prior to median fin development, myomere counts are useful in separating M. novaezelandiae larvae (78-80) from similarly pigmented morid (41-72), macrourid (10-16 + 70 > 100), gadid (39-64) and other known merlucciid larvae (48- 58) which they superficially resemble (Marshall and Iwamoto 1973; Fahay and Markle 1984; present study). Both M. novaezelandiae and most morid larvae show moderately pedunculate pectoral fins, a fea- ture common in gadiform larvae with delayed caudal development (Fahay and Markle 1984). Macrourid larvae, in contrast, have very promi- nently stalked pectorals and can further be sepa- rated from M. novaezealandiae and most morids by precocious development of the pelvic fin. Size at caudal flexion and the sequence of fin development are also useful in separating M. novaezelandiae from all Merluccius species. In Merluccius , notochord flexion generally begins at about 9 mm and the caudal fin is the first to form (Dunn and Matarese 1984; Fahay and Markle 1984). Macruronus novaezelandiae larvae do not begin caudal flexion until approximately 20 mm, and the caudal fin is the second last to form. Macruronus novaezelandiae larvae have 1-3 prominent melanophores along the ventral mid- line of the tail (although variable in appearance, see section on Trunk and Tail Pigmentation) and a double series of dorsal melanophores. When expanded, melanophores in these two regions coalesce to give the appearance of a broad postanal band. Postanal banding patterns are widespread in gadoid larvae (Fahay and Markle 1984); however, unlike many gadoid larvae, M. novaezelandiae lacks pigment at the notochord tip. At larger sizes M. novaezelandiae larvae de- velop long-based dorsal and anal fins confluent with the caudal fin. Other gadoid larvae with this configuration have markedly different pigmenta- tion (see Fahay and Markle 1984 for details). Ophidiiform larvae have confluent dorsal, caudal, and anal fins but can be separated from M. no- vaezelandiae by their lack of a separate first dor- sal fin and general lack of body pigment (see Gor- don et al. 1984). Development Embryonic development has not been treated in detail here as it is the subject of a manuscript in preparation by G. Patchell (Fisheries Research Centre, Wellington, New Zealand). The pelagic eggs of blue grenadier are spheri- cal, with an unsegmented yolk and a smooth cho- 120 BRUCE: LARVAL DEVELOPMENT OF BLUE GRENADIER rion. Late-stage eggs are 1.08-1.18 mm in diame- ter with a single oil droplet of 0.36-0.42 mm di- ameter (Fig. lA). Reared larvae hatch at 2.2-2.3 mm after 55-60 hours ( 14'-19"C). Newly hatched larvae have a posteriorly positioned oil droplet and adopt a head down position in rearing con- tainers. Yolk absorption was incomplete in speci- mens reared to 3.7 mm (6 days posthatch), al- though the smallest field-collected larvae (3.6 mm) had already completed yolk absorption. The anus opened laterally to the right in all reared larvae and 95*^ of field-collected larvae. The anus becomes symmetrical by 5.1 mm. A lateral anal opening in M. novaezelandiae is consistent with the developmental pattern reported for other gad- iform species (Marak 1967; Matarese et al. 1981; Fahay and Markle 1984; Dunn and Vinter 1984). Field-collected larvae are moderately elongate with the greatest body depth (16-229? body length) occurring at or near the pectoral fin base (Table 1). Head length as a proportion of body length (BL) remains relatively constant at about 22% BL throughout the larval phase, decreasing to about \1'7( BL in juveniles. Eye diameter de- creases from 99f BL in preflexion larvae to A'7c BL in juveniles. Depth at anus remains relatively constant at about 139^ BL in larvae and juveniles. Distances from the snout tip to the first dorsal fin and from the snout tip to the anal fin decrease slightly during development from about 279?^ BL to 2l7c BL and 519r BL to 469? BL respectively. in the number of melanophores and their degree of expansion. Although Badcock and Merrett (1976) suggested the appearance of melanophores can change on a diurnal rhythm, in the blue grenadier larvae examined, there was no conspic- uous relationship between time caught and me- lanophore expansion. Head Pigmentation Newly hatched larvae (2.2-2.3 mm) have melanophores concentrated in front and behind the eye (Fig. IB). Melanophores increase in num- ber and extend over the sides of the head and snout by 3.3-3.5 mm (reared larvae, Fig. IC). Melanophores migrate dorsally to the top of the head by 3.6 mm (Fig. ID). Eyes become pig- mented at this size in reared larvae. By 4.5 mm, the dorsal pigment on the head consists of a group of 3-11 melanophores scattered over the hind- brain and posteriorly to above the cleithrum. Pig- ment gradually extends over the midbrain, with 1 or 2 melanophores usually present between the eyes by 5.3 mm. Melanophores develop externally over these initial mid- and hindbrain spots and extend posteriorly as a double row to the dorsal fin anlage by 7.2 mm. Dorsal pigment gradually intensifies: melanophores increase in number and form a cap over mid- and hindbrains by 16.0 mm. Melanophores extend down between the eyes to the tip of the maxilla by 12.0 mm. Internal Table 1 — Body proportions of larvae and juveniles of Macruronus novaezelandiae (expressed as percentage NL or SL): mean, standard deviation, range. Preflexion Flexion Postflexion Juvenile Body proportions sample size 42 4 4 2 X SD range X SD range X SD range X SD range length (mm) 8.9 4.5 (3.6-19.0) 23.5 2.3 (20.6-26.1) 300 2.9 (27.6-34.2) 1890 1.4 (188.0-190.0) head length 227 1.7 (18.3-24.7) 23.7 1.0 (223-24.7) 22.3 1.4 (20.5-23.6) 17.6 0.4 (17.3-17.9) eye diameter 92 0.7 (8,1-10.3) 7.9 0.4 (7.3-8.3) 7.3 0.7 (6.5-8.0) 4.2 0.4 (4.4-5.2) snout length 6.1 0.9 (4.6-7.7) 6.2 0.6 (5.7-7.0) 5.9 0.5 (5.4-6.5) 4.9 0.1 (4.8-5.0) depth at pectoral 223 1.7 (21.0-24.4) 17.9 0.6 (17.2-18.5) 16.6 2.0 (13.7-18.5) 13.1 0.1 (13.0-13.2) depth at anus 120 2.4 (8.2-15.2) 13.1 0.3 (12.6-13.4) 13.0 0.8 (12.3-13.8) 12.7 0.6 (12.3-13.1) snout to first dorsal fin 27.5 1.3 (25.2-29.3) 26.6 0.8 (26.0-27.7) 25.3 0.8 (24.2-26.0) 20.7 0.2 (20.6-20.9) snout to anal fin 51.4 0.9 (50.0-52.6) 50.4 0.5 (49.6-50.6) 46.6 1.5 (45.0-48.4) 46.5 1.2 (45.7-47.4) Pigmentation Although pigmentation in M. novaezelandiae is variable, certain features persist that, when com- bined with meristic and morphometric informa- tion, enable identification. Variation in the ap- pearance of pigmentation is a result of differences pigment expands over the forebrain in larvae from 9.0 to 15.0 mm. Ventral pigment on the head first develops in 4.2 mm larvae as 3-5 melanophores between the dentaries. The number of melanophores increases to 10-12 by 12.0 mm. 121 FISHERY BULLETIN: VOL, 86, NO. 1 Figure 1. — Development of Macruronus novaezelandiae: A) Late stage egg 1.08 mm diameter, oil droplet 0.37 mm di- ameter, B) 2.2 mm.; C) 3.5 mm.; D) 3.6 mm.; E) 5.3 mm.; F) 7.2 mm.; G) dor.sal view of above; H) 12.0 mm.; I) 24.2 mm. postanal myomeres omitted. A-C reared specimens, D-I field-collected. BRUCE: LARVAL DEVELOPMENT OF BLUE GRENADIER 123 FISHERY BULLETIN; VOL. 86, NO. 1 The onset of dentary pigment is variable; no pigment may be present on some larvae as large as 7.0 mm. Most larvae develop 1 or 2 melano- phores over the posterior section of the dentary by 5.3 mm and add melanophores anteriorly along its length, with 5 or 6 usually present by 7.1 mm. Two melanophores are often present around the otic capsule by 7.0 mm, but they are obscured by overlying tissue in 10.0 mm larvae. Scattered melanophores develop over the pterotic region by 25.0 mm, but the operculum and preoperculum remain largely unpigmented, even in the largest specimen examined (34.2 mm). Trunk and Tail Pigmentation Newly hatched larvae have melanophores on the body above the yolk sac and ventrally on the tail. Some pigment is also present on the yolk sac near the developing gut and scattered over the oil droplet. Pigment forms a cap over the gas bladder by 4.2 mm. Melanophores are gradually added to the lateral surfaces of the gut throughout the lar- val period until the entire gut (including the ven- tral surface) becomes pigmented by 30.0 mm. Dorsal pigment first appears on larvae 3.8-4.5 mm as scattered melanophores at approximately 607c NL. Melanophores rapidly increase in num- ber and form a double row, extending from 51% to 67% NL in larvae of 5.0 mm. Lateral melano- phores may also develop above the body midline in this region. Concurrently, a similar double row of melanophores appears and extends posteriorly from the head (Fig. IG). The head and tail rows join by 10.5 mm. Melanophores appear posteri- orly in the caudal region by 29.0 mm forming a twin series one either side of the developing dor- sal fin. Pigment also appears internally on the dorsal surface of the vertebrae in larvae of 9.5 mm and extends anteriorly to approximately 50% SL and posteriorly to the last vertebrae by 34.0 mm. Single melanophores appear on the dorsal fin ray bases by 14.0 mm and are present on all bases by 29.0 mm. Pigment along the ventral midline of the tail appears in newly hatched larvae as a diffuse re- gion that extends posteriorly from the yolk sac to 75-82% NL. This contracts to 1-3 melanophores (most commonly 2) located 52-65% NL in larvae of 3.8-4.0 mm. Additional melanophores (up to 6) may appear later, but the initial 1-3 melano- phores persist throughout the larval period. In larvae larger than 7.0 mm, the initial 1-3 melanophores appear internally above the anal fin ray bases and are gradually obscured by both overlying musculature and external melano- phores. These ventral melanophores on the tail are a useful diagnostic character, although their appearance varies, depending on their degree of expansion. This variability in melanophore ap- pearance is particularly evident in small larvae where expanded ventral melanophores may ex- tend over the lateral surfaces of the body to al- most the dorsal area (Figs. ID, 2). Lateral pigment gradually intensifies through- out the larval period, excepting the area immedi- ately above the gut, which remains largely devoid of pigment even in the largest specimen (34.2 mm). Morphological Variability Macruronus novaezelandiae larvae showed some size variation in development. In general, specimens captured in ring net and RMT samples appeared to develop features at slightly smaller sizes than those taken from drop net samples. This is likely a result of difi"erential shrinkage of specimens caught by the different capture sys- tems. Hay (1981) reported that considerably more shrinkage occurred in Pacific herring when lar- vae were killed prior to fixation and that shrink- age increased with tow length. Ring net and RMT tows varied in duration from 15 to 110 minutes, with most larvae dead by the time the net was retrieved and the catch fixed. Drop net sampling, in contrast, lasted for, at most, 3 minutes dura- tion, and many larvae were still alive on fixation. Some variability in development can also be ex- pected in field-collected larvae as a reflection of past history (e.g., feeding success), although it is unlikely such variations would account for the observed differences between larvae caught by different techniques. Meristics and Osteology (Table 2) Head and Axial Skeleton In laboratory-reared larvae, jaw development was first visible after 3.5 days (posthatch) with a functional mouth present in larvae of 5.5 days (3.7 mm). Pigmentation of the eyes also occurred at this time suggesting that larvae were ready for first feeding. The smallest larva stained was a field-collected specimen 3.7 mm NL. The maxilla, premaxilla, dentary, and cleithrum were all ossi- fied in this specimen. 124 BRUCE: LARVAL DEVELOPMENT OF BLUE GRENADIER Figure 2. — Variability of ventral pigment on the tail in 4.9 mm larvae of Macruronus novaezelandiae. Ossification of branchiostegals begins in larvae of 4.6 mm, with the full complement of 7 ossified by 11.5 mm. Gill rakers are first discernible in larvae of 9.4-9.9 mm with the full complement of 7 + 22-23 present by 28.9 mm. Ossification of neural and haemal spines gener- ally precedes that of the vertebral centra. Ossifi- cation of centra, neural spines, and haemal spines occurs sequentially from anterior to posterior pro- ceeding slowly in larvae less than 9.0 mm in length and then more rapidly until the full com- plement is ossified by 23.2 mm. Elements associ- ated with the caudal complex are the last to os- sify. Fins Completion of fin development in M. novaeze- landiae occurs in the sequence: first dorsal and pelvic (almost simultaneously), anal, second dor- sal, caudal, pectoral. Pelvic fins first appear in larvae of 5.7— 5.8 mm as slight swellings either side of the gut. They do not form distinct buds until 6.9 mm. Ossification 125 FISHERY BULLETIN: VOL 86, NO 1 Table 2.— Meristic counts from cleared and stained larval and juvenile Macruronus novaezelandiae Specimens between dashed lines are undergoing notochord flexion, a = specimen damaged; b juveniles not stained. Length (mm) 37 3.9 4.2 4.6 4.8 5.2 6.0 74 9.4 9.9 11.5 16.3 17.4 19.8 dorsal 4 + 28 + 19 9 + 74 12 + 84 12 + 86 13 + 87 Fin rays anal pectoral pelvic Branchi- ostegal rays Gill rakers upper lower total Total centra Neural spines Haemal spines 18 4 60 73 76 86 4 8 8 8 1 1 3 3 5 6 6 7 7 7 7 8 6 12 15 15 15 8 6 12 16 18 20 5 42 38 55 70 71 73 1 1 2 2 2 3 6 55 54 58 70 70 74 37 34 42 55 55 57 Caudal elements 3 3 4 23.2 12+ 100 90 9 8 7 5 17 22 76 74 57 5 26.1 13 + 99 90 13 8 7 5 21 26 76 74 57 5 28.9 13 + 99 91 9 8 7 7 22 29 78 76 57 a 188 13 + 94 90 20 8 7 7 22 29 b b b b 190 13 + 96 90 20 8 7 7 23 30 b b b b may start as early as 9.4 mm with the full comple- ment (8 rays) present by 16.3 mm. Ossification proceeds from the outer to the innermost rays. The second dorsal fin anlage is visible in larvae of 5.7 mm. Bases are first visible by 6.9 mm, with ray ossification commencing by 7.3 mm. Al- though the anal fin anlage does not form until 6.9 mm, complete ossification is reached before that of the second dorsal. Distinct anal fin bases are first visible in 7.2 mm larvae and ossification has consistently begun by 9.9 mm. The full comple- ment of anal rays is present by 21.0 mm and for the second dorsal, by 23.2 mm. The first dorsal starts development slightly later than the second dorsal, although it is the first fin to complete ossification. The full comple- ment of 12 or 13 elements is present by 16.3 mm. Pectoral buds were first observed in larvae 4.5 days posthatch (3.2 mm). However, the pectoral fin is the last to complete development. Ossifica- tion of pectoral rays starts by 16.3 mm; a 34.2 mm specimen had only 15 ossified rays, still short of the 20 rays of juveniles. Sequence of ossification is from upper to lower. The caudal fin anlage first appears on the ven- tral surface of the notochord just anterior to the tip in larvae of 10.4 mm. Flexion begins at 20 mm and is usually complete by 28 mm. Ossification of all caudal elements was incomplete in a 34.2 mm specimen. Insufficient material of the appropriate size was available to define the completion of cau- dal ossification. The caudal complex in M. novaezelandiae is based on two ural centra, two epurals, a superior hypural (HP3 + 4), inferior hypural (HPl + 2), and a parhypural (Fig. 3). Eight to nine rays ar- ticulate with these elements — one or two rays on the second epural, three rays on the superior hy- pural, two on the inferior hypural, and one ray each on the first epural and the parhypural. Sin- gle rays also articulate with the elongate neural and haemal spines of the first preural centrum. X and Y bones are present although they are not readily distinguishable from dorsal and anal pterygiophores. Total caudal fin ray counts are low (12 or 13). Additional caudal structures occurred in one of the six specimens examined. This specimen had a twin haemal spine on the first preural centrum and greatly elongated haemal spines on preural centra 3-8 (1.3-1.4 times the length of corre- sponding neural spines, Fig. 3). DISCUSSION The general morphology and pigmentation of M. novaezelandiae larvae show broad similarities to Merluccius and to gadine gadids. Characteris- tic differences between M. novaezelandiae and Merluccius species occur in fin structure and the sequence of fin development. In Merluccius, the caudal fin is the first to form, followed by the pelvic. In Macruronus, caudal development is late with the caudal fin being the second last to 126 BRUCE: LARVAL DEVELOPMENT OF BLUE GRENADIER Additional Haemal Spine Figure 3. — Caudal osteology of a juvenile Macruronus novaezelandiae (181 mm SL). X = X bone, Y = Y bone, EP = epural, SH = superior hypural (hypurals 3 + 4), IH = inferior hypural (hypurals 1 + 21, PH = parhypural, U = ural centra, PU = preural centra. form. The pectoral fin in Macruronus larvae is more markedly stalked than in Merluccius. Fahay and Markle ( 1984) suggested that this pec- toral modification in larvae with delayed caudal development may be a compensatory response as- sociated with swimming. Although the larvae of the remaining mer- lucciid genera (Lyconus and Lyconodes) are cur- rently unknown, fin structure and position should be useful in separating these from Macruronus. Based on adult features, pelvic insertion should distinguish Macruronus (pelvics inserted behind pectorals) from Lyconus (opposite) and Lyconodes (abdominal). Additionally, Lyconus has only a single dorsal fin and no caudal fin. The caudal fin of M. novaezelandiae is similar to Muraenolepis in its confluence with dorsal and anal fins (Fahay and Markle 1984). This similar- ity extends to the undifferentiated X and Y bones and the total caudal fin ray count (12 or 13) re- ported by these authors. However, unlike Mu- raenolepis, M. novaezelandiae has radials fused to the spines of the first preural centrum, which is the more typical gadoid condition. Variability in the structure and appearance of bones associated with the caudal fin has been re- ported for other Macruronus species. Marshall ( 1966) observed double neural arches and "super- numary elements" in M. magellanicus. Indeed, variability in gadiform caudal structure ap- pears not to be unusual with examples in several taxa (Markle 1982). Unfortunately, insufficient specimens in the appropriate 35-150 mm size range were available to assess developmental characteristics of these variations in blue grenadier. ACKNOWLEDGMENTS I thank R. Thresher, J. Gunn, J. Leis, and A. Miskiewicz for their reviews of the manuscript. I also thank D. Furlani for sorting the samples and for her considerable patience in the laboratory. This work was supported by a grant from the Fisheries Industry Research Trust Account. NOTE: Since the acceptance of this paper, the embryological work by A. Patchell (see section on Development) has been published in New Zealand Journal of Marine and Freshwater Re- search Vol. 21, No. 2. That paper includes a simi- lar larval developmental sequence to that re- ported here. LITERATURE CITED Ahlstrom. E H , J L Butler, and B Y Sumida 1976. Pelagic stromateoid fishes (Pisces, Perciformes) of the eastern Pacific: kinds, distributions and early life histories and observations on five of these from the north- west Atlantic. Bull. Mar. Sci. 26:285-402. Badcock, J R . AND N R Merrett 1976. Midwaterfishes in the eastern North Atlantic. I. Vertical distribution and associated biology in 30° N, 23° W, with developmental notes on certain myctophids. Frog. Oceanogr. 7:3-58. Baker. A C, M R Clarke, and M. J Harrls 1973. The N.I.O. combination net (RMT 1 + 81 and further developments of rectangular midwatertrawls. J. Mar. Biol. Assn. U.K. 53:167-184. 127 FISHERY BULLETIN: VOL. 86, NO, 1 Cohen. D M 1986. Merlucciidae. /a; M. M. Smith and P. C. Heemstra (editors). Smiths' sea fishes, p. 324-326. Springer- Verlag, N.Y. Dunn. J R . and A C Matarese 1984. Gadidae: Development and relationships. In H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall, Jr., and S. L. Richardson (editors). Ontogeny and systematics of fishes, p. 283-299. Am. Soc. Ichth. Herpet. Spec. Publ. 1. Dunn. J R . and B M Vinter. 1984. Development of larvae of the safiron cod, Eleginus gracilis, with comments on the identification of gadid larvae in Pacific and Arctic waters contiguous to Canada and Alaska. Can. J. Fish. Aquatic Sci. 41:304-318. Fahay. M P . and D F Markle 1984. Gadiformies: Development and relationships. In H. G. Mo.ser. W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall, Jr., and S. L. Richardson (editors), On- togeny and systematics of fishes, p. 265-283. Am. Soc. Ichth. Herpet. Spec. Publ. 1. Gordon. D J . D F Markle, and J E. Olney. 1984. Ophidiiformes: Development and relationships. In H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall, Jr., and S. L. Richardson (editors). On- togeny and systematics of fishes, p. 308-319. Am. Soc. Ichth. Herpet. Spec. Publ. 1. Hay, D E 1981, Effects of capture and fixation on gut contents and body size of Pacific herring larvae. Rapp. P. -v. Reun. Cons. int. Explor. Mer 178:395-400. Heron. A C 1982. A free fall plankton net with no mouth obstruc- tions. Limnol. Oceanogr. 27:380-383. Hollister, G. 1934. Clearing and dyeing fish for bone study. Zoologica 12:89-101. Inada. T 1981. Studies on merlucciid fishes. Bull. Far Seas Fish. Res. Lab. 18:1-172. Marak.R R 1967, Eggs and early larval stages of the offshore hake, Merluccius albidus. Trans Am. Fish. Soc. 96:227-228. Markle, D F 1982. Identification of larval and juvenile Canadian At- lantic gadoids with comments on the systematics of gadid subfamilies. Can. J. Zool. 60:3420-3438. Marshall. N B 1966. The relationships of the anacanthine fishes, Macruronus, Lyconus and Steindachneria. Copeia 1966:275-280. Marshall, N B , and D M Cohen. 1973. Order Anacanthini (Gadiformes). Fishes of the western North Atlantic. Mem. Sears Found. Mar. Res. l(6):479-495. Marshall, N, B,, and T Iwamoto, 1973, Family Macrouridae, In D. M. Cohen (editor). Fishes of the western North Atlantic, Part 6, p. 496- 665. Mem. Sears Found. Mar. Res. Yale Univ. 1. Matarese, A C . S L Richardson, and J R Dunn 1981. Larval development of Pacific tomcod, Microgadus proximus, in the northeast Pacific Ocean with compara- tive notes on larvae of walleye pollock, Theragra chalo- gramnia, and Pacific cod, Gadus macrocephalus (Gadi- dae). Fish. Bull., U.S. 78:923-940. Monod, T 1968. Le complexe urophore des poissons teleosteens. Mem. Inst. Fond. Afr. Noire. 81:1-705. Patchell, G J 1982. The New Zealand hoki fisheries 1972-82. Fish. Res. Div. Occas. Publ. 38, 23 p. Svetovidov, a N 1948. Fauna of the USSR, Fishes, Gadiformies. Zool. Inst. Acad. USSR 9(4):l-222. Wilson, M A 1981. Blue grenadier spawning grounds. Fintas 4:9-10. 1982. Spawning blue grenadier caught off Cape Sorell. Fintas 4:13. 128 THE DISTRIBUTION, ABUNDANCE, AND TRANSPORT OF LARVAL SCIAENIDS COLLECTED DURING WINTER AND EARLY SPRING FROM THE CONTINENTAL SHELF WATERS OFF WEST LOUISIANA^ James H Cowan, Jr.^ and Richard F. Shaw^ ABSTRACT The larvae of six species of Sciaenidae were collected in continental shelf waters off west Louisiana on five midmonthly ichthyoplankton cruises from December 1981 to April 1982. Ranked in order of abundance these species were sand seatrout, Cynoscion arenarius; Atlantic croaker, Micropogonias undulatus; spot, Leiostomus xanthurus; black drum, Pogonias cromis; southern kingfish, Menticir- rhus amencanus; and banded drum. Larimus fasciatus. Total larva density was highest in April, and the high densities were associated with the coastal boundary layer, a horizontal density front caused by an intrusion of fresher water onto the inner shelf that probably issued from the Atchafalaya River east of the study area. Spawning by sand seatrout began in January, two months earlier than previously reported, and first occurred offshore of midshelf but moved shoreward as the season progressed. Analysis of length-frequency data suggest that spot probably began to spawn in Novem- ber, one month earlier than once thought. Both sand seatrout and Atlantic croaker larvae were captured at higher rates at night than during the daytime. Sand seatrout larvae appear to be somewhat surface oriented while spot may undergo vertical migration. Interpretation of the sciaenid data support a previously developed transport hypothesis involving gulf menhaden larvae and west- northwest alongshore advection within and just outside of a horizontally stratified coastal boundary layer. Members of the perciform family Sciaenidae are an important sport and commercial fishery re- source along the United States coast of the Gulf of Mexico and are perhaps the most prominent group of northern Gulf inshore fishes. Sciaenids exceed all other families in numbers of species (18) and in numbers of individuals or biomass; they are among the top four families with Mugili- dae, Engraulidae, and Clupeidae (Gunter 1938, 1945; Moore et al. 1970; Franks et al. 1972; Hoese and Moore 1977). Of the six species of sciaenids captured during this study, only the banded drum, Larimus fasciatus, is not commonly sought by both sport and commercial fishermen. Many of Louisiana's sciaenids spawn in coastal or offshore waters. They have pelagic eggs and young which are then transported into estuaries (Johnson 1978 for review). The seasonal impor- tance of Louisiana's estuaries as nursery grounds 'Louisiana State University Contribution No. LSU-CFI-86- 08. ^Center for Wetland Resources, Louisiana State University, Baton Rouge, LA 70803-7503; present address: Center for Envi- ronmental and Estuarine Studies, University of Maryland, Chesapeake Biological Laboratory, Box 38, Solomons, MD 20688-0038. ^Center for Wetland Resources, Louisiana State University, Baton Rouge, LA 70803-7503. for postlarval and juvenile sciaenids is well docu- mented (Cowan 1985 for review), and several summary works are available which contain tax- onomic and biological information on adult sciaenids (Pearson 1929; Suttkus 1955; Guest and Gunter 1958; Hoese and Moore 1977; Johnson 1978; Powles and Stender 1978; Barger and John- son 1980; Barger and Williams 1980; Mercer 1984a, b). In contrast, there is little information about sciaenid ichthyoplankton assemblages in Gulf continental shelf waters, their offshore and coastal distribution, or the oceanic current sys- tems which influence their estuarine recruit- ment. This study provides such early life history in- formation by determining larva distribution, abundance, and length frequency; by document- ing spawning location (depth and distance from shore) of winter and early spring-spawned sciaenids off west Louisiana; and by analyzing larval sciaenid distribution with respect to known water circulation patterns and a larval gulf men- haden, Brevoortia patronus, transport hypothesis in the shelf waters of the northwestern Gulf of Mexico (Shaw et al. 1985b). Recruitment implica- tions of the observed distribution, larva age struc- ture, and transport of sciaenids in Louisiana waters are also discussed. Manuscript accepted September 1987. FISHERY BULLETIN: VOL 86. NO. 1, 1988. 129 FISHERY BULLETIN: VOL 86. NO 1 METHODS AND MATERIALS Detailed sampling methodology has been pre- sented elsewhere (Shaw et al. 1985 a, b). Briefly, larval sciaenids were collected off west Louisiana on a sampling grid consisting of 37 stations on 5 transects (Fig. 1) during 5 midmonthly cruises from December 1981 to April 1982. Ichthyoplank- ton samples were analyzed from the 335 |xm mesh net side of an opening and closing, 60 cm, paired "bongo type" plankton sampler fitted with Gen- eral Oceanics' flowmeters (model no. 2030). Most plankton collections (125 of 187 total) consisted of 10-min stepped oblique tows from near bottom to surface. Nets were set closed and opened just prior to the stepped ascent. Each tow had five steps with a retrieval rate between steps of 20 m/minute; towing speed was about 1 m/second (2 knots). The object of the 10-min tow was to filter approximately 100 m*^ of water. This process increased the water volume filtered per unit depth at the shallow stations relative to deeper stations. This discrepancy is acceptable since the alternative would be to compare 17-s shallow- station oblique tows with 9-min deep-station tows at a uniform retrieval rate (Houde 1977). At se- lected stations (A-3, 6, 9; B-1; C-6; D-1; E-3, 6, 9; Fig. 1), only 10-min simultaneous surface and near-bottom horizontal tows (31 surface and 31 near-bottom) were made to determine if sciaenid larvae were vertically stratified. Larva total length (TL) was measured to the nearest 0.1 mm. Larva densities are reported as standardized catch rates at a station (density = larvae/100 m"^). A four-way analysis of variance (ANOVA) was performed on logio transformed [(no. larvae/100 m^) + 1] data to determine the spatial (vertical and horizontal), temporal, and diel patterns of species density and distribution. The four main effects tested were month (January-April); sta- tion depth group (d.g.) (d.g. 1 < 10 m, 10 m < d.g. 2 < 14 m, 14 m < d.g. 3 < 24 m and d.g. 4 > 24 m); day-night (2000 hours < night < 0500 hours); and horizontal tow type (surface vs. near-bottom). Data from the December cruise were not included as only the A transect was completed due to ad- verse weather conditions. Two methods of current estimates were utilized (following Shaw et al. 1985b): 1) instantaneous current profiles taken at each station and 2) continuous surface and near-bottom current ■♦Reference to trade names does not imply endorsement by the National Marine Fisheries Services, NOAA. meter measurements at two sites (H and S; Fig. 1). The instantaneous number of larvae trans- ported on each transect was calculated by using the equation D x U x M = number of larvae per meter per second where D = larva density darvae/m') from either oblique tows or from the mean of the horizontal tows (i.e., average of sur- face and near-bottom catch rates), U = depth- averaged water velocity (m/s) determined from instantaneous current meter profiles at each sta- tion, and M = water depth (m) at each station. Distribution diagrams and length-frequency histograms were generated for each cruise for the three most abundant sciaenid species. Inspection of these data along with current measurements allowed a comparison with the previously men- tioned transport hypothesis. RESULTS AND DISCUSSION Total Sciaenids A total of 5,225 larval sciaenids accounted for 9.1% of the fish larvae collected. In December through February, samples were dominated by Atlantic croaker, Micropogonias undulatus, and spot, Leiostomus xanthuriis. In March and April samples contained mostly sand seatrout, Cynoscion arenarius. In all, six species of sciaenid larvae were collected: sand seatrout (N = 4,100); Atlantic croaker {N = 567); spot (A^ = 264); black drum, Pogonias cromis {N = 68); southern king- fish, Menticirrhus americanus (N = 53); and banded drum (A'^ = 13). Additional Menticirrhus, not identifiable to species, accounted for 160 more specimens (Table 1). A more detailed examina- tion of the data on the three most abundant sciaenid species follows. Sand seatrout, Cynoscion arenarius A total of 4,100 sand seatrout larvae was col- lected making it the most abundant sciaenid taken during the study. Larval sand seatrout den- sities were highest in April (Table 1) with a mean of 46.1 larvae/100 m"^; mean density in February and March was 0.3 and 2.9/100 m'^, respectively, and 1 larva was collected in January. Larvae were distributed mostly over the midshelf in February but highest concentrations were later found inshore and towards the east (Fig. 1). Over the course of study, larvae were found in temper- atures and salinities ranging from 14° to 21°C and 130 COWAN AND SHAW: LARVAL SCIAENIDS COLLECTED OFF WEST LOUISIANA Figure l. — Density distribution of sand seatrout, Cynoscion arenarius, larvae by month, February-April 1982. Densities are as follows: o = 0: • >0-10; • >10-50; • >50-99; • >99-250; 250 100 m3 of water filtered from all plankton tow types. Also shown is the station sampling grid with moored current meter sites (H and S) and selected iso- baths. SP = Sabine Pass, CR = Calcasieu River, and MR = Mermentau River. 131 KISHKRY BULLETIN: VOL, 86, NO. 1 Table 1. — Total number, months of occurrence and monthly density of sciaenid larvae collected in west Louisiana shelf waters from December 1981 to April 1982. Months of occurrence and Total number density (No. /1 00 m3) Taxa Dec. Jan. Feb. IVIar. Apr. Cynoscion arenarius 4,100 — 1 larva 0.3 2.9 46.1 Micropogonias undulatus 567 20 2.2 32 22 02 Leiostomus xanthurus 264 7.8 1.3 0.9 0.1 — Menticirrhus sp. 160 — — — — 19 Pogonias cromis 68 — 1 larva 0.5 0.2 0.2 Menticirrhus americanus 53 — — — — 0.6 Larimus fasciatus 13 — — — — 0.2 from 15 to 36 ppt (Table 2) and at station depths ranging from 5 to 70 m, but most were collected inside the 18 m isobath. Larval sand seatrout density increased in April with many stations ex- hibiting densities in excess of 250 larvae/100 m'^; the larvae appeared to be associated with a freshet of water on the shelf, probably issuing from the Atchafalaya River east of the study area (Shaw et al. 1985a). The presence of riverine runoff on the shelf, which was most evident in March and April 1982, caused the development of an oceanic salinity front referred to as the coastal boundary layer, 10-35 km from shore (Wiseman et al. 1987). Observed spawning seasonality and location for sand seatrout are in part consistent with pre- viously published information (Shlossman and Chittenden 1981 for review). The presence of a 4 mm TL larva in January indicates some spawn- ing had taken place at least 2 months earlier than Table 2. — l^onthly data summaries at time of capture for three sciaenid larvae (Cynoscion arenarius, Micropogonias undulatus. and Leiostomus xanthurus) col- lected in west Louisiana shelf waters from December 1981 to April 1982. Total Species'month number Length range (mm TL) Temperature range ( C) Salinity range (ppt) Depth range (m) Cynoscion arenarius January 1 4 14 35 18 February 20 2.5-4.5 (mode = 2-3) 14-20 34-36 15-70 March 203 1.5-10.5 (mode = 2-3) 14-18 25-36 5-40 April 3,876 1.5-20.5 (mode = 2-3) 20-21 15-36 5-70 Micropogonias undulatus December 28 2.5-10.5 (mode = 3-4) 12-20 30-36 10-65 January 158 2.5-10.5 (mode = 4-5) 10-18 30-36 5-70 February 221 2.5-17.5 (mode = 14-15) 11-17 27-36 5-40 March 144 2.5-19.5 (mode =14-15) 14-20 25-36 5-115 April 16 11.5-18.5 (mode= 17-18) 20.5 22 7 Leiostomus xanthurus December 110 2.5-7.5 (mode = 3-4) 14-18 30-36 16-65 January 89 2.5-13.5 (mode = 4-5) 10-18 30-36 5-40 February 62 3.5-15.5 (mode= 12-13) 10-17 28-36 5-40 March 3 3.5-16.5 14-17.5 26-36 11-40 132 COWAN AND SHAW: LARVAL SCIAENIDS COLLECTED OFF WEST LOUISIANA previously reported. Monthly length-frequency data for sand seatrout show that larvae as large as 11 mm TL were first present in March samples (Fig. 2A). Based on the estimated growth rate determined for sand seatrout (Cowan 1985; Shaw et al., in press), an 11 mm larva could be as old as 65 days; this further supports January spawning. Sand seatrout are reported to spawn from March to August, during two discrete periods — one in March-May, the other August-September, with little spawning between the two peak periods (Hoese 1965; Daniels 1977; Shlossman and Chit- tenden 1981). An examination of distribution and length- frequency data (Figs. 1, 2A) suggests that most spawning initially took place in midshelf to off- shore waters at depths ranging from 15 to 80 m or to about 175 km from shore. As the season pro- gressed into March and April, spawning location, as determined by the presence of larvae <3.0 mm TL, was more inshore (5-18 m) with few small larvae occurring at depths >25 m. Other than the indication that spawning may move from offshore to inshore waters as the sea- son progresses, this spatial information agrees with the limited life history data available on sand seatrout. Most spawning has been shown to occur in the shallow waters of the Gulf of Mexico, primarily between 7 and 15 m in depth (Gunter 1945; Moffet et al. 1979; Shlossman and Chitten- den 1981). Running ripe C. arenarius have been captured in deepter waters (70-90 m) in February and March, but no spawning was indicated (Franks et al. 1972; Perry 1979). In a four-way ANOVA employed to determine patterns of larval sand seatrout density and dis- tribution, month was a highly significant main effect (P < 0.01; Table 3) reflecting spawning sea- sonality and the magnitude of the density in- crease in April. The test for interaction between month and day-night was employed to determine if daytime gear avoidance was evident as size and mobility of larva increased. Most sand seatrout collected, however, were small and no clear monthly modal increase in larva size was evident (Fig. 2A). The significant interaction (P <0.01) was probably due to an increased catch in oblique tows at night as the season progressed (0.0 in January, 64.9/100 m"^ in April). The significant interaction between month and depth group and the highly significant depth group main effect (P < 0.01) represents the shift in larva concentra- tion from midshelf early in the study, to a more coastal distribution in March and April (Fig. 1). Mean larva density was greatest in depth group 1 (23.7/100 m'^) followed by depth groups 2, 3, and 4 (12.7, 9.2, and 0.3/100 m^, respectively). The third main effect, day vs. night tows, was highly sig- nificant (P < 0.01); many more sand seatrout larvae were collected at night (averge catch rates in all night (74) tows combined = 21.6/100 m"^ vs. day (113) tows = 7.4/100 m^). Highest night- time catches occurred in oblique (49) tows (26.9/ 100 m'^) while the day-oblique-catch rate aver- aged 7.6/100 m'^ in 76 tows. Overall, average catch rate was highest in oblique (125) tows (14.6 larvae/100 m-^), followed by surface (31) tows (9.2/100 m^), and then bottom (1.9/100 m^; 31 tows). Intrepretation of the data suggests that Table 3. — Summary data from four-way analysis of variance done on logio transformed [(larvae 100 m3) + 1] data from ichthiyoplank- ton samples collected from January to April 1982. Tfie results are for A. Cynoscion arenarius. B, Micropogonias undulatus, and C. Leiostomus xanthurus. The four main effects tested were months (Jan. -Apr.), station depth group (d.g. 1 < 10 m, 10 m < d.g. 2 < 14 m, 14 m < d.g. 3 < 24 m and d.g. > 24 m), day - night (2000 hours s night < 0500 hours) and horizontal tow type (surface vs. near-bottom). Source df PR r2 = 0.75 A. Dependent variable: Log 10 [(Cynoscion arenarius 1^00 m3) + 1] Model 21 O.OOOr* r^^onth 3 0.0001" Depth group 3 0.0001" Day-night 1 0.0026" Horizontal tow type 1 0.2574 (NS) IVIonth vs. Day-night 3 0.0001" l\/lonth vs. Depth group 9 0.0001" Day-night vs. Tow type 1 0.4180 (NS) Error 177 Corrected Total 1 98 B. Dependent variable: Logio [(Micropogonias undulatusHOO m3) + 1] Model 21 0.0001" r2 = o.63 Month 3 0.0045" Depth group 3 0.3551 (NS) Day-night 1 0.0001" Horizontal tow type 1 0.4448 (NS) Month vs. Day-night 3 0.2168 (NS) Month vs. Depth group 9 0.0001" Day-night vs. Tow type 1 0.1288 (NS) Error 177 Corrected Total 1 98 C. Dependent variable: Logio [(Leiostomus xanthu^us/^00 m3) + 1] Model 21 0.0001" /-s = 0.51 Month 3 0.0001" Depth group 3 0.0033" Day-night 1 0.1875 (NS) Horizontal tow type 1 0.3216 (NS) Month vs. Day-night 3 0.2138 (NS) Month vs. Depth group 9 0.0001" Day-night vs. Tow type 1 0.0324* Error 177 Corrected Total 198 * = Statistically significant (P < 0.05). " = Highly significant (P < 0.01). (NS) = Not significant. 133 FISHERY BULLETIN: VOL. 86, NO. 1 00 CM II O ill Q CD O !• I I I ' I ' I 8Q Q O p O o ift ^ r^ 04 ^ 00 U1 II z z' < -J C L [ r° i E CM CM II z CD S 2 I I I S § CO 8 2 8 II z m' K to II z m [ [ — I— o -T — O i: [ < z o Cvj o CM II z if < S r I 1 So Q O p O O u^ ^ r-1 "N -- w'::. r" ^Kis;. ( :.... r:::„ r r— ' — 1 — •— I — •" r" e II z of < Z ' E (0 o II z (0 C 2 I UJ o -I w -I 9 I- O nVlOl dO iN30a3d II z of OL < o •o c (0 3 c 3 (O c o en o & CD oj 3 C n: c F QJ "O fe iJ F ^ S wi Z O .O CD U 0) 7 5 "O UJ -J O^S -1 < -c < O) H D O Cvi S 1- CO " ^ o ^ ^ E Q- D < c j|_ (U £ , E 0) O Q) a (0 ■o c: a> CO O (0 134 COWAN AND SHAW: LARVAL SCIAENIDS COLLECTED OPT WEST LOUISIANA sand seatrout larvae v^'ere somewhat surface oriented. Atlantic croaker, Micropogon ias undiilatus The second most abundant sciaenid taken was Atlantic croaker (A'^ = 567). Larval Atlantic croaker density was greatest in February at 3.2/ 100 m'^, but density was relatively constant from December through March (Table 1). Mean densi- ties for December, January, March, and April were 2.0, 2.2, 2.2, and 0.2/100 m^, respectively. Atlantic croaker was the only sciaenid collected in all months of the study. Their overall distribu- tion (all sizes combined) was generally uniform over most of the shelf (Fig. 3) except in March and April when they were more often found inshore. Recently spawned larvae (<3.0 mm TL) were also collected over much of the shelf at station depths ranging from 15 to 115 m or from about 20 to 200 km from shore. However, most small larvae were collected near midshelf about 65-125 km from shore. In December and January the majority of the larvae were small. By April, no recently spawned individuals were collected (Fig. 2B). For the study overall, larvae were found in salinities and temperatures ranging from 22 to 36 ppt and from 10 to 20.5°C (Table 2). Spawning by Atlantic croaker in Gulf of Mexico waters is reported to occur from September to March, with a distinct peak in October (Hoese 1965; Sabins and Truesdale 1974; White and Chittenden 1977; Benson 1982) and to occur pri- marily offshore over a wide area (Pearson 1929; Hildebrand and Cable 1930; Wallace 1940; Haven 1957; Bearden 1964; Hoese 1965; Nelson 1967). Atlantic croaker larvae, however, have been taken on the outer continental shelf off Texas from September to May (Finucane et al. 1979). As with sand seatrout, a four-way ANOVA was used to determine patterns in larva density and distribution (Table 3). Larval Atlantic croaker density by month was a highly significant main effect (P <0.01). Densities at the end of their spawning period were low, increased only slightly in February, and then dropped off by April (Table 1). The interaction between month and day-night was not significant. The highly significant inter- action between month and depth group was not surprising. Larvae were in more offshore waters early in the study while later becoming more abundant inshore (Fig. 3). However, as a main effect, depth gi'oup was not significant. Larval Atlantic croaker mean densities for depth groups 1 through 4 were 3.9, 0.8, 0.4, and 0.7/100 m^, respectively. Day-night, as a main effect, was highly significant (P < 0.01). Larval Atlantic croaker density was over 5 times higher at night (all tow types combined) than during the day (3.7 vs. 0.7/100 m'^). However, the interaction between day-night and horizontal tow type was not signif- icant. The fourth main effect tested, horizontal tow type, was not significant. Average catch rates at the surface and near-bottom were similar (1.0 and 1.8/100 m'\ respectively). Spot, Leiostomiis xcmthurus The third most abundant sciaenid collected was spot (A^ = 264). Density of spot larvae was highest in December at 7.8 larvae/100 m'^ (Table 1). How- ever, the high December value must be viewed with a consideration of the abbreviated cruise track for that month and the resultant reduction in spatial coverage. Mean densities for January to March were 1.3, 0.9, and 0.1/100 m"^, respectively. No spot larvae were collected in April. In general, larva density was low and their distribution was uniform over the shelf out to the 40 m isobath, about 130 km offshore (Fig. 4). Spot were col- lected in temperatures and salinities ranging from 10° to 18°C and from 26 to 36 ppt (Table 2), and at stations with depths ranging from 5 to 65 m. Larvae >7 mm TL in our mid-December (Fig. 2C) collections and small larvae (<3.0 mm TL) in all but the last cruise indicate that spawning probably began by at least November and contin- ued through March. Spawning occurred from near midshelf (about 65 km) out to 175 km from the coast. Data presented here partly concur with previously published information on spot spawn- ing periodicity. In the northern Gulf, spawning reportedly occurs from late December to March, peaking in January, and takes place well offshore in moderately deep water (Pearson 1929; Kilby 1955; Townsend 1956; Dawson 1958; Springer and Woodburn 1960; Pacheco 1962; Nelson 1967; Joseph 1972; Music 1974; Sabins and Truesdale 1974). A four-way ANOVA indicated that month, as a main effect for spot larvae, was highly significant (P < 0.01 ), which probably reflects the decreasing catch rates seen from January to March (Table 3). The interaction between month and depth group was also highly significant (P < 0.01) as was depth group as a main effect. Larval spot 135 FISHERY BULLETIN VOL 86. N(J 1 DEC -yT ,4^- -^ ^?^IV^ ° • : --^.^^^^ o o o . - o " o o o o o o o o o o o c o o - MAR 1 o 1 APR Figure 3— Density distribution of Atlantic croaker, Micropogonias undulatus, larvae by month, December 1981— April 1982. Densities are as follows: o = 0; • >0-10; • > 10-50; • >50;-99; • >99-250; # >250/100 m3 of water filtered from all plankton tow types. 136 COWAN AND SHAW: LARVAL SCIAENIDS COLLECTED OFF WEST LOUISIANA i.Tr'if" 1 ^u^ *^^\ -^ o o o ° O O . o O o o o o o o FEB o 1 1 o o o o o o o o o o o o O o o o o o . o o o o o o 1 MAR o Figure 4 — Density distribution of spot, Leiostomus xanthurus. larvae by month, December 1981-Marcti 1982. Densities are as follows: o = 0; >0-10; • > 10-50; • >50-99; • >99- 250; 9 >250/100 m3 of water filtered from all plankton tow types. 137 FISHERY BULLETIN: VOL. 86, NO. 1 densities were higher offshore in the early part of the study and then greater inshore during Febru- ary and March. Depth group 4 had the highest mean density (1.3/100 m-'^) followed by depth groups 1 and 3 (0.4/100 m'^ each) and depth group 2 (0.1/100 m^). Day-night comparisons proved nonsignificant as a main effect for spot larvae. The average catch rates for all day and night tows were identical (0.5/100 m^). However, the interaction between day-night and horizontal tow type was statistically significant (P < 0.05). In this case, vertical migration and stratification may be indicated. Average catch rate of spot lar- vae during the day at the surface was 0.1/100 m^, while near the bottom it averaged 1.6. Con- versely, nighttime average catch rate at the sur- face was 1.0/100 m-^ while the near-bottom rate averaged 0.04. These are very low densities but the vertical differences are an order of magnitude and their reversing pattern suggests that spot larvae were stratified and undergoing diel verti- cal migration. Daytime bottom and nighttime surface average catch rates were higher than for oblique (O) tows (day, = 0.45/100 m^; night, O = 0.49/100 m^). Average catch rates for surface and near-bottom tows, regardless of time of day, were 0.5 and 0.9/100 m\ respectively. As previously mentioned, other larval sciaenid species (i.e., black drum, banded drum, southern kingfish) were collected during these cruises. However, relatively few individuals were cap- tured (Table 1), making information on their dis- tribution inconclusive. TRANSPORT ANALYSIS Alongshore advection within and just outside the coastal boundary layer in the northwestern Gulf of Mexico has been hypothesized as the major mechanism transporting gulf menhaden larvae to the estuaries in western Louisiana, rather than across-shelf transport from directly offshore. In contrast, such direct across-shelf transport has been demonstrated for sciaenids and other species along the U.S. mid-Atlantic coast (Nelson et al. 1976; Norcross and Austin 1981; Miller et al. 1984). The data collected for sciaenid larvae (all species combined) were exam- ined in light of this Gulf hypothesis. Larval sciaenid densities were less than those for gulf menhaden but similarities in distribution were evident. Both larval sciaenid and gulf menhaden densities were highest at midshelf early in the study. By March and April the highest densities were found towards the east and inshore and were associated with a horizontal density front (coastal boundary layer) caused by an intrusion of fresher water onto the shelf. The along-transect length-frequency patterns exhibited by larval sciaenids and gulf menhaden were also similar. No apparent increase in size was seen until gulf menhaden larvae were on the inner shelf or sciaenids were on the mid- to inner shelf. The expected pattern of a gradual increase in larva size from offshore to inshore, which would result if there were significant across-shelf (south to north) transport, was not evident in ei- ther data set. Off the North Carolina coast, War- len (1981) and Miller et at. (1984) showed that ages and lengths of both spot and Atlantic croaker larvae increased systematically toward shore in an area where winter water currents fa- vored across-shelf (west to east) transport. During the winter of 1981-82, moored current meter data from sites H and S (Fig. 1) indicated that flow was directed primarily alongshore in the west-northwest direction. Several researchers have reviewed the circulation in the northwest- ern Gulf (Nowlin 1971; Kelly et al. 1982; Crout 1983). It was not until Cochrane and Kelly (1986) developed their comprehensive circulation model for the Louisiana-Texas continental shelf, how- ever, that the ocean current patterns, which led to the hypothesized larva transport model, were fully documented. Flow in nearshore coastal waters is westward all year except in summer when it usually reverses, while farther offshore flow is eastward all year (Cochrane and Kelly 1986). To quantify transport, larval sciaenid densities were combined with the vertically averaged, in- stantaneous current measurements. The resul- tant curves present the number of larvae trans- ported per unit time at each station (Fig. 5). Early in the winter, highest sciaenid larva transport (mostly Atlantic croaker and spot) was located midshelf. Later (March and April), transport val- ues were higher inshore and reflected the in- crease in larval sciaenid density (primarily sand seatrout). Overall, larva transport was primarily westward and ranged from about 0.05 to 4.0 lar- vae/meter per second. Although the oceanographic data collected were insufficient to precisely quantify onshore transport rates, an estimate was obtained by using the mean current vectors from the near sur- face meter at site H (Fig. 1) from 24 January to 12 May 1982 (14.33 cm/second alongshore westward 138 COWAN AND SHAW LARVAL SCIAENIDS COLLECTED OFF WEST LOUISIANA U) U1 \ ■« n eg k o I—-— MAR (/) r5Csi»-0'-fMr)'* / 1-' I »- o »- oj r) 09999 t-8,uj avAbvn oc a. < 3 8 8 3 8 3 009990 -I — ' — »■ o 3 CO CO 3 O a> c c CO c 5 uj T3 -^ (1) c r ^ t F c OJ Q) (/> ?. c 11 ^ CO O) - U5 O S < 2 o a. 2 2 <" to II 5 o JO (D t: <2 o ^ Q- 9 y CD 139 FISHERY BULLKTIN: VOL. 86, NO. 1 and 1.75 cm/second shoreward). Based on that av- erage shoreward advection rate we calculated that larvae could be passively transported 98 km in the onshore direction in 65 days. Examination of length frequency and age at capture data (Cowan, in press) suggest that larval Atlantic croaker arrive in nearshore coastal waters, on the average, 60-90 days after hatching. Most small, newly hatched Atlantic croaker larvae were col- lected approximately 100 km offshore. Although the onshore component of advective transport is small in comparison with the average alongshore component, the estimate of shoreward transport rate is reasonable when age of larvae is consid- ered. CONCLUSIONS- RECRUITMENT IMPLICATIONS Across-shelf transport appears to be an order of magnitude smaller than alongshore advective transport in the northwestern Gulf shelf waters during winter and spring. Sciaenid larvae col- lected offshore in the study area, at midshelf and beyond, would probably be lost to the estuaries in western Louisiana. Those offshore larvae would be transported towards north Texas estuaries, or back to the east if they were far enough offshore, since there is evidence for an easterly counter current (Kelly et al. 1982; Cochrane and Kelly 1986). Sand seatrout are common in west Louisiana estuaries (Herke et al. 1984) and were the most abundant sciaenid larvae collected in this study. They spawn, in general, more inshore (Fig. 1) than Atlantic croaker or spot. Conceivably, many of the sand seatrout collected in the study area inside the coastal boundary layer on the inner shelf would have recruited to Louisiana estuaries. Still, large numbers of postlarval sciaenids, other than sand seatrout, enter the estuaries in west Louisiana each year. Atlantic croaker and spot were the 3rd and 21st most abundant fish, of 117 species collected, in the Calcasieu River Basin, the largest estuary in west Louisiana (Herke et al. 1984). However, the distribution and transport analyses indicate that most spot and Atlantic croaker larvae directly offshore at least would not have recruited to the Calcasieu Basin. Interpretation of these data suggests that the source of the sciaenid postlarvae shown to season- ally recruit to the Calcasieu estuaries must be east of the study area. Interpretation of data sum- marizing several years of northern Gulf shrimp- trawl collections suggests that, during the spawn- ing season, a sufficient concentration of adults exists to the east of our study area (Darnell et al. 1983). In the fall and winter, high concentrations of Atlantic croaker, and to a lesser extent spot, have been found between the 20 and 40 m depth con- tours (65 and 125 km offshore) in an area east of the sampling grid. The area and timing of high concentration coincides with the reported spawn- ing location and period for both Atlantic croaker and spot. If indeed this concentration represents a spawning distribution, it would help explain why so few Atlantic croaker and spot larvae, relative to the number of juveniles seen in estuaries, were collected in this and previous Gulf of Mexico ichthyoplankton studies. Unless collections were made in or near the spawning area, single-station densities would be low as eggs and larvae were dispersed. Furthermore, this study demonstrates the need for understanding both biological (verti- cal distribution, age and growth, behavior, etc. of larvae) and physical (ocean currents, estuarine- shelf exchange, etc.) processes which may influ- ence estuarine recruitment. ACKNOWLEDGMENTS We would like to thank Wm. Wiseman, L. Rouse, Jr., and S. Dinnel for their discussion and assistance in interpretation of the physical oceanographic data. We gratefully acknowledge E. Turner, J. Geaghan, M. Fitzimmons, B. Thompson, and W. Herke for critically review- ing this manuscript. Funding was provided by a Louisiana Depart- ment of Wildlife and Fisheries, U.S. Department of Energy and LSU Center for Wetland Resources cooperative agreement No. DE-FC96-81P010313. Additional support was given by the Department of Marine Sciences, Louisiana Sea Grant College Program and the Coastal Fisheries Institute. LITERATURE CITATIONS Barger, L. E., and a. G. Johnson 1980. An evolution of marks on hardparts for age determination of Atlantic croaker, spot, sand seatrout, and silver seatrout. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-SEFC-22, 5 p. Barger, L E , and M. L. Williams. 1980. A summarization and discussion of age and growth of spot, Leiostomus xanthurus Lacepede, sand seatrout, Cynoscion arenarius Ginsburg, and silver seatrout, Cynoscion nothus (Holbrook), based on a literature review. U.S. Dep. Commer., NOAA Tech. Memo. 140 COWAN AND SHAW: LARVAL SCIAENIDS COLLECTED OFF WEST LOUISIANA NMFS-SEFC-14, 15 p. Bearden, C. M. 1964. Distribution and abundance of Atlantic croaker, Micropogon undiilatus, in South Carolina. Bears Bluff Lab. Contrib. 38, 27 p. Benson, N G (editor). 1982. Life history requirements of selected finfish and shellfish in Mississippi Sound and adjacent areas. Biol. Serv. Program, U.S. Dep. Inter., Fish Wildl. Serv. FWS OBS-78/12, 314 p. Cochrane, J D., and F. J Kelly, Jr 1986. Low-frequency circulation on the Texas-Louisiana continental shelf. J. Geophys. Res. 91(C9): 10,645- 10.659. CoWAN, J H. jR 1985. The distribution, transport and age structure of drums (family Sciaenidae) spawned in the winter and early spring in the continental shelf waters off west Louisiana. Ph.D. Thesis, Louisiana State Univ., Baton Rouge, 182 p. In press. Age and growth of Atlantic croaker, Micro- pogonias undulatus. larvae collected in the coastal waters of the northern Gulf of Mexico as determined by increments in saccular otoliths. Bull. Mar. Sci. 42(3). Crout, R L 1983, Wind-driven, near-bottom currents over the west Louisiana inner shelf. Ph.D. Thesis, Louisiana State Univ., Baton Rouge, 117 p, Daniels. K 1977. Description, comparison, and distribution of larvae oi Cynoscion nebulosus and Cynoscion arenarius from the northern Gulf of Mexico. M.S. Thesis, Louisiana State Univ., Baton Rouge, 48 p. Darnell. R N , R E Defenbaugh. and D Moore 1983. Northwestern Gulf shelf bio-atlas; a study of the distribution of demersal fishes and panaeid shrimp on soft bottoms on the continental shelf from the Rio Grande to the Mississippi River Delta. Open File Rep. No. 82-04, Miner. Manage. Serv.. Gulf Mex. OCS Reg. Off., Metairie, LA, 438 p. Dawson, C E 1958, A study of the biology and life history of the spot, Leiostomus xanthurus Lacepede, with special references to South Carolina. Bears Bluff Lab. Contrib. 28, 48 p. Finucane. J H , L A Colllns. L E Barger. and J D McEachran 1979. Environmental studies of the south Texas outer continental shelf. Ichthyoplankton'mackeral eggs and larvae. NOAA Final Rep. to Bur. Land Manage. Wash., D.C., 504 p. Franks. J S.J Y Christmas, W L Siler, R Combs, R Waller, and C Burns 1972. A study of nektonic and benthic faunas as related to some physical, chemical and geological factors. Gulf Res. Rep. 4:1-147. Guest. W. C, and G. Gunter. 1958. The seatrout or weak fishes (genus Cynoscion) of the Gulf of Mexico. Gulf States Mar. Fish. Comm. Tech. Summ. No. 1, 40 p. Gunter. G 1938. Seasonal variations in abundance of certain estuarine and marine fishes in Louisiana, with particular reference to life histories. Ecol. Monogr. 8:313-346. 1945. Studies on marine fishes of Texas. Publ. Inst. Mar, Sci., Univ. Tex. 1:1-190. Haven, D S 1957. Distribution, growth, and availability of juvenile croaker, Micropogon undulatus, in Virginia. Ecology 38:88-97. Herke. W H , B D Rogers, and J. A Grimes. 1984. The study of the seasonal presence, relative abundance, movements, and use of habitat types by estuarine-dependent fishes and economically important decapod crustaceans on the Sabine National Wildlife Refuge. La. Coop. Fish. Res. Unit Final Rep., two vols., 603 p. HiLDEBRAND. S F . AND L E CABLE 1930. Development and life history of fourteen teleostean fishes at Beaufort, N. C. Bull. U.S. Bur. Fish. 46:383- 488. HOESE. H D 1965. Spawning of marine fishes in the Port Aransas, Texas area as determined by the distribution of young and larvae. Ph.D. Thesis, Univ. Texas, Austin, 144 p, HOESE, H. D.. AND R. H MOORE 1977. Fishes of the Gulf of Mexico, Texas, Louisiana and adjacent waters. Texas A&M Univ. Press, College Station, 327 p. HOUDE.E D 1977. Abundance and potential yield of the round herring, Etrumeus teres, and aspects of its early life history in the eastern Gulf of Mexico. Fish. Bull., U.S. 75:61-89. Johnson, G. D. 1978. Development of fishes of the Mid-Atlantic Bight. An atlas of egg, larval and juvenile stages. Vol. IV, Carangidae through Ephippidae. Biol. Serv. Program, U.S. Dep. Inter., Fish Wildl. Serv. FWS/OBS-78/12, 314 p. Joseph, E B 1972. The status of the sciaenid stocks of the middle Atlantic coast. Chesapeake Sci. 13:87-100. Kelly. F J . Jr . J E Schmitz, R E. Randall, and J. D. Cochrane 1982. Physical oceanography. In R. W. Hann, Jr. and R. E. Randall (editors). Evaluation of brine disposal from the Bryan Mound site of the Strategic Petroleum Reserve Program, p. 1-144. Final report of the 18 month post-disposal studies. Texas A&M Univ. Res. Found., College Station. KiLBY, J 1955. The fishes of two Gulf coastal marsh areas of Florida. Tulane Stud. Zool. 2:175-247. Mercer, L P 1984a. A biological and fisheries profile of spotted seatrout, Cynoscion nebulosus. N.C. Dep. Nat. Resour. Community Dev., Div. Mar. Fish., Spec. Sci. Rep. 40, 87 p. 1984b. A biological and fisheries profile of red drum, Sciaenops ocellatus. N.C. Dep. Nat. Resour. Community Dev., Div. Mar. Fish., Spec. Sci. Rep. 41, 89 p. Miller, J M , J P Reed, and L J. Pietrafesa. 1984. Patterns, mechanisms and approaches to the study of migrations of estuarine-dependent fish larvae and juveniles. In J. D. McCleave, G. P. Arnold, J. J. Dodson, and W. H. Neill (editors). Mechanisms of migrations in fishes, p. 209-225. Plenum Publ. Corp., N.Y. Moffet, a W., L W. McEachron, and J. G. Key. 1979. Observations on the biology of sand seatrout (Cynoscion arenarius) in Galveston and Trinity Bays, 141 FISHERY BULLETIN VOL 86, NO. 1 Texas. Contrib. Mar. Sci. 22:163-172. Moore. D . H A Brusher. and L Trent. 1970. Relative abundance, seasonal distribution, and species composition of demersal fishes ofT Louisiana and Texas, 1962-1964. Contrib. Mar. Sci. 15:45-70. Music. J L.Jr 1974. Observations of the spot (Leiostomus xanthurus ) in Georgia's estuarine and close inshore ocean waters. Ga. Dep. Nat. Resour., Coastal Fish. OfT. Contrib. Ser. 28, 29 p. Nelson. W R 1967. Studies on the croaker, Micropogon undulatus Linneaus, and the spot, Leiostomus xanthurus Lacepede in Mobile Bay, Alabama. M.S. Thesis, Univ. Alabama, Tuscaloosa, 85 p. Nelson, W R , N C Ingham, and W E Schaaf 1976. Larval transport and year-class strength of Atlantic menhaden, Brevoortia tyrannus. Fish. Bull., U.S. 75:23-41. NoRCRoss. B L AND H M Austin 1981. Climate scale environmental factors affecting year class fluctuations of Chesapeake Bay croaker Micropogonias undulatus. Va. Inst. Mar. Sci. Spec. Sci. Rep. No. 110, 78 p. NOWLIN, W D, Jr. 1971. Water masses and general circulation of the Gulf of Mexico. Oceanol. Int. 6(2):28-33. PACHECO, a. L 1962. Age and growth of spot in lower Chesapeake Bay, with notes on distribution and abundance of juveniles in the York River system. Chesapeake Sci. 3:18- 28. Pearson, J C 1929. Natural history and conservation of redfish and other commercial sciaenids on the Texas coast. Bull. U.S. Bur. Fish. 44:129-214. Perry, A 1979. Fish of Timbalier Bay and offshore Louisiana environments collected by trawling. Rice Univ. Stud. 65:537-545. POWLES, H , AND B W STENDER 1978. Taxonomic data on the early life history stages of sciaenids of the South Atlantic Bight of the United States. S.C. Mar. Resour. Comm. Tech. Rep. No. 31, 64 p. Sabins, D S . AND F. M Truesdale. 1974. Diel and seasonal occurrence of immature fishes in a Louisiana tidal pass. Proc. Annu. Conf. Southeast. Assoc. Game Fish Comm. 28:161-171. Shaw, R F , J H Cowan, Jr , and T L Tillman 1985a. Distribution and abundance of Brevoortia patronus (gulf menhaden) eggs and larvae in the continental shelf waters of western Louisiana. Bull. Mar. Sci. 36:96-103. Shaw, R F , B D Rogers, J H Cowan, Jr , and W H Herke In press. Ocean-estuary coupling of ichthyoplankton and nekton in the northern Gulf of Mexico. Am. Fish. See. Symp. 3:77-89. Shaw, R F , W J Wiseman, Jr , R E Turner, L J Rouse, and R E CONDREY 1985b. Transport of larval gulf menhaden Brevoortia patronus in continental shelf waters of western Louisiana: a hypothesis. Trans. Am. Fish. Soc. 114:452-460. Shlossman, P a , AND M E Chittenden, Jr. 1981. Reproduction, movements and population dynamics of the sand seatrout, Cynoscion arenarius. Fish. Bull., U.S. 79:649-669, Springer, V G , and K D Woodburn 1960. An ecological study of the fishes of the Tampa Bay area. Fla. Board Conserv. Mar. Lab. Prof. Pap. Ser. No. 1, 104 p. Suttkus, R. D. 1955. Seasonal movements and growth of the Atlantic croaker (Micropogon undulatus ) along the east Louisiana coast. Proc. Gulf Caribb. Fish. Inst. 7:151-158. TOWNSEND, B C , Jr 1956. A study of the spot, Leiostomus xanthurus Lacepede, in Alligator Harbor, Florida. M.S. Thesis, Florida State Univ., Tallahassee, 43 p. Wallace, D H 1940. Sexual development of the croaker, Micropogon undulatus, and distribution of the early stages in Chesapeake Bay. Trans. Am. Fish. Soc. 70:475-482. Warlen, S. M 1981. Age and growth of larvae and spawning time of Atlantic croaker in North Carolina. Proc. Annu. Conf. Southeast. Assoc. Game Fish. Comm. 34:204-214. White, M. L., and M E Chittenden 1977. Age determination, reproduction and population dynamics of the Atlantic croaker, Micropogon undulatus. Fish. Bull., U.S. 75:109-123. Wiseman, W J , Jr , R E Turner, F J Kelly, Jr, L J Rouse, Jr., AND R F Shaw. 1987. Analysis of biological and chemical associations near a turbid coastal front during winter 1982. Contrib. Mar. Sci. 29:141-151. 142 BEHAVIOR OF SOUTHERN RIGHT WHALES, EUBALAENA AUSTRALIS , FEEDING ON THE ANTARCTIC KRILL, EUPHAUSIA SUPERBA William M Hamner,^ Gregory S. Stone,^ and Bryan S. Obst^ ABSTRACT Southern right whales, Eubalaena australis, were observed in 3 successive years on the western side of the Antarctic Peninsula. These whales do not appear to be from the well-documented Valdes, Argentina population. The whales we observed were feeding on Antarctic krill, Euphausia superba. When krill were at the surface right whales surface-skimmed at high speed, with upper jaw lifted above the water surface. In heavy weather one right whale "tail-sailed" at slow speed, with head submerged and apparently feeding. When krill were organized in subsurface schools, right whales engaged in subsurface feeding, diving repeatedly in place, apparently working a given school. One whale swam directionally to the only known large school of krill in the area and fed intensively, rested on the surface, then began a second feeding bout. Whales hyperventilated, false fluked, and fluked prior to feeding dives. These are the first detailed observations of feeding behavior of right whales in Antarctic waters and suggest that coastal Antarctica may have been (and may become again) a regular part of the summer feeding range of the species. Right whales are among the rarest of the great whales, having been hunted almost to extinction a century ago. The southern right whale, Eubal- aena australis, has been studied only recently and only during the austral winter when the whales aggregate inshore to bear calves and to mate (Clarke 1965; Payne 1976, 1986; Best 1981; Aguayo and Torres 1986). Because right whales were commercially extinct by the mid-1850's, very little has been learned about their ecology from the 20th century whaling industry. Informa- tion on feeding, migration, stock structure, and reproductive biology was collected for most other Antarctic mysticete whales during the heyday of whaling in this century (e.g.. Mackintosh 1965; International Whaling Commission reports 1964- present). The small number of surviving right whales (ca. 29c of historic levels in the Southern Hemisphere, Breiwick and Braham 1984) has made it difficult for researchers to study this spe- cies. Our current understanding of its feeding and calving ecology in the Southern Hemisphere comes from observations made primarily off Peninsula Valdes, Argentina (Payne 1986). In Antarctic waters south of lat. 60°S, more than 30 sightings of right whales have been re- ported previously (Berzin and Vladimirov 1981; Goodall and Galeazzi 1986; Ohsumi and Kasa- iDepartment of Biology, University of California, Los Ange- les, CA 90024. 2College of the Atlantic, Bar Harbor, ME 04609. Manuscript accepted September 1987. FISHERY BULLETIN: VOL. 86. NO. 1. 1988. matsu 1986). Most were in the vicinity of the South Orkney Islands, 9 were near the Antarctic Peninsula, 6 were in the Pacific sector of the Antarctic, and 2 were south of Africa. We have observed southern right whales during 3 consecu- tive austral summers near the western shore of the Antarctic Peninsula, and the sightings re- ported herein and in Stone and Hamner (in press) are the most southerly as well as the most de- tailed observations. We sighted one individual during the 1983- 84 austral summer (also recorded by Captain P. Lenie in the log of the RV Hero ; Goodall and Galeazzi 1986), two individuals in 1984-85, and eight in 1985-86, four of which we indi- vidually identified. In 1986 a fortunate combina- tion of fair weather and available ship time permitted us to make the first extended uninter- rupted observations on the behavior of right whales feeding on the Antarctic krill, Euphausia superba. METHODS Right whales are distinguished by the absence of a dorsal fin and regions of cornified skin (cal- losities) on the head, jaws, and chin. Individual whales were identified by standard methods, using video tapes and telephotographs of head callosities and scarring patterns on the head and back (Payne et al. 1983; Kraus et al. 1986). When possible, we dropped large disks of plywood of 143 FISHERY BULLP:TIN: VOL 86, NO 1 known diameter next to the whale and included it in the photograph to measure whale size. Behavioral patterns were observed from the ship's bridge and recorded on a portable computer as they occurred, using a program which timed entries of encoded behaviors or coments to the nearest second. Krill schools were recorded on a Simrad echosounder and the identity of the or- ganisms causing the echograms was verified by net samples taken with an Isaacs-Kidd midwater trawl and by divers' visual confirmation of krill schools near the surface. RESULTS 7 January 1984: At 0500 hour, north of Cape Murray near Two Hummock Island, one right whale was feeding at the surface with its upper jaw lifted above the water, swimming at high speed (estimated at 8-9 knots by the ship's cap- tain) in feeding runs of 15-20 seconds, which we recorded on video. Three humpback whales nearby were diving in one specific location. Th,e right whale repeatedly changed direction be- tween surface runs so that its horizontal direc- tional feeding excursions did not take it far from the vicinity of the vertically diving humpbacks. During these powerful filter-feeding runs enor- mous amounts of water were displaced, cascading beside and behind the right whale and producing a large wake. 15 January 1985: A cow and a calf were swim- ming slowly at the surface some 500 m from the eastern shore of Anvers Island. On approach by the ship the whales swam slowly into shallow water where we could not follow. It was near dusk and we could not get photographs for future iden- tification. 7 January 1986: We encountered one southern right whale and six humpbacks at 1830 hour at lat. 63°46'S, long. 61°13'W, between Trinity and Hoseason Islands. We photographed the head and body of the right whale for subsequent identifica- tion. We followed the whale for approximately 2 hours, recording diving times, surface intervals, and breathing rates. The whale frequently changed directions underwater and consequently we often failed to see the whale immediately when it resurfaced, so breathing rate data for this behavioral sequence are incomplete. The whale appeared to have captured krill on at least one dive because when the whale surfaced it repeat- edly and briefly opened and closed its mouth, with baleen visible, a behavior presumably associated with separation of krill and water prior to swal- lowing the prey (Watkins and Shevill 1976). About 50 cape petrels, Daption capensis, alighted on the water and fed at the surface around the whale. When the whale's jaw movements ceased, the birds soon stopped feeding, but they remained on the water and did not follow the whale when it swam away at the surface. 2 March 1986: We observed one right whale at the northern end of the Neumayer Channel, where a large iceberg was grounded on a 93- fathom rise 2 miles east of Iceberg Point. The wind was blowing from the north at 20-24 knots and a strong surface current was flowing south, producing a bow wave on the grounded iceberg. The right whale repeatedly swam NE of the berg, raised its tail high out of the water at 90° to the wind, submerged its head, and "sailed" downwind past the iceberg, a behavior previously noted for right whales in Argentina (Payne 1976). Soon after we first saw the whale, it stopped tail- sailing and began diving, still along the N-S tran- sect near the iceberg where it had been sailing. The presence of the ship did not cause the whale to alter its back-and-forth swimming rhythm or direction. We waited until the whale began one of its N-S transects past the iceberg and followed about 100 m behind it with the ship. A large school of krill was present on the east side of the iceberg. We recorded no other schools in the vic- nity. The whale was accompanied by three female fur seals. The seals constantly darted about the head of the whale when it surfaced after long dives and appeared to annoy the whale, because several times the whale repeatedly slashed its head sideways when the seals swam too close. 3 March 1986: At 1030 hour, we spotted a single southern right whale near the mouth of Andvord Bay on the Antarctic Peninsula. The whale was swimming SSW at about 3 knots, making short dives that lasted about 19 seconds {N = 12, SD = 9.0 seconds), with brief surface intervals that averaged 6.1 seconds (A^ = 12, SD = 2.6 sec- onds) (Fig. 1). The whale then stopped diving but continued to swim SSW toward the NE tip of Lemaire Island, swimming mostly at the surface for approximately 90 minutes. During this period the whale appeared unconcerned with the ship, which remained 50-100 m behind it, but when the whale neared an iceberg that was hard aground near Lemaire Island, it turned suddenly at a right angle to its prior course and swam be- tween the iceberg and the rocks. The ship was nonetheless able to follow the whale through the 144 HAMNER ET AL : FEEDING BEHAVIOR OF SOUTHERN RIGHT WHALES 30 60 90 120 150 180 1M1 35 65 95 125 155 185 10 15 20 25 1 40 45 50 55 70 75 80 85 U 100 105 no 115 130 135 140 145 160 165 170 175 1 V/ 190 195 200 Time in minutes 205 30 60 90 r 120 150 180 210 Figure 1. — Dive record of the right whale constantly observed for 3.5 hours on 3 March 1986. 44 minutes (arrowl: The ship approached to within 10 m of the whale for i.d. photographs and size measurement. 54 minutes: The whale hyperventilated and dove in an area without krill, then swam northward. 92 minutes: The whale began diving on scattered small krill schools while still traveling N and NNW. 140 minutes: The whale stopped and began diving on one large concentration of krill. 182 minutes: The whale rested on the surface, moving its jaws. 194 minutes: The whale began a second feeding bout ca. 400 m further south. narrow channel. Thereafter the whale ignored the ship and altered course to NNW, still swim- ming at the surface. The ship pulled ahead of the whale to measure its length and this may have caused the short dive noted at ^ = 43 minutes. However, the whale calmly surfaced again within 10 m of the ship and watched us while we pho- tographed it next to the wooden disk. The whale was 11m long. The whale then continued to swim at the surface to the NNW. During the period that the whale swam at the surface without diving the average time between breaths was 50.8 seconds (n = 55; SD = 10.8). At ^ = 54 minutes the whale began to hyperventilate, took three short dives, lifted the flukes partially clear of the water (false fluking I on the third dive, and then fully raised the flukes on the fourth dive, which lasted 210 seconds. The whale then remained at the surface for about 30 minutes while swimming northward. This pattern of hyperventilation prior to a long dive occurred before every long dive sequence which was preceded by a surface interval of at least 4 minutes (Fig. 2). We used this criterion to restrict our dive selection for this plot because there is some indication that there is also a brief recovery period involving hyperventilation after long dives. Of the last 19 breaths that were taken during the 90 seconds before the 5 long dives plot- ted in Figure 2, 18 were less than 30 seconds apart, with a mean interval of 15 seconds (A^ = 19, SD = 7.3). During the time preceding this 90-s hyperventilation period, the whale aver- aged 1 breath every 48 seconds iN - 17, SD = 12.4), not significantly different from the time of 50.8 seconds recorded between breaths during long surface intervals without dives. During hy- perventilation, therefore, breathing rate in- creased by about 3 x . At ^ = 92 minutes we began to record scattered small krill schools on the echosounder and the whale began to dive erratically, with some rea- sonably long dives, but most quite short. Of 23 dives, 13 (56%) were shorter than 10 seconds; the time averaged for all dives was 75.3 seconds 145 FISHERY BULLETIN: VOL 86, NO 1 70 r 3 2 Minutes prior to dive Figure 2. — Breath sequences prior to 5 dives lasting from 93 to 279 seconds. During the 90 seconds immediately preceding dives, the mean interbreath interval was 15 seconds. Prior to this 90-s period of hyperventilation the mean interbreath interval was 48 seconds. (A^ = 20, SD = 74.6). During this period of erratic diving the whale continued to swim N, then NNW. At about t = 140 minutes the whale stopped traveling NNW and began a series of 10 dives, the second being the longest, followed by successively shorter dives. These dives were fol- lowed by rather uniform surface intervals and all 10 lasted significantly longer (x = 183.5 seconds, N = 10, SD = 89.2) than the previous mean of 75 seconds. At the end of the dive sequence, at 182 minutes, the whale stopped swimming entirely and floated motionless at the surface, occasion- ally moving its jaws. This inactivity lasted about 12 minutes and then the whale began a second series of dives. We called these dive sequences, which consisted of regular length surface inter- vals interspersed with a series of longer dives of progressively decreasing duration, "feeding bouts" because at ^ = 140 seconds the whale had reached the only major aggregation of krill in the vicinity, as verified by sonic records (Fig. 3). We saw dense schools of krill and isolated krill at the surface from the bow of the ship, and we captured krill in three successive hauls with the 1 m Isaacs-Kidd midwater trawl. During the first feeding bout the whale slowed from its steady 3-knot swimming speed and moved slowly at about 1 knot, but no longer in any specific direction, finally swimming some 400 m to the south before beginning the second feeding bout. The whale had fed on krill earlier also, because it defecated during the feeding se- quence and the feces, as determined by later mi- croscopic examination, were composed entirely of well-digested euphausiids. The display of a false fluke prior to the high fluke initiating a long dive did not necessarily indicate presence or absence of krill, but when krill were present, it was a highly significant pre- dictor of the length of the dive. In the presence of prey, when the whale showed its flukes once (i.e., did not false-fluke preceding the dive), dive dura- tion averaged 91.2 ± 13.0 seconds {N ^ 5), whereas when both a false fluke and a high fluke preceded the dive, the dive averaged 234.7 ± 64.4 seconds (N = 1). In March and April 1986 we saw a total of eight southern right whales. Of these we distinguished four as individuals on the basis of video tapes and photographs of callosities and body markings. One of these whales was the same individual that we had observed in January some 70 miles fur- 146 HAMNER ET AL : FEEDING BEHAVIOR OF SOUTHERN RIGHT WHALES 0( I/) (V Qjia - ic) X U + ^ dRjt/dKj]a + p)-'] (1) where variable p represents the real price fisher- men expect to receive per unit of output in the Gulf of Mexico shrimp industry, X represents the quantity of shrimp expected to be harvested, Kj is the real stock of thej'^ category of fishing vessels, Tt and P/ represent the tax payment and principal payment due in period t expressed in constant 3The variables without any subscripts are expected at the time the investment is made to remain constant over time. FISHERY BULLETIN: VOL. 86, NO. 1, 1988 151 dollars, D, represents debt outstanding in period t, r is the real rate of interest on debt capital, p is the real after-tax opportunity rate of return on equity capital desired by fishermen, Qj is the real price paid for the^'^ category of fishing vessels at the retail level, a is the proportion of investment financed with equity capital, i^. is the investment tax credit rate, and Rjt represents the real level of replacement investment in the j^^ category of fishing vessels required in period t to offset losses in productive capacity due to wearout. The entire term on the left-hand side of the inequality sign in Equation (1) represents the present value of the additional net cash flows gen- erated by a permanent addition to the^'^ category of fishing vessels. It is assumed that both the in- terest and the principal payments, ((tP/SK^) and r(dD/BK, ), vary over time as further expenditures are required to maintain the productive capacity of this addition to the capital stock at its original level. The right-hand side of Equation (1) repre- sents the initial downpayment minus the invest- ment tax credit plus the present value of all fu- ture cash outlays required to maintain the stock of the j^^ category of fishing vessels at its new level. To maximize the present value of their equity. Gulf shrimp fishermen would continue to add to the stock of the j^^ category of fishing vessels until Equation (1) holds as an equality. Equiva- lently, maximization of the present value of owner equity requires thaf* pidX/dKj) = j^^ a ic + Z - A (1 - iy) (2) where F, represents the present value of the stream of capacity depreciation of they '^^ category of vessels and iy is the average income tax rate. The term Z represents the present value of the stream of after-tax interest payments and princi- pal payments on debt while A represents the present value of the stream of tax depreciation allowances that can be claimed for each dollar of investment as the stock of vessels is maintained at its new level. ^ The right-hand side of this ex- 4Equation (1) as well as the derivation of the implicit rental price of vessels (c) assume that fishermen expect real prices (p) and the marginal physical product of vessels dX/dK to remain at current levels. These and other assumptions which allow us to treat many components of Equation (1) as consoles are consis- tent with those employed in Penson et al. (1981) and Coen (1975). pression thus represents the implicit rental price of the J^^ category of fishing vessels. The concept of the implicit rental price of capi- tal has been widely employed in previous studies of investment behavior as a determinant of the capital stock which firms desire to hold (e.g., Coen 1968, 1975; Penson et al. 1981). Equation (2) sug- gests that the implicit rental price of they*^^ cate- gory of fishing vessels will increase if their pur- chase price, the cost of debt, and equity capital, or income tax rates increase. These effects, however, will be offset to some extent by an increase in the investment tax credit rate, the deductibil- ity of tax depreciation allowances, and interest expenses. Let us assume that output in this industry is a function, in part, of the stock of fishing vessels and that this production relationship is of the Cobb-Douglas form. Letting p^ represent the par- tial production elasticity associated with the stock of the 7'^ category of fishing vessels (Kj), the marginal physical product for these vessels can be expressed as follows: idX/dKj) = i^jiX/Kj). (3) Substituting Equation (3) into Equation (2), the desired stock of they'*^^ category of fishing vessels at the end of year t can be expressed as follows: K*jt^^j{pX/Cjh (4) where Cj represents the expected implicit rental price of they '^'^ category of fishing vessels given by the right-hand of Equation (2). Equation (4) im- plies that the desired stock ofj^^ category of fish- ing vessels is directly related to the expected real gross income from Gulf shrimp fishing and is in- ^The nominal value of Z (the present value of the stream of after-tax loan payments) in Equation (2) is equal to n il -iyW 2j d,(l + ^)''{1 + p)-' i = 1 n + ^ ^ (e - d,)(l -I- (J))-'(l -h p)-' ( = 1 while the real value of A (the present value of the stream of depreciation allowances is equal to i^(hl(p + A> + pi> + ^)) where d^ is equal to the nominal interest payment on a loan of one constant dollar, (J) is the inflation rate, e is the nominal amortized loan payment on a loan of one current dollar, and 8 is the tax depreciation rate given by 2ln where n is the service life of the vessel. 152 versely related to the expected implicit rental price of these vessels. Similar equations could be developed for other inputs used in the shrimp fishing effort. Desired Net Investment in Fishing Vessels New fishing vessels are acquired by Gulf shrimp fishermen both to expand their productive capacity and to replace losses in the productive capacity of existing vessels. This partitioning of observed gross investment into net investment and replacement investment for the j^^ category of fishing vessels can be expressed definitionally as follows: Njt = Kj, - K,, h - ^jt (5) where 7,7 represents the level of real gross invest- ment in the 7'^ category of fishing vessels in year t while Rf is the real replacement investment needed to offset annual capacity depreciation of these vessels. The variables K; and K/- 1 represent the productive capital stock of the 7'^ category of fishing vessels the end and the beginning of the year, respectively. Given Equations (4) and (5), the following relationship between the desired stock of the j^^ category of fishing vessels and current real net investment in these durable inputs can be defined: A^ jt QjiK*j, Kj,-i) (6) where < 0^ < 1 and where 0^ represents the par- tial adjustment coefficient that describes the speed of adjustment of actual stocks to desired levels for thej'^ category of fishing vessels. Sub- stituting Equation (4) into Equation (6) and as- suming an adaptive expectations hypothesis for (pXICj)t, the following compound geometric ex- pression is obtained: A^^, = 0^ \^jUpXICj)t + (1 - ^j)Njt-i + %{l - \j)Kjt-2 0A Jt-l + \^Jt (7) where \j is the adaptive expectations coefficient and [Xji represents the error term. Since Kjt-2 is equal to Kjt^i - Nf-i, Equation (7) reduces to the following estimating equation: A^,, = bjo + bji (pXICj)t + bj2 Kjt-i where 60 is the intercept, 61 = 8 px, 62 = -8X, 63 = (1 - \) (1 - 0) and \xt is once again the ran- dom disturbance term. The estimates of the 61 and 63 coefficients are expected to be positive while the value of 62 is expected to be negative.^ Equation (8) thus represents the general form of the equations to be econometrically investigated in this study. Data The time series data used in this study consist of annual observations for each variable in Equa- tion (8) over the 1965-77 period. This time period represents the only period for which investment expenditure information is available. The productive capital stock of the^'^'^ category of fishing vessels is comprised of a series of differ- ent vintages of vessels or Kj, = Ijt + (1 - hj^)Ijt.^ + (1 - A,i - hj2)\^_^ + . . . + (\-hj^-hj2- .. .- hjr,)Ijt-n (9) where /i^, is the fraction of thej*^^ category of fish- ing vessel's original productive capacity lost in the i^^ year of its service life. The value of A, is represented by (1 - <}>)'" ^, where = 2/n and n is the assumed service life.^ In a related matter, the present value of the stream of capacity de- preciation of a vessel {Fj) was computed as fol- lows: + bj3Njt-i + iXjt (8) 6The net investment model expressed in Equation (8) can be seen as a part of a simultaneous equation system that includes other investment equations as well as supply equations for all inputs and the production function for the fishing industry. The specification of the complete simultaneous system of equations and measurement of time series data needed to simultaneously estimate the 6, coefficients in Equation (8) are beyond the scope of this study. Since the disturbance terms for this set of invest- ment equations are likely correlated, the seemingly unrelated regression equations estimator was employed. The disturbance terms given by this estimator were also examined for autocorre- lation. The estimated rho coefTicient in this small sample was shown to be insignificant in all cases. Finally, the predicted rather than actual value of A'^, _ j was used in estimating Equa- tion (8) to address the pyossibility of correlation between the lagged dependent variable and the disturbance term. ''While a geometric decline in productive capacity has been assumed for fishing vessels, recent studies indicate that the productive capacity of equipment and machinery deteriorates at a lower rate in the early period than in latter years. Coen (1975) suggested that equipment and machinery deteriorate as they age, though not necessarily at a geometric rate. For farm trac- tors, a concave decay pattern represents the best proxy for the capacity depreciation pattern suggested by engineering consid- erations (Penson et al. 1981). The true pattern which underlines actual capital spending decisions in the fishing industry could not be examined due to inadequate data. 153 Fj = ^hj,a + p) (10) (=1 Data from the National Marine Fisheries Ser- vice were used as annual observations on the nominal value of gross investment in Gulf fish- ing vessels (U.S. Department of Commerce 1965- 77^). These values were deflated to real terms using the industrial price index. The quality of the time series for real net investment in fishing vessels, N/, depends on how well the annual values of/, reflect quality changes in vessels over time. The annual levels of the implicit rental price of vessels (c) were computed using the definition outlined on the right-hand side of Equation (2). Coen (1975) assumed that the real after-tax rate of return desired on equity capital, p, is constant over the economic life of the investment. Follow- ing the lead of Coen, a value for p of 5% was employed in this study. The real rate of interest on nonreal estate loans at commercial banks, r, along with annual rate of inflation equals the nominal rate of interest on debt capital. Annual values for all these variables were obtained from US. Department of Com- merce publications. The annual values for the fraction of investment expenditures that are debt financed (v];) used in computing Z were found by dividing the annual change in total debt in the fishing industry by the annual level of gross investment in durable inputs provided by the Na- tional Marine Fisheries Service. The time series for a, the fraction of capital expenditures fi- nanced with internal equity capital, was equal to one minus the percentage debt financed (l-il/). Investment tax credit rate, i^, was equal to 7% during the 1965-68 period, 0% during the 1969-70 period, and I09c during the 1971-77 period. The maximum corporate income tax was assumed to represent i^ for the Gulf shrimp fishery. The double-declining balance method was assumed in determining the present value of the stream of annual tax depreciation allowances in A . The time series data on prices paid for vessels, a component of the rental price, were measured using cost data collected from shrimp vessel builders. Griffin et al. (1978) have shown that vessel length, material of construction, and year of purchase were the most significant factors de- termining the price of a vessel. The equation esti- mated in that study was used to extend available vessel price information over the entire time pe- riod covered by this study. Econometric Results Statistical as well as economic criteria can be employed to evaluate the estimated equations for the various categories of fishing vessels. The eco- nomic criteria include the reasonableness of the elasticities for the economic variables and as well the partial production elasticities implied for the production function. Empirical estimates of the annual real net in- vestment model for steel, wooden, and fiberglass vessels indicate statistically significant coeffi- cients for all but one of the explanatory variables at the 10% level or less (Table 1). The lone excep- tion was the coefficient associated with the lagged capital stock variable in wooden vessel model, which was not significantly different from zero at less than the 20% level. All the coefficients associ- ated with the explanatory variables have the signs hypothesized earlier in this paper. Finally, the coefficients on the lagged dependent variable satisfy the constraint of being both greater than zero and less than one. Table 1 . — Estimated coefficients for the annual net investment model for fishing vessels. Gulf Shrimp Fleet, 1965-77. Vessel Constant (pX/c), K,-: N,-^ type (bo) (bi) {b2) (b3) Steel -95.3895 4.3770 -9.1302 0.6318 M6.21) (6.23) (2.34) Wooden 384.3848 0.1132 -3.1944 0.9089 (1.79) (0.93) (1.21) Fiberglass 0.3819 0.2990 -0.9529 0.2765 (5.37) (1.74) (2.88) 8U.S. Department of Commerce. 1965-77. Vessel charac- teristics data. National Marine Fisheries Service, NOAA, Wash., D.C. 1 Numbers in parentheses indicate absolute values of f -statistic. Economic criteria employed in evaluating the reasonableness of the empirical results and in comparing the investment behavior among vessel types include the partial production elasticity of fishing vessels (p). This elasticity is given by -61/62 which is computed using the estimated beta coefficients in Equation (7). It appears that steel vessels, with a partial production elasticity of 0.479, are highly productive and play an impor- tant role in the supply of fish. The fishing sector, unlike other sectors of the economy, depends pri- 154 marily on one major capital input — vessels. Therefore, the high partial production elasticities recorded for steel vessels and fiberglass vessels (0.314) are no surprise. Even though wooden ves- sels appear to incur more repair and maintenance costs, attract a lower quality crew, and, for that matter, are less efficient than the other vessel types, the low partial production elasticity of 0.033 is surprising. This low partial production elasticity may have been caused by the fact that the instrument for A'^^ - i in the wooden vessels equation used to address the issue of the relation- ship between the lagged dependent variable and the error term did a poor job of explaining Nf - i. An examination of the elasticities associated with the (pX/c )/ term, computed at the mean, re- veals that real net investment in steel vessels is the most sensitive to changes in prices, interest rates, taxes, and the other factors captured in this variable. An elasticity of 7.28 associated with this economic variable was computed for steel vessels. This means that a 1% change in the (pX/c)t vari- able causes real net investment in steel vessels to change by 1.289c. This high investment response to changes in these economic relationships could be attributed to the fact that steel vessels, by far the most productive (as evidenced by the high partial production elasticity reported earlier), are the most durable and the most capital intensive. The elasticity associated with the (pX/c)t term in the fiberglass and wooden vessel equations were 5.35 and 3.11, respectively. This would suggest that macroeconomic policy actions would have a substantially greater effect on real net invest- ment in steel vessels than, say, wooden vessels. Impact of Changes in Cost of Capital The impact of high real interest rates on the growth of selected sectors in the economy has been of great concern in the 1980s. The sensitiv- ity of annual real net investment in fishing ves- sels to changes in the real rate of interest is exam- ined in this section by simulating the estimated equations under annual real rates of 5 and 109?^. In the short run (3 years), an increase in the real rate of interest on debt capital from 5 to 10% would cause real net investment in fishing ves- sels to decrease by 3.04%. Annual real net invest- ment in these fishing vessels would decrease by 15.88% in the long run. As the real rate of interest on debt capital increases, it becomes more diffi- cult to justify the purchase of additional vessels owing to their rising marginal factor cost. Given the fact that 67% of the cost of new fishing vessels is normally financed with debt capital, it is not surprising that rising real interest rates on debt captial have a significant negative effect on the long run expansion of the Gulf fleet. The real cost of equity capital, which reflects the opportunity cost of the fisherman's own funds, has a less dramatic effect on annual real net in- vestment in fishing vessels. This can be at- tributed to the fact that only 33% of the cost of new fishing vessels are financed with equity cap- ital. The short run impact of an increase in the real cost of equity capital from 5 to 10% translates into only a 1.76% decrease in annual real net investment in fishing vessels in the short run. This same change in the cost of equity capital would result in a 12.32% decrease in annual real net investment in the long run. Summary and Conclusions This study evaluated aggregate investment be- havior by fishermen for steel, wooden, and fiber- glass fishing vessels in the Gulf of Mexico shrimp fishery and examined the implications of changes in the cost of acquiring debt and equity capital on the industry's investment response. This study showed statistical justification for the theoretical model of aggregate investment behavior for all three vessel types. It is quite evident that the cost of capital plays an important role in influencing the investment decisions in the Gulf shrimp fishing industry. Macroeconomic policies that lead to high real in- terest rates depress real net investment in this fishery. Capital expenditures for steel vessels are the most sensitive to changes in real interest rates while wooden vessels are the least sensitive. While low real interest rates are desirable for stimulating investment activities in the general economy, they would add to the overcapitaliza- tion problem which currently exists in the Gulf shrimp fishing industry. Finally, this study un- derscores the need to reinitiate efforts to collect data on gross investment expenditures for differ- ent categories of fishing vessels in the Gulf fleet. Literature Cited COEN, R M 1968. Effects of tax policy on investment in manufactur- ing. Am. Econ. Rev. 58:200-211. 1975. Investment behavior, the measurement of deprecia- ble and tax policy. Am. Econ. Rev. 65:59-74. 155 Griffin, W L.J P Nichols, R G Anderson. J E Buckner.and C M Adams 1978. Cost and returns data: Texas shrimp trawlers, Gulf of Mexico, 1974-75. Dep, Agr, Econ. Tech. Rep,, TAMU- SG-76-601, Texas A&M Univ. JUHL. R 1974. Economics of the Gulf of Mexico industrial and food- fish trawlers. Mar. Fish. Rev. 36(ll):39-42. Kmenta. J 1971. Elements of econometrics. MacMillan Company, N.Y. Penson, J B . R G Romain, and D W Huches 1981. Net investment in farm tractors: An econometric analysis. Am. J. Agr. Econ. 63:629-635. Prochaska. F J. andJ C Cato 1981. Economic conditions in the Gulf of Mexico shrimp industry: 1960-81. Food Resour. Econ. Staff Rep., Univ. Florida, Gainesville, FL. Watson, J W . Jr . and C McVea, Jr 1977. Development of a selective shrimp trawl for the southeastern United States shrimp fisheries. Mar. Fish. Rev, 39(101:18-24, Wilson, R R , R G Thompson, and R W Callen, 1970, Optimal investment and financial strategies in shrimp fishing, Dep. Agr. Econ. Tech. Rep, TAMU-SG- 70-218. Texas A&M Univ. John B Penson, Jr, Ernest O, Tetty Wade L Griffin Department of Agricultural Economics Texas A&M University College Station, TX 77843-2124 APPENDAGE II^JURY IN DUNGENESS CRABS, CANCER M AGISTER, IN SOUTHEASTERN ALASKA The Dungeness crab, Cancer magister, is com- mercially important along the western coast of the United States. Like many decapod crus- taceans, it can autotomize and regenerate ap- j>endages to heal wounds and limit injury. Studies of appendage injury may be useful in assessing the physical condition of crustacean populations and the impact of fishing on commer- cially important species. Incidences of appendage loss in the field have been reported for species of crabs other than C. magister (McVean 1976; McVean and Findlay 1979; Needham 1953). Ap- pendage loss was studied in adult Dungeness crabs in Washington (Cleaver 1949) and Oregon (Waldron 1958) and for juvenile crabs in the Co- lumbia River estuary (Durkin et al. 1984). In this study we examined adult Dungeness crabs in southeastern Alaska to determine the incidence of missing, regenerating, and damaged appendages. Temporal incidence of appendage in- jury was compared to the molting and mating periods of the crabs and to the commercial fishing season for Dungeness crabs. Materials and Methods Adult Dungeness crabs were collected from Icy Strait and the Excursion Inlet fjord near Glacier Bay, AK (lat. 135°30'N, long. 58°25'W), from May through November 1984-85. Data were obtained by monthly surveys of commercially caught crabs. Crab pots (Waldron 1958) were set at depths of 7 to 20 m and remained in the water for 3 to 11 days. All crabs were held in live tanks (<24 hours) before they were measured on board ship. In southeastern Alaska, pots are equipped with escape rings to permit release of crabs with carapace widths <165 mm, but sublegal-sized crabs are often found in the catch. Carapace width (excluding the 10th anterolat- eral spines), wet weight, and sex were recorded for each crab. Carapace condition was graded as soft-shell (recent molt), new-shell, worn-shell, or skip-molt (Somerton and Macintosh 1983). The number and identities of missing, damaged, or regenerating chelipeds and walking legs were recorded. An appendage with a cracked cuticle or missing dactyl was considered damaged. Ap- pendages smaller in length and diameter than intact appendages were designated regenerating. Combined missing, damaged, and regener- ating appendages are referred to as injured appendages. Interrelationships between variables were de- termined with Pearson correlations (SAS 1985). Means were compared with Student's ^ -tests, and chi-square analyses were used to determine if multiple autotomies occurred by chance (Steel and Torrie 1960). Data are presented as means ± 1 standard error of the mean. Results Males comprised 65% and females 35% of the 878 Dungeness crabs examined. Average cara- pace widths were 169 ± 0.6 and 159 ± 0.7 mm for males and females, respectively. Wet weights were 1,102 ± 9 g for males and 884 ± 14 g for females. The greatest number of female crabs was caught in July, and the greatest number of males in August. 156 FISHERY BULLETIN; VOL 86, NO, 1, 1988. Seventy-five percent of the crabs were intact, with no appendage injuries. Twenty-five percent of all crabs had injured limbs; 18% had missing, 5% had regenerating, and 2% had damaged ap- pendages. No relationship existed between cara- pace width and appendage injury. Most of the Dungeness crabs sampled were in the worn-shell condition (67%). Twenty-eight per- cent of all crabs were new-shelled. Only 1% were soft-shelled and 4% were skip-molts. Correlations between carapace condition and appendage in- jury were not significant. No significant differences existed in appendage injury between male and female Dungeness crabs. Injuries were bilaterally symmetrical ex- cept for the 3d walking leg which was missing more frequently on the left side (P < 0.05). Con- sidering only those crabs with missing legs, a total of 246 legs were missing with a mean of 1.5 ±0.1 missing legs/crab. Ninety-seven crabs had legs missing on the right side and 98 had legs missing on the left side. The maximum number of missing legs per crab was 5. Sixty percent of the crabs had 1 leg missing, 20% had 2 missing legs, and 12% were missing 3 or more legs. Of the crabs with missing legs, 63% were males and 37% were females. Forty Dungeness crabs had regenerating legs, with a mean of 1.2 ± 0.1 regenerating legs per crab. Sixty percent of those crabs had 1 regenerat- ing leg, 10% had 2 regenerating legs, and 3% had 3 or more regenerating legs. The maximum num- ber of regenerating legs per crab was 4. Of the crabs with regenerating legs, 73% were males and 27% were females. Seventeen crabs had damaged appendages with a mean of 1.1 ± 0.1 damaged appendages/crab. Of the crabs with damaged appendages, 82% were males and 18% were females. The observed number of Dungeness crabs with 2 or more missing appendages was significantly higher (P <0.01) than expected for both sexes, indicating that appendage loss was not due only to chance. Appendage injury was significantly correlated with date, with more injuries occurring later in the year. The number of Dungeness crabs with missing appendages was significantly correlated with date for both males and females (P < 0.01). The lowest percentage of crabs with injured ap- pendages occurred in July (4.8%, both sexes com- bined) and increased to a maximum of 34.3% in November. The percentage of male crabs with re- generating appendages did not vary significantly over time and was about 6% for all months. How- ever, the percentage of female crabs with regener- ating legs increased from 0% in May to 10.5% in November (P < 0.01). Male crabs with damaged appendages increased from 0% to 8.5% from May to October (P < 0.05) and then decreased to 1.7% in November. Chelipeds and 1st and 4th walking legs were injured most frequently. The hierarchy for fre- quency of injury for female crabs (chelipeds > 4th walking legs > 1st, 2d, and 3d walking legs) dif- fered slightly from the hierarchy for males (che- lipeds > 1st walking legs > 4th walking legs > 2d and 3d walking legs). Months in which high per- centages of crabs had injured chelipeds also had high percentages with injured 1st (males) and 4th (females) walking legs (Fig. 1). The temporal incidence of appendage injury in Dungeness crabs was compared to life history events and to the commercial crab fishing season in southeastern Alaska (Fig. 2). The season opened 15 June and closed 15 August, reopened 1 October and closed 28 February 1986. Ap- pendage injuries were low in July and increased 157% from July to August, a period of simulta- neous molting, mating, and fishing. An addi- tional increase in appendage injury of 43% oc- curred in October, even though the fishery was closed from 16 August to 30 September. Discussion Pot samples are biased towards larger sized Dungeness crabs because of the size of the mesh on the pot and the presence of two escape rings with diameters of 11 cm. However, 62% of the crabs collected for this study were either male crabs with carapace widths <165 mm or were females. Very few soft-shell crabs were caught, even though molting was occurring during part of the sampling period. Our dependence on commer- cial crabbers for data collection restricted us to sampling mostly during the open fishing season when most of the crabs were not in the soft- shelled condition. Twenty-five percent of the Dungeness crabs sampled in southeastern Alaska had appendage injuries. In other studies of Dungeness crabs in Washington, Oregon, and the Columbia River es- tuary, 18%, 32%, and 62%, respectively, of the crabs were injured (Cleaver 1949; Waldron 1958; Durkin et al. 1984). The crabs examined in our study were held for up to 24 hours in crowded tanks on board ship before being measured and 157 Appendage Injury in Male Crabs 30 25 in 01 oi (O c 01 a. a. < B A ppendage Injury in Fe m ale Crabs m 01 C31 ID T3 C 01 Q 30 Figure 1. — Monthly percentages of male (A) and female (B) Dungeness crabs with injured (missing + regenerating + damaged) chelipeds and walking legs. •o 9 60- 50- 4 0- 30- 20- 1 0- FISHING d" MOLTING 9 MOLTING J FISHIlie 1 15 31 May —I 1 — T"^ — I n — I 1 1 15 30 15 31 15 31 15 30 15 31 Jun Jul Aug Sep Oct s s s S s s s N s s s \ \ \ \ 15 30 Nov Males l\\l Females Figure 2. — Temporal relationships of percentages of male and female Dun- geness crabs with injured (missing + regenerating + damaged) append- ages to the crabs' molting and mating periods and the commercial crab fishing season in southeastern Alaska. were sometimes observed grasping other crabs, but very few autotomized limbs were found in the tanks. The estimate of appendage injury may be low if Dungeness crabs with injured appendages were less likely to enter pots than intact crabs. In an- other study, the observed number of Carcinus maenas missing 2 or more legs was higher than expected if multiple autotomies occurred by chance, which was interpreted to mean that in- jured C. maenas enter pots as readily as intact crabs (McVean 1976). Because there were more Dungeness crabs with 2 or more missing legs in our collections than would be expected if multiple 158 autotomies occurred by chance alone, our data could be similarly interpreted to suggest that Dungeness crabs with injured appendages showed little decrease in pot-entering ability and that our estimate of injury was accurate. For this interpretation of the chi-square results to be valid, one must assume that all injuries occurred before the crabs entered pots and that injury did not occur within the pots. Appendage injury in Dungeness crabs was bi- laterally symmetrical except for the 3d walking leg. Interestingly, Easton (1972) demonstrated that 3d walking legs of Hemigrapsus oregonensis were the most easily autotomized. In other stud- ies, both bilateral symmetry and asymmetry have been reported for different species of crabs (Durkin et al. 1984; Needham 1953). Asymmetri- cal appendage loss has been associated with crabs that move predominantly in one direction, while symmetrical leg loss occurs in crabs that move randomly (Needham 1953). The chelipeds, followed by the 1st and 4th walking legs, were most vulnerable to injury. Limb loss has been correlated in other studies with degree of exposure of the limb; the outermost limbs, the longest limbs and limbs with postures that afford little protection are most frequently lost (Needham 1953). Anterior limbs are lost more frequently than posterior limbs (Needham 1953). The chelipeds are the most anterior and one of the most exposed appendages on Dunge- ness crabs and are frequently used in aggressive threat postures. After loss of chelipeds, the 1st walking legs remain as the most anterior, ex- posed limbs and therefore, the most vulnerable. The 4th walking legs are the most posterior and also very exposed limbs on an intact crab. A significantly greater number of Dungeness crabs with 2 or more missing legs was observed than expected if multiple autotomies occurred by chance, indicating an increased susceptibility to subsequent appendage loss after initial injury (Needham 1953; Easton 1972). The correlations between appendage injury and date were significant but may not be biologically important. Although these correlations were sig- nificant, the r^ values (square of the coefficient of variation) were low. Increased appendage injury later in the year may be related to other factors. Soak times, the length of time pots were left in the water, were longer later in the year. Dunge- ness crabs may cannibalize other crabs while con- fined in pots (Waldron 1958). There may also be delays between time of injury and subsequent au- totomy and regeneration. Regeneration of legs in Dungeness crabs is usually completed after 2 or 3 molts (Cleaver 1949). Over the sampling period, only b^c of all crabs had regenerating appendages while 187^ had missing appendages. The discrepancy may be due to increased mortality of the crabs following in- jury (McVean and Findlay 1979), or by the effi- cient, yearly removal of legal-sized, injured crabs by the commercial fishery. When temporal incidence of appendage injury was compared with the opening and closing of the commerical Dungeness crab fishing season, con- siderable appendage injury occurred when the fishery was closed. Closure of the fishery tradi- tionally occurs during the crabs' mating period, when a high percentage of soft-shelled female crabs are present in the population. There was, however, some overlap in late July and early August in fishing, molting, and mating. Exclud- ing damage by humans, potential causes of ap- pendage injury are aggression between males competing for females, the cheliped-to-cheliped mating embrace of Dungeness crabs that can last up to a week, cannibalism, and the increased vul- nerability of females which molt prior to mating (Butler 1960; Durkin et al. 1984). Damage to Dungeness crabs can also result from other fish- ing gear such as trawls (Reilly 1983), but no other commercial fisheries occurred in the study area while the Dungeness crab fishery was closed. The results of our studies indicate that Dunge- ness crabs in southeastern Alaska are in com- parable condition to adult populations of Dunge- ness crabs examined in Washington and Oregon, in terms of appendage injury. Further studies are needed to investigate the effect of appendage in- jury on survival of Dungeness crabs and the con- tribution of handling injury and mortality of crabs in the commercial fishery. Acknowledgements T. Meyers and D. Erickson provided technical assistance for which we are very appreciative. We would like to thank T. Olsen, C. Kondzela, and D. Sterritt for assistance in data collection. This research was funded by Alaska Sea Grant project R/06-20. Literature Cited Butler, T. H. 1960. Maturity and breeding of the Pacific edible crab, 159 Cancer magister Dana. J. Fish. Res. Board Can. 17:641- 646. Cleaver, F. C. 1949. Preliminary results of the coastal crab {Cancer magister) investigation. Wash. Dep. Fish., Biol. Rep. No. 49A, 82 p. DuRKiN, J. T., K. D. Buchanan, and T. H. Blahm. 1984. Dungeness crab leg loss in the Columbia River estu- ary. Mar. Fish. Rev. 46(l):22-24. Easton, D. M. 1972. Autotomy of walking legs in the Pacific Shore crab Hemigrapsus oregonensis. Mar. Behav. Physiol. 1:209- 217. McVean, a. 1976. The incidence of autotomy in Carcinus maenas (L.) J. Exp. Mar. Biol. Ecol. 24:177-187. McVean, A., and I. Findlay. 1979. The incidence of autotomy in an estuarine popula- tion of the crab Carcinus maenas. J. Mar. Biol. Assoc. U.K. 59:341-354. Needham, a. E. 1953. The incidence and adaptive value of autotomy and of regeneration in Crustacea. Proc. Zool. Soc. Lond. 123:111-122. Reilly, p. N. 1983. Effects of commercial trawling on Dungeness crab survival. P. W. Wild and R. N. Tasto (editors), In Life history, environment, and mariculture studies of the Dungeness crab. Cancer magister, with emphasis on the central California fishery resource, p. 165-174. Calif Dep. Fish Game Fish Bull. 172. SAS Institute, Inc. 1985. SAS user's guide: statistics. 5 ed. Cary, NC, 956 P- Somerton, D. a., and R. A. MACINTOSH. 1983. Weight-size relationships for three populations in Alaska of the blue king crab, Paralithodes platypus (Brandt, 1850) (Decapoda, Lithodidae). Crustaceana 45:169-175. Steel, R. G. D., and J. H. Torrie. 1960. Principles and procedures of statistics with special reference to the biological sciences. McGraw-Hill Book Co., N.Y. 481 p. Waldron, K. D. 1958. The fishery and biology of the Dungeness crab (Cancer magister Dana) in Oregon waters. Fish Comm. Oreg., Contrib. 24, 43 p. Susan M. Shirley Thomas C. Shirley Juneau Center for Fisheries and Ocean Sciences University of Alaska- Fairbanks 11120 Glacier Highway Juneau. AK 99801 REEXAMINATION OF THE USE OF OTOLITH NUCLEAR DIMENSIONS TO IDENTIFY JUVENILE ANADROMOUS AND NONANADROMOUS RAINBOW TROUT, SALMO GAJRDNERI^ Otoliths are a potential source of taxonomic char- acteristics for identifying stocks offish (Ihssen et al. 1981). Differences in dimensions of the otolith nucleus have provided a basis for separating win- ter from summer races of steelhead, anadromous rainbow trout, Salmo gairdneri. In addition, otoliths provided data from which to distinguish steelhead from resident nonanadromous forms as well (McKern et al. 1974; Rybock et al. 1975). Neilson et al. (1985) studied the development of sagittal otoliths in resident rainbow trout and steelhead from south-central British Columbia, and identified sources of variability in the size of otolith nuclei. However, they were unable to find morphometric differences between the two forms of trout. They concluded that the usefulness of dimensions of the otolith nucleus for separating steelhead from resident rainbow trout was much more limited than that suggested by Rybock et al. (1975) for rainbow trout in the Deschutes River, Oregon. The difference in mean length of the otolith nuclei between the rainbow trout studied by Ry- bock et al. (1975) and those studied by Neilson et al. (1985) suggested either population differences or differences in defining the nuclear boundary. These disparate results, which led to opposite con- clusions, limit the usefulness of measurements of otolith nuclei for the racial identification of juve- nile rainbow trout until the source of these differ- ences is better understood. Consequently, to de- termine whether juveniles of the two forms could be distinguished by differences in dimensions of otolith nuclei, we measured the nuclei in sagittae from steelhead and resident rainbow trout col- lected from the same Deschutes River, OR, loca- tions used by Rybock et al. (1975). We used the definitions proposed by Rybock et al. and by Neil- son et al. (1985), and compared our measure- ments for the two forms with each other and with published values. Methods Resident rainbow trout and steelhead were col- iQregon State University Agricultural Experiment Station Technical Paper No. 8279. 160 FISHERY BULLETIN: VOL. 86, NO. 1, 1988. lected from three locations in the Deschutes River, OR. Resident rainbow trout, which were collected from the main stem near the mouth of Nena Creek in March 1985, were mature and smaller (280-450 mm FL) than the steelhead, and, based on analyses of scales and otoliths (McKern et al. 1974), had never entered salt- water. Juvenile progeny of steelhead were col- lected from Round Butte Hatchery on the Deschutes River in 1984. Wild juvenile rainbow trout (<200 mm FL) of unknown parental origin were collected in 1984 and 1985 from Bakeoven Creek, an important spawning tributary for steel- head in the Deschutes River. Sagittae removed from rainbow trout were stored in 90"^^ ethanol for up to two months. Be- fore they were viewed, one otolith from each pair was mounted (concave face up) with epoxy on a glass slide. The back of the slide was blackened with indelible ink. The otolith was ground by hand with 600 grit wet sandpaper and periodi- cally inspected under a light microscope at 100 x until the microstructure of the nucleus, as de- scribed by Neilson et al. (1985), was visible. The otolith was rinsed with 59f HCl for several sec- onds to remove scratches and improve resolution. To reduce bias, we coded each slide with a ran- dom number and ordered the slides sequentially for viewing. Otoliths were examined with a Zeiss^ dissecting microscope at 125 x. A camera lucida attachment enabled us to use a computer digitizer to measure three dimensions of the otolith. In measuring length and width of the central nu- cleus, we used the first growth increment encir- cling all the central primordia, which was the nuclear boundary defined by Neilson et al. (1985). In addition, we measured the maximum length along the longest axis through an area defined by the first metamorphic check, a narrow hyaline ring surrounding an opaque ring with a hyaline center, to replicate the measurements of Rybock et al. (1975). We used analysis of variance (ANOVA) to test for significant differences in each dimension of the otolith nuclei among groups in our study. Where adequate data were available, we tested for significant differences between groups in our study and similar groups described by Rybock et al. (1975) and Neilson et al. (1985) for mean di- mensions of otolith nuclei. Neilson et al. (1985) showed that the mean length of otolith nuclei for rainbow trout incubated at 6.5°C was signifi- cantly less than those for trout incubated at 9.5° or 15.0°C. Because of this discrepancy, we evalu- ated the potentially confounding effects of incuba- tion temperature on the comparisons of otolith dimensions between our samples and those of Ry- bock et al. (1975), by testing the hypothesis that water temperatures during 1967-69 were higher than those during 1982-83. We used a paired t- test of average daily water temperatures recorded by the U.S. Geological Survey on the 1st and 15th day of each month from 1 January to 1 August during 1967-69 and 1982-83 (U.S. Department of the Interior Geological Survey 1967, 1968, 1969, 1982, 1983 1. These dates represent the incu- bation periods for most of the resident rainbow and steelhead trout sampled in our study and by Rybock et al. (1975). Incubation temperature for steelhead at Round Butte Hatchery is from hatch- ery records. We estimated spawning and incuba- tion periods for resident rainbow and steelhead trout on the basis of reports of the Oregon Depart- ment of Fish and Wildlife (Fessler 1972) and per- sonal observations. Results For each dimension, we failed to reject the hy- pothesis (a = 0.05) that rainbow trout collected from different populations for our study had otolith nuclei of the same size (Table 1). There- fore, we concluded that these dimensions could not be used to discriminate between the resident and steelhead forms of rainbow trout sampled in our study. Water temperatures during 1967-69 were slightly greater than those during 1982-83 (t = 2.03, df = 14, P = 0.03). Mean difference be- tween the two periods was 0.8°C. Spawning dates for resident rainbow trout and steelhead differ; steelhead spawn from January to April and resi- Table 1. — Means, standard errors (in parentheses), and sample size for three otolith dimensions in resident rainbow trout and steel- head from three Deschutes River populations. 2Reference to trade name does not imply endorsement by the National Marine Fisheries Service, NOAA. Dimensions of nuclei Nucleus Nucleus Check Populations No of length width length compared fish (mm) (mm) (mm) Resident 44 0.173 0.070 0.323 rainbow trout (0.006) (0.003) (0.012) Hatchery 30 0.190 0.070 0.349 steelhead (0.006) (0.002) (0.009) Suspected 32 0.178 0.069 0.312 wild steelhead (0.006) (0.002) (0.007) 161 dent rainbow trout spawn from May to mid-July (Fessler 1972). Mean water temperature during the period of steelhead egg incubation was 8.4°C for 1967-69 and 7.6°C for 1982-83. Mean water temperature during the period when resident rainbow trout eggs were incubating in the main stem of the river was 12.6°C in 1967-69 and 11.9°C in 1982-83. Incubation temperature for steelhead at Round Butte Hatchery was 10°C and did not vary. The dimensions of otolith nuclei from resident rainbow trout and steelhead in our study were indistinguishable from those in fish from British Columbia. No significant difference (a = 0.05) in mean length of otolith nuclei existed between the British Columbia steelhead incubated at 9.5° or 15°C and suspected wild steelhead from Bakeoven Creek or Round Butte Hatchery steelhead incu- bated at 10°C. Among resident rainbow trout, the mean length of otolith nuclei for fish from the Deschutes River was also not significantly differ- ent from that for fish from British Columbia incu- bated at 9.5° or 15°C. Because Rybock et al. ( 1975) did not provide variances, we were unable to test the hypothesis that means from our study coin- cided with theirs. However, mean length and width of otolith nuclei in our study were 29 and 559c less, respectively, for resident rainbow trout and 49-70% less, respectively, for steelhead than those studied by Rybock et al. (1975). Discussion The similarity of our results to those of Neilson et al. (1985), who used similar methods, might be expected for different populations under similar genetic and environmental control. The disparate results of our study and that of Rybock et al. (1975) for the same populations after little ge- netic change (based on comparisons of unpub- lished, biochemical genetic data for these popula- tions from 1972 to 1974 and 1984 to 1986) and little environmental change partly reflected the use of different definitions for the nucleus. We defined the nuclear boundary as the first growth ring surrounding all the fused primordia, whereas Rybock et al. (1975) defined the nucleus as the hyaline area in the center of the otolith that is bounded by a metamorphic check formed at hatching; they resolved the check by rendering the otolith with HCl. We also measured the length of the check surrounding the nucleus, as- defined by Rybock et al. (1975), which we found either to correspond with the area enclosed by the first check or to increase in density of growth increments surrounding both the central and ros- tral primordia. The close similarity between our estimate for Round Butte Hatchery steelhead (0.349 mm) and the mean calculated by Rybock et al. (1975) for steelhead (0.354 mm) suggested sim- ilar checks. It is unclear, however, why values for resident rainbow trout for this dimension and the results of tests to discriminate races differed be- tween the two studies. Rybock (1973) noted that the nuclear check could not be distinguished in 29% of the otoliths and that the use of HCl may have caused the frequent confusion between the metamorphic check and other groups of daily growth rings. The grinding and polishing of otoliths greatly reduce this source of error. Neil- son et al. (1985) also discouraged the use of meta- morphic checks as boundaries because the causal links between checks and developmental events, such as hatching, have not yet been established. Neilson et al. (1985) demonstrated that nuclear length increased significantly with increase in in- cubation temperature from 6.5° to 9.5°C but not from 9.5° to 15°C. Although average water tem- peratures in the Deschutes River were 0.8°C lower during 1982-83 than in 1967-69, it is un- likely that such differences completely explain the greater estimates of mean length and width of otolith nuclei in the earlier study by Rybock et al. (1975). Rybock et al. (1975) calculated mean nu- clear lengths and widths of 0.354 and 0.230 mm for steelhead and 0.243 and 0.154 mm for resident rainbow trout in the Deschutes River. Our esti- mates were 29-70% less than their estimates for a 0.8°C difference; whereas under controlled con- ditions in British Columbia, mean nuclear length for resident rainbow trout at 6.5°C was 18% less for resident rainbow trout and 21% less for steel- head than the nuclear length for fish incubated at 9.5°C, a difference of 3°C (Neilson et al. 1985). Comparisons of otolith nuclear dimensions be- tween resident rainbow trout and steelhead incu- bated at similar temperatures would establish whether significant differences exist for these measurements between the two races from the Deschutes River. The use of a common definition of nuclear boundaries would allow better com- parisons between studies. However, given the dis- parate results of our study, which were similar to the results of Neilson et al. (1985), and the origi- nal study for steelhead and resident rainbow trout in the Deschutes River, as well as our fail- ure to discriminate between races using both nu- clear definitions proposed by Neilson et al. (1985) 162 and Rybock et al. (1975), we believe that popula- tion differences do not explain the differences in results between the studies of Rybock et al. (1975) and Neilson et al. (1985). Furthermore, our study provided strong evidence to support the conclu- sion of Neilson et al. (1985) that the usefulness of measurements of otolith nuclei to identify sym- patric juvenile progeny of resident rainbow trout and steelhead reared in the wild may be limited. Acknowledgments We thank Jeff Light for his advice on grinding and polishing otoliths to resolve their nuclear di- mensions and Eric Volk for his review of this manuscript. This research was funded by Bonneville Power Administration, U.S. Depart- ment of Energy, Agreement No. DE-A179- 83BP13499. Literature Cited Fessler, J L. 1972. An ecological and fish cultural study of summer steelhead in the Deschutes River, Oregon. Fed. Aid Fish. Prog. Rep. Proj. No. F-88-R-1. Oreg. State Game Comm., Portland, OR, 47 p. IHSSEN. P E . H E BooKE. J M Casselman, J M McGlade, N. R. Payne, and F M Utter 1981. Stock identification: materials and methods. Can. J. Fish. Aquat. Sci. 38:1838-1855. McKern. J. L., H. F HoRTON, and K V. Koski 1974. Development of steelhead trout (Salmo gairdneri) otoliths and their use for age analysis and for separating summer from winter races and wild from hatchery stocks. J. Fish. Res. Board Can. 31:1420-1426. Neilson. J D . G L Geen, and B Chan 1985. Variability in dimensions of salmonid otolith nu- clei: implications for stock identification and microstruc- ture interpretation. Fish. Bull., U.S. 83:81-89. Rybock. J. T 1973. Use of otoliths to differentiate juvenile steelhead trout from juvenile rainbow trout in the lower Deschutes River, Oregon. M.S. Thesis, Oregon State University, Corvallis, OR, 44 p. Rybock. J T , H F Horton, and J L Fessler. 1975. Use of otoliths to separate juvenile steelhead trout from juvenile rainbow trout. Fish. Bull., U.S. 73:654- 659. United States Department of the Interior Geological Survey 1967. Water resources data for Oregon, Pt. 2. U.S. Dep. Inter., Geol. Surv., Water quality records, 163 p. 1968. Water resources data for Oregon, Pt. 2. U.S. Dep. Inter., Geol. Surv., Water quality records, 145 p. 1969. Water resources data for Oregon, Pt. 2. U.S. Dep. Water quality records, 137 p. 1982. Water resources data for Oregon water year 1982. U.S. Dep. Inter., Geol. Surv., Eastern Oregon, Vol. 1, 206 p. 1983. Water resources data for Oregon water year 1983. U.S. Dep. Inter., Geol. Surv., Eastern Oregon, Vol. 1, 202 p. Kenneth P. Currens Carl B Schreck Hiram W. Li Oregon Cooperative Fishery Research Unit Oregon State University Corvallis. OR 973313 3Cooperators are Oregon State University, Oregon Depart- ment of Fish and Wildlife, and U.S. Fish and Wildlife Service. AGE-SPECIFIC VULNERABILITY OF PACIFIC SARDINE, SARDINOPS SAGAX, LARVAE TO PREDATION BY NORTHERN ANCHOVY, ENGRAULIS MORDAX To a large degree interannual variability in re- cruitment determines the size of pelagic fish pop- ulations. Recruitment to the Pacific sardine, Sardinops sagax, population off California varies from year to year over several orders of magni- tude and is unrelated to spawning stock size (Murphy 1966; MacCall 1979). Variable mortal- ity rates in the first year of life must determine year-class strength, although the sources of this variability are unknown. Mortality rates in the earliest stages are size specific with highest rates in the egg and yolk-sac stage (Ahlstrom 1954; Butler 1987) and may contribute to variability in year-class strength (Smith 1985). The sources of mortality of sardine larvae have yet to be investigated. In other pelagic larvae, mortality is due to either starvation or predation, and starvation is significant only during the brief period after the onset of feeding (O'Connell 1980; Hewitt et al. 1985; Theilacker 1986; Owen et al. 1987). In sardines, significant mortality occurs during the egg and yolk-sac stages (Ahlstrom 1954) and this mortality can only be due to preda- tion. Variable mortality in older larval and juve- nile sardines may also contribute to variability in recruitment, and this mortality, as in other fishes, may also be due to predation (Hunter 1984). The objective of this paper was to determine the size-specific vulnerability of Pacfiic sardine lar- vae to predation by adult northern anchovies, Engraulis mordax. The vulnerability of cape anchovy and northern anchovy larvae to FISHERY BULLETIN: VOL. 86, NO. 1, 1988. 163 cannibalism has been investigated by Brownell (1985) and Folkvord and Hunter ( 1986) and found to be an important source of mortality. In this paper the vulnerability of sardine larvae will be compared with that of anchovy larvae and differ- ences in the biology of sardines and anchovies will be discussed. Our approach was to observe the avoidance be- havior of Pacific sardine larvae in response to predatory attacks by northern anchovy adults. Adult northern anchovy were chosen as a preda- tor because the northern anchovy was the most abundant pelagic fish in the California Current region during the waning years of the sardine fishery and because its planktivorous diet in- cludes fish eggs and larvae (Loukashkin 1970; Hunter and Kimbrell 1980). Materials and Methods Experimental Fishes The Pacific sardine larvae used in the experi- ments were reared from eggs spawned in the lab- oratory. Adult Pacific sardines were collected off San Diego and held in 175 m"^ aquarium for six months. Males and females with developing go- nads were isolated in spawning tanks and in- jected with 250 mg human chorionic go- nadotropin and on the following day injected with 200 units pregnant mare serum and 20 mg salmon pituitary extract. On the third day fertil- ized eggs were collected from the spawning tank. Larval rearing procedures follow those described by Hunter (1976). Temperature in the rearing tanks was maintained at 21°C. Apparatus and Procedures Experimental apparatus and procedures were the same as those described by Folkvord and Hunter (1986) but will be briefly outlined here. Experimental predators were two groups of 5 adult northern anchovy (range of standard lengths 8.4-9.2 cm). Predators were maintained in two rectangular fiberglass tanks (0.75 x 2.15 X 0.83 m = 1.35 m"^) and fed adult brine shrimp except on days of experimental observation. Sea- water was supplied continuously to the tanks ex- cept during experiments. The, temperature in the observation tank ranged from 16.2° to 22.8°C (mean = 20.1°C). Two 100 W incandescent lamps produced 2,000-3,000 mc at the surface of each tank. A black plastic tent enclosing a window on one side of the tank provided a darkened observa- tion chamber. Each trial consisted of the encounter of three prey with the predators. Prey were introduced into the observation tank with a clear glass beaker. Initial feeding behavior of the predators is quite variable but becomes less variable as the predators become accustomed to prey in the tank. For this reason, prior to each experiment adult Artemia were introduced as prey for five consecu- tive trials to standardize predator behavior. After the preliminary trials with Artemia, three trials with sardine larvae were alternated with one trial with brine shrimp until 18 trials with sar- dine larvae were completed. Each experiment was concluded with a trial of brine shrimp to test for satiation. The number of observations for each larval size class was the total number of predator-prey inter- actions observed among larvae in that size class. The mean standard length was determined from 20 larvae sampled randomly from the rearing tank on the day of each behavior experiment. The numbers of observations for each size class (mean SL) were 8.0 mm, 41; 11.3 mm, 51; 12.1 mm, 114; 12.7 mm, 46; 14.1 mm, 81; 17.6 mm, 104; and 19.6 mm, 69. Experiments were not extended to larger sizes due to insufficient numbers of larvae. Classification of Behavior Prey behavior was scored only when the preda- tor attacked a prey. Four measures of predator- prey interactions were calculated: predator at- tack distance, the distance from which the predator responded to the prey and initiated its attack; frequency of avoidance response; fre- quency of escapes; and predation rate (percentage of larvae captured during the 5-min trials). An avoidance response was a change in speed or tra- jectory of a larvae before the predator had com- pleted its attack by closing its mouth. An escape was defined as a larval response in which the predator failed to capture the larvae in a single attack. Typically adult anchovy make a single attack on a prey item and do not pursue a prey that escapes (Folkvord and Hunter 1986) but rather continue searching the tank. Thus attacks on one prey item were recorded twice if the first attack was unsuccessful. Although predator at- tack distance was recorded, this measurement is highly subjective and comparison with the meas- urements of other observers is suspect. We did not analyze predator attack distance for this reason. 164 To compare response and escape behaviors of Pacific sardine with those reported by Folkvord and Hunter (1986) for northern anchovy at simi- lar stages of development, mean lengths were converted to ages using field growth rates back- calculated from otolith increment widths for each species. Confidence limits of the percentage of lar- vae responding or escaping attack were estimated assuming the binomial distribution. Results Probability of Response to Attack The youngest larval stages of both sardine and anchovy were the most vulnerable to predation. Only n7c of 8 mm sardine larvae (smallest size tested) responded to attack by adult anchovy. With increasing size more sardine larvae re- sponded to attack. At 20 mm, the largest size tested, 61% of the larvae responded to attack. The response rate of Pacific sardine larvae was consis- tently lower than that of northern anchovy larvae of similar lengths (Fig. 1). Although this differ- ence in responsiveness could be due to differences in the observer, it may also be explained by the difference in age of anchovy and sardine at the same length. Sardine larvae are about 6.2 mm when they begin feeding (age = 5 days from fertil- ization at 17°C), whereas first-feeding anchovy larvae are only about 4.3 mm (age = 5 days from fertilization at 17°C) (Zweifel and Lasker 1976). Sardine larvae also grow faster than anchovy lar- 100 o 80 1- z a. O 60 (A U OC 40 20 - ANCHOVY V.-- SARDINE 8 12 16 LENGTH (mm) 24 Figure 1. — Increase by size of the percentage of Pacific sardine larvae, Sardinops sagax, and northern anchovy larvae, En- graulis mordax, responding to attack by adult northern an- chovy and 95% confidence intervals. Data on anchovy larvae from Folkvord and Hunter (1986). vae at the same temperature (Butler and Rojas de Mendiola 1985). Thus, sardine larvae are younger at a given size than anchovy larvae. Since the latency of response to attack must be related to the development of the central nervous system (Webb 1981; Webb and Corolla 1981), it may be more appropriate to compare sardine lar- vae with anchovy larvae of the same age. For that reason lengths of the larvae of both species were converted to age using growth rates measured in the field (Methot and Kramer 1979; Butler 1987). Comparison of the percentage of larvae respond- ing to attack at a given age (Fig. 2) reveals no significant difference in the rate of development of response to attack. Thus, the escape response develops at the same rate in Pacific sardine and northern anchovy, and the difference in propor- tion of larvae responding at a given size (Fig. 1) is due to the difference of size at hatching and the difference in growth rates of the two species. 100 o z a. < O If) LU o OC UJ Q. 80 - 60 - 40 20 - _ h ANCHOVY . -'- - -| r .••■■ ..?' r .■■••• r J ^ L , [ kj •■••n 1 ' 1 SARDINE 1 10 20 AGE (days) 30 40 Figure 2. — Increase by age of the percentage of Pacific sardine larvae, Sardinops sagax, and northern anchovy larvae, En- graulis mordax, responding to attack by adult northern an- chovy and 95% confidence intervals. Size categories of reared larvae have been converted to ages using growth rates esti- mated from the field. Data on anchovy larvae from Folkvord and Hunter (1986). Probability of Escaping Attack The ability to successfully avoid attack in- creased with size of Pacific sardine as well as northern anchovy. Few small larvae of either spe- cie escaped attack by adult northern anchovy. Only 3% of 8 mm sardine larvae escaped attack and the percentage of larvae escaping increased to only 11% for 17 mm larvae and 13% for 20 mm 165 larvae (Fig. 3). The proportion of small anchovy larvae escaping attack was also low {67c ) but in- creased with size to 737r of 22 mm larvae (Folkvord and Hunter 1986). The numbers of lar- vae escaping attack were significantly different between anchovy and sardines at sizes larger than about 13 mm. Conversion of lengths to age using field growth rates does not eliminate the differences between sardine and anchovy (Fig. 4). Sardine larvae older than 20 days were more vul- nerable to predation than anchovy larvae of the same age (Fig. 4). 100 I— o z 5 z o a V) z lU O a. UJ Q. 80 60 40 20 ANCHOVY .f V..-- -SARDINE 8 12 16 LENGTH (mm) 20 24 Figures. — Increase by size of the percentage of Pacific sardine larvae, Sardinops sagax, and northern anchovy larvae, En- graulis mordax, escaping attack by adult northern anchovy and 95'^ confidence intervals. Data on anchovy larvae from Folkvord and Hunter (1986). 1 \J\J _ I i 80 - -] p .,.•'1 o z o Q. 60 V) \u a. ^ 40 o - . 1 ANC i- < HO r>-' 1 UJ 20 ~ < ■.C-^' J- SARDINE , ^ -l_. 1 1 F t r c ) 1 20 30 40 A GE ( days) Figure 4. — Increase by age of the percentage of Pacific sardine larvae, Sardinops sagax, and northern anchovy larvae, En- graulis mordax, escaping attack by adult northern anchovy and 95% confidence intervals. Size categories of reared larvae have been converted to ages using growth rates estimated from the field. Data on anchovy larvae from Folkvord and Hunter (1986). Discussion The proportion of Pacific sardine larvae re- sponding to attack and escaping attack increased with size and with age. Our results differ from those reported by Folkvord and Hunter (1986) for anchovy larvae in the rate at which sardine lar- vae respond and escape attacks at given sizes and ages. It should be noted that, although the methodology was the same, the observers were different. This difference could affect rate of re- sponse to attack. It also should be noted that the size of adult anchovy used by Folkvord and Hunter (1986) ranged from 83 to 89 mm SL, whereas the size range was 84-95 mm SL in our study and that the size of predator influences the number of larvae escaping (Folkvord and Hunter 1986). The slightly larger size of predators used in this study is not sufficient to explain difference in escapement, nor is the difference in observer likely to affect the rate of escapement since the observer's task is to examine whether the larvae are escaping or are being eaten. The greater vulnerability to predation of sar- dine larvae than anchovy larvae has interesting implications. In general, larger larvae are less vulnerable to predation than small larvae. Bailey (1984) and Bailey and Batty (1983) compared the vulnerability of cod, flounder, plaice, and herring larvae to predation by invertebrate predators. They found that herring larvae were the least vulnerable larvae because herring were more re- active and had the greatest escape speeds. Sar- dine larvae are larger at hatching and at a given age are larger than anchovy larvae. In our exper- iment sardine larvae react to predatory attacks at similar rates as anchovy larvae, but escape attack at a much lower rate. This difference in vulnerability to attack may be due to differences in swimming behavior. An- chovy larvae swim using beat and glide locomo- tion (Hunter 1972). The escape behavior is usu- ally a burst of swimming from a motionless position (Folkvord and Hunter 1986). We ob- served that sardine larvae, however, swim contin- uously and they respond to attack by changing direction and increasing speed. This difference in swimming mode may affect escape behavior in two ways. The escape behavior of sardine larvae may be less flexible than that of anchovy larvae because the direction the sardine larvae takes is largely determined by its trajectory. Since sar- dine larvae cruise, their scope for activity (escape behavior) may be limited. Anchovy larvae accel- 166 erate from a standing start and have the possibil- ity of moving in a number of directions. We spec- ulate that the beat and glide behavior of anchovy larvae may not only be hydrodynamically more efficient (Weihs 1974) but also may reduce the vulnerability to predation. Acknowledgments Roderick Leong maintained sardines in the lab- oratory and spawned the adults after manipula- tion with hormones. Sandor Kaupp reared sar- dine larvae in the laboratory. Arild Folkvord and Clelia Booman elaborated experimental proce- dures. John R. Hunter commented on the manuscript. Literature Cited Ahlstrom, E H 1954. Distribution and abundance of egg and larval popu- lations of the Pacific sardine. Fish. Bull., U.S. 56:83- 140. Bailey, K. M. 1984. Comparison of laboratory rates of predation on five species of marine fish larvae by three planktonic inverte- brates; effects of larval size on vulnerability. Mar. Biol. (Berl.) 79:303-309. Bailey. K M . and R S. Batty 1983. A laboratory study of predation on Aurelia aurita on larval herring (Clupea harengus ): experimental observa- tions compared with model predictions. Mar. Biol. (Berl.) 72:295-301. Brownell, C. L. 1985. Laboratory analysis of cannibalism by larvae of the Cape anchovy Engraulis capensis. Trans Am. Fish. Soc. 114:512-518. Butler, J. L. 1987. Comparisons of the early life history parameters of Pacific sardine and northern anchovy and implications for species interactions. PhD Thesis, University of California at San Diego, San Diego, 242 p. Butler, J L., and B. Rojas de Mendiola 1985. Growth of larval sardines off Peru. CalCOFI Rep. XXVI:113-118. Folkvord. A., and J R Hunter 1986. Size-specific vulnerability of northern anchovy, En- graulis mordax, larvae to predation by fishes. Fish. Bull., U.S. 84:859-869. Hewitt, R P . G H Theilacker. and N C H Lo 1985. Causes of mortality in young jack mackerel. Mar. Ecol. Prog. Ser. 26:1-10. Hunter. J R 1972. Swimming and feeding behavior of larval an- chovy, Engraulis mordax. Fish. Bull., U.S. 70:821- 838. 1976. Culture and growth of northern anchovy, Engraulis mordax, larvae. Fish. Bull., U.S. 74:81-88. 1984. Inferences regarding predation on the early life stages of cod and other fishes. Flodevigen rapportser., 1., ISSN 0333-2594 The propagation of cod Gadus morhua L, p. 533-562. Hunter. J R . and C A. Kimbrell 1980. Egg cannibalism in the northern anchovy, En- graulis mordax. Fish. Bull., U.S. 78:811-816. Loukashkin, a S 1970. On the diet and feeding behavior of the northern anchovy, Engraulis mordax (Girard). Proc. Calif Acad. Sci. 37 (Ser. 4):419-458. MacCalL, A. E. 1979. Population estimates for the waning years of the Pacific sardine fishery. Calif Coop. Oceanic Fish. In- vest. 10:72-82. Methot, R D , JR , AND D Kramer. 1979. Growth of northern anchovy, Engraulis mordax, larvae in the sea. Fish. Bull., U.S. 77:413-423. Murphy, G I. 1966. Population biology of the Pacific sardine (Sardinops caerulea). Proc. Calif Acad. Sci. 34:1-34. OConnell, C, p. 1980. Percentage of starving northern anchovy, En- graulis mordax, larvae in the sea as estimated by histo- logical methods. Fish. Bull., U.S. 78:475-484. Owen, R W , J L Butler, N C H Lo. A Alvarlno. G H Thei- lacker, J R. Hunter, J. Hakanson, and Y Watanabe. In press. Spawning and survival patterns of larval an- chovy in contrasting environments — a site intensive ex- periment. Cal. Coop. Fish. Invest. Rep. Smith, P E. 1985. Year-class strength and survival of 0-group clu- peoids. Can. J. Fish. Aquat. Sci. 42 (Suppl. 11:69-82. Theilacker. G H 1986. Starvation-induced mortality of young sea-caught jack mackerel, Trachurus symmetricus , determined with histological and morphological methods. Fish. Bull., U.S. 84:1-15. Webb, P. W 1981. Responses of northern anchovy, Engraulis mordax, larvae to predation by a biting planktivor, Amphiprion percula. Fish. Bull., U.S. 79:727-735. Webb. P W . and R T Corolla 1981. Burst swimming performance of northern anchovy, Engraulis mordax, larvae. Fish. Bull., U.S. 79:143- 150. Webb. P W , and D Weihs. 1986. Functional locomotor morphology of early life his- tory stages of fishes. Trans. Am. Fish. Soc. 115:127. Weihs, D. 1974. Energetic advantages of burst swimming of fish. J. Theoret. Biol. 48:215-229. ZWEIFEL, J E , AND R. LaSKER 1976. Prehatch and posthatch growth of fishes — a general model. Fish. Bull., U.S. 74:609-621. John L. Butler Darlene Pickett Southwest Fisheries Center National Marine Fisheries Service, NOAA P.O. Box 271 LaJolla.CA 92117 167 INDUCTION OF SPAWNING IN THE WEAKFISH, CYNOSC/ON REGALIS Reproductive activity in the weakfish, Cynoscion regalis, is associated with extensive, north-south migrations that result in spawning in the estuar- ies of the Middle Atlantic Bight during the late spring and early summer. Spawning is appar- ently related to increasing water temperature and day length, but there have been no experi- mental investigations of specific factors that con- trol this process in the weakfish. In contrast, gen- eral aspects of the reproductive biology of the species are well known (reviewed by Mercer 1983). Both males and females become sexually mature at 1 year of age, and remain sexually active throughout their lifespan (10^ years). Spawning involves external fertilization of eggs by pairs or small aggregations of fish. There has been limited study of larval develop- ment in the laboratory, and descriptions of growth and development of larval weakfish come entirely from field investigations (Lippson and Moran 1974). Weakfish larvae resulting from gametes stripped from sexually mature adults captured in the field have been reared for a few days on natural zooplankton diets (Public Service Electric and Gas Company 1984), but no informa- tion is available on mortality or growth rates on prescribed diets and rations. In contrast, Houde and Taniguchi (1981) and Taniguchi (1981, 1982) have conducted extensive investigations of the ef- fects of diet, ration, and temperature on growth and survival of larval spotted seatrout, Cynoscion nebulosus, under laboratory conditions. Similar studies have been conducted with other sciaenids including red drum, Sciaenops ocellatus (Holt and Arnold 1983; Holt et al. 1981), spot, Leiostomus xanthurus (Powell and Gordy 1980), and bairdiella, Bairdiella icistia (May 1974). The present paper describes a technique for in- duction of spawning in a laboratory population of weakfish and provides preliminary information on early development and growth of weakfish lar- vae. Methods Sixteen adult weakfish, C. regalis, (approxi- mately 30-45 cm and 0.5-1.5 kg) were captured in September 1984 in Delaware Bay by hook-and- line. Five of these fish were dissected and found to have regressed gonads. The remaining 11 fish were placed in two large tanks (2,000 L) con- nected to a 20,000 L recirculating system that delivered temperature-controlled seawater to each tank at 10 L/minute. Water in the recircu- lating system was replaced approximately monthly. Ordinary white room light was provided by two 1.25 m fluorescent lamps positioned 1 m above the surface of the water in each tank. Ini- tial conditions in the tank were similar to ambi- ent conditions in Delaware Bay in September: 18°-19°C, 30%c, and 12 hours light:12 hours dark. Temperature and salinity in the tanks were mea- sured daily and pH approximately weekly. The pH was maintained between 7.0 and 7.6 by additions of new seawater to the system; this was accomplished by replacing 40% of the water in the system with new seawater approximately monthly. Fish began to feed 5-7 days after capture. Diet consisted of an ad libitum ration of sliced squid with weekly additions of penaeid shrimp or fresh calf liver. After approximately one month in the laboratory, the fish were subjected to a prescribed regimen of temperature and photophase (Fig. 1). Temperatures were lowered and light phase shortened over a period of three weeks until con- ditions reached 8 hours light, 13°-14°C; this ap- proximated winter conditions on the continental shelf off Cape Hatteras where adult fish are known to overwinter (Merriner 1976). Fish were held under these conditions for 11 weeks after which temperature was gradually raised and light phase increased until spring conditions of 14 hours light and 22°-23°C were reached. Fish were held at these conditions throughout a period of extended spawning activity. After spawning ac- tivity ceased, six newly captured fish were placed in the system to replace fish that had died during the previous year, and the process of gradually changing temperature and photoperiod to winter conditions was repeated (Fig. 1). Fertilized eggs were buoyant and exited the tanks at the surface via stand pipes that emptied into a sump tank. The drain was located at the bottom of the sump tank allowing eggs to accu- mulate at the water surface. The presence or ab- sence of eggs in the sump was determined daily with a fine-meshed dip net. After collection, eggs from each spawning were allowed to hatch in 20 L plastic aquaria filled with 5.0 ixm filtered sea- water at 30%c and 23°C. Eggs exposed to gentle aeration under these conditions hatched in 24-36 hours. Larvae were cultured in 2 L beakers filled with filtered seawater (25°C, 30%f, and 5 mg/liter 168 FISHERY BULLETIN: VOL 86, NO. 1, 1988. MAY JUNE JULY AUG SEPT o LU IT a: Ld Q_ 25 24 23 22 21 20 H 19 18 17 16 15-1 14 13 12 II 10 ▼ T 16 C/) 3 O 1^ X X 8 1- (3 -z. 4 LU _l >- < O OCT NOV DEC JAN FEB MAR APR MAY Figure 1. — Temperature-photophase regime used to condition northern weakfish, Cynoscion regalis, to spawn in the laboratory. Values are 7-d means. Standard deviations are plotted for temperature. Arrows on time axis indicate dates of spawning. A = 1984-85. B = 1985-1986. 169 chloramphenicol) at a density of 25 larvae/liter. Larvae were fed rotifers, Brachionis plicatilis, be- ginning two days after hatching; on the seventh day after hatching, brine shrimp Artemia sp. were added to the diet. Early larval development from spawning to first feeding was determined for larvae from three separate spawnings. Newly hatched larvae (6-15 hours old) were pipetted into an aquarium and 10-15 removed immediately and preserved in 70% ethanol. Additional larvae were sacrificed daily for nine days. These samples were examined to relate early development and age. Results and Discussion Three adult fish died of undermined causes dur- ing the period of winter conditions in 1984-85. Four additional fish died after jumping from the tanks. The initial spawning by the remaining population of four fish occurred five weeks after spring conditions were achieved. This was fol- lowed the next day by another spawning. Spawn- ing continued for nine weeks at 10-14 d intervals. Spawning episodes usually consisted of produc- tion of fertilized eggs over 2 successive days. While actual spawning was never observed, it al- ways occurred between sunset and 08:00 the fol- lowing day. After cessation of spawning, the four fish were removed from the tanks and their sex deter- mined. Two of the fish drummed when handled and were clearly males. At least one (and proba- bly both) of the two remaining fish were females. The fish were returned to the system after exam- ination. The temporal sequence of the spawning events (two successive days at 10-14 d intervals) sug- gests that each female may have spawned as many as four times during the 9-wk period. This is in contrast to reports for natural populations of C. regalis (e.g., Merriner 1976), but has been reported for laboratory populations of other sciaenid fishes. For example, Arnold (1984) ob- served 82 spawning events in a laboratory popu- lation of 12 C. nebulosus over a 27-mo period and 52 spawning events in a similar population of six S. ocellatus over a 3-mo period. It is not clear why the C. regalis in our investigation ceased spawn- ing after several months of long days and high temperatures while the C. nebulosus in Arnold's system continued to spawn over a much longer period. Perhaps this is related to the smaller an- nual variation in photophase and temperature typical of C. nebulosus habitats. However, go- nadal resorption has been reported for at least one other sciaenid species iB. icistia) held in the laboratory for extended periods of long day-length (May 1974). Spawning from the 1985-86 conditioning pe- riod first occurred four weeks after spring conditions were reached with a second spawning nine days later. Further spawning in the 1985- 86 population did not occur because of an unidentified infection resulting in the death of all 10 fish in the system over a period of a few weeks. Autopsies revealed that all fish had highly developed ovaries or testes at the time of death. Newly hatched larvae (6-15 hours old) had a yolk sac, no mouth, and little development of the eyes. By 24-36 hours after hatching, the yolk sac had been virtually absorbed, the mouth was just beginning to form, and the eyes were not yet pig- mented. By 48-60 hours, larvae had a completely formed mouth and digestive system, and the eyes were pigmented. Larvae were capable of feeding at this stage. Chloramphenicol improved larval survival, which was as high as 24% over 11 days. This survival is comparable to that reported for other sciaenid larvae (Holt et al. 1981; Houde and Taniguchi 1981; Holt and Arnold 1983). However, growth was less than the maximum seen in the laboratory for C. nebulosus. After 11 days C. re- galis larvae in the present experiments had grown from a mean, posthatching size of 2.7 mm (36 fxg dry weight) to 4.5 mm (235 \i.g dry weight). In contrast Houde and Taniguchi (1981) found that one group of C. nebulosus larvae reached a size of 13.6 mm (7,082 fxg dry weight) in 12 days when reared on a concentrated ration of natural zooplankton at very low stocking density and high temperature (32°C). However, when fed a rotifer diet at comparable temperatures and stocking densities, growth of C. nebulosus larvae was somewhat less than that of C. regalis larvae in the present experiments. Our results show that the spawning cycle of weakfish can be manipulated to produce repeated spawnings without the aid of hormone injections. While the fish appear to resorb their gonadal tis- sue after several months of exposure to long day length and high temperature, the differential ma- nipulation of several groups of fish could allow year-round production of fertilized eggs. Further- more, survival and growth of larvae produced in this manner appear comparable to survival and 170 growth of other sciaenids reared in the labora- tory. Acknowledgments This study was supported by funds provided by the University of Delaware Sea Grant College Program and by the U.S. Department of Fish and Wildlife through the Delaware Department of Natural Resources and Environmental Control. Paul Cosby, Big John Ellsworth, Anne Marie Ek- lund, John Ewart, Paul Grecay, Anne Hastings, Kathleen Little, Leslie Picoult, and Peter Rowe were indispensible in the capture, care, and feed- ing of the adult and larval weakfish used in the investigation. Literature Cited Arnold, C R 1984. Maturation and spawning of marine finfish. In Carl J. Sindermann (editor). Proceedings of the seventh U.S. -Japan meeting on aquaculture, marine finfish cul- ture held at Tokyo, Japan, October 3-4, 1978, p. 25- 27. U.S. Dep., NOAA Tech. Rep. NMFS 10. Holt, G J , and C R. Arnold 1983. Effects of ammonia and nitrite on growth and sur- vival of red drum eggs and larvae. Trans. Am. Fish. Soc. 112:314-318. Holt, G J , R Godbout, and C R. Arnold 1981. Effects of temperature and salinity on egg hatching and larval survival of red drum, Sciaenops ocel- lata. Fish. Bull., U.S. 79:569-573. Houde, E D.. and a K Taniguchi 1981. Marine fish larvae growth and survival. Effects of density-dependent factors: spotted seatrout iCynoscion nebulosus ) and lined sole iAchirus lineatus ). Report to U.S. Environmental Protection Agency, Environmental Research Laboratory, Narragansett, R.I. EPA 600/3-81- 052. Natl. Tech. Inf Serv. PB82-101395, 66 p. LiPPSON, A J , and R L. Moran 1974. Manual for identification of early developmental stages of fishes of the Potomac River estu- ary. Environmental Technology Center, Martin Mari- etta Corporation, Baltimore, MD, 282 p. May,R C 1974. Effects of temperature and salinity on fertilization, embryonic development, and hatching in Bairdiella icis- tia (Pisces:Sciaenidae), and the effect of parental salinity acclimation on embryonic and larval salinity toler- ance. Fish. Bull., U.S. 73:1-22. Mercer. L P. 1983. A biological and fisheries profile of weakfish, Cynoscion regalis. N.C. Dep. Nat. Resour. Comm. Dev. Div. Mar. Fish., Spec. Sci. Rep. No. 39, 107 p. Merriner, J V. 1976. Aspects of reproductive biology of the weakfish, Cynoscion regalis, in North Carolina. Fish. Bull., U.S. 76:18-26. Powell, A B , and H R Gordy 1980. Egg and larval development of the sp)ot Leiosto- mus xanthurus (Sciaenidae). Fish. Bull., U.S. 78:701- 714. Public Service Electric and Gas Company. 1984. Salem Generating Station 316 (b) Demonstration. Appendix XI. Weakfish (Cynoscion regalis): A synthesis of information of natural history with reference to occur- rence in the Delaware River and Estuary and involve- ment with the Salem Generating Station. Public Ser- vice Electric and Gas Company, Newark, NJ. Taniguchi, A K. 1981. Survival and growth of larval spotted seatrout (Cynoscion nebulosus ) in relation to temperature, prey abundance and stocking densities. Rapp. P. -v. Reun. Cons. int. Explor. Mer 178:507-508. 1982. Growth efficiency estimates for laboratory-reared larval spotted seatrout fed microzooplankton or roti- fers. In C. F. Bryan, J. V. Conner, F. M. Truesdale (ed- itors), Proceedings of the Fifth Annual Larval Fish Con- ference, p. 6-11. Louisiana State University, Cooperative Fishery Research Unit, Baton Rouge, LA. Charles E Epifanio David Goshorn Timothy E. Targett College of Marine Studies University of Delaware Lewes, DE 19958 171 ERRATA Fishery Bulletin: Vol. 85, NO. 4 Botton, Mark L., and John W. Ropes, "Populations of horseshoe crabs, Limulus polyphemus , on the northwestern Atlantic continental shelf," p. 809. Page 809, Table 3, footnote was omitted: 'Depth at the end of this tow was 439 m Henwood, Tyrrell A., and Warren E. Stuntz, "Analysis of sea turtle captures and mortalities during commercial shrimp trawling," p. 814. Page 814, paragraph 1, line 3, the equation should read: E = (nets * length/30.5 m) * (min/60) INFORMATION FOR CONTRIBUTORS TO THE FISHERY BULLETIN Manuscripts submitted to the Fishery Bulletin will reach print faster if they conform to the following instructions. These are not absolute requirements, of course, but desiderata. CONTENT OF MANUSCRIPT The title page should give only the title of the paper, the author's name, his/her affiliation, and mailing ad- dress, including ZIP code. The abstract should not exceed one double-spaced page. In the text, Fishery Bulletin style, for the most part, follows that of the U.S. Government Printing Office Style Manual. Fish names follow the style of the Amer- ican Fisheries Society Special Publication No. 12, A List of Common and Scientific Names of Fishes from the United States and Canada , Fourth Edition, 1980. Text footnotes should be typed separately from the text. the literature cited. The accuracy of the literature cited is the responsibility of the author. Abbreviations of names of periodicals and serials should conform to Serial Sources for the BIOSIS Data Base. (Chemical Ab- stracts also uses this system, which was developed by the American Standards Association.) Common abbreviations and symbols, such as mm, m, g, L, mL, mg, °C (for Celsius), %, %o, and so forth, should be used. Abbreviate units of measure only when used with numerals. Periods are only rarely used with abbreviations. We prefer that measurements be given in metric units; other equivalent units may be given in parenthe- ses. FORM OF THE MANUSCRIPT Figures and tables, with their legends and headings, should be self-explanatory, not requiring reference to the text. Their placement should be indicated in the right-hand margin of the manuscript. Preferably figures should be reduced by photography to 53/4 inches (for single-column figures, allowing for 50% reduction in printing), or to 12 inches (for double- column figures). The maximum height, for either width, is 14 inches. Photographs should be printed on high quality or glossy paper. Do not send original drawings to the Scientific Editor; if they, rather than the photographic reductions, are needed by the printer, the Scientific Publications Office will request them. Each table should start on a separate page. Consis- tency in headings and format is desirable. Vertical rules should be avoided, as they make the tables more expen- sive to print. Footnotes in tables should be numbered sequentially in arable numerals. To avoid confusion with powers, they should be placed to the left of numer- als. Acknowledgments, if included, are placed at the end of the text. Literature is cited in the text as: Lynn and Reid (1968) or (Lynn and Reid 1968). All papers referred to in the text should be listed alphabetically by the senior author's surname under the heading "Literature Cited." Only the author's surname and initials are required in The original of the manuscript should be typed, double-spaced, on white bond paper. Please triple space above headings. We would rather receive good dupli- cated copies of manuscripts than carbon copies. The se- quence of the material should be; TITLE PAGE ABSTRACT TEXT LITERATURE CITED TEXT FOOTNOTES APPENDIX TABLES (Each table should be numbered with an arable numeral and heading provided). LIST OF FIGURES (Entire figure legends) FIGURES (Each figure should be numbered with an arable numeral; legends are desired) ADDITIONAL INFORMATION Send the ribbon copy and two duplicated or carbon copies of the manuscript to: Dr. Andrew E. Dizon, Scientific Editor Fishery Bulletin Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service, NOAA P.O. Box 271 La Jolla, CA 92038 Fifty separates will be supplied to an author free of charge and 50 supplied to his organization. No covers will be supplied. Contents — Continued CURRNES, KENNETH P., CARL B. SCHRECK, and HIRAM W. LI. Reexamina- tion of the use of otolith nuclear dimensions to identify juvenile anadromous and -^ nonanadromous rainbow trout, Salmo gairdneri 160 BUTLER, JOHN L., and DARLENE PICKETT. Age-specific vulnerability of Pacific sardine, Sardinops sagax, larvae to predation by northern anchovy, En- graulis mordax 163 EPIFANIO, CHARLES E., DAVID GOSHORN, and TIMOTHY E. TARGETT. Induction of spawning in the weakfish, Cynoscion regalis 168 GPO 791-008 kt^' Co. Fishery Bulletin •^'■^rEs o< P^^^^^^^^ Vol. 86, No. 2 ocr^8 ms April 1988 Migratibns- of jjoho' sa^lmon, PEARCY, WILLIAM G., and JOSEPH P. FISHI Oncorhynchus kisutch, during their first summer in the ocifta'n-?'*.^.^*^.;^^. T?: . . . 173 DUTIL, J.-D. and J.-M. COUTU. Early marine Hfe of Atlantic salmon/Sa?7m> salar, postsmolts in the northern Gulf of St. Lawrence 197 MURPHY, MICHAEL L., JOHN F. THEDINGA, and K V. KOSKI. Size and diet of juvenile Pacific salmon during seaward migration through a small estuary in southeastern Alaska 213 BOLZ, GEORGE R., and R. GREGORY LOUGH. Growth through the first six months of Atlantic cod, Gadus morhua , and haddock, Melanogrammus aeglefinus , based on daily otolith increments 223 NYMAN, ROBERT M., and DAVID O. CONOVER. The relation between spawn- ing season and the recruitment of young-of-the-year bluefish, Pomatomus salta- trix , to New York 237 JAHN, A. E., D. M. GADOMSKI, and M. L. SOWBY. On the role of food-seeking in the suprabenthic habit of larval white croaker, Genyonemus lineatus (Pisces: Sciaenidae) 251 WILLIAMS, AUSTIN B. New marine decapod crustaceans from waters influenced by hydrothermal discharge, brine, and hydrocarbon seepage 263 MARTIN, JOEL W., FRANK M. TRUESDALE, and DARRYL L. FELDER. The megalopa stage of the Gulf stone crab, Menippe adina Williams and Felder, 1986, with comparison of megalopae in the genus Menippe 289 SHENKER, JONATHAN M. Oceanographic associations of neustonic larval and juvenile fishes and Dungeness crab megalopae off Oregon 299 DAGG, M. J., P. B. ORTNER, and J. AL-YAMANI. Winter-time distribution and abundance of copepod nauplii in the northern Gulf of Mexico 319 HERRNKIND, WILLIAM F., MARK J. BUTLER IV, and RICHARD A. TANKERS- LEY. The effects of siltation on recruitment of spiny lobsters, Panulirus argus 331 KIRKLEY, JAMES E., and DALE E. SQUIRES. A limited information approach for determining capital stock and investment in a fishery 339 POLACHECK, TOM. Analyses of the relationship between the distribution of searching effort, tuna catches, and dolphin sightings within individual purse seine cruises 351 (Continued on back cover) Seattle, Washington U.S. DEPARTMENT OF COMMERCE C. William Verity, Jr., Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION William E. Evans, Under Secretary for Oceans and Atmosphere NATIONAL MARINE FISHERIES SERVICE James W. Brennan, Assistant Administrator for Fisheries Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. 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 instead of being issued individually. 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. SCIENTIFIC EDITOR, Fishery Bulletin Dr. Andrew E. Dizon Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service, NCAA P.O. Box 271 La Jolla, CA 92038 Editorial Committee Dr. Jay Barlow National Marine Fisheries Service Dr. William H. Bayliff Inter-American Tropical Tuna Commission Dr. George W. Boehlert National Marine Fisheries Service Dr. Robert C. Francis University of Washington Dr. James R. Kitchell University of Wisconsin Dr. William J. Richards National Marine Fisheries Service Dr. Bruce B. CoUette National Marine Fisheries Service Dr. Tim D. Smith National Marine Fisheries Service Mary S. Fukuyama, Managing Editor The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fisheries Service, NOAA, 7600 Sand Point Way NE, BIN C15700, Seattle, WA 98115. Second class postage is paid in Seattle, Wash., and additional offices. POSTMASTER send address changes for subscriptions to Fishery Bulletin, Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Although the contents have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has determined that the publication of this periodical is necessary in the transaction of the public business required by law of this Department. Use of funds for printing this periodical has been approved by the Director of the Office of Management and Budget. Fishery Bulletin CONTENTS / LIBRARY OCT 2 8 )98g ' Vol. 86, No. 2 I April1988 PEARCY, WILLIAM G., and JOSEPH P. FISHERi..,Migrfcns of coho salr^^n, : Oncorhynchus kisutch , during their first summer in thr nrrTHl i i , J.73 DUTIL, J.-D. and J.-M. COUTU. Early marine life of Atlantic salmon, Salmo salar, postsmolts in the northern Gulf of St. Lawrence 197 MURPHY, MICHAEL L., JOHN F. THEDINGA, and K V. KOSKI. Size and diet of juvenile Pacific salmon during seaward migration through a small estuary in southeastern Alaska 213 BOLZ, GEORGE R., and R. GREGORY LOUGH. Growth through the first six months of Atlantic cod, Gadus morhua , and haddock, Melanogrammus aeglefinus , based on daily otolith increments 223 NYMAN, ROBERT M., and DAVID O. CONOVER. The relation between spawn- ing season and the recruitment of young-of-the-year bluefish, Pomatomus salta- trix , to New York 237 JAHN, A. E., D. M. GADOMSKI, and M. L. SOWBY. On the role of food-seeking in the suprabenthic habit of larval white croaker, Genyonemus lineatus (Pisces: Sciaenidae) 251 WILLIAMS, AUSTIN B. New marine decapod crustaceans from waters influenced by hydrothermal discharge, brine, and hydrocarbon seepage 263 MARTIN, JOEL W., FRANK M. TRUESDALE, and DARRYL L. FELDER. The megalopa stage of the Gulf stone crab, Menippe adina Williams and Felder, 1986, with comparison of megalopae in the genus Menippe 289 SHENKER, JONATHAN M. Oceanographic associations of neustonic larval and juvenile fishes and Dungeness crab megalopae off Oregon 299 DAGG, M. J., P. B. ORTNER, and J. AL-YAMANI. Winter-time distribution and abundance of copepod nauplii in the northern Gulf of Mexico 319 HERRNKIND, WILLIAM F., MARK J. BUTLER IV, and RICHARD A. TANKERS- LEY. The effects of siltation op recruitment of spiny lobsters, Panulirus argus 331 KIRKLEY, JAMES E., and DALE E. SQUIRES. A limited information approach for determining capital stock and investment in a fishery 339 POLACHECK, TOM. Analyses of the relationship between the distribution of searching effort, tuna catches, and dolphin sightings within individual purse seine cruises 351 (Continued on next page ) Seattle, Washington 1988 For sale by the Superintendent of Documents, U.S. Government Printing OfTice, Washing- ton DC 20402 — Subscription price per year: $16.00 domestic and $20.00 foreign. Cost per single issue: $9.00 domestic and $11.25 foreign. Contents — Continued WATSON, CHERYL, ROBERT E. BOURKE, and RICHARD W. BRILL. A compre- hensive theory on the etiology of burnt tuna 367 BROWN-PETERSON, NANCY, PETER THOMAS, and CONNIE R. ARNOLD. Reproductive biology of the spotted seatrout, Cynoscion nebulosus, in South Texas 373 Notes CHEN, CHE-TSUNG, TZYH-CHANG LEU, and SHOOU-JENG JOUNG. Notes on reproduction in the scalloped hammerhead, Sphyrna lewini, in northeastern Taiwan waters 389 COLLINS, MARK R., and CHARLES A. WENNER. Occurrence of young-of-the- year king Scomberomorus cavalla, and Spanish, S. maculatus, materials in commercial-type shrimp trawls along the Atlantic coast of the southeast United States 394 DEW, C. BRAXTON. Stomach contents of commercially caught Hudson River striped bass, Morone saxatilis , 1973-1975 397 VERNET, MARIA, JOHN R. HUNTER, and RUSSELL D. VETTER. Accumula- tion of age pigments (lipofuscin) in two cold-water fishes 401 MULLIN, M. M., and E. R. BROOKS. Extractable lipofuscin in larval marine fish 407 Notices: NOAA Technical Reports published during the last 6 months of 1987 . . . 416 The National Marine Fisheries Service (NMFS) does not approve, recommend or endorse any proprietary product or proprietary material mentioned in this publi- cation. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased because of this NMFS publication. MIGRATIONS OF COHO SALMON, ONCORHYNCHUS KISUTCH , DURING THEIR FIRST SUMMER IN THE OCEAN William G. Pearcy and Joseph P. Fisher' ABSTRACT Marked juvenile coho salmon caught in fine-meshed purse seines during the summers of 1981-84 off Oregon and Washington generally demonstrated northward migrations from their rivers of ocean entrance. Northward movements in summer were preceded by southerly movements during spring, probably caused by southerly advection. Catch rates and sizes offish caught in different months and regions of the coast also indicated northerly movements of both yearling and subyearling coho salmon. Despite this movement, the average catch of juvenile coho salmon per purse seine set along the coasts of Washington and Oregon in late summer, including marked fish from the Columbia River, was still a substantial proportion of that in May and June soon after ocean entrance, suggest- ing that many coho did not migrate great distances. Additionally, recoveries of marked juvenile coho salmon by sports and commercial fishermen from Alaska to California and by scientists in Alaska were generally in the region of release. These data indicate that migrations of juvenile coho are of limited extent during their first summer in the ocean and are not strong support for an earlier conclusion that juvenile coho salmon from the Columbia River, Oregon, and California may form a large proportion of the stocks of this species that migrate northward along the coastal belt in Cana- dian and Alaskan waters each summer. Although there has been Httle research on juve- nile salmon during their first summer at sea, this phase of the life history may be critical to survival and recruitment to fisheries (Hartt 1980). High- est ocean mortality is thought to occur early in marine life (Foerster 1968; Parker 1968; Ricker 1976). Production (catch and escapement) of adult coho salmon, Oncorhynchus kisutch, in the Ore- gon Production Index (OPI) Area (from Leadbet- ter Point, WA, to Monterey Bay, CA) is usually accurately predicted in one year by the number of precocious males (jacks) returning to index streams in the previous year (Gunsolus 1978; Or- egon Department of Fish and Wildlife 1982; Pacific Fishery Management Council 1986). Hence survival from jacks to adults is fairly con- stant from year to year. Because survival rates from smolt to adult are variable (Nickelson 1986), however, variable year-class survival must occur before the time that jacks return, after only a few months in the ocean. This relationship, and the positive correlation between coastal upwelling and survival of OPI coho salmon (Gunsolus 1978; Scarnecchia 1981; Nickelson 1986), strongly sug- gest that the first few months in the ocean consti- tute the "critical period" in determining subse- quent adult survival. iCollege of Oceanography, Oregon State University, Corval- lis, OR 97331. Between 1976 and 1985 the production of coho salmon in the OPI area drastically declined, de- spite large increases in the number of public and private smolt releases (Oregon Department of Fish and Wildlife 1982; Nickelson 1986). Reduced upwelling and ocean productivity, perhaps cou- pled with density-dependent mortality, is one of the hypothesized causes for this decrease in sur- vival (Scarnecchia 1981; Peterman and Rout- ledge 1981; McCarl and Rettig 1983; McGie 1984; Nickelson 1986). To understand the mechanisms affecting survival of juvenile salmonids at sea, we must first know where salmon reside at the time of their high and variable mortality. Are the smolts highly migratory, immediately leaving local coastal waters and migrating into waters of the Gulf of Alaska (Hartt and Dell 1986), or are they nonmigratory, spending their early ocean life in local coastal waters? This paper summarizes research on the move- ments and migrations of coho salmon during their first summer in the ocean in the northeastern Pacific Ocean based on purse seine catches made mainly in coastal waters off Oregon and Wash- ington. A few records of migrations of tagged ju- venile (age .0)^ coho salmon were given by God- Manuscript accepted November 1987. FISHERY BULLETIN: VOL. 86, NO. 2, 1988. 2The numeral preceding and following the decimal indicate the number of winters spent in fresh water and in the ocean, respectively. 173 FISHERY BULLETIN: VOL. 86, NO. 2 frey (1965) and French et al. (1975), but by far the most comprehensive data were provided by Hartt (1980) and Hartt and Dell (1986). All of these studies, however, were based on recovery of ma- ture or maturing coho in the year following tag- ging and on tagging in northern waters from the Strait of Juan de Fuca to the Alaska Peninsula. The only other studies of juvenile salmon in the ocean off Washington and Oregon have been re- stricted to within 24 km of the Columbia River (Dawley et al. 1981) or to coastal waters during 1980 (Miller et al. 1983). Our 1979-85 research, covering large areas along the coast, provides ex- tensive and unique data on the movements of ju- venile coho salmon during their first summer in the ocean. METHODS Purse seines, our primary sampling method, were used to sample juvenile salmonids during 1979-85 (Table 1). Cruises were in coastal waters off Oregon in 1979-80, off Oregon and southern Washington in 1981, and off Oregon and the en- tire Washington coast during the summers of 1982-85 (Fig. 1). During July 1984, sets were also made from northern California (lat. 40°32'N) to northern Vancouver Island (50°26'N). Except for the exploratory cruises off Oregon in 1979 and 1980, purse seine sets were usually made along east-west transect lines (Fig. 1). Sets started at the 37 m (20-fathom) contour, and continued at 9.3 km (5-mi) intervals farther offshore, usually until no salmonids were captured. Repeat sets were sometimes made when fish with missing adipose fins were common, indicating the pres- ence of coded wire tagged (CWT) fish. In 1985, special sets were made in the vicinity of the Co- lumbia River plume. Detailed sampling data are provided in Pearcy (1984) and our cruise reports (Wakefield et al. 1981; Fisher et al. 1982, 1983; Fisher and Pearcy 1984, 1985). The mesh size of the seines were the same dur- ing all years, 32 mm (stretch), with 32 mm or smaller mesh in the bunts of the seines. The seine was 495 m long except in 1981 (457 m). Depths that seines fished, sometimes measured with a depth gauge on the lead line, varied among years from about 20 m to 65 m (Table 1). Generally, purse seine sets were "round hauls", where the seiner and the skiff made a circle with the net. The seine was fully pursed after about one-half its length was aboard (half-purse sets). All sets were "blind". We attempted to use sonar on some cruises to locate concentrations of sal- monids but were unsuccessful. Radar was some- times used to determine the distance between the seiner and the skiff when a semicircle was made with the net. Each round haul encompassed about 17,000 m2 (1981) or 19,000 m2 (1979-85). To de- termine the direction of movement of fish, eight "half-round" hauls, or "semicircular" sets, were made in 1979, where the entire net formed an open semicircle. Paired sets were made in close succession, with sets open in a northern and a southern direction, at four locations. The seine was open for the same duration (15-45 minutes, depending on location) in each paired set while Table 1. — Summary of number of purse seine sets and latitudinal range of sampling, 1979-85. Dates of No. of Purse seine Lengtfi Depthi Year cruises setsi Latitudinal range of sampling (m) (m) 1979 18-29 June 56 Cape Disappointment to Cape Arago 46°20' -43°18' 495 20 1980 20-28 June 36 Cape Disappointment to Alsea River 46°20' -44°30' 495 20 1981 16-25 May 63 Willapa Bay to Alsea River 46°35' -44°25' 495 20 9-18 June 67 Willapa Bay to Cut Creek 46°35' -43°1 1 ' 495 20 9-19 July 67 Willapa Bay to Alsea River 46°35' -44°25' 457 49 8-19 Aug. 66 Willapa Bay to Cut Creek 46°36' -44°1 1 ' 457 49 1982 19 May-2 June 62 Waatch Point to Siuslaw River 48°2r -44°00' 495 265 7-22 June 57 Quinault River to Yachiats 47°21 ' -44°20' 495 265 4-14 Sept. 42 Quinault River to Yachats 47°20' -44°19' 495 265 1983 16-27 May 56 Waatcfi Point to Yachats 48°2r -44°20' 495 249 9-27 June 58 Waatcfi Point to Four Mile Creek 48°20' -43°00' 495 249 15-24 Sept. 53 Waatch) Point to Coos Bay 48°20' -43°28' 495 249 1984 4-20 June 69 Waatcfi Point to Coos Bay 48''20' -43°27' 495 249 9 July-3 Aug. 65 Winter Harbor, B.C. to False Cape. CA 50°26' -40°32' 495 249 1-15 Sept. 63 Waatcfi Point to Siuslaw River 48°20' -44°00' 495 249 1985 29 May-25 June 112 Sea Lion Rock to Coos Bay 48°00' -43°27' 495 225 ^Quantitative sets. ^Measured with depth guage. 174 PEARCY and FISHER: MIGRATIONS OF COHO SALMON Figure l. — Locations of purse seine transects off the Oregon and Washington coasts. WAATCH POINT SEA LION ROCK DESTRUCTION ISLAND QUINAULT RIVER GRAYS HARBOR WILLAPA BAY CAPE DISAPPOINTMENT SEASIDE NEHALEM BEACH CAPE LOOKOUT WECOMA BEACH YAOUINA HEAD YACHATS SIUSLAW RIVER COOS BAY FOUR MILE CREEK 125 124" the vessel and skiff towed the seine only fast enough to maintain a constant net opening. The purse seine catches were either dip-netted from the bunt of the seine while it was alongside the vessel, brailed aboard, or hauled aboard in the bunt, depending on the composition and size of the catch. In 1979 and 1980, juvenile salmon preserved in formalin were identified ashore. In 1981-85, ju- venile salmon were identified to Species at sea, fork length (FL) was measured to the nearest mil- limeter and then they were individually wrapped in labelled plastic bags and frozen. All salmonids with marks or missing adipose fins were frozen. When large numbers of juvenile salmonids were caught in a set, most unmarked fish were re- leased after they were measured. In order to increase the numbers of marked fish released into our study area we marked about 1.5 million coho smolts in 1981 and 835,000 in 1982 using fluorescent pigment propelled by com- pressed air (see Phinney et al. 1967) prior to their 175 FISHERY BULLETIN: VOL. 86. NO. 2 transport from Oregon Aqua-Foods, Inc. (OAF) hatchery to their ocean release facilities at Yaquina Bay or Coos Bay, OR. In the laboratory ashore, species identifications were confirmed and individuals remeasured and reexamined for both fluorescent marks (under ul- traviolet light 1981-82) and missing adipose fins or other marks (1979-85). Coded-wire tags from the heads of salmonids with missing adipose fins were decoded by personnel from the Oregon De- partment of Fish and Wildlife, Clackamas Labo- ratory. Juvenile or age .0 (first year in the ocean) coho salmon were distinguished from adult or age .1 (second year in the ocean) coho salmon by exami- nation of size-frequency histograms and scales. The division between age .0 and .1 coho pro- gressed from approximately 300 to 420 mm FL from May to September, in most years. Most coho salmon migrated to the ocean a little over one year after hatching (age 1.0), but OAF released large numbers of subyearling (age 0.0) smolts into Yaquina Bay and Coos Bay. These two age groups of smolts were distinguished by the radial distance to the 21st circulus on scales removed from the preferred area (Clutter and Whitesel 1956) of the fish. The accuracy of this method for distinguishing known age 0.0 and 1.0 fish was approximately 85-90%. In the years 1981-85, scales from 52% of the 4,222 juvenile coho sam- pled were analyzed. The estimated numbers of age 0.0 and 1.0 fish represented in different geographic areas and cruises were then extrapo- lated from their proportions in each 10 mm length interval. Distances traveled and movement rates were estimated from actual distances between sites of release and entry into the ocean, and from straight-line distances between ocean entry and recapture locations for CWT or fluorescent marked juvenile coho salmon that were recovered in the ocean within 10 days of release. These dis- tances and swimming speeds are minimal esti- mates. In addition to purse seining, fine-meshed monofilament gill nets were used off the Oregon coast (ca. lat. 45°00'N, long. 124°21'W) during 24 and 25 July 1985, from the training vessel Oshoro Maru , to determine depth and direction of swim- ming of juvenile salmonids. Surface and subsur- face nets were used. The surface gill nets were 2,050 m long, and fished from depths of 0-6 m with 11 mesh sizes ranging from 29 to 121 mm (stretch). The subsurface nets were 500 m long and consisted of four mesh sizes ranging from 29 to 42 mm; they were suspended below large (300- 400 mm) mesh to fish at depths of 5-12 m. Four sets were made in an east-west direction with soak-times of about 4-9 hours. As the gill nets were hauled, the direction that each juvenile salmonid was heading when caught, and its depth in the net (upper, middle, or lower section) were noted. Each juvenile salmonid was given a con- secutive number and frozen for later identifica- tion. Comparisons of catch rates in the surface and subsurface nets were based on equal lengths of the four mesh sizes of the subsurface net, stan- dardized to 10 hours of fishing time. Information on the location of landings of marked juvenile coho by commercial and sports fishermen was provided by the Pacific Marine Fisheries Commission (PMFC), (PMFC 1980, 1981, 1984a, b, c, 1985a, b), from lists of non- standard recoveries (Johnson PMFC unpubl. data), and from state agencies. The actual num- bers of tagged fish, and the total numbers of tagged fish estimated from the proportions of the catch sampled are reported. To determine if juvenile coho salmon were sex- ually precocious "jacks", we examined testes from 542 juvenile males caught in July 1981 and 1984, in August 1981, and in September 1982, 1983, and 1984. All developed and some undeveloped testes (ribbonlike, with no thickening), as deter- mined by visual inspection, were weighed (123) and gonadal-somatic indices (GSI = testes wt./ body wt. X 100) were determined. RESULTS AND DISCUSSION Swimming Direction Of the 106 juvenile coho salmon captured dur- ing June in paired, half-round purse seine sets, all but two were in the sets open to the south (Table 2). This suggests that juvenile coho salmon Table 2. — Catches of coho salmon In semicircular purse seine sets open to the south (S) and north (N) off Oregon, June 1979. Location Km offshore Age S .0 N Age .1 S N Clatsop Spit Clatsop Spit Clatsop Spit Newport 12.6 18.5 18.5 9.4 57 37 7 3 2 2 6 15 37 6 8 9 98% 2% 28% 72% 176 PEARCY and FISHER: MIGRATIONS OF COHO SALMON were swimming to the north during this sampHng period. Maturing fish over 300 mm FL (age .1 echo) showed the opposite trend. Miller et al. (1983) made several hundred paired purse seine sets open to the south and north during three cruises off the northern Oregon-southern Washington coasts in 1980. During their May-June cruise, they caught 76% of the juvenile coho salmon, 80% of the chinook salmon, and almost all the steelhead trout in sets open to the south, indicating northward move- ment. We note a positive relation between the proportion of juvenile salmon caught in south- facing sets in their three cruises and strength of upwelling during these cruises (mean daily Bakun indices of 52, 39, and 19 m^ s"! 100 m'^ coastline at 45°N, 125°W in May, July, and Au- gust, respectively (Mason and Bakun 1986)), sug- gesting that surface currents to the south result- ing from Ekman transport may be cues for orientation of salmon smolts. Hartt (1980) and Hartt and Dell (1986) found that 83% of the combined species of juvenile salmonids caught in 19 paired purse sets along the coast from Cape Flattery, WA to Yakutat, AK were caught in sets held open to the southeast and only 17% in sets open to the northwest and north. They concluded that juvenile salmonids tended to migrate in a northwest direction along the coast during July-September. Of the 100 juvenile coho salmon (135-315 mm FL) caught in the gill nets set in an east-west direction off the Oregon coast in July 1985, 90 coho were caught as they approached the south- ern face of the gill net (heading north) and 10 in the northern face (heading south). Jaenicke et al. (1984) reported that 63% of the juvenile coho caught in a surface gill net fished off southeastern Alaska in July moved north at night, but only 6% moved to the north during the day. Available data indicate that most juvenile coho salmon caught off Oregon and southern Washing- ton, as well as juveniles farther to the north, are predominantely swimming in a northerly direc- tion during summer months. Depth Distribution One-half of the juvenile coho salmon caught in gill nets set off the Oregon Coast in 1985 were in the upper 2 m of the surface gill net (Table 3). Catches in the surface net exceeded those in the subsurface net, except for the last set that fished during daylight hours, indicating that juvenile coho salmon were most common in the upper 4 m of the water column. Other information on the vertical distribution of maturing coho and other species of salmon caught in gill nets or with longlines in oceanic waters also indicates that they usually swim near the surface, between and 20 m (Manzer 1964; Godfrey 1965; Godfrey et al. 1975). Machidori (1966), for example, fished gill nets from the sur- face to 50 m and caught 79% of the coho salmon in the upper 10 m of the gill net. Although catches in gill nets at different depths may be biased by ver- tical differences in avoidance reactions to the net or swimming speeds (Hartt 1975), acoustical methods have also shown that salmon are usually distributed near the surface (Susuki and Sonoda 1972; Lord et al. 1976). We conclude that most juvenile coho salmon in coastal waters and ma- turing coho in oceanic waters reside at depths above 20 m, the minimum depth that our purse seine fished. We recognize, however, that matur- ing coho and other species of salmon may feed in deeper water. Some salmon (including coho salmon) caught in surface gill nets in the oceanic waters of the Gulf of Alaska contained prey in their stomachs characteristic of mesopelagic depths (200-1,000 m), suggesting that some indi- viduals may feed well below the thermocline (Pearcy et al. in press). Table 3. — Catches of juvenile coho salmon in four gill net sets in 50 m lengths of 29, 33, 37, and 42 mm mesh at different depths and times 24-25 July 1985, each set adjusted to 10-h fishing duration. Depth in meters Surface net Subsurface net Times of set 0-2 2-4 4-6 5-7 7-9 9-12 0913-1702 2001-0104 0248-0701 0830-1737 2.5 25.0 42.8 1.1 3.8 11.5 2.4 2.4 1.3 9.5 6.6 1.3 16.6 7.7 1.9 5.5 Total catch Percent of total catch 51 50.3 16 12.5 1 1.7 11 12.3 15 18.0 6 5.2 North-South Trends in Catch per Set and Sizes of Juvenile Coho Variations in the average catches and sizes of juvenile coho salmon in purse seine sets in differ- ent regions of the Oregon-Washington coast dur- ing the summer provide indirect evidence for north-south coastal movements. Histograms showing average catches per set for 10 mm size 177 FISHERY BULLETIN: VOL. 86, NO. 2 groups of juvenile coho, classified as age 0.0 or 1.0 from scale analysis, are shown in Figure 2, for 1981-84 in three regions: (A) Cape Flattery, WA to Grays Harbor, WA (called Washington), (B) Willapa Bay, WA to Nehalem Bay, OR (Columbia River region), and (C) Cape Lookout, OR to Coos Bay, OR (Oregon) (Fig. 1). In May of 1981, 1982, and 1983, average catch per set of yearling (age 1.0) coho generally decreased from the southern to the northern regions. Catches v^ere highest off Oregon (Area C) or the Columbia River (Area B) and lowest off Washington (Area A) in May of 1982 and 1983. This trend was reversed later in the summer. In June of 1981, 1983, and 1984, lowest catches were found in the Oregon region. By August or September 1981-84, highest catches consistently occurred off the Columbia River or Washington and few yearling fish were caught off Oregon. These shifts in abundance sug- gest a northerly movement of age 1.0 smolts dur- ing the summer. Highest catch rates occurred in May and June of 1981 and 1982 when an average of over 10 juvenile coho salmon were caught in most sets. Subyearling or age 0.0 coho salmon released from private facilities at Yaquina and Coos Bays provide more direct evidence on movements. Sub- yearling coho salmon clearly demonstrated northward dispersal. They were most common in our catches of July 1981 and September 1982, 1983, and 1984 (Fig. 2). They were apparently more numerous than age 1.0 coho salmon in the Oregon region during June-August 1981 and September 1982, and in the Oregon and Columbia River regions in September 1983. The catches and proportions of age 0.0 coho salmon increased off Oregon during the summer because they were released from hatcheries later in the summer than yearling coho salmon. They were found in the most northern region sampled late in the 1981 6 - MAY 5 - 4 - 3 - 2 [- /SET J -TV. X 6 O ^ 5 " 4 3 - 2 - " 1 Jl wW.ljb.nv ,..y 20 30 JUNE JULY ....Mljlito^ i^i,i^»i AUGUST B C 11*11(^1 30 10 20 30 40 FORK LENGTH (cm) Figure 2. — Catch per purse seine set of age 1.0 (open) and age 0.0 (shaded) juvenile coho by 10 mm length groups during different months, 1981-84, for three regions of the Oregon-Washington coast: (A) Cape Flattery to Grays Harbor, WA, (B) Willapa Bay, WA to Nehalem Beach, OR, (C) Cape Lookout to California. 178 PEARCY and FISHER: MIGRATIONS OF COHO SALMON summer of all years. Their abundance, as that of age 1.0 coho salmon, also increased during the summer off Washington where they intermingled with age 1.0 coho salmon. Because the Oregon region included the release locations of all age 0.0 coho salmon, our figures provide no information on southward movements of these fish. The mean lengths of both age 0.0 and 1.0 coho salmon increased from the southern to the north- ern areas during most months. Larger age 0.0 and 1.0 coho salmon were caught off Washington than Oregon during the late summer, 1981-84 (Fig. 2), providing corroborative evidence for northward migration of coho salmon. Larger, and pre- sumably older, fish were found farther to the north than smaller fish. Despite northward movements, many yearling coho salmon did not migrate out of the sampling area, but remained in coastal waters off Oregon and Washington during the entire summer. Mean coastwide catch per set of yearling coho salmon in August 1981 and September 1982, 1983, and 1984 1982 I- LJ if) 6 5 4 3 21- 6 5 4 3 21- 51- J 4 MAY TTfllTTTTTTtlll l^[ JUNE t^t^ r i^i M ^ r T ^ f rr SEPTEMBER A ¥.« ^ ¥^ i W i T' i I ^ T^T*T^ IllllTTTfftfl ^ .UJJlmnifii B ^JMpMk T^P» | B .oM^ rp .iP, , c 1 — f^ 30 10 -r- 10 20 30 10 20 30 10 20 30 FORK LENGTH (cm) Figure 2.— Continued. 40 The Question of "Jacks" Are the juvenile coho salmon off Oregon and Washington in late summer relatively nonmigra- tory because they are sexually precocious? GSI were almost always <0.1% for those fish visually classified as "undeveloped". GSI's from fish with "developed" testes ranged from 0.29?^ to 1.0% in July 1981 and 1984 in fish >250 mm FL; from 0.3% to 5.6% in August 1981, mostly in fish >280 mm FL; and from 2.4% to 6.6% (except for one value at 0.6) in September 1982, 1983, and 1984 in fish >300 mm FL (Fig. 3). In August 1981, and clearly in September 1982, 1983, and 1984, two distinct groups of fish were evident: "jacks", with developing testes (GSI >0.3% August or GSI >2.0% September), and "nonjacks", which showed no development (GSI <0.1%). The total numbers of jacks and nonjacks in each 50 mm length group were estimated for the catch during August 1981 and September 1982, 1983, and 1984 from the ratio of jacks and nonjacks in the sample (Table 4). Only 8.4%, 4.8%, 5.2%, and 2.8% of all juvenile fish (male and female) were "jacks" in August 1981 and September 1982, 1983, and 1984, respectively. However jacks com- 181 FISHERY BULLETIN: VOL, 86, NO. 2 X LU Q < -z. o 3 - - 2 D 1981 n= 12 A 1984 n=9 - 1 n A A A ,D ^ * ° A 1 1 1 1 1 1 1 1 6 5 4 3 2 I 9 8 7 6 5 4 3 2 I - - D 1981 n=55 D D D D D °D D - D D ° a - .iiniLtniiM imiiiiiriijiiinjii .Viiiiiim ii. i i 1 1 t 1 1 1 1 1 1982 n=9 • 1983 n=27 A 1984 n=ll • - - A • A ^ A O o o - _ O . - ^ •a • 0° o - A - - , M . , r . m^m • m^}A ^AaftMB . an . . . , 1 , • , 20 25 30 35 40 45 FORK LENGTH (cm) Figure 3. — Gonadal-somatic index (testis wt/total body wt) x 100 of juvenile coho salmon vs. length of fish for July, August, and September 1981-84. Only data for those testes actually weighed are shown. prised a higher percentage of fish larger than 300 mm FL in August 1981 and larger than 350 mm FL in September 1982, 1983, and 1984. These results indicate that most juvenile coho salmon caught off Oregon and Washington were not sexu- ally precocious. Thus, the relatively large catches of juvenile coho salmon in late summer are ex- plained by lack of strong migrational tendencies of juvenile coho salmon in this region and not by a high proportion of precocious "jacks" that re- sided in this region as a prelude to re-entry of streams for spawning. 182 PEARCY and FISHER: MIGRATIONS OF COHO SALMON Table 4. — The percent of coho salmon jacks and males, by length groups, in the total catch, Augjst 1981 and September 1982-84. Fork Number Number Est. Est. % length of fish % of Total total jacks Date (mm) examined males jacks' catch jacks of total Aug. 1981 <200 71 60.6 111 201-250 63 60.0 1 115 2 1.7 251-300 55 47.3 4 104 8 7.7 301-350 16 81.3 4 22 6 27.3 351-420 10 80.0 8 19 15 78.9 Total 215 59.5 17 371 31 8.4 Sept. 1982 <200 56 55.4 125 201-250 21 76.2 54 251-300 22 63.6 109 301-350 33 69.7 3 97 9 9.3 351-420 13 53.8 5 28 11 39.3 Total 145 62.8 8 413 20 4.8 Sept. 1983 <200 16 62.5 18 201-250 23 47.8 25 251-300 23 56.5 71 301-350 39 61.5 4 77 8 10.4 351-420 2 1 3 2 66.7 Total 103 58.3 5 194 10 5.2 Sept. 1984 <200 6 33.3 15 201-250 38 47.4 69 251-300 90 50.0 2 128 3 2.3 301-350 27 55.6 12 31 2 6.5 351-420 6 33.3 2 7 2 28.6 Total 167 49.1 16 250 7 2.8 'Jack Is defined as a male whose testes wt. /total body wt. x 100 > 0.3% in August and >2.0% in Septem- ber. Movements of Marked Fish Direct evidence of movements of juvenile coho salmon was obtained from capture of marked fish containing coded wire tags or marked with fluo- rescent pigment. The generalized pattern of movements that emerges for 1981-85 is an initial movement of most juvenile coho salmon to the south soon after ocean entry in May and June and then a reversal of movement with most fish mi- grating to the north by August and September (Figs. 4-8). These trends are discussed for fish originating from the Columbia River, Oregon coastal, Washington coastal, and private hatch- eries. Columbia River Juvenile coho salmon originating from hatch- eries on the Columbia River were usually recov- ered south of the Columbia River in May. This trend was especially obvious in May 1982 when all 22 marked fish which were recovered moved south, some as far as 175 km (Fig. 5). In May 1981, all but one of 14 marked Columbia River fish were caught to the south, three as far as 180 and 204 km (Fig. 4). In May 1983, all four fish were taken south of the mouth of the Columbia River (Fig. 6). During June and July of all years, marked Co- lumbia River coho salmon were recovered in nearly equal proportions both north and south of the river mouth, except in June 1982 when 15 of 17 fish were found to the south (Figs. 4—8). By September, all marked Columbia River coho salmon were captured north of the river, includ- ing fish captured off the Quinault River in Sep- tember 1982 and off Cape Flattery in September 1984. Fish were also caught close to the mouth of the Columbia River in July, August, and Septem- ber, indicating that some marked juvenile coho salmon did not undertake extensive migrations at sea. In two sets on the Wecoma Beach Transect on 1 June 1982 we caught 17 marked juvenile coho salmon released between 30 April and 6 May from six hatcheries on the Columbia River. Based on downstream migration rates for these groups to Jones Beach (Dawley et al. 1985) and assuming similar rates from Jones Beach to the ocean, these fish had probably been in the ocean for <10 days before recapture. This indicates that some juve- 183 FISHERY BULLETIN: VOL. 86, NO. 2 AUG JULY JUNE 1981 Figure 4. — North-south movements of marked juvenile echo salmon captured in purse seines, May-August 1981. The width of the lines are approximately proportional to the number offish. Numbers at end of arrows indicate number offish captured. Arrows without numbers and thin lines represent single fish. Inshore-offshore movements are not shown. Dashed lines indicate latitudinal extent of sampling. nile coho salmon released from hatcheries at about the same time tended to stay together dur- ing their downstream migration in the Columbia River and during early residency in the ocean. Oregon Public Coastal Hatcheries We captured marked fish originating ft-om pub- lic Oregon coastal hatcheries both north and south of the latitude of ocean entrance in May. A total of five fish were found to the south, while 11 fish were found to the north in May (Figs. 4-8). With the exception of one coho salmon from the Umpqua River in June 1983 and two from the Rogue River in July 1984 (Figs. 6, 7), the other 25 fish taken after May were captured north of where they entered the ocean. Northerly move- ments into Washington waters occurred by June 1983 and 1985 (Figs. 6, 8). The southward movements of two juvenile coho salmon released from the Rogue River (south of Cape Blanco) and captured off northern Califor- 184 PEARCY and FISHER: MIGRATIONS OF COHO SALMON I I I I Ku I I I I 1 IV q -CAPE FLATTERY SEPT JUNEV MAY 1982 -rCAPE FLATTERY Figure 5. — North-south movements of marked juvenile coho salmon captured in purse seines, May, June, and September 1982. The width of the lines are approximately proportional to the number of fish. Numbers at end of arrows indicate number offish captured. Arrows without numbers and thin lines repre- sent single fish. Inshore-offshore movements are not shown. Dashed lines indicate latitudinal extent of sampling. JUNE 1983 Figure 6. — North-south movements of marked juvenile coho salmon captured in purse seines, May, June, and September 1983. The width of the lines are approximately proportional to the number of fish. Numbers at end of eirrows indicate number offish captured. Arrows without numbers and thin lines repre- sent single fish. Inshore-offshore movements are not shown. Dashed lines indicate latitudinal extent of sampling. nia during July 1984 are notable (Fig. 7). They were captured in our only cruise into California waters and represent the only recoveries of marked juvenile coho salmon originating from hatcheries south of Cape Blanco in all six years of sampling. Although ocean sampling was limited south of Coos Bay, if juvenile coho salmon from southern Oregon and northern California hatch- eries had migrated north of Coos Bay, we would expect them to be represented in our catches. The fact that they were not caught in this northern region, but two were caught after swim- ming to the south, suggests that juvenile coho salmon originating in streams south of Cape Blanco may migrate south, possibly occupying the region of intense coastal upwelling off north- ern California during their first summer in the ocean. The catch of over 70% of the adult coho 185 FISHERY BULLETIN: VOL. 86, NO. 2 ll SEPT UMPQUA RIVER JUNE 10-25 .coos BAY MAY 29 - JUNE 5 1985 ± CAPE BLANCO I I I I I — 43= JULY 1984 Figure 8. — North-south movements of marked juvenile coho salmon captured in purse seines, 29 May-5 June and 10-25 June. The width of the lines are approximately proportional to the number of fish. Numbers at end of arrows indicate number offish captured. Arrows without numbers and thin lines repre- sent single fish. Inshore-offshore movements are not shown. Dashed lines indicate latitudinal extent of sampling. Figure 7. — North-south movements of marked juvenile coho salmon captured in purse seines, June-September 1984. The width of the lines are approximately propor- tional to the number of fish. Numbers at end of arrows indicate number of fish captured. Arrows without num- bers and thin lines represent single fish. Inshore- offshore movements are not shown. Dashed lines indi- cate latitudinal extent of sampling. 186 PEARCY and FISHER: MIGRATIONS OF COHO SALMON salmon from the Rogue River in the troll fishery off California (R. Garrison'^) is further evidence for a southern distribution of this stock. Marked juvenile coho salmon from California hatcheries were reported from the sports fishery off southern Oregon, however, as will be shown later. Washington Coastal Hatcheries Juvenile coho salmon from Washington public hatcheries demonstrated southerly movements, sometimes into Oregon waters, during May 1981 and 1982 (Figs. 4, 5). During June 1982, 1984, and 1985, Washington coastal fish were found both north and south of ocean entry. Except for one fish that moved north in September 1984, no Washington coastal fish were taken in August or September of other years, suggesting that most Washington fish may have migrated out of our sampling area by late summer. Oregon Private Hatcheries All marked juvenile coho salmon originating from Yaquina and Coos Bays that we captured at sea were from private hatcheries. Those from Yaquina Bay were mainly age 0.0 smolts from OAF, those from Coos Bay were either age 1.0 smolts from Anadromous, Inc. or age 0.0 smolts from OAF. Forty-one recoveries of marked OAF fish released from Yaquina Bay were caught to the north while only 4 were to the south of Yaquina Bay. In general, more juvenile coho salmon from Yaquina Bay were captured in late than early summer, and distances traveled to the north were largest (up to 413 km) for fish caught in later summer (Figs. 4-8). All recoveries of marked Anadromous, Inc. and OAF fish released into Coos Bay were to the north in all years. Large northerly movements were demonstrated by some of these fish (Figs. 4, 6, 8). Since our sampling in the ocean usually did not extend south of Coos Bay, recoveries of these fish are biased to the north; however, strong northward movements of these stocks were indicated. Rates of Movement Recoveries of marked juvenile coho salmon in the ocean provided information on the minimum rates of movement from hatchery release to cap- 3R. Garrison, Oregon Department of Fish and Wildlife, Cor- vallis, OR 97330, pers. commun. December 1983. ture in the ocean. Some fish moved rapidly through estuaries into the ocean. We captured some tag-groups in the ocean only a few days after the median date of capture at Jones Beach (75 km from the ocean) as reported by Dawley et al. (1985): 5 fish after 2 days in 1981, 6 fish after 3-11 days in 1981, 8 fish after 1-14 days in 1982, and 5 fish after 3-8 days in 1983. Dawley et al. (1986) found average rates of movement of 14-23 km d"^ for marked groups of coho smolts from areas of release on the Columbia River to river km 75, and rates of movement that were 40% faster from river km 75 to the lower Columbia River estuary and to the ocean plume. These re- coveries and those reported by Miller et al. (1983) for yearling chinook salmon and steelhead trout indicate rapid movements of juvenile salmonids of over 20 km d^^ through the Columbia River estuary. Some juvenile coho salmon released from Yaquina Bay and Coos Bay also demonstrated rapid movements into and in the ocean, e.g., 17 Anadromous, Inc. fish were captured 11 km north of Coos Bay only two days after release in June 1983 (Table 5). Myers (1980) described an expo- nential decrease in the catches of juvenile coho salmon released from the OAF facility into Yaquina Bay; about one-half the fish from marked groups remaining in the bay after 1.7-9.0 days. Juvenile coho salmon apparently emigrate rapidly from estuaries into the ocean. Some of the marked fish recovered within 10 days of release demonstrated rapid movements down-rivers or in the ocean. Twenty-four fish traversed 10 km d~^ or more largely in the ocean, in both north and south directions (Table 5). Four fish released in bays or in the ocean moved over 18.8 km d~^ Two of these swam to the north, presumably against coastal currents. These speeds are equivalent to 1.7 body lengths (BL) per second or more and suggest that some fish must be traveling nearly straight courses during 24-h days, since 1-3 BL s"^ are thought to be optimal cruising speeds for small (<20 cm) pelagic fishes (Weihs 1973; Ware 1978). These maximum rates of movement for purse seine caught juvenile coho salmon are similar to those estimated by Hartt (1980) and Hartt and Dell (1986) for tagged sockeye salmon during their first summer in the ocean: 14-27 km d"^ for 11 Fraser River fish and 6-14 km d"^ for 10 Skeena River fish that were between about 8 and 23 cm in length during the migration period. 187 FISHERY BULLETIN: VOL. 86. NO. 2 Table 5. — Release Information and mean travel speeds in kilometers per day and body length (BL) per second for CWT and fluorescent-pigment marked age .0 coho salmon recovered in the ocean within 10 days of release. CR ^ Columbia River: OAF = Oregon Aqua Foods, Inc., (OAF Yaquina is 3.7 km from ocean; OAF Coos Is 14 km from the ocean), Anad. - Anadromous Inc. (7.4 km from the ocean). Median Direction Release Date days to Mean Mean Mean of Year No. location released recovery FL km/d BUS movement 1981 1 Big Cr. (CR) 8 June 3 153 22.2 1.7 S 1 Tanner Cr. (CR) 6 July 6 138 41.0 3.4 N 1 OAF Yaquina 11 May 7 140 4.0 0.3 S 1 OAF Yaquina 12 June 2 124 14.5 1.4 N 2 Anad. Coos 8 June 9 179 2.3 0.2 N 2 OAF Yaquina 10-15 June 2 124 3.7 0.3 N 1 OAF Yaquina 10-15 June 2 123 10.2 1.0 N 1 OAF Yaquina 10-15 June 2 126 18.8 1.7 N 11 OAF Coos 5-9 June 10 122 1.9 0.2 N 1 OAF Yaquina 10-15 June 5 124 24.0 2.2 S 1 OAF Coos 5-9 June 7 136 20.1 1.7 N 1983 17 Anad. Coos 26 June 2 156 11.2 0.8 N 1984 1 OAF-Offshore 7 June 7 143 11.3 0.9 N 1985 6 Tongue Pt. (CR) 24 IVIay 6 151 9.5 0.7 S 1 Offshore-22 km 30 l^ay 0.9 134 22.0 1.9 S 1 Naselle River 20 l^ay 9 143 7.8 0.6 S 1 Cowlitz River 31 IVIay-6 June 2.5 144 88.0 7.1 N Effects of Ocean Currents The tendency for juvenile coho salmon to move to the south early in the summer and to the north later in the summer (Figs. 4—8) may be related to advection of water and the size, orientation, and swimming speeds of fish. Generally, surface cur- rents are to the south off Oregon and Washington in the early summer owing to prevailing north- westerly winds (Hickey 1979; Huyer 1983). Southward flow averaging 17-34 km d"^ (Huyer et al. 1975, 1979) has been measured near the surface. May and June are periods of peak outflow of the Columbia River, so fish entering the ocean at this time, especially in the Columbia River plume, may be displaced to the south by advection of surface waters. Southward flow is at a maxi- mum in the coastal jet which is strongest (—22 km d~^) during the spring about 15-20 km from shore (Kundu and Allen 1976; Huyer et al. 1979). Since currents can be equivalent to 1.7 BL s"^ for a 15 cm smolt, advection alone could explain the southward movement of most marked Columbia River fish during May and June but not the rapid northward movement of fish during this period (see Figures 4-7). Coastal Oregon fish were often found to the north in May and June, but these fish were usu- ally substantially larger and generally released much earlier in the spring than Columbia River fish, and were presumably better able to swim against the current. Later in the summer when Columbia River hatchery fish had grown larger, movement was also predominately northward. In August and September southward velocities of surface coastal currents are diminished and the mean may be near zero (Huyer et al. 1975). Northward movements during the summer off Oregon and Washington generally cannot be ex- plained by passive drift and in most years must entail active, oriented swimming. The northern El Nino of 1982-83, which had severe effects on the growth and survival of adult and jack coho salmon (Pearcy et al. 1985; Johnson 1984; Pearcy and Schooner 1987; Fisher and Pearcy in press), also appeared to affect the distri- bution of juvenile coho salmon. During Septem- ber of 1983 nearly all the seine-captured juvenile coho were taken along our northernmost transect, off Cape Flattery, WA (Fig. 2). In other years juvenile coho salmon during late summer were common and more equally distributed from the Columbia River northward. In the summer of 1983 juvenile coho salmon may have moved far- ther north, or more likely those to the south may have experienced higher mortality, as a result of northerly currents (Huyer and Smith 1985), warm temperatures and low productivity (Pearcy et al. 1985; Chung 1985) that prevailed off Ore- gon. 188 PEARCY and FISHER: MIGRATIONS OF COHO SALMON Recoveries of CWT Juvenile Fish in Ocean Fisheries Data on ocean location of landings of juvenile CWT coho salmon reported in sports and commer- cial fisheries in the ocean along the west coast of North America, 1977-83, provide valuable infor- mation on ocean migrations of marked fish, al- though these data are biased by differences in legal minimum sizes, time and duration of open season, and effort in the different regions. The summary of all years shows that, except for Cali- fornia, most of the recoveries of juvenile coho salmon during their first summer in the ocean were in the general region of their ocean entry location (Table 6). Both the actual number offish reported and the estimated total numbers (in parentheses) support our earlier conclusion that many juvenile coho salmon off Oregon and Wash- ington are not highly migratory. All (20) of the actual recoveries of marked juvenile coho salmon that were released in southeastern Alaskan waters were from southeastern Alaska. Ninety- seven percent of the recoveries of marked fish released in British Columbia waters were landed in British Columbia; only two were landed in Alaska. Most (86%) marked juvenile fish from Puget Sound hatcheries were caught in the Sound, and more were recovered in British Co- lumbia fisheries (13%) than in coastal Washing- ton fisheries (<1%), probably due to the smaller size limits for coho in British Columbia as well as migratory patterns. Half of the actual numbers of recoveries of juvenile coho salmon liberated into Washington coastal waters were landed in Washington coastal ports, followed by British Co- lumbia (29%) and the Columbia River region (17%). Only one was landed in Alaska and two in Oregon ports (Garibaldi and south). Juvenile coho salmon originating from Columbia River hatch- eries had a broader distribution of recoveries in other regions. Only 40% of Columbia River fish were caught in this region, 41% were taken in northern regions, including two (1%) in Alaska. The remaining 19% were captured off Oregon. The majority (73%) of Oregon coastal fish were recovered off Oregon, followed by the Columbia River region, Washington coast, and British Co- lumbia. None was reported from Alaska, but 10 (2%) were from California ports. All marked Cali- fornia fish were recovered from the Columbia River region and farther south. Most (87%) were landed in Oregon. The few recoveries of Califor- nia fish off California is undoubtedly influenced by the larger size limits in this than other fish- eries. Table 6. — Recoveries of coded wire tagged juvenile coho in the ocean fisheries 1 977-83. Estimated total numbers are in parentheses Landing area »«.. 5- & ^ i O^ i #«"^ Release area CO' /I" / S.E. Alaska 20 (39) British 2 1,086 24 2 Columbia (2) (1,735) (90) (8) Puget Sound 201 1,352 9 1 2 (729) (5,262) (40) (2) (9) Washington 1 71 7 125 42 2 Coast (5) (316) (42) (451) (151) (77) Columbia 2 24 4 39 67 31 River (133) (14) (162) (310) (164) Oregon Coast 18 3 21 62 308 10 (107) (13) (83) (213) (1,137) (45) California 1 138 19 (4) (552) (200) 'Sports catches are not expanded. The estimated total number CWTs recovered in the sports fisheries. Preliminary data. expanded commercial catch + actual number 189 FISHERY BULLETIN: VOL. 86, NO. 2 In general, the legal size limits increased from north to south, which could result in more recov- eries of juvenile coho salmon in northern than southern regions. Thus, these data do not provide evidence that a large proportion of juvenile coho salmon from British Columbia and waters to the south made northward migrations into Alaskan waters before or during the commercial and sports salmon seasons. Movements of Washing- ton, Columbia River, and Oregon fish into British Columbian waters were common however. Hunter (1985) expanded the catches of CWT juvenile coho salmon caught along the west coast of North America during 1978-80 to the total landed plus estimated "drop-off' mortality (fish that were hooked and died without being landed) of both tagged and untagged hatchery groups. Calculations of the percentage returns from dif- ferent release and recapture areas are similar to ours (Table 6). The highest percentage of returns were from the areas of release for all areas except for California, and a higher proportion of the catches of Washington coastal, Puget Sound, Co- lumbia River, and Oregon coastal stocks were re- ported north than south of the area of release. Are Juvenile Coho Highly Migratory? Based on our observations on movements of marked fish, north-south and seasonal trends in abundance and size, and directional purse seine sets during the summer, we conclude that many juvenile coho salmon from Oregon and Washing- ton coastal streams and the Columbia River are transported by currents to the south in May and June but then migrate north later in the summer. The mean catches per set of yearling coho salmon in August and September are a large fraction of those in June, indicating that in the years studied many juvenile coho salmon in coastal waters of Oregon and Washington were not highly migra- tory. Moreover, more marked juvenile hatchery coho salmon were caught in ocean fisheries in the region of release than in distant waters. Recover- ies of juvenile coho salmon released from hatch- eries south of Cape Flattery were rare in northern waters off Alaska and relatively few were recov- ered in British Columbia (Table 6). In addition, the positive correlation between upwelling off Or- egon and survival of hatchery coho salmon from the Columbia River, Oregon, and California (Nickelson 1986) also argues for a close coupling of OPI coho salmon with a local, not a distant, environmental event during the time that year- class survival is determined. All of these trends suggest that most juvenile coho salmon from this area are not highly migratory and that many usu- ally remain in coastal waters near their sites of ocean entry during their first summer in the ocean, and perhaps during their entire ocean life. In years of unfavorable ocean conditions, how- ever, movements may be more extensive or mor- tality may be higher, as suggested by the very low catches of juvenile coho salmon in purse seine sets south of Cape Flattery during September 1983, the year of the recent strong El Nino. Although Pacific salmon are renown for their long foraging migrations in the subarctic Pacific, coho salmon demonstrate both nonmigratory and highly migratory behavior. Milne (1950) found immature coho salmon almost year-round in Georgia Strait and concluded that two types of coho salmon exist in British Columbia waters: "ocean" and "inshore" types, the "ocean" type spending most of its ocean life in coastal and off- shore waters and the "inshore" type in inside waters such as Georgia Strait. Healey (1978) caught "inshore" juvenile coho salmon in purse seines in Georgia Strait during summer, fall, and winter months. Similarily, large numbers of coho salmon originating from streams of Puget Sound remain in the Sound throughout their marine life (Haw et al. 1967). Young coho salmon have also been found in the winter and spring, many months after seawater entry in Yaquina Bay (Myers 1980) and other Oregon estuaries (J. Nicholas^). Hartt and Dell (1986), in their im- pressive study of juvenile salmonids of the north- eastern Pacific during 1956-70, recognized these two migratory patterns of coho salmon. They found juvenile coho salmon in waters off Vancou- ver Island and in the Strait of Juan de Fuca throughout the summer and fall, and concluded that some coho salmon spend their entire marine life in "inside" waters and make only limited ocean migrations. What Proportion of Juvenile Coho from Oregon and Washington Migrate North? The tagging experiments reported by Hartt and Dell (1986) and Godfrey (1965) provide convinc- ing evidence for long-distance migrations of coho ■IJ. Nicholas, Oregon Department of Fish and Wildlife, Cor- vallis, OR, 97331, pers. commun. May 1986. 190 PEARCY and FISHER: MIGRATIONS OF COHO SALMON salmon during their first summer in the ocean. Based on recoveries of maturing coho salmon that were tagged a year earlier at sea during April- October 1956-70, Hartt and Dell (1986) con- cluded that juvenile coho salmon from the Colum- bia River, Oregon, and California may form a large proportion of the coho stocks that migrate north along the coast each summer. Of the 70 recoveries of tagged fish that were released be- tween Kodiak Island and 56°N, 37% were recov- ered the following year in the area of the Colum- bia River and Oregon-California; of the 59 recoveries offish released between 56°N and Cape Flattery, 47% were recovered in these southern regions. In all, 25% of the recoveries were from Oregon-California, 16% from the Columbia River, 14% from Washington, 33% from British Columbia, and 12% from Alaska. Loeffel and Forster (1970) concluded that pat- terns of radioactive ^^Zn in juvenile coho salmon collected in the northeastern Pacific supported the concept of a northerly migration from Oregon and Washington into the Gulf of Alaska during the summer. They found that juvenile coho salmon captured off the west coast of Vancouver Island in June and July 1967 contained ^^Zn, pre- sumably originating from neutron activation of Columbia River water used to cool the nuclear reactors at Hanford, WA. ^^Zn levels decreased in fish caught farther to the north (54°42'N- 58°24'N) in July-September of 1967. The authors thought the low concentrations in northern sam- ples represented background levels and that fish with relatively high levels of ^^Zn had associated with the Columbia River plume and subsequently migrated north from the Oregon- Washington re- gion. They found low ^^Zn levels in 1968, how- ever, and no pronounced latitudinal gradients. Furthermore, they reported none of the many fin- marked juvenile coho salmon released from Ore- gon and Washington hatcheries in 1967 and 1968 north of Juan de Fuca Strait. Hence their evi- dence for northward movements of Columbia River or Oregon-Washington coho salmon was equivocal. During June and July 1984, research was con- ducted with the NMFS Auke Bay Laboratory in waters from northern California to southwest Alaska from the FV Pacific Warwind and Bering Sea , both making round hauls with the same size of purse seine, to sample juvenile coho in waters north of Oregon and Washington: 37 sets were made in coastal waters of British Columbia, and 39 were made in coastal waters and 29 in inland waters (bays, inlets, and ^ords) of southeastern Alaska. Of the 371 juvenile coho salmon captured in these regions, 77% were caught in inland waters of southeastern Alaska. The seven CWT juvenile coho salmon captured were all from Alaska inland waters and all originated from Alaska hatcheries (Auke Bay Laboratory 1984a). A later cruise in southeast Alaska by the Auke Bay Laboratory in August 1984 caught eight CWT coho salmon, also all from inland waters and from Alaskan hatcheries (Auke Bay Labora- tory 1984b). Of the 14 CWT juvenile coho salmon collected in other purse seines, gill nets, and special troll gear in waters of southeastern Alaska during 1982, 1983, and 1985, 12 originated from Alaska hatcheries and 2 originated from Washington hatcheries (Auke Bay Laboratory 1983; Jaenicke et al. 1984; Orsi et al. 1987). Table 6 shows that only 5 of 25 CWT juvenile coho salmon caught in Alaskan waters during 1977-83 were from hatch- eries south of Alaska, indicating that most juve- nile coho salmon caught in southeastern Alaska during the summer originated from Alaska and not from southern regions. Hartt (1980) and Hartt and Dell (1986) recog- nized that their data did not indicate the propor- tion of southern stocks that made northerly mi- grations, but they concluded that a large proportion is probable, since juvenile coho salmon were consistently caught in most seine sets throughout the area sampled. They estimated that the average density of juvenile salmonids in coastal waters between 56°N and 60°N off south- eastern Alaska during August and September 1964-68 was 1,500 km"^. The average density of juvenile coho salmon in this area during these two months was only 82 km~^ (Hartt and Dell 1986, app. A). During August and September 1981-84, the average density of juvenile coho salmon in our round hauls between Cape Flat- tery, WA and Cape Arago, OR to 37 km offshore was 350 km"^, several times the estimates of Hartt and Dell for the same months of the year. This suggests that juvenile coho salmon may be found in higher densities off Oregon and Wash- ington than southeastern Alaska during late summer, assuming that distributions and abun- dances in the late 1960s and early 1980s were similar. This trend for higher abundances of juve- nile coho salmon off Oregon and Washington than in coastal waters farther north was also found during July 1984 (Table 7), although average catches off Washington and Oregon were not as 191 FISHERY BULLETIN: VOL. 86, NO. 2 Table 7. — Average catches of juvenile coho salmon in purse seines sets in coastal waters along the west coast of North America during July 198412 Area No. per set No. per km2 Sitka-Juneau Ketchikan-Sitka Cape Scott-Dixon Entrance Vancouver Is. Washington Oregon No. California 1.29 0.58 1.90 1.80 3.76 2.59 68 3 100 95 198 136 'Cruise Report. Drum Seiner FV Bering Sea, Coastwide NWAFC/OSU Cooperative Study, Ecology of Juvenile Salmon in Coastal and Inside Waters of Soutfieast Alaska, 28 June-26 July 1984, NWAFC Auke Bay Laboratory, National Marine Fisfierles Service, NCAA, P.O. Box 115, Auke Bay, AK 99821 . 2Fisher and Pearcy (1984). large as in some earlier years owing to low sur- vival (Fig. 2; Pearcy 1984; Fisher and Pearcy in press). Comparisons of the estimates of total juvenile yearling coho salmon abundances off Oregon and Washington with the production of coho smolts in the Oregon Production Area (Columbia River to California) also suggests that many juvenile coho resided off Oregon and Washington during the summer. By expanding our catches per m^ to the region sampled, we estimated that the numbers of juvenile yearling coho salmon in areas surveyed by our purse seine sampling during August or September 1981-84 were 6.3%, 6.5% 5.1%, and 5.2%, respectively, of the numbers of hatchery and wild smolts released in the Columbia River and in Oregon (T. Lichatowich^). The areas in- cluded in these estimates were roughly 83%, 62%-, 51%, and 68% of the total area from Cape Flattery to Cape Arago out to 37 km offshore. Recognizing that the entire area was not sampled, that year- class strength of coho salmon in this region is probably established soon after ocean entrance (Fisher and Pearcy in press), and that early marine mortality may be inversely related to size (Parker 1968; Ricker 1976) so that much of the ocean mortality has occurred by late summer, these percentages probably represent a substan- tial portion of the surviving OPI coho smolts. In fact, they are several times higher than the smolt-to-adult survival of 1.3-2.8% for OPI public hatchery coho salmon (excluding Rogue River and California hatcheries) during 1981-84 (R. Kaiser^). 5T. Lichatowich, Oregon Department of Fish and Wildlife, P.O. Box 59, Portland, OR 97207, pers. commun. September 1987. 6R. Kaiser, Oregon Department of Fish and Wildlife, Hatfield Marine Science Center, Newport, OR 97265, pers. commun. September 1987. We conclude, therefore, that a major fraction of the juvenile coho salmon from Oregon and Wash- ington hatcheries did not undertake distant mi- grations into the Gulf of Alaska in recent years. This is not necessarily in conflict with Hartt and Dell's (1986) data, since they established the pres- ence of Oregon and Washington coho salmon in northern waters but not the proportion of total production that undertakes this migration. On the other hand, neither the stocks of coho nor oceanographic conditions have remained con- stant over the period from 1956 to 1985 when these two studies were conducted. Wild coho smolts exceeded hatchery smolts in the Oregon Production Area before 1961 (Nickelson 1986) but comprised <12% of the smolts in 1980-85 (R. Kaiser fn. 6). Perhaps wild smolts from the OPI area had different migratory patterns than hatch- ery smolts do today and migrated rapidly into northern waters soon after ocean entrance. This may explain Nickelson's (1986) finding that sur- vival of hatcherv, but not wild coho smolts, was significantly correlated with coastal upwelling off Oregon. Ocean conditions have also changed over this period. The late 1960's were accompanied by strong upwelling along the coast compared to weak upwelling in the early 1980's (Nickelson 1986; Mason and Bakun 1986). McLain (1984), Norton et al. (1985), and Royer (1985) illustrated that sea surface temperatures and sea levels in- creased in the northeastern Pacific between 1976 and 1984. These factors and associated changes in ocean circulation could explain differences in mi- gratory behavior of coho salmon between 1960's and 1980's. If currents provide orientational cues to migration, cues facilitating northerly move- ments may be reduced during years of weak up- welling and weak Ekman transport from the north. Ocean conditions may have modified mi- gratory patterns, as they possibly have for the migration of Fraser River sockeye salmon around Vancouver Island (Groot et al. 1984; Hamilton 1985). ACKNOWLEDGMENTS We thank D. Larden and his crew of the FV Pacific Warwind and Bering Sea for their cooper- ation and competence during purse seining opera- tions at sea; A. Chung, R. Brodeur, J. Shenker, W. Wakefield, D. Gushee, C. Banner, J. Long, K. Krefft, and C. Wilson for their help on cruises; J. Norton and the Oregon Department of Fish and 192 PEARCY and FISHER: MIGRATIONS OF COHO SALMON Wildlife Clackamas Laboratory for decoding coded-wire tags; K. Johnson for providing data on marked fish; W. McNeil and R. Severson for as- sisting in fluorescent spray-marking Oregon Aqua-Food's Inc. smolts; Northwest and Alaska Fisheries Center for the loan of a purse seine; Northwest and Alaska Fisheries Center Auke Bay Laboratory for the loan of electronarcosis equipment; and the Faculty of Fisheries, Hok- kaido University, and the TV Oshoro-Maru for conducting gill net research. A. Hartt, H. Jaenicke, R. Brodeur, R. Gowan, D. Hankin, and L. Botsford all provided helpful comments on the manuscript. This research was supported by Oregon State University Sea Grant College Program (Grant No. NA 81AA-D-00086, Project R/OPF-17), the Northwest and Alaska Fisheries Center (Con- tract 81-ABC-00192, 83-ABC-00102, 84-ABC- 0009, and NA-85-ABH0002J), the Oregon De- partment of Fish and Wildlife, and Oregon Aqua-Foods, Inc. LITERATURE CITED Auke Bay Laboratory. 1983. Cruise report. NOAA fishery research vessel John N. Cobb, JC-83-03. Juvenile salmon purse seining proj- ect in coastal and inside waters of southeastern Alaska. NWAFC Auke Bay Lab., Natl. Mar. Fish. Serv., NOAA, 30 p. 1984a. Cruise report drum seines. F/V Bering Sea. Coastwide NWAFC/OSU Cooperative study. Ecology of juvenile salmon in coastal and inside waters of south- eastern Alaska, 28 June-26 July 1984. NWAFC Auke Bay Lab., Natl. Mar. Fish. Serv., NOAA, 9 p. 1984b. Cruise report, NOAA RV John N. Cobb Cruise 84-03. Juvenile salmon purse seining project in coastal and inside waters of southeastern Alaska, July 31-Au- gust31, 1984. NWAFC Auke Bay Lab., Natl. Mar. Fish. Serv., NOAA. Chung, A 1985. Relationships between oceanographic factors and the distribution of juvenile coho salmon (Oncorhynchus kisutch) off Oregon and Washington, 1982-83. M.S. Thesis, Oregon State Univ., Corvallis, 116 p. Clutter. R I., and L E. Whitesel. 1956. Collection and interpretation of sockeye salmon scales. Int. Pac. Salmon Comm. Bull. 9:1-159. Dawley. E. M , R. D Ledgerwood, T H Blahm, C. W. Sims, J. T. DuRKiN, R A Kirn, A E Rankis, G E Monan, and F J. OSSIANDER 1986. Migrational characteristics, biological observa- tions, and relative survival of juvenile salmonids enter- ing the Columbia River estuary, 1966-1983. Final Report, Bonneville Power Administration Project No. 81-102. Northwest and Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Seattle, 256 p. Dawley. E M . R. D. Ledgerwood, and A. Jensen 1985. Beach and purse seine sampling of juvenile salmonids in the Columbia River estuary and ocean plume, 1977-1983. Vol. 2: Data on marked fish recover- ies. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/ NWC-75, 397 p. Dawley, E M., C. W Sims. R. D. Ledgerwood, D. R. Miller, and J G Williams 1981. A study to define the migrational characteristics of chinook and coho salmon in the Columbia River estuary and associated marine waters. Final Rep. Northwest and Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, 68 p. Fisher, J P , and W G Pearcy 1984. Studies of juvenile salmonids off northern Califor- nia, Oregon, Washington and Vancouver Is- land. 1984. Oreg. State Univ. Sea Grant Coll. Pro- gram (ORESU-T-OGl), 24 p. 1985. Studies of juvenile salmonids off the Oregon and Washington coast, 1985. Oreg. State Univ. Sea Grant Coll. Program (ORESU-T-004), 31 p. In press. Growth of juvenile coho salmon [Oncorhynchus kisutch ) off Oregon and Washington, USA in years of differing coastal upwelling. Can. J. Fish. Aquat. Sci. Fisher, J P , W G Pearcy, and A. W Chung 1982. Studies of juvenile salmonids off the Oregon and Washington coast, 1982. Oreg. State Univ. Sea Grant Coll. Program (ORESU-T-83-003), 41 p. 1983. Studies of juvenile salmonids off the Oregon and Washington coast, 1983. Oreg. State Univ. Sea Grant Coll. Program (ORESU-T-001), 29 p. FOESTER, R. E. 1968. The sockeye salmon Oncorhynchus nerka. Fish Res. Board Can. Bull. 162, 422 p. French, R. R., R G Bakkala. and D. F. Sutherland. 1975. Ocean distribution of stocks of Pacific salmon, Oncorhynchus spp., and steelhead trout, Salmo gaird- nerii, as shown by tagging experiments. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF 689, 89 p. Godfrey, H 1965. Salmon of the North Pacific Ocean. Part IX Coho, chinook and masu salmon in offshore waters. 1. Coho salmon in offshore waters. Int. North Pac. Fish. Comm. Bull. No. 16, 39 p. Godfrey, H., K A Henry, and S. Machidori. 1975. Distribution and abundance of coho salmon in off- shore waters of the North Pacific Ocean. Int. North Pacific Fish. Comm. Bull. No. 31, 80 p. Groot, C , L Margolis, and R. Bailey. 1 984 . Does the seaward migration of Eraser River sockeye salmon (Oncorhynchus nerka) smolts determine the route of return of adults? In J. D. McCleave, G. P. Arnold, J. J. Dodson, and W. H. Neill (editors). Mecha- nisms of migration in fishes, p. 283-292. Plenum Press, N.Y. GUNSOLUS, R. T. 1978. The status of Oregon coho and recommendations for managing the production, harvest, and escapement of wild and hatchery reared stocks. Oreg. Dep. Fish. Wildl. Fish Div. Proc. Rep., 59 p. Hamilton, K. 1985. A study of the variability of the return migration route of Eraser River sockeye salmon (Oncorhynchus nerka). Can. J. Zool. 63:1930-1943. Hartt, A C 1975. Problems in sampling Pacific salmon at sea. Int. North Pac. Fish. Comm. Bull. 32:165-231. 193 FISHERY BULLETIN: VOL. 86, NO. 2 1980. Juvenile salmonids in the oceanic ecosystem — the critical first summer. In W. J. McNeil and D. C. Himsworth (editors), Salmonid ecosystems of the North Pacific, p. 25-57. Oreg. State Univ. Press. Hartt, a. C , AND M. B Dell. 1986. Early oceanic migrations and growth of juvenile Pacific salmon and steelhead trout. Int. North Pac. Fish. Comm. Bull. 46:1-105. Haw. F . H. O. Wendler, and G Deschamps. 1967. Development of Washington State salmon sports fishery through 1964. Wash. Dep. Fish. Res. Bull. 7, 191 p. Healey, M. C. 1978. The distribution, abundance, and feeding habits of juvenile Pacific salmon in Georgia Strait, British Colum- bia. Dep. Fish. Environ. Can., Fish. Mar. Serv. Tech. Rep. No. 788, 49 p. HiCKEY, B M 1979. The California Current System - hypotheses and facts. Prog. Oceanogr. 8:191-279. Hunter. M. A. 1985. The 1976-1978 brood coho model. Wash. Dep. Fish. Prog. Rep. No. 222, 146 p. HUYER. A 1983. Coastal upwelling in the California Current Sys- tem. Prog. Oceanogr. 12:259-284. HUYER, A , R O PiLLSBURY, AND R L. SMITH. 1975. Seasonal variation of the alongshore velocity field over the continental shelf off Oregon. Limnol. Oceanogr. 20:90-95. HuYER, A., AND R. L. Smith. 1985. The signature of El Niiio off Oregon, 1982-1983. J. Geophy. Res. 90:7133-7142. HUYER, A., E. J. C. SOBEY, AND R. L. SMITH. 1979. The spring transition in currents over the Oregon continental shelf J. Geophy. Res. 84:6995-7011. Jaenicke. H. W., R D Brodeur, and T Fujii. 1984. Exploratory gillnetting from the Oshoro-Maru for juvenile salmonids off southeastern Alaska, 24-25 July 1982. Bull. Fac. Fish. Hokkaido Univ. 35:154-160. Johnson, S L. 1984. The effects of the 1983 El Nirio on Oregon's coho and Chinook salmon. Oreg. Dep. Fish Wild. Info. Rep. 84-8, 40 p. KuNDU, P. K., AND J S Allen 1976. Some three-dimensional characteristics of low- frequency fluctuations near the Oregon coast. J. Phys. Oceanogr. 6:181-199. LOEFFEL, R. E., AND W. O FORSTER 1970. Determination of movement and identity of stocks of coho salmon in the ocean using the radionuclide zinc- 65. Res. Rep. Fish Comm. Oreg. 2:1-12. Lord. G . W C Acker. A. C Hartt. and B J Rothschild. 1976. An acoustic method for high-seas assessment of mi- grating salmon. Fish. Bull., U.S. 74:104-111. Machidori, S 1966. Vertical distribution of salmon (Genus Oncor- hynchus) in the North-western Pacific. Hokkaido Reg. Fish. Res. Lab. 31:11-17. Manzer, J I 1964. Preliminary observations on the vertical distribu- tion of Pacific salmon (Genus Oncorhynchus ) in the Gulf of Alaska. J. Fish. Res. Board Can. 25:1085-1089. Mason, J. E , and A Bakun 1986. Upwelling index update, U.S. West Coast, 33N- 48N latitude. U.S. Dep. Commer., NOAA Tech. Memo. NMFS No. 67, 81 p. McLain, D R 1984. Coastal ocean warming in the Northeast Pacific, 1976-83. In W. G. Pearcy (editor), The infiuence of ocean conditions on the production of salmonids in the North Pacific, p. 61-86. Oreg. State Univ. Sea Grant Coll. Program (ORESU-W-001). McCarl, B. a., and R. B. Rettig 1983. Influence of hatchery smolt releases on adult salmon production and its variability. Can. J. Fish. Aquat. Sci. 40:1880-1886. McGlE.A M 1984. Evidence for density dependence among coho salmon stocks in the Oregon Production Index area. In W. G. Pearcy (editor). The influence of ocean conditions on the production of salmonids in the North Pacific, p. 37-49. Oreg. State Univ. Sea Grant Coll. Program (ORESU-W-83-001). Miller, D R., J G. Williams, and C. W. Sims. 1983. Distribution, abundance and growth of juvenile salmonids off Oregon and Washington, summer 1980. Fish Res. 2:1-17. Milne, D J 1950. The difference in the growth of coho salmon on the east and west coasts of Vancouver Island in 1950. Fish. Res. Board Can. Prog. Rep. Pac. Coast Biol. Sta. 85:80- 82. Myers, K, W. W. 1980. An investigation of the utilization of four study areas in Yaquina Bay, Oregon, by hatchery and wild juvenile salmonids. M.S. Thesis, Oregon State Univ., Corvallis, 234 p. NiCKELSON, T. E. 1986. Influences of upwelling, ocean temperature, and smolt abundance on marine survival of coho salmon (Oncorhynchus kisutch ) in the Oregon Production Area. Can. J. Fish. Aquat. Sci. 43:527-535. Norton, J , D McLain. R Brain, and D Husby 1985. The 1982-83 El Nirio event off Baja and Alta Cali- fornia and its effect on ocean climate context. In W. S. Wooster and D. L. Fluharty (editors). El Niiio North, p. 44-72. Univ. Wash. Sea Grant Program. Oregon Department of Fish and Wildlife. 1982. Comprehensive plan for production and manage- ment of Oregon's anadromous salmon and trout. Part II. Coho salmon plan. Oreg. Dep. Fish Wildl., var. p. Orsi, J A., A G Celewycz, D. G Mortensen. and K A Herndon 1987. Sampling juvenile chinook salmon (Oncorhynchus tshawytscha) and coho salmon (O. kisutch) by small trolling gear in the northern and central regions of south- eastern Alaska, 1985. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-115, 47 p. Pacific Fisheries Management Council 1986. Preseason Report 1. Stock abundance analysis for 1986 ocean salmon fisheries. Pac. Fish. Manage. Counc, var. p. Pacific Marine Fisheries Commission. 1980. 1977 Pacific salmonid coded wire tag recoveries. Regional Mark Processing Center for Pacific Coast States, Portland, OR. 1981. 1978 Pacific salmonid coded wire tag recoveries. Regional Mark Processing Center for Pacific Coast States, Portland, OR. 1984a. 1979 Pacific salmonid coded wire tag recoveries. Regional Mark Processing Center for Pacific Coast 194 PEARCY and FISHER: MIGRATIONS OF COHO SALMON States, Portland, OR. 1984b. 1980 Pacific salmonid coded wire tag recoveries. Regional Mark Processing Center for Pacific Coast States, Portland, OR. 1984c. 1981 Pacific salmonid coded wire tag recoveries. Regional Mark Processing Center for Pacific Coast States, Portland, OR. 1985a. 1982 Pacific salmonid coded wire tag recoveries. Regional Mark Processing Center for Pacific Coast States, Portland, OR. 1985b. 1983 Pacific salmonid coded wire tag recoveries. Regional Mark Processing Center for Pacific Coast States, Portland, OR. Parker. R R 1968. Marine mortality schedules of pink salmon of the Bella Coola River, Central British Columbia. J. Fish. Res. Board Can. 25:757-794. Pearcy. W G 1984. Where do all the coho go? The biology of juvenile echo salmon off the coasts of Oregon and Washington. In W. G. Pearcy (editor). Influence of ocean conditions on the production of salmonids in the North Pacific, p. 50- 60. Oreg. State Univ. Sea Grant Coll. Program (ORESU-W-83-001). Pearcy, W G , R D Brodeur, J M Shenker, W. W Smoker, and Y Endo In press. Food habits of Pacific salmon and steelhead trout, midwater trawl catches, and oceanographic condi- tions in the Gulf of Alaska, 1980-1985. Bull. Ocean Res. Inst. Univ. Tokyo. Pearcy, W . J Fisher. R Brodeur, and S Johnson 1985. Effects of the 1983 El Niiio on coastal nekton off Oregon and Washington. In W. S. Wooster and D. L. Fluharty (editors). El Nino North, p. 188-204. Univ. Wash. Sea Grant Program. Pearcy. W G , and A Schoener 1987. Changes in the marine biota coincident with the 1982-1983 El Nino in the northeastern subarctic Pacific Ocean. J. Geophys. Res. 92, No. 013:14417-14428. Peterman, R M . and R D Roittledge 1981. Experimental management of Oregon coho salmon (Oncorhynchus kisutch ): Designing for yield of informa- tion. Can. J. Aquat. Sci. 40:1212-1223. Phinney, D E , D M Miller, and M L Dahlberg. 1967. Mass-marking young salmonids with fluorescent pigment. Trans. Am. Fish. Soc. 96:157-162. Richer. W. E 1976. Review of the growth rate and mortality of Pacific salmon in salt water, and non-catch mortality caused by fishing. J. Fish. Res. Board Can. 33:1483-1524. Royer, T. C. 1985. Coastal temperature and salinity anomalies in the northern Gulf of Alaska, 1970-84. In W. S. Wooster and D. L. Fluharty (editors), El Nino North, p. 107-115. Univ. Wash. Sea Grant Program. SCARNECCHIA, D. L. 1981. Effects of streamflow and upwelling on yield of wild coho salmon (Oncorhynchus kisutch ) in Oregon. Can. J. Fish. Aquat. Sci. 38:471-475. SUSUKI, T., AND H SONODA 1972. On a trial fish finder for salmon and its experimen- tal results. Bull. Jpn. Soc. Sci. Fish. 38:463-469. Wakefield, W. W., J. P. Fisher, and W G Pearcy. 1981. Studies of juvenile salmonids off the Oregon and Washington coast, 1981. Oreg. State Univ. Sea Grant College Program Ref 81-13, 51 p. Ware, D M 1978. Bioenergetics of pelagic fish: Theoretical change in swimming speed and ration with size. J. Fish. Res. Board Can. 35:220-228. Weihs, D 1973. Optimal fish cruising speed. Nature 245:48-50. 195 EARLY MARINE LIFE OF ATLANTIC SALMON, SALMO SALAR, POSTSMOLTS IN THE NORTHERN GULF OF ST. LAWRENCE J -D DUTIL AND J.-M. COUTU' ABSTRACT Postsmolts of Atlantic salmon, Salmo salar, having spent some 2-4 months in the marine environ- ment were collected in the nearshore zone of the northern Gulf of St. Lawrence. From a back- calculated smolt length of 135 mm they had increased in length to 265 mm (212 g) on 1 September, and 306 mm (320 g) on 30 September. The rate of increase in length averaged 1.65 mm/day over more than 2 months. Individuals reached 35 cm and 500 g in late autumn. Postsmolts moved in small shoals near the surface and were possibly more active at dawn and dusk. In midsummer, stomach contents changed from insects and gammarids to sand lance, 40-100 mm in length; vertically migrating crustaceans also occurred in the stomachs in autumn. River origin of these postsmolts is not known. The possibility of their belonging to a particular subgroup of some north shore stocks is examined in relation to scale patterns and size of gonads. This occurrence of postsmolts near shore in late summer and presumably their late movement out of the Gulf of St. Lawrence indicate the directional nature of smolt migration to distant feeding areas should be reconsidered. Low sea temperature is hypothe- sized to trigger the movement out of the Gulf of St. Lawrence. Continual presence of postsmolts in a shallow layer at the surface could prove to be extremely valuable in forcasting movements and production. Many papers have been published concerning the biology of Atlantic salmon, Salmo salar, but very little has been said concerning the postsmolt stage. This stage has been defined as ". . .the juvenile salmon from the time that it leaves the river as a smolt until the onset of wide annulus formation on the scales at the end of the first winter in the sea" (Allan and Ritter 1977). This paper presents new data on the Atlantic salmon postsmolts (hereafter referred as postsmolts) in the northern Gulf of St. Lawrence, reviews our current knowledge on the biology of postsmolts, and points to biological and environmental fac- tors potentially limiting the success of their early life in marine environments. Published data on postsmolts are mainly lim- ited to stocks in the Baltic Sea. Routes of migra- tion have been described based on the locations and time of early recapture from smolt releases in Sweden and Finland (Carlin 1959; Larsson and Ateshkar 1979; Ikonen and Auvinen 1984, 1985; Jutila and Alapassi 1985). Data on predators are limited (Soikkeli 1973; Valle 1985), and most of the material concerns predation on smolts in streams and estuaries (Larsson 1985). Many analyses of stomach contents have been pub- 'Ministere des Peches et des Oceans, Gouvernement du Canada, 850, route de la Mer, C.P. 1000, Mont-Joli, Quebec, Canada G5H 3Z4. lished, particularly on smaller postsmolts (re- viewed by Christensen and Larsson 1979; Jutila and Toivonen 1985). Data on rate of growth (Ikonen and Auvinen 1985) and rate of mortality (Carlin 1959) are lacking. However, Baltic salmon spend their entire sea life in the brackish waters of the Baltic and nearly 80% of smolt pro- duction originates from hatcheries (Anonymous 1984). Thus the information derived from salmon in the Baltic should be extended to other stocks only with caution. Publications on postsmolts in the northern At- lantic and Gulf of St. Lawrence mentioned small salmon as bycatches of commercial fisheries and described the distribution of recaptures from smolt release programs. The earliest report on postsmolts in the Gulf of St. Lawrence claimed that small salmon, referred to as "ouananiche" by local fishermen, were regularly taken near shore in herring nets in autumn (Comeau 1909). Kendall (1935) also reported such incidental catches for the New England coast. Elson (1953) recorded a bycatch of more than 1,000 marked postsmolts from one locality in the Bay of Fundy in the period 1951-53 and reported their mean length. There are also limited records of postsmolts taken off France (Vibert 1953) and in the Gulf of St. Lawrence (Caron 1983) from smolts tagged in streams. Recently, information on movements has been derived from tag returns Manuscript accepted December 1987. FISHERY BULLETIN; VOL. 86, NO. 2, 1988. 197 FISHERY BULLETIN: VOL. 86, NO. 2 of smolts released in spring in New England and caught in summer in Canada (Meister 1984). Ru- mors of bycatch in herring nets along the coast of the northern Gulf of St. Lawrence in autumn pro- vided an occasion to acquire some knowledge con- cerning the elusive postsmolt. Production of salmon in the sea may well be limited by the success of smolts in the marine environment. Materials and Methods Postsmolts were collected between Bale Trinite and Riviere-au-Tonnerre in the northern Gulf of St. Lawrence, in 1982, 1983, 1984, and 1985 (Fig. 1). Fishermen contacted in summer 1982 col- lected postsmolts in late summer and autumn as bycatch in herring gill nets. Four of them were asked in 1983 to monitor the catch of postsmolts in experimental gill nets in late-September in 4 locations (Bale Trinite, Riviere Pentecote, Port Cartier, and Sept-Iles). We also monitored sta- tions in Bale Trinite in 1983 (23 September-11 October), in Baie Trinite and Port Cartier in 1984 (21 August-20 October), and in Sept-Iles in 1985 (20 August-4 October). Finally, smolts and early postsmolts were collected in seines in June and July, during an eel marking program in the estu- ary of Grande Trinite River at Baie Trinite (Fig. 1). Fishermen used standard herring gill nets in 1982, but custom-made gill nets were used in 1983, 1984, and 1985. Custom-made gill nets had 5 sections of increasing mesh sizes (50.8, 57.2, 63.5, 69.9, and 76.2 mm stretched) covering the range in mesh sizes of herring gill nets in the northern Gulf of St. Lawrence. Stretched mesh sizes were determined by measuring 10 meshes per section. Sections were 6 m deep and 10 m long. In 1983, 1984, and 1985, postsmolts were recorded by section individually. In 1984 and 1985, their position in the nets was recorded more precisely: floating lines had numbered buoys and a string divided the nets into 2 halves horizon- tally. Time of the catch was also recorded. The gill nets were usually visited at 2-h intervals be- tween 0600 and 1800, as sea conditions allowed. They were left fishing overnight. Gill nets were all set at the surface and near shore (<2 km). Mean air temperatures for 1982 to 1985 were drawn from Environment Canada meteorological summaries for Sept-Iles airport. Temperature recorders were also tied to nets in 1983, 1984, and R.MOISIE SEPT-ILES PORT CARTIER. R.PENTECdTE*/ NORTH SHORE GULF OF ST. LAWRENCE Figure 1. — Locations of the areas investigated in the northern Gulf of St. Lawrence (shaded area). 198 DUTIL AND COUTU: EARLY LIFE OF ATLANTIC SALMON 1985 to get hourly records of temperature near the surface, but portions of these records are miss- ing. Fish were preserved at -15°C for less than 4 months in 1982 and less than 2 months in 1983, 1984, and 1985. Fork length was measured to nearest mm and weight to nearest g. Condition index values were calculated as C/ = sin 1 100 W 0.5 CI = condition index; W = weight (g); L = length (mm); s = slope of the length- weight relationship for age groups combined. Scales, usually taken below the dorsal fin close to the lateral line, were cleaned and mounted on glass slides. Stomachs were preserved in formalin (10%) or alcohol (70%) to be analyzed later for their contents. The sex was determined visually and checked histologi- cally on a subsample of postsmolts collected in 1982. Gonads were weighed to nearest mg and only in 1982. Gonadosomatic index (GSI) was cal- culated as GSI = GW TW-GW 100 where GW = gonad weight; TW - total weight. Scales showing no focal regeneration were ex- amined for age determination by 2-4 readers. Readings were discussed and a consensus was reached in most cases. Reported ages are smolt ages plus 2-4 months. Fork length at smoltifica- tion (mm) was back-calculated as LS = (LC - 34) RS RC + 34 where LS = back-calculated smolt length; LC = body length of post smolt; RS = scale radius at smolt check; RC = scale radius of postsmolt. Length at scale formation is assumed to be 34 mm. Scales were also examined for any mark in the postsmolt zone that could be of potential in- terest. Hence the number of individuals showing a transition zone between the riverine and marine increments was determined in 1982, and the number of individuals showing a summer check was determined in 1982, 1983, 1984, and 1985. Scale radii were measured along the postero-anterior axis on 4 scales per individual in 1982, and on 1 scale per individual in 1983, 1984, and 1985. Projected scales were measured at con- stant magnification on a digitizing pad connected to a personal computer. Fork length at formation of a summer check (mm) was back-calculated for 1982 and 1983 as LF = (LC - 34) • RM RC + 34 where LF = back-calculated length; LC = body length of postsmolt; RM = scale radius at summer check; RC = scale radius of postsmolt. Stomach contents were examined in detail in 1982. Results are expressed as percentage of oc- currence (number of stomachs containing an item divided by the number of stomachs examined) and percentage by number (the count number of an item divided by the count number of all items). Food items that could not be identified are re- ferred to as remains. In as far as their condition allowed, prey fish were usually measured to the nearest cm. Stomach contents examined in 1983, 1984, and 1985 were consistent with the 1982 conclusions, so the results are not presented. Results The Age and Sex Composition Three-year-old female salmon dominated the catch both in 1982 and 1983 (Table 1). Postsmolts ranged from 2 to 5 years in 1982 and from 2 to 4 years in 1983. Five-year-old salmon made up less than 4% of the catch in 1982. Three-year-old and, to a lesser extent, 4-yr-old salmon dominated the catch both in 1982 and 1983 (82% of the catch); mean smolt age was 3.2 ± 0.7 years. Age composi- tion in 1983 differed from 1982 because of an in- crease in the percentage of 2-yr-old salmon (12% to 21%) and a decrease in the percentage of 5-yr- old salmon (x^ = 15.96, P < 0.01). The sex ratio (male:female) for 1982 and 1983 combined was Table 1 . — The age and sex composition of the catch of postsmolt Atlantic salmon along the north shore of the Gulf of St. Lawrence, 1982 and 1983. Sex Smolt age Year 2 3 4 5 Total 1982 Female 26 114 79 7 226 Male 19 67 49 7 142 1983 Female 25 58 15 98 Male 8 26 23 57 Total Female 51 172 94 7 324 Male 27 93 72 7 199 199 FISHERY BULLETIN; VOL. 86, NO. 2 0.62. This ratio did not change between years (x^ = 7.92, P < 0.01) and was significantly less than 1.0 (binomial test: z = -5.37, P < 0.001). There was no trend between sex ratio and smolt age in 1982 (x^ = 1.21, P = 0.75), but the percentage of males tended to increase with smolt age in 1983 (X^ - 12.67, P - 0.002). Length at Smoltification The mean length at smoltification was esti- mated by back-calculation at 135 ± 15 mm for postsmolts collected in 1982 and 1983. Scale ra- dius was linearly related to body length both in 1982 (4 scales measured per fish, P < 0.0001) and in 1983 (1 scale measured per fish, P < 0.0001); probability values are those of F-tests from anal- ysis of variance. Back-calculated smolt lengths were normally distributed (P = 0.80): 75% of the postsmolts ranged between 120 and 150 mm at smoltification. There were also 7% individuals in the 160-200 mm range. Differences in mean smolt length were not significant between the ages and the sexes both in 1982 (F = 1.84, P = 0.08) and in 1983 (F = 1.05, P = 0.39), and, pooling the ages and the sexes, between 1982 and 1983 (^ = 1.32, P >0.15). Rate of Increase in Size The rate of increase in size of postsmolts was very rapid in summer both in 1982 and 1983. Postsmolts collected in 1982 between mid-August and mid-October in = 383; mean date is 1 Sep- tember) averaged 265 ± 25 mm (range 195-328) and 212 ± 58 g (range 92-389). In 1983, postsmolts collected between mid-September and mid-October (n = 155, mean date is 30 Septem- ber) averaged 306 ± 17 mm (range 258-362) and 320 ± 57 g (range 192-565). There was no differ- ence in mean size between males and females and between age-classes in 1982 (P > 0.40) and 1983 (P > 0.20); probability values are those of F-tests from 2-way analysis of variance. From a mean smolt length of 135 mm and assuming smolts mi- grated to estuaries 15 June, postsmolts grew at a rate of 1.65 mm/day during a 2.5-mo (15 June — 1 September) and a 3.5-mo (15 June-30 Septem- ber) period in 1982 and 1983, respectively. This estimate is conservative because the rate of increase in length tended to decrease late in the sampling period. Postsmolts steadily increased in length and weight in the period mid-August to mid-September 1982 (Figs. 2, 3). From mid- September, the rate tended to slow down. The inclusion of data for 1983, collected later in the autumn, corroborates this observation indicating that conditions changed in late-September in 1982 and 1983. Length-weight relationships were examined for 1982 and 1983 separately. The analysis of co- variance showed that males and females exhib- ited the same length-weight relationship, both in 1982 (P - 0.53) and 1983 (P - 0.42). Similarly, length-weight relationship did not change be- tween age groups in 1982 (P - 0.06) and 1983 (P = 0.49). The covariance for 1982 was nearly significant because the slope for 5-yr-old postsmolts, based on 15 individuals, was larger than for other age groups. However, mean condi- tion index values by age revealed no significant difference between age groups in 1982 (P = 0.11) or 1983 (P = 0.28). Since there was also no significant difference in the length- weight relationship between 1982 and 1983 (P = 0.14), the data were pooled. Thus the length-weight relationship for postsmolts col- lected in this study can be described as a single regression: log W = (2.8280 • log L) - 4.5336 P < 0.001, n = 539 where W = weight (g); L = length (mm). Maturation of Gonads Postsmolts were all immature both in 1982 and 1983, but differences were observed between males and females in 1982. Testes averaged 48 ± 6 mg (95% C.L., n = 124) for a mean male gonadosomatic index value of 0.025% ± 0.003% (95% C.L., n = 124). There was no significant dif- ference in the mean value of either parameter between age groups (P > 0.52). Testis weight in- creased in time and as body length and body weight increased, but again there was no differ- ence between age groups (Table 2). However, the gonadosomatic index did not change in time (P = 0.10) or as postsmolts' size increased (P = 0.16 for body length and P = 0.10 for body weight), suggesting that changes in size of the testes were not allometric in male postsmolts in the autumn period. Regressions were tested and compared following Snedecor and Cochran (1967) and Sokal and Rohlf (1969). There was more variability in the data for fe- 200 DUTIL AND COUTU: EARLY LIFE OF ATLANTIC SALMON 360-. 350 340- 330 320H 310 300- 290- E E I- O z UJ 280 270 260 250 240 230 220 210 200 190- 180- • *♦ • . * A * 01 06 11 16 21 26 31 05 10 15 20 25 30 05 10 15 20 25 30 AUGUST SEPTEMBER OCTOBER DATE Figure 2. — Seasonal changes in length of Atlantic salmon postsmolts collected in 1982 and 1983. Table 2. — Functional regressions of gonad weight on date, body length, and body weight of Atlantic salmon male postsmolts {n = 124). Lengths and weights are log-transformed. Date Length (mm) Weight (g) Regression Covariance (age groups) P< 0.00011 P = 0.31 n.s. P< 0.00012 P = 0.56 n.s. P< 0.00013 P = 0.47 n.s. 1/ = 0.0168 X- 5.41 66 2y = 7.8496 X - 20.3350 3V = 2.6798 X- 7.5298. 201 FISHERY BULLETIN: VOL. 86, NO. 2 O) lU 460 440 420 400 380 360 340 320 300 280 260 240 220 200 180 160 140 120 100 (473) (565) (464) (488) 01 06 11 16 21 26 31 05 10 15 20 25 30 05 10 15 20 25 30 AUGUST SEPTEMBER DATE OCTOBER Figure 3. — Seasonal changes in weight of Atlantic salmon postsmolts coUecteci in 1982 and 1983. males. Though the size of ovaries did not differ significantly between age groups (P = 0.07), go- nadosomatic index values increased in older fe- male postsmolts (P = 0.0006). Mean values ranged from 133 mg (0.06%) for 2-yr-old post- smolts to 205 mg (0.10%) for 5-yr-old postsmolts (Table 3). The regressions of ovary weight on body length, body weight, and time were all significant except for 5-yr-old females (Table 4), and differed between age groups particularly in elevation (P < 0.05, n - 213). The regressions of gonadoso- matic index values on the same variables were Table 3. — Gonad weight (mg) and gonadosomatic index (GSI) (%) of Atlantic salmon female postsmolts: mean values and confidence limits (n = 213). Gonad Mean weight 95% C.L. interval GSI Age group Mean 95% C.L. interval 2 3 4 5 133 152 175 205 124-144 142-163 161-189 194-217 0.064 0.077 0.092 0.095 0.060-0.067 0.073-0.082 0.087-0.098 0.091-0.099 202 DUTIL AND COUTU: EARLY LIFE OF ATLANTIC SALMON not significant indicating that, in females as in males, changes in size of the gonads were not allometric (body length: P = 0.06; body weight: P = 0.73; time: P - 0.68). Scale Marks Scales used for age determination were also ex- amined for any mark that could be of use in stock identification. Many individuals exhibited a tran- sition zone on their scales. Circuli in this zone were more wide-spaced than circuli laid down in earlier summers, but they were more narrow- spaced than circuli formed in summer in a marine environment. This mark was present on 73% of the scales in both males and females (72.6 and 73.87f) but tended to decrease as smolt age in- creased: 84% (age 2 years), 75% (3 years), 68% (4 years), and 62% (5 years). However, the trend was not significant (x^ = 5.74, P > 0.10). Scales also exhibited summer checks in 1982, 1983, and 1984. The number of circuli between the smolt mark and the summer check was 10.9 ± 2.6 (mean ± SD) (range 7-15) in 1982 and 9.7 ± 2.2 (range 5-15) in 1983. The ratio between the radius to summer check and the radius to smolt mark averaged 2.00 ± 0.30 (range 1.49- 2.45) in 1982, and 1.67 ± 0.19 (range 1.26-2.32) in 1983. The overall percentage of occurrence was low in 1982 (3.6%) and involved only specimens collected in the area Bale Trinite-Pointe aux Anglais where 11 out of 13 postsmolts examined showed a summer check. The incidence of sum- mer checks on scales increased markedly in 1983 (55%) and 1984 (75%), and summer checks were no longer restricted in distribution. Examination of data concerning postsmolts possessing a sum- mer check showed no relationship between the check and measured biological variables except perhaps in 1982. The gonads of males having a summer check on their scales in = 5) were heav- ier than in males having no summer check (120 mg and 48 mg respectively). Their gonadosomatic index was also higher (GSI = 0.042% and 0.025% respectively). This was not so in females. Infor- mation on gonads is not available for 1983, 1984, and 1985. Back-calculated lengths at summer check for- mation were larger in 1982: 22 ± 2.5 cm in 1982 and 20 ± 2 cm in 1983. Postsmolts measured 265 mm on 1 September. With a mean rate of increase in length of 1.65 mm/day, the check must have been formed some 26 (1982) to 40 (1983) days earlier, i.e., in late-July. This is confirmed by ex- trapolation of the length-date plot (Fig. 2). No salmon were noted possessing 2 summer checks on their scales. Food Items Drastic changes took place in prey selection be- tween postsmolts collected in summer and in au- tumn. The stomach contents of 40 salmon ranging up to 188 mm in length (70 g) collected in the second half of July 1984 in the estuary of the Grande Trinite River, near Bale Trinite, revealed a low diversity in prey items, indicative of a tran- sition period between the riverine and marine en- vironment. Small fish remains were observed in only 5 out of 39 stomachs containing food re- mains. Invertebrates were observed in 38 stom- achs, classes Insecta and Crustacea occurring in 87 and 92% of the stomachs respectively. Food items consisted mainly of 2 families: Chironomi- Table 4. - Functional regressions of gonad weight on date, body length, and body weight of Atlantic salmon female postsmolts (n =213). Lengths and weights are log-transformed. Significance Functional regression Date Length (mm) Weight (g) 2: P = 0.0001 3 : P = 0.0001 4 : P = 0.0058 5 : P = 0.26 n.s. covariance : P 2 3 4 5 2 3 4 5 P < 0.0001 P < 0.0001 P < 0.0001 P = 0.34 n.s. covariance P < 0.0001 P < 0.0001 P < 0.0001 P = 0.28 n.s. covariance Y = 0.0075 X Y = 0.0044 X Y = 0.0052 X 2.7127 1.9052 2.0229 0.04 y = 3.8015 X- 10.1089 y = 2.9105 X - 7.8659 / = 3.1893 X -8.4701 P = 0.02 Y= 1.31 10 X- 3.9428 Y = 1 .0385 X - 3.2227 Y= 1.1763 X- 3.4659 P = 0.04 203 FISHERY BULLETIN: VOL. 86, NO. 2 dae (95% of insects by number) and Gammaridae (92% of crustaceans by number). Later in summer and autumn, postsmolts con- sumed mainly small fish. Stomach contents were analyzed for 373 out of 385 postsmolts collected in 1982. There were 109 stomachs with no food re- mains (29%). They were most prevalent in the first half of August: 46%, 1-15 August; 25%, 16- 31 August; 26%, 1-15 September; 25%, 16 Sep- tember-31 October. Fishes dominated the list of prey items as they occurred in 238 out of 264 (90%) stomachs containing food remains, includ- ing 200 postsmolts (84%) that fed exclusively on small fish. Fishes could be identified in 157 stom- achs. Diversity was low, capelin, Mallotus villo- sus, occurring in 16 stomachs (10%) and sand lance, Ammodytes americanus, in 145 stomachs (92%). Ammodytes americanus dominated in terms of percentage by number (94%). Postsmolts consumed smaller A. americanus, in the 40-100 mm range (Fig. 4). Stomachs examined in 1983, 1984, and 1985 confirmed these observations. Invertebrates were found in 69 out of 264 stom- achs (26%) containing food remains. Eighteen postsmolts had only invertebrates in their stom- achs (26%). In contrast with smolts in the estua- rine environment, postsmolts did not rely on in- sects; insects occurred in only 8 stomachs (3%), whereas crustaceans occurred in 61 stomachs (23%) (respectively, 12 and 88% by number). Fur- thermore, the class Amphipoda no longer domi- nated the crustaceans (Table 5). >- O z u o ui oc u. 20 40 60 80 100 120 140 160 LENGTH - CLASS (mm) Figure 4. — Length distribution of sand lance in the stomachs of Atlantic salmon postsmolts collected in 1982. Horizontal and Vertical Position in the Nets Postsmolts were not randomly distributed in the nets in 1984. They occurred most frequently (78% of the individuals) in the top half of the nets (binomial test; z = 4.62, P < 0.001; Siegel 1956). Furthermore, 25 out of 74 salmon occurred alone in the nets, but many also occurred in clusters. Positions of postsmolts are indicated by number of nearest buoy on the head-line for those sets having taken 2 salmon and more (stations visited at 2-h intervals usually but also nets set overnight) (Table 6). Distributions are likely not random in sets 3, 6, 7, 9, 10, and 11 and most particularly in sets 13 and 15. The catch was low in midday: 3 salmon between 0900 and 1200 and 5 between 1200 and 1500. This increased to 12 between 1500 and 1800. The majority were caught later than 1800 (36) and in the morning between 0600 and 0900 (20). Finally, positions in the nets were analyzed in terms of selectivity. The gear used in 1982 could not be controlled. Fishermen reported using Table 5. — Crustacean organisms in the stomachs of Atlantic salmon postsmolts collected in the penod August-October 1982, based on 39 stomachs containing identifiable crustaceans. Crustacean order Percentage of occurrence Percent- age by number Main organisms Euphausia- cea Decapoda Amphipoda 87 28 15 68 24 8 Meganyctiphanes norvegica Thysanoessa inermis Thysanoessa raschi Chionoecetes opilio (larvae) Table 6. — Positions of Atlantic salmon postsmolts in nets by num- ber of nearest buoy on head-line for catches of 2 salmon and more. Set No. of no. smolts Number of nearest buoy to each smolt 1 2 11 32 2 2 12 23 3 2 14 15 4 2 15 23 5 2 19 27 6 2 22 24 7 2 23 23 8 2 23 39 9 2 25 27 10 2 23 27 27 11 4 16 21 21 43 12 5 16 24 33 42 44 13 5 23 23 23 23 23 14 6 12 17 18 18 22 42 15 9 11 12 12 12 13 13 13 14 14 204 DUTIL AND COUTU: EARLY LIFE OF ATLANTIC SALMON 62 mm stretched mesh nets (range 57-70 mm); mesh size is regulated. The nets used in 1983, 1984, and 1985, had 5 sections of increasing stretched mesh sizes: 50.8, 57.2, 63.5, 69.9, and 76.2 mm. Catches were recorded for individual mesh sizes in 1983 combining the 4 stations. Modal length increased only slightly as mesh size increased. Furthermore, the distribution for the 69.9 mm mesh was skewed to the right, indicat- ing no larger sized postsmolts were present. There was also no catch in the 50.8 mm mesh, but this may have been because of the small mesh section being made of a coarser material in 1983. Postsmolts were frequent in = 49) in the 57.2 mm mesh, but the distribution for this mesh does not suggest the existence of smaller postsmolts as the size range is similar to that of the 63.5 mm mesh and covers the size range for the 5 meshes com- bined. Catches declined from a maximum in the 63.5 mm mesh (56) to 40 and 9 in the 69.9 and 76.2 mm mesh, respectively. Therefore it is un- likely that there was any bias except perhaps against smaller postsmolts. Fall Movement out of Coastal Reaches Postsmolts gradually left the nearshore area in late-September. Fishermen in 1982 started col- lecting postsmolts in mid-August. Their bycatch declined from mid-September and was low in Oc- tober. The catch declined partly because most commercial fishermen were asked to return no more than 20 salmon each, and they reached this limit early (Table 7). Fishing was initiated later in 1983 and took place over a shorter period (23 September-11 October), but a 68^ decline in the catch was observed between the period 29 Sep- tember-5 October and the period 5-11 October. Finally in 1984, postsmolts came near the coast near the end of August and moved out in mid- September (Table 8) so that no salmon were caught in the period 20 September-20 October. This movement out of the nearshore zone was associated with decreasing near-surface tempera- tures in autumn. Temperatures measured near the surface closely followed the mean air temper- atures recorded in Sept-Iles (Fig. 5). Since most postsmolts travelled near the surface, mean air temperatures were assumed to reflect prevailing conditions for postsmolts. In 1982, postsmolts were abundant until mean air temperatures de- clined below 5°C, i.e., in early October. The situa- tion was similar in 3 stations out of 4 in 1983, particularly in the Bay of Matamek River near Sept-Iles and in Riviere Pentecote, 2 stations closer to our monitoring station. In 1984, this de- cline in mean air temperature occurred earlier (26 September), but surface temperatures were also lower than air temperatures, and the catch declined (Table 8) as soon as surface tempera- tures fell below 5°C (mid-September.) Hence, low temperatures were associated with a movement of postsmolts out of the nearshore zone. However, the reverse is not true: results for 1983 and 1984 suggest that postsmolts do not necessarily move towards nearshore zones when prevailing tem- perature conditions are favorable. Variations in seasonal abundance are masked by a general decline in salmon abundance near shore from 1982 to 1985. Fishing effort could not be assessed in 1982, but the incidence of postsmolts in coastal herring nets in 1982 was such that it can reasonably be concluded they were more abundant than in 1983, 1984, or 1985. Relative abundance can be assessed for 1983, 1984, and 1985. Bale Trinite and Port Cartier stations were operated in 1983 and 1984 using similar nets at the same locations each year and showed that postsmolts were more abundant in 1983 than in 1984 (Table 9). Finally, a station was monitored near Sept-Iles in 1985 in an area where the best catches were made in 1982, using Table 7. — Time distribution of the commercial bycatch of Atlantic salmon postsmolts in 1982. Period Number 10-31 August 01-15 September 16-30 September 01-15 October 16-31 October 210 118 21 24 8 Table 8. — Catch of Atlantic salmon postsmolts by period and locality in 1984. Locality Baie Port Period Trinite Cartier 21-24 August 1 4 24-27 August 27-30 August 1 30 August-02 September 14 02-05 September 4 8 05-08 September 6 9 08-11 September 2 11-14 September 3 16 14-17 September 8 17-20 September 20 September-20 October 205 20 15 10 5 -5 20 15 10 P 5 uj cc < £ 20 Q. 15 10 5 -5 20 15 10 5 -5 FISHERY BULLETIN: VOL. 86, NO. 2 1982 UJ Sept-lles air temperature ^ 05 10 15 20 25 30 04 09 14 19 24 29 04 09 14 19 24 29 1983 Port Cartier sea temperature 05 lb i5 20 25 30 04 09 14 19 24 29 04 09 14 19 24 29 1984 Bale Trinite sea temperature 05 io i5 20 25 30 04 09 14 19 24 29 04 09 14 19 24 29 1985 Sept-lles sea temperature I r I ~ 05 10 15 20 25 30 04 09 14 19 24 29 04 09 14 19 24 29 AUGUST SEPTEMBER DATE OCTOBER Figure 5. — Mean daily air temperature in Sept-lles (— -), and mean daily sea temperature at 0.5 m ( — ) in Port Cartier (1983), at 6 m (---) in Baie Trinite (1984), and at 0.5 m (---) and 6 m ( ) in Sept-lles (1985). 206 DUTIL AND COUTU: EARLY LIFE OF ATLANTIC SALMON Table 9. — Relative abundance in terms of catch per unit of eHort (CPUE) at 2 stations run in 1983 and 1984 using similar nets at tfie same sites. Baie Trinite Port Cartier Period Catch E CPUE Catch E CPUE 1983 23 Sept.- 11 Oct. 54 38.5 1.4 40 20.0 2.0 1984 21 Aug. -31 Aug. 1 20.5 0.1 5 32.5 0.2 30 Aug. -17 Sept. 15 54.5 0.3 57 73.0 0.8 17 Sept.-23 Sept. 22.0 — 17.0 — 23 Sept.- 11 Oct. 55.5 — 45.5 — 11 Oct.-18 Oct. 9.5 — 21.0 — 'CPUE: catch of 1 net in 24 hours. similar nets to 1983 and 1984. The catch was nearly nonexistent: 5 salmon for 120 unit effort ( 1 unit effort, is 1 net x 24 hours). Thus, based on limited observations, numbers of salmon near shore in summer and autumn seem to be highly variable from year to year. DISCUSSION Postsmolts of Atlantic salmon stay much longer near our coasts than is usually believed. Though early months in the marine environment have been shown to have a marked influence on salmon runs 1 and 2 years later (e.g., Christensen and Larsson 1979; Scarnecchia 1983, 1984), postsmolt biology has been a neglected area of investigation. North American smolts are as- sumed to migrate rapidly out of the estuaries of their home rivers, to feeding areas located far out in the North Atlantic east of the Grand Bank (Templeman 1968; Reddin 1985) and north to Labrador and Greenland (Saunders 1966; Tem- pleman 1967). They return 1 or 2 years later to spawn in home rivers. This study, and early records, indicate that some postsmolts remain in coastal areas as late as autumn before moving offshore. This was clearly suggested by Comeau (1909) who stated that 0.5-1.5 lb postsmolts were regularly taken in autumn along the north shore of the Gulf of St. Lawrence. Fishermen inter- viewed in 1982 on the north shore of the Gulf of St. Lawrence, from Pointe-des-Monts to Blanc Sablon, declared incidental catches of postsmolts, mainly in the months of August and September. The majority declared taking postsmolts each year. Smolts were also shown to linger in estuar- ies of the north shore of the Gulf of St. Lawrence (this study; Power and Shooner 1966; Randall and Power 1979). Huntsman (1939) mentioned their occurrence in autumn in herring nets near the mouth of Gaspe Bay. Fall catches also occurred in New England (Kendall 1935). Recently, smolts released in New England were caught as post- smolts in coastal areas of Canada (Meister 1984). There are stocks in the Bay of Fundy (as the stocks in the Baltic) that do not go on extensive migrations in the North Atlantic (Huntsman 1939; Jessop 1976): postsmolts of these stocks are regularly taken in herring nets in Passa- maquoddy Bay and off Grand Manan Island (El- son 1953, 1964; Allen et al. 1972). Hence the pres- ence of postsmolts near shore in autumn (or in summer and autumn) is a characteristic of the marine life of Atlantic salmon in North America. Timing of migration has been described for hatchery-reared smolts released in Sweden (Larsson 1974) and Finland (Jutila and Alapassi 1985). Tags were returned mostly from a distance of less than 10 km between days and 10, 20-50 km between days 10 and 20, and 50-100 km 2 months past their release, in the brackish waters of the Gulf of Bothnia (Jutila and Alapassi 1985). Behavior of postsmolts is similar to that of 1- and 2-sea-year salmon. Postsmolts in this study occurred mainly near the surface as indicated by their distribution in the nets. LaBar et al. (1978) concluded that smolts migrated near the surface in the Penobscot estuary. Templeman (1967, 1968) also found salmon to occur near the surface in the Northwest Atlantic: 62% occurred in the top 0.6 m and 90% in the top 1.5 m in July and August 1965. Similarly in 1966, most salmon were taken in the top 1.5 m, the number caught decreasing sharply below 0.6 m. Similar observa- tions were made on Baltic salmon (Carlin and Lundin 1967; Christensen 1968). There is less in- formation available on schooling. Postsmolts did not regularly have a clustered distribution in nets, but considering that a net does not retain all the salmon striking it, there were still many in- stances of salmon moving in schools. Thurow (1968) came to the same conclusion for older salmon in the Baltic. Templeman (1967) pre- sented limited evidence for salmon in the North- west Atlantic, but reached negative conclusions later (Templeman 1968). Finally there are lim- ited data in the literature concerning the rhythm of activity of salmon in the marine environment. Christensen and Lear (1980) showed that in West Greenland best catches occurred early in the morning (0600-0800), decreased sharply between 0800 and 1000, and were low between 1000 and 1400. The nets were not set between 1400 and 207 FISHERY BULLETIN: VOL. 86, NO. 2 0600. Catches in this study were also low in mid- day. Either this is a reflection of 2 peaks of activ- ity, dawn and dusk, as is common in salmonids in freshwater, or this is possibly due to salmon avoiding the nets in high light levels. Thus postsmolts and 1- and 2-sea-year salmon appear to have a similar behavior at sea. They move in small schools close to the surface and are possibly more active at dusk and dawn. Food items in the stomachs of postsmolts changed markedly in summer and indicated a low diversity of prey. This is in contrast to findings reported for salmon in the Northwest Atlantic. Grande Trinite River postsmolts had fed mostly on chironomids and gammarids in late-July. Baltic salmon postsmolts fed mainly on aerial in- sects though small fish and crustaceans also oc- curred in the stomachs of postsmolts in the south- ern Baltic (reviewed by Christensen and Larsson 1979). Jutila and Toivonen (1985) also found aerial insects to be the dominant food items in the stomachs of small postsmolts ( <20 cm) in the Gulf of Bothnia (Baltic). They observed that post- smolts were not selective and must have fed near the surface (20 cm surface layer). Postsmolts col- lected later in the present study relied mainly on small sand lance. Insects and gammarids had been replaced by vertically migrating crus- taceans such as Meganyctiphanes noruegica (Kulka et al. 1982). Thurow (1968) estimated 25 cm to be the length threshold for piscivorous feed- ing by Baltic salmon. In the present study, this size was likely reached in the first half of August 1982. This coincides with a major change in stom- ach contents and a high percentage of stomachs containing no food. The data on postsmolts in July are too limited to suggest that sand lance abundance could limit the early success of postsmolts at sea, but potential relationships in late summer should be tested as was done for capelin by Reddin and Carscadden (1981). Data on 1- and 2-sea-year salmon indicate they will readily feed on a diversity of prey items, main items including Arctic squid, Gonatus fabricii; paralepids, Paralepis coregonoides ; and lantern fishes (Lampenyctus sp., Notoscopelus sp.) (Tem- pleman 1967, 1968; Lear 1980). Sand lance and capelin are dominant items in West Greenland and on the coast of Newfoundland (Lear 1972, 1980), and on the Grand Bank (Reddin 1985). Reddin ( 1985) observed major changes in stomach contents between salmon on the Grand Bank (sand lance and capelin) and east of the Grand Bank (Bathylagidae, Paralepis sp., and crus- taceans), emphasizing that salmon are not selec- tive predators. The rate of increase in mean length averaged 1.65 mm/day in the Gulf of St. Lawrence over the summer period in 1982 and 1983. This value is based on the hypothesis that smolts migrated to estuaries in mid-June. Smolt migration took place in the first half of June in Restigouche River in the southern Gulf of St. Lawrence (Peppar 1982) and in the second half of June in Grande Trinite River in the northern Gulf of St. Lawrence (Caron 1984). Downstream migration of smolts peaked at various dates in June in West- ern Arm Brook in western Newfoundland (Chad- wick 1981). The calculated rate of increase is also based on a mean back-calculated smolt length of 135 mm. Length at smoltification averaged 125- 130 mm in Grande Trinite River (mean and SD: 127.5 ± 12.3, n = 88, in 1981; 125.8 ± 10.9, n = 92, in 1982; see also Caron 1984). Matamek River and Moisie River smolts measured 125-150 mm (Schiefer 1972). They measured 150 mm in Little Codroy River (Murray 1968) and 174 mm in Western Arm Brook (Chadwick 1981). There are no data in the literature concerning the rate of increase in size of smolts and postsmolts in the marine environment. Postsmolts in the Bay of Fundy reached a mean length of 296 mm in mid- August 1952 (Allen et al. 1972), some 3 cm more than postsmolts in this study: 265 mm and 306 mm on 1 September and 30 September. How- ever the high value of the power exponent of the length-weight relationship as compared with salmon in Newfoundland and Labrador (Lear 1973) indicates postsmolts were not in poor condi- tion. There was possibly a decline in the rate of increase in length in mid-summer as suggested by the large proportion of scales showing a sum- mer check (false-annulus) in 1983. This occurred 10 circuli from the smolt check in mid-summer in postsmolts 20-22 cm in length, i.e., prior to this study period, and may have been produced as a response to a shortage of prey or to deteriorating environmental conditions. Elson (1953) also no- ticed the frequent occurrence of a slowing of growth 6-10 circuli out from the last parr an- nulus. The percentage of occurrence of the check varied between locations (1982) and between years (1982 < 1983 < 1984). Therefore it is not likely to be a response to a change in postsmolt biology such as a scheduled shift in prey selection. However, the summer check can be thought of as a potential tool for stock discrimination. Some 26-32 circuli are formed before the first sea an- 208 DUTIL AND COUTU: EARLY LIFE OF ATLANTIC SALMON nulus is formed (Lear and Misra 1978) at a length of 46-50 cm in the first half of April (Munro 1970). Presence of postsmolts near shore in late sum- mer in the northern Gulf of St. Lawrence, as re- ported in this study, and presumably their late movement out of the Gulf of St. Lawrence indi- cate that the directional nature of the migration should be reconsidered. There are some smolts that do not head towards the high seas as soon as they reach the estuaries. They seem to roam nearby unless prevailing conditions are not favor- able. Temperature can be hypothesized as trig- gering the late movement of postsmolts out of the northern Gulf of St. Lawrence. Saunders (1986) reviewed the thermal biology of Atlantic salmon and suggested the thermal range for salmon in the sea is lower than for juvenile salmon in fresh- water. Salmon occur mainly at temperatures ranging from 4° to 8°C in the Northwest Atlantic (Templeman 1968; May 1973; Reddin 1985). Post- smolt movements out of the nearshore area took place in a short period as temperature was de- creasing, between mid-September and mid- October. Postsmolts were more abundant in 1982 and 1983 as mean air temperature ranged between 4° and 10°C in early autumn. In 1984, sea temperature decreased rapidly from more than 15°C in late-August down to 2°C in mid- September; postsmolts vanished from the near- shore area as temperature declined below 4°C. However in 1985, they did not come near the coast though sea temperature ranged between 8° and 12°C. Saunders et al. (1975) reported the lethal temperature of salmon in seawater to be -0.7°C. This precludes the possibility of salmon over- wintering in the Gulf of St. Lawrence unless they return to freshwater, as do some salmon in the Koksoak River (Cote et al. 1984; Robitaille et al. 1984a, b), or move down to midwater, a behavior described for salmon in the Baltic in response to high temperatures (>12°C) at the surface (Aim 1958). Comeau (1909) reported postsmolts found in the stomach of seals off Pointe-des-Monts in January and February. Low sea temperature has been hypothesized as limiting the passage of Kok- soak River smolts (Ungava Bay) to West Green- land in some years, thereby resulting in an estu- arine population (Power 1969, 1981). This situation might also be hypothesized to occur in the Gulf of St. Lawrence. For instance in 1983, mean air temperature, not to mention minimal temperature, decreased from 4° to 0°C and less in a short period near the end of October. Masses of seawater carried by gyres east and west of Anti- costi Island, and presenting momentarily favor- able conditions, can get surrounded by masses of seawater at lower temperature. Should salmon rely on temperature as a cue for their movement out of the Gulf of St. Lawrence, then late mi- grants could not escape as conditions deteriorate. The origin of postsmolts collected in this study is not known. They may be a particular subgroup of some north shore stocks. Postsmolts in this study smoltified earlier and at a smaller size than stocks in northern Newfoundland (Chadwick 1981). However, their origin cannot be deter- mined based on smolt length or smolt age distri- butions. For instance, there is a general tendency for increasing smolt age with latitude, but there is much variability in the data at latitudes below 52°. Data for salmon stocks in rivers near 50° latitude range from 3 to 4 years (Power 1981). Furthermore, postsmolts in this study had an age distribution similar to that of salmon in the Port- aux-Basques (Newfoundland) drift net fishery. Port-aux-Basques salmon migrated to rivers all around the Gulf of St. Lawrence (Belding and Prefontaine 1938). Postsmolts in this study may also be from a particular subgroup of individuals, such as late- migrant smolts. Power and Shooner (1966) and Randall and Power (1979) observed remnants of the smolt migration feeding in river estuaries on the north shore of the Gulf of St. Lawrence in July and August. Furthermore coho salmon released in mid- to late-summer did not leave the general area of release (Mahnken and Joyner 1973). Since grilse are known not to migrate as far as 2- and 3-sea-year salmon in the Northwest At- lantic (Ruggles and Ritter 1980), postsmolts in this study can be thought of as potential grilse. However, the only indication in that direction that we have is the observation that some males having a summer check on their scales in 1982 had a higher gonadosomatic index than males having no summer check on their scales (0.042 and 0.025% respectively). However, this is based on a small number of postsmolts as few salmon exhibited a summer check in 1982 and, unfortu- nately, no gonads were preserved in 1983-85. There are some stocks maturing mainly as grilse among the north shore stocks, but grilse are nearly exclusively males in these stocks (Schiefer 1972; Caron 1984). Postsmolts in this study were 62% females. Female grilse are common in New- foundland (Chadwick 1981; Power 1981). There are no published data on Anticosti stocks. Poten- 209 FISHERY BULLETIN: VOL. 86, NO. 2 tial grilse have been reported from the Grand Bank in offshore fisheries in the Northwest At- lantic (Reddin 1985), and it has been suggested that stocks maturing as grilse in the Bay of Fundy may not leave the general area throughout their entire marine life (Jessop 1976). Knowledge on the marine biology of Atlantic salmon postsmolts has been a neglected area of research. Their continual presence in sea surface waters could prove to be extremely valuable in forecasting salmon movements and production (Chadwick 1982; Scarnecchia 1984). Potential studies include mortality rates in the period of transition (July) and the relationship between low temperatures (3°-4°C) and postsmolt migra- tion out of the Gulf of St. Lawrence. Biological data in general are needed to be included in mod- els forecasting salmon runs in the North Atlantic. ACKNOWLEDGMENTS The authors wish to thank Fisheries and Oceans personnel, including many summer stu- dents, having assisted in field work on the North Shore, particularly S. Cloutier, M. Fortin, M. Laverdiere, Y. Lavergne, B. Legare, B. Mercille, M. Michaud, M. Poirier, and others. J. Boulva believed in the need to study postsmolt biology and actively supported this study. G. Shooner, G. Morin, and team assisted in scale reading and in examining preserved postsmolts in 1982 and 1983. J. R. Robitaille assisted in scale data analy- sis. Fishermen between Pointe-des-Monts and Blanc Sablon kindly discussed the incidence of postsmolts in their nets and participated in sam- pling in 1982 and 1983. P. Bertrand of the Minis- tere du Loisir de la chasse et de la Peche (Quebec) shared his knowledge of the bycatch in herring nets. Many thanks also to the population of Bale Trinite. E. M. P. Chadwick, G. Power, and R. L. Saunders kindly reviewed an earlier version of the manuscript. LITERATURE CITED Allan. I R H . and J A Ritter. 1977. Salmonid terminology. J. Cons. int. Explor. Mer 37:293-299. Allen. K. R.. R. L. Saunders, and P. F. Elson 1972. Marine growth of Atlantic salmon (Salmo salar) in the Northwest Atlantic. J. Fish. Res. Board Can. 29:1373-1380. Alm, G 1958. Seasonal fluctuations in the catches of salmon in the Baltic. J. Cons. int. Explor. Mer 23:399-433. Anonymous 1984. Report of the Baltic salmon and trout working group. ICES CM. 1984/ Assess:ll. Belding, D L., and G Prefontaine 1938. Studies on the Atlantic salmon. II-Report on the salmon of the 1937 Port-aux-Basques (Newfoundland) drift-net fishery. Contrib. Inst. Zool. Univ. Montreal 3, 58 p. Carlin, B 1959. Results of salmon smolt tagging in the Baltic area. Rapp. P.-v. R^un. Cons. int. Explor. Mer 147:89-96. Carlin, B., and H.-E Lundin 1967. The size and position of the salmon caught in drift nets in the Baltic. Swed. Salmon Res. Inst. Rep. 14, 6 P- Caron. F. 1983. Migration vers I'Atlantique des post-saumoneaux (Salmo salar) du Golfe du Saint-Laurent. Nat. can. (Rev. Ecol. Syst.) 110:223-227. 1984. Rapport d'op6ration de la riviere Trinite, 1984. Gk)uvemement du Qu6bec, Ministere du Loisir, de la Chasse et de la Peche, Direction gen^rale de la faune, 26 P Chadwick, E M. P 1981. Biological characteristics of Atlantic salmon smolts in Western Arm Brook, Newfoundland. Can. Tech. Rep. Fish. Aquat. Sci. 1024, iv + 45 p. 1982. Dynamics of an Atlantic salmon stock (Salmo salar) in a small Newfoundland river. Ph.D Thesis, Memorial Univ. Newfoundland, St. John's, 267 p. Christensen, O. 1968. Report on an experimental fishery with salmon drift-nets. ICES CM. 1968/B:11. Christensen, O , and P.-O Larsson 1979. Review of Baltic salmon research. ICES Coop. Res. Rep. 89, 124 p. Christensen O., and W. H. Lear. 1980. Distribution and abundance of Atlantic salmon at West Greenland. Rapp. P.-v. R6un. Cons. int. Explor. Mer 176:22-35. Comeau, N. a. 1909. Life and sport on the north shore of the lower St. Lawrence and Gulf Telegraph Printing Company, Que- bec. COTfi, Y., I. Babos, and J A. Robitaille. 1984. Caract6ristiques scalim6triques des saumons du Koksoak (Ungava, Quebec). Nat. can. (Rev. Ecol. Syst.) 111:401-409. Elson, P F 1953. Growth and migration of PoUett River salmon. St. Andrews Biol. Stn. Annu. Rep., p. 146-148. 1964. Post-smolt Atlantic salmon in the Bay of Fundy. St. Andrews Biol. Stn. Annu. Rep., E20-E23. Huntsman, A G. 1939. Conference on salmon problems. Publ. Am. Assoc. Adv. Sci. 8:86-106. Ikonen, E., and H Auvinen 1984. Migration of salmon in the Baltic Sea, based on Finnish tagging experiments. ICES CM. 1984/M:4. 1985. Migration of salmon post-smolts (Salmo salar) in the Baltic Sea. ICES CM. 1985/M:19. Jessop. B M 1976. Distribution and timing of tag recoveries from native and normative Atlantic salmon (Salmo salar) released into Big Salmon River, New Brunswick. J. Fish. Res. Board Can. 33:829-833. 210 DUTIL AND COUTU: EARLY LIFE OF ATLANTIC SALMON JUTILA. E.. AND T ALAPASSI. 1985. Recaptures of salmon post-smolts (Salmo salar) during the first summer after release in Finnish tagging experiments. ICES CM. 1985/M:28. JUTILA. E , AND J TOIVONEN 1985. Food composition of salmon post-smolts (Salmo salar) in the northern part of the Gulf of Bothnia. ICES CM. 1985/M:21. Kendall. W C 1935. The fishes of New England, the salmon family. Part 2. The salmons. Mem. Bos. Soc. Nat. Hist. 9:1- 166. KULKA. D W , S Corey, and T D Iles 1982. Community structure and biomass of euphausiids in the Bay of Fundy. Can. J. Fish. Aquat. Sci. 39:326- 334. LaBar, G W , J D McCleave. and S M Fried. 1978. Seaward migration of hatchery-reared Atlantic salmon iSalmo salar) smolts in the Penobscot River estu- ary, Maine: open-water movements. J. Cons. int. Ex- plor. Mer 38:257-269. Larsson. P -O 1974. Migration of the Swedish west coast salmon stocks. Swed. Salmon Res. Inst. Rep. 3, 11 p. 1985. Predation on migrating smolt as a regulating factor in Baltic salmon, Salmo salar, populations. J. Fish Biol. 26:391-397. Larsson, P -O., and S Ateshkar. 1979. Laxsmoltens vandring fraan Lulealven. Fiskeritidskrift Finl. 1:8-9. Lear, W H 1972. Food and feeding of Atlantic salmon in coastal areas and over oceanic depths. ICNAF Res. Bull. 9:27-39. 1973. Size and age composition of the 1971 Newfoundland-Labrador commercial salmon catch. Fish. Res. Board Can. Tech. Rep. 392, 43 p. 1980. Food of Atlantic salmon in the West Greenland- Labrador Sea area. Rapp. P. -v. Reun. Cons. int. Explor. Mer 176:55-59. Lear, W H , and R K Misra 1978. Clinal variation in scale characters of Atlantic salmon (Salmo salar) based on discriminant function analysis. J. Fish. Res. Board Can. 35:43-47. Mahnken, C, and T Joyner 1973. Salmon for New England fisheries. Ill: Developing a coastal fishery for Pacific salmon. Mar. Fish. Rev. 35(10):9-13. May, AW 1973. Distribution and migrations of salmon in the Northwest Atlantic. Int. Atl. Salmon Found. Spec. Publ. Ser. 4:373-382. Meister. a L 1984. The marine migrations of tagged Atlantic salmon (Salmo salar) of USA origin. ICES CM. 1984/M:27. Munro. W R 1970. Notes on the salmon long-lining cruise by the R.V. "Jens Chr. Svabo" off Faroe, April 1969. ICNAF Res. Doc. 70/40, 8 p. Murray, A R 1968. Smolt survival and adult utilisation of Little Cod- roy River, Newfoundland, Atlantic salmon. J. Fish. Res. Board Can. 25:2165-2218. Peppar, J L 1982. Atlantic salmon smolt investigations, Restigouche River system. New Brunswick. Can. MS Rep. Fish. Aquat. Sci. 1641, vii + 15 p. Power, G. 1969. Salmon of Ungava Bay. Arct. Inst. N. Am. Tech. Pap. 22, 72 p. 1981. Stock characteristics and catches of Atlantic salmon (Salmo salar) in Quebec, and Newfoundland, and Labrador in relation to environmental variables. Can. J. Fish. Aquat. Sci. 38:1601-1611. Power, G., and G. Shooner. 1966. Juvenile salmon in the estuary and lower Nabisipi River and some results of tagging. J. Fish. Res. Board Can. 23:947-961. Randall, R G , and G Power. 1979. Atlantic salmon (Salmo salar) of the Pigou and Bouleau Rivers, Quebec. Environ. Biol. Fish. 4:179- 184. Reddin, D. G. 1985. Atlantic salmon (Salmo salar) on and east of the Grand Bank. J. Northwest Atl. Fish. Sci. 6:157- 164. Reddin, D. G., and J. E Carscadden. 1981. Salmon-capelin interactions. Can. Atl. Fish. Sci. Adv. Comm. Res. Doc. 81/2, 38 p. RoBiTAiLLE, J A . Y. Cote, G. Hayeur, and G. Shooner. 1984a. Particularit^s de la reproduction du saumon atlan- tique (Salmo salar) dans une partie du reseau Koksoak, en Ungava. Rapp. tech. can. sci. halieut. aquat. 1313, vii + 33 p. 1984b. Croisssance estuarienne du saumon atlantique (Salmo salar) dans le fleuve Koksoak, en Ungava. Rapp. tech. can. sci. halieut. aquat. 1314, vii + 23 p. RUGGLES. C P., AND J. A RlTTER 1980. Review of North American smolt tagging to assess the Atlantic salmon fishery off West Greenland. Rapp. P.-v. Reun. Cons. int. Explor. Mer 176:82-92. Saunders, R L. 1966. Some biological aspects of the Greenland salmon fishery. Atl. Salmon J., Summer: 17-23. 1986. The thermal biology of Atlantic salmon: influence of temperature on salmon culture with particular reference to constraints imposed by low tempera- ture. Inst. Freshwater Res. Drottningholm Rep. 63, 38 p. Saunders. R L . B C Muise, and E. B. Henderson. 1975. Mortality of salmonids cultured at low temperature in seawater. Aquaculture 5:243-252. Scarnecchia. D L 1983. Age at sexual maturity in Icelandic stocks of At- lantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci. 40:1456-1468. 1984. Climatic and oceanic variations affecting yield of Icelandic stocks of Atlantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci. 41:917-935. SCHIEFER, K 1972. Ecology of Atlantic salmon, with special reference to occurrence and abundance of grilse, in north shore Gulf of St. Lawrence rivers. Ph.D Thesis, Univ. Water- loo, Ontario, 129 p. SlEGEL, S 1956. Nonparametric statistics for the behavioral sci- ences. McGraw-Hill Book Co., N.Y., 312 p. Snedecor. G W . AND W G. Cochran 1967. Statistical methods. Iowa State Univ. Press, Ames, 10, 593 p. SOIKKELI, M. 1973. Tagged salmon smolts in the diet of the Caspian tern. Laxforskningsinst. Medd. 3:1-5. 211 FISHERY BULLETIN: VOL. 86, NO. 2 SoKAL, R R , AND F J RoHLF. 5:62-85. 1969. Biometry. The principles and practice of statistics Thurow, F. in biological research. Freeman and Co., San Francisco, 1968. On food, behavior and population mechanisms of CA., 776 p. Baltic .salmon. Swed. Salmon Res. Inst. Rep. 4, 16 p. Templeman, W. Valle, E 1967. Atlantic salmon from the Labrador Sea and off West 1985. Predation of birds on salmon and sea trout smolts Greenland, taken during A. T. Cameron cruise, July- and post-smolts. ICES CM. 1985/M:22. August 1965. ICNAF Res. Bull. 4:5-40. ViBERT, R 1968. Distribution and characteristics of Atlantic salmon 1953. Voyages maritimes des saumons et retour h la riv- over oceanic depths and on the bank and shelf slope areas iere natale. Bull. Fr. Piscic. 170:5-23. off Newfoundland, March-May 1966. ICNAF Res. Bull. 212 SIZE AND DIET OF JUVENILE PACIFIC SALMON DURING SEAWARD MIGRATION THROUGH A SMALL ESTUARY IN SOUTHEASTERN ALASKA Michael L. Murphy, John F. Thedinga, and K V. Koski' ABSTRACT To assess competition and predation among juvenile Pacific salmon iOncorhynchus spp.) migrating through the estuary of Porcupine Creek, a small stream in southeastern Alaska, their size and diet were determined in 1979 and 1981. Mean fork length (FL) during May and June increased from 32 to 73 mm (1.5 mm/day) for pink salmon, O. gorbuscha; from 39 to 51 mm (0.4 mm/day) for chum salmon, O. keta; and during June and July, from 99 to 165 mm (1.6 mm/day) for coho salmon, O. kisutch. Prey, in order of importance, included larval fish (mostly Gadidae), larval molluscs (Mesogas- trofKxla), and calanoid copepods for pink salmon; larval molluscs, larvaceans, and hyperiid amphipods for chum salmon; and fish (Clupea harengus pallasi, Ammodytes hexapterus, and Gadidae), insects, and larval decapods (Brachyrhyncha) for coho salmon. No pink or chum salmon were found in the coho salmon stomachs. Prey size for pink and chum salmon was similar (median, 0.4 mm long for both species), and much smaller than that of coho salmon (median, 2.3 mm). Diet overlap was greater between pink and chum salmon than between either species and coho salmon. Pink salmon, however, ate almost exclusively (95%) pelagic prey, whereas chum salmon ate both pelagic (74%) and epiben- thic (26%) prey. Rapid ecirly growth and differences in diet probably help minimize predation and competition among salmon during seaward migration. The early marine life stage of juvenile Pacific salmon iOncorhynchus spp.), during transition from freshwater to seawater, is important in de- termining brood-year survival and subsequent adult returns (Manzer and Shepard 1962; Parker 1968); their survival rate is lowest during this time (Parker 1968; Bax 1983). Salmon often school in large concentrations in estuaries as they migrate seaward, and are more likely to deplete food supplies and compete for food than after they disperse to the sea (Bailey et al. 1975; Feller and Kaczynski 1975). Survival depends on size (Parker 1971; Healey 1982), and competition for food can depress early growth (Peterman 1984) and prolong vulnerability to predators (Taylor 1977; Walters et al. 1978). Size and diet of juve- nile salmon in an estuary, therefore, determine the potential for predation and competition and can greatly affect survival. As salmon aquaculture expands and more juve- nile salmon are released into estuaries, competi- tion and predation among salmon may increase (Johnson 1974). To optimize hatchery production and avoid adversely affecting wild stocks, an 1 Northwest and Alaska Fisheries Center Auke Bay Labora- tory, National Marine Fisheries Service, NOAA, P.O. Box 210155, Auke Bay, AK 99821. Manuscript accepted November 1987. FISHERY BULLETIN: VOL. 86, NO. 2, 1988. understanding is needed of how different stocks of salmon interact in estuaries. This paper compares size and diet of juvenile pink, O. gorbuscha; chum, O. keta; and coho, O. kisutch, salmon to assess potential predation and competition between the species during their seaward mi- gration through the estuary of a small, pristine stream. STUDY AREA This study was conducted in the estuary of Por- cupine Creek, the only salmon stream flowing into Steamer Bay in southeastern Alaska (Fig. 1). The estuary (about 5.5 km long) consists of a 1.5 km stream reach that is periodically inundated by tides, and a 4 km series of three estuarine basins. At low tide, the inner and middle basins are small (2 and 7 ha, respectively) and shallow (14 and 16 m, respectively) compared with the outer basin (120 ha and 42 m deep). The littoral zone ranges from low-gradient mudflats to steep cobble. Bottoms of the basins are level and com- posed of shell, gravel, and mud. During low tide, the inner and middle basins are partially isolated from the outer basin and the main part of Steamer Bay by tidal rapids 1-3 m deep. Salinity is lower in the inner and middle 213 FISHERY BULLETIN: VOL. 86, NO. 2 Figure 1. — Aerial photo of study site in the inner part of Steamer Bay, southeastern Alaska, showing the Porcupine Creek estuary at low tide and location of smolt traps used by Thedinga (1985). basins (24-29%c) than in the outer basin (28- 30%c), but temperature does not differ between basins in spring and summer (11°-13°C from May to September 1981). Heavy tidal flushing, partic- ularly during spring tides, results in a diverse community within the estuary; e.g., eel grass, Zostera; Dungeness crabs, Cancer magister; bull kelp, Nereocystis; and rock scallop, Hinnetes. A detailed description of the study area is in Merrell and Koski (1978) and Koski (1984). Porcupine Creek, upstream of tidal influence, is 4.5 km long and has an average discharge of 214 MURPHY ET AL.: SIZE AND DIET OF JUVENILE PACIFIC SALMON about 0.5 m"^/second. Its watershed is forested by mature western hemlock, Tsuga heterophylla , and Sitka spruce, Picea sitchensis. Annually, 5,000-75,000 adult pink salmon and 200-4,000 chum salmon spawn in the creek from late July to October, and 250-600 adult coho salmon spawn from late September to November (Koski 1984). Pink and chum salmon fry typically migrate from Porcupine Creek from late March to mid-May (Koski^). Coho salmon smolts migrate from late April to early June, but over 90% usually migrate in late May (Thedinga 1985). METHODS Six stations, one each on the east and west sides of the three basins (Fig. 1) were sampled by a beach seine 37 m long, with 1.6 cm stretch mesh on the wings, and a central bag of 6 mm stretch mesh. The seine tapered from 2 m deep at the central bag to 1 m deep at each end. In 1979, only one station in each basin was seined about every 4 days from 16 May to 12 June. In 1981, all six stations were seined biweekly from 26 May to 7 July and monthly thereafter through 11 November. Seines were set parallel to and about 40 m from shore by a skiff, and retrieved from shore. Setting and retrieval were accomplished within 10 minutes. All fish caught were identified and counted. Fork lengths (FL) were measured to the nearest millimeter from a random sample of <25 salmon per species, station, and sampling period. Stom- ach contents were collected only in 1981 from <10 salmon per species and station in May, June, and July. Contents were collected from anesthetized fish by flushing the stomach with water from a syringe (Meehan and Miller 1978; Koski and Kirchhofer 1984) and preserved in 5% formalde- hyde. Prey were later identified, counted, mea- sured, and weighed. For diet analysis, the index of relative impor- tance (IRI) was calculated, where IRI = {% number + % weight) (% frequency of occurrence) (Pinkas et al. 1971). Diet overlap between salmon species was calculated (McCabe et al. 1983): 2 2 X'Y, C = 1 = 1 2xf . J^Yf 1=1 1 = 1 2K V. Koski, Northwest and Alaska Fisheries Center Auke Bay Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 210155, Auke Bay, AK 99821, pers. commun. October 1985. where C = overlap coefficient and X, and Y, are proportions of the total diet of salmon species X and Y, respectively, contributed by prey taxon i of s prey taxa. Diet overlap was calculated sepa- rately for proportions based on prey number and weight. Prey were also classified as epibenthic or pelagic to assess overlap in foraging mode (Feller and Kaczynski 1975). Epibenthic prey were poly- chaetes, gammarid amphipods, harpacticoid cope- pods, barnacle cyprids, and cumaceans. Pelagic prey were calanoid copepods, euphausiids, barna- cle nauplii, cladocerans, larvaceans, larval deca- pods (Brachyrhyncha), hyperiid amphipods, and fish (eggs, larvae, and juveniles). RESULTS Size In May 1979, pink salmon were the size of newly emergent fry, about 32 mm FL (Fig. 2). Average length increased 1.5 mm/day, to 73 mm on 12 June 1979. In 1981, pink salmon averaged 73 mm FL in late May and early June. Changes in average FL in 1981 could not be calculated because most migration occurred before sampling began. Average FL of chum salmon increased slower than that of pink salmon. Mean FL of chum salmon increased 0.4 mm/day in both years, from 39 mm to 51 mm in 1979, and from 60 mm to 78 mm in 1981 (Fig. 2). Chum salmon averaged about 10 mm FL longer in 1981 than in the same period in 1979. Chum salmon were not found in the estuary after early July, except for two fry caught in the outer basin in November. Average FL of coho salmon was nearly con- stant, between 85 and 110 mm, throughout May and early June in both 1979 and 1981 (Fig. 2). Average FL of coho salmon in the estuary during this period was influenced by an influx of Porcu- pine Creek migrants, which averaged between 75 and 96 mm FL (Thedinga 1985). After the migra- tion from Porcupine Creek in 1981 (=9 June), average FL increased 1.6 mm/day to 165 mm by 20 July. Average FL then decreased to 85 mm in 215 FISHERY BULLETIN: VOL. 86, NO. 2 uu 80 i> 60 40 • / / / - / (a) pink 20 4 - 'l 1 1 1 1 1 E 80 E O 60 u -» 50 Z O 40 < 30 0) ni ._^k, ^^ (^ CHUM 220 180 140 100 60 (c) COHO 3-^-tH^4 >- 4—.^ / / "n 1 1 1 1 r 15 MAY 1 JUN 15 JUN 1 JUL 15 JUL 1 AUG Figure 2.— Length of salmon in 1979 and 1981. Data shown are means and ranges for pooled samples from all stations on each sampling date. Data for pink salmon in 1981 are omitted be- cause of small sample sizes. (▲ = 1979, • = 1981.) August and 106 mm in September, after most smelts had left and a few new smelts entered the estuary. Diet A wide variety ef prey was eaten by the three salmon species, but usually only one or two prey taxa dominated the diet (Table 1). Pink salmon ate mostly larval molluscs (Mesogastropoda) and larval fish (mostly Gadidae) in May, and calanoid copepods in June. Chum salmon ate mostly larval molluscs in May; larval molluscs, larvaceans, and cladocerans in June; and hyperiid amphipods and larval decapods in July. Coho salmon ate mostly fish and insects in May and June, and fish and larval decapods in July. The identifiable fish prey of coho salmon consisted of 53% Pacific herring, Clupea harengus pallasi; 45% cod (Gadidae); and 2% Pacific sand lance, Ammodytes hexapterus. No identifiable pink or chum salmon were in the coho stomachs. Catch of coho, but not that of the other salmon, was significantly correlated (r = 0.46, P < 0.001) with aggregate catch of herring, sand lance, and cod, indicating that coho salmon con- gregated near prey schools. Diet overlap was higher between pink and chum salmon than between either species and coho salmon (Table 2). Diet overlap between pink and chum salmon was especially high in May when both species ate large numbers of larval molluscs. If based on prey weight, diet overlap between pink and coho salmon was negligible. If based on prey number, however, overlap was >50% in June when both pink and coho salmon ate large numbers of calanoid copepods. Diet overlap between chum and coho salmon was con- sistently low, especially when based on prey weight. Of the 12 most important prey taxa in May and June, when all 3 salmon species were present in the estuary, only 4 differed significantly iP < 0.05) in mean number per stomach between pink and chum salmon, whereas 9-10 differed significantly between the two species and coho salmon (Table 3). Compared with pink salmon, chum salmon ate more harpacticoid copepods, cladocerans, and insects. Coho salmon ate fewer small plankton and more fish than did the other salmon species. Coho salmon averaged fewer than 20 total prey items, compared to more than 100 in pink salmon and 200 in chum salmon. Coho salmon ate larger prey than did the other salmon (Fig. 3). Median prey length for coho salmon was 2.3 mm, compared with 0.4 mm for pink and chum salmon. Coho salmon generally selected larger individuals of each prey taxon — particularly larger calanoid copepods, gammarid amphipods, euphausiids, and larval decapods — than did pink and chum salmon (Table 4). Offish prey, coho salmon ate mostly juveniles, whereas pink and chum salmon ate mostly eggs and lar- vae. As they grew larger, all three salmon species selected larger prey. Numbers of hyperiid am- phipods, euphausiids, and fish larvae — all rela- tively large prey — were positively correlated with FL of pink or chum salmon, whereas num- bers of cladocerans and larvaceans — both rela- tively small prey — were negatively correlated with chum salmon FL (Table 5). Numbers of calanoid copepods and fish were positively corre- lated with coho salmon FL, whereas the number 216 MURPHY ET AL.: SIZE AND DIET OF JUVENILE PACIFIC SALMON Table 1 . — Stomach contents of juvenile salmon in Porcupine Creek estuary, 26 May-7 July 1981 . %A/ is percent by number, %W is percent by wet weight, %F0 is percent frequency of occurrence of fish with prey item /, and %IRI is percent of total sum of IRI for all prey taxa. IRI = (%/V + %W)%fO. Taxa are omitted if %IRI is s3 for all salmon species. Pink salmon Chum salmon Coho salmon Prey taxon %N %W %F0 %IRI %/V %W %F0 %IRI %/V %W %F0 %IRI 26-29 May 1981 Mollusc larvae 47 2 71 28 59 36 100 60 2 4 Calanoida 12 2 57 7 5 4 71 5 4 Harpacticoida 1 29 11 9 90 12 17 1 18 3 Cladocera 1 43 1 14 5 77 10 4 Cumacea 2 9 32 3 16 1 41 6 Euphausiacea 11 2 71 8 2 42 1 Decapod larvae Shhmp 9 5 71 8 1 13 Crabs 8 2 60 4 3 3 60 2 2 12 Fish eggs and larvae 8 86 57 44 21 13 2 1 6 4 Fish 22 62 48 38 Insects! 1 29 1 5 52 2 29 23 100 47 Total 98 99 100 95 95 97 89 93 94 (Number of stomachs) (7) (31) (27) 9-10 June 1981 Mollusc larvae 9 5 100 8 24 16 73 23 7 Calanoida 47 52 100 58 5 5 65 5 22 43 9 Harpacticoida 3 5 25 1 4 4 65 4 7 Cladocera 13 7 75 9 19 8 81 17 1 23 Decapod larvae Shhmp 1 1 75 1 1 20 Crabs 4 11 100 5 7 12 100 8 19 21 4 Larvacea, Oikopleura 9 5 100 8 20 13 85 23 7 14 1 Fish eggs and larvae 6 5 100 6 11 12 35 6 Fish juveniles 10 94 71 70 Insects^ 1 50 3 10 75 8 28 2 50 14 Total 92 92 96 93 91 94 87 96 98 (Number of stomachs) (4) (26) (14) 7 July 1981 Calanoida — — — — 4 1 43 3 16 24 5 Hypehidea — — — — 12 30 88 45 4 29 2 Decapod larvae Shrimp — — — — 61 40 14 18 1 6 Crabs — — — — 3 10 86 9 55 5 53 25 Fish eggs and larvae — — — — 3 57 2 Fish juveniles — — — — 9 94 50 65 Insects^ — — — — 6 4 100 14 5 24 2 Total — — — — 86 88 91 90 99 99 (Number of stomachs) (0) (7) (17) 1 Mostly adult Diptera. of insects was negatively correlated. As a conse- quence of the selection of larger prey as the salmon grew, total prey weight increased with salmon FL, whereas total prey number did not (Table 5). Although pink and chum salmon ate prey of similar size, they foraged differently (Fig. 4). Pink salmon consumed about 95% pelagic prey; chum salmon, only 74%. Individual taxa changed, but the importance of pelagic prey did not change significantly between sampling peri- ods, estuary basins, or salmon FL classes. Diet of coho salmon, on the other hand, varied widely depending on salmon FL, date, and loca- tion (Fig. 4). Pelagic prey increased from 1% of total prey for coho salmon <80 mm FL to 80% for those >100 mm FL. Coho salmon ate fewer pelagic prey in May, when most coho were in the inner basin and feeding mainly on insects, than in July when most were in the outer basin and feeding mainly on fish (Table 1). Analysis of vari- ance, however, showed that differences between basins and sampling periods were not significant 217 FISHERY BULLETIN: VOL. 86, NO. 2 Table 2. — Diet overlap (McCabe et al. 1983) based on number (n) and weight (W) of prey by sampling period for juvenile salmon in Porcupine Creek estuary, 26 May-7 July 1981. The number of stomach samples is in Table 1 . Sampling period Pink vs n chum W Pink vs n coho W Chum vs. coho n W 26-29 (vlay 9-10 June 7 July 0.87 0.47 0.40 0.38 0.07 0.54 0.09 0.00 0.13 0.08 0.29 0.00 0.10 0.01 Table 3. — Comparison of mean number of the 12 most important prey and total of all prey per salmon stomach from Porcupine Creek estuary, 26 N^ay- 1 June 1 98 1 . Means followed by the same letter are not significantly different (Kruskal-Wallis analysis of variance, P>0.05) compared within a row and across columns. The number of stomach samples is in Table 1 . Prey item Pink salmon Chum salmon Coho salmon Mollusc larvae 28 a Barnacle larvae 4a Calanoida 40 a Harpacticoida 2a Cladocera 10a Cumacea Oa Euphausiacea 5a Decapod larvae 11 a Larvacea 6a Fish eggs and larvae 3a Fish a Insects Oa 86 a 5a 12 b 19 b 37 b 5a 1 a 12a 20 a 11 a Oa 4b Total 115a 216 a b 4 b 6 b 19 b Table 4. — Mean length (mm) of prey items in salmon stom- achs from Porcupine Creek estuary, 26 May-7 July 1981. Mean prey length within prey taxa was significantly greater for coho salmon than for pink or chum salmon (sign test, n = 12 means, P=0.03 and P =0.003, for coho salmon vs. pink and chum salmon, respectively). The number of stom- ach samples is in Table 1 . Pink Chum Coho Prey item salmon salmon salmon Mollusc larvae 0.5 0.5 0.4 Barnacle larvae 0.4 0.7 0.4 Calanoida 1.2 1.2 4.8 Harpacticoida 1.3 1.1 1.3 Cladocera 0.6 0.6 0.7 Cumacea 2.4 2.5 2.8 Hyperiidea 3.3 2.4 3.0 Gammaridea 1.5 1.8 5.2 Euphausiacea 3.5 3.0 18.0 Decapod larvae 2.8 2.0 3.5 Larvacea 0.8 0.9 1 Fish, all life stages 2.0 1.2 23.9 Insects 2.8 2.2 3.5 T — I — I — I — I I I "1 — I — *T — I — T — I — I — r 0.5 2.5 4.5 6.5 8.5 10.5 12.5 >1 4 Prey Length (Interval Midpoint, mm) Figure 3. — Relative frequencies of length of prey eaten by pink, chum, and coho salmon in the Porcupine Creek estuary in 1981. Total prey measured were 687 in 11 pink, 5,634 in 63 chum, and 1,179 in 53 coho salmon. Table 5. — Spearman rank correlations between number of prey items and fork length of juvenile salmon. Because of the large number of correlations tested, significance levels were adjusted by multiplying the probability P by the number of tests for each salmon species. The number of stomach samples is in Table 1 . Prey item Pink salmon Chum salmon Coho salmon Mollusc larvae -0.32 Barnacle larvae -0.36 Calanoida -0.51 Harpacticoida -0.30 Cladocera -0.06 Cumacea -0.20 Hyperiidea -0.35 Euphausiacea 0.84 Decapod larvae 0.43 Larvacea -0.36 Fish eggs -0.33 Fish larvae and juveniles 0.86 Insects 0.61 Total prey number -0.40 Total prey weight 0.85 0.05 -0.02 0.10 -0.03 -0.34* 0.04 -0.37- 0.36* -0.19 -0.50* -0.18 -0.04 -0.11 -0.04 0.35* 0.18 0.22 0.33* 0.19 0.04 0.28 0.19 0.15 0.25 1 1 0.36* 0.71" 0.09 0.66** 'None present. 'None present in any stomachs. *Adjusted probability P < 0.05. **Adjusted probability P < 0.01. after adjusting for differences in coho salmon FL (Table 6). Thus, changes in diet were mainly re- lated to increasing size of coho salmon as they migrated through the estuary. DISCUSSION Both size and diet can affect predation and com- petition among juvenile salmon in an estuary. A 218 MURPHY ET AL.: SIZE AND DIET OF JUVENILE PACIFIC SALMON 100 80 60- 40- 20 -A) . Pink ' Chum Coho ^100^ Q. O ■5. 80 « 100 Figure 4. — Number of pelagic prey as percent of total prey eaten by individual salmon compared between sampling periods (A), estuary basins IB), and salmon fork length classes (C) in the Porcupine Creek estuary in 1981. Symbols are means; bars are ±2 SE of the means. Symbols in B and C are the same as in A . Pelagic prey are defined in the text. salmon's size mainly influences its vulnerability to predators, whereas its diet determines poten- tial competition for food. Size and diet, however, are not independent. Salmon change their diet as they grow, which helps relieve competition be- tween salmon of different size, and a poor diet slows their growth, which prolongs vulnerability to predation. Table 6. — Analysis of variance of percentage pelagic prey of coho salmon, with sampling period and estuary basin as factors and salmon fork length as covariate. Factors, covariate, and interactions were adjusted simultaneously before assessing significant (Kim and Kohout 1975). Source of variation df Mean square F P Length Sampling period Basin Residual Total 1 2 2 47 52 8,458 1,580 654 900 1,519 9.4 1.8 0.7 0.004 0.184 0.489 Because of similar diets, pink and chum salmon are potential competitors. Although diets of pink and chum salmon in the Porcupine Creek estuary were similar in prey size and some prey taxa, however, pink salmon fed almost solely on pelagic prey , whereas chum salmon foraged both pelagi- cally and epibenthically; such differences may help reduce competition. Competition probably also was reduced because, as the salmon grew larger, they switched to larger prey. Coho salmon probably did not compete for food with the other two species because the coho fed on larger, differ- ent prey. Rapid early growth of salmon is important in reducing vulnerability to predators (Parker 1971; Taylor 1977). For example, hatchery pink salmon fry raised for 60 days (to 40 mm FL) before release into an estuary in southeastern Alaska survived 68% better at sea than did fry released immedi- ately after emergence (Martin et al. 1981). Marine survival also is higher for year classes of larger (9-11 cm FL) than for smaller (6-8 cm FL) sockeye salmon, O. nerka , smolts (Foerster 1954). Coho salmon smolts from Porcupine Creek in 1978 averaged 99 mm FL and their survival was twice that of the 1979 smolts, which averaged only 91 mm FL (Thedinga 1985). Smolt size and migration timing, however, interact complexly to influence marine survival of coho salmon (Bilton 1978). Growth of juvenile salmon in estuaries usually inferred from changes in mean size IS (LeBrasseur and Parker 1964; Healey 1978), but these estimates are subject to bias. In this study, changes in mean size of fish in the catches on successive dates could underestimate real growth for two reasons: 1) small individuals may have migrated continuously into the estuary from 219 FISHERY BULLETIN: VOL. 86, NO. 2 freshwater and 2) larger individuals may have migrated continuously from the estuary to the sea. Conversely, growth could be overestimated if predators of salmon selected small individuals (Parker 1971). In addition, although the inner and middle basins are semiisolated from adjacent marine waters during low tide, juvenile salmon from adjacent waters could easily enter the estu- ary, especially the outer basin, during flood tide and mix with salmon from Porcupine Creek. Estimates of salmon growth in estuaries and nearshore marine waters are variable, but gener- ally range between 1 and 2 mm/day. LeBrasseur and Parker (1964) estimated pink salmon growth to be 0.9 mm/day during the first 30 days at sea, and Healey (1978) estimated pink salmon growth during summer to be 1.0 mm/day; our estimate was 1.5 mm/day. Our estimate for chum salmon at 0.4 mm/day was considerably less than that of Healey (1978) at 1.5 mm/day, also based on change in mean length; however, our estimate for coho salmon of 1.6 mm/day was similar to that of Healey (1978) at 1.2 mm/day. Summer growth back-calculated from scales of salmon from the Sea of Okhotsk was about 1.6 mm/day for pink and chum salmon (Birman 1969). Because of their initial small size, pink and chum salmon particularly are vulnerable to predators including juvenile coho salmon (Parker 1971). Several authors have suggested that a major share of pink salmon mortality in the first weeks at sea results from juvenile coho salmon predation (Parker 1971; Kaczynski et al. 1973; Hargreaves and LeBrasseur 1985), but such pre- dation has not been found in field collections. Parker (1971) demonstrated predation by juve- nile coho salmon on pink salmon fry in the labora- tory, and juvenile coho salmon are known preda- tors of salmon fi:y in fi*eshwater (Hunter 1959; Koski and Kirchhofer 1984). However, we have not found any published data that show predation by juvenile coho salmon on other salmon in estu- aries or marine waters. Predation by juvenile coho salmon on pink salmon fiy migrating ft-om freshwater does occur in the tidal-influenced reach of Porcupine Creek (Koski and Kirchhofer 1984), but such predation apparently does not ex- tend into the estuarine basins. Many fishes have been identified as predators of pink and chum salmon in estuaries, including Pacific herring (Thorsteinson 1960), sea-run cutthroat trout, Salmo clarki; cod, Gadus macrocephalus ; and sculpin, Leptocottus armatus, (Bax et al. 1977). We speculate that predation by coho salmon on salmon fiy may occur only under circumstances in which the coho salmon are combined with small fry as they migrate from freshwater. The period of vulnerability of pink and chum salmon fry to predation by juvenile coho salmon is probably relatively short. Within the first 3 weeks after entering the estuary, pink salmon fiy can grow larger than the prey fish of juvenile coho salmon. In the laboratory, juvenile coho salmon ate the smallest pink salmon available and did not eat any larger than about 50 mm FL (Parker 1971), which coincides with the largest fish eaten by coho salmon in our study. At a growth of 1 mm/day, pink salmon entering the estuary at 32 mm FL will outgrow predation by coho salmon smolts in 18 days. In Porcupine Creek, most pink and chum salmon migrated fi-om the stream sev- eral weeks before coho salmon, which enables them to grow large enough to avoid predation by coho salmon in the estuary. Thus, early migration and rapid growth of pink and chum salmon fiy probably are important in reducing predation by coho salmon. In the Porcupine Creek estuary, competition and predation probably were slight. Competition for food was minimal, as evidenced by the rapid salmon growth, because of differences in prey and foraging mode and because regular tidal flushing probably replenished food supplies, as in Traitors Cove, AK (Bailey et al. 1975). Natural stocking levels in the estuary also probably were below thresholds where competition for food would de- press survival. Predation by coho salmon on pink and chum salmon was avoided because the pink and chum salmon migrated earlier than coho salmon and rapidly grew too large for the coho to handle. Thus, in this natural system, com- petition and predation probably were unimpor- tant because of moderate stocking levels, rapid growth, and differences in diet and timing of mi- grations. In systems with hatchery inputs, how- ever, stocking levels would probably be higher and salmon size and timing of migrations differ- ent than in natural systems, which could increase competition and predation. Stocking levels and timing of hatchery releases of juvenile salmon in estuaries are important in minimizing competition and predation (Myers 1980). Hatchery releases should avoid combining large concentrations of pink and chum salmon fi-y so as not to deplete food supplies. Conversely, re- leases during low predator abundance and good growing conditions — high food availability and warm temperature — could increase grov^rth and 220 MURPHY ET AL.: SIZE AND DIET OF JUVENILE PACIFIC SALMON survival. Early releases of coho salmon could in- crease predation on pink and chum fry (Johnson 1974), especially fry <50 mm FL, if the releases coincide with fry migrations through the estuary. ACKNOWLEDGMENTS Our thanks to R. Brodeur, G. Grette, C. Hawkes, D. Kirchhofer, and R. Walter for help in study design and field sampling. Thanks also to J. Hard, J. Landingham, T. Merrell, Jr., D. Mortensen, and J. Pella for reviewing the manuscript. The Fisheries Research Institute of the University of Washington identified and measured prey from stomach samples. LITERATURE CITED Baily. J. E., B. L. Wing, and C R. Mattson 1975. Zooplankton abundance and feeding habits of fry of pink salmon, Oncorhynchus gorbuscha , and chum salmon, Oncorhynchus keta, in Traitors Cove, Alaska, with speculations on the carrying capacity of the area. Fish. Bull., U.S. 73:846-861. Bax, N J. 1983. Early marine mortality of marked juvenile chum salmon (Oncorhynchus keta) released into Hood Canal, Puget Sound, Washington, in 1980. Can. J. Fish. Aquat. Sci. 40:426-435. Bax, N J , E O. Salo, B. P Snyder, C. A Simenstad, and W. J. Kinney 1977. Salmon outmigration studies in Hood Canal: a sum- mary— 1977. In W. J. McNeil and D. C. Himsworth (ed- itors), Salmonid ecosystems of the North Pacific, p. 171- 201. Oregon State Univ. Press, Oregon State Univ. Sea Grant Program, Corvallis. BiLTON, H. T 1978. Returns of adult coho salmon in relation to mean size and time at release of juveniles. Fish. Mar. Serv. Tech. Rep. 832, 73 p. Dep. Fish. Environ., Pac. Biol. Stn., Nanaimo, B.C. BiRMAN, I B 1969. Distribution and growth of young Pacific salmon of the genus Oncoryhynchus in the sea. Probl. Ichthyol. 9:651-666. Feller, R. J., and V. W. Kaczynski. 1975. Size selective predation by juvenile chum salmon (Oncorhynchus keta) on epibenthic prey in Puget Sound. J. Fish. Res. Board Can. 32:1419-1429. FOERSTER, R. E. 1954. On the relation of adult sockeye salmon {Oncorhynchus nerka) returns to known smolt seawEird migrations. J. Fish. Res. Board Can. 11:339-350. Hargreaves, N B., and R J LeBasseur 1985. Species selective predation on juvenile pink (Oncorhynchus gorbuscha ) and chum salmon (O. keta ) by coho salmon (O. kisutch). Can. J. Fish. Aquat. Sci. 42:659-668. Healey, M. C 1978. The distribution, abundance, and feeding habits of juvenile Pacific salmon in Georgia Strait, British Colum- bia. Fish. Mar. Serv. Tech. Rep. 788, 49 p. Dep. Fish. Environ., Pac. Biol. Stn., Nanaimo, B.C. 1982. Timing and relative intensity of size-selective mortality of juvenile chum salmon (Oncorhynchus keta) during early sea life. Can. J. Fish. Aquat. Sci. 39:952- 957. Hunter, J G. 1959. Survival and production of pink and chum salmon in a coastal stream. J. Fish. Res. Board Can. 16:835- 886. Johnson, R. C. 1974. Effects of hatchery coho on native Puget Sound stocks of chum salmon fry. In D. R. Harding (editor). Proceedings of the 1974 Northeast Pacific Pink and Chum Salmon Workshop, p. 102-109. Dep. Environ., Fish. Vancouver, B.C. PCaczynski, V W , R. J Feller, J. Clayton, and R. J. Gerke. 1973. Trophic analysis of juvenile pink and chum salmon (Oncorhynchus gorbuscha and O. keta) in Puget Sound. J. Fish. Res. Board Can. 30:1003-1008. Kim, J O , AND F J. Kohout. 1975. Analysis of variance and covariance: subprogram ANOVA. 2d ed. In N. H. Nie, C. H. Hull, J. G. Jenkins, K. Steinbrenner, and D. H. Brent (editors). Statistical package for the social sciences, p. 398- 433. McGraw-Hill, N.Y. KOSKI, K V 1984. A stream ecosystem in an old-growth forest in southeast Alaska: Part I. Description and characteristics of Porcupine Creek, Etolin Island. In W. R. Meehan, T. R. Merrell, Jr., and T. A. Hanley (editors), Fish and wildlife relationships in old-growth forests. Proceedings, p. 47-55. Am. Inst. Fish. Res. Biol., Juneau, AK. Available from J. W. Reintjes, Rt. 4, Box 85, Morehead City, NC 28557. KosKi, K V , AND D. A. Kirchhofer. 1984. A stream ecosystem in an old-growth forest in southeast Alaska. Part FV: Food of juvenile coho salmon, Oncorhynchus kisutch in relation to abundance of drift and benthos. In W. R. Meehan, T. R. Merrell, Jr., and T. A. Hanley (editors). Fish and wildlife relationships in old-growth forests, Proceedings, p. 81-87. Am. Inst. Fish. Res. Biol., Juneau, AK. Available from J. W. Reintjes, Rt. 4, Box 85, Morehead City, NC 28557. LeBrasseur, R J , AND R R Parker 1964. Growth rate of central British Columbia pink salmon (Oncorhynchus gorbuscha ). J. Fish. Res. Board Can. 21:1101-1128. Manzer, J I., AND M. P. Shepard 1962. Marine survival, distribution and migration of pink salmon (Oncorhynchus gorbuscha) off the British Colum- bia coast. H. R. MacMillan Lectures in Fisheries, p. 113-122. Symposium on Pink Salmon 1960, Univ. British Columbia, Vancouver. Martin, R. M , W. R. Heard, and A. C. Wertheimer. 1981. Short-term rearing of pink salmon (Oncorhynchus gorbuscha ) fry: effect on survival and biomass of return- ing adults. Can. J. Fish. Aquat. Sci. 38:554-558. McCabe, G. T , JR , W D MUIR, R L. Emmett, and J. T. DURKIN. 1983. Interrelationships between juvenile salmonids and nonsalmonid fish in the Columbia River estuary. Fish. Bull., U.S. 81:815-826. Meehan, W R., and R A Miller. 1978. Stomach flushing: effectiveness and influence on survival £ind condition of juvenile salmonids. J. Fish. Res. Board Can. 35:1359-1363. 221 FISHERY BULLETIN: VOL. 86, NO. 2 Merrell, T. R.. Jr., and K V. Koski. 1978. Habitat values of coastal wetlands for Pacific coast salmonids. In P. E. Greeson, J. R. Clark, and J. E. Clark (editors). Wetland functions and values: the state of our understanding, p. 256-266. Proceedings of the Na- tional Symposium on Wetlands, American Water Re- source Association, Minneapolis, MN. Myers, K W. W. 1980. An investigation of the utilization of four study areas in Yaquina Bay, Oregon, by hatchery and wild juvenile salmonids. M.S. Thesis, Oregon State Univer- sity, Corvallis, 234 p. Parker, R R 1968. Marine mortality schedules of pink salmon of the Bella Coola River, central British Columbia. J. Fish. Res. Board Can. 25:757-794. 1971. Size selective predation among juvenile salmonid fishes in a British Columbia inlet. J. Fish. Res. Board Can. 28:1503-1510. Peterman. R M. 1984. Density-dependent growth in early ocean life of sockeye salmon iOncorhynchus nerka). Can. J. Fish. Aquat. Sci. 41:1825-1829. PiNKAS, L., M S OLIPHANT, and I. L. K. IVERSON. 1971. Food habits of albacore, bluefin tuna, and bonito in California waters. Calif Dep. Fish Game, Bull. 12, 105 p. Taylor, S. G. 1977. The effect of timing of downstream migration on meirine survival of pink salmon iOncorhynchus gor- buscha). M.S. Thesis, Univ. Alaska, Southeastern Senior College, Juneau, 40 p. Thedinga, J F. 1985. Smolt scale characteristics and yield of coho salmon, Oncorhynchus kisutch, smolts and adults from Porcupine Creek, southeastern Alaska. M.S. Thesis, Univ. Alaska, Juneau, 93 p. Thorsteinson, F. V 1960. Herring predation on pink salmon fry in a south- eastern Alaska estuary. Trans. Am. Fish. Sec. 91:321- 323. Walters, C. J., R. Hilborn, R. M. Peterman, and M J Staley. 1978. Model for examining early ocean limitation of Pacific salmon production. J. Fish. Res. Board Can. 35:1303-1315. 222 GROWTH THROUGH THE FIRST SEX MONTHS OF ATLANTIC COD, GADUS MORHUA , AND HADDOCK, MELANOGRAMMUS AEGLEFINUS, BASED ON DAILY OTOLITH INCREMENTS^ George R. Bolz and R. Gregory Lough^ ABSTRACT Daily growth increments of otoliths from larval and juvenile Atlantic cod and haddock were enumer- ated, and growth curves were derived describing the first six months of life. Growth for both species was best described by Gompertz-type curves. Inverse regressive methods were employed to construct general models with confidence limits for predicting age (days) for given standard lengths (mm) from hatching through the juvenile period. Microstructural analysis of the otoliths did not discern a settling check at the time when the fish would be expected to leave the pelagic lifestyle for the demersal one, which indicates that the transition is neither physiologically stressful nor abrupt. Fluctuations in the year-class strength of fish stocks are thought to be determined by the rate of mortality during the first year of life (Moser 1981; Lough et al. 1985; Neilson and Geen 1986; and others). Calculation of reliable mortality rates, assessment of the influences of size-selectivity, and establishment of precise hatching dates and times during a given year when loss to recruit- ment is greatest are dependent upon accurate age and abundance estimates. Recently, investigators have suggested that mortality during the postlar- val and juvenile periods may be as critical as that occurring in the egg and larval life stages (Cohen and Grosslein 1982; Sissenwine 1984). Investiga- tion of this hypothesis by the Northeast Fisheries Center (NEFC) has been ongoing since 1984. Enumeration of daily growth increments de- posited on fish otoliths provides the best method for the age determination of larvae and juveniles needed for generating growth curves and estimat- ing mortality (Essig and Cole 1986). An excellent review of past and current methodologies em- ployed in the study and application of otolith mi- crostructure may be found in Campana and Neil- son (1985). Atlantic cod and haddock are both spring spawners on Georges Bank (Sherman et al. 1984) and have pelagic eggs and larvae that undergo similar development. Transformation to the juve- nile life stage occurs around 20-30 mm SL, or 2-3 months from hatching (Fahay 1983). The transi- tion from the pelagic to demersal habitat of the adults takes place sometime after transforma- tion, usually by 6-8 cm in midsummer, and re- cent field observations by the NEFC indicates the transition is a gradual process with considerable variability. In an earlier study by Bolz and Lough (1983), growth curves were developed for larval Atlantic cod and haddock based on otolith analysis that defined growth from hatching (4-5 mm SL) through the first two months of life (ca. 20 mm SL). Juvenile Georges Bank Atlantic cod and had- dock are not fully vulnerable to bottom-trawl gear (Clark et al. 1982), and growth curves based on groundfish surveys conducted by the NEFC in the autumn and spring are inaccurate for fish younger than about six months of age. The pri- mary goal of the work reported here was to derive age-at-length curves for field-caught Atlantic cod and haddock describing their growth from hatch until they are fully available to capture by bottom-trawl survey gear. A secondary objective was to determine if a check ring, a wide incremen- tal band indicative of physiological or environ- mental changes, was deposited during the juve- nile's transition from the pelagic to the demersal mode of life. IMARMAP Contribution FED/NEFC 87-15, Northeast Fish- eries Center Woods Hole Laboratory, National Marine Fish- eries Service, NOAA, Woods Hole, MA 02543. ^Northeast Fisheries Center Woods Hole Laboratory, Na- tional Marine Fisheries Service, NOAA, Woods Hole, MA 02543. Manuscript accepted December 1987. FISHERY BULLETIN: VOL. 86, NO. 2, 1988. METHODS Atlantic cod and haddock larvae and juveniles were collected on six cruises conducted by the NEFC's RV Albatross IV and RV Delaware II on 223 FISHERY BULLETIN: VOL. 86, NO. 2 Georges Bank during the springs of 1981, 1983, and 1984 and the summers of 1984 and 1985. Sample dates and station locations where larvae and juveniles were collected for otolith analysis are given in Table 1. The samples were collected with either 1) a continuous double-oblique haul using a 61 cm bongo net sampler (0.505 and 0.333 mm mesh) deployed to a maximum depth of 100 m (Posgay and Marak 1980), 2) a 1 m MOCNESS^ fitted with nine 0.333 mm mesh nets which sam- pled discrete vertical strata from the bottom of the water column to the surface, 3) a 10 m MOC- NESS (3 mm mesh) with five nets fished in the same manner as the 1 m MOCNESS (Wiebe et al. 1976, 1985), or 4) a Yankee 36 otter trawl towed for 30 minutes (Grosslein 1974). Stations with high densities of Atlantic cod and haddock larvae and juveniles in good condition were selected dur- ing the cruises for otolith analysis. The fish were removed immediately following the haul and pre- served in 95% ethanol. In the laboratory, larvae and juveniles repre- sentative of the entire size-range collected were selected for analysis. The standard length, as well as several other morphometric measurements of each larva or juvenile, was measured to the nearest 0.1 mm prior to removal of their otoliths. The 2 sagittae, 2 lapilli, and, when possible, 2 asterisci were dissected from the fish and, except- ing juvenile sagittae, mounted whole on micro- scope slides with Permounf*. The growth incre- ments (Fig. lA) on most of these otoliths were ^Multiple Opening/Closing Net and Environmental Sensing System. ''References to trade names do not imply endorsement by the National Marine Fisheries Service, NOAA. Table 1 . — Station information for Atlantic cod and haddock specimens collected for otolith analysis by 61 cm bongo net (0.505 mm mesh) oblique hauls (6B5), 1 m MOCNESS (0.333 mm mesh) discrete vertical hauls (1M3), 10 m MOCNESS (3.0 mm mesh) discrete vertical hauls (10M), and Yankee 36 otter trawl (Y36) duhng the 1981, 1983, and 1984 survey seasons. Time GMT Bottom Number of Lat. Long. W (Night or depth M&ll Station N Date day) Gear (m) Cod Haddock 1981 Albatross IV 81-03 54 4ri0' 67=06' 24 April 1235(D) 6B5 62 19 — 55 4ri3' 67=02' 24 April 1330(D) 6B5 62 10 — 56 4ri8' 66=58' 24 April 1450(D) 685 66 16 — 57 41°22' 66=55' 24 April 1630(D) 6B5 66 13 — 58 41=26' 66=51 ' 24 April 1840(D) 6B5 71 12 — 160 41°22' 67=00' 26 April 0645(N) 1M3 63 32 — 1981 Albatross IV 81-05 190 40°57' 67=19' 22 May 0300(N) 1M3 76 — 8 197 40°55' 67=13' 25 May 1200(D) 1M3 80 — 16 205 40°55' 67=09' 26 May 1130(D) 1M3 80 — 6 211 41-11' 67=35' 27 May 1200(D) 1M3 49 — 27 215 4ri2' 67=36' 27 May 2330(D) 1M3 40 — 19 1983 Albatross IV 83-03 415 40°54' 67=32' 13 May 1816(D) 1M3 74 7 2 418 40°56' 67=35' 14 May 0456(D) 1M3 71 11 16 421 40°5r 67=34' 14 May 1026(D) 1M3 68 2 13 432 40°47' 67=26' 15 May 1636(D) 1M3 89 1 — 434 40°46' 67=24' 15 May 2229(D) 1M3 93 2 — 438 41°05' 67=47' 16 May 1147(N) 1M3 54 — 15 440 4r09' 67=54' 16 May 1646(N) 1M3 52 3 3 442 4r08' 67=48' 16 May 2222(N) 1M3 40 10 — 444 4r09' 67=55' 17 May 0504(N) 1M3 52 13 — 1984 Albatross IV 84-05 519 4ri9' 67=19' 18 June 0319(N) 10M 47 — 30 1984 Delaware II 84-07 76 40°53' 66=22' 15 Aug. 1744(D) Y36 66 1 1 85 4r50 66=26' 16 Aug. 1045(D) Y36 78 — 1 88 41=49' 66=23' 16 Aug. 1430(D) Y36 62 — 1 89 41°47° 66=18' 16 Aug. 1604(D) Y36 68 — 4 90 4r47' 66=24' 16 Aug. 1715(D) Y36 78 — 3 91 41=47' 66=30' 16 Aug. 1821(D) Y36 72 — 1 93 41°45' 66=30' 16 Aug. 2024(D) Y36 75 — 1 94 41=42' 66=25' 16 Aug. 2124(D) Y36 75 — 1 98 41=47' 66=1 1 ' 17 Aug. 0249(N) Y36 70 4 — 1984 Albatross IV 84-09 18 41=49' 66=16' 12 Sept. 0845(N) Y36 70 — 9 19 41=52' 66=21' 12 Sept. 0951 (N) Y36 90 1 12 224 BOLZ AND LOUGH: GROWTH OF ATLANTIC COD AND HADDOCK Figure 1. — A. Scanning electron micrograph for a portion of the sagitta from a 5-yr-old Atlantic cod, 79 cm SL, showing daily growth increments. Bar of photograph represents 100 (xm. B. Sagitta from 47-d-old Atlantic cod larva, 13.3 mm SL (630x). Bar of photograph represents 20 ji.m. nc = nuclear check, yc = yolk-sac check. 225 FISHERY BULLETIN: VOL. 86, NO. 2 discernible without any further preparation. Sagittae from fish >25 mm SL were mounted in epoxy resin and were ground, above and below, with carborundum paper (600 grit). The resulting thin section was secured to a microscope slide with epoxy resin and etched with 6% EDTA (pH 7.0). Both the grinding and etching procedures were monitored periodically by viewing the sagitta under a dissecting microscope. The sagittae were then viewed under a Zeiss compound microscope with transmitted light. The number of growth increments were counted from the image projected by a drawing tube onto a Zeiss MOP Digital Image Analyzer System. Under transmitted light each growth increment was composed of a light and dark ring (Fig. IB), which corresponded to the heavily calcified incre- mental zone and the organic-rich discontinuous zone of Watabe et al. (1982). Depending on the size of the otolith, magnifications used ranged from 400 X to 1,000 x . Three counts were made on one of the 2 sagittae from each larva or juvenile, and those otoliths with a repeatable increment count of >90% were used in the growi:h analysis. The other sagitta was counted once for compari- son, as were the 2 lapilli. The number of incre- ments on the 2 asterisci also were enumerated. It was found in the previous study (Bolz and Lough 1983) that the asterisci were not detectable at hatching, in contrast to the sagittae and lapilli, but appeared later in the larval period. This was reflected in the asterisci having on average 27 fewer growth increments than the sagittae. In those instances where the sagittae and lapilli were particularly difficult to read, the number of asteriscal increments plus 27 was consulted as an additional check. Maximum and minimum di- ameters and planar surface area of the entire otolith were measured routinely on all sagittae, lapilli, and asterisci. The differential shrinkage of Atlantic cod and haddock larvae and juveniles with respect to standard length was corrected using Theilacker's algorithm ( 1980), which is specified and discussed in Bolz and Lough (1983). All lengths referred to in the results and discussion portions of this paper are reported as corrected lengths. RESULTS Haddock Larval and Juvenile Growth From analysis of the 189 larval and juvenile haddock, ranging from 3.5 to 123.4 mm SL, we found that growth was best described by a Gompertz-type curve. Previous uses of the Gom- pertz growth curve and methodology for fitting the curve are described in Pennington (1979), Lough et al. (1982), and Messieh et al. (1987). The variance was stabilized by using the natural log form of the growth equation, and parameters were derived by nonlinear estimation techniques resulting in the relationship: ln(L) = 1.1987 + 4.8438(1 -e 0.0088R\ (1) where L = standard length in mm, and R = number of days (increments) from hatch. A plot of the Gompertz curve fitted to the natural log of standard length vs. age in days is shown in Figure 2. The predicted hatch-length fi-om the curve of 3.32 mm falls within the range of previous studies (Colton and Marak 1969; Fahay 1983). An aver- age growth rate of 0.24 mm/day (Table 2) for the first 30 days is also reasonable (Laurence 1978; Laurence et al. 1981) and agrees with the earlier study of Bolz and Lough (1983). As a generalized model the Gk)mpertz equation described haddock growth through the first six months (175 days), at which point it intersected (Fig. 3) the von Berta- lanffy growth curve generated fi'om an analysis of adult haddock by Clark et al. (1982): Table 2. — Mean standard length at age, 95% confidence limits, and growth rate (mm/day and %/day) of larval and juvenile had- dock from hatch through 200 days estimated from the Gompertz growth model fit. Mean 95% confidence limits Growth Growth Age (d) length (mm) rate rate Lower Upper (mm/day) (%/day) 3.32 3.22 3.41 0.14 4.22 10 4.99 4.88 5.11 0.20 4.01 20 7.27 7.13 7.41 0.26 3.58 30 10.25 10.07 10.42 0.34 3.32 40 14.03 13.80 14.26 0.42 2.99 50 18.70 18.39 19.02 0.51 2.78 60 24.34 23.89 24.80 0.61 2.55 70 30.98 30.34 31.63 0.71 2.32 80 38.63 37.74 39.53 0.82 2.12 90 47.27 46.08 48.51 0.91 1.95 100 56.88 55.30 58.50 1.01 1.77 110 67.37 65.35 69.45 1.09 1.63 120 78.67 76.14 81.28 1.17 1.49 130 90.66 87.56 93.86 1.23 1.35 140 103.23 99.52 107.08 1.28 1.23 150 116.26 111.88 120.81 1.32 1.14 160 129.63 124.54 134.92 1.35 1.04 170 143.21 137.38 149.28 1.36 0.95 180 156.88 150.29 163.76 1.37 0.87 190 170.54 163.17 178.25 1.36 0.80 200 184.09 175.92 192.63 1.35 0.73 226 BOLZ AND LOUGH: GROWTH OF ATLANTIC COD AND HADDOCK JZ -P cn c X) L. O "D C o -p cn 200.0 100.0 -■ 50.0 -■ 25. -- 12.0 -Q 0088R ln(L) = 1. 1987 + 4.8438(1 - g " ) -I — I — ^- -I — \ — U 4 — I — U J I \ L J I L 25 50 75 100 125 150 175 AgG in Days Figure 2. — Gompertz growth curve and equation fitted to plot of In standard length and number of otolith increments (estimated age in days) for 189 larval and juvenile haddock collected on Georges Bank. L = 738.0(1 - e-0.3763[(fl+D)/365-0.1649])^ (2) where D = Julian date of hatch. Based on the 1981 season, an average hatch-date of 15 April {D = 105) was employed in the present model. An average length of 19.9 cm would have been attained on 1 January, by fisheries science convention the date at which an individual is con- sidered to be 1-year-old. The predicted hatch-length of 4.02 mm was within known limits (Colton and Marak 1969). The average growth rate of 0.21 mm/day (Table 3) through the first month was slightly lower than that of haddock, which is consistent with previous findings (Bolz and Lough 1983). At approxi- mately 192 days the larval and juvenile growth curve intersected the von Bertalanffy curve calcu- lated for adult Atlantic cod by Penttila and Gif- ford (1976): Atlantic Cod Larval and Juvenile Growth Although there were few larger individuals amongst the 157 larval and juvenile Atlantic cod examined, the apparent pattern was similar to that seen in haddock. A Gompertz growth curve also provided a good fit when the natural log of standard length (range: 4.6-104 mm) was plotted (Fig. 4) against age in days (range: 7-151): ln(L) = 1.3915 + 6.2707(1 - e-o.oo53/?)_ (3) L = 1481 0(1 - e-01200((i?+D)/365-0.6160)^ (4) For purposes of the model a mean hatch-date of 15 March (D = 74) was assumed. An average At- lantic cod would be expected to have achieved a length of 26.1 cm by 1 January (Fig. 5). Predictability Since it is desirable, especially during field sur- veys when direct analysis of otoliths is impossi- ble, to be able to predict age from a given length, 227 FISHERY BULLETIN: VOL. 86, NO. 2 SL -t-> cn c Qi L. O X) c D -P tn 200.0 - * 100.0 - • 50.0 - ^^ -. 3783 <(R- 1055/365- ln? i^ -0. 0088R ln(L) = 1. 1987 * 4.8438C1 - e ) 3.0 n / • 1 Larval and Juvenile Gompertz Growth Curve 1 1 1 I 1 Apr May Jun Jul Aug Sep Oct Nov Dqc Jan Figure 3. — Haddock growth model with an assumed mean hatch-date of 15 April projected through mid-January. An average length of 19.9 cm would have been attained on 01 January, by fisheries science convention the date at which an individual is considered to be 1-year-old. inverse regression (Draper and Smith 1966) was performed on the Atlantic cod and haddock grovii;h curves to estabhsh confidence intervals for predicting age from a given standard length. In its reduced form the equation obtained for had- dock was X^ _ ln(l - (Xq ± Q.023^(((Xo - 0.2990)2/7.7959) + (1 + l/n))^^^)) Xi -0.0088 (5) where X^^ and Xi = upper and lower confidence limits, Xo=l-e-ooo88« and n = sample size. Figure 6 shows the fitted growth curve bracketed by 95% confidence intervals. Performing the same calculations on the At- lantic cod growth curve yielded the relation- ship: X„, _ ln(l - (Xq ± 0.022t{{iXo - 0.1918)^/0.9294) + (1 + l/n))^'^)) X, 0.0053 (6) where Zo = l-e-o 0053ft 228 BOLZ AND LOUGH: GROWTH OF ATLANTIC COD AND HADDOCK E JZ 4-> m c Q) XI O X) c D C/) 200.0 - 100.0 - 50. - 25.0 •- 12.0 6.0 - 3.0 - ln(L) = 1.3915 * 6.2707(1 - g-°-°°53Rj J I \ L J I L J 1 L J I L 25 50 75 100 125 150 175 AgG in Days Figure 4. — Gompertz growth ciirve and equation fitted to plot of In standard length and number of otolith increments (estimated age in days) for 157 larval and juvenile Atlantic cod collected on Georges Bank. Table 3. — Mean standard length at age, 95% confidence limits, and growth rate (mm/day and %/day) of larval and juvenile Atlantic cod from hatch through 200 days estimated from the Gompertz growth model fit. Mean 95% confidence limits Growth Growth Age (d) length (mm) rate rate Lower Upper (mm/day) (%/day) 4.02 3.79 4.27 0.13 3.37 10 5.56 5.31 5.82 0.18 3.24 20 7.55 7.30 7.82 0.23 3.05 30 10.11 9.85 10.37 0.29 2.87 40 13.32 13.03 13.62 0.36 2.70 50 17.31 16.88 17.75 0.44 2.60 60 22.19 21.48 22.92 0.54 2.43 70 28.08 26.95 29.26 0.64 2.31 80 35.11 33.40 36.91 0.76 2.19 90 43.39 40.91 46.03 0.90 2.10 100 53.05 49.58 56.77 1.04 1.98 110 64.18 59.48 69.26 1.19 1.87 120 76.90 70.69 83.65 1.35 1.77 130 91.27 83.26 100.05 1.52 1.68 140 107.38 97.24 118.57 1.70 1.58 150 125.27 112.65 139.30 1 88 1.51 160 144.99 129.52 162.30 2.06 1.43 170 166.55 147.85 187.61 2.25 1.35 180 189.95 167.61 215.26 2.43 1.28 190 215.17 188.79 245.24 2.61 1.22 200 242.18 211.34 277.52 2.79 1.16 Figure 7 shows the Atlantic cod growth curve bracketed by 95% confidence intervals. Tables 4 and 5 provide predicted ages of Atlantic cod and haddock for given standard lengths with 70% and 95% confidence limits. Otolith Growth In the earlier study of larval Atlantic cod and haddock (Bolz and Lough 1983), it was found that the sagittal rings (one incremental and one dis- continuous zone) were segregated into distinct re- gions separated by thicker, darker discontinuous zones referred to as checks or check rings. Two "heavy rings" were noted in the larvae: 1) a nu- clear check surrounding a central, amorphous core and 1 or 2 irregular rings, and 2) a yolk-sac check 2-8 increments farther outward (Fig. IB). The present study corroborated the existence of these two checks. Although each otolith was care- fully examined for the presence of a settling check, no regularly occurring heavy ring could be discerned beyond the yolk-sac check in either the haddock or the Atlantic cod juveniles. It should, 229 FISHERY BULLETIN: VOL. 86, NO. 2 E E \^ JZ c Q) L O "□ C o CO 200.0 100.0 50.0 25.0 12.0 6.0 3.0 -0.0053R In(L) = 1.3915 * 6.2707(1 - e ) Larval and Juvonile Gompertz Growth Cunva ln(L) = 7.6622 * Ind -. 1200((R*74)/3e5*.eiBro. Adult von Bortalanffy Growth Curve Panttila and Glfford (1976) Mar Apr Mav Jun Jul Aug Sep Oct Nov Dec Jan Figure 5. — Atlantic cod growth model with assumed mean hatch-date of 15 March projected through mid-January. An average length of 26.1 cm would have been attained by 01 January, by fisheries science convention the date at which an individual is considered to be 1-year-old. D Q CD < y 200 ; ^ 175 - ^ ^ ^ ^-'^^.^ 150 - ^ ^ l^ ^ - '^ 125 - ^ • lUU - /- / y ^ 75 — / / y / / y / y 50 — f 25 n 1 f 1 111) 1 1 . 1 < 1 1 1 1 1 1 1 1 I 1 1 1 1 I 1 1 1 1 1 25 50 75 100 125 150 175 230 Standard Length (mm) BOLZ AND LOUGH: GROWTH OF ATLANTIC COD AND HADDOCK 200 175 D Q W CD <; 150 - 125 _^ ^ .^--"''^ / ^^^-^""^ / ^^'''''^ ^ -^ . / '^''■^'^ -- 100 - ■^ ^...^ . ^ '" y /-^ • • -^ _ / ^X ^ '^ 75 - " / .X / 50 - i 3r> 25 n - itl / J — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — 1 — I — 1_. 1 1 1 1 1 25 50 75 100 — I — I — 1 I I I I I I I 125 150 175 Standard Length (mm) Figure 7. — Inverse regression of Atlantic cod growth curve with 95% confidence intervals for predicting age in days for a given standard length (mm). however, be noted that the size range available for study limited the search for a settling check to individuals >90 mm SL and does not preclude the possibility that thinner, less discernible checks may be found when greater numbers of juveniles 50-90 mm SL are analyzed. In both haddock and Atlantic cod diametral growth (fjim) of the sagittae, lapilli, and asterisci was linearly related to standard length (mm) throughout the larval and juvenile periods. The high correlation (r > 0.98) of this relationship and its good agreement with measurements made by Bergstad (1984) would allow the sagittal di- ameter to be used as a check on the predictability model outlined in Equations (5) and (6) for esti- mating age from standard length. Estimated ages for haddock larvae and juveniles based on maxi- mum otolith diameters may be obtained with the following equation: Figure 6. — Inverse regression of haddock growth curve with 95% confidence intervals for predicting age in days for a given standard length (mm). Y = 28.390 + 2.413Xi + 21.561X2 + 73.841X3 (7) where Y = estimated age in days, Xi = sagittal diameter in mm, X2 = lapillus diameter in mm, and X3 = astericus diameter in mm. Table 6 provides a comparison of estimated ages derived from otolith diameters with observed ages derived from the number of daily incre- ments. Although multiple regression analysis using the three otolith diameters yielded a high correlation coefficient (r = 0.9890) and nonsignif- icant ^-values, the 95% confidence limits are quite broad (±2 weeks) and should be used with caution. Use of the sagittal diameter alone (Y = 35.945 + 18.484Xi) provided a good fit (r = 0.9861) for juveniles >90 mm SL but was a poor age predictor for younger fish. If only the sagitta is available for analysis, the relationship: 231 Table 4. — Predicted age in days with 70% and 95% confidence limits of larval and juvenile haddock for a given standard length. FISHERY BULLETIN: VOL. 86, NO 2 Table 5. — Predicted age in days with 70% and 95% confidence limits of larval and juvenile Atlantic cod for a given standard length. Observed standard length (mm) Predicted age (d) 70% confidence limits 95% confidence limits Observed standard length (mm) Predicted age (d) 70% confidence limits 95% confidence limits Lower Upper Lower Upper Lower Upper Lower Upper 5 10.0 7.1 13.0 4.4 15.9 5 6.7 2.1 11.3 -2.1 15.7 10 29.3 25.8 32.8 22.7 36.2 10 29.6 24.5 34.8 19.9 39.8 15 42.3 38.4 46.3 34.9 50.1 15 44.5 39.0 50.1 34.0 55.5 20 52.5 48.2 56.9 44.4 61.1 20 55.8 49.9 61.8 44.7 67.6 25 61.1 56.5 65.8 52.4 70.4 25 65.0 58.9 71.3 53.4 77.5 30 68.6 638 73.7 59.4 78.7 30 72.9 66.5 79.6 60.8 86.0 35 75.4 70.3 80.8 65.7 86.1 35 79.9 73.2 86.8 67.3 93.5 40 81.7 76.2 87.4 71.4 93.0 40 86.1 79.2 93.3 73.1 100.3 45 87.5 81.7 93.5 76.7 99.5 45 91.8 84.7 99.2 78.3 106.5 50 92.9 86.9 99.3 81.6 105.6 50 97.0 89.7 104.7 83.2 112.2 55 98.1 91.8 104.8 86.3 111.4 55 101.9 94.3 109.8 87.6 117.6 60 103.1 96.5 110.0 90.7 117.0 60 106.4 98.7 114.6 91.8 122.6 65 107.8 101.0 115.1 95.0 122.4 65 110.7 102.7 119.1 95.7 127.3 70 112.4 105.3 120.0 99.0 127.6 70 114.7 106.6 123.3 99.4 131.8 75 116.8 109.4 124.8 103.0 132.7 75 118.6 110.3 127.4 102.9 136.1 80 121.1 113.5 129.4 106.8 137,7 80 122.3 113.8 131.3 106.3 140.2 85 125.3 117.4 133.9 110.4 142.6 85 125.8 117.1 135.0 109.5 144.1 90 129.5 121.2 138.4 114.0 147.4 90 129.2 120.3 138.6 112.5 147.9 95 133.5 124.9 142.8 117.5 152.2 95 132.4 123.4 142.0 115.5 151.6 100 137.5 128.6 147.1 121.0 156.9 100 135.6 126.4 145.4 118.3 155.1 105 141.4 132.2 151.4 124.3 161.6 105 138.6 129.2 148.6 121.0 158.6 110 145.2 135.8 155.6 127.6 166.2 110 141.5 132.0 151.7 123.7 161.9 115 149.0 139.3 159.8 130.9 170.8 115 144.4 134.7 154.8 126.3 165.1 120 152.8 142.7 163.9 134.1 175.4 120 147.2 137.3 157.7 128.7 168.3 125 156.6 146.1 168.1 137.2 180.0 125 149.9 139.9 160.6 131.1 171.3 130 160.3 149.5 172.2 140.4 184.6 130 152.5 142.3 163.4 133.5 174.3 135 164.0 152.9 176.3 143.4 189.2 135 155.1 144.8 166.1 135.8 177.3 140 167.6 156.2 180.4 146.5 193.9 140 157.6 147.1 168.8 138.0 180.2 145 171.3 159.5 184.6 149.5 198.5 145 160.0 149.4 171.4 140.2 183.0 150 175.0 162.8 188.7 152.5 203.2 150 162.4 151.6 174.0 142.3 185.7 155 178.6 166.0 192.8 155.5 207.9 155 164.8 153.8 176.5 144.4 188.5 160 182.3 169.3 197.0 158.5 212.7 160 167.1 156.0 179.0 146.4 191.1 165 185.9 172.6 201.2 161.4 217.5 165 169.3 158.1 181.4 148.4 193.7 170 189.6 175.8 205.4 164.3 222.4 170 171.5 160.2 183.8 150.4 196.3 175 193.3 179.0 209.6 167.3 227.4 175 173.7 162.2 186.2 152.3 198.8 Y = 11.875 + 112.654Xi, r = 0.9129 (8) DISCUSSION should be used for the larval and postlarval pe- riod. The equation for the estimated ages of larval and juvenile Atlantic cod (Table 7) is as follows: Y = 48.202 + 8.628Xi - 121.908^2 + 139.733Z3, r = 0.9292. (9) When using only the sagittal diameter, the fol- lowing relationship should be applied to larvae and postlarvae: Y = 19.364 + 89.560Xi, r = 0.8659. (10) Unlike the tedious laboratory methods needed for the enumeration of otolith increments, gross measurements on a limited number of juvenile otoliths could be performed at sea. Despite the tedious methodology required for enumerating daily growth increments in larval and juvenile otoliths, the present work suggests that it is feasible to construct age-length keys for Atlantic cod and haddock similar to those com- monly applied in adult population studies (Clark et al. 1982). The value of such growth data is based on several assumptions, however. Since all conclusions drawn depend upon it, reasonable as- surance of the day-increment relationship in the species being investigated is critical (Beamish and McFarlane 1987; Geffen 1987). Confidence in the growth models generated here for Atlantic cod and haddock may be found in the following inferences: 1) the predicted hatch lengths of 3.32 mm for haddock and 4.02 mm for Atlantic cod fall within known limits; 2) a high correlation for the length-at-age data with the rearing experiments of Laurence et al. (1981); and 3) the smooth- 232 BOLZ AND LOUGH: GROWTH OF ATLANTIC COD AND HADDOCK Table 6. — Estimated age in days based on otolith diame- ters with 95% confidence limits for larval and juvenile had- dock compared with observed age derived from number of daily increments. Observed age (d) Estimated age (d) 95% confidence limits Lower Upper 0.0 1.5 -12.3 15.4 10.0 11.3 -2.5 25.1 20.0 21.1 7.4 34.9 30.0 30.9 17.2 44.6 40.0 40.7 27.0 54.4 50.0 50.5 36.8 64.2 60.0 60.3 46.7 73.9 70.0 70.1 56.5 83.7 80.0 79.9 66.2 93.6 90.0 89.7 76.0 103.3 100.0 99.5 85.8 113.1 110.0 109.3 95.6 122.9 120.0 119.0 105.3 132.8 130.0 128.8 115.1 142.6 140.0 138.6 124.8 152.5 150.0 148.4 134.5 162.3 160.0 158.2 144.3 172.2 170.0 168.0 154.0 182.0 180.0 177.8 163.7 191.9 Table 7. — Estimated age in days based on otolith diame- ters with 95% confidence limits for larval and juvenile At- lantic cod compared with observed age derived from num- ber of daily increments. Observed age (d) Estimated age (d) 95% confidence limits Lower Upper 0.0 7.1 -6.9 21.1 10.0 15.8 2.1 29.5 20.0 24.5 11.0 38.0 30.0 33.2 19.9 46.5 40.0 41.9 28.7 55.1 50.0 50.6 37.5 63.8 60.0 59.3 46.2 72.5 70.0 68.0 54.8 81.2 80.0 76.7 63.4 90.1 90.0 85.4 71.9 98.9 100.0 94.1 80.4 107.9 110.0 102.8 88.8 116.9 120.0 111.6 97.2 125.9 130.0 120.3 105.5 135.0 140.0 129.0 113.8 144.1 150.0 137.7 122.1 153.3 160.0 146.4 130.3 162.5 170.0 155.1 138.5 171.7 180.0 163.8 146.6 180.9 ness with which the larval and juvenile curves flow into those independently developed for the adults (Clark et al. 1982; Penttila and Gifford 1976). The predictive models for Atlantic cod and had- dock have to be viewed as general in nature, and the widening of the confidence intervals with in- creasing length (Tables 4, 5) must be kept in mind. Natural variability of length-at-age and difficulty in the preparation and reading of otoliths increases as the fish becomes older and makes precise age determinations extremely dif- ficult. For example, the ability to predict correctly the age of an individual haddock at the 70% con- fidence level decreases from ±3 days at 5 mm SL to ±2 weeks at 175 mm. In spite of this problem, otolith aging of field-caught larvae and juveniles provides a degree of precision not possible with indirect methods based on size-frequency analy- ses (Ebert 1973). Refinement of the estimated means and the reduction and stabilization of the variance should result as a greater number of otoliths are analyzed in the future. Microstructural examination of larval Atlantic cod and haddock otoliths clearly delineated check rings related to hatching and yolk-sac absorption (Bolz and Lough 1983). Both of these transitions are abrupt, and the dark, thick discontinuous zones readily observable on the otoliths are a re- flection of metabolic disturbances undergone at these times. Although additional check rings were noted in 3 or 4 of the juvenile otoliths, there was no regularity with respect to age of their oc- currence. In these individuals the checks were probably the result of physiological trauma in- duced by disease or injury since calcium carbon- ate secretion ceases not only with the metabolic changes accompanying transitional phases but during times of stress (Morales-Nin 1987). It was suspected that a distinct check, similar to the set- tling check found by Victor (1982) in the bluehead wrasse, Thalassoma bifasciatum , would be found demarcating the transition from the pelagic to the demersal mode of life with its accompanying changes in diet and activity levels. No check rings were found in the transition period (50-100 days) on the otoliths analyzed. This suggests that an abrupt metabolic disturbance does not occur at this phase of the fish's life and that settling near the bottom takes place over an extended period of time (1-2 months) even for individual fish. This agrees with a preliminary finding for Scotian Shelf gadoids by Campana and Neilson (1985). However, in a recent study by Mahon and Neilson (1987) on the gut contents of Scotian Shelf had- dock, they concluded that the transition from pelagic to demersal life occurred relatively sud- denly, less that a month for the individual fish. Apparently, change to the demersal life stage is not stressful for Atlantic cod and haddock, at least as a metabolic manifestation recorded in their otoliths. 233 FISHERY BULLETIN: VOL. 86. NO. 2 When used in conjunction with length- frequency data collected throughout the year, the Atlantic cod and haddock growth curves pre- sented in this report should allow accurate esti- mates of the following: 1) peak hatching dates, 2) the number of cohorts produced within a given season, 3) intraseasonal changes in growth and mortality rates of cohorts, and 4) which part of the spawning curve the recruits originated from (Methot 1983). In the future year-to-year com- parison of deviations in these estimates could lead to the construction of viable recruitment models permitting the early prediction of year- class strength. ACKNOWLEDGMENTS We gratefully acknowledge the technical help and guidance provided by Susan Houghton with the scanning electron microscope portions of this paper and by Michael Pennington with the statis- tical analyses. LITERATURE CITED Beamish, R J . and G. A McFarlane. 1987. Current trends in age determination methodology. In R. C. Summerfelt and G. E. Hall (editors), Age and growth offish, p. 15-42. Iowa State Univ. Press, Ames. Bergstad, A. 1984. A relationship between the number of growth incre- ments on the otoliths and age of larval and juvenile cod, Gadus morhua L. In E. Dahl, D. S. Danielssen, E. Mok- ness, and P. Solemdal (editors). The propagation of cod Gadus morhua L, p. 251—272. Flodevigen rapportser. 1, 1984. BoLZ, G R., and R. G. Lough. 1983. Growth of larval Atlantic cod, Gadus morhua, and haddock, Melanogrammus aeglefinus , on Gteorges Bank, spring 1981. Fish. Bull., U.S. 81:827-836. Campana. S. E , and J. D. Neilson. 1985. Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Clark, S. H . W J. Overholtz, and R. C. Hennemuth. 1982. Review and assessment of the Georges Bank and Gulf of Maine haddock fishery. J. Northwest Atl. Fish. Sci. 3:1-27. Cohen. E. B., and M D Grosslein. 1982. Food consumption by silver hake (Merluccius bilin- earis) on Georges Bank with implications for recruit- ment. In G. M. Calliet and C. A. Simenstadt (editors), Gutshop '81, fish food habits studies, p. 286-294. Proc. Third Pac. Workshop, Washington Sea Grant, Univ. Wash. Press, Seattle. COLTON, J B , AND R R. MaRAK. 1969. Guide for identifying the common planktonic fish eggs and larvae of continental shelf water. Cape Sable to Block Island. U.S. Bur. Commer. Fish., Biol. Lab., Woods Hole, Mass., Lab. Ref. Doc. 69-9, 43 p. Draper, N R , and H. Smith. 1966. Applied Regression Analysis. Wiley, N.Y. Ebert, T a 1973. Estimating growth and mortality rates from size data. Oecologia 11:281-298. EssiG. R. J , and C F. Cole. 1986. Methods of estimating larval fish mortality from daily increments in otoliths. Trans. Am. Fish. Soc. 115:34-40. Fahay, M P 1983. Guide to the early stages of marine fishes occurring in the western North Atlantic Ocean, Cape Hatteras to the southern Scotian Shelf. J. Northwest Atl. Fish. Sci. 4:1-423. Geffen, a J 1987. Methods of validating daily increment deposition in otoliths of larval fish. In R. C. Summerfelt and G. E. Hall (editors). Age and growth offish, p. 223-240. Iowa State Univ. Press, Ames. Grosslein, M. D. 1974. Bottom trawl survey methods of the Northeast Fisheries Center, Woods Hole, Massachusetts, U.S. ICNAF Res. Doc. 74/86, Ser. No. 3332. Laurence, G. C. 1978. Comparative growth, respiration and delayed feed- ing abilities of larval cod (Gadus morhua) and haddock {Melanogrammus aeglefinus) as influenced by tempera- ture during laboratory studies. Mar. Biol. (Berl.) 50:1- 7. Laurence, G. C, A S. Smigielski, T. A. Halavik, and B R Burns. 1981. Implications of direct competition between larval cod (Gadus morhua) and haddock (Melanogrammus ae- glefinus ) in laboratory growth and survival studies at different food densities. In R. Lasker and K. Sherman (editors). The early life history of fish, p. 304-311. Rapp. P.-v. Reun. Cons. int. Explor. Mer 178. Lough, R. G., G R. Bolz, M. Pennington, and M D Grosslein 1985. Larval abundance and mortality of Atlantic herring (Clupea harengus L.) spawTied in the Georges Bank and Nantucket Shoals areas, 1971-78 seasons, in relation to spawning stock size. J. Northwest Atl. Fish. Sci. 6:21- 35. Lough, R. G., M. Pennington, G R. Bolz, and A A Rosenberg. 1982. Age and growth of larval Atlantic herring, Clupea harengus L., in the Gulf of Maine-Georges Bank region based on otolith growth increments. Fish. Bull., U.S. 80:187-199. Mahon, R., and J D Neilson. 1987. Diet changes in Scotian Shelf haddock during pelagic and demersal phases of the first year of life. Mar. Ecol. Prog. Ser. 37:123-130. Messieh, S. N., D S Moore, and P. Rubec. 1987. Estimation of age and growth of larval Atlantic her- ring as inferred from examination of daily growth incre- ments of otoliths. In R. C. Summerfelt and G. E. Hall (editors), Age and growth of fish, p. 433-442. Iowa State Univ. Press, Ames. Methot, R D, Jr. 1983. Seasonal variation in survival of larval northern anchovy, Engraulis mordax, estimated from the age dis- tribution of juveniles. Fish. Bull., U.S. 81:741-750. MoraleS-Nin, B. 1987. Ultrastructure of the organic and inorganic con- stituents of the otoliths of the sea bass. In R. C. Sum- merfelt and G. E. Hall (editors). Age and growth offish, p. 331-343. Iowa State Univ. Press, Ames. 234 BOLZ AND LOUGH: GROWTH OF ATLANTIC COD AND HADDOCK MOSER, H G. 1981. Morphological and functional aspects of marine fish larvae. In R. Lasker (editor), Marine fish larvae, p. 90- 131. Univ. Wash. Press, Seattle and Lond. NEILSON, J. D., AND G. H. Geen. 1986. First-year growth rate of Sixes River chinook salmon as inferred from otoliths: effects on mortality and age at maturity. Trans. Am. Fish. Soc. 115:28-33. Pennington. M R 1979. Fitting a growth curve to field data. In J. KOrd., G. P. Patil, and C. Taillie (editors). Statistical distribu- tions in ecological work, p. 419-428. Int. Coop. Publ. House, Fairland, MD. PENTTILA, J. A , AND V. M. GiFFORD. 1976. Growth and mortality rates for cod from the Georges Bank and Gulf of Maine areas. Int. Comm. Northwest Atl. Fish, Res. Bull. No. 12, p. 29-36. POSGAY. J A , AND R R MARAK 1980. The MARMAP bongo zooplankton sampler. J. Northwest Atl. Fish. Sci. 1:91-99. Sherman. K , W Smith, W Morse. M Berman, J Green, and L. Ejsymont. 1984. Spawning strategies of fishes in relation to circula- tion, phjloplankton production, and pulses in zooplank- ton off the northeastern United States. Mar. Ecol. Prog. Ser. 18:1-19. SiSSENWINE, M p. 1984. Why do fish populations vary? In R. May (editor), Exploitation of marine communities, p. 59-94. Springer-Verlag, N.Y. Theilacker, G H. 1980. Changes in body measurements of larval northern anchovy, Engraulis mordax, and other fishes due to han- dling and preservation. Fish. Bull., U.S. 78:685-692. Victor, B. C. 1982. Daily otolith increments and recruitment in two coral-reef wrasses, Thalassoma bifasciatum and Halicho- eres bivittatus. Mar. Biol. (Berl.) 71:203-208. Watabe, N . K. Tanaka. J Yamada, and J. M Dean. 1982. Scanning electron microscope observations of the organic matrix in the otolith of the teleost fish Fundulus heteroclitus (Linnaeus) and Tilapia nilotica (Linnaeus). J. Exp. Mar. Biol. Ecol. 58:127-134. Wiebe, P H , K. H Burt, S H Boyd, and A W Morton. 1976. A multiple opening/closing net and environmental sensing system for sampling zooplankton. J. Mar. Res. 34:313-326. Wiebe, P. H., A. W. Morton. A M Bradley, R H Backus. J. E. Craddock, V Barber, T. J. Cowles, and G R Flierl. 1985. New developments in the MOCNESS, an apparatus for sampling zooplankton and micronekton. Mar. Biol. (Berl.) 87:313-323. 235 THE RELATION BETWEEN SPAWNING SEASON AND THE RECRUITMENT OF YOUNG-OF-THE-YEAR BLUEFISH, POMATOMUS SALTATRIX, TO NEW YORK^ Robert M. Nyman^ and David O Conover^ ABSTRACT The association between oceanic spawning season and the recruitment of young-of-the-year (YOY) bluefish, Pomatomus saltatrix, to the inshore waters of New York was studied by estimating the spawn dates of recruited fish collected in the shore zone from the number of growth increments in their otoliths. Field collections on the south shore of Long Island showed that recruitment of 3-6 cm fork length fish occurred as a distinct pulse during the last week of May in 1985 and the second week of June in 1986. Length-frequency distributions were generally unimodal and most fish collected later could be attributed to this one recruitment episode. The frequency of otolith ring deposition in YOY bluefish was determined by marking the otoliths of field-caged fish with an injection of tetracycline, and then periodically subsampling these over the ensuing 61-day period. Regression analysis indi- cated a 1:1 relation between the number of days since marking and the number of rings beyond the mark. Back-calculation to the time of first ring deposition revealed that field-collected YOY bluefish from Long Island were spawned primarily in the March-April spawning season reported to occur south of Cape Hatteras. Relatively few fish were collected from the summer spawning season that reportedly occurs in the Middle Atlantic Bight. Almost all of these summer-spawned fish were collected from the Hudson River. The bluefish, Pomatomus saltatrix, supports a major recreational fishery along the Atlantic coast of the United States. In 1985, more bluefish by weight were caught than any other marine fish, accounting for over 24% of the total marine recreational catch (U.S. Department of Com- merce 1986). Despite the importance of bluefish to the recreational fishery, very little is known of its early life history. Bluefish are found over different portions of the continental shelf from Florida to Nova Scotia at various times of the year (Bigelow and Schroeder 1953; Wilk 1977; Gilmore 1985). Based on de- scriptions of the temporal and spatial abundance of larvae, Kendall and Walford (1979) suggested that there are primarily two distinct spawning periods and regions: a spring spawning in the South Atlantic Bight at the edge of the Florida Current (see also Collins and Stender 1987), and a summer spawning in the Middle Atlantic Bight iContribution No. 588 of the Marine Sciences Research Cen- ter, State University of New York, Stony Brook, NY. ^Marine Sciences Research Center, State University of New York, Stony Brook, NY 11794-5000; present address: Chesapeake Biological Laboratory, University of Maryland, Solomons, MD 20688. ^Marine Sciences Research Center, State University of New York, Stony Brook, NY 11794-5000. Manuscript accepted November 1987. FISHERY BULLETIN; VOL. 86, NO. 2, 1988. midway over the continental shelf (see also Morse et al. 1987). They further proposed that the spring-spawned larvae are transported north- ward in the slope waters and then move inshore, spending their first summer in the bays and estu- aries of the Middle Atlantic Bight. Summer- spawned larvae, according to Kendall and Wal- ford, spend their first summer at sea or enter the estuaries of the Middle Atlantic Bight only briefly before migrating southward with the onset of winter. A minor spawning season extend- ing from September to November off the coast of Georgia and Florida (Collins and Stender 1987) involves fish resident to the South Atlantic Bight (Kendall and Walford 1979). The purpose of this study is to evaluate Kendall and Walford's hypothesis by back-calculating the spawn dates of young-of-the-year (YOY) bluefish that have recruited to inshore waters, from the number of grov^i;h increments in their otoliths. First, we describe the timing and pattern of re- cruitment of YOY bluefish to one segment of the mid-Atlantic coastline: Long Island, NY. If spawning is episodic, and if YOY bluefish from each spavining period enter New York waters, then length-frequency distributions of field col- lections should be multimodal. Next, we verify that otolith increment deposition has a daily peri- 237 FISHERY BULLETIN: VOL. 86, NO. 2 odicity in P. saltatrix. Finally, the spawning sea- son(s) of YOY bluefish recruiting to New York is determined by ageing and back-calculating to the date of first ring deposition. METHODS Seine Collections The temporal abundance and length-frequency distribution of YOY bluefish was estimated by seining 2-4 times per month from April to Octo- ber at several sites on Long Island and in the Hudson River (Fig. 1). In 1985 and 1986, three sites in Great South Bay on the south shore of Long Island were sampled: Smith Point County Park, Fireplace Neck, and the Carmans River. Seining was conducted with a 0.6 cm mesh, 30 m net set from shore, either on foot or from a small boat. Water temperature was recorded at each site and date. In 1986, a site on the north shore of Long Island, Setauket Harbor, was also sampled. A few samples were taken in the fall by angling with rod and reel. All specimens were frozen for later measurement of fork length (FL) and weight, and extraction of otoliths. Additional specimens captured in 1986 from Jamaica Bay and the Hudson River were pro- vided by the New York Department of Environ- mental Conservation (NYDEC). Their sampling was conducted with a 60 m seine (1.2 cm mesh) set from a boat. Otolith Preparation and Analysis The sagittae were mounted concave side down on a glass microscope slide with cyanoacrylate (instant glue). Two layers of masking tape were applied on either side of the otolith. The slide was then turned upside-down and sanded on a strip of wet 1200 grit wet-dry sandpaper. The masking tape ensured that the otolith was sectioned on a consistent plane and helped prevent grinding past the nucleus. Once the nucleus was reached, the otolith was polished on wet felt, using levi- gated alumina polishing compound. Three repli- cate counts of each otolith were made under a Zeiss'* compound microscope with transmitted po- larized light at 125-312X. If the three counts dif- fered by more than 10% (which occurred in about 1 out of every 10 otoliths), an additional count was made and the outlier discarded. The three final counts were then averaged. The total length of each otolith was measured (nearest 0.1 mm) with a dissecting microscope using an ocular micrometer. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 7 4 "loo' 73° |00 i N SOUTH BAY JAMAICA BAY ATLANTIC OCEAN 73°]00' 72 °|00' Figure 1. — Map of the study area with sampling locations cis indicated. 238 NYMAN and CONOVER: YOUNG-OF-THE-YEAR BLUEFISH Frequency of Ring Deposition The frequency of growth ring deposition was determined by marking the otoliths of fish with tetracycline and then subsampling the marked fish at various periods of time thereafter (Cam- pana and Neilson 1982). Sixty YOY bluefish (7- 10 cm FL) were captured by seine in Flax Pond, Old Field, NY (Fig. 1) and were transported to the Flax Pond Laboratory of SUNY Stony Brook. The fish were anesthetized in a solution of MS-222 (30 mg/L) and given an intraperitoneal injection of tetracycline (100 mg/kg offish). After injection, all fish were placed in a 1.3 x 1.3 m cylindrical floating cage constructed out of 5 mm plastic mesh and anchored in Flax Pond. The fish were fed chopped Menidia menidia twice a day, and dead bluefish were removed daily. Samples of 5- 10 healthy fish were periodically taken from the cage using a dip net and frozen until the otoliths could be excised. The experiment was terminated 61 days after the injections. After preparation as described above, the tetracycline-treated otoliths were viewed on a Zeiss compound microscope using reflected ultra- violet (UV) light at 160-400 X. Tetracycline fluo- resces upon exposure to UV light, thus enabling the location of the marked ring to be determined. The UV light was then turned off, and the num- ber of rings from the mark to the edge of the otolith was counted under transmitted white light. Each otolith preparation was coded so that the reader did not know the true age. Three repli- cate counts were conducted on each otolith. RESULTS Temporal Abundance and Length Frequency Great South Bay The appearance of YOY bluefish in the shore zone was abrupt in both years of the study. In 1985, no YOY bluefish were caught in weekly samples until 28 May when a catch per unit effort (CPUE = no. fish per seine haul) of 14.0 was recorded (Fig. 2a). Corresponding water tempera- ture was about 20°C. CPUE declined steadily thereafter through October with two exceptions: the large collections on 10 and 28 July were each due to an unusually large number offish in single seine hauls in the Carmans River. In 1986, YOY bluefish were first caught on 10 June when the water temperature was 24°C. The maximum CPUE (45.3) was obtained on 16 June and was followed by a decrease in CPUE in subsequent collections (Fig. 2b). Length-frequency distributions in 1985 showed the progression of a single mode through mid- August (Fig. 3a). Newly recruited fish in late May were 3-6 cm FL. Subsequent samples showed an increase in the mean and range of fish lengths, probably due to somatic growi;h of the initial re- cruits. There was no evidence of new 3-6 cm re- cruits entering the shore zone later in the year (Fig. 3a). Although seining continued until November, very few YOY bluefish were caught after August. An additional sample {n = 8) taken on 16 September by angling from a pier on Great South Bay had a mean fork length of 17.8 cm and a range of 14.6-19.5 cm. Length-frequency data from 1986 (Fig. 3b) show a very similar pattern to that in 1985: a single length mode appears in June and these fish increase in size through the summer. Few YOY bluefish were caught in Au- gust, September, or October. Size at recruitment to the shore zone was simi- lar in both years of our study: mean length of the 1985 and 1986 year classes at first appearance in the shore zone was 4.6 and 4.5 cm respectively (Fig. 4). However, because the 1986 year class first appeared in the shore zone two weeks later than did the 1985 year class (Fig. 2), the mean lengths of 1986 year class were less than those of 1985 on comparable dates in June and early July. By mid-July, however, this difference in mean length of the two year classes was no longer ap- parent. Both year classes reached a size of about 13-14 cm by late August when they rarely ap- peared in our seine collections. Setauket Harbor In 1986, the YOY bluefish did not appear on the north shore of Long Island at Setauket Harbor until 3-6 weeks after they first appeared in Great South Bay. Collections at Setauket Harbor were small at first with only one individual being caught on 1 July and three on 8 July. It was not until 22 July that catches similar in number to those in Great South Bay were being obtained. These fish had similar mean lengths (10.2 cm, n = 87, on 22 July; 11.9 cm, n = 22, on 5 August; 13.9 cm, n = 17, on 20 August) to those on com- parable dates from Great South Bay (of. Fig. 4). Length-frequency distributions by date were uni- modal. 239 a. ^ 20 UJ '^ 16 X ^ 12 UJ CL a FISHERY BULLETIN: VOL. 86, NO. 2 (12) (2) -(2)(2)(4)(3) (3) (13) T (7) (9) (5) (7) 3) (7) (7) (10) (6) MAY 1 \ 1 JUNE JULY AUG SEPT OCT FlGlTRE 2.— Catch per unit effort (CPUE) of YOY bluefish from Great South Bay, NY, plotted with the mean water temperature Jamaica Bay and the Hudson River The length-frequency distributions of YOY bluefish from Jamaica Bay in June, July, and August were similar to those from Great South Bay (Fig. 5). Fish lengths in June were unimodal. Subsequent collections contained progressively larger fish that were also unimodal in length dis- tribution. Sampling in the Hudson River began on 16 July 1986 and continued through 8 October. The size ranges of YOY bluefish in the July and Au- gust samples were similar to those from Great South Bay, although the length distribution on 30 July appears bimodal (Fig. 6). The length distri- butions from the 10 and 23 September collections were especially broad. In particular, the 23 Sep- tember sample contained a group offish that were much smaller (10-14 cm FL) than the mean size at this time in Great South Bay (Fig. 4), together with a second group of larger fish that correspond more closely in size with those collected else- where (18-24 cm). Frequency of Ring Deposition In tetracycline-injected YOY bluefish, the number of rings beyond the tetracycline mark (Y) and the number of days after injection (X) had a 1:1 correspondence (Fig. 7). The relationship was described by the equation Y = 0.97 LY - 0.287 in = 27, r = 0.996). The slope did not differ signif- icantly from 1.0 (^test, P > 0.1). Growth rate of the caged fish was slightly greater than that of field fish and survival was high (80%) with mortalities occurring only in the first few days of the 61-d experiment. The in- crease in mean fork length was 1.7 mm/day among the caged fish, as compared with about 1.3 mm/day for fish from field collections during the same time period (Fig. 4). Hence, the caged fish did not appeEir to be adversely affected by confine- ment. 240 NYMAN and CONOVER: YOUNG-OF-THE-YEAR BLUEFISH b. '°'- ^ 26 o o a. 22 liJ 18 < 14 10, J- ^ 20|- UJ UJ en 16 iZ 12 3 a (3) 45.3 J (9) "(12) (8) (10(3) OOP — I r (3) (10) (10) (10) (10) (13) (7) (e)(4) MAY JUNE 1 r JULY AUG SEPT OCT Figure 2. — Continued — on each sampling date, (a) 1985; (b) 1986. Number of seine hauls is in (). Back-Calculated Date of First Ring Deposition A representative sample of 169 YOY bluefish (n - 88 from 1985, n = 81 from 1986) captured in Great South Bay were aged by counting the total number of otolith rings. The date of first ring deposition for each aged fish was then calculated based on the date of capture. In both 1985 and 1986, the dates of first ring deposition for YOY bluefish were predominantly in March and April (Fig. 8a, b). Four fish from each of the two apparent length modes in the 30 July collection from the Hudson River (Fig. 6) were aged to determine if these represented a difference in spawning season. The fish examined were 7.8—13.8 cm in size, and back- calculated dates of first ring deposition extended from 7 to 30 April. Hence, these fish could all be attributed to the same spring spawning period as those from the south shore of Long Island. However, YOY bluefish from the smaller (10- 14 cm) size class caught on 23 September in the Hudson River (Fig. 6) were also aged and their back-calculated dates of first ring deposition were found to be predominantly in June and July, and to a lesser extent in May (Fig. 8c). These dates differed greatly from those offish captured earlier in the year in the Hudson River, and along the south shore of Long Island. The relationship of ring number and fork length for each year was best described by the following equations: Y = 132.308X - 29.890 in 1985 and Y = 95.532X + 1.186 in 1986, where X is log fork length and Y is the number of rings (Fig. 9). The slopes of these regressions differed significantly (ANCOVA, P < 0.001). Total otolith length and fork length were highly correlated (r > 0.99) and increased isometrically. Total otolith length and ring number also had a high correlation (r = 0.91). 241 FISHERY BULLETIN: VOL. 86, NO. 2 a. o 'Z. UJ Z) o LU 15; 10; 28 MAY 85 5 \ II 1 32 22; 24 JUNE 1 85 12 ; 1 . 40 10 JULY 85 20 Jl. 15 10; 2 8 JULY 85 5 '■ II . •■iklMI 1, 15 1 10; 2 AUG 85 5 \ Li 3 6 9 12 15 18 21 24 FORK LENGTH (cm) >- o LU Z) o UJ (T l_L_ 5 10 10 JUNE 86 22 JULY 86 20 AUG 86 I illJUili , I X 3 6 9 12 15 18 21 24 FORK LENGTH (cm) Figure 3.— Length-frequency histograms for YOY bluefish (no.) from Great South Bay, NY. (a) 1985; (b) 1986. 242 NYMAN and CONOVER: YOUNG-OF-THE-YEAR BLUEFISH 20 p 18 - / 16 - ,/ E 1 / / o 14 — / X > -7^-— 1- 12 — <> / z V X / llJ ./ > _l 10 — 1 A ^ J o 8 - 1^ ^ 0- V i/ z 6 - /♦^ — ¥ < ^ • = 1985 UJ 4 2 1 1 • = 1986 1 1 1 JUNE JULY AUG SEPT DATE OF COLLECTION Figure 4. — Mean fork length (cm) of YOY bluefish (no.) captured with a beach seine in Great South Bay in 1985 and 1986. Vertical bars are 95% confidence intervals. The last sample in 1986 was caught by angling. 60 5 01 AUG 3 6 9 12 15 18 21 24 3 6 9 12 15 18 21 24 FORK LENGTH (cm) Figure 5. — Length-frequency distributions for YOY bluefish (no.) fi^m Jamaica Bay, NY, 1986. 243 FISHERY BULLETIN: VOL. 86, NO. 2 10 30 JULY *T*- -T 13 AUG • f " -y SEPT .uIL ii iJ I 23 SEPT 3 6 9 12 15 18 21 24 FORK LENGTH (cm) The analysis was then expanded to the remain- ing samples of YOY bluefish from Great South Bay that had not been aged. The above length-age equations were used to estimate date of first ring deposition from the dates of capture for all YOY captured in each year of sampling from Great South Bay. This exercise revealed that the vast majority of YOY bluefish in our collections from Long Island had dates of first ring deposition in late March, April, and early May (Fig. 10). The weighted mean date of first ring deposition was 8 April 1985 and 14 April 1986. The age-length equations for YOY bluefish from Great South Bay were not applied to collec- tions from Jamaica Bay or the Hudson River. Pre- liminary analyses suggested that the age-length equation for fish from the Hudson River differs substantially from those in Great South Bay, probably owing to a difference in growth rate. Geographic variation in the pattern of recruit- ment and in the age-length relationships of YOY bluefish are being further investigated. DISCUSSION Recruitment of YOY Bluefish to New York In both 1985 and 1986, the arrival of YOY blue- fish on the south shore of Long Island was abrupt. Within about a 1-wk period, CPUE went from 0.0 to 14-18 fish/seine haul. CPUE then remained high for the next two months until declining in August and September when the fish probably became too large to be efficiently sampled by our techniques. These data suggest that the YOY bluefish recruit to the shore zone as a sudden pulse. The timing of this recruitment event is ap- parently variable, differing by about two weeks among the two years of our study. The appear- ance of fish 3-6 weeks earlier on the south shore (Great South Bay) than on the north shore (Se- tauket Harbor) of Long Island suggested that these fish arrive from offshore waters to the south. Temperature probably influenced the time of arrival of YOY bluefish in the shore zone. In both years of our study, YOY bluefish appeared as temperatures reached about 20°-24°C. In Octo- ber, after temperatures dropped to the middle Figure 6. — Length-frequency distributions for YOY bluefish (no.) from the Hudson River, NY, 1986. 244 NYMAN and CONOVER: YOUNG-OF-THE-YEAR BLUEFISH Figure 7. — Relation between number of days since marking with tetracycline and the number of growth increments beyond the mark in YOY bluefish maintained in a field cage. q: < 2 O >- UJ CD (n I- z UJ 70 60 50 40 - 30 S 20 u z 10 Y= 0.97IX- 0.287 r = 0.996 1 1 1 1 1 10 20 30 40 50 60 DAYS SINCE INITIAL MARKING 70 O UJ o UJ on u. Figure 8. — Back-calculated date of first ring deposi- tion for YOY bluefish (no.) as determined by count- ing daily growth rings, (a) and (b) represent fish from Great South Bay in 1985 and 1986, respec- tively, (c) is for fish seined from the Hudson River on 23 September 1986. MAR APR MAY JUN JUL MAR APR MAY JUN JUL DATE OF FIRST RING DEPOSITION 245 FISHERY BULLETIN: VOL. 86, NO. 2 180 - 160 - 140 - Ul CD Z 1 — 1 cr Ll o o z 120 - 100 - 80 - 60 - 40 - 20 - 1986 X Xx , X %^^f0^ 1985 **^ ^ X ^x 1985 X 1986 - 1 1 1 1 1 r 1 0.4 0.6 0.8 10 12 LOG^O FORK LENGTH (cm) Figure 9. — Relation between log fork length (X) and number of otolith rings (7) for YOY bluefish from Great South Bay, NY. The regression equations are 7 = 132.308X - 29.890 (n = 88) for 1985 and y = 95.532A: + 1.186 (n = 81) in 1986. teens, we no longer captured YOY bluefish. Oben (1957) noted that in the Black Sea, YOY bluefish appeared in the shore zone at temperatures of 18°-24.5°C, and left the shore zone in October and November when temperatures dropped to 13°- 15°C. Length-frequency distributions of YOY blue- fish from the south shore samples showed only a single mode that attained progressively larger size through the summer and fall, and corre- sponded to the initial recruitment offish. If multi- ple spawning and recruitment events contributed YOY bluefish to Long Island, multimodal length- frequency distributions should have been ob- served. The unimodal distributions suggested that only one spawning period contributed the majority of YOY bluefish to Long Island. Interannual variation in the length-age rela- tionship of YOY bluefish was observed. Although recruitment occurred two weeks earlier in 1985 than in 1986, the empirical mean lengths at re- cruitment were similar (Fig. 4). Postrecruitment growth, however, was slower in 1985 than 1986 so that empirical mean lengths became similar by mid-July. Correspondingly, the slope of the length-age regressions differed significantly among years: YOY bluefish at an age of about 50-70 days had greater fork lengths in 1985 than in 1986 (Fig. 9), but the reverse was true among older, larger fish. Apparently, the growth rate of YOY bluefish prior to recruitment was higher in 1985 than 1986, but this pattern among the two years was reversed during the period of postre- cruitment growth. Validation of Daily Otolith Rings Our experimental results demonstrate that otolith ring deposition is daily in YOY bluefish. A regression slope of 0.971 indicates a one-to-one correspondence between number of days after in- jection and the number of rings beyond the tetra- cycline mark. This outcome is not particularly surprising because numerous studies have shown that increment production is daily, particularly in the early life history when somatic growth is rapid (Brothers et al. 1976; Campana and Neilson 1985; Jones 1985). Cases where ring periodicity is reportedly less-than-daily have involved subopti- mal growing conditions (Geffen 1982, 1987; Rice et al. 1985). In our study, the confinement of YOY bluefish in a field cage apparently had little effect on growth rate, or the production of daily growth increments. The field cage allowed for natural light, temperature, and salinity variations that the fish would normally have experienced in na- 246 NYMAN and CONOVER: YOUNG-OF-THE-YEAR BLUEFISH O u c: CD Z) cr QJ o u d CD Z) O" QJ 30 20 10 60 50 - 40 - 30 - 20 - 10 - 150 180 Julian Date of First Ring Deposition Figure 10. — Estimated date of first ring deposition for all YOY bluefish caught in Great South Bay in 1985 (n = 561) and in 1986 {n = 868) using the respective age-length equa- tions in Fig. 9. ture. Feeding periodicity was probably the pri- mary artifact of confinement that could have af- fected the rate of ring production in caged fish. However, Marshall and Parker (1982) showed that feeding periodicity did not significantly af- fect ring production in sockeye salmon, Oncorhynchus nerka. We were unable to determine directly the num- ber of days between spawning and first ring depo- sition because numerous attempts to capture running-ripe females for initiating experiments on eggs and larvae were unsuccessful. However, most species of fish deposit the first daily growth increment within a few days of hatching (Broth- ers et al. 1976; Radtke and Dean 1982; McGurk 1984; Radtke 1984; Davis et al. 1985). Recent ev- idence suggests this is also true in bluefish. Lar- vae captured off Long Island in 1987 had about seven otolith increments at a total body length of 4-5 mm (R. K. Cowen and D. O. Conover, unpubl. data). Based on the rate of development at 20°C in the laboratory observed by Deuel et al. (1966), 247 FISHERY BULLETIN: VOL. 86, NO. 2 larvae would reach this size in about 5-7 days posthatch. If so, first ring deposition would roughly coincide with hatching. Bluefish hatch in 48 hours at 20°C (Deuel et al. 1966), so the day of first ring deposition probably follows the date of spawning by about 2-4 days. Spawning Seasons Along the Atlantic Coast Published studies of larval bluefish distribu- tions along the Atlantic coast suggest the exis- tence of three temporally and spatially distinct spawning seasons: spring and fall spawning sea- sons in the South Atlantic Bight and a midsum- mer spawning in the Middle Atlantic Bight. In the only synoptic study covering most of the U.S. east coast, Kendall and Walford (1979) described two periods of high larval abundance: One peak occurred in March and April on the outer shelf of the South Atlantic Bight, and the other peak was in July and August midway over the continental shelf of the Middle Atlantic Bight. Subsequently, Powles (1981) and Collins and Stender (1987) also found the highest abundance of bluefish larvae in the South Atlantic Bight (Cape Canaveral to Cape Fear) to be in April and May. Collins and Stender, however, noted the existence of a lesser peak in larval abundance during September- November. This fall spawning season in the South Atlantic Bight was further confirmed by Finucane and Collins (in press) based on the gonad condition of bluefish from Georgia and the Carolinas. In the Middle Atlantic Bight off Vir- ginia, Norcross et al. (1974) found that eggs and larvae of bluefish first appeared in June, peaked in abundance in July, and persisted into August. Similar observations on the timing of the summer spawning season in the Middle Atlantic Bight were presented by Sherman et al. (1984) and Morse et al. (1987). Lassiter (1962) provided additional evidence of the existence of relatively discrete spawning seasons in bluefish. He showed that the dis- tribution of back-calculated lengths at age one has a distinctly bimodal pattern among adult fish from North Carolina. Size at age 1 tended to be either about 14 cm or 28 cm. Lassiter showed that the bimodal pattern could not be explained as a difference in growth rate, and suggested that there must be two distinct spawning seasons such that one group of fish had a first growing season about twice as long as the other. Spawn Dates of YOY Bluefish from New York Back-calculation to the day of first ring deposi- tion for YOY bluefish recruiting to Great South Bay in 1985 and 1986 demonstrated that these fish were spawned primarily in March and April (Figs. 8, 10). Fish that were spawned in July- August were rarely captured by us on Long Island in 1985 or 1986, despite continued sampling into October. Recruitment to Jamaica Bay and the Hudson River in July and August 1986 involved YOY bluefish of about the same size as those fi-om Great South Bay. Though the size range of fish fi-om the Hudson was slightly greater than those from Long Island, fish aged from each of the two modes appearing in the July Hudson River sam- ples (Fig. 6) were all spawned during April within about three weeks of each other. The apparent bimodality in July is probably a sampling arti- fact. Hence, Jamaica Bay and Hudson River fish collected in July and August can be attributed to the same spawning season as those fi^om Great South Bay. Length-frequency distributions ft"om the Hud- son River in September, however, contained a group of unusually small bluefish, and back- calculation showed that they were spawned pre- dominately in June and July (Fig. 8c). These fish probably resulted fi-om the summer spawning season in the Middle Atlantic Bight. Examina- tion of gonads from adult fish captured during 1986 suggested that the running-ripe males and mature females were most abundant during late June and July off Long Island (L. Chiarella and D. O. Conover, unpubl. data). Hence, at least, some summer-spawned YOY bluefish do recruit to the shore zone of the Middle Atlantic Bight. They were, however, much less abundant than spring-spawned YOY bluefish in our 1985 or 1986 samples. Spawning by bluefish in the spring is known to occur only in the South Atlantic Bight (Kendall and Walford 1979; Collins and Stender 1987). Water temperatures over the shelf north of Cape Hatteras are probably too low for bluefish to spawn in March and April: average shelf water temperatures in the Middle Atlantic Bight range from 5° to 14°C in March and April (Ingham 1986). Virtually no eggs and larvae (Morse et al. 1987) and comparatively few adult bluefish (Gilmore 1985) are captured in plankton or trawl surveys north of Cape Hatteras in March and 248 NYMAN and CONOVER: YOUNG-OF-THE-YEAR BLUEFISH April. Moreover, the time of arrival of YOY blue- fish on Long Island actually precedes the summer spawning season in the Middle Atlantic Bight. We therefore conclude that YOY bluefish recruit- ing to the Middle Atlantic Bight in late spring come from spawnings in the South Atlantic Bight. Larval Transport The physical mechanisms that account for the transport of bluefish larvae fi"om the South At- lantic Bight to New York are not clear. Spawning in the South Atlantic Bight occurs primarily over the outer half of the continental shelf (Powles 1981; Collins and Stender 1987), and some larvae may be entrained by the Gulf Stream and carried northward into the slope waters of the Middle Atlantic Bight (Kendall and Walford 1979). Neuston net collections in April have shown that bluefish larvae are periodically abundant on both sides of the Gulf Stream-shelf water interface off Cape Hatteras (Kendall and Walford 1979). Collins and Stender (1987) found a negative cor- relation between larval size and latitude in the South Atlantic, but their sampling may not have extended far enough north (i.e., they did not sam- ple above Cape Fear). If the Gulf Stream is responsible for the north- ward transport, a mechanism by which larvae avoid being advected too far offshore would ap- pear to be necessary. According to our results, the interval between spawning and recruitment to Long Island is about 45-60 days, whereas the surface flow of the Gulf Stream at lat. 36°N is about 104 km/day (Iselin 1936). Hence, larvae re- maining in the Gulf Stream for an extended pe- riod would be transported far off the shelf. Reten- tion near the shelf could be achieved by entering the slope waters at an appropriate time. The abrupt appearance of YOY bluefish in the shore zone suggests that the onshore migration is a temporally distinct event, perhaps triggered by vernal warming of the shelf. Because the circula- tion of the slope and shelf waters of the Middle Atlantic Bight is toward the southwest (Sherman et al. 1984), the cross-shelf migration must to some extent involve active swimming. Very few summer-spawned YOY bluefish were captured in our study. This may not be surpris- ing, however, because the prevailing currents over the midshelf off Long Island would carry lar- vae to the southwest. If so, summer-spawned fish would be found along the coast fi"om approxi- mately New Jersey to Cape Hatteras. We caution, however, against any general conclusion concern- ing the lack of summer-spawned fish in New York. There could, for example, be substantial year-to-year variation in the recruitment level of spring- and summer-spawned fish along any par- ticular segment of the U.S. coast. These issues are now being examined by extending our sampling to southern latitudes. ACKNOWLEDGMENTS We thank Robert Cerrato, Robert K. Cowen, and Peter Woodhead for reviewing the manu- script and Stephen Heins, Melanie Meade and Louis Chiarella for assistance in the field. Byron Young and Kim McKown of the NYDEC gra- ciously provided samples from the Hudson River and Jamaica Bay. An earlier version of this paper was submitted by R.M.N, to the Graduate School of the State University of New York at Stony Brook in partial fulfillment of the requirements for an M.S. degree in Marine Environmental Sci- ences. Initial funding was provided by grants from the Sport Fishery Research Foundation (D.O.C./R.M.N.) and the Montauk Marine Basin. Later funding was provided by grants to D.O.C. fi-om the NYDEC through the Dingell-Johnson Federal Aid in Sport Fish Restoration Act and by the New York Sea Grant Institute through the NOAA Office of Sea Grant, U.S. Department of Commerce, under Grant No. NA86AA-D- SG045. LITERATURE CITED BiGELOW, H. B., AhfD W. C. SCHROEDER. 1953. Fishes of the Gulf of Maine. U.S. Fish and Wildlife Service, Fish. Bull. 53:1-577. Brothers. E. B, C P. Mathews, and R. Lasker. 1976. Daily growth increments in otoliths from larval and adult fishes. Fish. Bull., U.S. 74:1-8. Campana, S. E., and J. D. Neilson. 1982. Daily growth increments in otoliths of starry floun- der (Platichthys stellatus ) and the influence of some envi- ronmental variables in their production. Can. J. Fish. Aquat. Sci. 39:937-942. 1985. Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Collins, M. R., and B. W. Stender. 1987. Larval king mackerel Scomberomorus cavalla, Spanish mackerel (S. maculatus), and bluefish (Po- matomus saltatrix ) off the southeast coast of the United States, 1973-1980. Bull. Mar. Sci. 41:822-834. Davis, R D., T. W. Storck. and S. J Miller. 1985. Daily growth increments in the otoliths of young-of- the-year gizzard shad. Trans. Am. Fish. Soc. 114:304- 306. 249 FISHERY BULLETIN: VOL. 86, NO. 2 Deuel. D G , J R. Clark, and A J Mansueti 1966. Description of embryonic and early larval stages of bluefish, Pomatomus saltatrix. Trans. Am. Fish. Soc. 95:264-271. FiNUCANE, J. H., AND L A. COLLINS. In press. Reproductive biology of bluefish, Pomatomus saltatrix, from the southeastern United States. North- east Gulf Sci. Geffen, A J 1982. Otolith ring deposition in relation to growth rate in herring (Clupea harengus) and turbot (Scophthalmus maximus) larvae. Mar. Biol. (Berl.) 71:317-326. 1987. Methods of validating daily increment deposition in otoliths of larval fish. In R. C. Summerfelt and G. E. Hall (editors). Age and growth offish, p. 223-240. Iowa State Univ. Press, Ames. Gilmore, J. 1985. Oceanic distribution, abundance, and migration of bluefish along the east coast of the United States. M.S. Thesis, State University of New York, Stony Brook, 60 p. Ingham, M C 1986. Sea surface temi>eratures in the northwestern At- lantic in 1985. NAFO SCR Doc. 86/75. ISELIN, C. O'D. 1936. A study of the circulation of the western North At- lantic. Pap. Phys. Oceanogr. Meteorol. 4:1-101. Jones, C. J. 1985. Determining age of larval fish with the otolith in- crement technique. Fish. Bull., U.S. 84:91-103. KENDALL, A. W , Jr. and L. a. Walford 1979. Sources and distribution of bluefish, Pomatomus saltatrix, larvae and juveniles off the east coast of the United States. Fish. Bull., U.S. 77:213-227. Lassiter, R. R. 1962. Life history aspects of the bluefish, Pomatomus saltatrix (Linnaeus), fi-om the coast of North Carolina. M.S. Thesis, North Carolina State College, Raleigh, NC, 91 p. Marshall. S L , and S. S. Parker. 1982. Pattern identification in the microstructure of sock- eye salmon (Oncorhynchus nerka) otoliths. Can. J. Fish. Aquat. Sci. 39:542-547. McGURK, M. D. 1984. Ring deposition in the otoliths of larval Pacific her- ring, Clupea harengus pallasi . Fish. Bull., U.S. 82:113- 120. Morse, W W , M P Fahay, and W. G. Smith 1987. MARMAP surveys of the continental shelf from Cape Hatteras, North Carolina, to Cape Sable, Nova Sco- tia (1977-1984). Atlas No. 2. Annual distribution pat- terns of fish larvae. U.S. Dep. Commer., NOAA Tech. Memo. NMFS-F/NEC-47, 215 p. NoRCROss, J J , S L Richardson, W H Massmann, and E. B. Joseph 1974. Development of young bluefish (.Pomatomus salta- trix ) and distribution of eggs and young in Virginian coastal waters. Trans. Am. Fish. Soc. 103:477-497. Oben, L C 1957. About the drifting approach of fingerling bluefish (Pomatomus saltatrix) (Linnaeus) to the shores of the Black Sea in the Karadaga region (1947-1954). Kara- dagskaia Biol. Stn. Tr. (4):155-157. POWLES, H. 1981. Distribution emd movements of neustonic young of estuarine dependent iMugil spp., Pomatomus saltatrix) and estuarine independent (Coryphaena spp.) fishes off the southeastern United States. Rapp. P. -v. R^un. Cons. int. Explor. Mer 178:207-209. Radtke, R L 1984. Formation and structural composition of larval striped mullet otoliths. Trans. Am. Fish. Soc. 113:186- 191. Radtke, R L., and J M Dean. 1982. Increment formation in the otoliths of embryos, lar- vae and juveniles of the mummichog, Fundulus heterocli- tus. Fish. Bull., U.S. 80:201-215. Rice, J A , L B Crowder. and F P Binkowski. 1985. Evaluating otolith analysis for bloater Coregonus hoyi: do otoliths ring true? Trans. Am. Fish. Soc. 114:532-539. Sherman, K , W. Smith, W Morse, M Berman, J Green, and L EllSYMONT. 1984. Spawning strategies of fishes in relation to circula- tion, phytoplankton production, and pulses in zooplank- ton off the northeastern United States. Mar. Ecol. Prog. Ser. 18:1-19. US. Department of Commerce. 1986. Recreational fishery statistics for 1985. U.S. Dep. Commer., NOAA, NMFS, 130 p. WILK, S J 1977. Biological £md fisheries data on bluefish, Po- matomus saltatrix (Linnaeus). NMFS, NEFC, Sandy Hook Lab. Tech. Ser. Rep. No. 11, 56 p. 250 ON THE ROLE OF FOOD-SEEKING IN THE SUPRABENTHIC HABIT OF LARVAL WHITE CROAKER, GENYONEMUS LINEATUS (PISCES: SCIAENIDAE) A. E. Jahn, D. M Gadomski, and M. L Sowby' ABSTRACT Fish larvae and their prey were sampled from discrete depths within the bottom meter and at middepth near the 15 m depth contour off southern California. The smallest white croaker larvae (<2.7 mm NL) occurred mostly at middepth. Mid-sized larvae (2.7 mm to the beginning of flexion) were almost all collected at the two depths nearest the bottom. All preflexion-stage larvae ate small (50-300 (im in length) prey, chiefly rotifers, copepod nauplii, tintinnids, and invertebrate eggs. Although small and mid-size larvae ate these items in different proportions, this difference could not be ascribed to vertical distribution. Diet of the largest larvae, flexion and postflexion (roughly 5-15 mm), consisted mainly of copepods and differed by >90% from diets of smaller larvae. Though largest larvae were only captured 50 cm above the bottom, their prey, with one exception (amphipods), were more abundant at or above 1 m. It was concluded that the observed suprabenthic concentration of older white croaker larvae was probably not motivated by food-seeking. Disparity between concentrations of food re- quired for survival and growth of laboratory- reared fish larvae and observations of average concentrations of food organisms in the ocean has led to the widely accepted idea that aggregations offish larvae and their food must frequently over- lap in nature (see reviews by Theilacker and Dorsey [1980] and Hunter [1981]). Direct and in- direct evidence for the importance of overlapping concentrations of larvae and their prey (Lasker 1975, 1978; Govoni et al. 1985; Buckley and Lough 1987) comes from sampling at fronts and discontinuities in the pelagic environment. One interface that attracts many zooplankters is the seabed itself (Hamner and Carleton 1979; Wish- ner 1980; Sainte-Marie and Brunei 1985). On the southern California continental shelf, the seabed serves as a surface of aggregation for larvae of numerous fish species (Brewer et al. 1981; Schlot- terbeck and Connally 1982; Barnett et al. 1984; Jahn and Lavenberg 1986) and other zooplankton (Clutter 1969; Barnett and Jahn 1987) and of large-zooplankton biomass (Jahn and Lavenberg 1986). While it is tempting to suggest a trophic advantage to the suprabenthic habit of the fish larvae, near-bottom concentrations of organisms actually eaten by larval fishes have yet to be demonstrated along the open coast. iNatural History Museum of Los Angeles County, 900 Expo- sition Boulevard, Los Angeles, CA 90007. In all cases reported, concentration in the near- bottom zone was greater in older larvae and, when observations permitted, greater during the day than at night (Brewer and Kleppel 1986; Jahn and Lavenberg 1986). The phenomenon is therefore thought to be behavioral. Possible ad- vantages of such behavior, including avoidance of midwater predators, maintenance of position on the shelf, and increased encounters with high concentrations of food, have been discussed else- where (Barnett et al. 1984; Brewer et al. 1984; Brewer and Kleppel 1986; Jahn and Lavenberg 1986). In discussing the near-bottom schooling behavior of a larval clupeoid in Japan, Leis (1986) stated, "knowledge of the biology of epibenthic fish larvae is too rudimentary to allow a clear assessment of the advantages and disadvan- tages. . . ." Whatever the advantages, a seemingly more answerable question about the near-bottom habit is what causes the larvae to behave as they do? In another study from Japan, Tanaka (1985) showed that juvenile red sea bream, Pagrus major, exploited suprabenthic copepod popula- tions, and he speculated that the distribution of prey was a template for the descent of the fish from midwaters and its subsequent migration into estuaries. The question addressed in the present study was whether the fine-scale layering of larval fishes was a direct response to that of their prey field. Because of the immediate behavioral aspect of Manuscript accepted Februar>' 1988. FISHERY BULLETIN: VOL. 86, NO. 2, 1988. 251 FISHERY BULLETIN: VOL 86, NO 2 the question posed, a 1-d study was thought appropriate. Though environmental conditions on this day might differ from "average", fish lar- vae were assumed to be capable of a constant array of behaviors. In other words, response (if any) of the larvae to the vertical distribution of their prey was assumed to be a deterministic rather than a statistical phenomenon. If their vertical distribution resembled that of their prey, then food-seeking would remain a plausible ex- planation for the near-bottom habit; if not, then other stimuli must be considered important in shaping these near-bottom concentrations of fish larvae. The sampling was planned for daylight hours, when most feeding by larvae was expected to occur (Hunter 1981; Govoni et al. 1983). Late win- ter was chosen because in this season peak larval abundances of several species of interest to us (northern anchovy, Engraulis mordax; white croaker, Genyonemus lineatus; California hal- ibut, Paralichthys californicus; and sometimes queenfish, Seriphus politus, often overlap (Lavenberg et al. 1986). A survey cruise in late February found moderate-to-high abundance of the first three species plus California sardine, Sardinops sa^ox, (all >0.2 m"'^, Lavenberg un- publ. data), and so this study was scheduled for 19 March 1985 off Seal Beach, CA (lat. 33°41'N, long. 118°05'W; for a map, see Jahn and Laven- berg 1986). As it happened, we chanced to encounter condi- tions that were less typical than those found on the February cruise. Only one fish species, white croaker, was abundant enough to merit analysis, and an uncommonly reported prey item, rotifera, was important for small larvae. The diet of various-sized larvae with respect to the abund- ance of prey organisms at an array of heights above the seabed was nevertheless useful in ques- tioning whether food-seeking shaped the ob- served larval distribution. METHODS Field At the hour of 0750 PST, an array of Interocean model S4^ electromagnetic current meters was set out over the 15 m isobath, with current meters 1, 4, and 8 m above the seabed. These meters were ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. set to record average current vectors and tem- perature at 5-min intervals. The vessel (RV West- wind) was then anchored some 200 m seaward of the current meter array. A Nielson model NCH fish pump, rated at 227 m'^ h^ at a 2 m head, was used to sample fish larvae and zooplankton. The end of the hose was tethered between a 200 kg flat steel weight and several subsurface floats, with a pulley arrangement such that divers could adjust the distance between hose mouth and seabed. A similar setup was previously found to give re- peatable, fine-scale resolution at vertical sep- arations of 25 cm (Jahn and Lavenberg 1986). Sampling heights above the bottom were 50 cm, 1 m, and 6.7 m. The 15.2 cm diameter hose was nearly horizontal at the tether point, so that nominal sampling strata were z ± 7.6 cm. Ves- sel surge, transmitted through the stiff hose, caused occasional downward excursions of some 10 cm. Accompanying each pump sample was a cast of water bottles for phytoplankton and microplank- ton analysis. Rigid arrays of horizontally held 4 L Niskin bottles (of. Owen 1981) were used to take water samples simultaneously from 25, 50, and 100 cm above the bottom. The bottle array was designed to be tripped by messenger, but poor performance led to diver-implemented use after the second cast. A midwater sample, 7.5 m below the surface, was obtained via a single Niskin bot- tle for each sample set. The sampling plan thus consisted of duplicate pump samples from each of three strata, each pump sample to be accompanied by a set of bottle samples from four standard heights, three within 1 m of the seabed and one at midwater column. One-liter samples from the bottles were fixed in Lugol's solution for later identification of phyto- plankton and microplankton. Pump samples of 15-min duration (approximately 35 m^) were mainly directed into an overboard, 330 |xm mesh plankton net for retention of large zooplankton and ichthyoplankton. Unexpected problems in reading an inline flowmeter required that vol- umes be estimated as 2.4 m"^ min"^, based on pre- vious experience with the pump under similar conditions aboard the same vessel. To collect smaller zooplankton, a 5 cm diameter hose led from the intake side of the fish pump to a 100 jjtm mesh plankton net. This small-meshed net was suspended over a watertight box, which was marked such that exactly 0.5 m'^ could be subsam- pled for animals too small to be quantitatively retained by the large net. This subsample, which 252 JAHN ET AL : FOOD-SEEKING LARVAL WHITE CROAKER was first seived through 330 p.m mesh, took about 10 minutes to obtain; the portion retained on the 330 |jim mesh was added to the contents of the large plankton net. All pump samples were pre- served in 5% formalin. Laboratory All fish larvae and eggs were sorted from the large zooplankton samples and identified. All specimens of white croaker, the only species abundant in all six collections, were measured with an eyepiece micrometer in units of 0.024, 0.062, or 0.159 mm, depending on magnification. Length was measured from tip of snout to end of straight (NL) or fiexed (FL) notochord or to the end of the hypural plate when this margin was vertical (SL). A further designation of de- velopmental stage indicated the amount of yolk present: "free embryos" (Balon 1975) had a relatively massive yolk sac and may or may not have had functional eyes and mouths; more advanced individuals with a much-reduced or totally resorbed yolk sac, fully pigmented eyes, and an apparently functional mouth were designated "feeding-stage" larvae, or simply "larvae". All larvae, plus a maximum of 20 free embryos with apparently functional mouths from each col- lection, were dissected for gut contents analysis by methods described in Arthur (1976) and Gadomski and Boehlert (1984). Length, rather than width, of prey items was measured, because it was considered a more conservative property of often crushed specimens and because our concern was not so much with what the larvae could eat (Hunter 1981) as with what they did eat. Lengths of prey items (of copepods, cephalothorax length) were recorded in 50 fxm classes up to 200 iJim, by 100 ji-rn classes from 200 \x.m to 1 mm, and by 0.5 mm classes at larger sizes. In a few cases, these size categories were inconvenient, and more in- clusive ranges were used. Water bottle samples of phytoplankton and mi- crozooplankton were prepared following proce- dures in Utermohl (1931). From a thoroughly agitated sample, a 50 mL subsample for net phytoplankton was taken and placed in a settling chamber overnight (about 14—18 hours). Cells were identified and counted in 10 ocular fields, and mean density (cells per liter) calculated as the number counted scaled by the proportion of the area of the 10 fields (20.6 mm^ total) to the area of the slide (510.7 mm^). Microzooplankton was filtered from a 500 mL subsample onto a 35 fxm mesh screen, washed from the screen into a 50 mL settling tube and allowed to settle overnight. All organisms >50 fxm were counted and identified to taxon and size category, using the same system as for larval fish gut contents. Densities were scaled to number per liter. The 100 ^jLm zooplankton samples were concen- trated to 200 mL, then subsampled twice using a 10 mL Stempel pipette. Organisms were identi- fied and classified to size categories as described above for larval fish prey. Counts from two sub- samples were averaged and expressed as number per m"^. Data Analysis The microzooplankton (from water bottles) data set consisted of six vertical profiles of four sampling heights each. Principal components analysis was used to look for vertical layering and time-correlated changes in the makeup of these assemblages. A list of taxa present in three or more samples from at least one sampling height was chosen. Abundances were log-transformed [logio(jc + D], and principal components com- puted from the covariance matrix. Component scores for each of the 24 samples were used to make plots in which two- and three-dimensional groupings were sought that could be clearly re- lated to sampling height or to the sequence in which the samples were taken. The taxa having high loadings on axes (components) identified with time and vertical trends were subsequently scrutinized individually. A similar analysis was done for phytoplankton, but omitted here in the interest of brevity. Gut contents were conveniently analyzed by lumping taxa into the 10 categories: dinoflagel- late, tintinnid, rotifer, polychaete larva, lamelli- branch larva, crustacean nauplius, copepodite and adult copepod, amphipod, invertebrate egg, and "other". Unidentifiable matter was ignored in all comparisons. To test for differences in diet between subsets of larvae, we used an adaptation of the "bootstrap" (Efron 1982). The test criterion was the percentage of prey comprised by a major item in one of the two groups of guts. The null hypothesis that two sets were not different was simulated by combining the two data sets and then, through repeated sampling, determining the probability of observing the criterion percent- age from such a mixture. 253 FISHERY BULLETIN: VOL. 86, NO 2 RESULTS General Observations The water column was very weakly stratified, with temperatures of 12.9 ± 0.1°C at 1 m, 13.0 ± 0.2°C at 4 m and 14.1 ± 0.1°C 8 m above the bottom during the time of biological sampling. Near the bottom, a turbid suspension limited vis- ibility to arm's length; the surface of the sediment was never clearly seen on any of the seven de- scents during the hours of 0930-1630. The mid- waters below about 3 m from the surface were densely populated with larvaceans (visually esti- mated and later confirmed to be about 10 L"^). Total diatom cell counts (principally Nitzschia spp.) were of order 10^ L~^ in all samples, bloom quantities suggestive of recent upwelling (cf. Tont 1981). Currents and Plankton During the hours of biological sampling, cur- rents ran steadily alongshore to the southeast, being deflected counterclockwise near the bottom and ranging from about 14 cm s"^ at 8 m to 6 m s"^ at 1 m above the seabed. At these current speeds, one may expect that the approximately 5-h period from beginning to end of biological sampling should correspond to a minimum spa- tial spread of 1-2.5 km. Distances of this order were previously found to be an important length scale of variation in larval fish abundance ( Jahn and Lavenberg 1986). Because the spatial dimen- sion of interest regarding distribution of larval fish prey was the vertical, we needed to quantify, at least partially, the effects of time (vertical migration?) and distance (advection) on the com- position and vertical dispostion of the plankton. Component 4 (11%) 2 - 1 - CM CM C 0) c o o. E o o - 1 - ■2 - FIGURE 1. — Projections of microzooplankton samples onto the first and fourth principal com- ponent axes. The initial digit represents profile number, M = midwater, B = near-bottom, final digit is proximity to bottom (1 = 25 cm, 2 = 50 cm, 3 = 100 cm), see Figure 3. 254 JAHN ET AL.: FOOD-SEEKING LARVAL WHITE CROAKER Accordingly, the microplankton data set, repre- senting six vertical profiles separated in time, was reduced to principal components for exami- nation of possible time effects. Twenty-four taxonomic/size categories of mi- crozooplankton were used to compute principal component scores for the 24 samples. The first four principal components accounted for 60% of the variance. No clear separation of midwater from near-bottom samples was seen. The first component, which accounted for 22% of the vari- ance, separated the near-bottom samples into two groups, morning to midday and afternoon (Fig. 1), leaving the midwater samples at intermediate projections. The midwater samples were in turn separated by the fourth component (11% of the variance) into time groups corresponding to those of the near-bottom set. No stratification by sam- pling height was seen within the near-bottom samples, and none of the other axes provided sep- aration by time. The highest loading variables on components 1 and 4 (Table 1) were various sizes of rotifer and, for component 1, three genera of tintinnids (Favella, Acanthocystis , and Da- dayella ). Much of the time-correlated variance structure depicted in Figure 1 thus appears to be due to change in the size composition of rotifers, described in a later section, and a decrease in these three tintinnids near the bottom in late af- ternoon (Table 2). An identical analysis of the Table 1. — Loadings of important variables on tfie first and fourth principal components of microplankton data. Component 1 Component 4 Variable Loading Variable Loading Rotifer 150-200 ^Lm 200-300 ^JLm Tintinnids Favella sp. Acanthocyctis Dadayella sp. sp. 0.572 -0.380 0.365 0.331 0.314 Rotifer, 100-150 M-m Egg, 50-100 \i.m Copepod nauplii, 150-200 ^JLm 0.388 0.349 -0.252 phytoplankton data found no trends in time or depth. Larval Fish Abundance Of 1,125 total fish larvae taken in the six pump samples, 666 (59%) were white croaker, a deep- bodied, robust larva (Watson 1982). More than half (338) of these had absorbed the yolk sac and were thus of feeding size. The second most abun- dant feeding-stage larva was an unidentified gob- iid type (84 specimens), but this taxon was not taken above 100 cm of the seabed and so was excluded from the gut analysis. Feeding-stage California sardine, northern anchovy, and Cali- fornia halibut — all relatively abundant O0.2 m""^) in the area three weeks earlier — each repre- sented <1% of the catch. Although the earlier survey employed oblique bongo net tows, past comparison of the Nielsen pump with bongo tows found no significant differences in diversity or abundance estimates based on similar-volume samples (R. Schlotterbeck^). We therefore think the differences between the February survey and our March samples were due mainly to a real change in the ichthyoplankton, from a typical late winter assemblage (McGowen 1987; Walker et al. 1987) to a more depauperate one. Vertical Distribution and Feeding Incidence of Larval White Croaker White croaker free embryos ranged in abun- dance from <0.1 m"'^ at 0.5 m to ~1 m"-^ at 1 m to >2 m-3 at 6.7 m above the bottom. Of 61 free embryos dissected, none had gut contents. Feeding-stage larvae of white croaker were only slightly more abundant at 6.7 m (1.9-2.2 m"3) than at 1 m and 0.5 m (1.1-1.6 m'^), but 3R. Schlotterbeck, Robert Schlotterbeck, Inc., 18842 Ridgeview Cr., Villa Park, CA 92667, pers. commun. April 1986. Table 2. — Density (cells per liter) of three tintinnids as a function of time and sampling height. Each set of three numbers gives the density of Favella spp. (F), Acanthocystis spp. (A), and Dadayella spp. (D). Time (PST) Height F 1030 A D 1130 1220 1313 1425 1454 (cm) FAD F A D FAD FAD FAD 750 100 50 25 6 2 2 4 20 4 24 6 18 6 4 4 22 30 18 24 26 12 44 14 12 44 6 4 12 18 48 2 24 22 54 2 12 2 26 4 10 2 26 46 12 14 8 4 24 4 42 6 18 26 6 12 255 there was a marked gradient in development with proximity to the seabed (Fig. 2). All the larvae at 6.7 m had unflexed notochords, and most were <2.5 mm NL. Feeding incidence (proportion of larvae with nonempty guts) was 78% at this height. At 1 m, modal larval length was 2.65 mm, FISHERY BULLETIN: VOL. 86, NO. 2 with a single postflexion specimen (Fig. 2); feed- ing incidence was 74%. At 0.5 m there were still some preflexion larvae, but a second length mode at 6.8 mm represented postflexion-stage larvae. Feeding incidence was 90%) at 0.5 m above the bottom, being somewhat greater among flexion 39 n O CO > c5 o J 3 ^ 13 -1 G. lineatus GUTS WITH CONTENTS GUTS EMPTY (excludes yolksac stage) 670 cm (n = 109) 2.0 2.5 3.0 in ' i 'i 100 cm (n=70) 2.0 2.5 3.0 ' I ' I n ' I ' I ' I 5 10 ^ fl#^ 2.0 2.5 3.0 50 cm (n=83) Length (mm) Figure 2. — Length frequencies of feeding-stage larvae of Genyonemus lineatus at three sampling heights. 256 JAHN ET AL.: FOOD-SEEKING LARVAL WHITE CROAKER and postflexion larvae (95%) than among preflex- ion larvae (82%). Gut Contents The white croaker larvae were divided into three size classes for analysis of gut contents with regard to height above the bottom: preflexion lar- vae <2.7 mm (size 1), preflexion larvae >2.7 mm (size 2), and flexion and postflexion larvae (size E o o 84 guts 373 prey E o o o 3). The largest preflexion larva was 4.6 mm NL, and the smallest flexion stage larva was 5.5 mm FL. The division of preflexion larvae at 2.7 mm retained all but one specimen at 6.7 m in size 1 while partitioning the preflexion larvae at 1 m and 0.5 m about equally into sizes 1 and 2 (Figs. 2, 3). Besides the 2.75 mm specimen at 6.7 m, a single flexion stage larva at 1 m was excluded by these criteria from the comparisons. 31 guts 82 prey 21 guts 66 prey DINOFLAGELLATES TINTINNIDS ROTIFERS POLYCHAETE LARVAE LAMELIBRANCH LARVAE NAUPLII COPEPODS AMPHIPODS INVERTEBRATE EGGS OTHER ^ ^ ^ Hii ESS E o o it) 8 guts 15 guts 52 guts 22 prey 26 prey 194 prey SIZE 1 SIZE 2 SIZE 3 Figure 3. — Percentage contribution of 10 food categories to the diet of larval white croaker at three heights above the bottom. Size 1 = preflexion larvae <2.7 mm NL; size 2 = preflexion larvae >2.7 mm NL; size 3 = flexion and postflexion-stage larvae. 257 FISHERY BULLETIN: VOL. 86, NO. 2 Most identifiable prey items fit into the nine categories: dinoflagellate, tintinnid, rotifer, poly- chaete larva, lamellibranch larva, crustacean nauplius, copepodite and adult copepod, am- phipod, and invertebrate egg (Fig. 3). The "other" category applied only to size-2 larvae at 1 m (1 Globigerina sp.) and to size-3 larvae (1 zoea, 3 larvaceans, and two large [1 mm] unidentified spheres). Guts of preflexion (sizes 1 and 2) larvae from the three sampling heights contained an array of small (<300 jxm) organisms that varied mainly in proportions from mostly rotifers (88%) in size-1 larvae at 6.7 m to a diverse mix of prey numeri- cally dominated by nauplii in size-2 larvae at 0.5 m (Fig. 3). Percent similarity (overlap) among the 5 groups of preflexion larvae ranged from 24 to 75%. The gut of the single size-2 larva at 6.7 m, not included in Figure 3, contained two tintin- nids. Size-3 larvae had a diet consisting chiefly of copepodite and adult copepods that overlapped only 8-9% with size-2 larvae and 1% or less with the three groups of size-1 larvae. The copepods eaten by size-3 larvae were mostly Corycaeus an- glicus (62% of all copepods), unidentified cope- podites (cyclopoid and calanoid, 25%), and Para- calanus parvus (9%). Polychaete larvae were identified only from the presence of setae in the guts, so the proportion (nominally 16% of all prey items) of this taxon in the diet is more an indica- tion of incidence than of numerical importance. Amphipods, mostly in the length range 1-1.5 mm, were found in white croaker larvae ranging from 6.5 mm FL to 10.3 mm SL. The gut of the flexion-stage larva at 1 m, not included in Figure 3, contained three C. anglicus and traces of poly- chaete setae. While there can be no doubt that flexion and postflexion larvae had a different diet than pre- flexion larvae, the pattern of decreasing propor- tion of rotifers with increasing size and proximity to the bottom among preflexion larvae was of questionable statistical significance. The first question asked was whether the very high per- centage of rotifers in the diet of size-1 larvae at 6.7 m was likely to have arisen by chance from a random sampling of size-1 larvae. Formally stated, Hq = "all size-1 larvae had the same per- centage of rotifers". The 123 nonempty guts were pooled, and random samples of 84 each were drawn. In 1,000 iterations, <4% of the samples had >88% rotifers, so it was concluded that lar- vae at 6.7 m ate significantly more rotifers than similar-sized larvae near the bottom. The remain- ing 75 preflexion larvae (sizes 1 and 2) are divided into 4 small groups at 0.5 and 1 m, so we next tested for a size effect by pooling across sampling height, such that the guts of the 39 near-bottom size-1 larvae contained 64% rotifers, and the 36 size-2 larvae had 36% rotifers. Bootstrapping as before, <2% of samples of 36 had <36% rotifers, so it was concluded that size-1 and size-2 larvae differed in this regard. Further testing (e.g., of a height effect within sizes) was not done because of small sample sizes and multiple testing consider- ations. Abundance and Vertical Distribution of Prey Rotifers, all identified as the brachionoid Tri- chocerca sp., figured importantly both in the diet of preflexion larvae and in the time-related vari- ance structure of the microplankton. As shown in Table 3, there was a change in the size spectrum of these animals that coincided approximately with the time of changing from near-bottom sam- pling to midwater sampling with the fish pump. It was only the largest category of rotifer (200-300 |xm, including the "toe") that was found in the guts of the larvae. The relative abundance of total rotifers in the plankton at the times and heights of pump sampling differed very little (25-33% of all organisms in the 100-300 ^JLm size class), but the percentage of rotifers in the 200-300 [xm class increased from 21% (near-bottom, morning) to 86% of all rotifers (midwater, afternoon). The dominance of rotifers in the diet of size-1 larvae in midwaters is thus likely related to the larger size of rotifer resident in the water column when that height was sampled. The most notable dietary difference among the larval size groups analyzed was the switch from small (50-300 (xm) to larger (0.5-2.5 mm) prey, principally the copepod Corycaeus anglicus (0.5- 0.8 mm), upon flexion of the notochord. The abun- dance of Corycaeus from the 100 ixm mesh pump samples (Table 4) shows that this prey item was equally or more abundant in midwater than near the bottom, where all the flexion and postflexion larvae were captured. (Within the bottom meter, the similar-sized but more transparent Para- calanus parvus outnumbered C anglicus by a fac- tor of 5-20.) The only prey found in numbers in these larvae that was restricted to the 0.5 m sam- ples was gammarid amphipods. Larger crus- taceans — cumaceans, crab and shrimp zoea, 258 JAHN ET AL.: FOOD-SEEKING LARVAL WHITE CROAKER Table 3. — Density (rotifers per liter) of Trichocerca sp. as a function of time and sampling heighit. Eachi set of three numbers gives the density of 100-150 M-m, 150-200 txm, and 200-300 ^.m rotifers. Height of simultaneous pump sample is given. Time (PST): Pump height: Sampling height (cm) Time (PST): Pump height: Sampling 1030 0.5 m 1130 0.5 m 1220 1 m 100- 150 pim 150- 200 ^.m 200- 300 pirn 100- 150 \i.m 150- 200 M-m 200- 300 M-m 100- 150 pLm 150- 200 M.m 1313 1 m 1425 6.7 m 1454 6.7 m 200- 300 M-m 750 8 4 4 10 20 100 8 14 66 14 22 10 50 24 18 8 40 34 25 2 12 16 46 30 height 100- 150- 200- 100- 150- 200- 100- 150- 200- (cm) 150 fim 200 (xm 300 M-m 150 ^.m 200 M.m 300 M-m 150 M-m 200 M-m 300 M-m 750 20 12 14 60 100 4 12 24 46 102 50 6 34 4 16 42 25 26 6 2 Table 4. — Abundance (animals m 3) of cope- podite and adult Corycaeus spp. in 100 [im mesh samples from the fish pump. Height (m) First sample Second sample 6.7 1 0.5 1,080 120 180 460 500 140 mysids, and euphausid furcilia larvae — were all abundant OlO m""^) in the 0.5 m, 330 ixm mesh samples but with the exception of the callianassa zoea mentioned above (from the gut of an 11 mm larva) were not found in these white croaker lar- vae. DISCUSSION The chief drawback of the pumping system used was its inability to obtain a simultaneous vertical profile. The sampling sequence left the possibility that differences among heights might be confounded by trends in time, as discussed by Jahn and Lavenberg (1986). Slight time effects were found among the vertical profiles of mi- croplankton, increasing the suspicion that the ap- parent vertical distributions of fish larvae and macrozooplanktonic prey might have horizontal components. To contradict the argument that food-seeking did not bring postflexion larvae near the bottom, one would need to invoke either an afternoon increase of some two orders of magni- tude in copepod abundance (Table 4) or else the presence of flexion and postflexion larvae throughout the water column in morning and midday followed by their sudden disappearance in the afternoon. A two-order-of-magnitude change in copepod species abundance over a distance of roughly 1 km (2 hours at 14 cm s"M is certainly possible; though zooplankton structures reported from the southern California continental shelf are gener- ally larger than this (Star and Mullin 1981; Bar- nett and Jahn 1987), there is always the possibil- ity of sampling the edge of a patch. Since no such edge was evident in the abundance or overall composition of microplankton or of phytoplank- ton, it seems unlikely that a macrozooplankton change of this order occurred. Moreover, the main copepod eaten, Corycaeus anglicus, is generally more abundant in midwater than near the bottom over the shallow shelf (A. Barnetf^), in accord with its apparent distribution in this study. As to a possible midwater abundance of postflexion white croaker larvae, no such concentration has ever been reported. In some nine vertical profiles taken in daylight over a 6-d period. Brewer and Kleppel (1986) took virtually all specimens >3.5 mm in their near-bottom sampler. White croaker appears similar to another abundant sciaenid, queenfish, in this regard (cf Jahn and Lavenberg 1986). 4A. Barnett, Marine Ecological Consultants, 531 Encinitas Blvd.. Encinitas, CA 92024, pers. commun. July 1987. 259 FISHERY BULLETIN: VOL. 86, NO 2 The only unequivocal instance in which a prey item of larval white croaker was vertically dis- tributed similarly to the larvae was the trace of amphipods found in the guts of competent (flexion and postflexion) larvae. At the lengths of larvae sampled (<12 mm) the prey were all planktonic and nearly all about equally abundant in mid- waters as near the bottom. The small numbers of amphipods eaten may indicate an incipient tran- sition to larger, suprabenthic crustacean prey. The size gap between the large prey of these com- petent larvae and the smaller prey of preflexion larvae is probably an artifact of the bimodal size distribution of sampled larvae. Though all of the prey eaten by size-1 (<2.7 mm) larvae were <300 |xm in length, the more varied diet of larger pre- flexion larvae contained some copepods as big as 500 |xm. There is therefore nothing in these data to suggest that the switch from microplanktonic to macroplanktonic prey is anything but a grad- ual transition as the larvae grow. Brewer and Kleppel (1986) also reported a change to copepod prey in white croaker larvae >6 mm. Our findings are further similar to those of Brewer and Kleppel in that there was no indi- cation that food-seeking had anything to do with the descent of larval white croaker from mid- waters to the near-bottom zone. The other defin- able dietary trend in this study (besides ontoge- netic change) was the high percentage of rotifers eaten by midwater preflexion larvae. This was apparently related to subtle but important differ- ences in the available planktonic prey — signifi- cantly, to a greater abundance of suitable-size ro- tifers — at the time the midwater stratum was sampled. It seems safest to conclude that white croaker larvae descend toward the bottom for reasons quite apart from seeking food (see discussions in Barnett et al. [1984], Brewer and Kleppel [1986], Jahn and Lavenberg [1986]) and simply eat what- ever they find there that suits them. Many poten- tial macroplanktonic prey also favor the near- bottom layer (Jahn and Lavenberg 1986; Barnett and Jahn 1987). Older larvae and their prey may occupy the near-bottom layer for different rea- sons, or it may be that a single advantage, or set of pressures, underlies the behavior of these di- verse planktonic and semi-planktonic taxa. Some species need to remain near shore, and living in the bottom boundary layer helps assure this. The boundary layer also tends to be more turbid than overlying waters and so may lessen an animal's jeopardy to visual (biting) planktivores. (The gen- erality of the latter explanation only holds if suprabenthic fish larvae are less important planktivores than other water-column inhabi- tants — see Gushing 1983.) Rotifers have never to our knowledge been re- ported as an important food of ocean-caught fish larvae, even though the genus Brachionus is com- monly cultured for feeding larval fish in the labo- ratory. Schmitt (1986) reported that small, laboratory-reared larval northern anchovy read- ily fed upon (unidentified) wild-caught rotifers. Rotifers are only occasionally abundant in neritic waters, and never in oceanic waters (J. Beers^). Their rarity notwithstanding, rotifers have the ability very rapidly to dominate marine mi- croplanktonic assemblages (Hernroth 1983), and their good food quality (Theilacker 1987) and high secondary productivity for a period of weeks might constitute a significant enhancement to growth and survival of a larval fish cohort. Our previous experience in handling larval white croaker specimens agrees with the findings of Brewer and Kleppel (1986) in that lamelli- branch larvae, easily seen through the body wall, are a common food for small white croaker larvae. In our study, this taxon was a minor constituent of the plankton and of the larval fish diet. We cannot say how unusual were the circumstances we encountered, but we know that in terms of diatom numbers and larval fish diversity these conditions were not typical of March on the south- ern California continental shelf That white croaker larvae appeared to find these conditions salubrious may be one reason this species is so successful in southern California (Love et al. 1984). ACKNOWLEDGMENTS Thanks go to S. Caddell, R. Feeney, T. Garrett, R. Lavenberg, J. McGowen, J. Petersen, J. Rounds, and Captain L. Nufer for able partici- pation in the field work. D. Carlson-Oda, J. Rounds, and S. Shiba helped process larval fish samples, and H. Schwarz helped prepare the manuscript. K. Zabloudil generously loaned the current meters, and R. Erdman assisted in proc- essing data therefrom. We also thank J. Beers and R. Brusca for help in accessing literature on rotifers. D. Cohen, R. Lavenberg, and J. Petersen 5J. Beers, Scripps Institution of Oceanography, La Jolla, CA 92093, pers. commun. November 1986. 260 JAHN ET AL.: FOOD-SEEKING LARVAL WHITE CROAKER reviewed the manuscript. The Southern Califor- nia Edison Company funded the study. LITERATURE CITED Arthur. D K 1976. Food and feeding of larvae of three fishes occurring in the California Current, Sardinops sagax, Engraulis mordax, and Trachurus symmetricus . Fish. Bull., U.S. 74:517-530. Balon, E K. 1975. Terminology of intervals in fish development. J. Fish. Res. Board Can. 32:1633-1670. Barnett. a. M . AND A E Jahn. 1987. Pattern and persistence of a nearshore planktonic ecosystem off southern California. Cent. Shelf Res. 7:1- 25. Barnett, A M , A E Jahn, P D. Sertic, and W Watson. 1984. Distribution of ichthyoplankton off San Onofre, California, and methods for sampling very shallow coastal waters. Fish. Bull., U.S. 82:97-111. Brewer, G D , and G S Kleppel 1986. Diel vertical distribution of fish larvae and their prey in nearshore waters of southern California. Mar. Ecol. Prog. Ser. 27:217-226. Brewer, G D , G S Kleppel, and M Dempsey. 1984. Apparent predation on ichthyoplankton by zoo- plankton and fishes in nearshore waters on southern California. Mar. Biol. (Berl.) 80:17-28. Brewer, G. D , R L Lavenberg. and G E. McGowen 1981. Abundance and vertical distribution of fish eggs and larvae in the Southern California Bight: June and October 1978. Rapp. P. -v. Reun. Cons. int. Explor. Mer 178:165-167. Buckley, L J., and R G Lough 1987. Recent growth, biochemical composition, and prey field of larval haddock tMelanogrammus aeglefinus) and Atlantic cod iGadus morhua ) on Georges Bank. Can. J. Fish. Aquat. Sci. 44:14-25. Clutter, R I 1969. The microdistribution and social behavior of some pelagic mysid shrimps. J. Exp. Mar. Biol. Ecol. 3:125- 155. Gushing, D H 1983. Are fish larvae too dilute to affect the density of their food organisms? J. Plankton Res. 5:847-854. Efron, B 1982. The jackknife, the bootstrap and other resampling plans. Soc. Ind. Appl. Math., Philadelphia, PA, 92 p. GaDOMSKI, D M , AND G W BOEHLERT 1984. Feeding ecology of pelagic larvae of English sole, Parophrys vetulus , and butter sole, Isopsetta isolepis, off the Oregon coast. Mar. Ecol. Prog. Ser. 20:1-12. GOVONI, J. J., D E HOSS, AND A J CHESTER 1983. Comparative feeding of three species of larval fishes in the northern Gulf of Mexico: Brevoortia patronus, Leiostomus xanthurus, and Micropogonias undulatus. Mar. Ecol. Prog. Ser. 13:189-199. GovoNi, J J , A J Chester, D E Hoss. and P B Ortner. 1985. An observation of episodic feeding and growth of larval Leiostomus xanthurus in the northern Gulf of Mexico. J. Plankton Res. 7:137-146. Hamner, W M , AND J. H Carleton 1979. Copepod swarms: attributes and role in coral reef ecosystems. Limnol. Oceanogr. 24:1-14. HERNROTH, L. 1983. Marine pelagic rotifers and tintinnids — important trophic links in the spring plankton community of the Gullmar Fjord, Sweden. J. Plankton Res. 5:835-846. Hunter. J R 1981. Feeding ecology and predation of marine fish lar- vae. In R. Lasker (editor). Marine fish larvae: morphol- ogy, ecology, and relation to fisheries, p. 34-77. Wash- ington Sea Grant Program, Seattle. Jahn. A E.. and R. J Lavenberg. 1986. Fine-scale distribution of nearshore, suprabenthic fish larvae. Mar. Ecol. Prog. Ser. 31:223-231. Lasker. R 1975. Field criteria for survival of anchovy larvae: the relation between inshore chlorophyll maximum layers and successful first feeding. Fish. Bull., U.S. 73:453- 462. 1978. The relation between oceanographic conditions and larval anchovy food in the California current: identifica- tion of factors contributing to recruitment. Rapp. P. -v. Reun. Cons. int. Explor. Mer 173:212-330. Lavenberg, R. J., G. E McGowen, A E Jahn. J. H. Petersen, AND T C Sciarrotta 1986. Southern California nearshore ichthyoplankton: a study of abundance patterns. CalCOFI Rep. 27:53- 64. Leis, J M 1986. Epibenthic schooling by larvae of the clupeid fish Spratelloides gracilis . Jpn. J. Ichthyol. 33:67-69. Love, M S , G E. McGowen, W Westphal, R. J Lavenberg, and L Martin 1984. Aspects of the life history and fishery of the white croaker, Genyonemus lineatus (Sciaenidael, off Califor- nia. Fish. Bull., U.S. 82:179-198. McGowen, G E 1987. Coastal zone ichthyoplankton assemblages off Southern California. Ph.D. Thesis, Univ. Southern California, Los Angeles, 304 p. Owen. R W. 1981. Microscale plankton patchiness in the larval an- chovy environment. Rapp. P. -v. Reun. Cons. int. Ex- plor. Mer 178:364-368. SaINTE-MaRIE, B . AND P BRUNEL. 1985. Suprabenthic gradients of swimming activity by cold-water gammaridean amphipod Crustacea over a muddy shelf in the Gulf of Saint Lawrence. Mar. Ecol. Prog. Ser. 23:57-69. Schlotterbeck, R E , and D W Connally 1982. Vertical stratification of three nearshore southern California larval fishes (Engraulis mordax, Genyonemus lineatus, and Seriphus politus). Fish. Bull., U.S. 80:895-902. SCHMITT. P D 1986. Prey size selectivity and feeding rate of larvae of the northern anchovy, Engraulis mordax Girard. CalCOFI Rep. 27:153-161. Star. J. L.. and M M Mullin. 1981. Zooplanktonic assemblages in three areas of the North Pacific Ocean as revealed by continuous horizontal transects. Deep-Sea Res. 28A:1303-1322. Tanaka, M. 1985. Factors affecting the inshore migration of pelagic larval and demersal juvenile red sea bream, Pagrus major, to a nursery ground. Trans. Am. Fish. Soc. 114:471-477. 261 FISHERY BULLETIN: VOL. 86, NO. 2 Theilacker. G 1987. Feeding ecology and growth energetics of larval northern anchovy, EngrauUs mordax. Fish. Bull., U.S. 85:213-228. Theilacker, G , and K Dorsey 1980. Larval fish diversity, a summary of laboratory and field research. FAO Intergov. Oceanogr. Comm. Work- shop Rep. No. 28, p. 105-142. TONT, S. A. 1981. Historical changes of diatom abundance off south- ern California as reflected in sea surface temperature, air temperature, and sea level. J. Mar. Res. 39:191-201. Utermohl. N 1931. Neue Wege in der quantitativen Erfassung des Planktons. Verb. Int. Ver. Limnol. 5:567-597. Walker, H J . Jr , W Watson, and A. M Barnett. 1987. Seasonal occurrence of larval fishes in the near- shore Southern California Bight off San Onofre, Califor- nia. Estuarine Coastal Shelf Sci. 25:91-109. Watson, W. 1982. Development of eggs and larvae of the white croaker, Genyonemus lineatus Ayres (Pisces:Sciaenidae), off the southern California coast. Fish. Bull., U.S. 80:403-417. WiSHNER, K F 1980. The biomass of the deep-sea benthopelagic plank- ton. Deep-Sea Res. 27:203-216. 262 NEW MARINE DECAPOD CRUSTACEANS FROM WATERS INFLUENCED BY HYDROTHERMAL DISCHARGE, BRINE, AND HYDROCARBON SEEPAGE Austin B. Williams^ ABSTRACT Five species of decapod crustaceans new to science are described. These are caridean shrimps of the family Bresiliidae — Alvinocaris markensis from a Mid-Atlantic Rift Valley hydrothermal field, A. muricola from a cold brine seep at the foot of the West Florida Escarpment in the Gulf of Mexico, and A. stactophila from a hydrocarbon seep on the continental slope of the northern Gulf of Mexico, with a key to the species of Alvinocaris ; a squat lobster of the family Galatheidae — Munidopsis alvisca from the Guaymas Basin and from the Juan de Fuca and Explorer ridges in the eastern Pacific; and a brachyuran crab of the family Bythograeidae — Bythograea mesatlantica from a Mid-Atlantic Rift Valley hydrothermal field. Species of both Alvinocaris and Bythograea are now known from the eastern Pacific and Mid-Atlantic. Munidopsis species are widely represented in the world oceans. Deep ocean hydrothermally active fields and waters influenced by brine and hydrocarbon seeps continue to yield species new to science. Such en- vironments were unknown until explored with the aid of submersible research vessels from which observations and collections could be ac- complished. The species of decapod crustaceans reported here come from hydrothermal fields in the Mid- Atlantic Rift Valley, the Guaymas Basin in the Golfo de California, and Juan de Fuca and Explorer Ridges in the northeastern Pacific, a cold brine seep at the foot of the West Florida Escarpment, and a hydrocarbon seep on the con- tinental slope of the northern Gulf of Mexico. These are scattered localities that exhibit diverse environmental conditions but that are bound together by the common thread of chemotrophic food chains (Childress et al. 1986; Brooks et al. 1987). The material from the Mid-Atlantic Rift Val- ley, West Florida Escarpment, and Guaymas Basin was observed and collected by scientists working with the aid of the DSRV Alvin and RV Atlantis II based at the Woods Hole Oceano- graphic Institution. That from the northern Gulf of Mexico came from the Minerals Management Service Northern Gulf of Mexico Outer Continen- tal Slope (MMS/NGOMCS) Regional Office Proj- ect, involving observation and collection of mate- rial by scientists from LGL Ecological Research iSystematics Laboratory, National Marine Fisheries Service, NOAA, National Museum of Natural History, Washington, D.C. 20560. Manuscript accepted January 1988. FISHERY BULLETIN: VOL. 86, NO. 2, 1988. Associates and Texas A&M University, with the aid of the submersible research vessel Johnson- Sea-Link and its support vessels. Specimens from Explorer and Juan de Fuca Ridges were collected with the aid of the Canadian DSRV Pisces IV and its support vessels. All specimens studied are deposited in the Crustacean Collection of the United States Na- tional Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. CARIDEA: BRESILIIDAE Alvinocaris Williams and Chace, 1982 Three species of bresiliid shrimps described below as new to science are placed in the hereto- fore monotypic genus Alvinocaris. Certain fea- tures of these species necessitate minor changes in the generic diagnosis by Williams and Chace (1982) as follows: Rostrum with or without ven- tral teeth. Telson with 2-5 pairs of principal spines on posterior margin. Strong median ster- nal spine between posterior pair of pereopods. Moreover, the branchial formula seems uni- formly fixed in this genus as well as in the genus Rimicaris Williams and Rona, 1986. The arrange- ment, figured in Williams and Chace (1982) and Williams and Rona (1986) may be described as follows: Phyllobranchs extensively developed in 2 se- ries; asymmetrically Y-branched pleurobranchs with relatively short ventral and progressively longer and more expansive dorsal ramus associ- 263 FISHERY BULLETIN; VOL. 86, NO. 2 ated with pereopods 1-5; smaller and more nearly symmetrical U-branched arthrobranchs of more nearly uniform size associated with third maxil- liped and pereopods 1-4. Key to Known Species oi Alvinocaris 1. Telson with terminal margin slightly con- cave in midline and bearing 3-5 pairs of principal spines A. lusca Williams and Chace Telson with terminal margin convex and bearing only 2 pairs of principal spines ... 2 2. Rostrum with ventral margin bearing 0-1 subterminal ventral spines A. stactophila new species Rostrum with ventral margin bearing 4 or more subterminal ventral spines 3 3. Abdominal segment 3 with pleural margin entire A. markensis new species Abdominal segment 3 with pleural margin obscurely serrate A. muricola new species Alvinocaris markensis new species Figures 1, 2, 7 Material— \JSnU 234286, Holotype 9 (crushed), USNM 234287, Paratypes, 2 9 (dam- aged), Mid-Atlantic Rift Valley about 70 km south of Kane Fracture Zone (see Leg 106 Ship- board Scientific Party [1986]; Ocean Drilling Pro- gram Leg 106 Scientific Party [1986]), 23°22.09'N, 44°57.12'W, 3,437 m, Alvin Dive 1683, MARK vent, Stn. 1, scoop, 30 May 1986, pilot D. Foster, observers S. Humphris and J. Ed- mond. From NSF Ocean Drilling Program-Leg 106, NSF Grant OCE-8311201 to J. F. Grassle, Woods Hole Oceanographic Institution, Woods Hole, MA. Measurements in mm. — Holotype 9, postor- bital carapace length 4.16, rostrum 2.3, maxi- mum carapace height 3.3, total length about 15.6. Paratype 9, same 2.7, 1.6, 2.4, 10.9. Description. — Integument extremely thin, 264 membranous, shining, with a few minute puncta- tions. Rostrum (Fig. la, 6) almost straight, slightly elevated above horizontal in distal half, sharply pointed tip usually reaching to between midlength of second and tip of third peduncular articles of antennule; dorsal margin raised into thin serrate crest containing 12-17 teeth, strongest in central sector of row, with about 1/3 length of crest continued onto carapace; ventral margin less prominent and armed with 5-8 sub- terminal teeth; tooth formulas examined, 17/8 (holotype), 17/5, 12/5 (apparently some subtermi- nal dorsal teeth fused); strong lateral carina broadened proximally and confluent with orbital margin. Carapace with acute antennal spine dis- tinct; pterygostomian spine acuminate and prominent. Indistinct antennal groove curving ventrad to intersect associated indistinct groove at about midlength of carapace and continuing posteriad. Abdomen of female (Fig. Ic) apparently broadly arched dorsally (all specimens examined are crushed), gradually tapering posteriorly, nar- rowest part of sixth somite about 1/2 width of first somite; pleura of 3 anterior somites broadly rounded, that of fourth somite drawn posterolat- erally to strong acuminate spine flanked dorsally by or 1 much more slender and smaller spine; posterolateral corner of fifth pleuron strongly acuminate to nearly right angled and flanked dorsally by 0-2 spines of variable size analogous to condition on somite 4, spine number possibly age related; sixth somite with middorsal length about 1.9 that of fifth, broad based midlateral spine overlapping base of telson, smaller pos- terolateral spine acute; fourth somite with small erect spine on sternite and fifth with analogous strong, posteriorly directed spine. Telson (Fig. \d) elongate subrectangular, length about 3.5 anterior width, 5.8 posterior width, and about 1.75 length of sixth somite, not including poste- rior spines; armed with 6-8 dorsolateral spines of nearly uniform size, sometimes bilaterally unequal in number; posterior margin convex, armed with 2 principal spines at each corner and 10-12 feathered strong setae on margin be- tween. Eyes (Fig. la , 6 ) with cornea imperfectly devel- oped, unfaceted though diffusely pigmented, glob- ular to ovate in outline and with prominent spine on anterodorsal edge. Antennular peduncle (Fig. la, 6) reaching end of antennal scale; basal article 1.3 length of sec- ond and about 3.0 length of third, stylocerite well WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS Figure 1. — Alvinocans markensis, holotype $ : a part of cephalothorax and anterior appendages, lateral; b rostrum, eye, anten- nular peduncle, antennal scale, dorsal of left side; c abdomen, lateral; cf tail fan; cheliped, e mesial, /"lateral; g pereopod 2, h chela. Scales = 2 mm: I ia, c , d, e , f, g, h)\2 ib). 265 FISHERY BULLETIN: VOL. 86, NO. 2 separated from peduncle, tapering to slender elongate tip reaching about 1/4 length of second article, basal article with distodorsal margin flanked by transverse row of setae, extended into strong lateral spine reaching level equal to that of stylocerite and closely appressed to second article; shorter second article with strong mesiodistal ap- pressed spine. Dorsolateral flagellum about 1.5 length of carapace, thickened proximal half bear- ing conspicuous ventral setae; ventromesial flag- ellum somewhat more slender in lateral view and shorter. Antennal scale (Fig. la, 6) about 2.5 as long as wide, distolateral tooth falling short of distome- sial apex of broadly rounded distal margin of blade; basal article with strong ventrolateral spine; flagellum slightly exceeding length of body (missing from holotype). Mandibles (Fig. 2g) similar, with 2-segmented palp, incisor process broad and armed with 8 marginal teeth, slender molar process simple, divergent, its narrowly rounded tip minutely setose. First maxilla (Fig. 2h) with proximal endite asymmetrically oval-triangular, distal margin bearing many long setae; distal endite with nar- rowed base but broadened distally, armed with many (about 37) short spines on mesial margin and with scattered longer spinules marginally and submarginally beyond either end of spine row; palp scarcely bifurcated, with long distal spine on obsolescent proximomesial branch and 1 shorter submarginal spine on distal branch. Second maxilla (Fig. 2i) with proximal endite represented by 2 similar lobes; distal endite sub- triangular, expanded mesiodistally and paral- leled laterally by narrow somewhat twisted palp, scaphognathite with anterior lobe rectangulo- ovate, fringed with uniformly long, silky setae on anterior and mesial borders, shorter setae along entire lateral margin; posterior lobe narrowly ovate-triangular, fringed on blunt tip and adja- cent mesial margin by strikingly long, tangled, strong setae preceded proximally by shorter setae similar to those on lateral margin. First maxilliped (Fig. 2j, partly flattened view) with irregularly fusiform endite, short palp much exceeded in length and size by leaflike exopod, epipod obscurely bilobed. Second maxilliped (Fig. 2k, I) somewhat pedi- form but flattened, mesial margin of articles bearing long, feathered setae, mesial surface of terminal article densely setose, tip of exopod ex- ceeding leaflike epipod. Third maxilliped (Fig. 2m, n) slender, 5- segmented, reaching beyond antennular pedun- cle; terminal segment trigonal in cross section, tapered distally, bearing 3 terminal spines, oblique tracts of dense setae along mesial surface; similar tract of setae on carpus and another less conspicuous group on merus-ischium, latter with distolateral spine at articulation with carpus; ex- opod much reduced, ovate-triangular, without lash. First pereopods (Fig. le, f) chelate, subequal; fingers curved ventrally and slightly laterad; dactyl much more slender than and slightly longer than fixed finger; mesial surface of each finger convex, lateral surface deeply concave; pre- hensile surfaces uniformly offset, closing without gape, each armed with row of almost uniform teeth so closely set as to be almost contiguous, line of sensory hairs mesial to cutting edges, acute tip of dactyl slightly spooned by elongate teeth slanted distad and curving around its exter- nal edge. Leg not reaching tip of third maxilliped and exceeding antennal peduncle. Palm of holo- type female inflated, length slightly greater than height and shorter than fingers (0.60); low ridge ending in small hooked spine on proximomesial surface near articulation with carpus. Carpus longer than palm; bearing oblique ventral crest ending in strong distoventral spine and flanked mesially by patch of setae on triangular raised area; rectangular distal notch above spine fol- lowed by oblique distomesial margin leading to poorly defined spine at condyle articulating with palm; distolateral margin with rounded ventral corner leading to sinuous border above it bearing 2 lobes near articulation with palm. Merus some- what swollen is distal half and bearing small dis- tomesial spine, distinct from ischium but fused to it. Second pereopod (Figs. 1^, h; 11) shorter and more slender than first, but reaching beyond an- tennal peduncle by about length of fingers. Fin- gers slightly shorter than palm, similar in size and shape; opposed edges without gape, each spineless proximally, but distal half pectinate with single row of spines directed obliquely distad and increasing slightly in size to end in notice- ably stronger spine crossing opposite member when closed. Carpus slender, about 0.9 length of chela; merus and ischium unarmed. Third to fifth pereopods (Fig. 2a-f) similar in length and structure, third reaching distal edge of antennal scale. Length articles of these legs in holotype 9, mm: 266 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS Figure 2. — Alvinocaris markensis , holotype 9: a pereopyod 3, b dactyl; c pereopod 4, d dactyl; e p)ereopK)d 5, f dactyl; g mandible; h maxilla 1; i maxilla 2; j maxilliped 1 partly flattened; maxilliped 2, k endopod, ventral, I part of exopod dorsal; maxilliped 3, m ventral, n dorsal, distal articles only. Scales: Uk, I) = 1 mm; 2 (6, d,f,g, A) = 1 mm; 3(a,c,e,i,j,m,n) = 2 mm. 267 FISHERY BULLETIN: VOL. 86, NO. 2 3rd 4th 5th ischio- merus 6.14 5.76 5.44 carpus propodus 2.56 2.18 2.30 3.78 4.12 5.44 dactyl 0.64 0.70 0.64 Each short dactyl armed with 5 spines on flexor surface, grading from small proximally to longest and strongest distally, often a sensory seta on extensor surface. Propodi with setae along flexor surface progressively more crowded distally. Carpi with distodorsal extension projecting as a stop along proximal part of propodal extensor sur- face. Third leg stronger, at least in merus- ischium, than fourth and fifth; merus of third and fourth with closely appressed ventral spine at 1/2 and 3/4 length, that of fifth with spines at 1/3 and 2/3 length, distal spine strongest in each case; ischium of third and fourth leg with 2 spines in line with those on merus. Pleopods well developed, pair 1 with endopods about half length of exopods, tapering to acute tip; appendices internae simple, that of pair 5 with blunt tip. Uropod (Fig. Id) with rami subequal in length, slightly exceeding distal end of telson, lateral ramus with movable spine mesial to smaller dis- tolateral tooth, diaeresis sinuous. Remarks. — Remarks are given in the account for A. stactophila. Etymology. — The name is taken from an acronym for the site of collection in the Mid- Atlantic Ridge Valley about 70 km south of an area known as the Kane Fracture Zone, "MARK", and the Latin genitive suffix "ensis". Alvinocaris muricola new species Figures 3, 4, 7 Material .—\JSnU 234288, Holotype c^ (ceph- alothorax and abdomen broken apart), USNM 234289, Allotype 9, West Florida Escarpment, 26°01'N, 84°54.61'W, 3,277 m, Alvin Dive 1754, 15 October 1986, pilot W. Sellers, observers R. Carney and B. Hecker. USNM 234290, Para- type 9 , West Florida Escarpment, same locality, Alvin Dive 1753, 14 October 1986, pilot P. Tib- betts, observers R. Carney and G. Knauer. All from Barbara Hecker, Lamont Geological Ob- servatory, Columbia University, Palisades, NY. Measurements in mm. — Holotype 6, postor- bital carapace length 6.4, rostrum 4.4, maximum carapace height 4.5. Allotype 9 , same, 6.4, ros- trum broken, 5.6. Description. — Integument thin, shining, mi- nutely punctate. Rostrum (Fig. 3a, b) almost straight to slightly upturned in distal half, sharply pointed tip reaching to proximal part of third peduncular article of antennule; dorsal mar- gin raised into thin serrate crest containing 17-21 teeth varying from obliquely erect in proximal part to nearly horizontal, shorter and more dis- tant distally, about 1/3 length of crest continued onto carapace; ventral margin much less promi- nent and armed with row of 6 correspondingly smaller subterminal teeth, sometimes obscure; lateral carina broadened proximally and conflu- ent with orbital margin. Carapace (Fig. 3a ) with broad based but slender, acuminate antennal spine; pterygostomian spine correspondingly acuminate and prominent. Prominent anterior antennal carina curving posteroventrally to in- tersect obliquely with carina extending from pterygostomian spine at about midlength of cara- pace, associated groove continuing indistinctly posteriad. Abdomen (Fig. 3d,e,f) of both male and female broadly arched dorsally, gradually tapering dis- tally, narrowest part of sixth somite less than 2/3 (0.60) width of first somite; pleura of 3 anterior somites broadly rounded, margin of third slightly serrated, that of fourth somite drawn posterolat- erally to strong spine flanked dorsally by 0-3 more slender and smaller spines and preceded on ventral margin by 0-2 small spines; number, po- sition, and shape of either lateral or ventral spines may be asymmetrical; posterolateral cor- ner of fifth pleura acuminate and flanked dorsally by 1 or 2 spines analogous to those on somite 4; sixth somite with middorsal length about 1.7 that of fifth, broad-based midlateral spine overlapping base of telson, smaller posterolateral spine acute; fourth and fifth somites each with strong, posteri- orly directed spine on sternite. Telson (Fig. 3^) elongate subrectangular, length about 3.0 ante- rior width, 6.8 posterior width, and about 1.4 length of sixth somite, not including posterior spines; armed with 7 dorsolateral spines of nearly uniform size; posterior margin convex, armed with 2 principal spines at each corner and 10 or 11 feathered strong setae on distal margin between. Eyes (Fig. 3a, b) with cornea imperfectly devel- oped, unfaceted though diffusely pigmented. 268 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS Figure 3. — Alvinocaris muricola, holotype d: a cephalothorax and anterior appendages, lateral; b rostrum, eye, antennular peduncle, antennal scale, dorsal of right side; c antennular peduncle, distal articles, mesial; abdomen with variations in spination of pleural margins fof allotype 1 1, d lateral, e margin of pleuron 3 from opposite side, /"(of holotjT)e d ) segments 4 and 5; g tail fan; h median sternal spine between fifth pereopods; i median sternal spine on abdominal segment 5; cheliped, j mesial, k lateral; I, m, n, o pereopods 2, 3, 4, 5. Scales = 1 mm: 1 (a-g,j-o); 2 (h, i). 269 FISHERY BULLETIN: VOL. 86, NO. 2 ovate in outline, though fused to each other mesially, and each with a short upturned spine on anterior surface. Antennular peduncle (Fig. 3a, 6, c) reaching beyond end of antennal scale; basal article 1.1 as long as second and about 2.5 as long as third, all measured on ventral margin; stylocerite well sep- arated from peduncle, tapering to slender elon- gate tip reaching tip of distolateral spine on basal article; latter exceeding distodorsal margin of ar- ticle, fringed by transverse subdistal row of setae, and closely appressed to second article, distome- sial spine much smaller; shorter second article with stronger mesiodistal spine. Dorsolateral flagellum about length of carapace, thickened in basal 2/3, with annulations, except at base, longer in female than in male and much longer than in whiplike distal part; ventromesial flagellum somewhat shorter and with annulations of vari- able but shorter length. Antennal scale (Fig. 3a, 6) about twice as long as wide, distolateral tooth strong, falling slightly short of broadly rounded distal margin of blade; basal article with acute distal spine ventrally; flagellum (broken in material studied) probably slightly exceeding length of body. Mandibles (Fig. 4a) similar, with 2-segmented palp, incisor process broad and armed with 8 mar- ginal teeth, slender molar process simple, diver- gent, its narrowly rounded tip minutely setose. First maxilla (Fig. 46) with proximal endite asymmetrically oval-triangular, distal margin bearing many long setae; distal endite with nar- rowed base but broadened distally, armed with many (about 30) short spines on mesial margin and with scattered longer spinules submarginally and marginally beyond either end of spine row; palp scarcely bifurcated, with long distal spine on obsolescent proximomesial branch and 1 shorter submarginal spine on distal branch. Second maxilla (Fig. 4c, c?) with proximal en- dite represented by 2 similar lobes; distal endite subtriangular, expanded mesiodistally and paral- leled laterally by narrow, somewhat twisted palp; scaphognathite with anterior lobe rectangulo- ovate, fringed with uniformly long, silky setae on anterior and mesial borders, uniformly shorter setae along entire lateral margin; posterior lobe narrow and acuminate, fringed on blunt tip and adjacent mesial margin by strikingly long, strong, tangled setae preceded proximally by shorter setae similar to those on lateral margin. First maxilliped (Fig. 4e, f) with irregularly fusiform endite, short palp concealed and much exceeded in length and size by leaflike exopod, epipod obscurely bilobed; indistinct mesial lobule on exopod possibly representing incipient lash. Second maxilliped (Fig. 4g, h) somewhat pedi- form but flattened, mesial margin of articles bearing long, feathered setae, mesial surface of terminal article densely setose, exopod barely ex- ceeding leaflike epipod. Third maxilliped (Fig. 4i, j) slender, 5- segmented, reaching beyond antennular pedun- cle; terminal article trigonal in cross section, ta- pered distally, bearing 3 spines, transverse tracts of dense setae along mesial surface; similar tract of setae on carpus and another conspicuous group mesiodistally on merus-ischium, latter with stout distolateral spine at articulation with carpus; exopod much reduced, subtriangular, without lash. First pereopods (Figs. 3^, k; If-k) chelate, subequal and sexually dimorphic, at least in fully mature individuals; fingers curved ventrally and slightly laterad; dactyl more slender than and with level of tip slightly shorter than or equal to that of fixed finger; mesial surface of each finger convex, lateral surface concave, with opposed sur- faces uniformly offset; closing without gape, each armed on prehensile edge with row of almost uni- form teeth so closely set as to be almost contigu- ous, acute tip of dactyl slightly spooned by elon- gate teeth slanted distad and curving around its external edge; line of sensory hairs mesial to cut- ting edges. Leg shorter than to almost equaling third maxilliped. Palm of holotype male inflated laterally, but apparently somewhat irregularly concave mesially, length 1.4 greatest height and longer than fingers; palm relatively shorter in allotype female, 0.3 length of fingers. Carpus shorter than palm, with oblique ventral crest end- ing in strong distolateral spine, flanked mesially by patch of setae on polygonal raised area. Merus somewhat swollen in distal half, distinct from is- chium but fused to it, neither armed. Second pereopod (Figs. 3/ ; 7e ) shorter and more slender than first, reaching about to end of anten- nal peduncle; fingers slightly longer than palm, similar in size and shape, opposed edges without gape, each pectinate with single row of teeth in distal half directed obliquely distad and increas- ing slightly in size to end in noticeably stronger tooth crossing opposite member when closed, but spineless proximally; carpus slender, about 1.16 longer than chela; merus and ischium unarmed. Third to fifth pereopods (Fig. 3m, n,o) similar in length and structure, third reaching to tip of or 270 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS 2. Figure 4. — Aluinocaris muricola, holotype 6: a mandible; 6 maxilla 1; maxilla 2, c ventral, d palp dorsal; maxilliped 1, e ventral, /"dorsal; maxilliped 2, g ventral, /i dorsal; maxilliped 3, j ventral, _/ dorsal; ^ endopod of pleopod 1; /appendix masculina, pleopod 2. Scales: I (e, f, g,j , o, k, I) = 1 mm; 2 (c, d) = 1 mm; 3 (/) = 0.2 mm; 4 (a, 6) = 0.2 mm. 271 FISHERY BULLETIN; VOL 86, NO 2 slightly beyond antennal scale. Length articles of these legs in holotype 6 , mm: ischium merus carpus propodus dactyl 3rd 1.76 3.84 2.33 2.34 0.35 4th 1.60 3.52 1.98 3.04 0.06 5th 1.60 3.20 2.43 4.96 0.42 Each short dactyl armed with about 4-6 corneous spines on flexor surface, grading from small prox- imally to longest and strongest distally; carpus of each with distodorsal extension projecting as a stop along proximal part of propodal extensor surface; third leg stronger, at least in merus- ischium, than fourth and fifth, but propodus suc- cessively longer from third to fifth; merus of each with ventral spine at 1/3 and 2/3 length; ischium of third, fourth, and fifth leg with 2 spines in line with those on merus. Pleopods well developed; first pair with en- dopods about 1/2 length of exopods in both sexes, narrowed into distomesial projection in male (Fig. 4k) but evenly tapered in female; appendix mas- culina (Fig. 41) of second pair in male (holotype) armed with 7 slender spines extending beyond level of simple slender appendix interna; en- dopods of third to fifth in male and second to fifth in female with simple slender appendix interna, but that of fifth blunt tipped. Uropod (Fig. 3^) with rami subequal in length, lateral ramus slightly exceeding distal end of tel- son, and with movable spine mesial to smaller distolateral tooth, diaeresis sinuous. Remarks. — Remarks are given in the account for A. stactophila. Etymology. — The name is from the Latin "muria", brine, and "cola", inhabiting, for associ- ation of the species with cold brine seeping from the base of the West Florida Escarpment. Alvinocaris stactophila new species Figures 5, 6, 7 Material .—\]SnM 234291, Holotype d, USNM 234292, Allotype 9, USNM 234293, Paratypes, 5 c?, 2 9; north central Gulf of Mexico about 129 km (80 miles) S of Louisiana, 27°46.94'N, 9r30.34'W, 534 m, Johnson-Sea- Link Dive 1879, 28 September 1986, Bush Hill hydrocarbon seep. From Linda H. Pequegnat and Randall Howard, LGL Ecological Research Asso- ciates, Bryan, TX, supported by partial funding for Minerals Management Service-Northern Gulf of Mexico contract 14-12-0001-30212. Measurements in mm. — Holotype 6, postor- bital carapace length 7.0, rostrum 2.7, maximum carapace height 5.3, total length about 25. Allo- type 9 , same 6.8, 2.0, 5.9, 24. Paratype 6 , same 4.2, 1.9, 3.2, total length not measured; paratype 9, same, 4.9, 2.0, 4.1. Description. — Integument thin, shining, mi- nutely punctate. Rostrum (Fig. 5a, b) almost straight, imperceptibly elevated above horizontal in distal half, sharply pointed tip usually reach- ing proximal level of second peduncular article of antennule, but sometimes to proximal part of third peduncular article; dorsal margin raised into thin serrate crest containing 12-17 teeth, strongest in central sector of row, with about 1/2 length of crest continued onto carapace; ventral margin less prominent and armed with or 1 subterminal tooth; sample tooth formulas 11/1, 12/0, 14/1, 17/1; lateral carina broadened proxi- mally and confluent with orbital margin. Cara- pace (Fig. 5a, b) with buttressed acuminate an- tennal spine distinct; pterygostomian spine acuminate and prominent. Prominent anterior antennal carina curving posteroventrally to in- tersect obliquely with carina extending from pterygostomial spine at about midlength of cara- pace, associated groove continuing indistinctly posteriad. Abdomen (Fig. 5d) of both male and female broadly arched dorsally, gradually tapering dis- tally, narrowest part of sixth somite less than 1/2 (0.44) width of first somite; pleura of 3 anterior somites broadly rounded, that of fourth somite drawn posterolaterally to acuminate spine flanked dorsally by 0-3 much more slender and smaller spines; posterolateral corner of fifth pleuron varying from strongly acuminate to nearly rectangular and flanked dorsally by 2-5 spines analogous to those on somite 4; sixth somite with middorsal length about 1.8 that of fifth, broad-based midlateral spine overlapping base of telson, smaller posterolateral spine acute; fourth and fifth somites each with strong, posteri- orly directed spine on sternite. Telson (Fig. 5e) elongate subrectangular, length about 2.8 ante- rior width, 5.2 posterior width, and about 1.7 length of sixth somite, not including posterior spines; armed with 5-8 dorsolateral spines of 272 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS Figure 5 — Alvinocaris stactophila, holotype 6: a cephalothorax and anterior appendages, lateral; 6 rostrum, eye, antennular peduncle, antennal scale, dorsal of left side; c antennular peduncle, distal articles, mesial; d abdomen, lateral; e tail fan. Allotype 9: cheliped, /"mesial, g lateral; h, i,j, k pereopods 2, 3, 4, 5. Scale = 2 mm. 273 FISHERY BULLETIN: VOL. 86, NO. 2 nearly uniform size, occasionally unequal in number on either side; posterior margin convex, armed with 2 principal spines at each corner and 8-12 feathered strong setae on distal margin be- tween. Eyes (Fig. 5a, 6) with cornea imperfectly devel- oped; unfaceted though diffusely pigmented in adults, but with internal facetlike pattern in smaller individuals; ovate in outline though fused to each other mesially, and each with an up- turned spine on anterodorsal surface. Antennular peduncle (Fig. 5a, b, c) reaching beyond end of antennal scale; basal article 1.3 length of second and about 2.2 length of third, all measured on dorsal margin; stylocerite well sepa- rated from peduncle, tapering to slender elongate tip variably reaching as far as midlength of sec- ond article; basal article with distodorsal margin exceeded by rostral tip though extended into strong lateral spine reaching level equal to that of stylocerite and closely appressed to second article, much smaller distomesial spine slightly diver- gent; shorter second article with stronger mesiodistal spine. Dorsolateral flagellum about twice length of carapace, thickened in basal half; ventromesial flagellum somewhat shorter. Antennal scale (Fig. 5a, b) about twice as long as wide, distolateral tooth strong, falling short of broadly rounded distal margin of blade; basal ar- ticle with small distal spine ventrally; flagellum slightly exceeding length of body. Mandibles (Fig. 6a) similar, with 2-segmented palp, incisor process broad and armed with 7 mar- ginal teeth, slender molar process simple, diver- gent, its narrowly rounded tip minutely setose. First maxilla (Fig. 66) with proximal endite asymmetrically oval-triangular, distal margin bearing about 25 long setae; distal endite with narrowed base but broadened distally, armed with many short spines on mesial margin and with scattered longer spinules submarginally and marginally beyond either end of spine row; palp scarcely bifurcated, with long distal spine on ob- solescent proximomesial branch and shorter adja- cent submarginal spine and tangled setae on dis- tal branch. Second maxilla (Fig. 6c) with proximal endite represented by 2 similar lobes; distal endite sub- triangular, expanded mesiodistally and paral- leled laterally by narrow somewhat twisted palp; scaphognathite with anterior lobe rectangulo- ovate, fringed with uniformly long, silky setae on anterior and mesial borders, uniformly shorter setae along entire lateral margin, posterior lobe narrow and acuminate, fringed on blunt tip and adjacent mesial margin by strikingly long, tan- gled strong setae preceded proximally by shorter setae similar to those on lateral margin. First maxilliped (Fig. 6d, e) with irregularly fusiform endite, short palp concealed and much exceeded in length and size by leaflike exopod, epipod obscurely bilobed; indistinct mesial lobule on exopod possibly representing incipient lash. Second maxilliped (Fig. 6f, g ) somewhat pedi- form but flattened, mesial margin of articles bearing long, feathered setae, mesial surface of terminal article densely setose, exopod barely ex- ceeding leaflike epipod. Third maxilliped (Fig. 6h, i, j) slender, 5- segmented, reaching beyond antennular pedun- cle; terminal article trigonal in cross section, tapered distally, bearing 3 spines, transverse tracts of dense setae along mesial surface; similar tract of setae on carpus and another less conspic- uous group on merus-ischium, latter with stout distolateral spine at articulation with carpus; exopod much reduced, subtriangular, without lash. First pereopods (Figs. 5f, g; 7c, d) chelate, subequal and sexually dimorphic, at least in fully mature individuals; fingers curved ventrally and slightly laterad; dactyl more slender than fixed finger, tips varying slightly in relative length; mesial surface of each finger convex, lateral sur- face concave; prehensile surfaces uniformly off- set, closing without gape, each armed with row of almost uniform teeth so closely set as to be almost contiguous, line of sensory hairs mesial to cutting edges, acute tip of dactyl slightly spooned by elon- gate teeth slanted distad and curving around ex- ternal edge. Leg exceeding third maxilliped by length of fingers in holotype male, but shorter than third maxilliped in other individuals. Palm inflated in holotype male, length 1.4 greatest height and longer than fingers; palm relatively shorter in allotype female and other individuals examined, 0.3 length of fingers. Carpus shorter than palm in holotype but longer than palm in remainder of specimens examined, bearing oblique ventral crest ending in strong distolateral spine and flanked mesially by patch of setae on polygonal raised area; notch above spine smoothly concave and opposing low ridge ending in small rounded spine on heel of palm; shallowly concave anteromesial margin of carpus leading dorsally to 2 low rounded lobes. Merus somewhat swollen in distal half, distinct from ischium but fused to it, neither armed. 274 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS 1. 2. Figure 6. — Alvinocans stactophila, allotype 9: a mandible; b maxilla 1; c maxilla 2; maxilliped 1, d ventral, e dor- sal; maxilliped 2, /"ventral, g dorsal; maxilliped 3, h ventral, ; dorsal, y exopod. Paratype 6: k endopod of pleopod 1; I appendix masculina, pleopod 2. Scales: 1 id-g, k) = \ mm; 2 (a, 6) = 0.5 mm; 3 (/) = 0.3 mm; 4 (c) = 1 mm. 275 FISHERY BULLETIN: VOL 86, NO. 2 Figure 7. — Parts of Alvinocaris chelae viewed by SEM. A. stactophila: fingers of small chela, a mesial, b dorsal; fingers of large chela showing finely toothed opposed edges near tips, c mesial, teeth flush with convex surface, d lateral, teeth marginal on spooned tips, with points rounded on dactyl, acute on fixed finger. A. muricola: e fingers of small chela, lateral; /large chela and distal part of carpus, lateral. Scales: 100 p.m, d; 200 ^.m, a-c; 500 |j.m, e; 1 mm, f. 276 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS Figure 7. — Continued. — Alvinocaris muricola : fingers of large chela showing finely toothed opposed edges neair tips, g teeth flush with convex mesial surface, h spooned lateral surface of same, points rounded on dactyl, acute on fixed finger; close-up lateral view of teeth and associated sensory setae, teeth of fixed finger in foreground and of dactyl in background, i near distal end of fingers, y near midlength of fingers; k sensory seta showing 2 rows of sensillae on concave surface. A. markensis: I fingers of small chela, mesial view of distal part. Scales: 3 jjim, k\ 20 (jim, ;; 30 \i^m,j; 50 p.m, h\ 100 M-m, /; 200 ^.m, g. 277 FISHERY BULLETIN: VOL. 86, NO. 2 Second pereopod (Figs. 5/i; 7a, b) shorter and more slender than first, reaching to between mid- length and end of antennal peduncle; fingers slightly longer than palm, similar in size and shape, opposed edges without gape, each pecti- nate with single row of teeth in distal half di- rected obliquely distad and increasing slightly in size to end in noticeably stronger tooth crossing opposite member when closed, but spineless prox- imally; carpus slender, about 1.2 longer than chela; merus and ischium unarmed. Third to fifth pereopods (Fig. 5i,j,k) similar in length and structure, third reaching beyond an- tennal scale by about 0.3 length of propodus. Length articles of these legs in allotype 9 , mm: 3rd 4th 5th ischio- merus 4.48 4.89 4.16 carpus propodus 2.30 2.18 2.24 3.20 3.39 4.22 dactyl 0.48 0.48 0.48 Each short dactyl armed with about 6 corneous spines on flexor surface, grading from small prox- imally to longest and strongest distally; carpus of each with distodorsal extension projecting as a stop along proximal part of propodal extensor sur- face; third leg stronger, at least in merus- ischium, than fourth and fifth; merus of third and fourth with ventral spine at 1/3 and 2/3 length, distal one strongest, fifth without spines; ischium of third with 2 spines in line with those on merus, that of fourth and fifth spineless. Pleopods well developed, pair 1 with endopods about half length of expods in both sexes, endopod of male (Fig. 6j) with asymmetrical mesial exten- sion, that of female tapering to acute tip; pair 2 with appendix masculina of male (Fig. 6^) bear- ing distal cluster of about 9 strong straight spin- ules extending beyond level of simple slender ap- pendix interna. Uropod (Fig. 5e) with rami subequal in length, slightly exceeding distal end of telson, lateral ramus with movable spine mesial to smaller dis- tolateral tooth, diaeresis sinuous. Etymology . — The name is from the Greek "stactos", oozing out or trickling, and "philos", to love, for association of the species with hydrocar- bons seeping from the substrate. Remarks. — Alvinocaris lusca and the three new species of Alvinocaris described here exhibit minor differences that are highlighted in the key to species given above, but their similarities seem far more significant; i.e., general body appear- ance and strength of integument, shape of ros- trum (although that of A. stactophila sometimes lacks ventral teeth), shape and general armature of tail fan, blindness, and general structure of appendages, including mouthparts. Some minor differences that may be mentioned are features such as number of incisor teeth on the mandible, number of spines on the first maxilla, shape of the second maxilla, lack of meral spines on pereopod 5 in A. stactophila, unequal distribution of spines on ischia of pereopods 3-5 in the three species, and shape of the endopod of male pleopod 1 and appendix masculina (though males of A. marken- sis are not yet known). Each of these species lives in a distinctive ben- thic environment, but all share similarities that suggest dependence on a chemotrophic bacteria- based food chain (Childress et al. 1986). Van Dover et al. (in press) provide evidence from mor- phology, behavioral and gut content analyses of the similar Rimicaris exoculata Williams and Rona that indicates a bacterial diet grazed from surfaces of hydrothermal chimneys, although di- rect observations of bacteria within the digestive tract could not confirm this. The distinctively spoon-shaped chelae of the first pereopods of both Alvinocaris and Rimicaris species, with unbroken comb of exceedingly fine teeth on the prehensile edges, could be an adaptation for scooping or sweeping bacteria toward the mouthparts. Williams and Chace (1982) described the first chelae of A. lusca as convex on the extensor sur- face and concave on the flexor surface, but they also said (p. 142) that the outer surface of the fingers is convex and the inner surface concave. The latter is misleading because in full extension the convex side of the chela is mesial and the concave side lateral. It is not yet known how these appendages are used, but certainly the chelae can be folded compactly against the leg's proximal articles, and in the related Rimicaris exoculata and R. chacei (Williams and Rona 1986) these legs seem very mobile. Sensillae flanking prehen- sile surfaces of the fingers seem well adapted to aid feeding on finely particulate matter. More- over, the species of Rimicaris have exceedingly setose mouthparts. In species of both genera, the second pair of pereopods have much smaller chelae with fingers bearing long sensory setae and spines on the pre- hensile edges that are seemingly adapted for 278 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS grasping. For mobile animals of this morphologi- cal makeup, the most likely feeding methods in the stated environments would seem to be bacte- rial concentration, along with secondary preda- tion and scavenging. ANOMURA: GALATHEIDAE Munidopsis alvisca new species Figure 8 Materm/. —USNM 234294, 9 Holotype, USNM 234301, 9 Paratype, Guaymas Basin, Golfo de Cahfornia, 27°00'N, lir25'W, 2,008 m, Aluin Dive 1616, 8 August 1985, pilots J. Hardi- man and R. Wilkes, observer J. F. Grassle. From J. F. Grassle, Woods Hole Oceanographic Institu- tion, Woods Hole, MA. USNM 234295, 6 Paratype, Explorer Ridge, Magic Mountain, 49°45.6'N, 130°16.16'W, 1,818 m, Pisces IV Dive P-1494, Coll. No. 1877, Gulati Gusher-base, 1 July 1984, pilots-observers. Wit- combe, Johnson, Tunnicliffe. USNM 234296, 9 ovig. Paratype, Explorer Ridge, Upper Magic Mountain, 49°45.5'N, 130°16.1'W, 1,812 m, Pisces IV Dive P-1497, Coll. No. 1873, Lunch Hour Vent, 4 July 1984, pilots-observers. Shep- herd, Juniper, Johnson. USNM 234297, 9 ovig. Paratype, same. Coll. No. 1875, Crab Vent. USNM 234298, 9 ovig. Paratype, same. Coll. No. 1875. USNM 234299, 2 9 ovig. Paratypes, Juan de Fuca Ridge, Limbo Vent ( = 3 m from Holland's Hillock Axial Seamount), 45°55'N, 130°03'W, 1,545 m, Pisces IV Dive P-1732, Coll. No. 1934, 2 August 1986, pilots-observers, K. Shepherd, R. Embley, J. Franklin. From Verena Tunnicliffe, Biology Department, University of Victoria, B.C., Canada. Measurements in mm. — Holotype 9, carapace length including rostrum 23.7, margin of orbit to posterior edge of carapace 18.6, maximum cara- pace width 15.7; Paratype 9 234301, same, 27.9, 20.8, 17.3; Paratype 6 234295, same, 13.8, 10.2, 8.4. Description. — Carapace (Fig. 8a, c) exclusive of rostrum distinctly longer than broad, moder- ately arched transversely; anterior and posterior cervical grooves apparent, depression in anterior part of cardiac region; short moderately devel- oped rugosities on each anterior branchial region, but more distinct and transversely developed rugae on each posterior branchial region, with tendency to being continuous across anterior and central part of cardiac region; posterior margin with median concavity. Rostrum narrowly tri- angular, concave dorsal surface smoothly curv- ing to upturned tip exceeding eyestalks by more than twice their length, distinct carina bearing almost imperceptible scalelike rugae diminish- ing to obsolescence on gastric region. Frontal margin with broad angle lateral to eyestalk followed by concave raised and sparsely orna- mented margin ending in antennal spine followed in turn by almost rectangular but acute antero- lateral angle. Lateral plate obliquely rugose, pro- jecting anteriorly below antennal peduncle, its rather angular tip minutely but bluntly bi- spinose. Abdomen (Fig. 86) unarmed; transverse ridge of segments 2 and 3 smooth, that of segment 4 obsolescent; segments 5 and 6 smooth. Eyes (Fig. 8a , c ) moderate in size; well exposed, smoothly ovate cornea cupped within broad based movable ocular peduncle extended into elongate mesiodorsal spine, directed obliquely upward at low angle and ornamented with tiny irregular obsolescent spinules, and much shorter mesioven- tral spine. Basal article of antennular peduncle with dis- tal margin irregularly crenulate, slender dorso- lateral spine and broader lateral spine flanked by cluster of irregular small spinules, an obsolescent mesiodorsal spine present. Antennal peduncle with fixed basal article extended into stout, flat ventral spine and shorter crenulate lateral spine; succeeding articles short, second bearing stout lateral angle, third unarmed, fourth with scal- loped distal margin, its dorsomesial projection spinelike. Third maxilliped (Fig. 8e ) with ischium shorter than merus, bearing mesial crest armed with finely uniform, evenly spaced corneous teeth. Basis with 2 low spines in line with crest on is- chium. Merus with obsolescent spine at pos- teromesial corner, mesial margin usually with another at level of propodo-carpal joint, followed after an interval by an obscure tubercle, and then by a more prominent spine at base of convex dis- tal margin; stronger spine at anterolateral cor- ner; lateral margin broadly arched. Carpus, propodus, and dactyl folded on merus-ischium and about as long as those two articles together, dense setation on dorsal surface of each, and dis- tally on prehensile surface of propodus and dactyl. Sternite (Fig. 8d) at base of third maxil- 279 liped with convex crenulate anterior margin on mesial lobe, lateral lobe angular. Epipods absent from pereopods. Chelipeds (Fig. SP subequal, ornamented with variably ciliate rugosities tending to arrange- ment in longitudinal tracts; ischium with mesial FISHERY BULLETIN: VOL. 86, NO. 2 fidge bearing subterminal spine and obsolescent irregular subsidiary spines; merus rough, bear- ing 3 mesial spines, 1 distodorsal spine, and a smaller distoventral spine; carpus with 2 spines in dorsolateral row paralleled by less prominent ventrolateral row; palm with spines on prominent Figure 8. — Munidopsis alvisca , holotype 9 : a carapace, eyes and right antenna, dorsal; b abdomen, somites 2-4 in folded position; c part of cephalothorax and anterior appendages, lateral; d stemites at base of third maxilliped and chelipeds; e left third maxilliped, merus and ischium; g left second pereopod. Paratype 9 234301: /"right cheliped. Scales: 1 (a, 6) = 5 mm; 2 (c) = 5 mm; 3 if, g) = 3 mm; 4 (d, e) = 1 mm. 280 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS dorsal ridge, stronger on right than on left; fin- gers longer than palm, spooned at tips, prehensile edges close fitting, entire, but small basal tooth of fixed finger opposed by notch in prehensile edge of dactyl. Walking legs rather long, first walking leg (Fig. 8^) reaching almost to tip of chela, second and third reaching about to base of dactyl on pre- ceding leg; corresponding articles of respective legs approximately equal in length except for meri which decrease posteriorly; each merus with rounded, rugose dorsal crest ending in distal spine; each carpus with longitudinal dorsal and dorsolateral rib ending in more or less well- developed spine, and often with secondary spine(s) on distal margin between them; each propodus slender, bearing small movable spine distolaterally at base of dactyl; each dactyl slen- der, acute corneous tip preceded by row of 12 or more movable spines on prehensile edge. Slender fifth leg with well-developed cleaning brush on more or less flattened dactyl opposed by similar setae on distal end of propodus. Variation . — There is minor variation in orna- mentation of the specimens available for study, but none of it is associated with the disjunct dis- tribution in the Golfo de California and the north- eastern Pacific. Remarks. — The specimens reported here were taken around hydrothermal vent sites discussed by Canadian American Seamount Expedition (1985), ASHES Expedition (1986), and Tunni- cliffe et al. (1985, 1986). Munidopsis has been sighted at three other sites along Juan de Fuca Ridge, but the only specimens collected are those listed above (V. Tunnicliffe^). Comparisons of Munidopsis alvisca with previ- ously described species of the genus are aided by reference to A. Milne Edwards (1880), Milne Ed- wards and Bouvier (1897), Chace (1942), Sivert- sen and Holthuis (1956), and Ambler (1980). Lack of epipods on the pereopods immediately sepa- rates M. alvisca from species such as M. crassa Smith, 1885 and M. subsquamosa Henderson, 1885 which it superficially resembles. Both of the latter species have relatively prominent rugae and spines on the cephalothorax and legs whereas M. alvisca has fairly smooth ornamentation on 2Verena Tunnicliffe, Department of Biology, University of Victoria, P.O. Box 1700, Victor, B.C., Canada V8W 2R2, pers. commun. 1987. these body parts, except for minor development of spines on the lateral carapace margin anteriorly. The rostrum of all of these species is narrowly triangular, curves moderately upward to the tip and bears a middorsal carina, but the carina in M. alvisca bears almost imperceptible scalelike rugae and diminishes to obsolescence on the gas- tric region whereas in both M. crassa and M. sub- squamosa the carina is varyingly rugose, rather strongly so in the former, and maintains this or- namentation onto the gastric region. Moreover, M. crassa bears tiny irregular marginal spines on the rostrum. Spination of the merus of the third maxilliped is far weaker in M. alvisca than in the other two species discussed, and both the anterolateral spine of the ischium and the crenulate margin of the crest on the ischium are less developed than in them. On the other hand, M. alvisca possesses both mesiodorsal and mesioventral eye spines whereas M. subsquamosa and M. crassa lack the mesioventral spine. More distant comparisons seem inappropriate because of different body proportions and orna- mentation, rostral width, length, elevation and spination, and structure of the eye and third max- illiped. The keys for identification by both Chace (1942) and Pequegnat and Pequegnat (1970), for example, though strictly applicable to species of the Atlantic basin, would ally M. alvisca with M. aries (A. Milne Edwards, 1880), a much larger species with broader cephalothorax and rostrum, eyes almost hidden from dorsal view, and with less transverse ornamentation. The revised ver- sion of this key by Pequegnat and Pequegnat (1971) would place M. alvisca in the couplet space occupied by M. sundi Sivertsen and Holthuis, 1956, a species with superficially similar shaped cephalothorax, but densely clothed with short setae. Etymology. — The name is an acronym taken from names of the deep submersible vessels used in collecting the species, Alvin and Pisces IV. BRACHYURA: BYTHOGRAEIDAE Bythograea tnesatlantica new species Figures 9, 10 Materia/.— USNM 234300, Holotype 9, Mid- Atlantic Rift Valley about 70 km south of Kane 281 FISHERY BULLETIN; VOL. 86, NO. 2 Fracture Zone (see: Kong et al. [19851; Leg 106 Shipboard Scientific Party 1 19861; Ocean Drilling Program Leg 106 Scientific Party [1986]), 23°22.09'N, 44°57.12'W, 3,437 m, Aluin Dive 1683, MARK vent, Stn. 1, scoop, 30 May 1986, pilot D. Foster, observers S. Humphris and J. Ed- mond. From NSF-Leg 106-Ocean Drilling Pro- gram, NSF Grant OEC-8311201 to J. F. Grassle, Woods Hole Oceanographic Institution, Woods Hole, MA. Measurements in mm. — Carapace Length 13.8 Width 23.3 Depth of cephalothorax 8.1 Frontoorbital width 7.7 Propodus lower margin R 15.5 L 15.2 Dactyl length P':ilm 8.3 8.3 1 aini Height 7.9 7.8 Thickness 4.9 5.1 Description. — General aspect similar to that of B. thermydron, cancroid, depressed. Carapace (Figs. 9, lOd) broad, transversely elliptical, its rounded lateral angles displaced somewhat ante- riorly; almost flat in middle dorsally, very slightly arched from anterior to posterior and near lateral margins; anterolateral region pro- duced, margin not toothed; surface finely granu- late anteriorly and laterally, smooth but minutely punctate to unaided eye over posterior 2/3 to 3/4; regions indistinct. Frontoorbital width ca 1/3 carapace width. Front almost evenly rounded and somewhat de- flexed, projecting over folded antennules, shallow median depression continued onto protogastric region giving faint suggestion of bilobation; mar- gin beaded with line of fairly uniform granules, closely packed on anterior and anterolateral parts but diminishing almost to obsolescence near or- bits. Arcuate tract of scattered punctations sweeping across anterolateral, hepatic, orbital, protogastric, and metagastric regions. Trans- verse tract of rather prominent granules at rear edge of protogastric region. Carapace with smooth part behind these anterior areas micro- scopically granular and punctate anteriorly, grading posteriorly into almost featureless sur- FlGURE 9. — Bythograea mesatlantica , holotype 9: dorsal. Scale = 3 mm. 282 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS Figure 10. — Bythograea mesatlantica . holotype 9: right chela, a frontal view, b fingers viewed from tips; c left chela, frontal view; d left side of cephalothorax in frontal view showing anterolateral pigmented spot, eye, antennules, antennae, and mouthparts in situ; e mouth field showing third maxilliped turned to side, second maxilliped, first maxilliped with lacinia bearing tiny "portunid lobe" at its mesial comer, partly hidden mandibles, and palate with patch of fine setae to either side of midline; /"abdomen showing somites 3-6 and telson; g oviducal openings and parts of associated stemites. Scales: 1 (a-d, f-g) = 2 mm; 2 (e) = 1 mm. 283 FISHERY BULLETIN; VOL. 86, NO. 2 face. Protogastric, mesogastric, metagastric, and cardiac regions poorly indicated; epibranchial line indicated by small, light colored blotches originating posterior to lateral angle at each side and arching anteromesially over branchial re- gion, then posteromesially toward mesogastric region. Posterior margin concave and paralleled by obsolescent postmarginal groove becoming more pronounced along posterolateral margin. Subhepatic and subbranchial areas orna- mented with small granules, coarsest along upper part of hepatic region but becoming finer and more numerous near base of chelipeds. Orbits sunk into essentially smooth transverse concavity in anterolateral region confluent later- ally at either side with a prominent irregularly oval tan colored spot having very finely granu- late, shallowly concave surface; somewhat in- flated and irregularly granular suborbital area almost fully visible in dorsal view, reaching level of front, tilted anteroventrally from frontal plane lateral to and almost at same level as epistome. Eyestalks projecting anterolaterad, barely movable, depressed and broadened to fit snugly in orbit; unpigmented cornea terminal, subcircular, narrower than eyestalk and anterolaterally ori- ented. Epistome (Fig. lOe) projecting well beyond front, its anterior margin cut into 6 unequal lobes; rather narrow and advanced submedian lobes, separated by narrow deep notch, much broader intermediate lobes and somewhat less broadened lateral lobes less advanced. Antennules folding transversely, stouter than antennae, large bulbous basal articles contigu- ous, concealed beneath front, interantennular septum represented by minute remnant at upper and lower edge of antennular fossa; slender penultimate and terminal articles of peduncle nearly equal in length, former slightly hollowed laterally, latter slightly longer and more slender. Flagella short; mesial 7-segmented ramus slen- der; slightly shorter lateral ramus curved, multi- segmented, thick at base but tapering to point, dense mesial brush of long sensory setae in chord of curve. Antennal insertion mesial to eyestalk; pedun- cle mesial to eyestalk, extending anteriorly or anterolaterally in situ; fixed article broad but short; first free article slender, ca 1.7 length of second article; latter broadened distally; terminal article short, its diameter only slightly greater than that of flagellar base; flagellar length ex- ceeding midline of front. Mouth field (Fig. lOe) divergent anteriorly, sides of its frame broadest posteriorly and some- what swollen and granular at anterolateral cor- ners, maximal inside anterior width about 1.4 minimal inside posterior width. Third maxil- lipeds filling mouth field except for narrow gap of nearly uniform width between ischia of en- dognaths and rather irregular gap anteriorly be- tween meri-carpi of endognaths and epistome; ex- ognaths overlapping sides of mouth frame. Endognaths with exposed surface bearing sparse, sometimes linear, setose punctations; exposed surface of ischium nearly smooth; elongate polyg- onal in outline but primarily rectangular, great- est (distal) width 1.1 narrowed part ca 1/2 length from base; mesial margin straight through most of its length but curved at each end, tooth- less, bearing many stifl" straight setae, submar- ginal zone somewhat thickened and flanked later- ally by shallow longitudinal groove; anterior margin nearly perpendicular to mesial margin except for anteriorly projecting truncate lobe at inner corner; lateral margin concave; posterome- sial margin obliquely convex; basi-ischial suture line visible posterolaterally. Merus slightly narrower than ischium; low granules with tips directed anteromesially along distal margin; ir- regularly quadrate perimeter flanked by submar- ginal thickened zone and groove similar to mesial counterpart on ischium except on straight proxi- mal side, anterolateral angle broadly rounded, anteromesial angle at insertion of palp oblique; mesial margin doubled anteriorly for reception of folded palp, its ventral (exposed) side broadly angled proximal to carpopropodal articulation; posteromesial corner fitted to projecting lobe of ischium, dorsal (hidden) side produced behind carpus, its margin setose. Palp large, dactyl reaching posteriorly about 1/4 length mesial mar- gin of ischium. Carpus expanded distally, nar- rowed proximally, bent nearly at right angle near insertion and obscurely crimped inside angle; dense tuft of setae on distooral surface. Propodus wider than carpus, longer than broad, asymmetri- cally ovate in ventral view; distal (longest) mar- gin convex, densely beset with rows of strong ser- rated setae, longest distally; distal tuft of such setae on dorsal surface. Linear dactyl slightly bent away from midline in distal part and setose as propodus, especially on prehensile edge. Ex- ognath narrow, not extending to full length of merus; ventral surface slightly curved mesially to fit closely against lateral side of endognathal is- chium, with dorsomesial flange (widest distally) 284 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS fitting beneath latter; palp conspicuous, flagel- lum densely beset with setae in hollow of curve. First maxilliped with lacinia of endopod broad, its distal edge 3-lobed and conspicuously though not heavily setose; oblique mesial margin of strongly advanced anterolateral lobe confluent with broader gradually rounded and much less advanced intermediate lobe, latter in turn fol- lowed by still less advanced tiny mesial lobe, sep- arated by a notch and directed anteromesially; tuft of setae preceding notch. Endostome large, divided by low median sagit- tal ridge bifurcated somewhat anteriorly and merging into projecting endostome; each half of palate shallowly concave, crossed by low longitu- dinal ridge slightly offset at its midlength and trending anteromesially from near base of large mandibular palp; ridge flanked laterally by irreg- ular patch of velvety pubescence; smooth lateral 2/3 of palate receiving large efferent branchial channels. Chelipeds (Figs. 9, 10a, b, c) heavy, subequal; integument punctuate on upper and extensor sur- faces, obscure granulation on upper surfaces of palms and on ridges or raised areas elsewhere; chelae inflated, lower margin of palm arched downward, its rather pronounced keel merging into fixed finger; swollen palm with shallow exca- vation proximally for reception of carpus in flexed position, inner surface glabrous but drawn into moderate and slightly granular elevation slightly in front of proximal excavation. Fingers tan col- ored in preservation (70% ethanol) and darkest proximally, color of fixed finger not extending onto palm; fingers not gaping, prehensile edges entire except for obsolescent proximal tooth on fixed finger of each hand; dactyl longer than rela- tively straight fixed finger, arching down distally to close in distal notch of spooned tip of fixed fin- ger. Carpus with extensor surface inflated, right carpus with internal margin rounded, that of left obscurely angled. Merus broadened mesially into cristate flange angled distally for reception of car- pus, strong granules in single line along inner margin, outer surface rounded, strewn with obso- lescent punctations and granules, latter most prominent along distoventral tract. Walking legs rather long, flattened, length de- creasing posteriorly in order 3, 2, 1, 4; each with dense patches of short darkened setae inter- spersed with sparer longer setae on extensor sur- face of carpus and propodus (as well as its lateral side on legs 1 and 2), distoventral corner of car- pus, and more extensively on dactyl; fifth legs somewhat more flattened than others, propodi relatively broader and not densely setose later- ally. Mean maximum length of propodi about twice width. Dactyls slightly longer than propodi, narrowly lanceolate, shallow longitudinal grooves on anterior and posterior surfaces ob- scured by dense setae, tip stout, corneous. Merus of each with upper margin finely granular, ante- rior lower margin present throughout length but posterior lower margin obsolescent proximally. Sternum broadest between legs 1 and 2, nar- rower posteriorly, glabrous beyond outline of ab- domen. Abdomen (Fig. lOf) ovate in outline, fully seg- mented and densely fringed with plumose setae; somite 1 slightly arched dorsally to fit contour of adjacent carapace, somites 2-4 of about equal length, somites 5 and 6 progressively longer; ab- domen with greatest width at 4; telson nearly as broad as somite 6, outline broadly arched distally. Somites 2-5 bearing large, well-developed bi- ramous pleopods, outer curved branch lying near edge of abdomen and heavily beset with short setae laterally and mesially, inner branch more sparsely equipped with ovigerous setae and jointed. Female openings (Fig. 10^^) large, obscurely subtriangular in outline. Color in preservation predominantly off-white except for fingers, matted setal tracts laden with brownish finely particulate matter. Remarks. — Brachyuran crabs that resemble Bythograea were observed and reported by Rona et al. (1986). Bythograea mesatlantica differs in several re- spects from Pacific members of the genus, B. thermydron Williams (1980) and B. microps de Saint Laurent (1984). Among obvious differences from B. thermydron, the new species has even less ornamentation on the carapace; it lacks a distinct suborbital plate separated by a suture, and the suborbital area is inflated, not flat and inclined; there is a transverse concavity lateral to each eyestalk that terminates near the very dis- tinctive brown spot in the cuticle at either side of the carapace; the eyestalk itself is shorter and thicker than in B. thermydron and the shape and position of the cornea differs. The ischium of the third maxilliped is relatively shorter than in B. thermydron and bears only sparse setiferous punctations on the external surface, it lacks tiny granules on the truncate lobe at the anterolateral 285 FISHERY BULLETIN: VOL. 86, NO. 2 corner, the submarginal thickened zone and groove are less distinct; the merus is not tilted dorsally in normal position, and the palp is rela- tively shorter and club shaped rather than curved like a knife edge along the prehensile edge. The lacinia of the first maxilliped is more angular anterolaterally and has a smaller "portunid lobe". The epistome is less lobulate than in B. thermy- dron and the concave palatal area has much less setose covering. The nearly toothless chelae have brown fingers and there is no dense patch of setae on the inner side of the palms. Comparisons with B. microps are necessarily less complete because of the brief description of the latter. The eyes are certainly not slender and retracted in B. mesatlantica ; the chelipeds are not noticeably dimorphic, and they are relatively smooth rather than strongly granular and pilose on the external surface as in B. microps. The distinctive exocular spots on the carapace seem similar to those noted on the chelipeds of Hypsophrys noar Williams (1974, 1976) and Mu- nidopsis lentigo (Williams and Van Dover 1983). Their function is unknown. Etymology. — From the Greek "mesos", middle and "Atlantic", with reference to the Mid- Atlantic Rift habitat. ACKNOWLEDGMENTS Contributors of specimens and crew members of submersible vessels who helped to make the col- lections are owed a special debt of gratitude for securing the rare material described here. Donors are acknowledged individually in each of the spe- cies accounts. I am indebted to Keiko Hiratsuka Moore for rendering the excellent illustrations, to Ruth Gibbons for helping to produce the SEM micrographs, and to F. A. Chace, Jr., R. B. Man- ning, and B. B. Collette for critical review of the manuscript. LITERATURE CITED Ambler, J. W. 1980. Species of Munidopsis (Crustacea, Galatheidae) occurring off Oregon and in adjacent waters. Fish. Bull., U.S. 78:13-34. ASHES Expedition 1986. Pisces submersible exploration of a high- temperature vent field in the caldera of Axial Volcano, Juan de Fuca Ridge. Eos 67(44):1027. Brooks, J M., M C Kennicutt II, C. R Fisher, S. A Macko, K. Cole, J. J. Childress, R. R. Bidigare, and R. D. Vetter. 1987. Deep-sea hydrocarbon seep communities: Evidence for energy and nutritional carbon sources. Science 238:1138-1142. Canadian American Seamount Expedition 1985. Hydrothermal vents on an axis seamount of the Juan de Fuca Ridge. Nature 313:212-214. Chace, F A . Jr 1942. Reports on the scientific results of the Atlantis expeditions to the West Indies, under the joint auspices of the University of Havana and Harvard University. The Anomuran Crustacea. I. Galatheidae. Torreia, Havana, 11:1-106. Childress, J. J , C R Fisher, J M Brooks, M C. Kennicutt II, R Bidigare, and A E Anderson 1986. Methanotrophic marine molluscan (Bivalvia, Mytilidae) symbiosis: mussels fueled by gas. Science 233:1306-1308. DE Saint Laurent, M. 1984. Crustaces Decapodes d'un site hydrothermal actif de la dorsale du Pacifique oriental (13° Nord), en provenance de la campagne frangaise Biocyatherm. C. R. Hebd. Stances Acad. Sci. Ser. 3 Sci. Nat. (Paris) 299(9):355-360, plate 1. KoNG. L., W B. F. Ryan, L. A. Mayer, R. S. Detrick, P J. Fox, and K Manchester. 1985. Bare-rock drill sites, O.D.P. Legs 106 and 109; evidence for hydrothermal activity at 23°N on the Mid-Atlantic Ridge. EOS 66(46):936. Leg 106 Shipboard Scientific Party (R. S. Detrick, J. Honnorez, A. C. Adamson et al.). 1986. Mid-Atlantic bare-rock drilling and hydrothermal vents. Nature 321:14-15. Milne Edwards, A 1880. Reports on the results of dredging, under the supervision of Alexander Agassiz, in the Gulf of Mexico and in the Caribbean Sea, 1877, '78, '79, by the United States Coast Survey Steamer "Blake," .... VIII. — Etudes pr6liminaires sur les Crustac6s. Bull. Mus. Comp. Zool. Harvard 8:1-68, 2 plates. Milne Edwards, A., and E. L. Bouvier. 1897. Reports on the results of dredging, under the supervision of Alexander Agassiz, in the Gulf of Mexico (1877-78), in the Caribbean Sea (1878-79), and along the Atlantic coast of the United States (1880), by the U.S. Coast Survey Steamer "Blake" .... XXXV. Description des Crustac6s de la famille des Galatheides recuellis pendant I'exp^dition. Mem. Mus. Comp. Zool. Harvard 19:1-141. Ocean Drilling Program Leg 106 Scientific Party. 1986. Drilling the Snake Pit hydrothermal sulfide deposit on the Mid-Atlantic Ridge, lat 23°22'N. Geology 14(12):1004-1007. Pequegnat, L H., and W. E. Pequegnat. 1970. Deep-sea anomurans of Superfamily Galatheoidea with descriptions of two new species. In W. E. Pequegnat and F. A. Chace (editors). Contributions on the biology of the Gulf of Mexico. Tex. A&M Univ. Oceanogr. Stud. 1(5):125-170. Pequegnat, W. E , and L H. Pequegnat. 1971. New species and new records of Munidopsis (Decapoda: Galatheidae) from the Gulf of Mexico and Caribbean Sea. Contributions on the biology of the Gulf of Mexico. Tex. A&M Univ. Oceanogr. Stud. 1 (Suppl.):l-24. RoNA, P. A., G. Klinkhammer, T. a. Nelson. J. H. Tefrey, and H Elderfield. 1986. Black smokers, massive sulphides and vent biota at the Mid-Atlantic Ridge. Nature 321:33-37. 286 WILLIAMS: NEW MARINE DECAPOD CRUSTACEANS Saint Laurent, M., de. See De Saint Laurent. SiVERTSEN, E., AND L. B HOLTHUIS 1956. Crustacea Decapoda (The Penaeidea and Stenopodidea excepted). Rep. Sci. Results "Michael Sars" North Atl. Deep-Sea Exped. 1910, 5(12):l-54, plates 1-4. tunnicliffe. v , m botros, m e de burgh, a dinet, h p. Johnson. S. K Juniper, and R E McDuff 1986. Hydrothermal vents of Explorer Ridge, northeast Pacific. Deep-Sea Res. 33:401-412. TUNNICLIFFE, V , S K JUNIPER, AND M E DE BURGH 1985. The hydrothermal vent communities on Axial Seamount, Juan de Fuca Ridge. Bull. Biol. Soc. Wash. 6:453-464. Van Dover, C. L., B Fry. J F Grassle, S Humphris, and P. A. RONA. In press. Feeding biology of the Mid-Atlantic Ridge hydrothermal vent shrimp: Functional morphology, gut content analyses, and stable isotopic composition. Mar. Biol. (Berl.). Williams, A B 1974. A new species of Hypsophrys (Decapoda: Homolidae) from the Straits of Florida, with notes on related crabs. Proc. Biol. Soc. Wash. 87:485-492. 1976. Integumental organs of unknown function on chelipeds of deep-sea crabs, genus Hypsophrys. J. Morphol. 150:889-900. 1980. A new crab family from the vicinity of submarine thermal vents on the Galapagos Rift (Crustacea: Decapoda: Brachyura). Proc. Biol. Soc. Wash. 93:443-472. Williams, A B., and F A. Chace, Jr. 1982. A new caridean shrimp of the family Bresiliidae from thermal vents of the Galapagos Rift. J. Crust. Biol. 2:136-147. Williams, A B , and P A. Rona. 1986. Two new caridean shrimps (Bresiliidae) from a hydrothermal field on the Mid-Atlantic Ridge. J. Crust. Biol. 6:446-462. Williams, A. B , and C. L. Van Dover. 1983. A new species of Munidopsis from submarine thermal vents of the East Pacific Rise at 21°N (Anomura:Galatheidae). Proc. Biol. Soc. Wash. 96:481-488. 287 THE MEGALOPA STAGE OF THE GULF STONE CRAB, MENIPPE ADINA WILLIAMS AND FELDER, 1986, WITH A COMPARISON OF MEGALOPAE IN THE GENUS MENIPPE Joel W. Martin,' Frank M. Truesdale,^ and Darryl L. Felder^ ABSTRACT The laboratory-reared megalopa stage of the Gulf stone crab, Menippe adina , is described and illus- trated and compared with megalopae of three other species of Menippe . The megalopa of M . adina differs from that of Af. nodifrons in having serrate spines on the ventral margin of the dactylus of pereiopod 5 and from that of M. rumphii in having spines on the dactyli of pereiopods 2-5 and a more quadrate carapace. The megalopa of the morphologically similar A/ . mercenaria was also reared in the laboratory, and selected characters are described and compared with the megalopa of A/, adina; megalopae of the two species differ only slightly. Megalopae of M. adina taken from field collections made off South Texas, U.S.A., were compared with and were found to be consistent with laboratory- reared M . adina megalopae. Stone crabs of the genus Menippe are large xan- thid crabs common along the eastern coasts of the United States and Mexico from North Carolina to Yucatan, the Bahamas, Cuba, and Jamaica (Rathbun 1930; Felder 1973; Williams 1984; Williams and Felder 1986). Recently the "common" stone crab, Menippe mercenaria (Say, 1818), was divided into two species: Menippe mer- cenaria (Say) (restricted), known from the east coast of the United States, the Caribbean, and the west coasts of Florida and Yucatan, and Menippe adina Williams and Felder, 1986, known from the northwestern Gulf of Mexico; hybridization of the two species occurs in northwest Florida (see Williams and Felder 1986). These two species (primarily M. mercenaria ) support an important stone crab fishery in the southern United States and Mexico (Williams and Felder 1986) and con- sequently have been the subject of numerous in- vestigations. Despite this interest, the complete larval developments of both commercial species of Menippe remain unknown. For M. mercenaria (Say), Hyman (1925) described a prezoea and first zoeal stage, and Porter (1960) described six zoeal stages reared in the laboratory. Unfortunately, Porter did not describe the megalopa stage, pre- iLife Sciences Division, Natural History Museum of Los An- geles County, 900 Exposition Boulevard, Los Angeles, CA 90007. 2School of Forestry, Wildlife, and Fisheries, and Louisiana Agricultural Experiment Station, Louisiana State University, Baton Rouge, LA 70803. 3Department of Biology and Center for Crustacean Research, University of Southwestern Louisiana, Lafayette, LA 70504. sumably because he considered it a postlarva and not a true larval stage. An unpublished but often- cited report by Kurata"* included descriptions of the zoeal stages of M. mercenaria and a brief sketch of the megalopa; Kurata's description of the megalopa did not include morphology of the pleopods, pereiopods, or mouthparts. Because of recent interest in the phylogenetic significance of the brachyuran megalopa (see Rice 1981a, in press; Martin in press) and postlarval stages (Martin et al. 1984; Felder et al. 1985), and because of the potential importance of stone crab larval biology to aquaculture, it is surprising that the megalopae of M . mercenaria and M . adina remain undescribed. The present paper describes the laboratory-reared megalopa of the Gulf stone crab, Menippe adina Williams and Felder, and compares it with field collections of the same spe- cies from south Texas, laboratory-reared megalo- pae of M. mercenaria, and all previously de- scribed megalopae of the genus Menippe: Menippe mercenaria (Say, 1818) (as described by Kurata fn. 4); Menippe nodifrons Stimpson, 1859 (as de- scribed by Scotto 1979); and Menippe rumphii (Fabricious, 1798) (as described by Kakati 1977). MATERIALS AND METHODS A large ovigerous M . adina was collected from Manuscript accepted February 1988. FISHERY BULLETIN: VOL. 86, NO. 2, 1988. 4Kurata, H. 1970. Studies on the life histories of decapod Crustacea of Georgia. Part IIL Larvae of decapod Crustacea of Georgia. Unpubl. rep., 274 p. University of Georgia Marine Institute, Sapelo Island, GA. 289 FISHERY BULLETIN: VOL. 86, NO. 2 shallow waters of the northern Gulf of Mexico near Grande Terre, LA, in May 1982 and held in a small aquarium at room temperature. After the eggs hatched, the zoeal larvae were given fresh seawater and newly hatched Artemia nauplii daily. Exuviae as well as dead and some living megalopae were preserved in 70% ethanol. Draw- ings were made with the aid of a Wild^ M-5 stereoscope and a Wild M-11 compound stereo- scope, both with camera lucida; accuracy was ver- ified with a Nikon Optiphot. Measurements were made with an ocular micrometer. Ten laboratory- reared megalopae were examined, measured, dis- sected, and compared with megalopae from field collections made in 1973 off south Texas. Com- parisons with M. mercenaria are based on laboratory-reared M . mercenaria megalopae from two females collected on 13 August 1987 from the Indian River system, north of Ft. Pierce, FL. Eggs of these two females hatched on 21 August 1987, and the megalopa stage was first reached after 17 days in mass culture aquaria (30%o salinity, 25°C, 12h:12h light/dark regime). Descriptions of setation for all appendages proceed from proximal to distal. Specimens examined under the scan- ning electron microscope (SEM) were prepared according to procedures outlined by Felgenhauer (1987) but without postfixation in osmium tetrox- ide and with 100% ethanol, rather than amyl ac- etate, as the transitional fluid. Sibling megalopae and field collections have been deposited in the U.S. National Museum of Natural History, catalogue No. USNM 229962 (laboratory- reared M. adina), USNM 229961 (field-collected M. adina), and USNM 229963 (laboratory-reared M. mercenaria). RESULTS Carapace (Figs., lA, B, C, 3 A).— Length 1.67 mm, width 1.45 mm (A'^ = 10). Subquadrate, with 2 lateral prominences on each side; dorsoven- trally thick, with minute tubercle centrally lo- cated. Posterior border fringed with numerous short setae; lateral margin with few scattered setae. Rostrum ventrally deflexed, nearly verti- cal, with deep medial depression, rounded anteri- orly. Angular interorbital prominences extend ventrally between orbit and antennule. Chroma- tophores variable in placement, but almost al- ways found in areas indicated in Figure IB. Eyes (Figs. lA, B, C, 3A). — Large, exposed; eye- stalks sometimes with 2 or 3 short, simple ante- rior setae, always with posterodorsal chroma- tophore. Abdomen (Fig. lA, B). — Subequal in length to carapace. All pleura with rounded posterolateral angles. All somites with sparse setae dorsally; somites 2-5 always with elongated chroma- tophores. Telson (Fig. IG). — Broadly rounded with vari- able setation, occasionally with pair of small pos- terior spines (as in Figure IB). Antennule (Fig. IK). — Biramous; peduncle 3- segmented, with variable setation. Basal seg- ment of peduncle large, bulbous, always with large chromatophore; middle segment subcylin- drical with 0-2 distal setae; distal segment ovoid with scattered short setae. Lower ramus 1-segmented with 6-8 setae; upper ramus 5-segmented with aesthetascs arranged in tiers, usually 0, 7, 8, 6, 4 subterminal plus 3 terminal, with short setae sometimes present on segments 2 and 4 (note: all aesthetascs not illustrated). Antenna (Fig. IJ). — Flagellum 12-segmented (sometimes 11), with 3 peduncular articles and 8 or 9 flagellar articles (see Rice, in press, for cor- rect number of antennal segments in megalopae); setation variable, usually 2, 3, 2, 0, 0, 2, 4, 0, 4 or 5, 1, 4, 4. Mandibles (Fig. 2F). — Asymmetrical, with broadly rounded spade-shaped cutting edges; palp 2-segmented with setation 0, 11-14. Maxillule (Fig. 2E). — Protopodite with 1 or 2 long plumose setae on posterodorsal margin; en- dopodite 2-segmented with setation 1, 2 subtermi- nal plus 2 terminal; basal endite with 29-35 spines and setae; coxal endite with 13-16 spines and setae. Maxilla (Fig. 2D).— Scaphognathite with 70-78 fringing setae and 0—6 setae on blade; endopodite unsegmented with or 1 distolateral seta and 4 or 5 basal plumose setae; basal endite bilobed with setation variable, usually 8-10, 9-11; coxal en- dite bilobed with setation usually 7, 9 or 10. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Maxilliped 1 (Fig. 20). — Exopodite 2-seg- mented, with setation 2 or 3, 5-7. Endopodite 290 MARTIN ET AL.: MEGALOPAE OF STONE CRABS iMENIPPE) A-C D-K- Figure l. — Megalopa of the Gulf stone crab, Menippe adina. A, entire animal, lateral view; B, same, dorsal view; C, frontal view of rostrum and eyes; D, pleopod 1; E, pleopod 4; F, pleopod 5; G, telson and posterior part of sixth abdominal segment; H, dactylus of pereiopod 3; I, dactylus of pereiopod 5; J, antenna; K, antennule. Both scale bars = 1.0 mm. 291 FISHERY BULLETIN: VOL. 86. NO. 2 Figure 2. — Megalopa oi Menippe adina, mouthparts. A, third maxilliped; B, second maxilliped; C, first maxil- liped; D, maxilla; K,- maxillule; F, mandible. Scale bar = 0.5 mm. 292 MARTIN ET AL.: MEGALOPAE OF STONE CRABS (MENIPPE) unsegmented with 6-8 setae arranged as shown. Basal endite setation 28-33; coxal endite setation 15-17. Epipodite with 22 or 23 long, minutely plumose setae, appearing simple under low mag- nification. Maxilliped 2 (Fig. 2B). — Exopodite 2-segmented, with setation 3, 5-8. Endopodite 4-segmented, with setation usually 5, 2 or 3, 5 or 6, 9 or 10; distal segment with 4 or 5 stout serrate setae. Epipodite with 9 or 10 long minutely plumose setae. Maxilliped 3 (Fig. 2A). — Exopodite 2-seg- mented, with setation or 1, 6-8. Endopodite 5-segmented, with variable setation, usually 18- 20, 15 or 16, 5-9, 6-8, 7-10; ischium with scal- loped medial border. Epipodite with 18 long minutely plumose setae on distal two-thirds plus 8-12 plumose setae on proximal one-third. Pro- topodite setation variable. Pereiopods (Figs. lA, B, H, I, 3B, C, D).— Che- lipeds long, stout, subequal; dactylus with 4 irreg- ular teeth; immovable finger with 3 teeth (Fig. 3B); tips of fingers overlap distally when approxi- mated. No recurved hook on basi-ischium (Fig. 3B). Second to fourth pereiopods similar; dactylus with 5 (rarely 4) serrate spines ventrally (e.g., Figs. IH, 3C, D); propodus with long ven- trodistal spine (Fig. IH). Fifth pereiopod dactylus (Fig. II) with 3 long pectinate setae, 1 markedly toothed and concave (Fig. 3E), on distal ventral border and 3 or 4 serrate spines ventrally. Pleopods (Fig. ID, E, F). — Decreasing in size posteriorly. Pleopod 1 (Fig. ID) with 19-22 plumose setae; endopodite with 3 or 4 hooked setae. Pleopod 4 (Fig. IE) with 19-21 plumose setae; endopodite with 3 or 4 hooked setae (Fig. 3F). Pleopod 5 (uropod) (Fig. IF) with 12-14 plumose setae; basal segment lacking setae or with 1 or 2 setae (field collections); endopodite absent. Color. — Overall coloration rose-orange, with dark blue-black chromatophores located as shown in Figure lA, B. DISCUSSION The genus Menippe de Haan, 1833, presently contains about 8 species, only 3 of which occur in North America. The megalopa stage is now known for 3 species in the genus: M. rumphii (Fabricious, 1798), M. nodifrons Stimpson, 1859, and M. adina Williams and Felder, 1986. In addition, selected characters of M. mercenaria (Say, 1818) are pre- sented here for comparison; some characters of that species are also obtainable from an unpub- lished report by Kurata (fn. 4) (see Table 1). Laboratory-reared megalopae of M. adina were virtually identical to megalopae presumed to be- long to M. adina that were collected off south Texas. Even meristic counts of the mouthpart se- tation agreed exactly, with the only observed dif- ferences being that field-collected megalopae were slightly larger and occasionally bore 1 or 2 setae on the basal segment of the uropod. Thus, we feel that our laboratory conditions have not adversely affected development or introduced ab- normal characters, and we have used these field collections for the SEM figures of M. adina mega- lopae (Fig. 3). We expected to find that characters of the megalopa of M. adina are similar to those de- scribed by Kurata (fn. 4) for the morphologically similar (in adulthood) M. mercenaria, a species known to hybridize with M. adina (see Williams and Felder 1986). In general this is true. How- ever, some characters reported by Kurata differ from our observations on M. adina and from our laboratory-reared megalopae of M. mercenaria (Fig. 4). Kurata mentioned (but did not illustrate) Table 1 . — Comparison of characters in megalopae of the genus Menippe. Dash ( — ) indicates information not available from reference. Size! (mm) CL CW Setation Spinafion Setation pleopod 5 Palp of mandible Epipod of Dactylus of Menippe Maxilliped 1 Maxilliped 2 Maxilliped 3 Pereiopods 2-4 Pereiopod 5 Reference adina mercenaria mercenaria nodifrons rumphii 1.67 1.45 1.70 1.55 1.7-8 — 1.50 1.31 1.60 1.55 0, 11-14 0, 11-13 0, 10-13 0,0,9 22-23 20-23 12-20, 226 22 9-10 7-10 up to 10 8 18 18-20 18 18 5 4 4-5 5 4 4 12-13 11-13 11-12 11 12 Present study Present study Kurata3 Scotto 1979 Kakati 1977 ^CL = carapace length; CW = carapace width. 2From a megalops hatched from a stage 6 (rather than the typical stage 5) zoea. 3See text footnote 4. 293 FISHERY BULLETIN: VOL. 86, NO. 2 Figure 3. — Scanning electron micrographs (SEM) of selected characters of Menippe megalopae (presumably M. adina ) collected in south Texas. A, dorsal view of carapace (x 25); B, ventral view of chelipeds showing dentition of the fingers and lack of recurved hook on ischium (x 37); C, dactyli of second (upper figure) and third pereiopods (x 230); D, higher magnification of ventral dactylar spine indicated by arrow in C (x 1,900); E, endopod of third abdominal pleopod showing 4 dentate hooklike setae (x 2,200); F, serrate setae (only 2 of 3 shown) of dactylus of pereiopod 5 (x 2,300). 294 MARTIN ET AL.: MEGALOPAE OF STONE CRABS iMENIPPE) Figure 4. — Scanning electron micrographs (SEM) of selected characters of the laboratory-reared megalopa of Menippe merce- naria. A, dorsal view of carapace (x 37); B, ventral view of chelipeds and third maxillipeds (x 55); C, dactylus of second pereiopod showing ventral serrate spines (x 270); D, ventral spines on dactylus of fourth pereiopod with fewer spinules than anterior pereiopod spines (x 2,000); E, dactylus of fifth pereiopod showing four serrate "sensory" setae with one (arrow) more obviously serrate (x 190); F, higher magnification of concave serrate setae indicated by eutow in E (x 1,500). 295 FISHERY BULLETIN: VOL. 86, NO. 2 "about 9 small spines" on the ischium of the third maxilliped; we found no spines on M. adina or M. mercenaria (Figs. 3B, 4B), but it is possible that Kurata was referring to the acute borders of the scalloped medial margin (our Figure 2A), in which case the 2 species are similar. Spination of the ischium of pereiopods 1-3 differs also; Kurata described 5 or 6, 2 or 3, and 1 small spine on the ischia of pereiopods 1, 2, and 3, respectively, whereas we did not notice this condition in M. adina or M. mercenaria (see Figures 3B, 4B). Fi- nally, Kurata (fn. 4: pi. 74, fig. E) illustrated no spines on the ventral surface of the fifth pereiopod dactylus of M. mercenaria; these spines are obvi- ous on both species (Figs. II, 4E). We found few differences between megalopae of M. adina and M. mercenaria. General morphology of the cara- pace and chelipeds, spination of the dactylus of the pereiopods, and setation of the pleopods agreed almost exactly (compare Figures 3 and 4). Ventral dactylar spines on the posterior walking legs of M. mercenaria were not so serrate as in M. adina and were sometimes armed with only 2 or 3 large spinules rather than the numerous spin- ules seen in M. adina (e.g., Fig. 3D) and in the more anterior legs of M. mercenaria (see Figure 3C, D, 4C). Also, in all but 1 of the 9 megalopae of M. mercenaria examined there were 4 (rather than 3) long serrate setae on the dactylus of the fifth pereiopod (Fig. 4E). As in M. adina, one of these setae was more serrate and concave than were the other long setae (Fig. 4E, F). However, we have not examined mouthpart morphology of M. mercenaria in the detail in which we described M. adina, and so it is possible that additional characters will be found to separate these 2 species at the megalopa stage. The megalopa of M. adina is very similar to that of M. nodifrons as described by Scotto (1979). Although the 2 species differ in setation of some of the mouthparts, this setation may differ from side to side in a given individual. The salient character that serves to separate megalopae of these 2 species is the presence in M. adina of 4 stout serrate spines on the dactylus of pereiopod 5. Scotto (1979) figured only setae (and no spines) on the dactylus of the fifth pereiopod in M. nod- ifrons and the dactylar spines on other pereiopods apparently are not serrate (Scotto 1979, fig. 9c, pereiopod 3). The megalopa of M. rumphii described by Kakati (1977) differs from that of M. nodifrons, M. mercenaria, and M. adina in having a more ovoid carapace with the rostrum only slightly de- flexed. Kakati did not describe the dactyli of pereiopods 2-5 for M. rumphii, but his figure of pereiopod 2 (1977:639, fig. 2, p. 50) does not show stout ventral spines on the dactylus. Possibly Kakati overlooked these spines; if not, the ab- sence of these spines on pereiopod 2 would further serve to separate the megalopa of M. rumphii from those of A/, nodifrons, M. mercenaria, and M. adina. All 4 species have been described as hav- ing a rose-orange coloration in life. Although Rathbun (1930) and Monod (1956) synonymized M. rumphii with M. nodifrons, de- scriptions of the zoeal stages of M. rumphii and M. nodifrons by Kakati (1977) and Scotto (1979), re- spectively, show that larvae of the 2 species differ considerably. In the first zoeal stage, M. rumphii exhibits elongated posterolateral processes on ab- dominal segment 5 that extend posteriorly to more than half the length of the telsonal furcae, which lack spines. The first zoea of M. nodifrons has similar posterolateral processes but these do not extend posteriorly beyond the fork of the tel- son; the telsonal furcae bear 1 dorsal and 2 lateral spines each. These differences are not apparent in later zoeal stages, but their presence in the first zoeal stage and the differences noted in the mega- lopa stage may be reason to question the syn- onymy of these 2 species. Xanthid larvae are known to be variable, and it is often difficult to reconcile larval and adult groupings based on morphology. Larvae of some morphologically disparate (as adult) species are very similar, whereas zoeal stages for species in some genera differ markedly in their morphology (see Martin 1984; Martin et al. 1984, 1985; Mar- tin and Abele 1986). Because of the known mor- phological variability of xanthid larvae, charac- ters presented for taxonomic purposes here and elsewhere (e.g., Martin 1984) must be used with caution. It is not our intent to promote descriptions of single stages in the life cycles of brachyuran crabs. However, in those cases where a descrip- tion of a single stage adds appreciably to our knowledge of phylogeny (e.g.. Rice 1981b) or fills a gap in the larval biology of a commercially im- portant species complex (present study), we feel such a description is justified. A detailed compari- son of zoeal stages of the two species is planned for the near future. ACKNOWLEDGMENTS We are grateful to D. H. Wilber and A. B. This- 296 MARTIN ET AL.: MEGALOPAE OF STONE CRABS (MENIPPE) tie for constructive criticism of the manuscript, to B. E. Felgenhauer for help in preparing the fig- ures, to N. N. Rabalais and B. Cole for assistance in rearing larvae of M. mercenaria, and to L. G. Abele for providing space and facilities for the research. We thank S. Silvers and K. Riddle of the Florida State University Electron Microscopy Center for their expert assistance. We also thank C. Dugas, Louisiana Department of Wildlife and Fisheries, for collecting the ovigerous specimen of M. adina. This work was supported in part by the National Science Foundation grants No. BSR- 8414347 and BSR-8615018 to J. W. Martin and L. G. Abele, and by a research grant to D. L. Felder, N. N. Rabalais, F. M. Truesdale, and D. W. Foltz from the Louisiana Education Quality Support Fund under grant No. 86-LUM(2)-084- 09. LITERATURE CITED Felder, D L 1973. An annotated key to crabs and lobsters (Decapoda, Reptantia) from coastal waters of the northwestern Gulf of Mexico. La. State Univ. Cent. Wetland Resour., Publ. LSU-SG-73-02. Baton Rouge, LA. Felder, D. L., J. W. Martin, and J. W. Gov. 1985. Patterns in early postlarval development of deca- pods. In A. M. Wenner (editor), Larval growth, p. 163- 225. Crustacean Issues, vol. 2. Balkema Press, Rotter- dam. Felgenhauer, B E. 1987. Techniques for preparing crustaceans for scanning electron microscopy. J. Crustacean Biol. 7:71-76. Hyman. O W 1925. Studies on the larvae of crabs of the family Xanthi- dae. Proc. U.S. Natl. Mus. 67:1-22. Kakati, V S. 1977. Larval development of the crab, Menippe rumphii (Fabricious), as observed in the laboratory (Crustacea, Brachyura). Proc. Symp. Warm Water Zooplankton, UNESCO/NIO Spec. Publ., p. 634-641. Martin. J W 1984. Notes and bibliography on the larvae of xanthid crabs, with a key to the known xanthid zoeas of the west- em Atlantic and Gulf of Mexico. Bull. Mar. Sci. 34:220- 239. In press. Phylogenetic significance of the brachyuran megalopa: evidence from the Xanthidae. In A. A. Fin- cham and P. S. Rainbow (editors), Aspects of decapod crustacean biology. Symp. Zool. Soc. Lond., Vol. 59. Martin. J W.. and L G Abele 1986. Notes on male pieopod morphology in the brachyuran crab family Panopeidae Ortmann, 1893, sensu Guinot (1978) (Decapoda). Crustaceana 50:182- 198. Martin, J W., D L. Felder, and F M Truesdale. 1984. A comparative study of morphology and ontogeny in juvenile stages of four western Atlantic xanthoid crabs (Crustacea: Decapoda: Brachyura). Philos. Trans. R. Soc. Lond., Ser. B, 303:537-604. Martin, J. W., F. M. Truesdale. and D. L. Felder. 1985. Larval development of Panopeus bermudensis Rathbun, 1891 (Brachyura, Xanthidae) with notes on zoeal characters in xanthid crabs. J. Crustacean Biol. 5:84-105. MONOD, Th 1956. Hippidea et Brachyura ouest-africains. M6m. Inst. Fr. Afr. Noire 45:1-674. Porter, H. J 1960. Zoeal stages of the stone crab, Menippe mercenaria Say. Chesapeake Sci. 1:168-177. Rathbun, M. J. 1930. The cancroid crabs of America of the families Eu- ryalidae, Portunidae, Atelecyclidae, Cancridae, and Xan- thidae. U.S. Natl. Mus. Bull. 152:1-609. Rice, A. L. 1981a. The megalopa stage in brachyuran crabs. The Podotremata Guinot. J. Nat. Hist. 15:1003-1011. 1981b. The zoea of Acanthodromia erinacea A. Milne Ed- wards; the first description of a dynomenid larva (Deca- poda, Dromioidea). J. Crustacean Biol. 1:174-176. In press. The megalopa stage in majid crabs, with a re- view of spider crab relationships based on larval charac- ters. In A. A. Fincham and P. S. Rainbow (editors). As- pects of decapod crustacean biology. Symp. Zool. Soc. Lond., Vol. 59. Scotto. L. E. 1979. Larval development of the Cuban stone crab, Menippe nodifrons (Brachyura, Xanthidae) under labora- tory conditions with notes on the status of the family Menippidae. Fish. Bull., U.S. 77:359-386. Williams, A. B 1984. Shrimps, lobsters, and crabs of the Atlantic coast of the eEistem United States, Maine to Florida. Smithson. Inst. Press, 550 p. Williams, A. B., and D. L. Felder