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 0 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>\ij 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 0 1 .0 o ■o c o to I 0 - - 2 0 — — 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 - 0 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 0 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 0 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- 0 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 iTi I I I I I I I I I I I I I I I I I I iMi 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 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 0 166 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) • 0 • 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 0 16 0 0 0 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 0 ?? 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) 0 1701 (0,0711) 00809 (00190) 28 1 174 Feb. 1980 LA 324 0954 2.131 (0 138) 0 1592 (00105) 00710 (0 0031) 302 12.7 Winter 1980-81 Dec. 1980 LA 191 0931 2888 (0.188) 0 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) 0 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) 0 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) 0 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; = 0 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; = 0 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 0 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 0 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 0 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 0.5- -1- 1 1 ; 1 \ 1 1 1 ^ 1 0.5- 0 -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 -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- 0 5 0 -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 -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 0 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 0 17 17 Dasyatis sabina 6 7 13 Elops saurus 4 4 8 Orthopristis chrysoptera 1 7 8 Bagre marinus 1 4 5 Lagodon rhomboides 0 5 5 Paralichthys lethostigma 1 3 4 Rhinoptera bonasus 0 4 4 Chilomycterus schoepfi 0 3 3 Ophichthus gomesi 3 0 3 Cynoscion arenanus 0 2 2 Negapnon brevirostris 0 1 1 Trachinotus carolinus 0 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) 0 CO in 0° (£) CM CO •^ ■^ '- 0 0 0^ T- 0 0 CO cp CM cp CD 0 o^ d 0 t^ 0 CO -* 1^ 0 0 d in 0 in d 0 0 CD 0 CM 0 0 CO 0 d 0 0 d rg ~ ? in — in in 00 , — T— t^ ° 0 CO to CM 0 CD 0 r^® CO 0 in cp CD 0 0 cp r^ 0 0 ^ 0 CO 0 ,_ 0 tT '^co d c\j d ■t d 0 0 in 0 5 0 0 in 0 d d d d r? in CO CM £" LO CJ) CO ■* t^ CO "" CO OM cr> CO CO 0 t-" r- 0 (D -"-cfi d in d in d CM d 0 TJ- 0 CO 0 CD 0 d d d d ? en CD ^° 00 CO in "" CO 0 0 coP ■ 0 in 0 0 r-~ 0 d d d d CO ST CM G^ CM in ■* CD 0 r~- 0)0 1^ CM r- 0 in 0 a)P CO 0 0 cp '^ d 0 d •.- 0 °in CM in 0 .,^ 0 yL Pco 0 eg d CO d CO d 0 0 CD 0 0 0 CM d d d d d <5" CD CO — in (D CO in ■^ ■^ 0)° (D in O) CM .^ 0 OlP '^ 0 CO 0 0 CD CO cp CO 0 °co '«■ CM r^ 0 CM ^O) 0 CO d , — d co d CM 0 CO 0 ■ 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 0 + 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; 0 250 100 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 0 o . - o " o o 0 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^ *^^\ -^ 0 o ■ o 0 o ° O O . o O 0 o o o o o 0 o FEB o 1 1 o o ■ o o 0 o o o o o o 0 o o O 0 o o o o o . o o 0 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 = 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 0 ^ t F c 0 OJ 0 Q) (/> ?. c 11 0 ^ 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 0 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 0 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 < 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 0 , ^ -l_. 1 1 F t r c ) 1 0 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 >- 0 < 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 0 0 0 2 2 6 15 37 6 8 0 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 0 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 0 0 0 2.4 0 1.3 0 9.5 6.6 1.3 0 16.6 7.7 0 1.9 0 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 0 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^ri^iM^rT^f rr SEPTEMBER A ¥.« ^ ¥^ iWiT'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 0 9 8 7 6 5 4 3 2 I 0 - - 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 0 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 0 111 0 0 201-250 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 0 125 0 0 201-250 21 76.2 0 54 0 0 251-300 22 63.6 0 109 0 0 301-350 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 0 18 0 0 201-250 23 47.8 0 25 0 0 251-300 23 56.5 0 71 0 0 301-350 39 61.5 4 77 8 10.4 351-420 2 0 1 3 2 66.7 Total 103 58.3 5 194 10 5.2 Sept. 1984 <200 6 33.3 0 15 0 0 201-250 38 47.4 0 69 0 0 251-300 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) 0 0 0 0 0 0 British 2 1,086 24 2 0 0 0 Columbia (2) (1,735) (90) (8) Puget Sound 0 201 1,352 9 1 2 0 (729) (5,262) (40) (2) (9) Washington 1 71 7 125 42 2 0 Coast (5) (316) (42) (451) (151) (77) Columbia 2 24 4 39 67 31 0 River (133) (14) (162) (310) (164) Oregon Coast 0 18 3 21 62 308 10 (107) (13) (83) (213) (1,137) (45) California 0 0 0 0 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 0 98 Male 8 26 23 0 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 0 0 27-30 August 0 1 30 August-02 September 0 14 02-05 September 4 8 05-08 September 6 9 08-11 September 0 2 11-14 September 3 16 14-17 September 0 8 17-20 September 0 0 20 September-20 October 0 0 205 20 15 10 5 0 -5 20 15 10 P 5 uj 0 cc < £ 20 Q. 15 10 5 0 -5 20 15 10 5 0 -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. 0 22.0 — 0 17.0 — 23 Sept.- 11 Oct. 0 55.5 — 0 45.5 — 11 Oct.-18 Oct. 0 9.5 — 0 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 0 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 0 4 0 Calanoida 12 2 57 7 5 4 71 5 0 0 4 0 Harpacticoida 1 0 29 0 11 9 90 12 17 1 18 3 Cladocera 1 0 43 1 14 5 77 10 0 0 4 0 Cumacea 0 0 0 0 2 9 32 3 16 1 41 6 Euphausiacea 11 2 71 8 0 2 42 1 0 0 0 0 Decapod larvae Shhmp 9 5 71 8 0 1 13 0 0 0 0 0 Crabs 8 2 60 4 3 3 60 2 2 0 12 0 Fish eggs and larvae 8 86 57 44 0 21 13 2 1 6 4 0 Fish 0 0 0 0 0 0 0 0 22 62 48 38 Insects! 1 0 29 0 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 0 0 7 0 Calanoida 47 52 100 58 5 5 65 5 22 0 43 9 Harpacticoida 3 5 25 1 4 4 65 4 0 0 7 0 Cladocera 13 7 75 9 19 8 81 17 1 0 23 0 Decapod larvae Shhmp 1 1 75 1 0 1 20 0 0 0 0 0 Crabs 4 11 100 5 7 12 100 8 19 0 21 4 Larvacea, Oikopleura 9 5 100 8 20 13 85 23 7 0 14 1 Fish eggs and larvae 6 5 100 6 11 12 35 6 0 0 0 0 Fish juveniles 0 0 0 0 0 0 0 0 10 94 71 70 Insects^ 0 1 50 0 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 0 24 5 Hypehidea — — — — 12 30 88 45 4 0 29 2 Decapod larvae Shrimp — — — — 61 40 14 18 1 0 6 0 Crabs — — — — 3 10 86 9 55 5 53 25 Fish eggs and larvae — — — — 0 3 57 2 0 0 0 0 Fish juveniles — — — — 0 0 0 0 9 94 50 65 Insects^ — — — — 6 4 100 14 5 0 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 0 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 0 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 0 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) 0 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. 0 -- 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. 0 - 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) 0 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, 0 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) 0 0 0 0 (3) (13) T (7) (9) (5) (7) 3) (7) (7) (10) (6) 0 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) 0 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 0 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 0 - 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 0 60 50 - 40 - 30 - 20 - 10 - 0 150 180 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%) 0 2 - 1 - CM CM C 0) c o o. E o o 0 - 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 0 2 4 20 4 24 6 18 6 4 0 4 22 30 18 24 26 12 44 14 12 44 0 6 0 0 4 12 18 48 2 24 22 54 2 0 12 0 2 26 0 0 4 10 2 26 46 12 14 8 4 24 0 0 4 0 0 0 42 6 18 0 0 26 0 6 12 0 0 0 255 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 0 J 3 0 ^ 13 -1 G. lineatus GUTS WITH CONTENTS ■ GUTS EMPTY 0 (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 0 ^ 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 0 4 4 0 10 0 0 20 100 8 14 0 0 66 0 14 22 10 50 24 18 0 0 8 40 0 34 0 25 0 2 12 16 46 0 0 30 0 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 0 0 0 12 14 0 0 60 100 0 4 0 12 24 0 46 0 102 50 6 0 34 0 4 16 0 0 42 25 0 26 0 0 0 6 0 2 0 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 0 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 0 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 0 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 0 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 0 4 4 0 0 0 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. 1986. Analysis of stone crabs: Menippe mercenaria (Say), restricted, and a previously unrecognized species de- scribed (Decapoda: Xanthidae). Proc. Biol. Soc. Wash. 99:517-543. 297 OCEANOGRAPHIC ASSOCIATIONS OF NEUSTONIC LARVAL AND JUVENILE FISHES AND DUNGENESS CRAB MEGALOPAE OFF OREGON Jonathan M. Shenker' ABSTRACT The larval and juvenile fishes and crabs inhabiting the neustonic zone within 50 km of the coast were sampled biweekly from April through July 1984, with a Manta net (mouth 1.0 m wide x 0.7 m deep), and a large neuston trawl (mouth 3.5 m wide x 1.0 m deep). The Manta net was an efficient sampler for larval fishes and crabs, while the neuston trawl collected larger juvenile fishes that had rarely been observed in previous studies. Nocturnal sampling accounted for nearly all the ichthyoneuston and zooplankton taken. Dunge- ness crab megalopae were the most abundant species. Although present throughout the survey, the great majority were found in very large aggregations along visible convergence zones or in association with Velella velella. Discrete groups of abundant larval and juvenile fishes were found prior to upwelling [Parophrys vetulu3 , Scorpaenichthys marmoratus , Hemilepidotus spinosus, Hexagrammos sp., and A noplopoma fimbria ) and after its onset (Engraulis mordax and Sebastes spp.). These species had distinct zonal (east-west) distribution patterns and were generally associated with, or affected by, hydrographic characteristics such as convergences, upwelling, and the Columbia River plume. Recent studies of the ichthyoplankton off the northwest coast of the United States have con- tributed new information on the temporal and spatial occurrences of larvae of coastal and pelagic fish species (Richardson 1973; Richardson and Pearcy 1977; Laroche and Richardson 1979; Richardson et al. 1980; Kendall and Clark 1982a, b; Clark 1984; Bates 1984; Mundy 1983; Brodeur et al. 1985; Boehlert et al. 1985). These surveys focused on larvae occurring below the surface layer of the ocean, although concurrent neustonic samples were occasionally collected (Kendall and Clark 1982 a, b; Clark 1984; Bates 1984; Richard- son^). Comparison of simultaneous surface and subsurface samples demonstrated that many spe- cies were found in both depth strata, while an additional group of species was collected only from the neustonic zone. Brodeur et al. (1987) examined the larval fish and invertebrate compo- nents of the neuston in the northeast Pacific, and determined that these organisms were frequent prey items of juvenile salmonids. ^Oregon State University, College of Oceanography, Hatfield Marine Science Center, Newport, OR 97365; present address: University of California, Bodega Marine Laboratory, P.O. Box 247, Bodega Bay, CA 94923. 2Richardson, S. L. Oregon's coastal ichthyoneuston - a pre- liminary report. Unpubl. rep. Presented at American Soci- ety of Ichthyologists and Herpetologists, Williamsburg, VA, June 1975. Manuscript accepted November 1987. FISHERY BULLETIN; VOL. 86, NO. 2, 1988. Standard plankton and neuston nets used in these studies were effective in collecting the rela- tively slow-moving early larvae. However, net avoidance by larger larvae and juvenile stages (Barkley 1972; Murphy and Clutter 1972) sug- gests that use of traditional collecting gear is in- appropriate for more mobile fishes. This paper describes the results of a neustonic survey con- ducted off the Oregon coast in the spring and sum- mer of 1984 that focused on the larger ichthy- oneuston. Conventional sampling gear and a new net designed specifically for sampling juvenile fishes were used to characterize the temporal and spatial distribution patterns of both larval and juvenile ichthyoneuston and Dungeness crab, Cancer magister, megalopae. MATERIALS AND METHODS Sampling was conducted at approximately 2-wk intervals from early April through July 1984, on an east-west transect along the 44°40'N parallel, 3 km north of Newport, OR (Fig. 1). Sta- tions were located at distances of 1, 5, 10, 15, 20, 30, 40, and 50 km from shore. The stations were occupied twice during each 24-h cruise, once dur- ing the day and once at night. On one cruise (8-10 June), the 50 km station was occupied for 27 hours to assess diel variation in abundance of neustonic organisms. In response to several 299 FISHERY BULLETIN: VOL. 86, NO. 2 45° 00' -1 1 1 r- 440 50' 90 80 70 60 50 40 30 15 5 • • • •) 20 10 It YAQUINA BAY 125° 00' 124° 30' 124° 00' Figure 1. — Location of stations along the transect off Newport, OR. The numbers indicate distance offshore in kilometers. weeks of strong upwelling, during which very few larvae were collected close to the coast, the sam- pling scheme was modified for the two July cruises by eliminating some of the inshore sta- tions and extending the transect as far offshore as 90 km. Only night samples were collected during July. One of the primary considerations of the survey was to collect samples while minimizing net avoidance by the target organisms. Reduction of vessel-induced disturbance was accomplished by deploying the collecting gear from the ends of 12 m-long outriggers on the FV Cumberland Trail, a chartered 23 m commercial scallop-fishing ves- sel. Two different nets were used to collect the sam- ples. Larval fish and zooplankton were collected with a Manta neuston net (Brown and Cheng 1981), modified to have a mouth 1.0 m wide x 0.7 m deep. The Manta frame was equipped with a green-colored 0.333 mm mesh net, PVC cod end bucket, and General Oceanics model 2030 digital flowmeter'^. A two-point bridle was attached to the upper corners of the frame. Drag on the net while towing kept the entire bridle and towing wire assembly out of the water. The second net was designed to collect the larger juvenile fishes that were assumed to avoid the smaller Manta net. The neuston trawl was constructed with a mouth 3.50 m wide x 1.05 m deep (Fig. 2a). The frame consisted of 43 mm (out- 3References to trade names do not imply endorsement by the National Marine Fisheries Service, NOAA. side diameter) heavy-duty galvanized pipe, with towing points welded at the four corners. Flota- tion for the frame was provided by three inflat- able spar buoy floats. These 40 cm diameter floats had hollow tubes running through their centers and were fitted onto the upper bar of the frame. The 8.5 m-long net was made of 4.8 mm green- colored woven mesh, with a 15 cm-wide cloth collar around the mouth, PVC cod end bucket and General Oceanics model 2030 flowmeter. The net had a mouth slightly larger than the frame (3.70 m X 1.20 m) so the net could be laced around the outside of the frame and flotation buoys. The ends of six 12 mm polypropylene rope riblines running the length of the net were shackeled to the frame for additional support. A 4-point bridle of 6.4 mm wire was attached to the corners of the neuston trawl (Fig. 2b). Drag forces kept the en- tire bridle and towing wire (except for the two short segments attached to the bottom of the frame) out of the water. A length of 5 cm polypropylene rope was attached between the upper towing points on the frame, as a bridle for use in deploying and retrieving the net. A line attached to the retrieval bridle was passed to a hydraulic capstan through a block tied in the rig- ging over the deck. During a tow, this retrieval line was slackened, but remained attached to the vessel, and floated in a broad arc behind the net mouth. At each station, the neuston trawl and Manta net were towed simultaneously from the port and starboard outriggers, at approximately 1-1.5 m/ second. Tows were generally made either against or with the direction of the prevailing swells. The Manta net was usually fished for approximately 8-9 minutes/tow and filtered about 300-400 m^. The neuston trawl generally filtered 2,000-3,000 m'^ during a 10-11 minute tow. Additional tows with the neuston trawl were frequently made at a station to assess small-scale patchiness and to sample visible "structures" in the surface layer, such as convergence zones marked by foam lines and rafts of the pleustonic hydroid Velella uelella. On several cruises, onset of high winds and rough sea conditions prevented use of the Manta net during the daj^ime sampling of the transect as we returned toward shore. However, we were able to fish the neuston trawl in winds up to an estimated 40—45 km/hour, and in white-capped seas of 2-3 m. Neuston trawl samples were col- lected at alternate stations during the periods of adverse weather. 300 SHENKER: OCEANOGRAPHIC ASSOCIATION OF NEUSTONIC MEROPLANKTON a) 0.4m 1.05m 3.50m 1.0m 6.0m b) Figure 2a. — Front view of the neuston trawl; B = inflated 40 cm diameter spar buoys; C = cloth collar around mouth of net; R = polypropylene rib lines. Figure 2b. — Side view of the neuston trawl showing construction of the towing bridle. Floats were removed for clarity. Several very large catches of Dungeness crab megalopae in the neuston trawl were subsampled by volume, with the majority of the megalopae being returned alive to the ocean. Samples from both nets were preserved in 10% formalin/sea- water and sent to the laboratory for analysis. All larval and juvenile fishes were identified to the lowest practical taxonomic level. After extrapo- lating catch sizes from subsample counts, the data were normalized to produce density esti- mates of organisms/1,000 m^^. As an indicator of patchiness of abundant taxa, the coefficient of dis- persion (CD. = S'^/X ) was calculated for the sam- ples in which each taxon was present. Hydrographic and meteorological data were collected at each station. Surface water tempera- ture was determined with a handheld bucket thermometer. Water samples were collected for laboratory analysis of surface salinity using a Model 8400 Guildline Autosalinometer. A 200 mL surface water sample was filtered through a 0.3 fxm pore size Gelman A/E glass fiber filter for determination of chlorophyll a concentration. The filters were stored in dessicant over dry ice at sea, and then in a freezer, before acetone extraction of the chlorophyll a and analysis with a Turner De- signs Model 10 Fluorometer. During the daytime, Secchi depths were estimated, using a 30 cm disc, and surface irradiance was measured with a handheld General Electric Model 214 Light Meter. Other data collected included weather and sea state conditions at each station. RESULTS Hydrographic Data The neustonic realm was characterized by dis- tinct temporal and spatial patterns that reflect the highly dynamic springtime oceanographic 301 FISHERY BULLETIN: VOL. 86, NO. 2 processes of the Oregon coastal zone. Daily up- welling indices derived from barometric pressure data (Fig. 3; A. Bakun"*) indicated that in April and May, local winds varied in direction and strength, with one extended period of stormy weather with southwest winds (negative index values). Temperature and salinity profiles during these months showed relatively little variation along the transect within a sampling period (Fig. 4a, b, c). Chlorophyll a peaks and low salinity near shore may have been due to the influence of water exiting Yaquina Bay. The highest chloro- phyll a values were observed following periods of northwest winds, and the highest temperatures occurred after a period of very light winds and high insolation. The hydrographic regime was altered by the onset of northwest winds in June (Fig. 3), which induced the upwelling of cold, more saline nutrient-rich water along the coast. Strong up- welling [defined by Small and Menzies (1981) as a daily index value >50] was sporadic in June but nearly continuous in July. Temperature, salinity, and chlorophyll a profiles rapidly changed in re- sponse to the upwelling-favorable winds (Fig. 5a, b, c). During the 8-10 June cruise, low temperature and high salinity water was found within 5 km of the coast. Low salinity, warm water more than 15 km offshore apparently was the plume of the Co- lumbia River. By 19-20 June, continued offshore transport of the surface layer pushed the inshore edge of the plume to 40 km off the coast. Low chlorophyll a concentrations, low temperatures, and high salinity near shore on 19-20 June indi- cated the occurrence of active upwelling. Upwelling persisted through the end of July. Temperature and salinity profiles were similar to those observed during the second June cruise, with the patterns of increasing temperature and decreasing salinity farther offshore continuing to the limits of sampling of 90 km on 8 and 9 July and to 70 km on 23 and 24 July. During the up- welling in July, a dramatic phytoplankton bloom occurred in the surface waters within 30 km of the coast. Chlorophyll a concentrations here were denser and broader in offshore extent than the surface and subsurface biomass concentrations previously observed off Oregon (Small and Men- zies, 1981). + 200 « T3 C c 0) a 3 -200- U I ||'|||l||l ll'l' f |'l|| '1 11 1 T lll.i„lil yt4 y^ /N /K yfs yT\ xTs xl\ "•A. Bakun, Pacific Environmental Group, National Marine Fisheries Service, Monterey, CA 93940, pars, commun. August 1984. M J J 1984 Figure 3. — Daily upwelling indices for spring-summer 1984. Positive values >50 indicate occurrence of winds inducing strong upwelling. Arrows indicate dates of sampling cruises. Ichthyoneuston A total of 107 Manta net and 142 neuston trawl samples collected 48 taxa of larval, juvenile, and adult fishes. Larvae <10 mm and a few juveniles were collected by the Manta net, while large numbers of juvenile fishes up to 60-70 mm were taken by the neuston trawl. Size-frequency data for three of the commonest species (Fig. 6 a, b, c) illustrate the relative ability of the nets to cap- ture different sizes of fishes. Nighttime sampling (59% of the Manta net tows and 62% of the neuston trawl tows) ac- counted for 93.3% of all fishes in the Manta net catch and 96.5% of the fishes taken in the neuston trawl (Table 1). Four Manta net and 8 neuston trawl tows made during twilight collected 5.4% and 2.9%, respectively, of the total number of fishes taken with each gear type (Table 2). Only 1.3% of the Manta net catch and 0.6% of the neu- ston trawl catch was made during daytime (Table 2). Only two species were not collected predomi- nantly at night. Larval Pacific saury, Cololabis saira, were taken in both day and night Manta net samples. Larval northern lampfish, Steno- brachius leucopsarus , were abundant only in the twilight samples from early June (Tables 1, 2). Arbitrary criteria on frequency of occurrence (in >15% of the night samples of either net) or abundance (peak densities >40/l,000 m"^) indi- cated that nine taixa of larval, juvenile, and adult fishes were dominant components of the ichthy- oneuston. Most taxa were characterized by dis- 302 SHENKER: OCEANOGRAPHIC ASSOCIATION OF NEUSTONIC MEROPLANKTON 14 o o < if) E >- X Q_ O O _J X o O O 4/6 • • 4/16-17 A A 5/4-5 ▲ A 5/18-19 10 20 30 40 DISTANCE FROM SHORE (km) % 50 Figure 4a-c. — Nighttime environmental conditions observed on each cruise during the pre-upwelling period, a = temperature; b = salinity; c = chlorophyll a concentration. 303 FISHERY BULLETIN: VOL. 86, NO. 2 o o a: 01 LJ Q. UJ o o Z _J < Dl >- X Q_ O q: o _j o 30 40 50 60 70 80 DISTANCE FROM SHORE (km) 100 Figure 5a-c. — Nighttime environmental conditions observed on each cruise during the upwelling period, a = temperature; b = salinity; c = chlorophyll a concentration. 304 SHENKER OCEANOGRAPHIC ASSOCIATION OF NEUSTONIC MEROPLANKTON tinct temporal shifts in occurrence: Five taxa were abundant only prior to the onset of continu- ous upwelling in mid-June, three taxa were spo- radically abundant following the beginning of up- welling, and one species (Ronquilus jordani) was present tliroughout the survey. Pre-upwelling Species The numerically dominant species in the nearshore region in spring was larval English sole, Parophyrys vetulus. These larvae were patchily distributed (CD. = 32.5) within 30 km offshore, reaching densities of 75/1,000 m'^ in Manta net collections. Larval abundance declined rapidly after April (Table 1). Size-frequency dis- tributions indicated that the Manta net was more effective than the neuston trawl for collecting fish <15 mm (Kolmogorov-Smirnov test, P<0.01). Eye migration in the larger fish (20-24 mm) was nearly complete, indicative of their impending shift to a benthic existence. Similar-sized juve- niles were taken nearshore in April in benthic tows (B. Mundy^). Cabezon, Scorpaenichthys marmoratus , larvae were most abundant in Manta catches in the spring (Table 1), although newly hatched larvae were collected near shore through June. The smallest larvae (4-6 mm.) were taken only at the 1-15 km stations, while larvae up to 10 mm were found along the entire transect. These larvae were patchily distributed, with April densities reaching 207/1,000 m^ (CD. = 92.7). Frequency of occurrence and larval density declined through early July. Large juvenile cabezon (26-38 mm) were collected by the neuston trawl (Fig. 6a). These juveniles were encountered all along the transect from April to mid-June. Two dominant taxa found at all stations prior to upwelling were greenling (Hexagrammos sp.) and brown Irish lord, Hemilepidotus spinosus . Ju- venile greenlings were the most frequently col- lected species, occurring in AA.9>% of the night- time neuston trawl samples (Table 1), although they were rarely taken by the Manta net (Fig. 6b). They were the most evenly dispersed of the abun- dant fish collected (CD. = 2.2), and never ex- ceeded densities of 10/1,000 m^. The two largest catches of greenling were made in association with distinct hydrographic features. Nineteen ju- a) Scorpaenichthys marmoratus 100 -, 50 - I J Manta net Neuston trawl TT' 4 6 8 10 20 26 32 38 b) Hexagrammos sp. 40 —I 30 - 20 - 10 - c) Anoplopoma fimbria 5B. C. Mundy, Southwest Fisheries Center Honolulu Labora- tory, National Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 96822, pers. commun. May 1984. T r 10 20 30 40 50 60 70 80 90 100 IK) 120130 Standard Length (mm) Figure 6a-c. — Length-frequency data of larval and juvenile fishes collected by each net. a = Scorpaenichthys marmoratus ; b = Hexagrammos sp.; c = Anoplopoma fimbria. veniles were collected from a convergence zone near shore in mid-May (along with 150,000 Dun- geness crab megalopae) and 20 juveniles were taken from Columbia River plume water 50 km offshore in early June. Mean lengths of greenling from the neuston trawl samples increased from 305 FISHERY BULLETIN: VOL. 86, NO. 2 o o> c a> £1 E :d Tr c CO II en a^ c 0) >> o c 0} 3 3 (/) o QJ o o £ o O >. o c 0) 3 (T 0) tn o a> C/5 OJ y- in d IT) CD O) >- t- CO CO CD CO CO CO T- •.- CD CD in in uo CO CD d 1 s CO 'J- CD 1 (D 1 O CO Cvj ^ CO C\J PO CD - 1 00 in - CO - CM ;; 1 CM in r^ CM 'I- o CO CD CM CO CO d d CM d ■•- '- d CO o CD O CD CO CO CO in CD CO CM CD CM t r^ r^ r- o tt ■^ I^ CD CD CO T- O r~ O) CO o in CD CM T- CM CO CO CD CO CO T- CM CM •-- •r- CO O in t^ CD co CD in o • E CD to O ™ ^,^ 55 2 CO 3 C CO <0 S iS 5 3 CD ^ O CD £ 0) CO 3 o -c 5 c £l ■^ "^ fc X: -o O CO 3 .E O CD CO o -5 o ^ CO CXJ CD si u en CO 3 'cB -3J 3 CJ CD 0) CO m ? "C Oi ^ P CO ^^ CO CD If 3 CO qj C CO O Q. CO c CO 3 en o (0 >< CO (D CO £ (O c (D > _ 3 . . -Q E ■^ 0) (0 to _ __ E ^.fe 9? S S §-■§ -5 c "o t 3J to to E S. CO CO CO < -C X to CD Q. >< I 306 SHENKER: OCEANOGRAPHIC ASSOCIATION OF NEUSTONIC MEROPLANKTON 13 c o O CD < <£> o 0) CO d CO CO 00 CVJ 1^ CO •* CO c\j r-- •■- C\J CO r^ (j> ia n O r- <£> ID un o '- ■^ CM en CO ■* CO d d CO uo I/) T- uo ■>- OJ •- ■.- O) T- ^ ^ •^ 00 uo CD in d oj d (b o CO CO CO o uo (D (D CM UO 00 CO LO O O) n Tt o 1 ^ Tl- CD CO CM in CM CO t-~^ If) 1^ CO CM ■■- CO O CD CM 00 1— (D CD 2i 1 1 '^ 1 "^ ■V h- 1 "^ CM 1 1 '- ^ CM 1 LL 1 1 1 1 o 00 1 i ^ CO CD r^ 1 •* •^ 1 1 in 1- 1- 1 CM ^' < 1 1 r^ r^ ■* in 1 c^ r-^ 1 1 CM in CD 1 CO d d d T-' d d d (6 - I - CO in ■- CO in uri CD r-- '- ••-■<«■ oo CvJ CD CD d d CO in ■•- CM d ■* o CM CO CO in d 1 CO 00 CO 1 ^~ Cvj CM '" CO CO ■* CO CM in o 1 °° CO CM CM CO CD in 1 CO ■* O) CM CM CM CD r-- B o c C3) c CO O o CO tT3 x: 5. t^ ,„ CO g « CO nj cio
  • CB :9 E t; U) O I CTJ S '^ ?15 O -C ^ Q. m 0) c 3 -3 C 00 CO o E Q c (0 CO 5 CO 05 c c fO 5 CO CO c CO CO Q. < C CO en 0) o 0) d CO C\J <£> CM ■^ CD LT) CD CD CO (N T- >- <>i •* y- ■■- •* TT CD in CO C\j T- 1- c\i CO o in CO o o ••- T- O £31 O in in in o CO CO CO CVJ CO in CO CO in CM CO CJ5 CO in en CO IT) CO '- CO CO h- ■.- ,- r- CD CO '-CD CD O CO CO CJ5 CD -^ in CO o ■>*■ in in CM Cvj ■■-■•- CO CO co in o CM CO in in o CD oo CD CM CM ■^ CM N- in in CO I o 6 0) CO CO -^ •D 3 CD Ol a>uj LU i CD to 2 - CO o -o If 3 CO "3" y C CO o CO c: "cD CO CO CU -c CO QJ 0) CO u "9 § o ^£ -^ E < s '^ C (0 CO jo ^^ . _ , O CO CO CO CO CO •Q I CO E o CL It 9^ E ~ ''J CD X CO (U >< I: I CO 3 'cO CD c: o "S c O CD Ob o CO & CJ o CO 3 o o o c in cii CO £ o 00 CM CO CD CO ^ 3 §o - c 9f E o ^ "a li O i; F ICO 2.i CO o w 9 SJ Q.B a? CD ■§_^;_ - S o ^ 2 9- k 3 .° ^ "O co t^ 00 CM 0 CD CJ> '* CM 00 CD 0 0 CD in CU cO CM y- 0 d ■^ G> CM CO •\ r 1 - 0 \ < 1 1 E — ffl- -ffl- -E- -m 1.000 •■ 16-17 April 1.000 ■• 100D 1.000- 8-9 July / A ^"^A 100- ■ / A 10- • 1 0^ ■i — 1 1 1 1 — 1 1 1 1.000 ■ 100• 10 1 + 0 n- / t I H H A 18-19 May I.OOOt 100- 10 0 5 10 15 20 30 40 50 1 ■• 0 23-24 July A— i H 1- 0 5 10 15 20 30 40 Distance From Shore (km) 50 Figure 7.— Daytime and nighttime densities of megalopae collected by the Manta net. Rough weather prevented the use of the Manta net during the day of 19 May; neuston trawl data for this portion of the cruise are presented. Circled letters refer to events discussed in the text. 310 SHENKER; OCEANOGRAPHIC ASSOCIATION OF NEUSTONIC MEROPLANKTON the Manta net in the daytime were lower than at night for each station (Fig. 7). The differences between day and night were particularly appar- ent on bright sunny days and at the offshore sta- tions. Only one of 13 samples of megalopae col- lected when light intensity exceeded 4,000 lux had a density over 200/1,000 m^, while 12 of 24 day samples collected at lower light levels ex- ceeded this density. Highest daytime densities (1,600-2,400/1,000 m^ in neuston trawl samples) were observed on 19 May when a sudden shift in weather before daybreak resulted in very dark, overcast and rainy skies, with 40-45 km/hour winds and 2-3 m breaking waves. Water clarity may also affect the daytime abundance estimates of megalopae in the neuston along the transect. Daytime densities were al- ways low seaward of 30 km where Secchi depths exceeded 9 m, while higher densities were ob- served in the more turbid near-coastal waters with Secchi depths of 4-9 m. The 27-h occupancy of the 50 km station on 8-9 June provided additional evidence of a distinct diel pattern in utilization of the neustonic habitat by megalopae (Fig. 8). Abundance of megalopae peaked during the twilight periods of dawn and dusk. A significant decrease in abundance was noted in the middle of the night, while midday samples did not collect any megalopae. In addition to the wide distribution of megalo- pae along the transect, sampling occasionally de- tected the aggregation of very large numbers of megalopae (denoted by letters over station blocks n 1,000 o o o 0) D a. o o 0) 0) E 100 -- 10 o o Manta Net • • Neuston // Trawl // I // 1 -: . 0 I ' ' f^m^^^^^m=^ I i I . ii I T-rr '•' i 18:00 24:00 06:00 12:00 18:00 Time of Day Figure 8. — Density of megalopae collected by h)oth nets during a 27 h sampling p>eriod 50 km ofTshore on 8-9 June. in Figure 7). These aggregations were typically associated with hydrodynamic or biological struc- tures in the surface layer. Event A: Simultaneous tows with the Manta net and neuston trawl caught megalopae in densi- ties of 373 and 325/1,000 m^. The neuston trawl was then redeployed and towed along a foam line marking a surface convergence. An estimated 60,000 megalopae were collected from the conver- gence zone (24,000/1,000 m^). Event B: A series of five consecutive tows with the trawl net was made. Each tow covered a linear distance of approximately 1 km in a west- ward direction, with an estimated distance of 300 m between tows. Megalopae were very sparse in the simultaneous Manta and trawl tows at the beginning of the series, with densities of 0 and 1.9/1,000 m"^, respectively. The subsequent neu- ston trawl tows produced density estimates of 79, 3,555, 2,414 and 187/1,000 m^. Apparently, the megalopae were aggregated over an area extend- ing at least 5 km in the east-west direction. Event C: Following a period of southwest winds, a large raft of Velella velella was found around the inshore station. Velella velella were generally found at offshore stations but at densi- ties far lower than the 36 kg/1,000 m^ level taken here. Although the overall extent of the raft could not be determined at night, the density of megalo- pae was estimated by both nets at 21,500/1,000 Event D: The density of megalopae sampled by the Manta net was 790/1,000 m^, while a simultaneous neuston trawl tow along a conver- gence zone only 30 m away collected 150,000 megalopae (46,400/1,000 m^). Census estimates of megalopal abundance were computed from nighttime samples for each date (Fig. 9). A modification of Smith's regional census estimate (1972) was used to compute the total number of megalopae occurring in a 1 m wide X 1 m deep track along the surface from the coast to 55 km offshore: S (^A) j=i where C^ = estimate of abundance of megalopae per meter of coastline during cruise k Aj = volume of water in the track sur- rounding station i, computed from 311 FISHERY BULLETIN: VOL. 86, NO. 2 10' 0) o Q. _o a 0) 0) E 3 10 4 .. -O O O Transect Census • Convergence Zones 10- 4/6 4/16 5/5 5/18 6/10 6/20 7/8 7/23 Sampling Date Figure 9. — Census estimates of Dungeness crab megalopae at night from Manta net samples for each cruise in a 1 m wide x 1 m deep track extending from the coast to 55 km offshore (30-95 km offshore on 7/8). Single samples from convergence zones sampled with the neuston trawl contained more megalopae than the entire census estimates for the respective cruises. the midpoints of the distances be- tween the adjacent stations Z), = density of megalopae at station i. Although census estimates varied between cruises from approximately 4,000-78,000, these estimates ignored the concentrations of megalo- pae in convergence zones or other aggregations that were sampled by the neuston trawl. The con- vergences alone contained 5-15 times more megalopae than the entire census estimates for their respective cruises. DISCUSSION The neustonic habitat off Oregon is a dynamic environment that supports an abundant and di- verse fauna including Dungeness crab larvae and numerous fish species. The large mouth size of the neuston trawl and minimal disturbance from the vessel probably contributed to larger catches of surface-dwelling ichthyofauna than have been made in previous studies in the same region (Boehlert et al. 1985). Several species of neustonic organisms were as- sociated with specific oceanographic features such as convergence zones and water masses like the Columbia River plume. These aggregations of Dungeness crab megalopae and several fish spe- cies accounted for the major portion of the total catch of each taxon. Recent studies by Shanks (1983, 1985) and Kingsford and Choat (1985, 1986) further emphasize the importance of aggre- gation of neustonic organisms in oceanic conver- gences, surface slicks, and around floating ob- jects. This information clearly demonstrates that randomized or grid sampling plans will often fail to detect micro- or meso-scale features of the environment. Surveys may thus severely under- estimate the abundance of species, and impor- tant data on ecological characteristics such as predator-prey interactions may not be obtained. This is particularly important in understanding the role of spatial co-occurrence of patches of lar- val fishes and their prey on larval survival and growth (Lasker 1975, 1981; Grover and 011a 1986). Diel patterns in abundance were striking. Most of the neustonic organisms were collected at night, and their absence from daytime collections may be attributed to vertical migration out of the surface layer or to visual avoidance of the nets. The fact that very few larval or juvenile sablefish and greenlings were collected in subsurface sam- ples in earlier studies (e.g., Richardson and Pearcy 1977; Richardson et al. 1980; Kendall and Clark 1982a; Clark 1984), suggests that these fishes are obligate inhabitants of the neuston, and their low abundance in daytime hauls indi- cates substantial visual avoidance of the nets. Conversely, other species may be facultative neuston that undergo diel migrations into the surface layer (Hempel and Weikert 1972), where they can most easily be assessed. Dungeness crab megalopae, in particular, disperse to at least 50 m during the day, but concentrate at the surface at dawn and dusk (Booth et al. 1985). Still other species (the pseudoneuston) may have depth ranges that overlap with the surface layer. Previous studies on northwest ichthyoplankton have defined basic coastal, transitional, and off- shore assemblages of species during different sea- sons (Richardson and Pearcy 1977; Richardson et al. 1980; Kendall and Clark 1984a, b; Clark 1984). In general, the transitional region roughly parallels the shelf break (approximately 30-40 km offshore in the vicinity of Newport, OR). Richardson et al. (1980) ascribed the consistency of these zonal assemblages to the spawning habits of adults and larval transport in the alongshore coastal circulation. According to Parrish et al. (1981), the spatial and temporal patterns of spawning, and durations of pelagic larval stages of these species, should correspond with the sur- face drift patterns of the region to minimize lar- val advection out of suitable habitats. In the Pacific Northwest, species with larvae adapted to the nearshore zone should spawn from fall 312 SHENKER: OCEANOGRAPHIC ASSOCIATION OF NEUSTONIC MEROPLANKTON through early spring when net surface drift is primarily northward and onshore. Oceanic lar- vae, however, should be spawned following the spring-summer transition when upwelling re- sults in southward and offshore transport. The onshore/offshore distributions of several abundant fish species in 1984 were similar to those described earlier, and their seasonal occur- rence usually coincided with the constraints dis- cussed by Parrish et al. ( 1981). However, four spe- cies had distributions that crossed the previously described zonal boundaries (Fig. 10). Early cabezon larvae appeared to disperse offshore as they grew. In contrast, early juvenile brown Irish lords were collected offshore in the early spring, while the larger juveniles spread inshore across all stations. Larval greenlings have only been col- lected close to the coast in late fall (Bates 1984), while the juveniles collected in this study were found at all stations prior to upwelling. Juvenile ronquils were also distributed along the transect, during both non-upwelling and upwelling condi- tions. The dispersion of organisms across the transi- tion zone may have been accomplished by several mechanisms. Physical transport of the organisms by mesoscale hydrographic events (e.g., eddies, offshore jets, and meanders in the alongshore cur- rents) (Ikeda and Emery 1984; Mooers and Robinson 1984; Abbott and Zion 1985; Davis 1985) undoubtedly play important roles in the onshore/offshore dispersal of planktonic organ- isms. Transport with these events may be acceler- ated or hindered by diel vertical migration into water layers with different zonal flow patterns. Additionally, several species (especially green- ling and sablefish) are rapid swimmers whose mo- bility can contribute to their dispersal or aggrega- tion. The rapid change in the composition of the ichthyoplankton fauna following the onset of up- welling has not been previously observed. Fishes abundant prior to upwelling may respond to the change in the environment in different ways. The disappearance of greenling and brown Irish lord juveniles suggests upwelling triggered settle- ment of these fishes to the demersal habitat uti- lized by older juveniles and adults. A possible stimulus for this transition is the breakdown of the thermocline, which has been identified as a 1 ' 90 8*0 70 60 UPWELLING 1 1 1 1 PRE-UPWELLING V." 45° nn' Sablefish 1 10-20 nnm 20-30 mm Brown Irish Lord S (vy: Greenlinq p-. /■■'■■ 4-6 mm 1 - ^[Vaquina A^y BAY & :■ 1 6-10 mm Cabezon English Sole • • • • • • 50 40 30 20 15 10 km • 5 / 4 4" _ Rockfish SDD. 3 0' 1 1 1 1 Blue Lanternfish Northern Anchovy 1 125° 00' 124° 30' 124° 00' Figure 10. — Summary of the zonal distributions of larval and juvenile neustonic fishes in the pre- and early upwelling period (April through mid-June) and the strong upwelling period (July). 313 FISHERY BULLETIN: VOL. 86, NO. 2 barrier to settlement to neustonic juvenile red hake, Urophycis chuss, (Steiner et al. 1982; Steiner and 011a 1985). In contrast, juvenile sablefish apparently re- main in the water column off Oregon following the transition to upwelling conditions. Small ju- veniles (<50 mm) have been captured as far as 250 km offshore in the spring (Kendall and Clark 1982a), and specimens up to 250 mm have been collected from the surface waters in late summer with a purse seine (Brodeur and Pearcy 1986). The very rapid growth rates of young juveniles, reaching 2 mm/day (Boehlert and Yoklavich 1985; Shenker and 011a 1986), undoubtedly re- sulted in their increasing ability to avoid the towed neuston nets used in this study. The most abundant larvae occurring after the onset of upwelling were northern anchovy and rockfish. As observed in previous studies (Richardson 1973, 1980; Richardson et al. 1980), these larvae were primarily found offshore. The occurrence of these species, along with a high abundance of vertically migrating blue lantern- fish, on the periphery of the Columbia River plume 60-90 km offshore on only one cruise, fur- ther illustrates the patchy nature of neustonic distributions. Although the fish fauna in 1984 was generally characterized by discrete temporal and/or spatial limitations, the occurrence of Dungeness crab megalopae transcended these limits through the 4-mo study. Megalopae were the most abundant organisms collected throughout the survey, but their abundance varied widely between adjacent stations, with occasional very dense patches. Sim- ilar patchy distributions of megalopae were ob- served off British Columbia by Booth et al. (1985), who measured horizontal patch dimensions of 2-4 km. Several dense swarms of megalopae were observed in Bodega Bay, CA, during 1985 (Shenker and Botsford^). These patches were long, sinuous aggregations extending 10-20 m along the surface, approximately circular in cross section and about 1 m in diameter. Densities were visually estimated to be on the order of thousands of megalopae per m^. The temporal occcurrence of crab larvae in the plankton spans two distinctly different oceanic regimes, and the larvae are potentially trans- ported long distances by the seasonal currents. ^Shenker, J. M., and L. W. Botsford, University of California, Bodega Marine Laboratory, P.O. Box 247, Bodega Bay, CA 94923, unpubl. data. Larvae typically hatch in mid-winter, pass through 5 zoeal stages in approximately 90 days, and then a month-long megalops stage before set- tling to the bottom (Reilly 1983a). After hatching, zoeae are released into the northerly flowing Davidson Current. Despite a general onshore component of flow of the Davidson Current, older zoeal stages have been found progressively far- ther offshore (Lough 1976; Reilly 1983a). About the time of the spring transition in the alongshore currents from a northerly to southerly direction (Huyer et al. 1975), zoeae metamor- phose into megalopae. To survive, these megalo- pae must be transported back toward the shore, and settle to the bottom in depths shallower than 25 m (Reilly 1983b). Larvae have been found at least 100 km from shore (Lough 1976), although it is unclear if these larvae make it back to shore, or are simply lost from the population. Again, an apparent anomaly exists between the directions of movement of megalopae and surface waters, where the Ekman layer moves offshore in response to upwelling winds. These discrepancies may be explained by mesoscale mixing processes, as cited earlier. Alternatively, the diel vertical movements of the larvae can move them into dif- ferent water masses with different directions of zonal movement. The data from this study and previous research on vertical migration indicate that at least the early zoeal stages and megalopae move to the sur- face during twilight, and below the surface dur- ing the day (Reilly 1983a). Surface abundance estimates of megalopae obtained in the 27-h sam- pling on 8-9 June (Fig. 8), and by Booth et al. (1985), decreased during the middle of the night. This movement away from the surface is an ex- ample of "midnight scattering", perhaps result- ing from the lack of a light cue to orient plank- tonic organisms to the surface (Owen 1981). If midnight scattering is typical for megalopae, abundance estimates made by sampling along a transect throughout the night (Figs. 7, 9) proba- bly underestimate the actual abundance of mega- lopae utilizing the surface habitat. Megalopae have been collected as deep as 50-70 m during the day (Booth et al. 1985; Shenker and Botsford fn. 6). These observations correlate with Jacoby's (1982) laboratory demonstration that megalopae are positively phototactic to dim light, but avoid bright light. The present study indicates that the phototac- tic response of megalopae may assist their return to shore in several ways. Downward movement 314 SHENKER: OCEANOGRAPHIC ASSOCIATION OF NEUSTONIC MEROPLANKTON during sunny days, when upwelling wind stress and offshore Ekman transport is generally strongest, can move the larvae into the slow on- shore flow below the surface Ekman layer (Peter- son et al. 1979). In contrast to the usual northerly upwelling winds, occasional storms blow from the southwest, driving the surface layer onshore. The occurrence of megalopae at the surface during the day was most pronounced on these dark stormy days (especially on the second May cruise), and thus may facilitate their onshore transport. Observations of megalopae entering into em- bayments and nearshore areas from Washington to northern California in 1984 and 1985 indicated a dramatic increase in the abundance of megalo- pae and an extended seasonal occurrence in the plankton, as compared to previous years (Shenker and Botsford fn. 6; Armstrong^). This high abun- dance of crab larvae may presage an upswing in the cyclical crab fishery along these coasts. Numerous mechanisms have been proposed as causes of the 10-yr cycles in crab abundance. Al- though some hypotheses have been discounted, several models have survived scrutiny as possible causes of the cycles (see Botsford 1986 for review). Potential mechanisms of environmental forcing of the cycles focus on larval transport and sur- vival. Johnson et al. (1986) detected periodic 10- yr cycles in the occurrence and strength of south- ward stress during the late larval period that significantly correlated with commercial crab catch 4 and 5 years later. This lag corresponds to the time between larval settlement and growth into the adult fishery (Botsford 1984). Model simulations by Botsford (1986) indicated that nonlinear effects of wind on larval transport can produce the cyclical swings in crab abun- dance. However, the models do not preclude the possibility that density-dependent phenomena (such as cannibalism by adult crabs on newly set- tled juveniles, and predation by nemertean worms on egg masses) may act in concert with the environmental forcing to produce the observed cycles. The water's surface is the only oceanic habitat that is easily accessible to observation using tech- niques ranging from satellite and aerial scanning to shipboard visual sighting of targets and contin- uous monitoring of environmental parameters. Because of this accessibility, micro- and meso- scale patterns in distribution of neustonic taxa '^D. A. Armstrong, School of Fisheries, WH-10, University of Washington, Seattle, WA 98195, pers. commun. October 1984. and their associations with hydrographic and bio- logical characteristics of the surface zone can be determined more easily than in other environ- ments. The neustonic realm thus offers an excel- lent opportunity to investigate the mechanisms of transport of the early stages into appropriate nursery habitats, and the availability of food for growth that are required for successful recruit- ment into adult stocks (Hjort 1914; Lasker 1975, 1981; Frank and Leggett 1982; Sinclair et al. 1984). ACKNOWLEDGMENTS This research was supported by the Northwest and Alaska Fisheries Center of the National Marine Fisheries Service. 1 thank Captain Leland Oldenberg and his crew on the FV Cumberland Trail, and my colleagues R. Brodeur, A. Chung, J. Fisher, J. Hennessey, L. Krasnow, B. Mundy, C. Paczkowski, and E. Rexstad for their assis- tance on the sampling cruises. W. Laroche helped identify juvenile Sebastes. W. G. Pearcy provided valuable advise on the design and operation of this project, and comments on the manuscript. Helpful suggestions on this paper were also given by R. Brodeur, A. W. Kendall, and two anony- mous referees. LITERATURE CITED Abbott. M. R., and P. M. Zion. 1985. Satellite observations of phytoplankton variability during an upwelling event. Cont. Shelf Res. 4:661-680. Barkley, R. A. 1972. Selectivity of towed net samplers. Fish. Bull., U.S. 70:799-820. Bates. R. D. 1984. Ichthyoplankton off Washington, Oregon and northern California, October-November, 1981. Natl. Mar. Fish. Serv., Northwest and Alaska Fish. Cent. Proc- ess. Rep. 84-12, 41 p. Boehlert, G. W., D M. Gadomski, and B C. Mundy. 1985. Vertical distribution of ichthyoplankton off the Or- egon coast in spring and summer months. Fish. Bull., U.S. 83:611-622. Boehlert, G. W., and M. M. Yoklavich. 1985. Larval and juvenile growth of sablefish, Anoplopoma fimbria, as determined from otolith incre- ments. Fish. Bull., U.S. 83:475-481. Booth. J., A. Phillips, and G S Jamieson. 1985. Fine scale spatial distribution of Cancer magister megalopae and its relevance to sampling methodol- ogy. In B. R. Melteff (editor), Proceedings of the sympo- sium on Dungeness crab biology and management, p. 273-286. Univ. Alaska Sea Grant Coll. Program, Rep. No. 85-3. Fairbanks, AK 99701. Botsford. L. W. 1984. Effect of individual growth rate on expected behav- 315 FISHERY BULLETIN: VOL. 86, NO. 2 ior of the northern California Dungeness crab (Cancer magister) fishery. Can. J. Fish. Aquat. Sci. 41:99-107. 1986. Population dynamics of the Dungeness crab (Cancer magister). In G. S. Jamieson and N. Bourne (editors), North Pacific workshop on stock assessment and man- agement of invertebrates. Can. Spec. Publ. Fish. Aquat. Sci. 92:140-153. Brodeur, R D. D M Gadomski, W G Pearcy. H P. Batchelder, and C B Miller 1985. Abundance and distribution of ichthyoplankton in the upwelling zone off Oregon during anomalous el Nino conditions. Estuarine Coastal Shelf Sci. 21:365- 378. Brodeur, R. D. B C Mundy, W G Pearcy, and R W. WiSSEMAN. 1987. The neustonic fauna in coastal waters of the north- east Pacific: Abundance, distribution, and utilization by juvenile salmonids. Oreg. State Univ. Sea Grant Coll. Program, Rep. ORESU-T-87-001. Corvallis, OR 97331. Brodeur, R D . and W. G. Pearcy. 1986. Distribution and relative abundance of pelagic non- salmonid nekton off Oregon and Washington, 1979- 1984. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 46, 85 p. Brown, D M , and L Cheng 1981. New net for sampling the ocean surface. Mar. Ecol. Prog. Ser. 5:225-227. Clark, J B 1984. Ichthyoplankton off Washington, Oregon and northern California, May-June, 1981. Natl. Mar. Fish. Serv., Northwest and Alaska Fish. Cent. Process. Rep. 84-11, 46 p. Davis, R E 1985. Drifter observations of coastal surface currents dur- ing CODE: the method and descriptive view. J. Geophys. Res. 90:4741-4755. Frank, W T , and W C. Leggett. 1982. Coastal water mass replacement: its effect on zooplankton dynamics and the predator-prey complex as- sociated with larval capelin (Mallotus villosus ). Can. J. Fish. Aquat. Sci 39:991-1003. Grover, J J , AND B L Olla. 1986. Morphological evidence for starvation and prey size selection of sea-caught larval sablefish, Anoplopoma fim- bria. Fish. Bull., U.S. 84:484-489. HEMPEL, G , AND H Weikert 1972. The neuston of the subtropical and boreal North- eastern Atlantic Ocean. A review. Mar. Biol. (Berl.) 13:70-88. Hjort, J 1914. Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Rapp. P. -v. Reun. Comm. int. Explor. Sci. Mer 20:1- 228. HUYER, A , R D PiLLSBURY, AND R L. SMITH 1975. Seasonal variation of the alongshore velocity field over the continental shelf off Oregon. Limnol. Oceanogr. 20:90-95. Ikeda, M . AND W J Emery 1984. Satellite observations and modeling of meanders in the California Current system off Oregon and Northern California. J. Phys. Oceanogr. 14:1434-1450. Jacoby, C a. 1982. Behavioral responses of the larvae oiCancer magis- ter Dana (1852) to light, pressure, and gravity. Mar. Behav. Physiol. 8:267-283. Johnson, D F , L W Botsford, R D Methot, Jr , and T C. Wainwright 1986. Wind stress and cycles in Dungeness crab (Cancer magister) catch off California, Oregon and Washing- ton. Can. J. Fish. Aquatic Sci. 43:838-845. Kendall, A W . Jr , and J Clark 1982a. Ichthyoplankton off Washington, Oregon, and Northern California. April-May, 1980. Natl, Mar. Fish. Serv., Northwest Alaska Fish. Cent. Process. Rep. 82-11, 44 p. 1982b. Ichthyoplankton off Washington, Oregon and Northern California. August, 1980. Natl. Mar. Fish. Serv., Northwest and Alaska Fish. Cent. Process. Rep. 82-12, 43 p. Kingsford, M J , AND J H Choat 1985. The fauna associated with drift algae captured with a plankton-mesh purse seine net. Limnol. Oceanogr. 30:618-630. 1986. Influence of surface slicks on the distribution and onshore movements of small fish. Mar. Biol. (Berl.) 91:161-171. Laroche, J L, AND S. L Richardson 1979. Winter-spring abundance of larval English sole, Parophrys vetulus , between the Columbia River and Cape Blanco, Oregon, during 1972-1975, with notes on occurrences of three other pleuronectids. Estuarine Coastal Mar. Sci. 8:455-476. Lasker, R 1975. Field criteria for survival of anchovy larvae: the relation between the inshore chlorophyll layers and suc- cessful first feeding. Fish. Bull., U.S. 71:453-462. 1981. The role of a stable ocean in larval fish survival and subsequent recruitment. In R. Laisker (editor), Marine fish larvae. Morphology, ecology and relation to fisheries, p. 80-87. Wash. Sea Grant Program, Univ. Wash. Press, Seattle. Lough, R G 1976. Larval dynamics of the Dungeness crab, Cancer magister, off the central Oregon coast, 1970-71. Fish. Bull., U.S. 74:353-375. MooERS, C. N K , and a. R. Robinson. 1984. Turbulent jets and eddies in the California Current and inferred cross-shore transport. Science 223:51-53. MUNDY, B. C. 1983. Yearly variation in the abundance and distribution of fish larvae in the coastal zone off Yaquina Head, Ore- gon, from June 1969 to August 1972. M.S. Thesis, Ore- gon State Univ., Corvallis, OR, 158 p. MuRPHY, G I , AND R I Clutter 1972. Sampling achovy larvae with a plankton purse seine. Fish. Bull., U.S. 70:789-798. Owen, R W 1982. Fronts and eddies in the sea: Mechanisms, interac- tions and biological effects. In A. R. Longhurst (editor). Analysis of marine ecosystems, p. 197-233. Acad. Press, N.Y. Parrish, R H , C S Nelson, and A Bakun. 1981. Transport mechanisms and reproductive success of fishes in the California Current. Biol. Oceanogr. 1:175- 203. Peterson, W T , C B Miller, and A. Hutchinson. 1979. Zonation and maintenance of copepod jwpulations in the Oregon upwelling zone. Deep-Sea Res. 26:467- 494. Reilly, P N 1983a. Dynamics of Dungeness crab. Cancer magister. 316 SHENKER: OCEANOGRAPHIC ASSOCIATION OF NEUSTONIC MEROPLANKTON larvae off central and northern California. In P. W. Wild and R. N. Tasto (editors!, Life history, environment, and mariculture studies of the Dungeness crab. Cancer magister . with emphasis on the central California fish- ery, p. 57-85. Calif Dep. Fish game, Fish Bull. 172. 1983b. Predation on Dungeness crabs, Cancer magister, in central California. In P. W. Wild and R. N. Tasto (editors), Life history, environment, and mariculture studies of the Dungeness crab. Cancer magister, with emphasis on the central California fishery, p. 155- 164. Calif Dep. Fish. Game, Fish Bull. 172. Richardson. S L. 1973. Abundance and distribution of larval fishes in wa- ters off Oregon, May-October 1969, with special empha- sis on the northern anchovy, Engraulis mordax. Fish. Bull., U.S. 71:697-711. 1980. Spawning biomass and early life of northern an- chovy, Engraulis mordax, in the northern subpopulation off Oregon and Washington. Fish. Bull., U.S. 78:855- 877. Richardson, S L . J L Laroche, and M. D. Richardson. 1980. Larval fish assemblages and associations in the northeast Pacific Ocean along the Oregon coast, winter- spring, 1972-1975. Estuarine Coastal Mar. Sci. 11:671-699. Richardson, S L , and W G Pearcy. 1977. Coastal and oceanic fish larvae in an area of up- welling off Yaquina Bay, Oregon. Fish. Bull., U.S. 75:125-146. Shanks, A L 1983. Surface slicks associated with tidally forced inter- nal waves may transport pelagic larvae of benthic inver- tebrates and fishes shoreward. Mar. Ecol. Prog. Ser. 13:311-315. 1985. Behavioural basis of intemal-wave-induced shore- ward transport of megalopae of the crab Pachygrapsus crassipes. Mar. Ecol. Prog. Ser. 24:289-295. Shenker, J. M . AND B L Olla. 1986. Laboratory feeding and growth of juvenile sablefish, Anoplopoma fimbria. Can. J. Fish. Aquat. Sci. 43:930-937. Sinclair, M. F.. M. J. Tremblay, and P. Bernal. 1984. El Nino events and variability in a Pacific mackerel {Scomber japonicus ) survival index: supj)ort for Hjort's second hypothesis. Can. J. Fish. Aquat. Sci. 42:602- 608. Small. L. F . and D W Menzies. 1981. Patterns of primary productivity and biomass in a coastal upwelling region. Deep-Sea Res. 28:123-149. Smith, P E 1972. The increase in spawning biomass of northern an- chovy, Engraulis mordax. Fish. Bull., U.S. 70:849-874. Steiner, W W., J. J. LuzcKovicH. and B L. Olla. 1982. Activity, shelter usage, growth and recruitment of juvenile red hake, Urophycis chuss. Mar. Ecol. Prog. Ser 7:125-135. Steiner, W W., and B L Olla 1985. Behavioral responses of prejuvenile red hake, Uro- phycis chuss, to experimental thermoclines. Environ. Biol. Fish 14:167-173. 317 WINTER-TIME DISTRIBUTION AND ABUNDANCE OF COPEPOD NAUPLII IN THE NORTHERN GULF OF MEXICO M J Dagg^, P B Ortner2, and F. Al-Yamani^ ABSTRACT Copepod nauplii were collected from continental shelf waters in 3 regions of the northern Gulf of Mexico during winters between 1981 and 1984, off Cape San Bias, Florida, off the Mississippi River delta, and off of Galveston, Texas. Some statistically significant iP < 0.05) patterns in the abundance and distribution of nauplii were observed: there was significant interannual variability in naupliar concentrations within the region around the Mississippi River delta; naupliar concentrations in the upper 10 m decreased in the onshore-offshore direction in 2 of 4 comparisons; naupliar concentrations in the upper 10 m differed regionally in 2 of 3 comparisons; and naupliar concentration was correlated with chlorophyll concentration in 4 of 5 comparisons. Maximum concentrations of nauplii (number per m^) within a water column were 2-10 times greater at stations influenced by the Mississippi River plumes than in the other 2 regions. This condition in attributed to vertical stratification imparted to the water column by the inflowing low salinity water from the Mississippi River. We conclude that the physical stratification provides a mechanism for the establishment of high concentrations of nauplii that otherwise would not exist in the winter months on the continental shelf. Microzooplankton are important diet items for larval fish (Arthur 1976; Gamble et al. 1981; Checkley 1982; Govoni et al. 1983; Houde and Lovdal 1984; Stoecker and Govoni 1984), and gut analyses indicate that copepod nauplii are fre- quently the dominant prey form found in the lar- vae of many fish species (Duka and Gordina 1973). Available concentrations of micro- zooplankton are considered an important deter- minant of larval survival rates in the ocean be- cause this relationship has been demonstrated in the laboratory (Laurence 1974; Houde 1978) and because field studies have demonstrated a rela- tionship between regions or periods of high mi- croplankton and high larval abundance (Arthur 1977; Lasker 1978). Because survival is enhanced by increased food availability, oceanographic processes that result in increased concentration or production of prey items are important. The gulf menhaden, Brevoortia patronus (Clu- peiformes), supports the largest volume fishery in the United States (U.S. Department of Commerce 1983). Spawning occurs in the wintertime, from October to March, in continental shelf waters of the northern Gulf of Mexcio (Fore 1970; Lewis and Rothmayr 1981; Warlen 1988), primarily off iLouisiana Universities Marine Consortium, Chauvin, LA 70344. ^Applied Oceanic Marine Laboratory, Ocean Coastal Divi- sion, NOAA, 4301 Rickenbacker Causeway, Miami, FL 33149. of Mississippi and Alabama to the east of the Mis- sissippi River delta, and ofi" of Louisiana to the west of the delta. During February and December 1982, concentrations of menhaden larvae in the region near the Mississippi River delta were greater in plume waters than outside and were much higher at the plume front (Govoni and Hoss"^). Furthermore, gut contents of these larvae indicated that nauplii were the most abundant prey items. Much of the continental shelf water of the northern and western Gulf of Mexico is vertically unstratified during the winter, and winter is also the season of minimum primary productivity in these regions, as it is in shelf waters of the north- east and southeast United States. However, coastal waters and nearshore regions influenced by freshwater inputs can be physically stratified during winter when low salinity plumes disperse over higher salinity shelf waters. We postulated that shelf waters influenced by fresh water from the Mississippi River plumes would be regions of increased production and concentration of micro- zooplankton. The purpose of this paper is to exam- ine this hypothesis by describing the vertical and horizontal distribution and abundance of copepod nauplii in shelf waters of the northern Gulf of Manuscript accepted December 1987. FISHERY BULLETIN: VOL. 86, NO. 2, 1987. ^Govoni and Hoss. Unpubl data. Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Service, NOAA, Beaufort, NC 28516. 319 FISHERY BULLETIN: VOL. 86, NO. 2 Mexico and by comparing these with oceano- graphic and hydrographic indicators of water column stratification and productivity. METHODS Samples were collected during 5 winter cruises between November and March, during 1981-84, (Table 1). Microplankton samples and hydro- graphic data were collected from 3 general re- gions of the northern Gulf of Mexico: off of Galve- ston, TX, in the region of the Mississippi River delta, and off of Cape San Bias, FL (Fig. 1). Sam- ples were not collected from all 3 regions on each cruise. Water samples for chlorophyll and nutrient analyses were collected with Niskin bottles. For each chlorophyll analysis, between 25 and 150 mL of sea water was filtered onto a GF/F or GF/C glass fiber filter and homogenized by grinding in 90% aqueous acetone. Fluorescence of the filtrate, brought up to a volume of 10.0 mL, was deter- mined before and after acidification with 2 drops of 10% HCl using a Turner Designs'' Model 10 fluorometer. Chlorophyll and pheopigment con- centrations as chlorophyll equivalents were de- termined from chlorophyll (fxg/L) = Table 1.— Station information, 1981-84. Kifo- fa) V KiRfa -fo) pheopigment (|JLg/L) where K is the machine calibration constant, [„ and fa are the fluorescence readings before and after acidification, R is the acid ratio, and u is the volume of seawater filtered, in mL (Strickland and Parsons 1968). Samples for nutrient analyses were frozen. Ni- trate and nitrite were analyzed according to method number 353.2 described by EPA publica- tion number EPA 600/4-79-020 (Environmental Protection Agency 1979). Temperature and salinity were measured by several methods. During cruise I, temperature was measured by the temperature sensor on the MOCNESS^ net, and salinity was measured with a refractometer. During cruises II and III, tem- perature was measured with expendable bathy- Depth Lat. Long. Station Date Time (m) (N) (W) 1-1 2-13-81 0000 18 29 '05' 94°07' 1-2 2-13-81 1200 18 2906' 94' 07' 1-3 2-14-81 0001 90 28 04' 93"03' 1-4 2-14-81 1200 90 28 04' 93"03' 1-5 2-15-81 0000 180 27' 54' 92'^51 ' 1-6 2-15-81 1200 180 25' 54' 92'50' 1-7 2-16-81 0001 180 28 34' 89"37' 1-8 2-17-81 1200 180 28 35' 89 38' 1-9 2-18-81 0001 90 2850' 89 "16' 1-10 2-18-81 1200 90 28 50' 89 "16' 1-11 2-19-81 0000 18 28^56' 89'29' 1-12 2-19-81 1200 18 28°56' 89'^29' 1-13 2-20-81 0000 90 29 09' 85' 56' 1-14 2-21-81 1200 90 29 08' 85"56' 1-15 2-22-81 0000 180 29°28' 8607' 1-16 2-22-81 1200 180 29°28' 86°07' 1-17 2-23-81 0001 18 29°36' 85°47' 1-18 2-23-81 1200 18 29^35' 85 47' 1-19 2-24-81 0001 18 30°12' 87"06' 11-1 12-05-82 0700 25 28°53' 89' 29' 11-2 12-05-82 1255 62 28°53' 89^32' 11-3 12-05-82 1945 33 28' 59' 89°34' 11-4 12-06-82 2120 887 28 20' 89°27' 11-5 12-07-82 0715 885 28°20' 89'27' 11-6 12-07-82 1220 834 28' 19' 89 25' 11-7 12-08-82 1820 443 28°32' 89°53' 11-8 12-09-82 0830 402 28°34' 89°53' 11-9 12-09-82 1215 461 28^32' 89 '53' 11-10 12-13-82 1245 44 29'26' 85°53' 11-11 12-15-82 1930 19 28°54' 89°29' 11-12 12-16-82 0900 68 28°51' 89°32' III-1 11-19-83 1200 27 28°54' 89°29' III-2 11-20-83 0800 35 28°47' 89°59' III-3 11-20-83 1946 732 28°38' 89°00' III-4 11-21-83 0555 27 28°54' 89°30' III-5 1 1 -22-83 1100 48 28°51' 89°30' III-6 1 1 -23-83 0500 194 28°31' 89'=37' III-7 1 1 -24-83 1030 50 28=50' 89°31' III-8 11-25-83 0820 45 28^55' 89°34' III-9 11-25-83 1945 40 28°56' 89°36' 111-10 11-26-83 1200 16 28°54' 89°29' 111-11 11-27-83 0900 9 29'02' 89°30' 111-12 11-28-83 0840 29 28^48' 89^57' 111-13 11-28-83 1250 44 28°56' 89°51' 111-14 11-29-83 0950 24 29''03' 89°39' 111-15 1 1 -30-83 0930 44 28°47' 89°58' 111-16 12-01-83 0800 38 28°52' 89°29' IV- 1 3-14-83 53 27' 58' 95°53' IV-2 3-19-83 59 28=18' 90^41 ' V-1 2-21-84 0850 18 28 54' 90°25' V-2 2-22-84 0830 18 28' '54' 90°25' V-3 2-23-84 0910 18 28°54' 90°25' V-4 3-1-84 0835 18 28°54' 90°25' V-5 3-1-84 0830 18 28^54' 90°25' V-6 3-3-84 0730 18 28'54' 90°25' '^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. ^Multiple opening-closing net and environmental sensing system. thermographs (Sippican Instruments). Salinity was measured with a YSI Model 33-S-C-T sali- nometer on cruise II, and with a Beckman Model RS5-3 salinometer on cruise III. For both of these cruises, bottle samples, analyzed in the labora- tory with a Guildline Model 8400A Autosal, were used to check the shipboard salinity measure- 320 DAGG ET AL.: DISTRIBUTION OF COPEPOD NAUPLII c o c o C/2 ca D O 321 FISHERY BULLETIN: VOL, 86, NO. 2 ments. On cruise IV, a Plessey Model 9040 CTD was used for both temperature and salinity. On cruise V, salinity samples were stored in bottles and analyzed in the laboratory using a Guildline Model 8400A Autosal, and temperature was measured with reversing thermometers. Microzooplankton were collected using 5 L (cruises I, II, and V) or 30 L (cruises III and IV) Niskin bottles. During cruise I, 1 L samples of untreated water were collected from 3 depths and preserved for later analysis. During subsequent cruises, samples were collected from up to 8 depths; four liters were poured gently through a 20 |xm sieve, backwashed with filtered seawater into a sample jar, and preserved in a 5% formalin- seawater solution. In the laboratory, all organ- isms in each preserved sample were identified and counted. In selected samples, the length and width of the first 50 copepod nauplii were mea- sured using an ocular micrometer. In addition, during cruise I larger nauplii were collected with 0.1 m^ 64 |jLm mesh nets nested inside the 1 m^ 333 ^JLm mesh nets of a MOCNESS net (Wiebe et al. 1976). Estimates of naupliar abundance from these collections were used separately in the analyses for patterns in distribution and abun- dance. Lastly, from cruise I, the total number of copepods in the shallow-water samples, total number of copepods from the 333 jxm nets, and the zooplankton displacement volume from the 333 (xm nets were analyzed for patterns in distribu- tion and abundance. RESULTS Samples from the Mississippi River delta re- gion, collected during 1981 (cruise I), 1982 (cruise II), and 1983 (cruise III), were compared to deter- mine if there was significant interannual vari- ability within a region. Whole water samples and samples retained on 20 ixm mesh sieves were pooled for this analysis under the assumption that nauplii in a whole water sample would be retained on a 20 jjim mesh screen. The 64 [im net samples were excluded from this analysis. This test was made using average naupliar concentra- tions in the upper 10 m only; naupliar concentra- tions were typically lower in deeper waters, and we did not have the same number of deep and shallow stations for each of the 3 sampling years. A 1-way analysis of variance indicated significant variation between years within the Mississippi River delta region {P <0.01). All 3 years were statistically different. These results indicated to us that regional or other spatial comparisons can only be made using samples collected during the same year. Interannual comparisons for other re- gions were not possible. Samples were collected from all 3 regions dur- ing cruise I. Four tests for regional differences during this cruise were made. First, regions were compared based on naupliar concentrations in the upper 10 m only (Table 2); based on whole water collections F was significant at P < 0.05 and based on samples collected with 64 \xm mesh nets F was significant at P < 0.09. Second, regions were compared based on naupliar concentrations from all depths (Table 2); based on whole water collections F was significant at P < 0.01 and based on samples collected with 64 (xm mesh nets F was significant at P < 0.22. Thus, whole water samples collected on cruise I indicated significant regional differences in naupliar concentration but samples collected with 64 fxm mesh nets did not. In addition, there were significant regional differences in the concentrations of total copepods (P < 0.01) and in zooplankton displacement vol- ume (P < 0.05) during cruise I (data not shown). Regional comparisons were not possible during other cruises because all 3 regions were not sam- pled in years other than 1981. Where possible, we also tested for onshore- offshore gradients in naupliar concentrations in the upper 10 m. Stations were categorized accord- ing to bottom depth as shallow, <18 m, interme- diate, 18-90 m, and deep, >90 m, and compari- sons were made between these 3 depth categories. A highly significant onshore-offshore difference existed during cruise I for nauplii collected with the 64 |xm MOCNESS nets (Table 3). During cruise III samples were only collected in the Mis- sissippi River delta region but onshore-offshore differences within this region were significant (Table 3). No significant onshore-offshore gradi- ents were observed in naupliar concentrations de- termined from whole water samples collected from all 3 regions during cruise I, or in nauplii from only the Mississippi River delta region dur- ing cruise II (Table 3). Although regional differences in average con- centrations of nauplii were not strong, the verti- cal distributions were different. For example, sev- eral stations at approximately the 50 m isobath are compared in Figure 2. At station IV- 1 (Fig. 2a) ofT the coast of Texas, there was no marked vertical heterogeneity in naupliar concentra- tions, the maximum concentration was low (23 nauplii/L), and the average concentration was 322 DAGG ET AL.: DISTRIBUTION OF COPEPOD NAUPLII Table 2. — Regional comparisons of naupliar concentrations during cruise I. Samples were collected by 2 methods: whole water sam- ples were collected with Niskin bottles, and 64 (i.m mesh nets were used to collect the net caught samples. Cruise 1 Depths Sample type Probability of F upper 10 m whole water 28 <0.05 upper 10 m 64 ^m nets 126 <0.09 all depths whole water 57 <0.01 all depths 64 |jLm nets 222 <0.22 Table 3. — Onshore-offshore comparisons of naupliar concentra- tions from the upper 10 m of each station. Tests for cruise I are based on naupliar concentrations from all 3 regions pooled then separated into 3 depth categories. During cruises II and III. nauplii were only collected from the Mississippi River delta region. Cruise I Regions Sample type Probability of F all whole water 28 <0.14 all 64 ^m nets 126 <0.01 Mississippi R. 20 jjim sieve 28 <0.25 Mississippi R. 20 M-m sieve 49 <0.01 NAUPLIUS CONCENTRATION (number/ liter) Q. UJ Q I »- Q. UJ Q c 25 50 75 1, . 1 1 10- ^ / 4 1 If 1 1 1 30- * ^ 50- Hi 112 n3 (0 Figure 2. — Naupliar concentrations (number/liter) from selected stations near the 50 m isobath on the Texas continental shelf (a, station IV-1), the Louisiana Shelf (b, station IV-2, c, stations II-l, II-2, II-3), and the northwest Florida shelf (d, station 11-10). 323 FISHERY BULLETIN: VOL. 86, NO. 2 low (11 nauplii/L). The same pattern was ob- served off Cape San Bias (Fig. 2d); there was no marked vertical heterogeneity, and the maxi- mum and average concentrations were low, 14 and 12 nauplii/L. In contrast, nauplii at stations near the 50 m isobath in the region near the Mis- sissippi River delta typically showed marked ver- tical heterogeneity in abundance. At station IV-2 (Fig. 2b), the maximum concentration was 187 nauplii/L and the minimum was 24 nauplii/L. At station II- 1, II-2, and II-3 in the Mississippi River delta region (Fig. 2c) naupliar abundances were also vertically heterogenous, although concentra- tions were lower than at Station IV-2. In general, based on all our vertical profiles (most not shown), regions influenced by the Mississippi River can contain high concentrations of nauplii in the surface layer, while subsurface concentra- tions are similar to concentrations in the other 2 regions. Vertical structure in the distribution of nauplii appears to be related to physical structure. At the stations off the coasts of Texas and Florida (Fig. 2a, 2d), temperture and salinity were essentially vertically homogenous (Fig. 3a, 3d) whereas off the Mississippi River delta temperature and 20%. 24 28 32 36 24%. 28 16*0 18 Figure 3. — The vertical distribution of temperature (•) and salinity (x) at the stations represented in Figure 2. 324 DAGG ET AL.: DISTRIBUTION OF COPEPOD NAUPLII salinity were vertically stratified (Fig. 3b, 3c). The same pattern was seen in the chlorophyll a distributions (Fig. 4) and the nitrate distributions (Fig. 5). Correspondence between chlorophyll a concen- tration and naupliar abundance was computed; the correlation coefficient for a linear regression of chlorophyll concentration on naupliar abun- dance for all 5 cruises in = 269) was 0.39, signifi- cant at the 1% level. However, analysis of covari- ance showed that the relationship between chlorophyll and naupliar concentration was sig- nificantly different between cruises, and there- fore the data from the 5 cruises should not be pooled. Although a significant relationship be- tween chlorophyll concentration and naupliar concentration typically exists, it is not consistent from cruise to cruise. Regression parameters and correlation coefficients for each cruise are shown in Table 4. Pairwise tests indicate that the slope of the regression for cruise V was significantly different from all the others, cruise IV was differ- 0 I 0 • CHLOROPHYLL a (ug/liter) 2.0 3 0 4.0 10 2.0 Figure 4. — The vertical distribution of chlorophyll a at the stations represented in Figure 2. NITRATE (ug-at/liter) Figure 5. — The vertical distribution of nitrate at the stations represented in Figure 2. 325 FISHERY BULLETIN: VOL. 86, NO. 2 Table 4. — Parameters and coefficients for linear regressions of chlorophyll concentration on nauplius concentration for each cruise. Y = a + b(X) where y = naupliar concentration as number per liter, and X = chlorophyll concentra- tion as fjig per liter. Significance level tested is 0.01. Significance Significance of Cruise n Intercept Slope of slope /-2 correlation 1 57 12.9 9.7 <0.01 0.40 <0.01 II 70 5.4 15.4 -0.01 0.53 <0.01 III 80 3.7 9.2 <0.01 0.34 <0.01 IV 20 6.9 39.4 <0.01 0.94 <0.01 V 42 26.1 1.1 NS 0.04 NS ent from all others, cruise II was different from all the others, and cruises I and III were not dif- ferent from each other but different from cruises II, IV, and V. The slopes of all relationships ex- cept cruise V were significantly different from zero. The size-frequency distribution of nauplii did not vary in any systematic manner (Fig. 6). For example, at station IV-1, the bulk of the nauplii were between 20 and 100 \x.m at all depths sam- pled during both the daytime and nighttime peri- ods. The size-frequency distribution was essen- tially identical at station IV-2, off Louisiana (data not shown). At some stations, copepodid stages <600 ixm in length (50-200 \xm in width) were counted in ad- dition to nauplii because they are potential prey items for menhaden larvae. Copepodites were usually not as abundant as nauplii, especially at stations with high naupliar abundances. On occa- sion, they were as abundant or, at specific depths, more abundant than the nauplii (Table 5). The ratio of nauplii to copepodites varied widely, be- tween 16.3 and 0.3, so it is not possible to assign a constant factor to naupliar abundances to esti- mate the increase in available prey attributable to copepodid stages. Other microzooplankton were usually not as abundant as nauplii or copepods. In our samples, various forms of eggs reached a maximum density of 6 eggs/L (cruise III station 6 at 3 m, data not shown). Pelecypod larvae at one station (III-ll) were abundant, reaching a maximum of 29 lar- vae/L at 4 m, compared with 22 nauplii, 7 cope- podites, and 2 eggs/L in the same sample. Only on this one occasion were organisms other than cope- podites or nauplii the dominant form of micro- zooplankton. Table 5. — Abundance (number/liter) of other microzooplankton in addition to nauplii at selected stations during cruise III. Station Depth (m) Nauplii Cope- podids <600^m Cope- podids >600 ii.vn Larva- ceans Eggs Other III-3 6 17.3 7.5 1.0 7.8 0.5 1.8 9 29.3 12.0 0.5 11.0 0 1.8 12 17.5 8.0 1.0 6.8 1.5 2.3 20 7.8 2.8 0 1 0 2.0 0.5 III-7 0 24.0 6.3 4.0 0 1.0 1.0 3 15.3 12.3 1.0 0 0.3 1.3 5 3.0 5.0 1.0 0 0.8 0.3 13 4.8 5.8 0.3 0 0 0.5 22 2.0 1.0 0.3 0 0 0 111-14 0 15.8 12.0 1.3 0 1.3 2.0 5 44.8 3.8 0.3 0.5 3.0 4.0 10 40.0 6.8 0.8 1.3 4.3 0.3 15 11.5 8.0 3.8 1.0 1.8 0.3 20 12.0 35.0 6.8 0.3 0.3 0.8 111-16 0 15.8 25.3 5.0 0 0 12.8 6 8.0 16.0 2.3 0 1.0 2.5 9 30.0 14.0 2.0 0 0.3 2.5 12 29.0 13.3 0.5 0 0.8 1.3 15 17.0 12.5 1.0 0 0.8 0.5 25 31.8 12.8 0.8 0 0 5.0 326 DAGG ET AL.: DISTRIBUTION OF COPEPOD NAUPLII NIGHT 20- ^i=H -T 1 DAY 0 m 20 -1 1 1 I 20- 0^— -^ 20 -I 1 1 1 52 ^ 0 -I 1 1 20-1 E 13 C o LlI o :z. o o CO _J CL 0 20- 0 20 0 20n r~| . somple lost -1 1 1 1 20-1 n -I 1 r 20- :i-^ 20 5m X4- lOm 15m 20 m 25m 30m 20n 0 20- 0 20- 0 20- 0 20- 0 20- ^aXL 5=L r. 35m -I ' 1 -T r ■! 1 1 I o o o o o o o ^ CD ^ (I> O ^ BODY WIDTH (mm) 40m 20 =F=L o o o o o o o •3- CD oj i£ o ^ BODY WIDTH (mm) 45 m Figure 6.— The size-frequency distribution of nauplii from station IV- 1. 327 FISHERY BULLETIN: VOL. 86, NO. 2 DISCUSSION Menhaden larvae between 3 and 20 mm long collected during wintertime from 3 regions in the northern Gulf of Mexico contained a variety of diet items, including dinoflagellates, tintinnids, copepod eggs, nauplii, copepodids, juvenile pele- cypods, and pteropods (Govoni et al. 1983, 1985). Dinoflagellates and tintinnids constituted the main diet items of larvae <5.0 mm long but were replaced by copepod nauplii, then copepodids and small adult copepods as larvae grew larger. In another study, Stoecker and Govoni (1984) found that copepod nauplii were the dominant items in the diet of menhaden larvae 7.0-9.0 mm long, although only 12 fish were examined. Later studies, however, have verified that nauplii are the dominant items in the diet of a wide size range of larval menhaden collected in the boundary between the plume of the Mississippi River and oceanic waters (Govoni and Hoss fn. 3) The maximum width of food items in menhaden larvae <5 mm long was about 200 |xm (Govoni et al. 1983). This width increased up to about 400 (xm for larvae 10 mm long. In another study (Stoecker and Govoni 1984) maximum prey size was calculated in a different manner; the average width of the largest prey type was used. With this index the estimated maximum width of prey items was 50 ^JLm for larvae <5 mm long and about 140 |jLm for larvae 9-11 mm long. In our study, most of the nauplii were between 40 and 80 |jLm body width, and most of the copepodites were between 50 and 200 p-m body width. These cope- pod developmental stages were in the size range of diet items typically found in guts of intermedi- ate size (7-11 mm) larvae of the gulf menhaden and we believe our samples are reasonably repre- sentative of the prey concentrations available to these larvae. Other prey items for larvae in this size range were common only on rare occasions. The importance of high prey concentrations to successful feeding and survival offish larvae has frequently been noted (Hunter 1981). Concentra- tions required to give high survival in laboratory experiments are seldom found in the ocean but careful attention to culture techniques can result in reasonably high survival at prey densities that are close to or overlap the maximum natural con- centrations (Houde 1978). At stations near the river delta that are strongly affected by the river plumes, maximum concentrations of nauplii were typically in the range of 20-50 nauplii/L. At sta- tions farther down plume, away from the delta. maximum naupliar concentrations were higher, up to 187 nauplii/L at station IV-2 for example. It is not clear whether this is a seasonal pattern, a pattern due to the down-plume development of the food web, or a pattern attributable to some other factors. However, the concentrations ob- served in waters with salinities lowered by the river plumes are mostly within the range of 10- 100 nauplii/L that is frequently reported for oceanic and coastal waters for microzooplankton, or the 50-100 nauplii/L range for coastal and es- tuarine areas (see Houde 1978; Hunter 1981 for summaries). The addition of copepodite stages and other potential prey items to the naupliar abundances in this study would increase the esti- mates of available prey somewhat but usually not more than 20%. At least during the wintertime, the waters surrounding the Mississippi River delta do not appear to contain exceptionally high concentrations of copepod nauplii, compared to other coastal and estuarine areas. The vertical distribution of nauplii was fre- quently similar to that of chlorophyll. Regres- sions of naupliar concentration on chlorophyll concentration were significant for 4 of the 5 cruises although not the same in each case. As- suming chlorophyll is a reasonable indicator of phytoplankton abundance and thus food abun- dance for nauplii, then nauplii appear to be ag- gregated at the depth of highest food availability. Year to year variability in the relationship indi- cates that other factors are also important in de- termining the abundance and distribution of nau- plii. Because eggs released from adult female copepods would sink out of the surface water be- fore hatching, it is probable that active swimming by nauplii plays a part in the aggregation process, perhaps enchanced by physical convergence proc- esses. In our study we noted that there were interan- nual differences in naupliar concentrations, there were sometimes significant regional differences, and there were sometimes significant onshore- offshore differences. Although we anticipated finding higher concentrations of nauplii in the region near the Mississippi River delta, this pat- tern was not always observed. This finding might indicate that there is not an important difference between the 3 regions as far as larval food availability is concerned. Alternatively, it might indicate that there are subregions within the larger Mississippi River delta region that contain a food environment for fish larvae that allows for enhanced survival and growi;h, but that we failed 328 DAGG ET AL.: DISTRIBUTION OF COPEPOD NAUPLII to clearly identify and segregate these regions. The overall region did not always appear signifi- cantly better than the other regions because only parts of it are better. On the continental shelf in the Gulf of Mexico, as in other shelf regions of the United States, the typical pattern is for the water to be vertically well-mixed during the wintertime (Parker 1968). Autumn cooling breaks down the thermal strati- fication that has existed throughout the summer and allows the isothermal water column to be easily mixed. The major exception to this general pattern in the Gulf of Mexico is the shelf region influenced by the large volume of freshwater runoff from the Mississippi River. During this pe- riod, the freshwater influx is of sufficiently large volume (average flow for December to March in 1975 through 1979 was 17,290 m^/second (U.S. Army Engineer District, New Orleans Corps of Engineers, 1980)) to physically stratify the shelf waters hundreds of km downstream from the delta. We suggest that this salinity-induced strat- ification is a vital component of the recruitment success of Gulf menhaden because it provides an environment in which prey aggregations can occur. It has been suggested that vertical stratifi- cation by other small-scale physical phenomena (e.g., Langmuir circulation) allows significant patchiness of prey items to exist (Lasker 1975). This patchiness provides small regions of compar- atively high food concentrations for fish larvae, and results in improved feeding success. In this study, copepod nauplii were aggregated vertically at stations with physical stratification and were nearly homogenously distributed at stations lack- ing physical stratification. Maximum naupliar concentrations (no. per m^) at stratified stations were typically 2-10 times greater than at non- stratified stations. In conclusion, we believe that the large fresh- water inflow of the Mississippi River during the wintertime spawning period of the Gulf men- haden contributes to the feeding success and survival of larval fish by providing physical strat- ification which in turn results in a vertical strat- ification of phytoplankton and microzooplankton in layers or patches of relatively high concentra- tions. ACKNOWLEDGMENTS Part of this work was done in cooperation with personnel from the National Marine Fisheries Service at Beaufort, NC and at Pascagoula, MS. In particular, we would like to thank D. Hoss for providing time and space in his field programs, and J. Govoni for discussion of several portions of the manuscript. S. Cummings, L. Hill, E. Walser, J. White, J. Turner, P. Morgan, C. Neill, and E. Turner assisted in collecting and analyzing samples. Part of this work was done in cooperation with D. Checkley and A. Amos at the University of Texas; their contributions are greatly appreciated. Thanks also to the captains and crews of the F.R.V. Oregon II, the RV Longhorn, and the R. J. Russell. This work was supported by the Louisiana Universities Marine Consortium, by NSF Grant OCE-81 19848, and by the Ocean Assessment Division, National Ocean Services, NOAA. LITERATURE CITED Arthur, D. K 1976. Food and feeding of larvae of three fishes occurring in the California Current, Sardinops sagas, Engraulis mordax, and Trachurus symmetricus. Fish. Bull., U.S. 74:517-530. 1977. Distributions, size, and abundance of microcopte- pods in the California Current system and their possible influence on survival of marine teleost larvae. Fish. Bull., U.S. 75:601-611. Checkley, D M, Jr. 1982. Selective feeding by Atlantic herring (Clupea harengus ) larvae on zooplankton in natural assem- blages. Mar. Ecol. Prog. Ser. 9:245-253. DUKA, I A , AND A. D GORDINA. 1973. Abundance of ichthyoplankton and feeding of fish larvae in the western Mediterranean and adjacent areas of the Atlantic Ocean. Hydrobiol. J. 9:54-59. Environmental Protection Agency. 1979. Methods of chemical analysis of water and wastes. Nitrogen, nitrate-nitrate. U.S. Environ. Prot. Agency, Cincinnati, OH 45268. Fore, P L 1970. Oceanic distribution of eggs and larvae of the Gulf menhaden. Rep. Bur. Comm. Fish. Biol. Lab., Beaufort, NC, Fish. Wildl. Serv. Circ. 341. Gamble, J C , P Maclachlan, N T Nicholl. and I. G. Baxter. 1981. Growth and feeding of Atlantic herring larvae reared in large plastic enclosures. Rapp. P. -v. R6un. Cons. int. Explor. Mer 178:121-134. Govoni, J J , A J Chester. D E Hoss, and P B Ortner 1985. An observation of episodic feeding and growth of larvae Leiostomus xanthrus in the northern Gulf of Mex- ico. J. Plankton Res. 7:137-146. 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 xanthrus, and Micropogonias undula- tus. Mar. Ecol. Prog. Ser. 13:189-199. Houde, E D 1978. Critical food concentrations for larvae of three spe- cies of subtropical marine fishes. Bull. Mar. Sci. 28:395-411. 329 FISHKRY BULLKTIN: VOL. 86, NO 2 HOUDE. E D . AND J. A LOVDAL. 1984. Seasonality of occurrence, foods and food prefer- ences of ichthyoplankton in Biscayne Bay, Florida. Estuarine Coastal Shelf Sci. 18:403-419. Hunter. J R 1981. Feeding ecology and predation of marine fish lar- vae. In R. Lasker (editor). Marine fish larvae. Morphology, ecology, and relation to fisheries, p. 34- 77. Univ. Wash. Press, Seattle. Laurence, G C 1974. Growth and survival of haddock (Melanogrammus aeglefinus) larvae in relation to planktonic prey concen- tration. J. Fish, Res. Board Can. 31:1414-1419. 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 failure. Rapp. P.-v. Reun. Cons. int. Explor. Mer 173:212-230. Lewis, R. M., and C M Rothmayr 1981. Spawning and sexual maturity of gulf menhaden, Brevoortia patronus. Fish. Bull., U.S. 78:947-951. Parker. C A 1968, The effect of a cold air outbreak on the continental shelf waters of the northwestern Gulf of Mexico. Tech. Rep. 68-3T, 89 p. Off. Nav. Res. Stoecker. D K , and J J Govoni. 1984. Food selection by young larval gulf menhaden (Bre- voortia patronus). Mar. Biol. 80:299-306. Strickland, J D H , and T R. Parsons. 1968. A practical handbook of seawater analysis. Bull. Fish. Res. Board Can. 167, p. 201-203. US Department of Commerce 1983. Fisheries Statistics for the United States, 1982. US Army Engineer District, New Orleans Corps of Engi- neers. 1980, Stages and discharges of the Mississippi River and its tributaries and other watersheds in the New Orleans District, 1975-1979. Warlen, S. M. 1988. Age and growth of larval menhaden, Brevoortia pa- tronus, in the northern Gulf of Mexico. Fish. Bull., U.S. 86:77-90. 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. 330 THE EFFECTS OF SILTATION ON RECRUITMENT OF SPINY LOBSTERS, PANULIRUS ARGUS William F Herrnkind.' Mark J. Butler IV,' and Richard A. Tankersley^ ABSTRACT Several surveys in the Florida Keys indicated fewer juvenile spiny lobsters, Panulirus argus, in an area where their primary habitat, stands of benthic algae Laurencia spp., was heavily silted as compared with similar, less silted habitat. We tested several hypotheses explaining this relationship: 1) planktonic postlarval lobster abundances are lower in the silted area, 2) siltation of algae impedes postlarval settlement or subsequent juvenile habitat selection, or 3) siltation increases mortality at the time of metamorphosis. We also compared the time-to-metamorphosis for settling pueruli within silted and nonsilted algae, analyzed the physical character of algal silt in low-silt and high-silt regions and measured the abundances of epifauna constituting prey of juvenile spiny lobsters. Plank- tonic postlarval abundances were substantially higher in the high-silt area thus rejecting hypothe- sis 1. Likewise, results from laboratory experiments testing the effect of algal siltation on postlarval time-to-metamorphosis and early postsettlement survival showed no short-term increase in mortal- ity. Limited postlarval settlement and avoidance of silted algal habitats by juveniles, as determined in substrate choice experiments, probably accounts for the paucity of young spiny lobsters in heavily silted localities. In addition, although juvenile spiny lobsters are nonselective predators, lower prey availability in silted algae probably promotes transciency which, in turn, causes increased mortality by predation while juveniles are exposed. Large-scale siltation exacerbated by human activity must be viewed as potentially deleterious to spiny lobster recruitment. The western Atlantic or Florida spiny lobster, Panulirus argus, is the focus of an intense com- mercial and recreational fishery in south Florida, particularly the Florida Keys. Besides severe fishing pressure, spiny lobster populations are subject to a variety of other factors that poten- tially limit population size. For example, habitat degradation, like that resulting from chronic sil- tation, may affect not only adult lobsters but the postlarval settlement stage as well. During 1983 and 1984 we sampled numerous sites in a region of about 40 km^ east of Big Pine Key which was chronically heavily silted and held low numbers of newly settled spiny lobsters despite extensive benthic algal growth typical of settlement habi- tat. We hypothesized that postlarval spiny lob- sters either do not settle in silted habitat or settle there but do not survive. In either case, we sup- posed that the heavy siltation reduced the carry- ing capacity of otherwise suitable habitat, poten- tially reducing regional recruitment where siltation is widespread. •Department of Biological Science, The Florida State Univer- sity, Tallahassee, FL 32306-3050. ^Department of Biological Science, The Florida State Univer- sity, Tallahassee, FL; present address: Department of Biology, Wake Forest University, Winston-Salem, NC 27109. Sediment particle size, composition, and stabil- ity influence larval settlement in a variety of marine benthic invertebrates (Crisp 1974, 1976; Gray 1974; Rhoads 1974; Pearson and Rosenberg 1978). For example, heavy siltation and sediment instability, created by natural biogenic rework- ing of the substrate and subsequent resuspension of sediments by turbulence reduces the abun- dance of suspension feeding infauna (Rhoads and Young 1971; Aller and Dodge 1974). In addition, siltation from human activities (e.g., dredging, shoreline development, boat traffic, etc.) can de- grade benthic community structure via anaero- biosis, direct burial, toxic poisoning, or increased turbidity (Morton 1977; Allen and Hardy 1980; Jones and Candy 1981; Cortes and Risk 1985). Most available information concerns sessile or in- faunal species, but little information exists for mobile, epibenthic forms (Pearson and Rosenberg 1978). Although the habitat selection and bur- rowing behavior of some shrimps and juvenile clawed-lobsters has been investigated in relation to substrate character (Ruello 1973; Howard and Bennett 1979; Aziz and Greenwood 1982; Botero and Atema 1982; Pottle and Elner 1982; Roach 1983; Herrnkind and Butler 1986), we know of no research describing the effect of siltation on deca- Manuscnpt accepted February 1988. FISHERY BULLETIN: VOL. 86, NO 2, 1988. 331 FISHERY BULLKTIN: VOL. 86, NO. 2 pod recruitment. Our research focuses on the im- pact of siltation on spiny lobster postlarvae and early benthic juveniles, stages that are morpho- logically and behaviorally distinct from adults. Late stage P. argus phyllosome larvae drift in the oceanic plankton for 6-9 months after hatch- ing and metamorphose offshore into nonfeeding pueruli (postlarvae) that swim inshore and settle in benthic vegetation (Marx 1986). Nev^ly settled pueruli metamorphose into cryptically colored benthic juvenile instars after about one week. Pueruli preferentially settle in highly architec- tured benthic algal assemblages where subse- quent survival and growth depend upon available prey and physical refuge from predators (Herrnkind and Butler 1986). Ubiquitous, widely distributed stands of bushy red algae, Laurencia spp., provide these essential conditions and prob- ably serve as the most important regional settle- ment and nursery habitat for juvenile spiny lob- sters (Marx and Herrnkind 1985a, b; Herrnkind and Butler 1986; Marx 1986). The early instars remain within the algae for several months until attaining about 20 mm carapace length (CL) (Andree 1981; Marx and Herrnkind 1985a) when they begin to occupy crevices in rubble or under sponges, coral, and exposed seagrass rhizome mats. In Florida, postlarval settlement is year- round with vernal, autumnal, newmoon, and oc- casionally aperiodic peaks (Little 1977; Little and Milano 1980; Marx 1986). The spatial pattern of settlement is poorly known although new recruits are widely dispersed within algal habitats; diver surveys have yielded estimates of one juvenile per 36 m^ of profuse algal growth (Marx and Herrn- kind 1985a). Yet because postlarvae settle con- tinuously and juveniles grow rapidly, a single hectare of the above habitat is estimated to nur- ture about 1,000 spiny lobsters annually (Marx and Herrnkind 1985a; Marx 1986). There is no compelling evidence suggesting that benthic stage lobsters immigrate into Florida waters from other Caribbean areas, although their planktonic larvae presumably do so (Lyons 1980; Marx 1986). Recruitment is thus primarily limited to postlarval influx. Therefore, precise knowledge of the factors influencing postlarval recruitment and recruit mortality is essential to managing the intensive Florida spiny lobster fishery. Here we report on studies undertaken to inves- tigate the impact of the observed algal siltation on spiny lobster recruitment. We compared field abundances of both pueruli and early juveniles in a representative silted and unsilted area, exam- ined the relationship between siltation and avail- able epifaunal prey, determined the impact of silt load on puerulus survival from settlement through metamorphosis, and tested the prefer- ence of settling pueruli and algal-dwelling juve- niles for silted and unsilted algae. METHODS Postlarval-Juvenile Abundance in Silted and Unsilted Habitats During June through August 1985, we com- pared the natural abundance of newly settled ju- venile spiny lobsters (6-20 mm CL) in previously sampled, chronically silted and unsilted areas. The silted site (No Name Key) was located ap- proximately 30 m off the western shore of No Name Key (Monroe County, FL, U.S.A.) and the unsilted site (Burnt Point) 30 m off the northwest shore of Grassy Key at Burnt Point (Fig. 1). The benthic habitat at both sites was similar and characterized by nearly contiguous stands of algae (Laurencia spp.) at depths of 1.5-3.0 m. In- tensive visual search in algal clumps by divers was used to estimate the relative number of ben- thic juveniles. Because newly settled spiny lob- sters are almost exclusively found associated with algal clumps (Marx and Herrnkind 1985a), catch per unit effort (CPUE) as search time within algae, gives a more suitable estimate for our pur- poses of comparing abundance than density/area per se. Modified Witham-type postlarval collec- tors (Witham et al. 1964, 1968; Little and Milano 1980; Marx and Herrnkind 1985a) were used to compare postlarval abundance among sites. Twelve collectors were initially deployed at both sites and visited approximately every 2 weeks for 3 months. Collector results are reported in CPUE to standardize catch records biased by the loss of collectors and different sampling durations. Algal Silt Content and Prey Content To determine the amount of silt and macro- fauna contained in algal clumps at the silted and unsilted sites, we bagged individual clumps (ap- proximately 25 cm diameter) of Laurencia in the field for subsequent laboratory processing. Care was taken to ensure that loose silt present on the surface of the algae was not disturbed during col- lection. Ten clumps, ranging in displacement vol- ume from 55 to 300 mL, were collected at each site. Algal samples were rinsed through a series 332 HERRNKIND ET AL RECRUITMENT OK Sl'lNY LOBSTERS Florida Keys ^ -r«*. ^B'' LITTLE PINE KEY / NAME KEY r BAHIA ' HONDA KEY ^NEWFOUND HARBOR KEYS BURNT POINT GRASSY KEY Figure l. — Map of field sites in the middle and lower Keys, Monroe County, FL, U.S.A. Insets provide more detail of the areas surrounding our field sites (•). Diagonal lines depict approximate extent of the heavily silted area we surveyed. of sieves (500 |x, 250 |jl, and 63 jx; U.S. Standard Sieve Series), but only the two smallest size frac- tions were retained because subsamples >500 |x consisted entirely of shell and algae fragments. Silt samples were dried for 48 hours at lOO'C and then weighed. The amount of algal-entrained silt at the two sites was compared using a two-sample ^-test. Organic weight of the silt was derived by digesting three silt samples in SO^f hydrogen per- oxide for 1 week, then oven drying the remaining silt at 60°C for 48 hours (Cortes and Risk 1985). The fraction of carbonates in the silts was deter- mined by dissolving the three samples in 5% hy- drochloric acid for 1 week, then drying the sam- ples as above (Cortes and Risk 1985). Percent organics and carbonates (by weight) in the silts at the two sites were compared in two-sample ^-tests on arcsin transformed data. We counted the number of epifaunal prey in silted and unsilted clumps to determine the possi- ble influence of siltation on juvenile spiny lobster food abundance. The reported estimates of prey abundance are means of two separate counts per clump; 5 clumps per treatment were processed. The volume of each Laurencia clump was deter- mined by water displacement, and all silt load 333 FISHERY BULLETIN; VOL. 86, NO. 2 and prey abundance estimates standardized by clump volume. Prey abundance data were ana- lyzed using a two-way fixed-effects ANOVA on log transformed data and Bonferroni pairwise multiple comparisons. Habitat Selection/Settling Experiments We tested postlarval settlement and juvenile habitat selection in laboratory experiments using clumps of Laurencia spp. with high- and low-silt loads (referred to hereafter as silted and un- silted); the null hypothesis being equal selection of both habitats. Experiments were conducted in fourteen 75.7 L aquaria with subgravel filters and circulating current of 3 cm s ^ Light was provided by skylights and fluorescent lights with a photoperiod of approximately 14L:10D. Two 20 cm diameter algal clumps, one silted and one un- silted, were situated 25 cm apart at opposite ends of each aquarium and at least 5 cm from aquar- ium walls. The number of natural prey in both silted and unsilted clumps far exceeded the num- ber eaten daily by a juvenile. To further control food availability in experiments with juveniles we added equal amounts (10 mg) of Tetramin^ fish food to each clump, providing an overabun- dance of food available ad libitum. If juveniles chose one type of algal clump over the other, then their selection was most likely based on the pres- ence or absence of silt, because food abundance and quality were similar, if not strictly identical, in both types of algal clumps. Pueruli neither feed nor respond to the differential abundance of po- tential prey (Herrnkind and Butler 1986). Silted algae was collected from the No Name Key site (see section on Algal Silt Content and Prey Con- tent); unsilted algae was collected just offshore of the Sea World Marine Science and Conservation Center on Long Key. Fresh algal clumps were used in each experimental replicate. An experi- ment was initiated by introducing a single puerulus or juvenile spiny lobster to the center of an aquarium through a 5 cm diameter PVC pipe. Once a spiny lobster settled to the substrate, the pipe was slowly withdrawn allowing the lobster to move freely about the aquarium. This tech- nique prevented "tailflipping" by lobsters and fa- cilitated active selection of habitats. Twenty-four hours later we located the lobsters and recorded their positions, as in previous experiments ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. (Herrnkind and Butler 1986). Fourteen spiny lob- sters were tested on day 1, 14 more on day 2, and so on until our stock of animals was depleted. Each lobster was used only once. All pueruli were collected on the incoming tide from the plankton in interisland channels. Pueruli were either tested immediately or allowed to metamorphose for later use in experiments requiring juveniles. Data were analyzed with log-linear Goodness-of- fit tests. Metamorphosis Experiment The effect of siltation on the survival and time- to-metamorphosis of pueruli was tested experi- mentally in an outdoor, flow-through seawater system. One freshly collected puerulus was placed in each of 46 seawater-filled 1 L plastic beakers, 23 containing 5 cm diameter clumps of unsilted algae and the other 23 an equal amount of silted algae. Each container was independently sup- plied with flowing, filtered seawater. Algal clumps were replaced daily. Pueruli do not feed and their habitat selection operates independent of food availability (Herrnkind and Butler 1986), thus no food was added to the containers. Sea- water temperature in the beakers remained be- tween 26° and 28°C; photoperiod was approxi- mately 14L:10D. Pueruli were monitored daily and their survival and time-to-metamorphosis recorded. Pueruli were collected as transparent postlar- vae from the plankton in interisland channels which concentrate oceanic postlarvae as they move into Florida Bay nursery areas. Time-to- metamorphosis values represent the elapsed time (in days) from puerulus collection until metamor- phosis into the first benthic stage. Values are likely to differ among collections as different co- horts of pueruli arrive inshore. There are cur- rently no techniques available to determine the actual age of pueruli (i.e., time since metamor- phosis from the phyllosoma stage), but estimates of duration of the puerulus stage range from 2 weeks to 1 month (Lyons 1980; Calinski and Lyons 1983). Differences in time-to- metamorphosis between the two treatments were analyzed via a two-sample ^-test. Juvenile Spiny Lobster Prey Selection Experiments Laboratory experiments were conducted to de- termine juvenile spiny lobster prey preference 334 HERRNKIND ET AL.: RECRUITMENT OF SPINY LOBSTERS and rate of consumption of algal epifauna. Prey were obtained by rinsing large clumps oiLauren- cia through a 100 |jl sieve. Prey included small gastropods, amphipods, isopods, and ostracods ranging in size from 1 to 9 mm (Marx and Herrn- kind 1985a). Prey were individually counted and placed in 1 L plastic beakers containing 750 mL of seawater. One starved (24-h) juvenile lobster (6- 8 mm CL) was introduced to each container, al- lowed to feed for 12 hours, and was then removed; the remaining prey were counted. Fifteen spiny lobsters were tested in each experiment; each lob- ster was used once. Three experiments were con- ducted using different prey combinations (Table 1), but the total number of prey available re- mained similar and exceeded the amount a single lobster could consume in 12 hours. Electivity in- dices calculated for each experiment were used in multiple comparison tests to determine whether juvenile spiny lobsters fed preferentially or ran- domly (Johnson 1980). Table 1 . — Relative prey availabilities (A) and predator usage (U) values (percentages) in the three juvenile spiny lobster prey selec- tion experiments. N = ^5 lobsters per experiment. F-values calcu- lated from Johnson (1980) indicate whether prey choice differed significantly from random in each trial: none of the tests were signif- icant at P = 0.05. Experiment 1 2 3 Prey type A U A U A U Gastropoda Tncolia sp. Batlillaria spp. Tegula spp. Atys spp. Amphipoda Isopoda Copepoda Ostracoda Decapoda F-values df 43.5 615 714 83.0 58.0 54.5 8.7 10 4 3 6.5 00 — — — 14.3 32.6 33.9 10.0 0.2 7.2 9.6 8.3 0.46 4,11 36 0.40 3,14 7.0 17.4 11.6 6.0 0.09 4,11 7.1 20.0 13.1 5.3 RESULTS Postlarvae were more abundant on collectors at the silted site than at the unsilted site. Diver surveys revealed that higher numbers of algal- dwelling juveniles (<20 mm CL) resided at the unsilted site despite greater influx of pueruli into the silted area (Table 2). Only one juvenile spiny lobster collected at the silted site was <20 mm CL, most were considerably larger (25—35 mm CL) than those at the unsilted site, and some were possibly large enough to have immigrated there from adjacent unsilted areas. Silt recovered from the algal clumps was largely calcareous and formed a cohesive cast around the algal filaments. Classifying algal- bound silts by particle size would yield irrelevant values because sieving caused fragmentation of aggregated particles. Therefore we report only the total dry weight of the silts. Algae at No Name Key carried a higher silt load than algae at Burnt Point ix = 125 vs. 65 g/L algae, respec- tively; t = 2.90, df = 18, P <0.01). Silt at No Name Key was characterized as 12.5 ± 7.4% or- ganic and 28.1 ± 27.6% carbonate by weight, whereas Burnt Point silt was 16.9 ± 1.06% or- ganic and 66.3 ± 14.5% carbonate (means ± 1 SD). There was no significant difference in silt compo- sition between sites (organics: t =0.83, df=4, P > 0.05; carbonates: t = 1.97, df = 4, P > 0.05), although sample sizes at each site were small and sample variance substantial. Algae at both sites contained primarily gastropods, amphipods, and isopods, although significantly more gastropods and echinoderms occupied unsilted clumps (Fig. 2, Table 3; P < 0.05 in Bonferroni pairwise multiple comparisons). Significantly more pueruli settled in unsilted algal clumps than in silted clumps during labora- tory settlement choice experiments (38 vs. 11, re- spectively; G = 15.72, P < 0.001). Juvenile spiny lobsters responded similarly in the habitat selec- tion experiment (54 vs. 24; G = 11.78, P < 0.001). We excluded the open sand habitat in aquaria from our analysis because 1) juvenile spiny lob- sters are never found residing on open sand in the field, presumably due to a lack of food and refuge Table 2. — (A) Postlarval lobster abundances at silted and unsilted Florida Bay study areas. Postlarval catch per unit effort (CPUE) was estimated from Witham collector catches. (B) Juvenile lob- ster (8-20 mm CL) abundances at the two study sites in 1985 and adjacent areas sampled dunng 1983 and 1984. Juvenile CPUE was estimated via diver surveys. A. POSTLARVAL ABUNDANCE CPUE 1 Location Description June July August No Name Key silted Burnt Point unsilted 0.08 0.32 0.01 0.05 0.15 0.10 B. JUVENILE ABUNDANCES Location Description Diving hours CPUE2 No Name Key (1983-85) silted Burnt Point (1983-85) unsilted 28 20 0.04 0.50 ippiip - no ot postlarvae no. o( collectors no. of days between collections 2CPUE no. of lobsters collected no. of diver hours 335 FISHERY BULLETIN: VOL. 86, NO. 2 < o z LU CC Z) < _l u. O < 9 > CC Hi m 1200 - 1000 - 800 - 600 - - 400 - 200 - GASTRO* OSTRA AMPHI ISO DECA ECHINO* Figure 2. — Abundances (x ± 1 SE) of the six most common prey found in algal clumps at the silted No Name Key and unsilted Burnt Point sites. Five clumps were collected at each site. Values are standardized by clump volume. Asterisks denote significant difference in Bonferroni multiple comparison tests (P = 0.051. Abbreviation key: GASTROpod, OSTRAcod, AMPHlpod, isopod, DECApod, ECHlNOderm. Table 3. — Two-way fixed-effects ANOVA testing for differences in the total number of individuals among six prey categones (see Table 1) at two sites, one silted (No Name Key) and one unsilted (Burnt Point). Data were log transformed. Source df SS Site 1 0.841 18.78 0.001 Prey type 5 6 242 27.89 0.001 Site X prey type 5 0.847 3.78 0.006 Error 48 2.149 (Marx and Herrnkind 1985b; Herrnkind and But- ler 1986), 2) most spiny lobsters recovered from open sand were actually in corners, indicating edge-seeking behavior rather than selection for sand per se, and 3) only 10.6% of 142 spiny lob- sters tested were found on sand even though it constituted 68% of the exposed substrate in aquaria. Silt had no effect on puerulus survival through metamorphosis to the first benthic instar (13% vs. 9% mortality in silted and unsilted algae, respectively), or time-to-metamorphosis (Fig. 3; t = 0.37, P > 0.05). The total number of prey items consumed in the juvenile prey selection experiments ranged from 19 to 57 prey per lobster per 12 hours. Juveniles fed randomly from the three different prey combi- nations and frequencies offered to them (Table 1). O z C/) 20 O I Ql CC o 10 " LU y^ — O O ' //^ SILTED " U % . // UNSILTED • 0 1 2 3 4 5 6 DAYS FlGt_iRE 3. — Cumulative number of spiny lobster postlarvae metamorphosing as a function of time in one of two treatments: silted algae (Laurencia spp.) or unsilted algae. Twenty-three postlarvae were tested in each treatment; two postlarvae died in the unsilted treatment and three in the silted treatment. DISCUSSION The relative paucity of newly settled spiny lob- sters in the heavily silted region around No Name Key over a 3-yr period indicates that low recruit- ment to benthic habitat is typical there. The ab- sence of juveniles was apparently not due to a 336 HERRNKIND ET AL : RECRUITMENT OF SPINY LOBSTERS lack of postlarval influx, which was higher than that at the unsilted Burnt Point site, but instead to low rates of postlarval settlement. Results from our habitat selection experiments support this hypothesis because settlement was significantly lower in silted algal clumps than in unsilted clumps. Previous studies showed that postlarvae selectively settled in highly architectured materi- als, like algal clumps (Herrnkind and Butler 1986). Thus, heavy silt covering an otherwise pre- ferred habitat either masks the stimuli triggering settlement or contains stimuli that elicit rejection by pueruli. This question remains for further study. We cannot conclusively ascertain from our laboratory experiments the mechanisms govern- ing habitat choice in the field where silted and unsilted habitats may not be adjacent, as they were in our aquaria. Yet for many species with planktonic larvae, these kinds of experiments, coupled with field observations of more general patterns of behavior and abundance, provide valuable insights into natural processes (Sulkin 1986). Twenty percent of the pueruli we tested in lab- oratory tanks settled in silted clumps despite the general rejection of this habitat. Pueruli settling in silted algae probably metamorphose normally into the first benthic instar, as indicated by the equivalent time-to-metamorphosis and early post- settlement survival in both silt levels tested. However, subsequent residency by juveniles pre- sumably is limited, given their great mobility and preference for unsilted algae with high food levels (Marx and Herrnkind 1985b; Herrnkind and But- ler 1986). We found that prey abundances were significantly lower in silted algae. Thus, even if pueruli settle in the silted habitat the subsequent juveniles may leave to obtain adequate food. Fre- quent interclump movement by juvenile spiny lobsters, searching either for food or unsilted habitat, would predictably result in increased predatory mortality. Susceptibility to predation is much greater for juveniles in the open, than it is for individuals amidst algal clumps or dense seagrass (Herrnkind and Butler 1986). Thus, ju- venile residency patterns and susceptibility to predation may, in addition to locally low settle- ment, contribute to the paucity of lobsters in the silted habitat. The algal-bound silt load at No Name Key was roughly twice that at Burnt Point where spiny lobster recruitment was considerable. Our cur- sory surveys from Key Largo to Boca Chica Key indicate that similar silt levels are common. though geographically variable in Florida Bay. Benthic algae, including La//re/icm spp., serve as sediment traps (Scoffin 1970) and demonstrate a remarkable resistance to siltation, growing pro- fusely even in heavily silted areas. Silt in these areas is primarily calcareous, most of it probably a byproduct of sediment processing by deposit feeding shrimp (particularly Callianassa), an- nelids, and sea cucumbers. We did not evaluate the geographic extent of siltation relative to spiny lobster settlement in Florida Bay. However, the demonstrated aver- sion to settling in naturally silted algae, charac- teristic of the region around No Name Key, strongly suggests that low postlarval recruitment and juvenile abundances would occur in similar conditions elsewhere. The sparse juvenile popula- tion at our silted site, one-tenth that of the un- silted site, suggests deleterious impact of high chronic silt levels in areas of potential recruit- ment. Human activities also cause siltation (Mor- ton 1977; Allen and Hardy 1980). We noted that algal stands adjacent to heavily trafficked boat channels typically were more heavily silted than adjacent areas. We suspect the effect of manmade siltation to be similar to that from natural causes. Although it is now generally accepted that Flor- ida Bay shallows serve as the main nursery grounds for the south Florida spiny lobster popu- lation (Marx 1986), the regional distribution of settlement and early juvenile habitation remains to be mapped. Future wide-area surveys by con- cerned researchers and agencies should include sampling of new spiny lobster recruits as well as silt levels. Meanwhile, sizable human activities such as channel construction, dredging, spoil dumping, coastal development, and mineral min- ing must be viewed as potentially deleterious to spiny lobster recruitment. ACKNOWLEDGMENTS We thank the Sea World Center for Marine Sci- ence and Conservation on Long Key, FL, for logis- tical support and the use of their facilities. John Hunt and Jim Marx of the Florida Department of Natural Resources Marine Research Laboratory in Marathon provided field assistance and helpful advice throughout the project. Comments by P. Greenwood, D. Wilber, and two anonymous re- viewers substantially improved the manuscript. This research was supported by a Sea Grant award (R/LR-B-16) to W. F. Herrnkind. Addi- tional support was provided M. J. Butler via 337 FISHERY BULLETIN: VOL. 86, NO. 2 Marine Science Fellowships from the Aylesworth Foundation and the International Women's Fish- ing Association (IWFA). LITERATURE CITED Allen, K D . and J W Hardy 1980. Impacts of navigational dredging on fish and wildlife: a literature review. U.S. Fish Wild. Serv., Biol. Serv. Program FWS/OBS-80/07, 81 p. Aller. R C , AND R E Dodge 1974. Animal-sediment relations in a tropical lagoon, Discovery Bay, Jamaica. J. Mar. Res. 32:209-232. Andree. S W 1981. Locomotory activity patterns and food items of ben- thic postlarval spiny lobsters, Panulirus argus. MS Thesis, Florida State University, Tallahassee, FL, 50 p. Aziz. K A . and J G Greenwood 1982. Response of juvenile Metapenaeus bennettae Racek & Doll 1975 (Decapoda: Penaeidae) to sediments of differ- ing particle size. Crustaceana 43:121-126. BOTERO, L , AND J ATEMA 1982. Behavior and substrate selection during larval set- tling in the lobster Homarus americanus. J. Crust. Biol. 2:59-69. Calinski, M D , AND W G Lyons 1983. Swimming behavior of the puerulus of the spiny lobster Panulirus argus (Latrielle, 1804) (Crustacea: Pal- inuridae). J. Crust. Biol. 3:329-335. Cortes, J N . and M J Risk 1985. A reef under siltation stress: Cahuita, Costa Rica. Bull. Mar. Sci. 36:339-356. Crisp, D J 1974. Factors influencing the settlement of marine inver- tebrate larvae. In P. T. Grant and A. M. Mackie (editor), Chemoreception in marine organisms, p. 177-265. Acad. Press, N.Y. 1976. Settlement responses in marine organisms. In R. C. Newell (editor). Adaptation to environment: essays on the physiology of marine animals, p. 83-124. Butterworths, Lend. Gray, J S 1974. Animal-sediment relationships. Oceanogr. Mar. Biol. Annu. Rev. 12:223-261. HERRNKIND, W F , AND M J BUTLER IV. 1986. Factors regulating postlarval settlement and juve- nile microhabitat use by spiny lobsters Panulirus argus. Mar. Ecol. Prog. Ser. 34:23-30. Howard, A E., and D B Bennett 1979. The substrate preference and burrowing behavior of juvenile lobsters (Homarus gammarus (L.)). J. Nat. Hist. 13:433-438. Johnson, D H 1980. The comparison of usage and availability measure- ments for evaluating resource preference. Ecology 61:65-71. Jones, G , and S Candy. 1981. Effects of dredging on the macrobenthic infauna of Botany Bay. Aust. J. Mar. Freshwater Res. 32:379-398. Little, E J 1977. Observations on recruitment of postlarval spiny lobsters, Panulirus argus, to the South Florida coast. Fla. Mar. Res. Publ. No. 29, 35 p. Little. E J , Jr , and G R Milano 1980. Techniques to monitor recruitment of postlarval spiny lobsters, Panulirus argus, to the Florida Keys. Fla. Mar. Res. Publ. No. 37, 16 p. Lyons, W. G 1980. Possible sources of K'lorida's spiny lobster popula- tion. Proc. Gulf Caribb. Fish. Inst. 33:253-266. Marx. J M 1986. Recruitment and settlement of spiny lobster pueruli in south Florida. Can. J. Fish. Aquat. Sci. 43:2221- 2227. Marx, J M , and W F Herrnkind 1985a. Macroalgae (Rhodophyta: Laurencia spp.) as a habitat for young juvenile spiny lobsters, Panulirus argus. Bull. Mar. Sci. 36:423-431. 1985b. Factors regulating microhabitat use by young ju- venile spiny lobsters, Panulirus argus: food and shel- ter. J. Crust. Biol. 5:650-657. Morton, J. W. 1977. Ecological effects of dredging and dredge spoil dis- posal: a literature review. U.S. Fish Wild. Serv. Tech. Pap. No. 94, Wash., DC. Pearson, T H , and R Rosenberg 1978. Macrobenthic succession in relation to organic en- richment and pollution of the marine environ- ment. Oceanogr. Mar. Biol. Annu. Rev. 16:229-311. Pottle, R A . and R W Elner 1982. Substrate preference behavior of juvenile American lobsters, Homarus americanus, in gravel and silt-clay sediments. Can. J. Fish. Aquat. Sci. 39:928-932. Rhoads, D C 1974. Organism-sediment relations on the muddy seafloor. Oceanogr. Mar. Biol. Annu. Rev. 12:263-300. Rhoads, D C , and D K Young 1971. Animal-sediment relations in Cape God Bay, Mas- sachusetts. II. Reworking by Molpadia oolitica (Holothuroidea). J. Mar. Biol. 11:266-261. Roach, S G 1983. Survivorship, growth, and behavior of juvenile lob- sters Homarus americanus Milne-Edwards in controlled environments in nature. Nova Scotia Dep. Fish. Mar. Res. Tech. Rep. Ser. 83-02, 60 p. RUELLO. N V 1973. Burrowing, feeding, and spatial distribution of the school prawn Metapenaeus macleayi (Haowell) in the Hunter region (Australia). J. Exp. Mar. Biol. Ecol. 13:189-206. SCOFFIN, T p. 1970. The trapping and binding of subtidal carbonate sed- iments by marine vegetation in Bimini Lagoon, Ba- hamas. J. Sediment Petrol. 40:249-273. SULKIN, S D 1986. Application of laboratory studies of larval behavior to fisheries problems. Can. J. Fish. Aquat. Sci. 11:2184— 2188. WiTHAM, R , R M iNGLE, AND H W SiMS jR. 1964. Notes on postlarvae o{ Panulirus argus. Q. J. Fla. Acad. Sci 27:289-297. WiTHAM, R , R M iNGLE, AND E A JOYCE jR 1968. Physiological and ecological studies of Panulirus argus from the St. Lucie estuary. FL Board Conserv. Tech. Ser. 53:1-31. 338 A LIMITED INFORMATION APPROACH FOR DETERMINING CAPITAL STOCK AND INVESTMENT IN A FISHERY^ James E Kirkley^ and Dale E. Squires^ ABSTRACT There have been few empirical studies on the level of capitalization and investment in fisheries because the necessary data are often inadequate. Specifically, data on capital stock and investment in a fishery are not routinely collected and compiled or are limited in scope. In this study, a method is provided for estimating the aggregate capital stock and investment in a fishery utilizing the available information. Data on acquisition and list prices and vessel characteristics for a sample of New England vessels are obtained. The data are then used to estimate an hedonic cost function which specifies the acquisition price as a function of vessel characteristics. The resultant equations are subsequently used, with information on vessel characteristics for all New England vessels, to esti- mate aggregate capital stock and investment. The results indicate that substantial investment oc- curred in the otter trawl and scallop dredge fisheries, particularly since the Magnuson Fisheries Conservation and Management Act. Moreover, the results demonstrate that the number and change in the number of vessels are inadequate indicators of the level of capital stock and investment in a fleet comprised of vessels with heterogeneous characteristics. The common property nature of fisheries is recog- nized as causing excess capitalization and har- vesting capacity (Gordon 1954). The theoretical argument is that since fishermen do not have to pay for the utilization of common property fish stocks, new vessels enter a fishery until net rev- enue is driven to zero. This common-property fea- ture results in more capital and investment than is economically optimum. Public regulation of fishing industries is usually advocated to redress the excess entry, economic inefficiency, and loss of economic rents in a common property fishery (Scott 1979; Sissenwine and Kirkley 1982). In essence, overcapitalization and excess in- vestment are perceived to be the reasons for many of the major fisheries problems (Hilborn 1983). Economists argue that management of overcapi- talization is necessary to realize the benefits of fisheries (Cunningham et al. 1985). Alterna- tively, gains from fisheries management require control of overcapitalization and excess invest- ment (Charles 1983a, b). Clark et al. (1979) and Charles (1983a, b) ex- tended the static theory of optimal fisheries in- vestment as developed by Gordon (1954). They ^Virginia Institute of Marine Science (VIMS) Contribution No. 1467. ^College of William and Mary, Schools of Marine Science and Business Administration, Gloucester Point, VA 23062. 3Southwest Fisheries Center La Jolla Laboratory, National Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. also demonstrated the relationships between in- vestment, capital stock, overexploitation, and fisheries management. They concluded that in- vestment should be more conservative to prevent overexploitation. The economic literature provides substantial justification for the need to solve the problems of overcapitalization and excess investment. Yet, few empirical studies document the level of capi- talization and investment."* Moreover, there ap- pears to be no attempt by any U.S. agency to routinely collect and compile statistics on either the stock of capital or the level of investment in U.S. fisheries. Overcapitalization and excess in- vestment, though, continue to be suggested as the reasons for many of the economic problems of fisheries. Since data are often inadequate, many empiri- cal studies on fisheries consider capital stock and investment in terms of the number of vessels. That is, capital stock and investment in a fishery are tyically measured in terms of number of ves- sels. In the absence of appropriate information, the number of vessels may be the only basis for examining overcapitalization and excess invest- ment. Alternatively, if a fleet has identical-sized vessels, fixed inputs or vessel characteristics, and Manuscript accepted February 1988. FISHERY BULLETIN: VOL. 86, NO. 2, 1988. '•Tettey et al. ( 1986) provided an exception, but their analysis is restricted to the shrimp fishery and is based on cost data obtained from boat builders. 339 FISHERY BULLETIN: VOL. 86, NO. 2 gear, the number of vessels may be used to indi- cate the stock of capital and investment. How- ever, Kendrick (1961) demonstrated that the total number of operating units, such as the number of plants, is an inadequate measure of capital stock. Also, few fleets have identical-sized vessels, char- acteristics, or gear. As a consequence, the number of vessels does not provide adequate information to indicate the level of capitalization and invest- ment. This study presents an approach for estimating the aggregate or industry level of capital stock and investment in a fishery comprised of hetero- geneous vessels using available information. Data available include acquisition and list prices for some but not all vessels and boat characteris- tics for all vessels in a fleet. The approach was developed and used to obtain estimates of capital stock in support of the United States and Cana- dian maritime boundary dispute. It is viewed as an initial step towards examining the problems of excess capitalization by providing estimates of aggregate capital stock and investment. THE DEFINITION AND MEASUREMENT OF CAPITAL AND INVESTMENT Capital The concept of capital has created problems for economists for quite some time. The term has been used in so many different contexts that it is a source of enormous confusion (Hirshleifer 1970). Sloan and Zurcher (1968), in their "Dictionary of Economics", define capital as "One of the major factors of production consisting of property from which an income is derived, ex- pressed in terms of money. Popularly, the term is frequently used interchangeably with capital good. A distinction is sometimes made between money capital, or that part of the capital held in the form of money and bank deposits, and prop- erty capital, or that part of the capital held in the form of evidences of ownership such as stocks, bonds, and mortages." Hirshleifer (1970) presented three meanings of capital: 1) real capital or capital stock, 2) capital value, and 3) liquid capital. Real capital is de- fined as a collection of capital goods or an aggre- gate of heterogeneous capital inputs. It is one of the major productive commodities or economic factors of production. Capital value is the net dis- counted value of expected future income streams associated with a capital good. Liquid capital is the level of current funds available or intended for investment. In the case of fisheries, real capital or capital stock is the form of capital which should be exam- ined with respect to the problems of overcapital- ization and excess harvesting capacity. It is the relevant measure of capital goods used or avail- able for production (National Academy of Sci- ences 1979). Moreover, capital stock is the con- cept of capital used to define and measure the services of capital inputs which are required to harvest fish; that is, it is the concept of capital required to define the economic production tech- nology of a fishery. Thus, this study is concerned with the concept of real capital stock. ^ The measurement of capital stock in a fishery, however, presents several problems. First, capital inputs are usually quite heterogeneous and can- not be easily aggregated without restrictive as- sumptions about the form of the catch equation or fishermen's behavior. Conceptually, it should be possible to combine all the different types of capi- tal goods by weighting each type by its average compensation (i.e., the rental price). However, in- formation at this level of detail is typically not available. Second, fisheries agencies, particularly in the United States, generally do not collect and compile information necessary to calculate capi- tal stock and investment. Third, in order to com- pare changes in capital stock over time, the meas- ure of capital stock must be converted to some base period value by deflation. In empirical economic studies of traditional in- dustries, the common practice is to measure the capital input or stock by converting the purchase or acquisition price or the book values of capital to base period values by the use of a price index (National Academy of Sciences 1979). Varian (1984) noted: "The usual procedure is to measure capital value and then deflate by a price index; in some sense, this should measure the level of cap- ital stock." In this study, the acquisition price is used as the measure of the stock of capital or capital value; the real stock of capital is obtained by dividing the capital value by the producer price index for heavy machinery. Investment The definition and measurement of investment is more straightforward than is the definition and 5The market value of capital, acquisition price, and cost of capital are used interchangeably in this study. 340 KIRKLEY AND SQUIRES: CAPITAL STOCK AND INVESTMENT IN A FISHERY measurement of capital. In simple terms, invest- ment is the exchange of money for some form of property. Although there are several types of in- vestment (Branson 1972), attention is restricted in this study to an examination of net investment (i„) which is defined as the difference between the stock of capital in two periods of time (Baumol 1977). This is the same concept of investment ex- amined by Tettey et al. (1986). Replacement investment is another type of in- vestment. The sum of replacement investment and net investment equals total investment. Re- placement investment is measured by multiply- ing the rate of depreciation times the level of the capital stock. Replacement investment, however, is not estimated in this study since the method and rate of depreciation must be arbitrarily as- sumed. METHOD OF ANALYSIS The Hedonic Approach While the definition and measurement of real capital stock and investment are conceptually straightforward, limited data and heterogeneous capital inputs complicate the measurement of capital stock and investment. In particular, data are usually only available for a small number of vessels which are often similar in their character- istics; these data are inadequate for calculating actual or observed capital stock and investment in a fishery. However, data are available on the acquisition price and characteristics of vessels which permits estimation of the stock of capital by an hedonic approach. In turn, estimates of cap- ital stock from the hedonic approach can be used to estimate net investment in a fishery. The hedonic approach hypothesizes that the price of a commodity is influenced by its charac- teristics (Rosen 1974). The hedonic price or char- acteristics function in market equilibrium re- flects both the distribution of marginal rates of substitution over households and the distribution of marginal rates of transformation over firms. In effect, the hedonic hypothesis states that goods may be valued for their attributes and that im- plicit or hedonic prices exist as a function of the attributes. Griliches (1971) estimated hedonic cost func- tions of U.S. automobiles using advertised or list prices and actual transaction prices. Griliches, however, was concerned with explaining quality differentials. Ladd and Martin (1976) examined prices and demands for input characteristics using a Neoclassical input characteristics model. Triplett (1986) estimated quality adjusted price indexes for computers. Thus, there is a history of deriving prices of commodities as a function of their characteristics. The hedonic approach offers several attractive properties for estimating capital stock in fishing industries. First, the hedonic method incorpo- rates changes in the quality of capital over time (Triplett 1986). This is because changes in quality should be reflected in vessel acquisition prices. The hedonic approach, thus, permits a quality ad- justed measure of capital stock to be obtained. In contrast, vessel count ignores changes in the quality of capital. Second, the hedonic approach allows easy aggregation of the heterogeneous cap- ital inputs frequently observed in a fishing fleet because the heterogeneous capital is measured by value rather than physical measures. In the hedonic approach, cost (C) or input price may be expressed as a function of the associated characteristics of the commodity or input, C = f(CHi, CH2, ...,CH„) , (1) where C is cost, CH, is the tth characteristic (Braeutigam et al. 1982). The characteristic cost equation can be obtained either from a dual speci- fication or by determining the reduced form equi- librium equation. The Dual and Reduced Form The dual specifies cost as a function of input prices and their characteristics. The partial derivative of the dual cost function with respect to input prices yields the demand for the inputs as a function of prices and input characteristics. The reduced form is obtained by solving structural equations, equations which explain the behavior and interrelations of endogeneous variables, of demand and supply for the endogeneous variable, price, in terms of exogeneous variables and dis- turbance terms. ^ In this study, the reduced form is directly speci- fied and estimated for two reasons: First, all of the input price data necessary for estimating a dual are not available. Second, estimation of the struc- tural equations is complicated by the need to use a limited dependent variable method of estima- ^See Rosen (1974) for additional information on obtaining reduced form equations for hedonic prices. 341 FISHERY BULLETIN: VOL. 86, NO. 2 tion since demand is a discrete choice, and there are individuals or firms which do not purchase a vessel. Rather than attempt to deal with these difTi- culties which do not appear to have widely ac- cepted solutions, it is assumed that there are underlying structural equations of demand and supply which permit the derivation of the reduced form equation. Moreover, the direct specification and estimation of the reduced form is the more commonly used approach to estimate the value of a commodity as a function of its characteristics (Bockstael et al. 1986). Last, the use of the re- duced form should not create a problem since the hedonic function is a joint envelope in character- istic space of demander's bid schedules and sellers offer schedules, and thus, the observed implicit price locus is a reduced form equilibrium vector reflecting both supply and demand influences. As a result, shifts in both market demand and supply are incorporated into the reduced form equi- librium vector. The Capital Stock Model As previously indicated, it is common practice to use the acquisition price as a measure of capital stock. By the hedonic approach, the capital stock or acquisition price of a fishing vessel is postu- lated to be a function of several vessel character- istics. The following general specification is con- sidered: C = f(LN, CS, YEAR, FI, AGE, DNO, HPWR, GRT), (2) where C is the list or advertised price or the ac- quisition price of a vessel, LN is length, CS is hull type (1 = steel, 0 = other), FI is fishery or gear tj^e,' AGE is age of vessel, DNO is a dummy variable for new and used vessels (DNO = 1 for new and 0 for used), HPWR is engine horsepower, and GRT is gross registered tonnage. The vari- able year equals the year of the observation. It is included to incorporate technological changes '^Gear type is an integer valued variable set equal to the gear codes used by the Northeast Fisheries Center. This may influ- ence the results similarly to the use of nonbinary dummy vari- ables (Kmenta 1971). The use of this variable implies that the cost difference between gear types is equal to a scalar multiple of the lowest integer valued gear code. Thus, statistical rejection of cost differences due to gear type is actually a rejection of differences not being equal to scaleir multiples. However, the large number of gear codes prohibits their separate treatment. over time, such as the introduction of electronics and stern trawling, and structural changes in the industry such as changes in organization of capi- tal and fish markets and changes in public regu- lation. Functional Form Several functional forms could be specified for Equation (2). Alternatively, generalized Box-Cox (Box and Cox 1962) transformations could be used to determine the functional form. However, the selection of functional form is mostly concerned with obtaining the relationship between the de- pendent variable and individual independent variables. This study is primarily concerned with estimating the capital stock. Moreover, it may be quite difficult to estimate capital stock from Box- Cox transformations or other functional forms since the conditional expectation is often of a com- plex form (Smallwood and Blaylock 1986). Also, Dadkhah (1984) noted that predictions based upon some transformations and functional forms result in biased predictions and asymmetric confi- dence intervals. Since this study is primarily concerned with estimating capital stock and investment rather than estimating the implicit prices for character- istics, a linear functional form is proposed for Equation (2). The capital stock equation is speci- fied as a second-order polynominal in order to provide an approximation to an unknown, under- lying hedonic function. This is equivalent to specifying a generalized quadratic flexible func- tional form and imposing zero valued restrictions on the cross product coefficients or interaction terms (Lau 1978). Flexible functional forms are widely used in economic studies concerned with determining the underlying economic structure because they impose very little structure on the economic equations of concern (Blackorby et al. 1978). Postulated Relations The postulated relations between cost and each of the explanatory variables is as follows: 1) the coefficients for the first order terms of vessel ton- nage, length, and engine horsepower should be positive since costs should increase as the size characteristics increase; 2) the coefficient for steel-hulled vessels should be positive since steel vessels are generally more expensive than wood or fiberglass vessels; 3) the dummy variable coef- 342 KIRKLEY AND SQUIRES: CAPITAL STOCK AND INVESTMENT IN A FISHERY ficient for new vessels should be positive since new vessels tend to cost more because they are newer and usually incorporate more recent tech- nological advances; 4) the second-order coeffi- cients for the size characteristics are expected to be negative since cost is believed to increase at a decreasing rate in response to increases in the size of a vessel. THE DATA Data used to estimate the capital stock equa- tion were obtained from three sources: First, in- formation on vessel acquisition price and associ- ated characteristics were obtained for 164 new vessels from the Northeast Regional Office of the National Marine Fisheries Service.^ Second, data on list prices and vessel characteristics for 946 used vessels were obtained from various trade magazines and vessel brokers. Third, detailed vessel and fishery data were obtained from the U.S. Coast Guard master vessel listing and the Northeast Fisheries Center. Data were obtained for the years 1965 through 1981. The first two sets of data were used to estimate the capital stock equations. The third data set was used to estimate aggregate capital stock and investment in selected New England fisheries using the estimated capital stock equations; that is, vessel characteristics for all of the vessels in the New England fleet were inserted into the esti- mated capital stock equations to obtain estimates of capital stocks per vessel in a given year. The estimated capital stock per vessel was summed over all vessels in a year to obtain total or aggre- gate capital stock. Net investment was then cal- culated as the annual change in total capital stock. The use of both the list and acquisition prices, however, presents a problem. The list price re- flects the supply price or the price at which a vessel is offered for sale. The acquisition price reflects the price determined by the equilibrium between demand and supply. As a consequence, estimates of capital stock and investment may be in error. It is not known by how much the acquisition price differs from the list price for used vessels. Thus, the magnitude of the error cannot be calcu- lated. Since the list price is not less than the ac- quisition price, estimates of capital stock based on Equation (2) are likely upwards biased. Counter to this problem is the argument that by the time a vessel is ready for fishing, required capital mod- ifications or improvements may result in the cap- ital stock being close in value to the list price. There are no completely satisfactory economic justifications for using both the list and acquisi- tion prices. Two possible justifications are that Griliches (1971) and others used both prices, and capital improvements or modifications may be nearly equal to the difference between the list and acquisition prices. Another possible justifica- tion is the accepted use of cost data obtained from boat builders as in Griffin et al. (1978) which may impose similar problems plus the statistical prob- lem of measurement error. In the remainder of this paper, the term "acquisition price" is used although the estimation and calculation of capital stock and investment are based on both price series. EMPIRICAL RESULTS Equation (2) was estimated by ordinary-least- squares with the dependent variable, the acquisi- tion price, measured in both nominal and real terms; the real price is the nominal price deflated by the producer price index for heavy machinery. As previously stated, deflation is necessary to es- timate the real capital stock. The estimated coef- ficients and statistical results are presented in Tables 1 and 2. Equation (2), however, was also estimated using data for several groups of years between 1965 and 1981. The reason for considering differ- ent time periods was that prior knowledge of New England fisheries suggested that various changes occurred in the fisheries during the selected time periods.^ Moreover, economic studies of New Eng- land fisheries have often been criticized for as- suming stable relationships over time or over dif- ferent cross sectional units. Failure to incorporate changes in the estimated relationship between cost and vessel characteristics over time may re- sult in biased parameter estimates. Changes which have possible ramifications for the estimated relationships include 1) depressed resource conditions and the presence of foreign fishing between 1965 and 1971; 2) management by the International Commission for Northwest ^Data are confidential and may not be available to other re- searchers or the general public. ^Additional information on possible structural changes in New England fisheries is available in Kirkley et al. (1982), Dewar (1983), Doeringer et al. (1986), and Kirkley (1986). 343 FISHERY BULLETIN: VOL. 86. NO. 2 CO y^ in Tj- O) CO CM ^ tr Oi O) CO CO 00 r~. 00 3 d d d d d d d o o r;^ •* — o> ^ CO r-- CO t^ ^ S LU to ^ CM <0 ■* CO en in CO in CM O T- OO ^ < CO ^ in in tv. CO r^ CM 1 3, CD U) > 05 >. > E ? CO o C:in cop 5 P CJ> 00 CM Op ay CO f in CM CM ^ r^ o CO 8 e£ Cvj o CM Tt CM CO CO T— CO ^— s y— in •>*' "g. CO in « aJ C\J — T- --^ r^ ^-^ ■* CO ,_. CO CO ^ C/D ® E s^ o2 g^ in 05 d op in CD CO 8 CM •<* ^ w r- — . Ol S.9 CM ■•- ^ C <^ C C^J d T— y— y— CD ^ 1 ^^ 1 '" ""^ ^~^ "■"'^ §< C\J ^_^ t^ ^^ in ^__ • CJ> CO 00 t^ in o in C. 5£i O T- CM CM d 5 CM T-^ o ^ CD in CO o CM CM CM • -g (0 LU > 1 d ^ei gS2 00 55 CM 1 1 <5 in CO rii s To c CM CO 1 o (NJ 2 — CO ^^ CO ^ CO in CM CM CO oT CM In 5 CO c z •,- CO <75 r>J (O o 5 O d ■«T CM o d C\) o .^ (D o y— CM c o o CD r: ^ n3 tf) o> — ■* .~. C\J ^^ CO , ^ in ,,_^ q ,_^ T— , , CJ c- tn 2 '^ 3 ST J3 z -I §? d° CO '^ g^ o in in O CD o CD O in 0) (0 CM O CO CO »- ■* •* d o CM CM CM o Cvj CD 0 o > 1 "^ CO ^-' CD "-^ d o d CO d CO d CM m nj © > Ol => ss e o a. CO 1 T— T^ CM T^ T^ ih (/) I ^-^ ""^ ^-^ ' — ' ' — ' ""^ B ra CT-g ® ^ « o x: nj 1 oo ,— in 05 CD CO S§oo CC CO — CM O o -~- 1-^ t^ OO m § ?;" (5 CO d O CD ■^ in CO CtD o CO CM in S =* 0 o (D I O -; "^ ,^ to )-c CM (J) CO S « >, . w CM £. 1 2. y- d^ CM ^ ^ ^ 2= CD S Q) 1 1 « t: .c (0 5i c — c OB ffl ra .i" i; E Csj ® d CM — h- § C35 S" in d ■* CM « £ tr ^ CO (0 Q. 1 S| co [ O) d i CM CM d 1 in CM « 5 < — 2 S UJ m IE 3) ^ CS 1 ■ — ' 2. w ^ 01 mo.-® ^ CNj x: -o > QC c ra » in r^ ■«r CM t CO CO UJ Tj- Tt CM ^^ to Cvj • CM o H o CO 2. 72i ^-S to d CM .^ o CM ?^ CM 1 o> o o o o o o "" q! X) 0 c ■>s^ _^ CO d ^ d p^ LO •§ "5 - >, « GO - — - 5= CM CM ® 00 CO CO « CO CO 00 CM o CM CD o c o ■.- o 22 -* iP CO CM OO CM in y-i CM r^ C*J Tj- t-- ~^ 1 in O 1 CO 1 i 1 1 ?% S E 1^ S E t;: ^ — > ^ CO CO ■V O) T— h- o •S "^ m "C^ 2 CO CO 00 l~~ CO (DPS" CM CM T— y— (0 ^CM -« in t^ a> ll^i - (j> 05 O) <7> O) 05 CD o o CO CD O o a. to o "(0 0) c o CO 13 CT (t> 0 'c o (D B E CD (0 eg m < DC (0 o CD CO CO 00 d 05 00 corr- y- ^^ eg—, h-^^ t^ — in"' "-o cMin -"jco •-o I "^ I ~^ "^ I "^ CO — . o CO -^ CO — ~ " Eo «> ^ CO °^ CM ^ Tt -I- 1- co CO CO t^ 00 CO ••-t^ r^TT t^OO -^(D ■r-05 OOO) CMCD r^o '- _ ■•- _ 00 _ o 00 C. CO C- CM — - o CM ^^ 1- ^^ t^ "tj- CD r^ ^-05 in 00 in ^ CD .^ CD C CO C CM 00 05 oco r^co yf (£> coin o r^ •.- -^ eg Tj- ,- Pm coeg cb^ CM'— 05® 'i'^ -^o '-COCO-q^CO.^'-.^ \ y^ jC3 — I — I -^ ^ ^ ^ 000 CD ~ CO — . CM . — . §s gte «s CO o o 10 in '*• I ^^ •* -^ CO -^ o o CO ^^ CO . — 05 ^^ CO o S o CO 05 -^ CD "^ eg CO o in eg T- CM ^-' eg "-' CO ^ ^ CM o r- CO P ^ '. 00 O CO O '- '- _ CM _ CO ^ CO o CO '"' ~ d> y- O CO CM '- P in O CO 05 fs. eg £. n CM CO CO ^ — ^ o ^ — 7^ in • o Si o r: CO CO ■ CO • P. '- in CD ■^ 'X- f^ 'X- CD ;r- ■ in _; in ■ o 05 CD CO '^ ^ CO in Z; CD ■ r- ■ I O '- '- •t,^ CO. h^^ co,^ CM—, in,^ ooco '-r^ 'S-'- r^Tj- •^r^ cdcd CO ,nCO rSI^ CO*" d"^ lri° CD CO in CO ^^ CM r^ in CD O I '- CO CD O CO CD^ in^ Tt^ 05^ ihcD COCO -^in oocM cocD cocg cor^ oin '- y- f-T*: C*'d ■^ y^ y- CO ■* -^ d o ;r o r~ CO o 1^ '^ 00 ■ '- T ■* '- 00 .r- o eg CO ;;;- r^ 00 eg CO CO eg f^ — - I o T in r-~ r^ • — CO eg CO CN h- CM I eg in in CD 00 •«1- CM '^ eg 05 06 in --- 00 CM CO CO CO CD »- eg '- 05 00 1^ eg 1^ CO '- in r-~ 05 t^ t^ t^ t^ I I I I in CM CD CO o '- CD r^ r- h~ CO 00 05 05 05 05 05 05 a -C5 _: (0 Q) 1- O) m tfi > , > 11 %o ™ §> en to CO ® O S — at 1^ — ..LU ^5. S,|S O _:_; 52 o S 5 5 i "P eo e — §5 £ >3 CO I II S 3 — £ ■^ £ c ^ .<» m C ! CL -O T- C ^ — ■- o c <" ^ o £ ™ B-g M 2 ^ ^^ o 10 o n £ c o o =: £ 2 o o ■D 5 s — 03 to o) 2 i? * o o (0 o CM £ QC c O cj) <0 IS I- re QC t>j CD CC -§ . . Q. «I |c o *^ re $ T3 Q. ? I £^ re 0) > 2 re s 05 0) ? n (T> r 0 0 re 0 >. m < r 05 c CL ,_ < .^ Hi en >- 0 tn TT ? ? rr T5 <1) C 3 re re > 1 (ft r ■n 0 r re 9) r n m re c 2 0) F (Tl E •V 05 ■0 re as CO 2 0 ■0 c ^ 0 (0 Z in _i j- CO '^ 344 KIRKLEY AND SQUIRES: CAPITAL STOCK AND INVESTMENT IN A FISHERY Atlantic Fisheries and increased fuel prices be- tween 1972 and 1975; 3) the formulation and im- plementaiion of the Fisheries Conservation and Management Act in 1976 and 1977; 4) extremely restrictive regulations in 1978 and 1979; and 5) a long dock strike in New Bedford, frozen harbors, and foreign purchases from American fishermen in 1981. In addition, there was substantial entry of new steel-hulled vessels between 1975 and 1977 and 1979 and 1980. Chow (1960) tests of the equality of regression equations were conducted to further examine the possibility that the statistical relationships be- tween cost and vessel characteristics changed over time. The Chow test is an F-test of the stabil- ity of the coefficients (Maddala 1977). The results of these tests are presented in Table 3. The results of the tests indicate that the esti- mated relationships between cost and character- istics were not the same for the selected time peri- ods. Alternatively, the hypothesis of the equality of the regression equations for different periods of time could not be accepted at any reasonable level of significance. However, these results only verify that different models should be estimated for dif- Table 3. — Results of tests for the equality of regression equations. Periods Critical values' tested F-statistic2 0.05 0.01 1965-71 1972-75 1976-77 1978-79 1980-81 39.38 '^12,1050 = ''•^5 f^12,1050 = 2.18 1965-71 1972-75 8.45 1^12.325 = ■'•'75 f^12,325 ^ 2.18 1965-75 1976-81 20.55 f^12.1086 = ''•^5 f^12.1086 = 2.18 1965-75 1976-79 21.33 1^12.668 = ■'■''5 '^12.668 = 2.18 1976-77 1978-79 3.60 '^12,329 = l-^S '^12,329 = 2.18 1976-77 1978-81 3.22 ^12.737 = l-^S f^1 2.737 = 2.18 1976-79 1980-81 3.35 f^12.737 = ■'•75 f^12.737 = 2.18 1980 31981 2.76 f^12,384 = "I -^5 f^l2,384 = 2.18 'F-statistIc for denominator degrees of freedom equal to infin- ity. ^Restricted residual sum of squares obtained by pooling data over all time periods and estimating Equation (2). Unrestncted residual sum of squares obtained by estimating Equation (2) for each penod being examined and then adding the residual sum of squares for each equation. ^Residual sum of squares obtained from estimating Equation (2) without the time vanable ferent periods of time. They are not conclusive proof of the selection of the years for a particular group of years. A more accurate determination of the years to be included in each group requires considerably more estimation and hypothesis testing which is beyond the intent of this study. As indicated by the ^-statistics in Tables 1 and 2, the age of the vessel, whether or not the vessel is new or used, and the size characteristics appear to be the more statistically significant explana- tory variables. These results are consistent with the results of Griffin et al. (1978). The statisti- cally significant negative coefficients for the size characteristics during the years 1972-75 are par- ticularly interesting. These years coincide with the Arab oil embargo when fuel prices increased; expected increases in future operational costs may have deterred new entry. The results also indicate that the value of a vessel declines as it becomes older. The coefficient for hull construc- tion, CS, suggest that steel-hulled vessels are more expensive than are those of other materials such as fiberglass, ferro cement, wood, and alu- minum. There does not appear to be a difference in the value of a vessel based on gear type. This may be a result of specification problems with the dummy variable for gear t)^e (see footnote 7). It is of further interest to examine the elastic- ities of cost with respect to the vessel characteris- tics. ^° These are presented in Table 4. As indi- cated, vessel tonnage, length, and age have the greatest influence on the acquisition price. The elasticity for length is consistent with the manner in which vessels are sold; that is, vessel prices are largely stated in terms of their length. There are, however, inconsistences in the esti- mated capital stock equations and elasticities. In '0 Elasticities indicate the p)ercentage change in cost result- ing from a 1% change in the value of a characteristic. Table 4. — Estimated elasticities of cost with respect to vessel characteristics.' Elasticities Year GRT2 HPWR LN AGE 1965-71 0.43 0.06 2.88 -0.32 1972-75 0.78 -0.04 -0.08 -0.08 1976-77 1.14 0.09 0.11 -0.11 1978-79 -0.14 0.30 2.38 -0.40 1980 0.16 0.30 1.43 -0.30 1981 -0.64 0.56 2.38 -0.41 ' Elasticities based on nominal estimates and observed mean values of cost and vessel charactenstlcs. 2GRT is gross registered tonnage; HPWR is engine horse- power; LN is lengtti; AGE is age ol vessel. 345 FISHERY BULLETIN: VOL. 86, NO. 2 particular, not all of the estimated coefficiencts have the desired or expected sign. For example, the coefficient for tonnage in the 1981 estimate is negative and that for the square of tonnage is positive. In contrast, a positive coefficient for ton- nage was expected since vessel cost should in- crease with vessel tonnage, and the coefficient for tonnage squared was expected to be negative be- cause vessel costs are believed to increase at a decreasing rate. The incorrect signs could be the result of multi- collinearity between vessel characteristics. Aux- iliary regressions, though, do not indicate severe multicollinearity. The R^ values are <0.35. If multicollinearity is a problem, its major effect is on the estimated relationships between cost and individual characteristics, the estimated elastic- ities, and the previously discussed Chow tests. Multicollinearity is not thought to be a problem for several reasons. First, Chow tests in the pres- ence of multicollinearity usually result in the ac- ceptance of the null hypothesis of the equality of regression equations or that differences in the es- timated coefficients are statistically insignificant (Maddala 1977). Second, multicollinearity does not pose a problem for estimating capital stock and investment; the estimated equations still per- mit estimation of the conditional mean.^^ 11 For additional information on the problems of multi- collinearity with respect to analysis and prediction, see Kmenta (1971). The equations for the real or deflated value of capital were used to estimate total capital stock and investment for the New England fleet be- tween 1965 and 1981. The real capital stock equa- tions were applied to vessel data available from the Northeast Fisheries Center, and total capital stock was calculated as the sum of the stock over all vessels. Aggregate net investment was calcu- lated as the difference in total capital stock be- tween consecutive years. Prior to estimating total capital stock and in- vestment, it was necessary to define vessel type by gear to avoid double counting which might occur since vessels frequently switch port and gear. Three categories of gear types were estab- lished: 1) otter trawl, 2) scallop dredge, and 3) all others, which include lobster trawl and pots, her- ring gear, harpoons, etc. Vessels were assigned a gear type based on a plurality of days at sea. Esti- mates of capital stock and investment are pre- sented in Table 5. Several limitations should be considered when evaluating the estimates. First, there is the previ- ously discussed problem that estimates of capital stock may be biased or overestimated. Second, estimates are only for New England vessels or those which are believed to be homeported in New England. Third, the capital stock and investment series pertain to New England and not the United States; that is, estimates should not be inter- preted as net changes in the capital stock and investment in U.S. fisheries. Fourth, under- Table 5. — Estimates of real capital stock and investment in New England harvest- ing sector, 1965-81.1 Number of vessels Capital stock Investment Index of investment Year Trawl Dredge Other Trawl Dredge Other - $ Millie inc --_--. Mio ■ 1965 594 25.96 16.29 1.98 1966 619 28.84 14.75 2.82 2.88 -1.54 0.84 100 1967 628 39.37 11.42 1.63 10.53 -3.33 -1,19 276 1968 610 28.65 15.43 1.68 -10.72 4.01 0.05 -301 1969 602 29.22 12.92 1.29 0.57 -2.51 -0.39 -107 1970 607 25.61 8.29 1.37 -3.61 -4.63 0,08 -374 1971 620 23.65 12.80 1.47 -1.96 4.51 0,10 122 1972 655 27.01 7.74 4.70 3.36 -5.06 3,23 70 1973 666 28.42 7.12 5.13 1.42 -0.62 0,43 56 1974 695 21.26 4.19 6.40 -7.17 -2.93 1,27 -405 1975 737 29.04 5.15 8.48 7.78 0.96 2.08 496 1976 783 20.73 7.72 8.16 -8.31 2.57 -0.32 -278 1977 836 28.86 14.19 7.86 8.13 6.47 -0.30 656 1978 881 37,45 12.85 8.65 8.59 -1.34 0,79 369 1979 1,107 45.41 27.30 11.25 7,96 14.45 2,60 1147 1980 1,260 52.26 34.30 11.69 6.85 7.00 0,44 656 1981 1,246 43.35 30.15 11.60 -8.91 -4.15 -.09 -603 'Real Investment and capital are deflated with producer pnce Index for tieavy machinery, 1967 = 100. 346 KIRKLEY AND SQUIRES: CAPITAL STOCK AND INVESTMENT IN A FISHERY tonnage vessels or those which are <5 gross registered tons are excluded since data are unavailable; thus, total capital stock and invest- ment in New England fisheries are underesti- mated. Despite these limitations, the results provide information on investment and capital stock in New England fisheries which has not been avail- able. As is characteristic of open-access fisheries, investment in New England fisheries appears to have increased over time. This is quite evident subsequent to the passage of the Magnuson Fish- eries Conservation and Management Act of 1976. While possibly true, the observation that invest- ment increased when the act was passed should not be construed to imply that the act was respon- sible for increased investment. Positive expecta- tions regarding fish stocks or economic conditions could also be responsible for increases in capital stock and investment. The largest increase in investment occurred in 1979 which coincided with the greatest influx of vessels. It also coincided with increased fish stocks in 1978 and 1979 and high vessel profits of the previous year. In 1979, 226 vessels entered the New England fisheries; 92 vessels were newly constructed. Disinvestment also occurred, but the reasons are unknown. ^^ Disinvestment may be associated with vessel sinkings, foreclosures on vessel mort- gages, and exit from the region. Alternatively, if vessel prices reflect the expected net discounted present value of earnings from a fishery, the dis- investment or even changes in investment may be due to changes in economic expectations. Dis- investment occurred between 1968 and 1970, 1974, 1976, and in 1981. Disinvestment between 1968 and 1970 may re- flect expectations of declining future earnings. Disinvestment in 1974 may reflect the effects of high fuel prices. Disinvestment in 1976 and 1981 may be associated with fishermen's expectations of declining fish stocks, the effects of manage- ment, and high fuel prices. The reasons for investment are also unknown. Increasing levels of investment between 1977 and 1980 may be due to economic expectations associ- ated with fisheries management, fish stocks, and future earnings. The real capital stock increased approximately 6% per year between 1965 and 1981. The largest increase occurred in 1978 and 1979 when there was a substantial increase in the number of new steel-hulled vessels and entry of vessels from the mid-Atlantic region. Capital stock declined in 1981 when some of the mid-Atlantic vessels left New England and the stocks of fish, particularly sea scallops, declined (Northeast Fisheries Cen- ter 1985). The results also indicate why a vessel count should not be used to indicate capital stock or investment. The number of vessels increased be- tween 1969 and 1970 and 1975 and 1976, but investment decreased in 1970 and 1976 (Table 6). Alternatively, the number of vessels increased in each year relative to 1965, but the real capital stock between 1969 and 1976 declined with re- spect to 1965. Table 6. — Indices of capital stock based on constant dollar value and vessel count. Capital stock Constant dollar value Vessel count Year Trawler! Dredge' Total Trawler! Dredge! Total 1965 100 100 100 100 100 100 1966 111 91 105 102 94 104 1967 152 70 119 109 77 106 1968 110 95 103 104 106 103 1969 113 79 98 104 106 101 1970 99 51 80 108 74 102 1971 91 79 86 108 83 104 1972 104 48 89 108 74 110 1973 109 44 92 105 79 112 1974 82 26 72 107 49 117 1975 112 32 96 105 66 124 1976 80 47 83 109 202 132 1977 111 87 115 107 202 141 1978 144 79 133 116 189 148 1979 175 168 190 137 313 186 1980 201 211 222 158 440 212 1981 167 185 192 158 413 210 i2Hirshleifer (1970) provided additional information about disinvestment. 'Gear type assigned by plurality of days absent. The inadequacy of using vessel count is more pronounced when comparing the indices based on gear type. As shown in Table 6, the number of trawl vessels in every year exceeded the number of vessels in 1965 while capital stock was less than that of 1965 for four years. The two scallop dredge indices also indicate dissimilar changes between vessel count and real capital stock. It may, though, be possible than an index based on vessel count and weighted by vessel size would more closely compare to the real capital stock index. The construction of these indices would re- quire considerable additional analysis and deter- mination of the weights necessary for aggrega- tion. 347 FISHERY BULLETIN; VOL. 86, NO. 2 SUMMARY AND CONCLUSIONS This study developed procedures for estimating the capital stock and net investment using only the information generally available. That is, data on acquisition and list prices for some but not all vessels and characteristics for most vessels in a fleet were used to estimate capital stock and in- vestment in a fishery. An hedonic approach which specified the acquisition price as a function of ves- sel characteristics was suggested as a possible method for obtaining information necessary for estimating capital stock and net investment. It was argued that the approach was consistent with procedures used by economists to estimate the capital stock and investment in traditional indus- tries. Estimates were based on 1,110 observations ob- tained from NMFS and classified advertisements for the period 1965-81. Vessel characteristics pertaining to size and age were shown to be the more significant characteristics for explaining the value of a vessel. The corresponding informa- tion was then used to estimate the capital stock and investment in the New England fish- eries. An interesting result was that although invest- ment increased over time, there also was disin- vestment in the fisheries. It was suggested that this was possibly the result of vessel sinkings, exit of vessels from the region, foreclosures, and expectations of declining future net returns. It also was demonstrated that a vessel count should not be used to indicate capital stock or investment. The number of vessels generally in- creased over time, but the level of capital stock did not coincide with these changes. This suggests that more attention should be given to developing economic measures of capital stock and invest- ment. A question still in need of attention, however, cannot be answered from this study. That is, given the level of investment, can an optimal uti- lization and allocation of resources be deter- mined. The answer is clearly no. This requires management authorities to specify the objectives of fisheries management and a detailed bioeco- nomic model. There is a need for more research on invest- ment, and in particular, the determination of the optimum rate of investment. This includes re- search on the social discount rate, reasons for in- vestment, and the marginal productivity of capi- tal. More important, additional research is required to better address the issues of fisheries management, particularly whether or not public expenditures should be allocated to managing fisheries. ACKNOWLEDGMENTS Comments by William DuPaul, Wade Griffin, Ray Hilborn, Maurice Lynch, Ivar Strand, and two anonymous reviewers are gratefully ac- knowledged. Any errors remain the responsibil- ity of the authors. The article does not necessarily reflect the opinion of either the National Marine Fisheries Service or the Virginia Institute of Marine Science. LITERATURE CITED Baumol, W J 1977. Economic theory and operations analysis. Prentice-Hall, Englewood Cliffs, N.J. Blackorby, D., D Primont, and R R. Russel 1978. Duality, separability, and functional structure: the- ory and economic applications. North-Holland Publish- ing Company, Amsterdam. BocKSTAEL, N. E., W. M Hannemann, and I. E. Strand. 1986. Measuring the benefits of water quality improve- ments using recreation demand models. Volume II of ben- efit analysis using indirect or imputed mtirket methods. EPA contract No. CR-811043-01-0. U.S. Environmen- tal Protection Agency, Washington, D.C. Box, G E P , and D R Cox 1962. An analysis of transformations (with discussion). J. R. Stat. Soc, Ser. B:21 1-243. Braeutigam, R R., a. F Daughety, and M. A. Turnqutst. 1982. The estimation of a hybrid cost function for a rail- road firm. Rev. Econ. Stat. 64:394^04. Branson, W H 1972. Macroeconomic theory and policy. Harp)er & Row, Publishers, N.Y. Charles. A. T. 1983a. Optimal fisheries investmentxomparative dy- namics for a deterministic seasonal fishery. Can. J. Fish. Aquat. Sci. 40:2069-2079. 1983b. Optimal fisheries investment under uncertainty. Can. J. Fish. Aquat. Sci. 40:2080-2091. Chow, G C 1960. Tests of equality between subsets of coefficients in two linear regressions. Econometrica 28:591-607. Clark, C W., F. H. Clarke, and G R Munro. 1979. The optimal exploitation of renewable resource stocks: problems of irreversible investment. Econometrica 47:25-47. Cunningham, S., M. R. Dunn, and D. Whitmarsh. 1985. Fisheries Economics: an introduction. St. Mar- tin's Press, N.Y. Dadkhah, K. M 1984. Confidence interval predictions from a logarithmic model. Rev. Econ. Stat. 66:527-528. Dewar, M. E. 1983. Industry in trouble: the Federal government and 348 KIRKLEY AND SQUIRES: CAPITAL STOCK AND INVESTMENT IN A FISHERY New England fisheries. Temple University Press, Philadelphia, PA. DOERINGER, P B.. P I Moss, AND D. G. TERKLA. 1986. The New England fishing economy. The Univer- sity of Massachusetts Press, Amherst, MA. Gordon, H S. 1954. The economic theory of a common property re- source: the fishery. J. Polit. Econ. 62:124-142. Griffin, W. L.J P Nichols, R. G. Anderson. J E Buckner. and C M Adams. 1978. Costs and returns data: Texas shrimp trawlers. Gulf of Mexico 1974-75. Texas A&M Univ., Dep. Agric. Econ. tech. Rep. TAMU-SG-76-601. Griuches, Z 1971. Hedonic price indexes for automobiles: an econo- metric analysis of quality changes. In Price indexes and quality changes, p. 103-130. Harvard University Press, Cambridge, MA. HILBORN. R. 1985. Fleet dynamics and individual variation: why some people catch more fish than others. Can. J. Fish. Aquat. Sci. 32:2-13. Hirshleifer. J. 1970. Investment, interest and capital. Prentice-Hall, Englewood Cliffs, NJ. Kendrick, J 1961. Some theoretical issues in capital measurement. Am. Econ. Rev. 51(2):102-111. KMENTA, J. 1971. Elements of econometrics. Macmilliam Publish- ing Co., Inc., N.Y. KiRKLEY, J E , M P Pennington, and B. E. Brown. 1982. A short-term approach for analyzing the effects of harvesting quotiis. J. Cons. int. Explor. Mer 40:173- 175. KiRKLEY, J. E 1986. The relationship between management and the technology in a multi-species fishery. Ph.D. Thesis, Univ. Maryland, Collage Park, MD. Ladd, G W . AND M B Martin. 1976. Prices and demands for input characteristics. Am. J. Agric. Econ. 58:21-30. Lau, L J 1978. Applications of profit fiinctions. In M. Fuss and D. McFadden (editor), Production economics: a dual ap- proach to theory and applications, p. 133-216. North- Holland Publishing Company, Amsterdam. Maddala. G S. 1977. Econometrics. McGraw-Hill Book Company, N.Y. National Academy of Sciences 1979. Measurement and interpretation of productivity. National Research Council, Washington, D.C. Northeast Fisheries Center. 1985. Status of the fishery resources off the northeastern United States for 1985. U.S Dep. Commer., NCAA Tech. Memo. NMFS-F/NEC-42, 137 p. Northeast Fish- eries Center, National Marine Fisheries Service, NOAA, Woods Hole, MA. Rosen, S. 1974. Hedonic Prices and implicit markets:product differ- entiation in pure competition. J. Polit. Econ. 82:34-55. Scorr, A. 1979. Development of economic theory of fisheries regula- tion. J. Fish. Res. Board Can. 36:725-741. SISSENWINE, M P , and J E. KiRKLEY 1982. Fishery management techniques:practical aspects and limitations. Mar. Policy 6:22-42. Sloan, H S., and A. J. Zurcher. 1968. Dictionary of economics. Barnes & Noble, Inc., N.Y. Smallwood, D M., and J. R. Blaylock. 1986. Forecasting performance of models using the Box- Cox transformation. Agric. Econ. Res. 38:14-24. Tettey, E O . W L. Griffin, J. B. Penson, and J. R Stole. 1986. Implications of tax policy on investment in a com- mon property resource. N. Am. J. Fish. Manage. 6:100- 104. Triplett, J 1986. The economic interpretation of hedonic methods. Surv. Curr. Bus. 66:36-40. Varian, H. R. 1984. Microeconomic analysis. 2d, ed. W. W. Norton & Company, N.Y. 349 ANALYSES OF THE RELATIONSHIP BETWEEN THE DISTRIBUTION OF SEARCHING EFFORT, TUNA CATCHES, AND DOLPHIN SIGHTINGS WITHIN INDIVIDUAL PURSE SEINE CRUISES Tom Polacheck^ ABSTRACT The fine scale distribution of searching effort within individual purse seine cruises in the eastern tropical Pacific is analyzed in relationship to sightings of spotted dolphin, Stenella attenuata, and tuna catches. The data for these analyses were derived from detailed observations made by National Marine Fisheries Service observers aboard U.S. purse seiners. A clustering algorithm is developed which separates the activity of a vessel into areas where sets were common and areas where they are infrequent. Within clusters of high set densities, vessels tend to concentrate their searching effort. Vessels searched proportionately greater distances relative to the physical distances between sets while within clusters than when outside clusters. Encounter rates with schools of spotted dolphins tend to be either much higher or much lower within defined clusters than outside them. Clusters with low encounter rates were clusters in which non-dolphin associated tuna sets predominated. Because of this dichotomy in the magnitude of the dolphin encounter rates within clusters, overall encounter rates appeared to have relatively small biases if the concentration of searching effort within clusters is ignored. The average catch of tuna per set was higher within the defined clusters than between them. The overall results suggest that fine scale geographic effects need to be considered when using data from purse seiners to examine changes in relative abundances of either dolphins or tuna. Catch and effort data underlie most indices of abundances used for assessing the status of com- mercially exploited fish stocks. The validity of using catch and effort data from commercial har- vests has long been questioned because of the likelihood that fishermen concentrate their effort in areas of high fish densities (Helland-Hansen 1909). Yet, almost nothing is known about the allocation of effort by individual vessels. If data are available on the activity of individual vessels, these data have been mainly used for standard- ization of effort. The catch and effort activities within individual cruises have not been examined in detail (in part because the data for such an examination generally do not exist). The purpose of the present paper is to examine the searching behavior of tuna purse seiners in the eastern trop- ical Pacific (ETP) based on detailed data compiled by the National Marine Fisheries Service (NMFS). The main questions addressed are whether seiners concentrate their effort and what is the relation between searching behavior, en- iSouthwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038 and Department of Biology, University of Oregon, Eu- gene, OR 97403; present address: Northeast Fishery Center Woods Hole Laboratory, National Meirine Fisheries Service, NOAA, Woods Hole, MA 02543. Manuscript accepted January 1988. FISHERY BULLETIN: VOL. 86, NO. 2, 1988. counter rates for dolphins, and tuna catches. The maximization of profit is presumably a strong influence on the behavior of commercial fishermen. Upon leaving port, the catch rate real- ized by a fisherman is probably the most impor- tant factor affecting his profits. Given this orien- tation, it is reasonable to assume that fishermen have developed strategies of when and where to fish that increase their catch rates beyond that achieved by random search. There is little empir- ical information to support this assumption other than correlations, which have been noted for some fisheries, between the spatial distribution of catch rates and effort for the fleet as a whole (Gul- land 1955; Calkins 1963). In a fishery where the detection of fish depends upon visual cues, searching would be expected to be located in the vicinity of previously located fish, if there is a tendency for the underlying pop- ulation to be spatially clustered. The search path for a vessel in such a fishery might be expected to look something like the hypothetical one depicted in Figure 1. The amount of crisscrossing or zigzagging in the vicinity of a catch and the area over which the search extends would be expected to vary between fishermen. The solution to the optimal searching strategy for such a situation is nontrivial and depends upon information on the 351 FISHERY BULLETIN; VOL. 86, NO. 2 Figure l. — A hypothetical cruise track for a vessel searching for clustered prey. The location of catches are indicated by xs. underlying distribution of the fish population (Koopman 1980). The question of optimal search- ing strategies for fishermen has been receiving increased attention (Pazynich 1966; Salia and Flowers 1979; Clark and Mangel 1983). These studies are primarily theoretical at this time and their application to actual fisheries requires knowledge of the spatial distribution of the fish population. Locating schools of tuna (e.g., yellowfin, Thun- nus albacares, and skipjack, Katsuwonus pelamis) in the purse seine fishery in the ETP depends on visual cues. Fishermen use a variety of cues including birds which feed on the same prey as tuna, disturbances on the surface of the water, floating debris which frequently have as- sociated tuna, and schools of dolphins which are often associated with tuna (primarily yellowfin). Fishermen have names for the different types of sets depending upon what is associated with the tuna school. They refer to sets associated with floating debris as log sets, sets associated with dolphins as porpoise sets, and sets not associated with other animals (except possibly birds) as school sets. Fishermen when not engaged in a set usually spend their day actively searching for signs of tuna. They use 25 x binoculars to scan the water while the boat cruises at speeds generally between 10 and 12 knots. The distribution of schools of tuna, as well as schools of the most commonly associated dolphins (i.e., the spotted dolphin, Stenella attenuata), ap- pears to be spatially and temporally clustered within the ETP considered as a whole (Calkins and Chatwin 1967, 1971; Blackburn and Wil- liams 1975; Suzuki et al. 1978; Au et al. 1979^; Polacheck 1983). At finer geographic scales, there is little available information although Au et al. (fn. 2) suggested that schools of spotted dolphin tend to be locally concentrated in areas of conver- gences and fronts. Given the above observation, it is not surpris- ing that the detection or encounter process for tuna or dolphins does not, in general conform to a Poisson process when the distribution of search- ing times, searching distances, or physical dis- tances between nearest encounters are analyzed (Polacheck 1983; Allen and Punsely 1984). In such analyses it is impossible to separate or dis- tinguish the effects of nonrandom search from nonrandom distributions of tuna or dolphins. In the harvesting of tuna schools associated with dolphins, fishermen chase and capture the associated dolphins (Perrin 1968, 1969), and some dolphins may be incidently killed. NMFS, as part of its responsibility under the Marine Mammal Protection Act of 1972 for managing and monitor- ing the status of dolphin populations, placed trained observers aboard tuna purse seiners. From the data collected by these observers, ap- proximate cruise tracks can be drawn by connect- ing all positions that were recorded. Many of these approximate cruise tracks (e.g.. Figure 2) have superficially a strong similarity to the hypo- thetical one depicted in Figure 1. It was this sim- ilarity that provided the impetus for the analyses presented below. METHODS NMFS observers aboard tuna purse seiners col- lected a wide variety of information, both on the sightings of marine mammals and fishing opera- 2Au, D. W. K., W. L. Peryman, and W. F. Per- rin. 1979. Dolphin distribution and the relationship to envi- ronmental features in the eastern tropical Pacific. Natl. Mar. Fish. Serv., Southwest Fish. Cent., Adm. Rep. LJ-79-43, 59 p. 352 POLACHECK: DISTRIBUTION OF SEARCHING EFFORT Figure 2. — Examples of the approximate cruise track for two tuna purse seiners. Diamonds represent the location of a set or chase. No geographic coordinates are given, and the orientation of figures were rotated at random so as not to compromise any proprietary fishing information. The distance between tick marks equals 300 nautical miles. 353 FISHERY BULLETIN: VOL. 86, NO. 2 tions. Included in the recorded data are informa- tion on all changes in a vessel's activity; the loca- tion, tjqae, and catch of all purse seine sets; and the location and identity of all marine mammal sightings. In these data, a vessel's activity is clas- sified into one of five mutually exclusive cate- gories: searching, running, setting, chasing, or resting. Searching is defined to be whenever a vessel is moving and the crew are actively search- ing for signs of tuna; running, anytime the vessel is moving but not actively searching for signs of fish (e.g., moving locations at night); chasing, anytime schools of dolphins are being pursued before the net has begun to be set. More detailed descriptions of the available data, collection pro- cedures and their preparation for the analyses below can be found in Polacheck (1983, 1984^). The analyses in this paper were part of a larger project on the use of these observer data for as- sessing the relative abundances of dolphin stocks. As such, the emphasis in this paper is on the encounter rate for the most important dolphin species for the fishery (spotted dolphin), although catch rates for tuna are also considered. The re- sults presented in this paper are based on two different approaches for analyzing the data. The first method is a set of nearest neighbor calcula- tions, and the second is a cluster analysis. The nearest neighbor calculations were per- formed in order to get an indication whether ves- sels tend to search in the vicinity of a previous encounter (either a sighting of marine mammals or a set on tuna). In these calculations, the phys- ical distance between either the next or preceding encounter is compared with the distance to the nearest other encounter made within the entire cruise. Also, the proportion of times in which the nearest encounter is not either the next or preced- ing one is calculated. For a vessel that never re- turned to the area of an encounter, this propor- tion would equal 1. Similarly, if a vessel never returned to the area of an encounter, the ratio of the distance between either the next or preceding encounter and the distance to the nearest other encounter within an entire cruise would also equal 1. Note that the expected values for these proportions with random search are not necessar- ily 1. The expected value will be dependent both on the distribution of potential encounters and 3Polacheck, T. 1984. Documentation of the time sequen- tial files created ft-om the tuna boat observer data bases for analyzing relative abundances. Natl. Mar. Fish. Serv., South- west Fish. Cent., Adm. Rep. LJ-84-33, 26 p. the definition of random search (see Discussion). These calculations were performed separately for sets and chases for tuna and for the sightings of spotted dolphin. In performing these nearest neighbor calculations, the first and last encounter during a cruise were not included. The other main approach used for examining the data is a form of cluster analysis. When the sequences of distances between sets and chases within any cruise were examined, they appeared to be spatially and temporally clustered in the sense that sets and chases in which the distance to the next set or chase was small tended to be clumped sequentially. This observation led to the development of an algorithm for clustering sets and chases that were spatially and temporally related. Standard clustering algorithms were not appropriate in this situation because of the prob- lem of scaling spatial and temporal distances within a common metric (i.e., how much time should be equal to a given distance). Note that the term "clustered" or "clustered distribution" is used in this paper to refer to any distribution in which high- and low-density areas are more frequent than would be expected if the distribution was generated by a Poisson process. The term is not meant to refer to any particular nonhomogeneous process. A cluster is considered as an area of high density and should not be con- strued as referring to a discrete unit. The primary purpose of the clustering al- gorithm was to define areas which a fisherman might have thought to have a high density of potential fishing targets so that the searching be- havior of a vessel could be compared between these areas and outside them. This analysis ex- ploits the fact that the physical distance between events is partially independent of the distance that a vessel travels to locate them. Since the purpose of the algorithm was to define areas of potentially good fishing, chases of dolphin, as well as sets, have been included as events in the clus- tering algorithm. (Sets made for the purpose of washing the net were not used.) The clustering algorithm began with consideration of the dis- tance between the first and second set and/or chase. If this distance was less than a specified amount, then these two events were placed in the same cluster, and the distance between these two and the third event were examined. This specified amount will be referred to as the clustering parameter. If the distance between the third event and either of the events within the cluster was less than the value of the clustering parame- 354 POLACHECK: DISTRIBUTION OF SEARCHING EFFORT ter, this third event was included in the cluster and the fourth event was examined. Whenever a set or chase was found for which the distance be- tween it and all members of the last defined clus- ter exceeded the clustering parameter, a new cluster was formed. This process was repeated until all sets and chases within a cruise were placed in a cluster. Using this algorithm, all the activities within a cruise could be considered to occur either between or within clusters. Isolated sets or chases (i.e., clusters containing only a single event) were con- sidered as occurring between clusters. Distances travelled (i.e., distance searched and distance run) within a cluster were defined as the dis- tances travelled after the first set or chase until the last set or chase in that cluster. Distances travelled between clusters were defined from the last event of the previous cluster to the first event of the subsequent cluster. The distances travelled until the location of the first set or chase and after the last one were not included because of the large distances involved in reaching the fishing grounds. The location of a cluster was estimated by cal- culating the centroid for all sets and chases within it. The size of a cluster was estimated by determining the radius of the smallest circle with a center at the centroid that encompassed all sets and chases within it. The sensitivity of this algorithm to the value of the clustering parameter was examined for val- ues of 50, 75, 100, and 150 miles. For most of the results, only clusters with at least three members are considered as clusters. Clusters with only two members have been excluded from most of the summary results describing a cluster and also in the comparisons of results between and within clusters. This was done because two physically close events did not seem to warrant being called a cluster. Yet, given the relative difficulty in lo- cating potential sets, two close events might be considered as areas of potentially good fishing. Clusters with only two members contained 18% of all sets and chases when the cluster parameter equalled 50 miles and 6% of all sets when the value equalled 150 miles. The overall results and conclusions are robust to whether or not clusters with two members are included or excluded. An average intercluster distance for a cruise was calculated in order to get an indication of the stability of the clustering algorithm to the value of cluster parameters. The intercluster distance was defined as the distance from the nearest member of a cluster or isolated set to the next set and represents the minimum value that the clus- ter parameter would have to be for a cluster or isolated set to be combined in a single cluster with the next set. Encounter rates for schools of spotted dolphin for each cruise were calculated between and within clusters as the total number of sightings divided by the total distance searched. In these rates, if the first chase or set within a cluster was based on a sighting of spotted dolphin, this sight- ing was included in the encounter rates between and not within a cluster. For the analysis of these encounter rates, clusters were classified accord- ing to the percentage of the total number of sets and chases within a cluster that involved schools of dolphins. In calculating the number of events that occurred within a cluster, sequential non- dolphin sets in which no searching was done be- tween them were counted as a single event. This was done to reduce the effect of multiple sets on the same floating object counting as a large clus- ter. In order to see whether the searching behavior within the defined clusters resulted in biased esti- mates of encounter rates for dolphins if the clus- ters were ignored in the estimates, two different estimates for the overall encounter rate for a cruise were calculated and compared. The first estimate, which will be referred to as the unad- justed rate, was simply the total number of en- counters divided by the total distance searched for an entire cruise. This would be an unbiased estimate if search was in fact random. The second estimate, which will be referred to as the adjusted encounter rate, was calculated as the weighted average of the encounter rate within and between clusters. The weights for the encounter rate within clusters were equal to the diameter of the cluster. The weights between clusters equalled the total distance searched between clusters. In effect, this adjusted encounter rate is an estimate of what the encounter rate would have been if a vessel had made a straight line crossing of each cluster. (In the calculation of these adjusted en- counter rates, clusters with two members were treated the same as other clusters to simplify the calculations.) The analyses in this paper are based on 35 cruises. This represents a subset of the cruises in 1979 with NMFS observers, which in turn is a subset of all purse seine cruises for tuna in the ETP during 1979. Analyses were restricted to 1979 because only for this year have the data 355 FISHERY BULLETIN: VOL. 86, NO. 2 been carefully edited for positional errors. Cruises from 1979 were excluded either because of insufficient positional data or because the ves- sel made port stops during the middle of the cruise. Inclusion of the time spent searching on the way in and out of port could distort the results on searching and encounter rates between clus- ters. Preliminary analyses suggested that there is little difference in the results between vessels that went into port and those that did not. However, to avoid further complicating the analyses, these cruises have not been included (see Polacheck 1983 for more detail). RESULTS Is Effort Concentrated? The percentage of sets or chases for which the nearest one was not the preceding or next set or chase ranged from 12 to 77 among cruises (Fig. 3). The mean percentage was 41 (SE = 1.9, n = 35). The average ratio of the physical distances be- tween the next or preceding set or chase com- pared to the distance to the nearest set or chase within a cruise was 1.45 (SE = 0.044, n = 35) and ranged from 1.00 to 2.24. Consideration of the same statistics for the distances between sight- ings of spotted dolphin indicates an even more concentrated pattern. The mean percentage of sightings for which their nearest neighbor was not either the next or preceding one equalled 80 (SE = 2.8, n = 34, note one cruise recorded no sightings of spotted dolphins) and ranged from 25 to 100 (Fig. 4). The average ratio of the distance between the next or preceding sighting compared with the nearest sighting was 4.05 (SE = 0.261, n = 34) and ranged from 1.14 to 7.82. For these sighting statistics, low percentages and ratios near 1.00 are found in vessels with few sightings (Fig. 4). These results suggest that in general ves- sels return to the area of a previous sighting and/ or set and search in that area at least 41% of the time. There appears to be large differences among vessels in their ability and success at locating potential fishing targets. Thus, the average dis- tance searched between sets or chases varies by about a factor of 4 among cruises, while the aver- age physical distance between sets or chases varies by about a factor of 6 (Fig. 5). There is little relationship between the average distance searched between sets and chases and the aver- age distance to the next one. However, vessels 15 > u z U4 3 O m a u. 10 0 I 1 1 1 1 1 1 1 10 20 30 40 50 60 70 80 PERCENT OF THE SETS AND CHASES IN WHICH THE NEAREST ONE WAS NEITHER THE PRECEDING OR NEXT ONE Figure 3. — The frequency distribution for the percentage of the sets and chases within a cruise for which the nearest other set or chase was neither the preceding or next one. 15 10 UJ O 111 cc u. 5 - — I — 60 100 30 40 50 60 70 80 90 PERCENT OF SIGHTINGS IN WHICH THE NEAREST ONE WAS NEITHER THE PRECEDING OR NEXT ONE Figure 4. — The frequency distribution for the jjercentage of sightings of schools of spwtted dolphins, Stennella attenuata , within a cruise for which the nearest other sighting was neither the preceding or next one. Shaded portions represent cruises with less than 15 sightings (note one cruise had no sightings of spotted dolphins). 356 POLACHECK: DISTRIBUTION OF SEARCHING EFFORT with the largest distances between sets also have large average search distances, but the converse is not true (Fig. 5). The importance of dolphin fishing as compared to log and school fishing ap- pears not to be related to these differences among cruises (Fig 6). It should be noted that the points in Figure 5 must lie above the straight line with a slope of 1.0. The expected relationship between the variables in Figure 5 depends upon both the underlying searching process and the spatial dis- tribution of potential sets. While a positive rela- tionship would be expected if both of these are random, a more precise definition is beyond the scope of this paper. The purpose in presenting Figures 5 and 6 is to display the range of differ- ences in the success of vessels in locating poten- tial fishing targets. The large variation among cruises suggests that all vessels may not be using the same searching strategy and is an important factor to keep in mind when considering the re- sults from the clustering algorithm. ^ 220 1 ■5 200 S 180 u oc o »- Ui (A X u UI 160 140 120 O 100 I- D UI < UI s 180 160 140 120 100 80 60 40 20 0 • . • • •• • '. 10 20 30 40 50 60 70 80 PERCENT OF THE SETS AND CHASES INVOLVING DOLPHINS 90 Figure 5. — The relationship between the average distance searched between sets and chases within a cruise and the aver- age distance to the nearest one. Figure 6. — The relationship between the average distance searched between sets and chases and the proportion of sets and chases in which schools of dolphins were associated for a cruise. 357 FISHERY BULLETIN: VOL. 86, NO. 2 Table 1. — Statistics summarizing the results of the clustering algorithm showing the sensitivity to a range of values for the clustering parameter. Standard deviations are given as opposed to standard errors to provide an indication of the variation between cruises. All means are the simple average for 35 cruises. Cluster parameter 50 75 100 150 Percent of sets within clusters with at least 3 members Mean SD Range 64 14.1 30-90 75 13.5 31-97 81 13.3 31-97 88 9.2 60-100 Average cluster radius for clusters with at least 3 members (nautical miles) Mean SD Range 35 8.5 19-56 55 19.7 27-142 77 40.4 34-225 116 66.0 38-390 Percent of the search- ing within clusters with at least 3 members Mean SD Range 25.8 12.6 9-61 39 16.4 16-91 47 16.1 19-91 59 10.6 31-91 Average intercluster distance (nautical miles) Mean SD Range 210 65.9 134-425 261 75.9 155-465 297 82.2 173-513 356 100.9 196-691 30 r 20 z UJ O 10 A N = 767 30r 20 - 10 B N = 570 50 40 r 30 - UJ O 20 oc Ul a. 10 75 100 125 150 175 200 225 250 >250 75 100 125 150 175 200 225 250 275 >275 40r C N = 448 _L I I _L _L _L 30 20 10 D N = 309 _L _L _L _L _L _1_ 100 125 150 175 200 225 250 275 300 >300 INTER-CLUSTER DISTANCE (nautical mNaa) 150 175 200 225 250 275 300 325 350 >350 MTEH-CLUSTCR MSTANCC (nauUcal nritM) Figure 7. — The frequency distribution of intercluster distances. Distances from all cruises have been pooled. The values of the clustering parameter are A) 50, B) 75, C) 100, and D) 150. 358 POLACHECK: DISTRIBUTION OF SEARCHING EFFORT by a value of 50 for the clustering parameter and from 196 to 691 miles for a value of 150 miles (Table 1). The frequency distribution of these in- tercluster distances (Fig. 7) is an indication of the stability of the clusters to the value of the cluster- ing parameter. Thus, for a value of 100, over 65% of the clusters have an intercluster distance exceeding 175 miles. This suggests that 65% of clusters will be stable up to a value of 175 miles for the clustering parameter. (Note, this is not strictly true. If the set preceding the first mem- ber of a cluster was less than 175 miles away and this set was also less than 175 miles from the first set of the next cluster, this set plus these two clusters would be combined in a single cluster for a value of the cluster parameter less than 175.) These statistics describing the characteristics of the defined clusters suggest that the algorithm used to create them successfully separates the ac- tivities of a cruise into areas where sets are com- mon and areas where they are infrequent. The major differences in the clusters with different values for the clustering parameters result from the merging of two relatively close clusters or the inclusion of an isolated set or chase near the boundary of a cluster (e.g., for 80% of the cruises, the actual number of clusters decreases or re- mains the same over a range of 50-150 miles for the clustering parameter). However, the fact that many of these descriptive statistics vary continuously with the value of the clustering parameter suggests that these defined clusters do not represent distinct units, but areas of high concentration in a continuously grading system. Cruises vary greatly with respect to the amount of variability they exhibit in response to changes in the value of the clustering parameter (Table 2). Such variability is to be expected since no single searching strategy is used by all vessels and vessels may change their strategy during the course of a cruise. In addition, the spatial distri- bution of potential sets probably also varies with time and space. These sources of variability among cruises, combined with the relatively small sample sizes within a cruise, may be part of the reason that the descriptive statistics charac- terizing clusters vary continuously with the value of the clustering parameter. The lack of any sharp demarcation in the clus- ters as a function of the clustering parameter, combined with the large amount of variability exhibited among different cruises, creates a prob- lem in presenting results based on the clustering algorithm. Consequently, whenever summary statistics are presented, results are given for a range of values for the clustering parameter. Table 2. — Examples of the effect of changes in the value of the clustering parameter for 5 arbitrarily selected cruises. Average Percent of cluster radius Percent of Number of sets and for clusters the search- Average Value clusters chases with with at least ing within Intercluster of the with clusters with 3 members clusters with distance cluster at least at least (nautical at least (nautical Cruise parameter 3 members 3 members miles) 3 members miles) 1 50 4 90 32 61 171 75 1 96 142 91 298 100 1 96 142 91 298 150 1 96 142 91 298 2 50 10 85 27 34 210 75 7 97 61 60 294 100 4 97 105 68 397 150 4 98 126 71 443 3 50 5 30 27 9 402 75 5 31 36 16 450 100 5 31 46 19 479 150 5 60 68 36 614 4 50 4 78 30 36 425 75 4 80 39 41 465 100 4 82 38 45 513 150 4 82 38 45 513 5 50 1 65 45 38 302 75 3 79 45 43 362 100 3 79 45 43 362 150 3 79 45 43 362 359 FISHERY BULLETIN: VOL. 86, NO. 2 Whenever it seemed important to display the re- sults by cruise, a value of 100 miles for the clus- tering parameter was used. It should be empha- sized that the general patterns appear to be independent of the exact value of the clustering parameter. Is Searching DiflFerent in Clusters? To determine whether vessels tended to in- crease their searching effort in the vicinity of sets, the ratio of the total distance travelled (i.e., the distance searched plus the distance run) to the actual physical distances between sets was compared between and within clusters. This ratio is about 1.7 times greater in clusters than be- tween clusters (Table 3). Also for no more than 7 cruises is the ratio within clusters less than the ratio between clusters (Table 3). By a sign test (Snedecor and Cochran 1967), these results imply that this ratio is significantly greater than 1.00 (P < 0.005). In addition, the proportion of the total distance travelled that is devoted to searching is much greater when a vessel is between than when it is within clusters (Table 4). The fact that vessels run, and are not actively looking for tuna, propor- tionately more between than within clusters is an indication that clusters are areas in which a ves- sel has decided to remain. Vessels tend only to run at night. Usually, as long as the vessel is moving during daylight, the crew will be search- ing. Thus, large amounts of running tend to occur when vessels are actively moving to new areas. Table 3. — Comparison between and within clusters of the ratio of the total distance travelled (i.e., distance searched plus distance run) to the actual physical distance between sets and chases. Means are the average values of the ratio for the 35 cruises being considered. Value of the cluster parameter 50 75 100 150 Between Mean 1.79 1.70 1.65 1.59 SD 0.064 0.062 0.060 0.065 Range 1.24-2.70 1.13-2.71 1.12-2.68 1.11-2.81 Within Mean 3.08 2.84 2.79 2.68 SD 0.343 0.254 0.250 0.239 Range 1.13-11.03 1.16-8.67 1.09-8.67 1.12-8.67 Number of cruises out 128 131 132 131 of 35 in which the ratio is greater within than between clusters ip < 0.05 of observing this many ratio greater within clusters than between if they were in fact equal based on a sign test (Snedecor and Cochran 1967). Table 4. — Comparison between and within clusters of the ratio of the distance run without searching to the distance searched. Means are the average values of this proportion for the 35 cruises being considered. Value of the cluster parameter 50 75 100 150 Between clusters Mean ratio SE Range 0.85 0.063 0.35-1.94 0.93 0.068 0.26-1.94 0.97 0.067 0.26-2.05 1.06 0.095 0.26-3.45 Within clusters Mean ratio SE Range 0.29 0.040 0.01-0.88 0.33 0.036 0.04-0.98 0.38 0.042 0.04-1.13 0.46 0.048 0.05-1.06 Number of cruises out of a total of 35 in which the propor- tion is greater be- tween clusters than within them 134 134 134 134 ip < 0 001 of observing the portion being greater between clusters than within 34 out of 35 cruises, if in fact the proportion was equal based on a sign test (Snedecor and Cochran 1 967). 360 POLACHECK: DISTRIBUTION OF SEARCHING EFFORT However, some running can still be expected within clusters because vessels mark favorable logs with radio transmitter in order to return to them at a latter time. It should be noted that three or more sets or chases rarely occur on the same day (Fig. 8). Therefore, the proportionately smaller amount of running that occurs within clusters is not an artifact resulting from clusters that do not span more than a single day. Are Clusters Areas of High Spotted Dolphin Densities? Encounter rates {total number of sightings of spotted dolphins divided by the total distance searched) tend to be much greater within clusters dominated by sets in association with dolphins than either within clusters dominated by non- dolphin sets or while searching between clusters (Table 5). The clustering algorithm would tend to ensure that such differences are likely to occur, but the magnitude of the differences is large and 900 800 700 600 > o 500 LLI o lu 400 a. 300 200 100 ° 1 2 3 4 5 NUMBER OF SETS OR CHASES Figure 8. — The frequency distribution of the number of sets or chases during a day for all days on which there was at least one chase or set. Table 5. — Comparison of encounter rates (number of schools of spotted dolphins, Stenella attenuata, per 100 miles searched) while searching within clusters classified according to the percentage of sets involving dolphins, and while searching between clusters. Unweighted means are the average value of the encounter within each cruise. Weighted means are based on the encounter rate weighted by the distance searched by a cruise and are equivalent to the total number of sightings divided by the total distance searched pooled across all vessels. Within clusters in which the percent of the cluster parameter of sets on dolphins ranged from Between Value 0-25 25-75 75-100 clusters 50 Unweighted mean 0.17 1.45 2.74 0.52 SE 0.070 0.182 0.219 0.046 Weighted mean 0.10 1.04 2.46 0.50 SE 0.047 0.120 0.205 0.046 Number of cruises 27 22 30 35 Total distance searched 17.9 10.0 23.1 148.2 (thousands of miles) 75 Unweighted mean 0.15 1.32 2.32 0.47 SE 0.051 0.191 0.188 0.044 Weighted mean 0.12 0.88 2.16 0.46 SE 0.039 0.111 0.166 0.044 Number of cruises 25 22 31 35 Total distance searched 28.2 15.1 33.0 125.4 (thousands of miles) 100 Unweighted mean 0.14 1.15 2.08 0.46 SE 0.046 0.181 0.174 0.045 Weighted mean 0.15 0.81 1.82 0.45 SE 0.043 0.117 0.148 0.044 Number of cruises 25 20 31 35 Total distance searched 35.1 16.0 45.0 106.7 (thousands of miles) 150 Unweighted mean 0.10 0.73 1.83 0.43 SE 0.040 0.106 0.159 0.049 Weighted mean 0.11 0.68 1.63 0.40 SE 0.037 0.096 0.114 0.041 Number of cruises 24 18 32 35 Total distance searched 33.9 27.2 57.1 85.7 (thousands of miles) 361 FISHERY BULLETIN: VOL. 86, NO. 2 relatively insensitive to the value of the cluster- ing parameter. However, the fact that a large per- centage of the clusters tend to be dominated by either dolphin or non-dolphin sets is not a neces- sary consequence of the clustering algorithm and suggests that the two types of methods for locat- ing and catching tuna tend to be spatially and temporally distinct. Encounter rates are substan- tially lower in clusters dominated by non-dolphin sets than when a vessel is searching between clusters (Table 5). This result is also not a neces- sary consequence of the clustering procedure and suggests that these clusters not only define areas of high densities of spotted dolphin schools but also areas of low densities. Encounter rates could be lower in non-dolphin areas because of differences in detectability not related to the density of schools. For example, Hammond'* suggested that the crew may scan closer to the vessel when searching in non- dolphin areas (see also Polacheck 1983). It seems unlikely that such factors could account for all of the differences between the encounter rates in Table 5. Differences in detectability due to differences in weather conditions between and within clus- ters could also affect the results in Table 5. En- counter rates do decrease at higher Beaufort sea states (Polacheck 1983). However, little search- ing occurs above Beaufort state 4. The difference in encounter rates at Beaufort 0-2 compared with Beaufort 3-4 (about a factor of 1.28) is insuffi- cient to explain the difference in Table 5 (Polacheck 1983). Moreover, areas of non-dolphin sets tend to be in nearshore areas with calmer seas and fishermen do not consider Beaufort 4 conditions as being too rough to fish. Do Tuna Catches Differ Between and Within Clusters? The average tons of tuna caught per set tend to be greater for sets which occur within clusters than sets between clusters (Fig. 9). For all values of the clustering parameter, the average catch per set was greater within than outside of clusters in approximately 70% of the cruises (Table 6). Para- metric statistical comparisons of the average tons per set are not appropriate because of the large differences in the average catch per set among vessels (Fig. 9). A nonparametric sign test (Sned- cor and Cochran 1966) suggests that the differ- ences in catch per set between and within clusters are significant at least at the 0.05 probability level for all values of the clustering parameters that were considered. 60 t- 56 CO QC liJ t- V) u _l o z UJ UJ I- LLI UJ Ui QC LU a I- X o < O 180 30 60 90 120 150 180 >180 40 30 lU O 20 c 10 C N=214 X _L X X _r 30 60 90 120 150 180 >180 CLUSTER RADIUS (nautical mllaa) 40r- 30- 20 - 10 ~ - ] D N=184 1 1 1 gj;;: 1 1 t 1 I 30 60 90 120 150 180 >180 CLUSTER RADIUS (nautical miles) Figure 11. — The frequency distribution for estimated radii for all clusters with at least three members. Estimates from all cruises have been pooled. The values of the clustering parameter are A) 50, B) 75, C) 100, and D) 150. 365 FISHERY BULLETIN. VOL. 86, NO. 2 a 2° stratification would be necessary to avoid biases in these areas. ACKNOWLEDGMENTS I wish to thank Peter Frank and Tim Smith for their generous support and useful comments dur- ing the research presented here. This paper would not exist without the data collected by the Na- tional Marine Fisheries Service observers. I wish to acknowledge these observers and the staff of the Southwest Fisheries Center of the National Marine Fisheries Service for collecting, editing, and initially preparing these data. A number of the staff of the Southwest Fisheries Center pro- vided support and technical assistance at various stages in this research. The number of people in- volved is too large to acknowledge all of them individually. Special thanks go to Phil Tarentino, Michael Lichter, and Susan Boyer. The art department of the Southwest Fisheries Center were responsible for the final preparation of some of the figures. Veronica van Kouwen helped with the preparation of the final manuscript. The manuscript is a portion of a Ph.D. Thesis submitted to the University of Oregon, Eugene, OR. LITERATURE CITED Allen, R., and R. Punsley. 1984. Catch rates as indices of abundance of yellowfin tuna, Thunnus albacares , in the eastern Pacific Ocean. Inter-Am. Trop. Tuna Comm. 18:303-379. Blackburn, M., and F. Williams. 1975. Distribution and ecology of skipjack tuna, Katsu- wonus pelamis , in an offshore area of the eastern tropical Pacific. Fish. Bull., U.S. 73:382-411. Calkins, T. P. 1963. An examination of fluctuations in the "concen- tration index" of purse seiners and bait boats in the fish- ery for tropical tuna in the eastern Pacific. Inter-Am. Trop. Tuna Comm. Bull. 8:255-316. Calkins. T. P., and B M Chadwin. 1967. Geographical distribution of yellowfin tuna and skipjack catches in the eastern Pacific Ocean, by quarters of the year 1963-1966. Inter-Am. Trop. Tuna Comm. Bull. 12:435-508. 1971. Geographical catch of yellowfin and skipjack tuna in the eastern tropical Pacific Ocean, 1967-1970, and fleet and total catch statistics. Inter-Am. Trop. Tuna Comm. Bull. 15:285-377. Clark. C. W., and M. Mangel. 1983. Search theory in ecology and resouce manage- ment. In M. Freedman (editor). Proceedings of the In- ternational Conference on Population Biology, p. 380- 388. Springer Lecture Notes in Biomathematics, Vol. 52. Springer Verlag, N.Y. Cooper. E 1979. Home range criteria based on temporal stability of areal occupation. J. Theor. Biol. 73:687-695. GULLAND, J. A 1955. Estimation of growth and mortality in commerical fish fxipulations. Fish. Invest. Minist. Agric. Fish. Food (G.B.), Series 2. 18:1-46. Helland Hansen, B 1909. Statistical research into the biology of the haddock and cod in the North Sea. Rapp. P-v R6un 10:1-62. Koopman, B O 1980. Search and screening. Pargamon Press, N.Y., 509 p. Orbach, M K 1975. The cultural system of the tuna seinermen of San Diego, California. Ph.D Thesis, Univ. California, San Diego, CA, 411 p. Pazynich, G I 1966. Primery primeneniya teorii statisticheskiskh resh- enii k zadacham taktiki promysla (Examples of the appli- cation of statistical-solution theory to problems in the fishing tactics). Noe Khoz. 6:80-81. (Translation available in the library of National Marine Fisheries Service Northeast Fishery Laboratory, Woods Hole, MA.) Perrin, W. F. 1968. The porpoise and the tuna. Sea Frontiers 14:166- 174. 1969. Using porpoise to catch tuna. World Fishing 18:42-45. POLACHECK, T 1983. The relative abundance of dolphins in the eastern tropical Pacific based on encounter rates with tuna purse seiners. Ph.D Thesis, Univ. Oregon, Eugene, OR, 444 p. Sall\, S. B., AND J M. Flowers. 1979. Elementary applications of search theory to fishing tactics as related to some aspects of fish behavior. F.A.O. Conference on Fish Behavior in Relation to Fish- ing Techniques and Tactics, p. 343-355. Sanderson, G. 1966. The study of mammal movements - a review. J. Wildl. Manage. 30:215-235. SCHOENER, T W 1981. An empirically based estimate of home range. Theor. Popul. Biol. 20:281-325. Snedecor, G W , AND W C. Cochran 1967. Statistical methods. 6th ed. Iowa State Univ. Press, Ames, 593 p. Suzuki, Z , P K. Tomlinson, and M Honma. 1978. Population structure of Pacific yellowfin tuna. Inter-Am. Trop. Tuna Comm. Bull. 17:277-411. SwiHART, R. K . and N. a. Slade. 1985. Influence of sampling interval on estimates of home range size. J. Wildl. Manage. 49:1019-1025. 366 A COMPREHENSIVE THEORY ON THE ETIOLOGY OF BURNT TUNA Cheryl Watson,' Robert E. Bourke,^ and Richard W Brill^ ABSTRACT Over the past 14 years, the Hawaii handline fishery has experienced phenomenal growth in the catch for large yellowfin tuna, Thunnus albacares, and bigeye tuna, Thunnus obesus. These fish are primarily caught for the sashimi (raw consumption) market but have been continually plagued with a product quality problem known as "burnt tuna" or, in Japanese, yake niku. Not only does this problem significantly reduce the value of the catch, it also limits export markets and expansion of this low-capital, high-return fishery to other areas of the Pacific. Previous research and suggestions for mitigating burnt tuna have centered on the hypothesis that it is caused by high muscle temperature and low pH, which is the result of a violent struggle during capture. A new, more comprehensive hypothesis is presented: Burnt tuna is actually caused by the p)ost- mortem activation of enzymes known as calcium-activated proteases and by the enhancement of the effect of these enzymes by high blood catecholamine levels. Previously unexplainable observations, such as the propensity of female fish to become burnt more often during the summer months, the efficacy of brain destruction in preventing burnt tuna, and the lack of effect of cooling on the incidence of burnt tuna, are explainable in light of this new hypothesis. One of the largest fisheries in Hawaii is the hand- line fishery for large (>50 kg) yellowfin tuna, Thunnus albacares, and bigeye tuna, T. obesus, caught primarily for raw consumption as sashimi. Yearly landings increased from 89 short tons (ex- vessel value, $131,000) in 1973 to 615 short tons (ex-vessel value, $2.1 million) in 1984 (Yuen 1979; Hudgins and Pooley 1987). The total eco- nomic value of the fishery has been estimated as high as $5 million yearly (Ikehara'*). In Hawaii, the night handline fishery is known as ika shibi from the Japanese words for squid and tuna, and the daytime fishery is known as palu ahi from the Hawaiian words for chum and yellowfin tuna. There is also growing international interest in this tjTDe of fishing because of its low initial capi- tal investment, low operating and fixed expenses, strong export markets, and high profitability (Strong 1979; Gibson 1981; Jerrett 1984). Boats can be as small as 6 m and require only one- or two-man crew. Catch rates in Hawaii have ranged from two fish per hook per night (Yuen 'Pacific Gamefish Research Foundation, 74-425 Kealakehe Parkway, #15, Kailua-Kona, HI 96740; present address: De- partment of Physiology, John A. Bums School of Medicine, Uni- versity of Hawaii, Honolulu, HI 96822. 2465B Kawailoa Road, Kailua, HI 96734. 3Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 96822-2396. ^Ikehara, W. N. 1981. A survey of the ika-shibi fishery in the State of Hawaii, 1980. Southwest Fish. Cent. Adm. Rep. H-82-4C, 12 p. Southwest Fisheries Center Honolulu Labora- tory, National Marine Fisheries Service, NOAA, Honolulu, HI 96822-2396. 1979) to one-half fish per hook per night (Bourke^), illustrating the profitability for the in- dividual fisherman. Unfortunately, handline (as well as primarily recreational trolling) fishermen are plagued by a product quality problem known as "burnt tuna" or, in Japanese, as yake niku (literally translated as "cooked meat"). When fish are intended for raw consumption, product quality is of utmost impor- tance. Prime quality tuna flesh should be red, translucent, and firm and have a delicate flavor. Burnt tuna is pale, exudes a clear fluid, and has a soft texture and a slightly sour taste. Although perfectly palatable when cooked or canned, burnt tuna is considered unsuitable for raw consump- tion and commands only a fraction of the price of prime quality fish. Fish are usually exported whole to preserve freshness, so burnt tuna often is not detected until shipping costs have been in- curred. This discourages exports of tuna caught from areas or via fishing techniques with high incidences of the problem. Burnt tuna aflects from 5 to 100% of the tissue from an individual fish and ranges from mild to severe. Approximately 25% of the fish caught by the Hawaii handline fishery are burnt, as are 50% of the large yellowfin tuna caught by com- mercial and recreational trollers (Bourke fn. 5). ^Bourke, R. E. 1985. Hilo ikashibi fishery 1984 survey: Problems of a maturing fishery and a potential solution to the burnt tuna problem. Poster presentation at 1985 Tuna Confer- Manuscript accepted January 1988. fishery BULLETIN: VOL. 86, NO. 2, 1988. 367 FISHERY BULLETIN: VOL. 86, NO. 2 Annual losses to the handline fishery are an esti- mated 169^ of the value of the total catch (Cramer et al. 1981). The problem has also been reported in fish intended for raw consumption caught by purse seine (Nakamura et al. 1977). Surprisingly, the problem rarely occurs in fish caught by longlining (Williams 1986). PAST RESEARCH ON BURNT TUNA The Japanese were the first to investigate the causes of burnt tuna and possible mitigating strategies (Itokawa 1968, 1969). The first con- trolled laboratory investigations were those of Nakamura et al. (1977) and Konagaya and Kona- gaya (1978, 1979). Nakamura et al. (1977) con- cluded that high muscle temperature and low muscle pH caused myofibrillar protein denatura- tion and also noticed that, once denaturation began, it continued even if the tissue was kept at 0°C. Because of the relatively high thermostabil- ity of tuna myofibrillar protein and because yake niku occurs in species (e.g., frigate mackerel and sardine) that do not generate high muscle tem- peratures during struggling, Konagaya and Kon- agaya (1978) concluded that acid denaturation of myofibrillar proteins at moderate temperatures was the underlying cause. Cramer et al. (1981) studied handline-caught yellowfin tuna in Hawaii and found that the oc- currence of burnt tuna did not correlate with muscle temperature at time of landing and corre- lated only loosely with extracellular muscle pH. Ikehara^ conducted an engineering study to de- velop methods to cool large yellowfin tuna more rapidly, the presumption being that rapid cooling would prevent muscle degradation. Although suc- cessful in developing a technique to increase cool- ing of deep muscle temperature, as shown in Fig- ure 1, there was no apparent correlation between rate of cooling and incidence of burnt flesh. In spite of this lack of directly observed correlation, some publications designed for fishermen still stress that high muscle temperatures and low muscle pH are the prime causes of burnt tuna (Gibson 1981). Others have expressed doubt as to the validity of this hypothesis (Jerrett 1984). The high muscle temperature-low pH hypothe- sis appears to fit with what is known about tuna physiology, in that these fishes are capable of pro- ducing muscle temperatures significantly above ambient (Carey et al. 1971; Carey 1973) and ex- hibit some of the highest rates of muscle glycoly- sis (production of muscle lactate and concomitant production of acidity) observed in nature (Hochachka et al. 1978; Hochachka and Momm- sen 1983). Yet some observations do not fit this hypothesis. For example, burnt tuna occurs more frequently in summer, more frequently in female fish, and more frequently in fish fought for short periods of time than in fish fought for long periods (>7 minutes or <2 hours) (Davie and Sparksman 1986; Nakamura^). Furthermore, burnt tuna oc- curs rarely in longline-caught fish and in fish subjected to brain or spinal column destruction immediately following capture (Nakamura et al. 1977; [Suisan Sekai] 1977; Cramer et al. 1981; Davie and Sparksman 1986; Nakamura fn. 7). The hypothesis that burnt tuna is caused by high muscle temperature and low muscle pH does not seem to directly fit with any of these observa- tions. A NEW ANALYSIS OF THE BURNT TUNA PROBLEM At the biochemical level, the high muscle temperature-low pH h3rpothesis would predict that the observed drop in extracellular pH would be accompanied by a similar drop in intracellular pH and activation of lysosomal proteases. These proteases would then degrade actin and myosin, the dominant muscular proteins, resulting in the '^R. Nakamura, Department of Animal Sciences, University of Hawaii, Honolulu, HI 96822, pers. commun. October 1987. o LJ I— < a: o I/) Q. LiJ □ UKN 1 30.0 ^^=!5-- NOT BURNT ^^> "~^^^^^ •-« .^^_^^6.2) 20.0 \ s s \ ~~^~~^ (64.4) (M.9) \s._^ N. (103.0) ^ - ^ ~(62.1) "(61.7) 10.0- ■ X,,^^ 9.0 8.0 (99.8 ) 7.0- 1 — 1 1- — 1 1 H 1 1 1 1 10 11 ELAPSED TIME (h) ^Ikehara, W. N. 1981. Development of a small-boat chill- ing system for the reduction of burnt tuna. Final Report for a Pacific Tuna Development Foundation contract, var. pag. Figure 1. — Data are from Ikehara (text fn. 6). Initial and final deep muscle temperatures are plotted on semilogarithmic axes to linearize the rate of temperature change. Numbers in paren- theses are the body weight (kg) for each fish. 368 WATSON ET AL : ETIOLOGY OF BURNT TUNA undesirable changes seen in burnt tuna. The lyso- somal proteases, specifically cathepsins B, D, and L, have an optimum pH of about 5 (Dahlmann et al. 1984). However, recent work by Abe et al. (1985, 1986) has shown that, intracellularly, tuna muscle is well buffered and that a fall in extracel- lular pH is not necessarily accompanied by an equivalent fall in intracellular pH. Using is- chemic rat gastrocnemius muscle, Hagberg (1985) found that a 1 pH unit drop (7.30-6.36) in extracellular pH was accompanied by a drop in intracellular pH of only 0.4 unit (7.00-6.60). Thus, the acidic intracellular environment that would favor the action of the lysosomal proteases probably is not present in burnt tuna. Also, with the exception of calcium-activated neutral protease, all other muscle proteases (cathepsins B, D, L; alkaline serine protease; neutral trypsin- like protease; and alkaline cysteinyl protease) de- grade myosin (Dahlmann et al. 1984). Yet Hochachka and Brill (1987) found burnt muscle had no increase in 3-methyl-histidine, a specific marker for myosin breakdown. Decomposition of other myofibrillar proteins is, therefore, impli- cated. Electronmicrographs of postmortem burnt tuna muscle (Davie and Sparksman 1986) showed a consistent, rapid disintegration of Z-discs and ir- regularities in the sarcoplasmic reticulum (SR). The changes in burnt muscle were not different in kind from postmortem changes seen in unburnt tissue, but were a result of a significant increase in the rate of disintegration. Selected destruction of the Z-discs, troponin and tropomyosin, and the SR is characteristic of a pair of proteases known as calcium-activated neutral proteases (CANP's) (Sugita et al. 1984; Suzuki et al. 1984). These proteases are cytoplasmic, ubiquitous, activated by increased intracellular calcium levels, and ac- tive at pH 5.5-8.0 (Sakamoto and Seki 1985; Koohmaraie et al. 1986; Seki and Kimura 1986; Zeece et al. 1986b). The intracellular pH most likely found in burnt tuna muscle is, therefore, more consistent with CANP action than with lysosomal proteases whose activity requires a pH closer to 4.5 (Hochachka and Brill 1987). In addi- tion, while cathepsin D's activity is greatly re- stricted at 15°C, CANP is still active at 5°C (Koohmaraie et al. 1986; Zeece et al. 1986a). A new etiology of burnt tuna proposed by Hochachka and Brill (1987) is summarized in Fig- ure 2. Their hypothesis predicts that low intracel- lular ATP concentrations lead to the leaking of Ca"^^ into the cell and increases in intracellular INTENSE MUSCLE ACTIVITY I LACK OF 02 + ATP i METABOUC COLLAPSE OF MEMBRANE i RISE IN INTRACELLULAR CALCIUM i 02 i PSE i LLU i IN 0 i ACTIVATION OF CANP i UNDESIRABLE MUSCLE TEXTURE Figure 2. — Biochemical reac- tions involved in the development of burnt tuna as proposed by Hochachka and Brill (1987). Ca"^^ concentrations. These increases, in turn, ac- tivate CANP, which specifically attacks troponin, tropomyosin, SR, and mitochondria. The break- down of the latter two intracellular organelles releases more calcium into the cytoplasm, thus further increasing the activity of CANP. The effect of brain and spinal cord destruction on reducing the incidence of burnt tuna and of similar muscle degradation seen in other fish spe- cies can be explained by this new hypothesis, which assumes the initial drop in ATP is the root cause of the elevated intracellular calcium (Amano et al. 1953; Fujimaki and Kojo 1953; Konagaya and Konagaya 1978; Ikehara fn. 6). Brain destruction maintains elevated muscle ATP levels after capture (Boyd et al. 1984). Data recently collected on the use of brain and spinal cord destruction in large yellowfin tuna (Fig. 3) 40 I- I 35 m g 30 m I 20 o Sl5 z o UJ Q. 10-- 5- 1 4 FISHERMEN NOT USING BRAIN DESTRUCTION A -- ^ ^\ /\ SINGLE FISHERMAN ^ ■^ \ / \ ^ r / \ 1 \ r~^^^^ \ 1 \ / \ 1 \ / \ 1 \ 1 \ 1 l\ \ J \ 1 / \ \ 1 \ / 1 ^ \ \ 1 \ 1 \ --^ \ 1 / 1 U "^ "■--. ^' \ 1 I BRAIN DESTRUCTION USED \ 1 1 / \ (■ 1 1— +- H 1 1 1 1 1 \ JAN MAR MAY JUL MONTH SEP NOV Figure 3. — Incidence of burnt tuna occurring in large yellowfin tuna caught by 14 commercial handline fishermen operating out of Hilo, HI, January-September 1984, and by 1 fisherman who began using brain destruction in May. His incidence offish becoming burnt dropped dramatically in spite of the normal summer increased incidence seen in the catch of 14 other fisher- men. 369 FISHERY BULLETIN: VOL. 86, NO. 2 clearly show the effectiveness of this technique in preventing burnt tuna. A drop from 30 to < 10% of the tuna becoming burnt was noted when the fisherman changed his killing technique from shooting to the use of a brain spike. A MORE COMPREHENSIVE HYPOTHESIS ON THE ETIOLOGY OF BURNT TUNA The hypothesis presented by Hochachka and Brill (1987) does not explain all the observed fac- tors leading to tunas' different propensity to be- come burnt. It cannot explain why tunas strug- gling for <7 minutes rarely become burnt (Nakamura fn. 7), nor can it account for the obser- vations of Davie and Sparksman (1986) and Bourke (unpubl. data), who found that tuna struggling for extended (>1 hour) periods of time have a lower probability of becoming burnt. It also cannot explain why female fish become burnt more often than males (Nakamura et al. 1977), or why fish caught in summer become burnt more often (Fig. 3). Our revised hypothesis is presented in Figure 4. We have incorporated into the Hochachka and Brill (1987) hypothesis, the action of the neuro- transmitters-hormones norepinephrine (NE) and epinephrine (E) (collectively referred to as cate- cholamines). Periods of intense physical activity or capture stress increase the levels of circulating catecholamines by about 10-200 times in fish other than tunas (Nakano and Tomlinson 1967; Ling and Wells 1985). Most likely a similar situa- tion occurs in tuna during hooking, fighting, and capture. The importance of elevated circulating catecholamines in this particular schema comes INTENSE MUSCLE ACTIVITY i LACK OF 02 + ATP i METABOUC COLLAPSE OF MEMBRANE i RISE IN INTRACELLULAR CALCIUM i ACTIVATION OF CANP STRESS OF CAPTURE NE & E RELEASE ■INCREASE OF CANP ACTION ■ from their potentiating effect on the action of CANP. Catecholamines cause the phosphoryla- tion of troponin, resulting in a more rapid and prolonged proteolysis of this muscle structure (Toyo-oka 1982); in other words, high circulating levels of catecholamines greatly increase the ef- fectiveness of the enzymes responsible for a tuna becoming burnt. ^ The gills are the organ primarily responsible for the degradation of catecholamines, and circu- lating catecholamines are rapidly cleared from the blood (Nekvasil and Olson 1986). We estimate circulating catecholamines in tuna have a half- life of 30 minutes or less. Therefore, we hypothe- size that, in a longline-caught fish remaining in the water for several hours after hooking (Davie and Sparksman 1986), plasma catecholamines are reduced to low levels before the fish is landed and killed, thus resulting in a low percentage of burnt fish. Similarly, fish caught by rod and reel or by handline and fought for several hours would have the low circulating catecholamine levels and low propensity to become burnt. Indeed, Gibson (1981) recommends that handline-caught fish be attached to a buoy and left for an hour prior to being brought on board and killed, as a measure to prevent burnt tuna. That this technique pre- sumably lowers blood catecholamine levels prior to death could explain its efficacy. On the other hand, we hypothesize that a fish landed during the peak of its blood catecholamine concentra- tions would have the greatest propensity for be- coming burnt. This indeed appears to be the case. The mechanism that ties catecholamines to the observed seasonality and increased number of fe- male fish becoming burnt lies in the biochemical structural similarity of catecholamines and re- productive steroids, particularly estrogen. These steroids reach a peak during spawning season, which is May through October for yellowfin tuna in Hawaii (June 1960). A corresponding increase occurs in the percentage of the tuna catch that becomes burnt during this season (Fig. 3). In the presence of tyrosine hydroxylase, estrogen is con- verted to catecholestrogen and consequently com- petes for the same degradative enzymes in the gills, therefore slowing the clearance of cate- cholamines from the blood (Nekvasil and Olson 1986). Female fish, with high circulating estro- gen levels, would then be expected either to reach UNDESIRABLE MUSCLE TEXTURE Figure 4. — Comprehensive theory of the etiology of burnt tuna proposed in this paper. ^Although Cortisol is also known to be released during stress, its mechanism of action is much slower than NE or E, making it unlikely that it could exert catabolic effects within 15 min- utes. 370 WATSON ET AL.: ETIOLOGY OF BURNT TUNA higher maximum circulating catecholamine lev- els or to maintain high levels for longer periods, thus explaining the greater number of female fish that become burnt and the seasonality of the oc- currence of burnt tuna. As pointed out by Davie and Sparksman (1986) and Hochachka and Brill (1987), burnt tuna is a quantitative change in tissue decomposition, not a qualitative one. High blood catecholamine lev- els at the time of death act as accelerators in this inevitable metabolic cascade. However, because we are dealing with a change in rate, the action of catalysts can make all the difference as to whether or not a fish becomes burnt due to CANP. FUTURE RESEARCH Under normal conditions, CANP is inhibited by calpastatins. This class of proteins is distinct from the other cysteine protease inhibitors, the cys- tatins. Although the cystatins can inhibit cathep- sins B and H, they are unable to alter CANP activity. Conversely, calpastatin is only effective against calcium-activated neutral protease (Bar- rett et al. 1986; Parkes 1986). Unfortunately, al- most nothing else is known about the relationship of CANP to its endogenous inhibitor, either struc- turally or physiologically. Certainly investiga- tions into the role of calpastatin would prove valuable for this research. Immediate future research will concentrate on tracking the specific action of CANP in burnt tuna and attempting to stop this action with Ca^ ^ chelating agents such as EGTA (ethyleneglycol- bis-(aminoethyl ether )-N,N,N',N'-tetraacetic acid) or by use of its intracellular inhibitor, cal- pastatin. Also, blood catecholamine levels of stressed and unstressed fish will be measured, along with metabolic clearance rates of nore- pinephrine and epinephrine. When viewed as a process of metabolic deregu- lation of CANP, the rapid deterioration of tuna muscle ceases to be an isolated muscular phe- nomenon. The Z-disc disintegration characteris- tic of burnt tuna is also present in cardiac muscle injury due to ischemia and muscular dystrophy (Sugita et al. 1984). Given the highly conserved nature of muscle tissue, an understanding of burnt tuna may also provide insights into the metabolic processes of human disease. ACKNOWLEDGMENTS This research was supported by a grant to Pacific Gamefish Research Foundation from the Department of Land and Natural Resources, Di- vision of Aquatic Resources, State of Hawaii. LITERATURE CITED Abe, H , R W Brill, and P W Hochachka. 1986. Metabolism of L-histidine, camosine, and anserine in skipjack tuna. Physiol. Zool. 59:439-450. Abe, H., G P Dobson, U Hoeger, and W. S. Parkhouse. 1985. Role of histidine-related compounds to intracellular buffering in fish skeletal muscle. Am. J. Physiol. 249:R449-R454. Amano, K., M Bito, and T Kawabata 1953. Handling effect upon biochemical change in the fish muscle immediately after catch — I. Difference of gly- colysis in the frigate mackerel killed by various meth- ods. Bull. Jpn. Soc. Sci. Fish. 19:487-498. Barrett, A J . N D Rawlings, M E. Davies. W. Machleidt, G Salvesen, and V Turk 1986. Cysteine proteinase inhibitors of the cystatin super- family. In A. Barrett and G. Salvesen (editors), Proteinase inhibitors, p. 515-569. Elsevier Sci. Publ. BV, North Holland. Boyd, N S , N D Wilson, A R Jerrett, and B I Hall 1984. Effects of brain destruction on post harvest muscle metabolism in the fish kahawai (Arripis trutta ). J. Food Sci. 49:177-179. Carey, F. G 1973. Fishes with warm bodies. Sci. Am. 228:36-44. Carey, F G , J M Teal, J W Kanwisher, and K D Lawson. 1971. Warm-bodied fish. Am. Zool. 11:137-145. Cramer, J L, R. M Nakamura, A E Dizon, and W N. Ikehara. 1981. Burnt tuna: Conditions leading to rapid deteriora- tion in the quality of raw tuna. Mar. Fish. Rev. 43(6):12-16. Dahlmann, B , L Kuehn, and H. Reinauer 1984. Proteolytic enzymes and enhanced muscle protein breakdown. In W, H. Horl and A. Heidland (editors), Proteases, p. 505-517. Plenum Press, N.Y. Davie, P S , and R. I. Sparksman. 1986. Burnt tuna: An ultrastructural study of post- mortem changes in muscle of yellowiin tuna iThunnus albacares) caught by rod and reel and southern bluefin tuna (Thunnus maccoyii) caught on handline or long- line. J. Food Sci. 51:1122-1128, 1168. FUJIMAKI, M , AND K KOJO 1953. Handling effect upon biochemical change in the fish muscle immediately after catch — II. Changes of acid- soluble phosphorus compounds of frigate mackerel mus- cle. Bull. Jpn. Soc. Sci. Fish. 19:499-504. Gibson, D J M 1981. A handbook on processing southern bluefin tuna for the fresh chilled sashimi market in Japan. Minist. Agric. Fish., Wellington, var. pag. Hagberg, H 1985, Intracellular pH during ischemia in skeletal mus- cle: Relationship to membrane {wtential, extracellular pH, tissue lactic acid and ATP. Pfluegers Arch. 404:342-347. Hochachka, P. W., and R W Brill In press. Autocatalytic pathways to cell death: A new analysis of the tuna bum problem. Fish Physiol. Biochem. 3. 371 FISHERY BULLETIN: VOL. 86, NO. 2 HOCHACHKA, P W . W C HULBERT, AND M. GUPPY. 1978. The tuna power plant and furnace. In G. D. Sharp and A. E. Dizon (editors), The physiological ecology of tunas, p. 153-174. Acad. Press, N.Y. HOCHACHKA, P. W., AND T. P MOMMSEN. 1983. Protons and anaerobiosis. Science (Wash. D.C.) 219:1391-1397. HUDGINS. L L.. AND S. G. POOLEY. 1987. Growth and contraction of domestic fisheries: Hawaii's tuna industry in the 1980s. In D. J. Doulman (editor). Tuna issues and perspectives in the Pacific is- lands region, p. 225-241. East-West Center Press, Hon- olulu. ITOKAWA, S 1968. On maintaining the freshness of tuna seine catches. Reports of the Mie Prefecture Owase Fishery Experiment Station, p. 1-15. 1969. On experiments in prevention of burnt flesh in yel- lowfin tuna. Rep. Mie Pref Owase Fish. Exp. Stn. p. 1-9. Jerrett. a R. 1984. Detection and prevention of burning in southern bluefin tuna. Part A: Research results and discus- sion. Fish Processing Bull. 4, 22 p. + app. June, F C 1960. Spawning of yellowfin tuna in Hawaiian waters. Fish. Bull., U.S. 54:47-64. KONAGAYA, S., AND T KONAGAYA 1978. Denaturation at moderate temperatures of myo- fibrillar protein of red-meat fish: A possible cause of yake-niku. Bull. Tokai Fish, Res. Lab. 96:67-74. 1979. Acid denaturation of myofibrillar protein as the main cause of formation of "yake-niku", a spontaneously done meat, in red meat fish. Bull. Jpn. Soc. Sci. Fish. 45:245. KOOHMARAIE. M . J. E SCHOLLMEYER, AND T R DUTSON. 1986. Effect of low-calcium-requiring calcium activated factor on myofibrils under varying pH and temperature conditions. J. Food Sci. 51:28-32, 65. Ling, N., and R. M. G Wells 1985. Plasma catecholamines and erythrocyte swelling following capture stress in a marine teleost fish. Comp. Biochem. Physiol. 820:231-234. NaKAMURA, K , Y FUJII. AND S ISHIKAWA 1977. Experiments on the prevention of "burning" of tunas — L An examination of causes of occurrence. Bull. Tokai Reg. Fish. Res. Lab. 90:39-43. (Engl, transl. by W. G. Van Campen, 1980, 8 p., Transl. No. 46; available Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396.) NaKANO. T . AND N TOMLINSON 1967. Catecholamine and carbohydrate concentrations in rainbow trout iSalmo gairdneri ) in relation to physical disturbance. J. Fish. Res. Board Can. 24:1701-1715. Nekvasil, N P , AND K. R. Olson 1986. Extraction and metabolism of circulating cate- cholamines by the trout gill. Am. J. Physiol. 250:R526- R531. Parkes, C. 1986. Calpastatins. In G. Barrett and A. Salvesen (edi- tors). Proteinase inhibitors, p. 571-587. Elsevier Sci. Publ. BV, North Holland. Sakamoto, S . and N Seki 1985. Limited hydrolysis of carp myosin by cal- pains. Bull. Jpn. Soc. Sci. Fish. 51:1551-1557. Seki, N , and Y Kimura 1986. Degradation of troponin and tropomyosin by cal- pain. Bull. Jpn. Soc. Sci. Fish. 52:1673-1680. Strong. R D 1979. 1978-79 PTDF funded feasibility study to establish an ika-shibi fishery on Guam — Final Report. Pac. Tuna Dev. Found., May 1979, var. pag. SuGiTA. H., S Ishiura. K Kamakura, H. Nakase, K. Hagiwara. I Nonaka, and K Tomomatsu. 1984. Ca-activated neutral protease in physiological and pathological conditions. In S. Ebashi, M. Endo, K. Ima- hori, S. Kakiuchi, and Y. Nishizuka (editors). Calcium regulation in biological systems, p. 243-256. Acad. Press, Tokyo. [Suisan Sekai] 1977. How to effectively kill tunas in order to maintain quality and higher prices. Suisan Sekai 9:52- 57. (Engl, transl. by T. Otsu, 1978, 4 p., Transl. No. 25; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822- 2396). Suzuki. K., S. Kawashima, and K. Imahori. 1984. Structure and function of Ca2+-activated protease. In S. Ebashi, M. Endo, K. Imahori, S. Kaki- uchi, and Y. Nishizuka (editors). Calcium regulation in biological systems, p. 213-226. Acad. Press, Tokyo. Toyo-OKA, T 1982. Phosphorylation with cyclic adenosine 3:5' monophosphate-dependent protein kinase renders bovine cardiac troponin sensitive to the degradation by calcium- activated neutral protease. Biochem. Biophys. Res. Commun. 107:44-50. Williams, S C 1986. Marketing tuna in Japan. Queensl. Fish. Ind. Train. Comm., Brisbane, Queenl., Aust., 60 p. Yuen, H S H 1979. A night handline fishery for tunas in Hawaii. Mar. Fish. Rev. 41(8):7-14. Zeece, M G , K. Katoh, R M Robson, and F. C Parrish, Jr. 1986a. Effect of cathepsin D on bovine myofibrils under different conditions of pH and temperature. J. Food Sci. 51:769-772, 780. Zeece, M G , R M Robson. M L Lusby, and F C Parrish, Jr. 1986b. Effect of calcium activated protease (CAF) on bovine myofibrils under different conditions of pH and temperature. J. Food Sci. 51:797-803. 372 REPRODUCTIVE BIOLOGY OF THE SPOTTED SEATROUT, CYNOSCION NEBULOSUS , IN SOUTH TEXAS^ Nancy Brown-Peterson,^ Peter Thomas,^ and Connie R. Arnold' ABSTRACT The spotted seatrout, Cynoscion nebulosus, spawns from April through the end of September in shallow bays in South Texas. All females were sexually mature at 300 mm SL and males at 200 mm SL. Histological examination of the testes showed that spermatogenesis began in January and contin- ued until September, although spermatozoa remained in the testes until November. Spermatogenesis was more active in the peripheral lobules of the testes than in the central lobules during the latter half of the spawning season. The sequence and timing of final oocyte maturation (FOM) was investi- gated for the first time in this species. Lipid coalescence began at dawn and germinal vesicle migra- tion started by midmoming. By early afternoon, oocj^tes were hydrated and spawning occurred at dusk. Evidence of multiple spawning was examined. Morphological and histological data showed that oocytes were continually recruited from March through the end of September, and the percentages of vitellogenic and fully yolked oocytes did not decline during the spawning season. An average of only 15.5% of the vitellogenic oocytes underwent FOM and hydration during a single spawn. Postovulatory follicles were found in many fish with mature ovaries throughout the reproductive season. Laboratory studies showed that this species is capable of repeated spawns. Batch fecundity was best predicted by ovary-free body weight of the fish and averaged 451 ±43 eggs/g ovary-free body weight. Estimates of spawning frequency ranged from every other day to once every three weeks. The spotted seatrout, Cynoscion nebulosus , sup- ports important commercial and recreational fisheries throughout its range (Chesapeake Bay, Virginia to Tampico, Mexico; Tabb 1966). Whereas aspects of the reproductive biology of this sciaenid species have been documented throughout its range (i.e., Chesapeake Bay: Brown 1981; Georgia: Mahood 1975; Florida: Moody 1950, Klima and Tabb 1959, Tabb 1961; Mississippi: Overstreet 1983; Louisiana: Hein and Shepard 1979; Texas: Pearson 1929, Miles 1950, 1951), a comprehensive study of reproduc- tion in spotted seatrout does not exist for any area. An extensive knowledge of the reproductive life history of a species is necessary to understand certain aspects of its reproductive physiology and endocrinology. Most previous studies have con- centrated on the size at sexual maturity and the extended spawning season of this species (Pear- son 1929; Moody 1950; Tabb 1961; Mahood 1975). A protracted spawning season is generally char- iContribution No. 696, University of Texas at Austin, Marine Science Institute, Port Aransas, TX 78373. 2Marine Science Institute, University of Texas at Austin, Port Aransas, TX; present address; Florida Department of Nat- ural Resources, Indian River Lagoon Aquatic Preserves, 4842 S. U.S. Highway 1, Fort Pierce, FL 34982. 3Marine Science Institute, University of Texas at Austin, Port Aransas, TX 78373. Manuscript accepted February 1988. FISHERY BULLETIN: VOL. 86, NO. 2, 1988. acteristic of multiple spawners (Nikolskii 1969). In addition, Overstreet's (1983) histological data suggest that C. nebulosus may be a multiple spawner. However, the possibility that spotted seatrout are multiple spawners has not been thor- oughly discussed in the literature, and previous estimates of fecundity (Sundararaj and Suttkus 1962; Overstreet 1983) have not considered the multiple spawning nature of C. nebulosus or have included estimates of batch fecundity. In the present study, the reproductive biology of C. nebulosus in South Texas was investigated, and particular attention was given to evidence for multiple spawning and estimates of batch fecun- dity and spawning frequency. The spawning sea- son, spawning sites, time of spawning, and per- centage of running ripe fish were documented. In addition, the temporal pattern of final oocyte maturation was determined. The histological ap- pearance of the gonads was examined, and partic- ular attention was given to the presence of post- ovulatory follicles in ovarian tissue, an indicator of multiple spawning in the northern anchovy, Engraulis mordax (Hunter and Goldberg 1980). The size-frequency distribution of vitellogenic oocytes was examined, since this can indicate multiple spawning (deVlaming 1983). Batch fe- cundity was determined and spawning frequency 373 FISHERY BULLETIN: VOL. 86, NO. 2 was estimated from both field-caught and labora- tory fish. MATERIALS AND METHODS Collection of Samples Spotted seatrout were collected in Redfish Bay or Lydia Ann Channel near Port Aransas, TX, U.SA. at depths of 1-3 m (Fig. 1). Fish were cap- tured using a 300 m gill net (82 mm stretch) or a 300 m trammel net (outer panels, 178 mm stretch; inner panel, 89 mm stretch) over shallow beds of turtle grass, Thalassia testudinum or shoal grass, Halodule wrightii, bordered by a 1-2 m drop-off into a channel. Fish <300 mm SL were captured with hook and line in 1-2 m of water. Samples were collected weekly or twice monthly from March 1982 through early May 1985. No samples were taken in November and December 1983, October and December 1984, and February 1985. During 1982, samples were col- lected at dusk only; in 1983, 1984, and 1985, sam- ples from dawn, midday, and midnight were also taken. During each sampling period, salinity and temperature were recorded as well as the time of capture. Analysis of Fish and Gonads Total length (TL) and standard length (SL) were measured to the nearest mm for each speci- men and total body weight (WT) was determined to the nearest 10 g. Gonads were removed and weighed to the nearest 0.1 g (gonad weight, GW) and the gonadosomatic index (GSI) was calcu- lated, using the formula: GSI = (GWAVT) x 100. Reproductive stage of the gonads was assessed macroscopically using the criteria in Table 1. The macroscopic criteria used were similar to those used by Overstreet (1983) and Macer (1974). A small portion of tissue was removed from the anterior or midsection of one gonad from each fish and preserved in Davidson's fixative for histolog- ical analysis (Jones 1966). Tissues were dehy- drated and embedded in paraffin. Seven micron sections were cut and stained with Harris' hema- toxylin and counterstained with eosin. Reproduc- tive stage of each sample was assessed microscop- ically using the criteria in Table 1. Fish with ^Area enlarged Gulf of Mexico Figure 1. — Location of spotted seatrout sampling sites in South Texas. Asterisks denote sampling sites. 374 BROWN-PETERSON ET AL.: REPRODUCTIVE BIOLOGY OF THE SPOTTED SEATROUT mature gonads were considered sexually mature. The microscopic criteria used were similar to those described by Yamamoto (1956), Hyder (1969), and Macer (1974). The percentages of each stage of oocyte, atretic structures, and postovula- tory follicles were determined from histological slides of the ovaries of females >305 mm SL by lected fields of view. An oocyte was counted if >50% of the cell was in the field of view. Two hundred and fifty to 500 oocytes were counted for each histological specimen. A small portion of tissue (<0.5 g) was removed from the center of each ovary of females >305 mm SL collected in 1984 and 1985, placed in clearing slides of the ovaries of females >305 mm SL by SL collected in 1984 and 1985, placed in clearing counting all the oocytes in three randomly se- solution (6 parts ethanol, 3 parts formalin, 1 part Table 1 . — Criteria used to describe gonadal reproductive stages of male and female spotted seatrout collected in South Texas. Stage Sex Macroscopic appearance Microscopic appearance Immature Regressed M Early developing F M Developing F Mature Ripe Running ripe or spawning Partially spent Spent F Ovary small, thin, light pink, slight vasculariza- tion. GSI range: 0.39-0.55. M Testes small, thin, light grey, appearance similar to mesentery. F Ovary small, light pink, vascularization more obvious than in immature fish. GSI range: 0.39-1.16. Testes small, thin, white, slightly larger than imma- ture fish. GSI range: 0.04-0.20. Ovary similar to regressed fish but slightly larger. GSI range: 0.80-1.26. Testes similar to regressed fish but slightly larger. GSI range: 0.16-0.48. Ovary visably enlarged, light yellow, highly vascular, approximately 60% of the length of the body cav- ity. GSI range: 0.95-2.6. M Testes thickened, creamy white, no free milt ex- pelled when cut. GSI range: 0.21-0.68. F Ovary large, brilliant yellow orange, oocytes visible to naked eye, vascularization prominent. Ovaries 80-90% length of body cavity. GSI range: 2.5- 9.6. M Testes creamy white, thicker, more firm and elon- gated than developing testes. A small amount of milt expelled when cut. GSI range: 0.54-1.42. F Oocytes hydrated, ovary looks clear, takes up al- most entire body cavity, highly vascular- ized. GSI range: 12.5-19.9. M Testes white, swollen, milt does not flow with light pressure, but flows freely when cut. GSI range: 1.07-1.62. F Hydrated oocytes expelled with little to no pressure, ovary fluid, fills almost entire body cavity. GSI range: 7.7-17.6. M Testes creamy white, milt freely flowing with slight pressure. GSI range: 1.41-3.84. F Ovary looks similar to mature condition but more flacid; occupies smaller percentage of body cav- ity. GSI range: 2.0-5.3. M Tests looks the same as running ripe but smaller. GSI range: 0.85-1.77. F Ovary flacid but still highly vasculahzed, no longer than 50% of body cavity, pinkish in color. GSI range: 1.33-2.63. M Testes flacid, width reduced, white. GSI range: 0.13-0.80. Only primary oocytes present; no atretic oocytes. Lamellar margin thin. Only primary spermatogonia present. Primary chromatin nucleolar and early perinuceolar stage oocytes. Some atretic oocytes. Lamellar margin thicker than immature, more convoluted. Primary and secondary spermatogonia present. Appearance of late perinucleolar oocytes and ph- mary cortical alveoli stage oocytes. No atretic oocytes present. Many secondary spermatogonia and primary sper- matocytes. Oocytes in secondary cortical alveolar stage and yolk granule stage. Many oocytes still in primary cortical alveolar stage. Early yolk globular stage present. All stages of spermatogenesis present, with few pri- mary spermatogonia and free spermatozoa. Pri- mary and secondary spermatocytes predominate. Oocytes in yolk globular stage most common. Yolk and oil globules begin to encroach on nucleus. Largest oocytes range from 300 to 375 y.m. No primary spermatogonia present, some second- ary spermatogonia. Secondary spermatocytes and spermatids most numerous. Spermatozoa in central lobules. Many hydrated oocytes, irregularly shaped, eosinophilic. Other oocytes in yolk granular and yolk globular stages. Spermatozoa, spermatids and secondary spermato- cytes predominate. Few to no spermatogonia. Central lobules filled with spermatozoa. Same as ripe ovaries but fewer hydrated oocytes. Atretic and postovulatory follicles may be present. Same as ripe fish, but spermatozoa evident in sperm ducts. Similar to mature ovary, except atretic oocytes in several stages of degeneration always evident. Postovulatory follicles occasionally present. No spermatogonia, few primary spermatoocytes. Most active spermatogenesis in peripheral lob- ules. Spermatozoa partially filling lobules and abundant in sperm ducts. Massive atresia of all remaining vitellogenic oocytes. Many primary and chromatin nucleolar oocytes present. Lamellar membrane highly convoluted. Spermatozoa present in some lobules. Most lobules small, with only primary and secondary sper- matogonia. 375 FISHERY BULLETIN: VOL, 86, NO. 2 glacial acetic acid), and vigorously shaken for 30 seconds. Within a few minutes the cytoplasm cleared and the germinal vesicle could be easily observed microscopically. Ovarian fragments were taken from females collected at the spawn- ing site over a 24-h period, placed in clearing solu- tion and then examined under low-power magni- fication to determine the stage of final oocyte maturation. Oocyte Size-Frequency Distributions and Estimates of Batch Fecundity To determine fecundity and the frequency dis- tribution of oocyte diameters, a 2-15 g piece of tissue was removed from the midsection of the ovaries of 57 fish and weighed to the nearest 0.01 g. The tissues were placed in a modified Gilson's solution (Bagenal 1966) for 3-12 months and periodically shaken to separate the oocytes from connective tissues. Ovaries containing hy- drated oocytes were examined after three months since hydrated oocytes of spotted seatrout began to disintegrate when left in Gilson's solution for a longer period of time. The volumetric method was used to estimate fecundity (Bagenal and Braum 1971). The oocyte samples were suspended in 500-1,500 mL of water and three replicate 0.5 or 1 mL subsamples were taken. All the oocytes >30 ^JLm were counted, and those >80 |jim in diameter (the growing oocytes) in each sample were measured to the nearest 15 ^m using an ocular micrometer. A total of 556-1,110 growing oocytes were mea- sured in each sample. The number of resting oocytes (oocyte diameter 30-80 ^JLm) was deter- mined by diluting the original oocyte suspension 1:10, and counting three replicate subsamples. Altogether, the frequency distributions of oocytes from 48 fish were analyzed (3 in developing stage; 9 in mature, spawning not imminent stage; 14 in mature, just prior to spawning stage; and 22 in running ripe stage). Fecundity was calculated fol- lowing Macer's (1974) formula and expressed as relative fecundity of number of eggs per gram ovary-free body weight. Batch fecundity (BF) is defined as all oocytes >350 ixm which were undergoing final oocyte maturation that formed a distinct batch, and all hydrated oocytes. This defi- nition of batch fecundity is in agreement with Hunter and Macewicz's (1985) statement that oocytes undergoing final oocyte maturation may be included as hydrated oocytes when hydration occurs very rapidly. Spawning of Fish in the Laboratory Four female and two male spotted seatrout were maintained in a 30,000 L recirculating sys- tem. The tank, filtration system and feeding regime of the fish has been described previously (Arnold et al. 1976). The salinity ranged from 25 to 30%(. Spawning was induced by increasing the temperature and photoperiod from wintertime settings of 13°C, 9L:15D to 26°C and 15L:9D (Arnold et al. 1976). The filter boxes were checked daily for the presence of buoyant, newly fertilized eggs. Statistical Analysis Simple linear regression, oneway analysis of variance, and analysis of covariance were com- puted for the data using SPSS packaged programs (SPSS 1981). RESULTS Size at Maturity Some female spotted seatrout were sexually mature after they reached 231 mm SL and >90% of the females had reached sexual maturity at 271 mm SL (Table 2). By 300 mm SL, all female spot- ted seatrout were sexually mature. Fish 300 mm SL or larger made up 85.4%, and immature fish comprised 6.5%, of all the females sampled. Male spotted seatrout reached sexual maturity at a much smaller size than females. The size at Table 2. — Number and percentage of mature female and male spotted seatrout by 10 mm size categories collected in South) Texas, April 1982-N^ay 1985. Maturity was judged by histological and macroscopic inspection. Standard Female Male length (mm) N % mature N % mature 201-210 5 0 5 100 211-220 4 0 4 100 221-230 4 0 1 100 231-240 4 50 3 100 241-250 6 83 3 67 251-260 6 100 5 100 261-270 5 80 7 100 271-280 11 91 10 100 281-290 14 100 11 100 291-300 24 96 20 100 301-310 60 100 40 100 311-320 94 100 68 100 321-330 115 100 80 100 >330 945 100 507 100 Total 1,297 764 376 BROWN-PETERSON ET AL.: REPRODUCTIVE BIOLOGY OF THE SPOTTED SEATROUT which most male spotted seatrout attain sexual maturity could not be determined, since the smallest fish collected by the sampling methods (200 mm SL) were all sexually mature (Table 2). Season and Time of Spawning Histological and macroscopic examination of the gonads (see Table 3 and Figure 3) and mean GSI values (Fig. 2) show that spotted seatrout have an extended reproductive season in South Texas. Mean GSI values of males and females increased by April 1982-85 and remained ele- vated through the end of September. A 5°C in- crease in water temperature at the sampling sites to 23°C during the first week of April 1982-85 was paralleled by an increase in GSI to 2.1 in males and 4.5 or greater in females. The pattern of seasonal changes in mean GSI values of males was relatively consistent during the three-and-a- half years of sampling. Mean GSI values began to increase in mid-February, reached a maximum of 1.9-2.4 by April, and slowly declined during the spawning season until they dropped rapidly to regressed levels of 0.2 by the first half of October (Fig. 2). The seasonal patterns in mean GSI val- ues of females were also similar from 1982 to 1985. In all four years, mean GSI increased in April and subsequently declined in May. In 1982- 84, mean GSI increased again later in the season prior to the final decrease to regressed levels in October (Fig. 2). Thus, mean GSI values appear to be bimodal, with one period of peak spawning activity in April, and the second period of peak spawning activity varying between August 1982 and July 1983 and 1984. Male and female spotted seatrout in spawning condition (males with freely flowing milt, females with ovulated oocytes) were consistently captured during a 2-h period around dusk over shallow (1 m) beds of Thalassia testudinum or Halodule wrightii bordered by a channel 2 m deep. The salinity at the spawning sites ranged from 20 to 37%c. Although actual spawning was not ob- served, collection of newly fertilized eggs from the spawning area at dusk confirmed that spawning was taking place (Holt and Holf*). Spawning fish were not captured over beds of scattered H. wrightii or T. testudinum that were not immedi- ately adjacent to a channel. Females with freely ^S. A. Holt and G. J. Holt, University of Texas at Austin, Marine Science Institute, Port Aransas, TX 78373, pers. com- mun. 1983. t 33 1-25 17 1-9 I I I I I I I I I I I I |'i''rV;'i^|"r|T| I M I M ' M 1 M ' I M I I ly'V' I I I I M M I I I Ml I I I I I I I I I ' i ' i ' i °C MM J SNJMMJ 1982 1983 N J M M' J S N 1984 J M M 1985 Month Figure 2. — Bi-monthly mean water temperature and mean gonadosomatic index (GSI) of male and female spotted seatrout collected in South Texas from March 1982 to May 1985, including ± 1 SE of the mean. Sample size for each data point: 3-61 for females (sample size <15 in only 9 cases) and 2-52 for males (sample size <9 in only 9 cases). 377 FISHERY BULLETIN: VOL. 86, NO. 2 flowing oocytes were only captured at dusk and milt appeared to flow more freely in males at dusk than at other times of the day. Gonadal Development in Males Testicular recrudescence began in January with the appearance of primary spermatocytes (Fig. 3), and by February the majority of the males were in the early developing or developing reproductive stage. In March, 89% of the males had testes containing free spermatozoa (Fig. 3), although only 27% were running ripe. Over 94% of all males captured from April until the end of August were running ripe. Spent males first ap- peared in August and all the males during Octo- ber and November were either in the spent or regressed condition. In December, all males cap- tured had regressed testes containing only pri- mary and secondary spermatogonia (Fig. 3). Histological observations revealed that the testes are the common unrestricted spermatogo- nia type, as described by Grier (1981). Spermato- genesis occurred throughout the entire testis dur- ing the majority of the year. However, from late February through early August, spermatogenesis was more advanced in the central lobules then in the peripheral lobules. By August there was no spermatogenic activity in the central lobules, al- though spermatogenesis continued in the periph- eral lobules until mid-September. The only period of the year when there was no active spermato- genesis was from October through late January, although primary and secondary spermatogonia were present from October through the end of April (Fig. 3). Primary spermatocytes were com- mon from late January until March, although they did not disappear from the testes entirely until the end of July. Secondary spermatocytes first appeared in mid-February, were common through the end of July and did not disappear until early September (Fig. 3). Spermatids and spermatozoa were present in the central lobules by late February, and from March through mid- August the central lobules were swollen with spermatozoa (Fig. 3). In late August and Septem- ber many of the central lobules appeared to be partially empty of spermatozoa, although some spermatozoa were still present up to mid-Novem- ber (Fig. 3). The peripheral lobules contained ac- tive spermatogenic cysts of primary and second- ary spermatocytes during June and July and the lumens of these lobules began to fill with sperma- tozoa. The peripheral lobules were swollen with spermatozoa during August and September, but by mid-October no spermatozoa remained. Gonadal Development in Females Gonadal recrudescence was observed in a small percentage of the females captured in January and by March 94% of the females were undergo- ing ovarian development (Table 3). Fish with ma- ture ovaries were found from March until the end of September, while running ripe fish were cap- tured fi-om April through the end of September (Table 3). Partially spent fish (females that ap- peared to have spawned at least once but still contained vitellogenic oocytes) were captured from May through the end of September. No com- 1° Spermatogonia 2° Spermatogonia 1° Spermatocytes 2° Spermatocytes Spermatids Spermatozoa A -| r J J Month M M A S — r N D Figure 3. — Seasonal cycle of spermatogenesis in spotted seatrout collected in South Texas, as determined by histological and macroscopic observations. Data from April 1982 through mid-May 1985 are combined. 1° - primary, 2° - secondary. 378 BROWN-PETERSON ET AL : REPRODUCTIVE BIOLOGY OF THE SPOTTED SEATROUT Table 3. — Percentage of female spotted seatrout In seven repro- ductive stages by month, as assessed by histological and macro- scopic examination of the ovaries. Data from April 1982 to May 1985 are combined. REG ^ Regressed, E DEV = Early Develop- ing, DEV = Developing, MAT = Mature, RR = Ripe and Running Ripe, P SP = Partially Spent, SP = Spent. N Percent in each reproductive stage Month REG EDEV DEV MAT RR PSP SP January 70 98 2 February 51 55 43 2 March 124 4 56 31 9 April 372 4 6 78 12 May 220 4 84 5 7 June 104 1 56 19 24 July 114 1 48 30 20 1 August 100 58 32 7 3 September 37 14 37 30 11 8 October 8 87 13 November 27 100 December 51 98 2 Total 1,278 tions: 150-190 (xm) varied from a lov^^ of 7.3% in June to a late season high of 16.1%, and averaged 13% of the total number of oocytes in the ovary. Yolk globular oocytes, the largest oocytes present (diameter: 200-375 p.m), were most common. The percentage of oocytes in the yolk globular stage ranged from a high of 30.8% in April to an end-of- season low of 22.6% in September and averaged 26.9% during the reproductive season. In addition to the actively growing oocytes in the ovary, a small percentage of atretic oocytes was always found in fish with mature ovaries. Atretic oocytes are defined here as vitellogenic oocytes undergo- ing alpha stage atresia. The percentage of atretic oocytes ranged from a low of 2.2% in March to a midsummer high of 8% in July and averaged 5.4%. Additionally, postovulatory follicles (POF) were observed in the ovaries offish captured from pletely spent fish were captured before July and few were captured during the remainder of the reproductive season. Histological observations offish with regressed ovaries collected from late September to mid- February showed only primary chromatin nucle- olar and early perinucleolar oocytes. Atretic oocytes were present from September until mid- December; no atretic oocytes were observed from mid-December through the end of February. The appearance of late perinucleolar and primary cor- tical alveoli stage oocytes in late January, Febru- ary, and early March represented the initial stages of ovarian recrudescence. Ovarian devel- opment was proceeding rapidly by early March and oocytes in the secondary cortical alveoli and yolk granule stages were common. During the reproductive season, histological observations of the ovary showed a heterogeneous morphology. Oocytes in all stages of growth (from resting to the fully grown yolk globular stage) were distributed throughout the ovaries of all ma- ture, running ripe, and partially spent fish col- lected from March through the end of September. Vitellogenesis was probably continuous from February until the end of September, as shown by the continual presence of large numbers of oocytes in the cortical alveoli and yolk granule stages. The relative percentages of three types of post cortical alveolar stage oocytes in fish in the mature reproductive stage did not change markedly from March through the end of Septem- ber (Fig. 4). The percentage of oocytes in the yolk granule stage (diameter in histological prepara- Q o < 1- z m o DC LU Q. LU > I- < _l LU IT O z LU =) o LU DC P771 Yolk granule frr^ Yolk globule ^g Atretic N I 5 I 58 I 21 I 17 I 27 I 29 I 9 I 36-1 32 28 24H 20 16 12- 8- 4- M M J J MONTH Figure 4. — Frequency of three different types of oocytes in fe- male spotted seatrout ovaries in the mature reproductive stage (MAT, Table 3) as assessed by histological observation. Fish were collected from South Texas during the reproductive sea- son. Frequency refers to the percentage of each type of oocyte relative to the total number of all types of oocytes counted. Data from April 1982 through the end of September 1983 are com- bined and the number (A^) of individuals examined each month is indicated. Bars indicate mean percentage of oocytes ±1 SE of the mean. Yolk granule oocyte diameters ranged from 150 to 190 p-m, and yolk globule oocyte diameters ranged from 200 to 375 fim in histological preparations. Atretic oocytes are defined here as oocytes in alpha stage atresia. 379 FISHERY BULLETIN: VOL. 86, NO. 2 April through September, although the percent- age of POF relative to the total number of oocytes was always small (<5%). The histological appearance of the zona radiata and follicle layers changed as the oocytes devel- oped. The zona radiata (chorion) was thin and nonstriated in perinucleolar and cortical alveoli oocytes. The zona radiata started to take on its characteristic striated appearance when oocytes reached the yolk granular stage and became no- ticeably thicker and more striated as oocytes grew into the yolk globular stage. In contrast, the granulosa and thecal layers appeared to decrease in thickness as oocyte development proceeded. Oocytes in the perinucleolar and cortical alveoli stages had thick, well-developed granulosa and thecal layers. In yolk granular and yolk globular oocytes, both follicle layers were noticeably thin- ner and the thecal layer was not always continu- ous around the oocyte. Frequency distributions of growing oocytes (di- ameters in Gilson's solution >80 fxm; perinucle- olar to hydrated stage oocytes) from ovaries of 48 fish were analyzed, and Figure 5 shows typical patterns in fish from four reproductive stages. In a fish captured in March with developing ovaries, growing and vitellogenic oocytes (oocytes during the phase of active vitellogenin uptake) ranging from 80 to 185 jxm in diameter were present (Fig. 5A; 625 oocytes counted). Fish in the mature re- productive stage in which spawning did not ap- pear to be imminent, as indicated by GSI values <4, had growing and vitellogenic oocyte diame- ters ranging from 100 to 320 M-m (Fig. 5B; 556 oocytes counted), comprising 21% of the total number of oocytes >30 fxm in the ovary. Fish just prior to spawning, as indicated by GSI values >7, had a distinct batch of oocytes with diameters Figure 5. — Frequency distributions of growing oocyte diame- ters O80 p.m) in female spotted seatrout collected from South Texas. Growing oocytes comprised 21% of the total number of oocytes >30 ^.m in the ovary. Oocyte diameter refers to the diameters of oocytes after preservation in Gilson's solution. Fre- quency refers to the percentage of each size of oocyte relative to the total number of oocytes in the subsample. Each graph repre- sents data from a single fish at a different reproductive stage. The arrow at 350 jim indicates the minimum size necessary for final maturation to occur. A. Developing stage in March, GSI = 1.6. B. Mature stage in which spawning is not immi- nent, GSI = 3.7. Fish in this stage were collected from April through the end of September. C. Mature stage just prior to spawning, GSI = 7.4. Fish in this stage were collected from April through the end of September. D. Running ripe stage, GSI = 17.6. Fish in this stage were collected from April through the end of September. >350 fxm that were undergoing final oocyte mat- uration (Fig. 5C; 585 oocytes counted). However, there were no other distinct modes of vitellogenic oocytes. In running ripe fish, the batch of large oocytes hydrated to a diameter of 520 ixm or greater (Fig. 5D; 1,110 oocytes counted). The re- sults (Figs. 5A-D) clearly demonstrate that C. nebulosus has a continuous distribution of grow- ing and vitellogenic oocytes. The oocyte frequency distribution of the vitellogenic oocytes <320 \i.vc\ in diameter remaining in running ripe fish ap- I- 2- lU o LU LU DC 2-J 80 200 320 440 560 680 OOCYTE DIAMETER (nm) 380 BROWN-PETERSON ET AL.: REPRODUCTIVE BIOLOGY OF THE SPOTTED SEATROUT peared similar to the oocyte distribution in non- spawning fish (Fig. 5B, D). The percentage (21%) of growing and vitellogenic oocytes in the ovary remained constant throughout the reproductive season in fish in the mature and running ripe stages. Final Oocyte Maturation Final oocyte maturation (FOM) was highly syn- chronized in spotted seatrout and occurred only in oocytes >400 ixm ^. Figure 6A shows a photomi- crograph of a histological section of a spotted seatrout ovary in the mature reproductive stage that was not undergoing final oocyte maturation. Many oocytes were in the yolk globular stage and appeared to be fully grown. The first readily ob- servable stage of FOM in "cleared" oocytes was lipid coalescence (Fig. 6B). The oil droplets in the oocytes began to coalesce around the germinal vesicle (nucleus) and subsequently formed one to three large oil droplets. This stage was not always observed in histological preparations since many of the oocytes were not sectioned through their centers. The yolk globules remained discrete dur- ing lipid coalescence. After the lipids had coa- lesced, the germinal vesicle (GV) began to mi- grate to the periphery of the oocyte (germinal vesicle migration, or GVM). GVM could be seen in both histological sections (Fig. 6C) and in "cleared" oocytes (Fig. 6D). The oil droplet occu- pies the center of the oocytes shown in Figure 6C, D. Histological observation of this stage (Fig. 6C) showed that the yolk globules were not coalesced, the oil droplets had coalesced to form one or two large droplets and the GV had begun to lose its integrity and often appeared semicircular. At the completion of GVM, the nuclear (germinal vesi- cle) membrane broke down (GVBD) and the nu- clear material intermingled with the cytoplasm of the oocyte. Hydration occurred shortly thereafter, followed by ovulation and spawning of the fully mature oocyte. Final oocyte maturation occurred within 10 hours in spotted seatrout in the natural environ- ment (Fig. 7). A total of 209 fish were collected over eight 24-h periods from April through Au- gust in 1984 and 1985. Forty-three percent of the fish collected between the hours of 0500 and 1500 were undergoing FOM. Lipid coalescence was first observed at dawn (0545), and GVM started at 0900. By 1430, all fish undergoing final matu- ration had hydrated oocytes and ovulation and spawning commenced at dusk (1830) and contin- ued until 2100. None of the fish collected from 2100 to 0500 were undergoing FOM. Batch Fecundity The significant positive relationship {P < 0.001) between BF and ovary-free body weight can be best described by the following equation: BF = 459WT - 56,066, r^ = 0.56 (Fig. 8), while curvilinear equations best described the relation- ship between BF and SL and TL. The coefiicients of determination in all cases were <0.56. A one-way analysis of variance showed that mean BF (number of eggs per gram ovary-free weight) did not vary significantly during the April through September spawning season. Mean relative batch fecundity was highest in Septem- ber, lowest in May, and varied little during April, June, and July (Table 4). A prominent batch of oocytes was present only in females that were in all stages of final oocyte maturation or were running ripe (Figs. 5C, D; 7). The average batch size calculated from 14 fish containing hydrated oocj^es and no postovulatory follicles was 451 ± 43 eggs/g ovary-free body weight. This number averaged 15.5 ± 2.5% of the number of growing and vitellogenic oocytes in the ovary. Table 4. — Monthly mean batch fecundity expressed as number eggs/g ovary-free body weight of spotted seatrout in South Texas. All means were not statistically dif- ferent. Month N Mean fecundity ±1 se April 19 477 ± 42 May 2 320 ± 72 June 3 435 ± 109 July 5 409 ± 76 August 3 361 ± 60 September 3 560 ± 79 5A 400 (xm live oocyte equals a 350 ^.m oocyte preserved in Gilson's solution. Both measurements represent oocytes begin- ning FOM. Spawning Frequency To estimate the spawning frequency of spotted seatrout in South Texas, the percentage of run- ning ripe females captured monthly from April through the end of September in 1982 through 1985 was examined. Only fish captured at dusk and >305 mm SL were included in this analysis. The percentage of spawning females ranged fi-om 381 FISHERY BULLETIN: VOL. 86, NO. 2 B «Q •#«, > «l ^IIf ^ Qy »: 382 BROWN-PETERSON ET AL.: REPRODUCTIVE BIOLOGY OF THE SPOTTED SEATROUT Figure 6. — Photomicrographs of oocytes from spotted seatrout in the mature reproductive stage collected from South Texas. A. Histological section of an ovary in the mature stage that is not undergoing final oocyte maturation (FOM). Oocyte development is continuous, with fully grown yolk globular oocytes co-occurring with oocytes in earlier stages of develop- ment (Magnification 120x). B. "Cleared" oocytes in the lipid coalescence stage, the first stage of FOM (magnification 40 x). C. Histological section of an oocyte undergoing germinal vesicle migration (GVM). (Magnification 160 x ). D. "Cleared" oocytes in the GVM stage (magnification 40 X). Key: GV = germinal vesicle or nucleus, L = Lipid droplets, YG = yolk globules, YGO = yolk globular oocyte, ZR = Zona radiata. Lipid Coalescence KV\\\\\\\N GVM 14L:10D Hydration K\\\\\\1 ion and Spawning kWWN ^ ^^^^^M 0000 04'00 III 0800 1200 Time I — r^ 1600 2000 Figure 7. — Time-course of final oocyte maturation in spotted seatrout collected from South Texas. Lipid coalescence is the initial stage in final oocyte maturation. Data obtained from 46 fish undergoing final oocyte maturation in April and May 1985. GVM = germinal vesicle migration. 14L: lOD = hours of light and dark. Figure 8.— Relation between batch fecundity (BF) and ovary-free body weight (WT) of spotted seatrout from April 1984 to May 1985. Thirty-three fish with oocytes >350 \i.m that were undergoing final oocyte maturation and formed a distinct batch of oocytes or had hydrated oocytes are included. Weight (g) 383 FISHERY BULLETIN: VOL. 86, NO. 2 a low of 8% in May to a high of 45% in September (Table 5). An average of 27.5% of the females captured during the spawning season were run- ning ripe, corresponding to an average spawning frequency of once every 3.6 days or 50 times dur- ing the 6-mo spawning period. The spawning frequency was also estimated from females that were undergoing final oocyte maturation (FOM) between the hours of 0600 and 1400 in April, May, July, and August 1984 and 1985 (Table 5). For comparison, the actual per- centage of running ripe females captured during the same months is also presented. An average of 42.8% of the females examined for final oocyte maturation from April through August were undergoing FOM. Therefore, average spawning frequency was once every 2.3 days, or 80 times during the 6-mo spawning season. Spawning frequency was also estimated from the percentage of females captured with ovaries containing postovulatory follicles (POF). An av- erage of 13.1% of the females captured had ovaries which contained POF ranging from 12 hours to 2 days old. This would correspond to a spawning frequency of once every 7.6 days or 24 times during the spawning season. Finally, the spawning frequency of four female C. nebulosus in the laboratory under conditions of controlled temperature and photoperiod were ex- amined (Table 6). The fish spawned from 1 to 10 times each month for 17 months, an average of 17 spawns per individual over a 12-mo period (Table 6). This would correspond to a spawning fre- quency of once every 21 days. DISCUSSION Sexual Maturity and Spawning Season Male spotted seatrout reached sexual maturity at a smaller size than females which appears to be a fairly common phenomenon in spotted seatrout throughout its range (see Mercer 1984 for a re- view). Both male and female spotted seatrout in South Texas reached sexual maturity at a size similar to that reported for other groups of C. nebulosus along the Gulf Coast (Moody 1950; Klima and Tabb 1959; Overstreet 1983) and at a smaller size than along the East Coast (Tabb 1961; Brown 1981). Running ripe females were captured only be- tween one hour before and two hours after sunset in South Texas, suggesting a high degree of syn- chrony in spawning fish. Collections of newly fer- Table 5. — Percentage of spotted seatrout spawning (RR) or under- going final oocyte maturation (FOf^) in South Texas. Data on spawning fish collected April 1982-f^ay 1985. Data on fish under- going FOM collected April 1984-IVIay 1985. Spawning fish Fish undergoing FOM 1982-85 1984-85 Month N # RR % RR N # FOM % FOM % RR April 256 42 16 33 14 42 37 May 140 11 8 21 10 48 13 June 101 20 20 — — — — July 82 33 40 7 3 43 52 August 101 36 36 8 3 38 — September 20 9 45 — — — — Mean 27.5 42.8 25.5 Table 6. — Number of spawns and average water temperature by month for four female spotted seatrout contained in a 30,000 L tank in Porl Aransas, TX, under conditions of con- trolled temperature and photoperiod, July 1974-November 1975. Expenmental proce- dures descnbed in Arnold et al. (1976). Month Number of spawns °C July 8 26.75 August 7 26.75 September 4 26.0 October 6 25.5 November 4 24.5 December 8 24.0 January 9 23.5 February 10 24.0 March 10 24.0 April 3 25.75 May 1 25.0 June 0 22.0 July 6 25.75 August 2 25.0 September 2 23.0 October 9 23.5 November 10 24.0 Total 99 Average: 1 7 spawns/ 1 2 months/ female tilized eggs during a 25-h period after sunset (Holt et al. 1985) provides supporting evidence for this spawning synchrony. Both histological data and GSI values showed spotted seatrout have an extended spawning sea- son in South Texas. Gronadal recrudescence began in January in male and in February in female spotted seatrout. Spawning commenced in April and continued until the end of September. Other studies of C. nebulosus along the Gulf and East coasts of the United States have also reported long spawning seasons (Mercer 1984). Addition- ally, other sciaenids have extended spawning sea- sons (Merriner 1976; White and Chittenden 1977; DeVries and Chittenden 1982; Love et al. 1984), 384 BROWN-PETERSON ET AL.: REPRODUCTIVE BIOLOGY OF THE SPOTTED SEATROUT suggesting a prolonged spawning season is a com- mon reproductive strategy among sciaenids liv- ing in temperate and subtropical waters. The duration of the spawning season may be related to water temperature. Perhaps this appar- ent association with temperature is related to the viability and development of spawned spotted seatrout eggs. Twenty-three degrees may be the minimum temperature necessary for successful spawning, as indicated by both the failure to cap- ture running ripe fish at lower temperatures and by data from spotted seatrout induced to spawn under laboratory conditions (Arnold et al. 1976; Table 6). However, since spawning ceased in Sep- tember when water temperatures were well above 23°C, possibly a decrease in photoperiod in combination with a decrease in temperature pro- vides the necessary cue for termination of spawn- ing. Hein and Shepard (1979) suggested photope- riod may be an important regulating factor in C. nebulosus spawning. Data from laboratory- spawned spotted seatrout (Arnold et al. 1976) also support this speculation. The seasonal pattern of mean GSI values dur- ing the spawning season was relatively consistent for both males and females over the Sj-yr period, 1982-85 (Fig. 2). The bimodality of the female GSI data suggests the possibility of two peaks in spawning activity, although the timing of the sec- ond peak varies from year to year, thus demon- strating the need for several consecutive years of data. Bimodal spawning peaks have been previ- ously reported for the species by Hein and Shep- ard (1979) in Louisiana, Stewart (1961) in Florida, and Brown (1981) in Chesapeake Bay, VA. However, mean GSI values should be used with caution when attempting to predict actual peaks in spawning activity (deVlaming et al. 1982). Fecundity Accurate annual fecundity measurements are difficult to determine for multiple spawning fishes with an extended spawning season. Meth- ods of calculating annual fecundity from mea- surements of the total number of growing oocytes at the beginning of the spawning season (Bagenal 1966), or other approaches based on the total number of oocytes at the beginning of the spawn- ing season minus egg retention at the end of the spawning season (Conover 1985), are inappropri- ate for multiple spawning species such as spotted seatrout which show continuous recruitment of oocytes during the reproductive season. Thus, the previous estimates of annual fecundity in spotted seatrout, which did not take continuous recruit- ment of oocytes into consideration and measured either total fecundity (Overstreet 1983) or fecun- dity of growing and vitellogenic oocytes (Sun- dararaj and Suttkus 1962), probably underesti- mated annual fecundity. Furthermore, BF has not previously been calculated for this species. Although no monthly differences in BF were apparent (Table 4), sample sizes were too small to draw any definite conclusions from these data. The relatively low coefficient of determination (0.56) is similar to values reported by Conover (1985) for Atlantic silversides, Menidia menidia, another multiple spawning species. Perhaps a more accurate estimate of annual fecundity than previously reported for spotted seatrout can be obtained by multiplying BF by the number of spawns during the reproductive season. Unfortu- nately, as discussed later, estimates of spawning frequency vary considerably, so it is not possible to make an accurate estimation of the annual fe- cundity. However, available data indicate that average annual fecundity may be greater than 10 million eggs. Multiple Spawning Histological examination of the testes revealed that spermatogenesis ceased earlier in the central lobules, although they contained spermatozoa one-and-one-half months longer than the periph- eral lobules. It is possible that the same central lobules act as storage areas for spermatozoa pro- duced by the more spermatogenically active pe- ripheral lobules during the second half of the spawning season, as suggested by Hyder (1969) for Tilapia . This may represent a strategy in mul- tiple spawning fish with a prolonged spawning season that allows for a constant supply of sper- matozoa while investing a minimal amount of en- ergy into spermatogenesis. Several lines of evidence indicate that female C. nebulosus also spavni several times during the reproductive season. The relatively high percent- age of running ripe females, fish undergoing FOM and partially spent fish captured through- out the spawning season, and the absence of com- pletely spent fish until the last third of the spawn- ing season, suggest that an individual does not spawn all the vitellogenic eggs in the ovary at one time. Indeed, oocyte size-frequency analysis shows a continuous distribution of growing and 385 FISHERY BULLETIN: VOL. 86, NO, 2 vitellogenic oocytes (Fig. 5) and fecundity esti- mates show that only about 15% of the growing oocytes undergo FOM prior to a spawn. Histolog- ical data shows that the percentage of vitellogenic oocytes in the ovary remains constant throughout the spawning season (Fig. 4), which suggests that new oocytes may be recruited into the vitellogenic phase as rapidly as mature oocytes are released. Convincing histological evidence of multiple spawning is the presence of postovulatory follicles (POF) from May through the end of September in ovaries containing many vitellogenic oocytes. Hunter and Goldberg (1980) characterized pos- tovulatory follicles in laboratory-spawned En- graulis mordax, a multiple spawning fish, and found POF in all females that had spawned in the laboratory one or two days previously. Finally, laboratory studies also show that spotted seatrout are capable of multiple spawning under relatively constant environmental conditions (Table 6). Tucker and Faulkner (1987) also found that six female fish kept in raceways outdoors at the am- bient summer temperature and photoperiod spawned repeatedly. Spawning Frequency It is especially difficult to determine the spawning frequency of wide-ranging, multiple- spawning marine fishes such as C. nebulosus that are not group-synchronous spawners. One method to estimate spawning frequency is to count the number of distinct batches of vitel- logenic oocytes in the ovary (Shackley and King 1977). However, only one distinct batch of vitel- logenic or hydrated oocytes can be distinguished in spotted seatrout ovaries at any one time (Fig. 5) and the reliability of this method has been questioned (deVlaming 1983). Therefore, three techniques were used to estimate spawning fre- quency in spotted seatrout. Spawning frequency was estimated to be once every 3.6 days from the percentage of running ripe fish caught on the spawning grounds. Al- though this is probably an overestimate owing to sampling bias, the error may not be susbtantial, since the spawning grounds are also the feeding grounds for this species (Moody 1950) and many nonspawning individuals were captured. The time of sample collection did not significantly in- fluence the estimate of spawning frequency. High spawning frequencies were also obtained (every 2.3 days) when fish were captured 6-12 hours prior to spawning and examined for signs of final oocyte maturation (Table 5). The spawning fre- quency of other sciaenids fishes has been esti- mated by this technique (DeMartini and Foun- tain 1981; Love et al. 1984). Additionally, Hunter and Macewicz (1985) suggested that this method produces a useful first approximation of spawning frequency. The proportion of fish having POF in the ovary has also been used to determine spawning fre- quency (Hunter and Goldberg 1980; Albeit et al. 1984; Hunter and Macewicz 1985). Spotted seatrout were found to have POF throughout the spawning season, although the age of the POF was often difficult to determine. Furthermore, de- tailed laboratory studies have not been under- taken to accurately age POF in spotted seatrout. However, the once a week estimate of spawning frequency obtained using this method is similar to spawning frequencies reported for two other sciaenids, the queenfish, Seriphus politus, (De- Martini and Fountain 1981) and the white croaker, Genyonemus lineatus, (Love et al. 1984). Spawing frequency estimates from POF are prob- ably more reliable than estimates based on the number of spawning fish since sampling bias is less likely to occur when capturing fish with POF. Another method used to quantify spawning fre- quencies in various species is direct observation of spawning in the laboratory or in "controlled" field situations, such as impoundments (Gale and Deutsch 1985; Hubbs 1985; Heins and Rabito 1986). Spotted seatrout spawned an average of once every three weeks per individual under con- trolled temperature and photoperiod in the labo- ratory (Table 6). Tucker and Faulkner (1987) found that six female spotted seatrout kept out- doors at ambient temperature and photoperiod averaged one spawn per individual every 2.3 weeks. The same spawning frequency was noted for an individual female, although that same in- dividual later spawned three times in four days (Tucker and Faulkner 1987). Thus, spotted seatrout appear to be capable of the high spawn- ing fi"equencies estimated from field-caught fish with hydrated ovaries, although this frequency is probably not sustained throughout the entire spawning season. In general, the spawning fre- quencies in both laboratory studies are lower than those estimated from actual spawning fish in the field. However, it is unclear whether this is due to an overestimation of spawning frequency in the field or to a decline in spawning frequency owing to confinement in the laboratory. 386 BROWN-PETERSON ET AL.: REPRODUCTIVE BIOLOGY OF THE SPOTTED SEATROUT ACKNOWLEDGMENTS We thank John Trant, John Gourley, John Smith, and Wayne Wofford for their help with the field work and other aspects of the research. Funding was provided by grants from the Sid Richardson Foundation and the Caesar Kleberg Foundation. LITERATURE CITED Alheit. J . V H. Alaron, and B. J. Macewicz. 1984. Spawning frequency and sex ratio in the peruvian anchovy, Engraulis ringens. CalCOFI. Rep. 25:43-52. Arnold. C R . T D Williams, W. A. Fable, Jr., J. L Lasswell, AND W H Bailey. 1976. Methods and techniques for spawning and rearing spotted seatrout (Cynoscion nebulosus ) in the laboratory. Proc. 30th Conf Southeast. Assoc. Game Fish Comm. 30:167-178. Bagenal. T B. 1966. The ecological and geographical aspects of the fe- cundity of the plaice. J. Mar. Biol. Assoc. U.K. 46:161- 186. Bagenal, T. B., and E. Braum. 1971. Eggs and early life history. In W. E. Ricker (edi- tor), Methods for assessment of fish production in fresh waters, p 159-181. IBP (Int. Biol. Prgramme). Handb. 3, 2d ed., Blackwell Sci. Publ. Oxford, England. Brown, N J. 1981. Reproductive biology and recreational fishery for spotted seatrout, Cynoscion nebulosus, in the Chesa- peake Bay area. MA Thesis, College of William and Mary, Gloucester Point, VA, 120 p. Conover, D O. 1985. Field and laboratory assessment of patterns in fe- cundity of a multiple spawning fish: The Atlantic silver- side, Menidia menidia. Fish. Bull., US. 83:331-341. DeMartini, E. E . AND R K. Fountain. 1981. Ovarian cycling frequency and batch fecundity in the queenfish, Seriphus politus: attributes representa- tive of serial spawning fishes. Fish. Bull., U.S. 79:547- 560. deVlaming, V. 1983. Oocyte development patterns and hormonal in- volvements among teleosts. In J. C. Rankin, T. J. Pitcher, and R. T. Duggan (editors). Control processes in fish physiology, p. 176-199. Wiley & Sons, N.Y. deVlaming, V , G Grossman, and F Chapman. 1982. On the use of the gonosomatic index. Comp. Biochem. Physiol. 73A:31-39. DeVries, D a., and M. E Chittenden, Jr. 1982. Spawning, age determination, longevity and mor- tality of the silver seatrout, Cynoscion nothus , in the Gulf of Mexico. Fish. Bull., U.S. 80:487-500. Gale, W. F., and W G Deutsch. 1985. Fecundity and spawning fi-equency of captive tesse- lated darters - fractional spawners. Trans. Am. Fish. Soc. 114:220-229. Grier, H J. 1981. Cellular organization of the testis and spermatoge- nesis in fishes. Am. Zool. 21:345-357. HEIN, S , AND J SHEPARD. 1979. Spawning of spotted seatrout in a Louisiana estuar- ine ecosystem. Proc. Annu. Conf. Southeast. Assoc. Fish. Wildl. Agencies 33:451-465. Heins, D. L , AND F. G. Rabito, Jr. 1986. Spawning performance in North American min- nows: Direct evidence of the occurrence of multiple clutches in the genus Notropis. J. Fish Biol. 28:343- 357. Holt, G. J., S. A. Holt, and C. R Arnold. 1985. Diel periodicity of spawning in sciaenids. Mar. Ecol. Prog. Ser. 27:1-7. HUBBS, C. 1985. Darter reproductive season. Copeia 1985:56-68. Hunter, J R , and S. R Goldberg. 1980. Spawning incidence and batch fecundity in north- em anchovy, Engraulis mordax. Fish. Bull., U.S. 77:641-652. Hunter J. R., and B. J. Macewicz. 1985. Measurement of spawning frequency in multiple spawning fishes. In R. Lasker (editor). An egg produc- tion method for estimation spawning biomass of pelagic fish: application to the northern anchovy, Engraulis mor- dax, p. 79-94. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 36. Hyder, M. 1969. Histological studies on the testis of Tilapia leucos- ticta and other species of the genus Tilapia (Pisces: Teleostei). Trans. Am. Microsc. Soc. 88:211-231. Jones, R. McC. 1966. Basic microscopic techniques. 5th ed. Chicago Univ. Press. (See p. 249). Klima, E F , AND D. C. Tabb. 1959. A contribution to the biology of the spotted weak- fish Cynoscion nebulosus from northwest Florida, with a description of the fishery. Fla. State Board Conserv. Tech. Ser. 30, 24 p. 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 (Sciaenidae), off Califor- nia. Fish. Bull., U.S. 82:179-198. Macer, C. T. 1974. The reproductive biology of the horse mackerel, Trachurus trachurus (L.) in the North Sea and English Channel. J. Fish Biol. 6:415-438. Mahood, R K. 1975. Spotted seatrout in coastal waters of Geor- gia. Proc. Annu. Conf Southeast. Assoc. Game Fish Comm. 29:195-207. Mercer, L. P 1984. A biological and fisheries profile of spotted seatrout, Cynoscion nebulosus. N.C. Dep. Nat. Res. Comm. Dev. Div. Mar. Fish., Spec. Sci. Rep. No. 40, 87 p. Merriner, J V. 1976. Aspects of the reproductive biology of the weakfish, Cynoscion regalis (Sciaenidae), in North Carolina. Fish. Bull., U.S. 74:18-26. Miles, D W 1950. The life histories of the spwtted seatrout, Cynoscion nebulosus and the redfish, Sciaenops ocellatus. Tex. Game Fish Comm. Mar. Lab. Annu. Rep. 1949-1950, 38 P- 1951. The life histories of the seatrout, Cynoscion nebulo- sus, and the redfish, Sciaenops ocellatus: Sexual develop- ment. Tex. Game Fish Comm. Mar. Lab. Annu. Rep. 1950-1951, 11 p. 387 FISHERY BULLETIN: VOL. 86, NO. 2 Moody, W. D. 1950. A study of the natural history of the spotted seatrout, Cynoscion nebulosus, in the Cedar Key, Florida area. Q. J. Fla. Acad. Sci. 12:147-172. NiKOLSKn, G. V 1969. Theory offish population dynamics as the biological backgtound for rational exploitation and management of fishery resources. Oliver & Boyd, Edinb., 323 p. OVERSTREET, R. M. 1983. Aspects of the biology of the spotted seatrout, Cynoscion nebulosus, in Mississippi. Gulf Res. Rep. Suppl. 1:1-43. 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. Shackley, S E , AND P E King 1977. The reproductive cycle and its control; frequency of spawning and fecundity in Blennius pholis L. J. Exp. Mar. Biol. Ecol. 30:73-83. SPSS Inc. 1981. SPSS user's guide. 2d ed. SPSS, Inc., 988 p. Stewart, K W 1961. Contributions to the biology of the spotted seatrout (Cynoscion nebulosus) in the Everglades National Park, Florida. M.S. Thesis, Univ. Miami, Coral Gables, FL, 103 p. Sundararaj, B I , and R D Suttkus 1962. Fecundity of the spotted seatrout Cynoscion nebulo- sus (Cuvier) from Lake Borgne area, LA. Trans. Am. Fish. Soc. 91:84-88. Tabb, D C 1961. A contribution to the biology of the spotted seatrout, Cynoscion nebulosus of East Central Florida. Fla. State Board Conserv. Tech. Ser. 35, 21 p. 1966. The estuary as a habitat for spotted seatrout, Cynoscion nebulosus. Am. Fish. Soc. Spec. Publ. 3:59- 67. Tucker, J W , Jr , and B E Faulkner. 1987. Voluntary spawning patterns of captive spotted seatrout. Northeast Gulf Sci. 9:59-63. White. M L . and M E Chittenden, Jr 1977. Age determination, reproduction and population dynamics of the Atlantic croaker, Micropogonias undula- tus. Fish. Bull., U.S. 75:109-123. Yamamoto, N. 1956. Studies on the formation of fish eggs. I. Annual cycle in the development of ovarian eggs in the flounder, Liopsetta obscura. J. Fac. Sci. Hokkaido Univ. Ser. VI, Zool. 12:362-373. 388 NOTES NOTES ON REPRODUCTION IN THE SCALLOPED HAMMERHEAD, SPHYRNA LEWINI, IN NORTHEASTERN TAIWAN WATERS The reproductive mode of hammerhead sharks (family Sphyrnidae) is placental viviparity, fol- lowing Teshima's (1981) designations. Three spe- cies of hammerhead sharks are known from northeastern Taiwanese waters: Sphyrna lewini (Griffith and Smith), S. mokarran (Riippel), and S. zygaena (Linnaeus). The scalloped hammer- head, S. lewini, commonly found from Pung-Chia Island to Guei-Shan Island (Fig. 1) is one of the most abundant shark species in this area. Based on data from the Nan Fan Ao Fish Market (lo- cated in northeastern Taiwan), 460-510 t/year from 1982 through 1984 were landed, represent- ing one-fourth of the total catches of sharks in this area (the total landing being 1,760-2,240 t/ year). Catches peaked in spring and winter and were lower in summer and autumn. This species is also distributed in western and southern waters of Taiwan; however, the catch from those areas were smaller. The employees of the southern and western fish markets explained that the scalloped hammerhead is occasionally seen in small numbers in those areas, but there are no landing data. It is also common in coastal warm temperate and tropical seas throughout much of the world (Compagno 1984). Although 26°N Figure 1. — Sampling area of Sphyrna lewini. FISHERY BULLETIN; VOL. 86, NO. 2, 1988 25°N the scalloped hammerhead is one of the most valuable food resources in Taiwan, many facets of its life history, particularly reproduction, are not well known. This study provides information about certain aspects of reproduction in the scal- loped hammerhead in northeastern Taiwan waters. Materials and Methods From September 1982 to June 1983 and from December 1983 to September 1985, shark speci- mens were examined and material collected monthly at Nan Fan Ao Fish Market. These sharks had been caught by drift longlines set near the surface to around 100 m, or by surface har- poon. A total of 674 scalloped hammerhead sharks were examined at the fish market (Table 1). Data recorded included total length measured from the tip of the snout to the tip of the upper lobe of the caudal fin (straight line measure), body weight, clasper length (measured from cloaca to the tip of claspers), ovarian egg diameter and number, and condition of the uteri. In addi- tion, uterine embryos were counted, sexed, and measured for total length. Counts of litter size included any uterine eggs as well as embryos. Stages of maturity for both females and males were categorized simply as "mature" or "imma- ture". Females having threadlike uteri and tiny ovarian eggs were called immature, while those with eggs larger than 25 mm in diameter, with uteri containing embryos or eggs or with empty but expanded and flaccid uteri, were designated as mature. Males with rigid claspers were classi- fied as being mature. Clasper length relative to total length also gave an indication of maturity. Table 1 . — The number of specimens examined in this study. Number of specimens Immature Mature Month Male Female Male Female Jan. 13 11 18 93 Feb. 3 3 17 43 Mar, 4 13 4 44 Apr. 8 34 18 68 May 3 6 9 33 June 3 2 4 20 July 4 3 11 60 Aug. 3 1 6 19 Sept. 0 7 3 15 Oct. 0 1 1 3 Nov. 0 6 1 38 Dec. 1 8 1 10 Total 42 95 91 446 Results The reproductive organs of scalloped hammer- head closely resemble those of the bonnethead shark, S. tiburo (described by Schlernitzauer and Gilbert 1966). As with the bonnethead shark, only the right ovary of the scalloped hammerhead is functional, supplying both oviducts. As the ovarian eggs ma- ture (>25 mm diameter), they pass through the common ostium into the oviducts, where they are fertilized. The eggs then become encased in the embryonic membrane as they pass through the nidamental gland and descend into the uterus. In the uterus, embryonic development proceeds, nourished by a yolk sac. After a period, the uter- ine compartments develop, which enclose the em- bryo, and a yolk-sac placenta is implanted. After the yolk is exhausted, the embryo is nourished until birth by the placenta through the umbilical stalk. Based on the condition of the uterus and ovary, female scalloped hammerheads became mature at a larger size, around 210 cm, than males. All females over 230 cm were mature. Based on the rigidity of the claspers, male scal- loped hammerheads reached their first maturity at a total length of 198 cm, while all those over 210 cm were mature. The clasper length of males increases rapidly relative to total length, until the sharks reach around 200 cm, at which size the clasper length/ total length relationship plateaus, suggesting that sexual maturity has been attained (Fig. 2). Also, this size marked the approximate transi- tion point from fiaccid to rigid claspers (Fig. 2). It takes roughly 10 months of development from egg formation to ovulation (Fig. 3). In Octo- ber and November, eggs were very small, measur- ing about 2 mm in diameter. By July they had increased in size to about 30-38 mm and num- bered 40-50 per female, and by August and Sep- tember they had grown to about 40-45 mm. In one mature female, in September we counted 28 uterine eggs and 4 ovarian ones measuring 45 mm in diameter. This suggested that the 4 ovarian eggs were ready for ovulation. Because ovarian eggs larger than 30 mm in diameter seemed near ovulation, we concluded that ovula- tion occurred between July and October. The parturition season lasts from May to July, and the gestation period of this species was esti- mated to be roughly 10 months (Fig. 4). No em- 390 30 26 22 X I 18 LU 5 14 < •• • • • •• • • • •• • o o CO CDO ooo o o o 10 6 - o J. _L • : rigid O flaccid ■ ■ 120 140 160 180 200 220 240 TOTAL LE^4CTH (cm) 260 280 Figure 2. — Relationship between total length and clasper length in Sphyrna lewini. Jan heh Mtr Apf M«y Jun lul Au* Sep < >cl No* Dec Figure 3. — Monthly increase in the ovarian egg diameter of Sphyrna lewini. Error bars represent 2 SD; numerals, sample size. bryos were found in July and August, though some mature sharks had large ovarian eggs measuring about 30-40 mm in diameter during those months. The first uterine eggs, and embryos measuring around 2 cm were found in September. Uterine eggs were observed until November. After nine more months of rapid growth, the em- bryos attained a total length of around 45 cm in the period between May and June. These embryos were regarded as full term because they were eas- ily separated from placenta implying that partu- rition would occur soon. 391 O X UJ 50 40 30 20 10 -i_ -L. M M J J MONTH N [) Figure 4. — Monthly growth of embryos of Sphyrna lewini. Solid triangles indicate the uterine eggs. Error represent 2 SD; numerals, sample size. Counts of the number of uterine embryos as well as eggs of 110 gravid females (230-320 cm TL) ranged between 12 and 38 (mean 25.8). The relationship between the total number of uterine embryos or eggs and the size can be described by the regression equations (Fig. 5): N = -26.105 + 0.179L, where N is litter size, and L is total length in cm of the female. There is considerable variation in number of embryos with length, and the correla- tion coefficient r is low (0.567), but obviously fe- cundity is related to the size of the parent. Examination of all females carrying developing embryos showed that occasionally some uterine eggs failed to develop. The number of nondevelop- ing uterine eggs carried by a single female was 1-4. As with litter size, the number of ovarian eggs increased with the length of the adult female. The embryos from 51 gravid females were sexed; of a total of 1,281 embryos, 637 were fe- males. The sex ratios of embryos differed by litter; for instance, some individuals had predominantly male (13:5) or predominantly female (23:10) lit- ters. But, as a whole, sex ratio was about 1:1. The ratio of males to females for immature sharks in northeastern Taiwan waters was about 1:2 (42:91), but decreased to about 1:5 (95:446) for mature individuals (Table 1). There was no ap- parent increase in the relative number of males in the catch during the parturition season. Discussion Scalloped hammerhead sharks are abundant in the coastal seas around northeastern Taiwan, es- pecially during the spring and winter, and are captured by harpoon or drift longlines at or near the surface. Rarely individuals smaller than 120 cm TL are captured because fishing gear and strategy are likely selective for larger fish. Clarke (1971) reported that in Hawaii, scalloped hammerhead pups usually stay close to the bot- tom. Approximately equal numbers of male and fe- male scalloped hammerhead sharks are born, al- though a much higher proportion of females than males are caught in the studied area. We are un- certain whether scalloped hammerhead females are more vulnerable to the fishing gear or are simply more numerous in this area. Similar pre- dominances in females in the catch have been found in the scalloped hammerhead in the Gulf of California (Klimley 1981; Klimley and Nelson 1981). During the summer in 1979 off Baja Cali- fornia, females outnumbered males by 1.6 x at Isla Cerrolvo, 3.8 x at the El Bajo Seamount, and 40 38- 32 28 - o > a: 00 UJ 24 S 20 z 18 - •t - • • • • • • • • - • • • / • • • •• • X • • •/ • • ~ • ■ • • •/ • • ••• • " • • " y • • \ • ~ • / • •• • \ • X •• N = -26.105 + 0.179L / (r=0.5665) - y • • , (n=110) • • • • • : J l_ • 12- 240 280 280 300 320 TOTAL LENGTH (cm) Figure 5. — Relationship between fecundity (number of uterine eggs or embryos) and total length of Sphyrna lewini females. 392 3.1 X farther north at Isla Las Animas (Klimley 1981; Klimley and Nelson 1981). Similar dispari- ties in sex ratios were also observed in sandbar sharks by Springer ( 1960). He felt that adult male sandbar sharks live over a large geographical and depth range, perhaps in deep cool oceanic waters inaccessible to the fishermen's gear, while fe- males occur in warmer inshore water where they are more accessible to fishermen. He suggested that the males move inshore only to mate. We had insufficient data to test this suggestion in our sample area. Castro (1983) reported that S. lewini from North American waters probably mature at about 180 cm; he did not mention whether this length referred to male, female, or both. Bass et al. (1975) reported that male scalloped hammerhead at Mozambique matured between 140 and 165 cm, reaching a maximum length of at least 295 cm; females matured at about 212 cm, reaching at least 309 cm. Compagno (1984) reported "maxi- mum" sizes ranging from 370 to 420 cm. Our largest males were 305 cm TL, and the smallest mature male was 198 cm TL. The largest female observed was 324 cm, and the smallest mature female was 210 cm. It seems that in our sample area, males reached maturity at a somewhat larger size than in the other studied areas while females were about the same. The close relationship of growth pattern of uterine embryos and ovarian eggs implies that eggs are transferred into the uterus and fertilized immediately after parturition. If the estimate of 10 month gestation period is correct, adult fe- males give birth once each year. The length at birth of scalloped hammerheads has been reported to be around 50 cm from Natal and southern Mozambique coastal waters (Bass et al. 1975), 43 cm from northeastern United States to Chesapeake Bay waters (Casey 1964), 38-45 cm in North American waters (Castro 1983), and 42-55 cm combined from all oceans (Compagno 1984). Our largest uterine embryos measured about 47 cm TL. Pupping season appears to be during the sum- mer in Taiwan as well as in Mozambique (Bass et al. 1975) and North America (Castro 1983). In Hawaii, Clarke (1971) found newborn pups throughout the year but with increased numbers in summer. Fisheries Center, National Marine Fisheries Ser- vice, NOAA, for reading the first draft of this manuscript and offering useful comments. We also thank Jin-Jehn Wu, Department of Fish- eries, National Taiwan College of Marine Science and Technology, for his help in collecting speci- mens. Financial support was received from Na- tional Science Council (contract: NSC 72-0409- B019091 and NSC 73-0409-B019-02). Literature Cited Bass, A. J . J D. D' Aubrey, and N Kistnasamy. 1975. Sharks of the east coast of southern Africa. III. The families Carcharinidae and Sphymidae. Oceanogr. Res. Inst., Durban, Invest. Rep. No. 38, 100 p. Casey, J G 1964. Angler's guide to sharks of the northeastern United States, Maine to Chesapeake Bay. Bur. Sport Fish. Wildl., Circ. No. 179, 32 p. Castro, J I. 1983. The sharks of North American waters. Texas A&M Univ. Press, 180 p. Clarke, T A 1971. The ecology of the scalloped hammerhead shark, Sphyrna lewini, in Hawaii. Pac. Sci. 25:133-144. Compagno, L J. V. 1984. FAO species catalogue. Vol. 4. Sharks of the world. An annotated and illustrated catalogue of shark species known to date. Psirt 2. Carchariniformes. FAO Fish. Synop. No. 125, 4(2):251-655. Klimley, A. P 1981. Grouping behavior in the scalloped hammer- head. Oceans 24(4):65-71. Klimley. A P and D R. Nelson 1981. Schooling of scalloped hammerhead shark, Sphyrna lewini, in the Gulf of California. Fish. Bull., U.S. 79:356-360. Schlermitzauer, D a , and P W. Gilbert. 1966. Placentation and associated aspects of gestation in bonnethead shark, Shyrna tiburo. J. Morph. 120:219- 231. Springer, S. 1960. Natural history of the sandbar shark, Eulamia mil- berti. U.S. Fish. Wildl. Serv., Bull. 61:1-38. Teshima, K. 1981. Studies on the reproduction of Japanese smooth dogfish, Mustelus manazo, and M. griseus. J. Shi- monoseki Univ. Fish. 29:113-119. Che-Tsung Chen Tzyh-Chang Leu Shoou-Jeng Joung Graduate School of Fisheries National Taiwan College of Marine Science and Technology Keelung, Taiwan Republic of China Acknowledgments We sincerely thank Susumu Kato, Southwest 393 OCCURRENCE OF YOUNG-OF-THE-YEAR KING, SCOMBEROMORUS CAVALLA , AND SPANISH, 5. MACULATUS, MACKERELS IN COMMERCIAL-TYPE SHRIMP TRAWLS ALONG THE ATLANTIC COAST OF THE SOUTHEAST UNITED STATES^ King mackerel, Scomberomorus cavalla, and Spanish mackerel, S. maculatus , are migratory scombrids that support large recreational and commercial fisheries along the southeast coast of the United States (Manooch 1979). Recent evi- dence indicates that both species may be overex- ploited in portions of their range, prompting the South Atlantic Fishery Management Council to impose catch limits and landing quotas^. Many aspects of the biology and ecology of adult mack- erels in this region have been studied (Manooch et al. 1978; Collette and Russo 1984), and the larval stages have also received attention (Fahay 1975; Collins and Stender 1987). However, little is known concerning the distribution and occur- rence of juvenile (young-of-the-year) mackerels along the Atlantic coast of the southeastern United States, nor does it seem to be widely known that large numbers of these young fishes may be included in the bycatch of a major fishery. This report provides preliminary information on both of these topics. Methods During 1980-82 and 1985-86 the Marine Re- sources Monitoring and Assessment Program (MARMAP) at the South Carolina Marine Re- sources Research Institute conducted trawl sur- veys of the nearshore fish fauna in the South At- lantic Bight (Cape Hatteras, NC, to Cape Canaveral, FL). Before 1986, trawl gear consisted of two semiballoon shrimp trawls with an 18.3 m footrope, a 12.2 m headrope, 4.1 cm stretch mesh in the cod end, and 1.5 x 0.9 m doors towed at 2.5 knots (4.6 m/second). In 1986, paired "tongue" trawls with a 22.9 m footrope, 4.1 cm mesh, and 3.0 X 1.0 m doors were towed at approximately 2.5 knots. Sampling strategy and length of tow (20 minutes to 1 hour) varied between cruises. Station depths were 3-18 m in 1980-81 and 3-9 m in subsequent years. In each sample, all mack- erels were identified to species and measured (fork length), and number and total weight were recorded for each species. We conducted two addi- tional cruises in 1986 to test tongue trawl nets equipped with trawl efficiency devices (TED's). A National Marine Fisheries Service (NMFS) TED equipped with finfish deflector"^ was installed in one net, a Georgia TED^ in a second net, and a third net acted as a control. The TED's were fished against each other and against the control for a total of 30 tows on the first cruise (July-Au- gust 1986) and 15 tows on the second (September- October 1986). Tongue nets, also known as bib, falcon, cobra, or mongoose trawl nets, have become widely used by commercial shrimpers in some areas during recent years (Edwards 1987). The major differ- ence between these and other common towed gears used in the penaeid shrimp fishery is a mod- ified and elongated headrope that is held well above the footrope by attachment to the trawl warp with a third bridle. A large float, usually attached to the center of the headrope, produces a high, vertical mouth opening. The result is a net that fishes a larger portion of the water column than other common nets with similar footrope and door configurations. Results During years when semiballoon nets were used, catch per unit effort of mackerels was relatively low (king mackerel: 0.2-0.4 individuals/net-hour; Spanish mackerel: 0.3-2.2 individuals/net-hour). Tongue trawl nets on four 1986 cruises gave cruise-specific average catches of 2.5-8.7 individ- uals/net-hour for king mackerel and 12.4-115.2 individuals/net-hour for Spanish mackerel, using catches of control [unmodified] nets for TED cruises (Table 1). Both species were taken as late as 30 October, except for two king mackerel caught in December 1982. Approximately 79% of king and 91% of Spanish mackerel were caught in depths <9 m during the two years in which sam- ple depths extended to 18 m. Mackerels taken in shrimp trawls were almost entirely juvenile fishes. Rather than pool catches between cruises that often differed in time of year, geographic area, and sampling strategy, ex- amples of length frequencies of Spanish and king iContribution No. 244 of the South Carolina Marine Re- sources Center. ^South Atlantic Fishery Management Council, Council Meet- ing Summary, 27-29 April 1987, Charleston, SC. 3Described and illustrated in the Federal Register, vol. 52, no. 124; Monday, June 29, 1987 - Rules and Regulations; p. 24244- 24262. 394 FISHERY BULLETIN: VOL. 86, NO 2, 1988. Table 1. — Results of nearshore cruises using two types of shrimp trawls. Cruise dates Net type Sampling! strategy Sample area2 Total net-hour No. Spanish No. kings July-Sept. 1980 semiballoon SR + NR NC, SC, GA, FL 114 250 40 Apr.-June 1981 semiballoon SR NC. SC. GA. FL 77 21 13 Sept. 1982- Jan. 1983 semiballoon SR NC, SC, GA, FL 73 27 27 Aug. -Sept. 1985 semiballoon SR + NR NC, SC, GA, FL 59 120 36 Aug. -Sept. 1986 tongue NR NC, SC, GA, FL 41 1,421 104 Oct. 1986 tongue SR NC, SC, GA 55 681 481 July-Aug. 1986 tongue NR SC 20 2,303 84 Sept.-Oct. 1986 tongue NR SC 20 327 80 ^SR = stratified random sampling; NR = nonrandom sampling. 2At least some portions of the waters of these states were sampled. mackerels taken in tongue nets on two cruises are presented in Figures 1 and 2, respectively. Com- parisons of the catch rates of Spanish mackerel during the TED evaluations in July- August 1986 showed that significantly fewer were taken in the tongue net equipped with a NMFS TED than the control net (Mann-Whitney: P < 0.001) or the Georgia TED-net (P < 0.05) (Table 2). There were no significant differences in catches of king mack- erel (Table 3), or in catches of either species be- tween nets on the September-October cruise, per- haps due to the smaller sample sizes. Figure 1. — Length-frequency distribution of Spanish mackerel taken in shrimp tongue trawl nets during July-August 1986 along the southern At- lantic coast of the United States in = 2,303). FORK LENGTH (cm) 395 130 -I 24 28 32 36 FORK LENGTH (cm) r 40 I 44 Figure 2. — Length-frequency distribution of king mackerel taken in shrimp tongue trawl nets during October 1986 along the southern Atlantic coast of the United States (n = 481). Table 2. — Catches of Spanish) mackerel during evaluations of tongue nets equipped with trawl efficiency devices (TED'S). C = 22.9 m footrope length tongue net; NMFS = 22.9 m footrope length tongue net with a NMFS TED; GA = 22.9 m footrope length tongue net with a Georgia TED. Tow times = 1 hour. No. of Net No. of Spanish mackerel Net No. of Spanish mackerel Cruise Comparison tows type Total X ±SD Range type Total X ±SD Range July-Aug. C-NMFS 10 C 1,104 110 ±27 68-144 NMFS 519 52 ±23 12-73 C-GA 10 C 1,199 120 ±65 40-276 GA 1,219 122 ± 74 44-300 NIVIFS-GA 10 NMFS 880 88 ±53 35-176 GA 1,650 165 ±142 67-533 Sept.-Oct. C-NMFS 5 C 104 21 ± 15 2-41 NMFS 53 11 ± 11 2-27 C-GA 5 C 206 41 ±38 1-103 GA 155 31 ±28 0-76 NI^FS-GA 5 NMFS 108 22 ±23 4-58 GA 149 30 ±25 3-59 Table 3. — Catches of king mackerel during evaluations of tongue nets equipped with trawl efficiency devices (TED'S). C = 22.9 m footrope length tongue net; NMFS = 22.9 m footrope length tongue net with a NMFS TED; GA = 22.9 m footrope length tongue net with a Georgia TED. Tow times = 1 hour. Comparison No. of tows Net type No. of king mackerel Net type No. of king mackerel Cruise Total X ± SD Range Total X ± SD Range July-Aug. C-NMFS 10 C 41 4± 4 1-13 NMFS 46 5± 9 0- -30 C-GA 10 C 43 4± 5 0-15 GA 41 4 ± 4 0- -13 NMFS-GA 10 NMFS 23 2± 3 0-8 GA 26 3± 5 0- -17 Sept.-Oct. C-NMFS 5 C 26 5± 7 0-15 NMFS 14 3± 3 0- -5 C-GA 5 C 40 8± 8 4-22 GA 22 4 ± 5 1- -14 NMFS-GA 5 NMFS 24 5± 2 2-8 GA 20 4± 4 2- -11 Discussion Although it is possible that juvenile mackerels were more abundant in 1986 than in previous years, the increased catches of these fishes in tongue nets over semiballoon nets suggests that the former are much more efficient in capturing these fishes. Preliminary data from a gear com- parison cruise in 1987 indicate that tongue nets do catch more pelagic fishes than semiballoon nets even after adjusting for differences in footrope lengths (G. Sedberry"*). Unfortunately, ■*G. Sedberry, Marine Resources Research Institute, South Carolina Wildlife and Marine Resources Department, P.O. Box 396 these tests were conducted during March-April when juvenile mackerels are rare in the coastal waters of South Carolina. In 1986 we collected juvenile mackerels during July through October. Because of incomplete tem- poral sampling, we do not know if they were present earlier and later in the year in this re- gion. Based on the occurrence of early larval stages, spawning of both mackerels in the South Atlantic Bight extends from May through at least September (Collins and Stender 1987). If growth rate estimates of ca. 3 mm/day for juveniles are correct (M. R. Collins, unpubl. data), king mack- erel spawned in early May could be recruited into the bycatch of the shrimp fishery in June. Late- spawned fish from the previous year may also be present at this time. In South Carolina, the open season for commercial trawling for penaeid shrimps in state waters usually extends from June through December, which coincides with the presence of juvenile mackerels in the heavily fished nearshore waters. In addition, mackerels were much more abundant in tows made in depths <9 m, which includes the preferred shrimping areas, than in deeper waters. This may be due either to greater abundance in these depths or to greater catchability in response to the fact that the trawl nets fish a larger portion of the water column in shallower areas. It is difficult to accurately estimate the bycatch of mackerels in the commercial shrimp fishery owing to lack of current, detailed information from throughout the region on number of vessels, effort expended, gears used, and areas fished. However, our catch rates suggest that the impact of tongue nets on mackerel stocks may be signifi- cant. As the current status of these stocks is such that strong restrictions have been imposed on both the recreational and commercial fisheries, it is unfortunate that the situation may be exacer- bated by a potentially large bycatch of juvenile mackerels in the shrimp fishery. More informa- tion is needed on the ecology and behavior of young mackerels, and their vulnerability to vari- ous gears, in order to resolve this conflict. Acknowledgments We wish to express our appreciation to the many people who assisted in the tedious tasks of collecting and processing trawl samples over the years, including the captains and crews of the vessels involved. This work was sponsored by the National Marine Fisheries Service (Southeast Fisheries Center), the South Carolina Wildlife & Marine Resources Department, and the Sport Fishery Research Foundation. Literature Cited COLLETTE. B B . AND J L RUSSO 1984. Morphology, systematics, and biology of the Spanish mackerels (Scomberomorus , Scombridae). Fish. Bull., U.S. 82:545-692. Collins, M R . and B W Stender. 1987. Larval king mackerel (Scomberomorus cavalla), Spanish mackerel (S. maculatus), and bluefish iPo- matomus saltatrix ) off the southeast coast of the United States, 1973-1980. Bull. Mar. Sci. 41:822-834. Edwards. M L. 1987. The shrimp trawl. Natl. Fisherman 67(13):89-92. Fahay, M P 1975. An annotated list of larval and juvenile fishes cap- tured with surface-towed meter net in the South Atlantic Bight during four RV Dolphin cruises between May 1967 and February 1968. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-685, 39 p. Manooch. C S , III 1979. Recreational and commercial fisheries for king mackerel, Scomberomorus cavalla , in the South Atlantic Bight and Gulf of Mexico, U.S.A. In E. L. Nakamura and H. R. Bullis (editor), Proceedings of the mackerel colloquium, p. 33-41. Gulf States Mar. Fish. Comm., Brownsville, TX. Manooch. C S , III. E L. Nakamura, and A B Hall 1978. Annotated bibliography of four Atlantic scombrids: Scomberomorus brasiliensis , S. cavalla, S. maculatus, and S. regalis. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 418, 166 p. Mark R Collins Charles A. Wenner Marine Resources Research Institute South Carolina Wildlife and Marine Resources Department P.O. Box 12559 Charleston, SC 29412 12559, Charleston, SC 29412, pers. commun. May 1987. STOMACH CONTENTS OF COMMERCIALLY CAUGHT HUDSON RIVER STRIPED BASS, MORONE SAXATILIS , 1973-75 The Hudson River estuary is a detritus-driven ecosystem. Only a few of the 100 or more reported fish species function as tertiary piscivores more typical of a grazing food chain. Of these few spe- cies, which include the American eel, Anguilla rostrata, and the summer-transient juvenile FISHERY BULLETIN: VOL. 86, NO. 2. 1988. 397 bluefish, Pomatomus saltatrix, striped bass, Mo- rone soxatilis, are probably the most important in terms of biomass and commercial value. How- ever, commercial fishing for the American eel and striped bass has been banned in the Hudson since 1976 because both species are contaminated with higher levels of toxic polychlorinated biphenyls (PCB's) than most other Hudson River fish spe- cies. To date, only one published paper has touched upon the diet of prespawning Hudson River striped bass larger than 400 mm TL (Gardinier and Hoff 1982), and their results for fish of that size are based on less than 10 samples for which fish remains were identified to species or family level. The present paper describes the findings of stomach content analysis for 510 striped bass, most of which were prespawning adults larger than 400 mm TL (Fig. 1) collected during the spring months of 1973-75 by commercial gill nets in the Tappan Zee region of the Hudson River. Description of Sampling Area and Methods The Tappan Zee gill-net fishery is located ap- proximately 90 m south of the Tappan Zee Bridge at river km 43.5 (measured from the Manhattan Battery) in a relatively shallow section that is 4.5 km wide, a location through which many of the spawners move on the way to upstream spawning grounds. The staked gill nets were set perpendic- ularly to the north-south current in water depths of 2.4-9.8 m and spanned a distance of approxi- mately 1.6 km out to the dredged channel on the east side of the river. Stretched mesh sizes were 11.4, 11.8, 12.1, 12.7, 13.3, and 14.0 cm. Water temperatures during the sampling period ranged from approximately 6° to 19.5°C, and the salinity was essentially that of freshwater, rarely rising above 300 mg/liter. Sampling dates included the following periods in each year: 28 March-9 May 1973, 8 April-20 May 1974, and 6 April-19 May 1975. Fish were collected from the nets at 8-12 h intervals, placed on ice, and taken directly to a laboratory for total length (TL) measurements and stomach content analysis. Prey items were identified with the aid of a binocular dissecting scope and frequency of occurrence was noted. Results Stomach Content Analysis A summary table of stomach content data, pooled for the three years of analysis (Table 1), shows that adult, prespawning striped bass were highly piscivorous and that 89% of the stomachs containing identified food items contained fish and 21% contained invertebrate remains, primar- ily those ofCrangon species. Clupeid species were the most prevalent fish, with blueback herring, Alosa aestivalis, predominating. Most of the clu- peids were adult, prespawning herring, approxi- mately 200 mm TL. NUMBER OF FISH = 510 140 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 TOTAL LENGTH (mm) Figure 1. — Length distribution of Hudson River striped bass examined for stomach con- tents. 398 Table 1— Stomach contents of adult prespawning striped bass collected by commercial gill net during spring 1973, 1974, and 1975. % Frequency of % Frequency of occurrence Number of occurrence based on based on number of stomachs total stomachs stomachs containing in which analyzed identified food items Food item item occurred (n = 510) (n =201) Fish 179 35.1 89.1 Clupeidae 73 14.3 36.3 Alosa pseudoharengus 13 2.5 6.5 Alosa aestivalis 29 5.7 14.4 Brevoortia tyrannus 3 0.6 1.5 Clupeidae 28 5.5 13.9 Morone spp. 9 1.8 4.5 Morone americana 6 1.2 3.0 Morone saxatilis'^ 1 0.2 0.5 Morone spp. 2 0.4 1.0 Other fish 29 5.7 14.4 Ammodytes americanus 14 2.7 7.0 Microgadus tomcod 7 1.4 3.5 Osmerus mordax 2 0.4 1.0 Syngnathus fuscus 2 0.4 1.0 Anchoa mitciiilii 2 0.4 1.0 Ictalurus catus 1 0.2 0.5 Scomber scombrus 1 0.2 0.5 Unidentified fish remains 68 13.3 33.8 Invertebrates 43 8.4 21.4 Crangon spp. 38 7.5 18.9 Polyctiaeta 5 1.0 2.5 Unidentified contents 11 2.2 Not applicable Total stomachs analyzed 510 Number (percent) containing identified food items 201 (39.4) Number (percent) containing food material (identified and unidentified) 212 (41.6) Number (percent) empty 296 (58.4) 'Two yearling striped bass in female collected 15 April 1974. The next most abundant fish prey species in the composite 1973-75 sample was the American sand lance, Ammodytes americanus. Sand lance were present in striped bass stomachs only during 1975, when the frequency of occurrence was about ll9c (Fig. 2). Atlantic tomcod, Microgadus tom- cod, were found in striped bass stomachs only during 1973 when the frequency of occurrence was about 20%, somewhat higher than the fre- quency of clupeid species for that year (Fig. 2). These fish were yearling, postspawning Atlantic tomcod averaging about 125 mm TL. The Morone species, white perch, M. americana, and striped bass, had a frequency of occurrence of 4.5% for the composite sample and were present in striped bass stomachs every year (Fig. 2). Discussion Similar to findings in this study, Trent and Hassler (1966) found that blueback herring were predominant in the diet of adult, prespawning striped bass from the Roanoke River, NC, and Manooch (1973) found that herring ranked second after Atlantic menhaden, Brevoortia tyrannus , in the spring diet of adult striped bass from Albe- marle Sound, NC. Data from Gardinier and HofF (1982) do not indicate that blueback herring or clupeids were important in the diet of prespawn- ing Hudson River striped bass collected during 1976 and 1977, but fish remains in bass stomachs were generally not identified to the family or spe- cies level in their study. American sand lance are rarely found over muddy bottoms (Bigelow and Schroeder 1953; Leim and Scott 1966) or within estuaries (Mass- man 1960; Norcross et al. 1961), but they ranked second only to blueback herring in the diet of Hudson River striped bass. Sand lance were not reported from the Hudson River prior to 1975 when adults and larvae were relatively abundant in February-April ichthyoplankton collections south of the Tappan Zee Bridge (Dew and Hecht 1976). Data from the present study (Fig. 2) and from early spring ichthyoplankton sampling dur- ing 1976 (Dew, unpub. data), indicate that sand lance are not abundant every year in the Hudson, but it is probable that during some years adult sand lance serve as important alternate prey for prespawning striped bass. 399 o o o o >- o 3 o 8? 100 80 .. 60 40 20 ^ CLUPEIDS [^ MORONE SPP. ^ ATLANTIC TOMCOD □ SAND LANCE □ UNIDENTIFIED FISH REMAINS N = 36 i it -iii 1973 N = 39 i Fl i I i A N =126 1974 1975 N = 201 P i i ^±.„L^ 1973- 1975 Figure 2. — Percent frequency of occurrence offish prey categories based on number of stomachs containing identified food. The fact that Atlantic tomcod were found only during 1973 may be due to a sampling artifact because 1973 was the only year in which March samples were taken, and all tomcod prey were found during March 1973. In March, gravid alewives, Alosa pseudoharengus , and blueback herring have not yet moved into the estuary and, prior to the arrival of these apparently preferred prey species, Atlantic tomcod may serve as an alternate forage fish. Thus the winter-spawning Atlantic tomcod may be a regular diet item at a time when alternate prey species are not avail- able. It is evident from Table 1 that there are several alternate fish prey for striped bass in the Hudson River (e.g., blueback herring, sand lance, alewife, Atlantic tomcod, white perch, and striped bass), and it is expected that striped bass predation upon these species would be of a compensatory nature. In a simple predator-prey situation where there are no alternate prey, predation is often depensatory (Neave 1953) because a population of predators would tend to seek out and consume a high proportion of small year classes and a lower proportion of strong year classes. In a more com- plex community where there are several species of prey, predation may be compensatory if preda- tors change their feeding habits in response to the availability of food (Ivlev 1961; Forney 1971). That is, if several equally suitable prey species were available within a system, the predator spe- cies would tend to feed more heavily on the most abundant species, thus acting in a compensatory manner. The ratio of white perch to striped bass in stom- ach samples identified to the species level is 6:1 (Table 1). This ratio is based on a small sample size, but it closely approximates the long-term average ratio of 5.9:1 for annual CPUE values for white perch and striped bass collected in several hundred bottom trawls from the Haverstraw Bay region (Milepoint 36) during 1971-77 (Lawler, Matusky and Skelly Engineers unpubl. data). In other words, yearling and older white perch and yearling striped bass may be equally attrac- tive prey species, and their frequency of oc- currence in adult striped bass stomachs may depend primarily upon the frequency of ran- dom encounters rather than active prey selec- tion. If this were true, the frequency of can- nibalism should be greatest during those years when the abundance of yearling striped bass rela- tive to white perch (and other species) was great- est. Acknowledgments Financial support for this project came from Lawler, Matusky and Skelly Engineers; Con- solidated Edison of New York; Central Hudson Gas and Electric; and Orange and Rockland Utilities. 400 Literature Cited BiGELOW. H. 3., AND W. C. SCHROEDER 1953. Fishes of the Gulf of Maine. Fish. Bull. 53:1- 577. Dew. C B . and J H Hecht 1976. Observations on the papulation dynamics of At- lantic tomcod Microgadus tomcod in the Hudson River estuary. Proc. Fourth Symposium On Hudson River Ecology, Paper 25. Hudson River Environmental Soci- ety, Bronx, N.Y. Forney, J L 1971. Development of dominant year classes in a yel- low perch population. Trans. Am. Fish. Soc. 100:739- 749. GaRDINIER. M N , AND T B HOFF 1982. Diet of striped bass in the Hudson River estuary. New York Fish Game J. 29(2):152-165. IVLEV, V S. 1961. Experimental ecology of the feeding of fishes. Yale Univ. Press, New Haven, 302 p. Leim. a H . AND W B Scott 1966. Fishes of the Atlantic coast of Canada. Fish. Res. Board Can. Bull. 155, 485 p. Manooch, C S. 1973. Food habits of yearling and adult striped bass, Mo- rone saxatilis (Walbaum), from Albemarle Sound, North Carolina. Chesapeake Sci. 14(2):73-86. Massmann, W H 1960. Additional records for new fishes in Chesapeake Bay. Copeia 1960:1-70. Neave, F. 1953. Principles affecting the size of pink and chum salmon populations in British Columbia. J. Fish. Res. Board Can. 9:450-491. Norcross, J J , W H Massmann, and E. B. Joseph 1961. Investigations of inner continental shelf waters off lower Chesapeake Bay. Part II. Sand lance larvae, Ammodytes americanus. Chesapeake Sci. 2(l-2):49- 59. Trent, L., and W. W. Hassler. 1966. Feeding behavior of adult striped bass, Roccus sax- atilis , in relation to stages of sexual maturity. Chesa- peake Sci. 7(4):189-192. C. Braxton Dew Lawler, Matusky, and Skelly Engineers One Blue Hill Plaza Pearl River, NY 10965 Present address; Northwest and Alaska Fisheries Center Kodiak Laboratory National Marine Fisheries Service, NOAA P.O. Box 1638 Kodiak, AK 99615 ACCUMULATION OF AGE PIGMENTS (LIPOFUSCIN) IN TWO COLD-WATER FISHES In fisheries management, age structured models are the preferred method for meeting the key ob- jectives of estimating optimal yields and deter- mining the effect of fishing on stock structure (Gulland 1978). However, fev^^ species of commer- cial marine fishes exist in which age can be deter- mined with certainty (Boehlert 1985). The con- centration of age pigments (lipofuscin) (Ettershank 1984) in fish tissues (Agius and Agbede 1984) may be a measure of fish age that could be used to validate other ageing techniques and might also improve estimates of age of long- lived species where other techniques are difficult to apply. Lipofuscin originates in biological membranes through lipid peroxidation (Tappel 1975). Lipo- fuscin accumulation has been documented for a wide variety of animals, from mammals to the bread-mold Neurospora (Ettershank et al. 1983 and references therein). The rate of accumulation has been shown to be constant during the lifetime of laboratory-raised mice (Reichel 1968; Miquel et al. 1978), dogs (Munnell and Getty 1968), fiesh- flies (Ettershank et al. 1983); man (Strehler et al. 1959); and also wild populations of mice (Dapson et al. 1980). On the other hand, the rate of lipofus- cin accumulation has been shown to vary with level of activity and lifespan (Sohal and Donato 1978). It is expected that the rate in which lipo- fuscin accumulates with age in a natural popula- tion of fishes, or stock, with its free genetic inter- change and likely common habitat, would be fairly uniform (but see Smith 1987). In addition, if measurements are made in nonmitotic and constantly metabolizing tissues such as brain or myocardium, the variation in concentra- tion due to environmental effects is least like- ly. In this paper we present the results of a prelim- inary study designed to assess the usefulness of extracted lipofuscin as a method of ageing fishes. Two species of cold-water fishes are in- cluded: the rainbow trout, Salmo gairdneri, reared in captivity, and of known age; and the Dover sole. Microstomas pacificus, a long-lived fish, in which age is not known with certain- ty. We present spectral characteristics of the ex- tracted lipofuscin from several tissues and the change in concentration of lipofuscin with fish age. FISHERY BULLETIN: VOL. 86, NO. 2 401 Materials and Methods Specimens of Dover sole were collected on De- cember 1986 ofTPoint Conception. Ages were esti- mated by J. Butler, E. Lynn, and M. Drawbridge of the Southwest Fisheries Center using otolith sections. Ages ranged from 2.7 to 44.7 years (av- erage of three readings). Specimens of the species Salmo gairdneri were collected in July 1986 at Hot Creek Hatchery. Ages ranged from 3 months to 3 years. Samples were frozen after collection and kept at -80°C until analysis. Studies by Nicol (1987) showed that this form of preservation yielded the lowest fluorescence when compared with ethanol and formalin-preserved samples. There was no indication that the level of fluorescence changed as a function of time of preservation or by inter- action with the extracted lipofuscin, as in the case for formalin. To excise the brain, the top of the skull was opened and the four brain lobes were removed as a unit, without the optic nerve. To excise the heart, we separated the muscle from connective tissue, blood vessels, and fat de- posits. Three methods were compared for maximum lipofuscin extraction. The first two methods (Tap- pel 1975; MacArthur and Sohal 1982) were specif- ically developed for lipofuscin extraction and employed chloroformrmethanol (2:1) as the ex- tractive solvent; they differ in the optimal volume-to-weight ratio (30:1 and 20:1, respec- tively), temperature of extraction, and number of times the chloroform phase is washed with water. The third method was developed for lipid extrac- tion in fishes (Bligh and Dyer 1959) and uses chlo- roform:methanol:water (1:2:0.8) as the extractive solvent. Subsamples (n = 3) of cerebellum of Stenella sp. collected at the eastern tropical Pacific and kept at -30°C were defrosted, dried with lint-free paper, weighed, and extracted fol- lowing the three methods as described originally. Tissues from Dover sole were extracted basi- cally following the MacArthur and Sohal (1982) technique, with two additional steps to ensure a complete washing out of flavoproteins and pho- tooxidation of retinol when present (Fletcher et al. 1973). Tissue was freeze dried prior to analy- sis. Ten mL of chloroform: methanol (2:1, v/v) was added to a homogenizer containing the sample for a final solvent:sample ratio of about 120:1 (vol/ dry weight). The sample was ground with a teflon pestle attached to an electric drill and later sub- merged in a 40°C water bath for 1 minute. A 2 mL subsample was taken from the homogenate. Three mL of deionized water was added, the sam- ple was shaken, and the emulsion centrifuged 10 minutes at 1,912 g and 0°C. After centrifugation, the hyperphase was decanted and a second rinse performed. After decanting the hyperphase a sec- ond time, 1 mL of the hypophase (chloroform con- taining lipofuscin) was sampled and transferred to a polypropylene tube. Three mL of chloroform were added to the sample (for a total of 4 mL) which was then exposed to UV irradiation (254 nm) for 1-2 minutes to photooxidize retinol. This last step was routinely performed for liver tissue. Samples were then transferred to glass tubes, sealed, and kept in the refrigerator until analysis. Sample fluorescence was measured at the emis- sion peak (430-440 nm) in a quartz cuvette with a Perkin Elmer Fluorescence Spectrophotometer^ Model MFP-44A. The sample was excited at the peak of fluorescence excitation (—360 nm). The intensity of the fluorescent emission (at 430 nm) was normalized to the intensity of the quinine solution standard (1 mg L"^ in IN sulphuric acid) and expressed in fluorescence units. Lipofuscin in Dover sole is expressed as 1) total lipofuscin content per organ and 2) weight- specific lipofuscin concentration, calculated by di- viding the total lipofuscin content by the dry weight of the entire organ. Lipofuscin ft-om rainbow trout tissues was ex- tracted following the Bligh and Dyer (1959) tech- nique, without modifications. Whole tissues were ground in water with a tissue homogenizer to give a final concentration of 100 mg of tissue (wet weight) in 0.8 mL of homogenate. A sample of 0.8 mL was taken and solvents were added to give a final ratio of 1:2:0.8 (chloroform:methanol:water) with a solvent to sample ratio of 20:1 (vol/w). The sample was then filtered through a 2.4 cm glass fiber filter (Whatman GF/C), the tissue re- extracted with 1 mL of chloroform and refiltered. The extract was then washed with 3 mL of water, shaken, and centrifuged for 10 minutes at 1,912 g and 0°C. One mL of the chloroform hypophase was subsampled and fluorescence estimated as described above. Lipofuscin in rainbow trout is expressed as 1) total lipofuscin content per organ and 2) weight-specific lipofuscin concentration (total lipofuscin content of the organ divided by its wet weight). iReference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 402 Results EfiBciency of Extraction Lipofuscin was successfully extracted and quantified from three different fish tissues: brain, heart, and liver. Wavelengths of fluorescence ex- citation and emission maxima of extracted lipo- fuscin in chloroform are presented in Figure 1. All maxima are within the range cited by Shel- dahl and Tappel (1973); fluorescence excitation maxima were between 340 and 370 nm and fluo- rescence emission maxima between 420 and 470 nm. Retinol (wavelengths of maximum fluorescence excitation at 325-340 nm and maximum fluores- cence emission at 475 nm) was photooxidized by exposing the chloroform extract to UV irradiation (254 nm). As expected, retinol was found in liver and sometimes it was present in brain and heart tissues. For example, wavelengths of fluorescence excitation and emission maxima in liver of Dover sole shifted from 352 to 365 nm and from 470 to 440 nm, respectively, after UV irradiation (see Figure IC). Mullin and Brooks (1988) also found that this UV irradiation is effective in oxidizing retinol although they did not find significant retinol interference in fish tissue. It seems liver tissues may require UV irradiation to oxidize retinol while brain and heart tissue should be checked for retinol presence before irradiation. The extractive efficiency of the three methods tested are compared in Table 1. All methods ex- tracted similar fluorescent compounds from the brain oi Stenella sp., as the fluorescence excita- tion and emission spectra were similar. The MacArthur and Sohal (1982) method extracted Table 1. — Comparison of the three methods of lipofuscin extrac- tion in brain tissue Fluorescence excitation at 360 nm and fluores- cence emission at 440 nm. Results are presented as fluorescence units per mg of wet tissue, where the fluorescence emission signal is normalized to the intensity of emission of a standard quinine sulfate solution (1 mg L Hn IN sulphuric acid). The three methods were significantly different from each other, P < 0.05, Newman- Keuls range test (Zar 1974). ANOVA: F ratio = 21.81, 2 df, P < 0.05. Fluorescence units Fluorescence maxima Average 10-3 SD 10-3 n Excitation (nm) Emission (nm) f^acArthur and Sohal (1982) Tappel (1975) Bligh and Dyer (1959) 4.30 2.83 1.98 0.39 0.37 0.45 3 3 3 362 362 360 445 440 445 _L _L _L 3E0 360 400 440 480 520 >- (/) Z LiJ UJ U z: UJ o UJ (r o Z) B 365 1 435 1 1 1 A 1 AZ) 1 1 320 360 400 440 480 520 C 365 430 450 1 u ^l r t(l) // \(3) N / \ r \ \ \ N / W/ \ \ \ S / A \ X X X \^ X 1 1 it 1 1 1 1 320 360 400 440 480 520 WAVELENGTH (nm) Figure 1. — Spectral characteristics of lipofuscin extracted from Dover sole, Microstomus pacificus, tissues (uncorrected spec- tra). In chloroform. (A) Brain: (1) Lipofuscin fluorescence exci- tation spectrum (emission wavelength 440 nm); (2) lipofuscin fluorescence emission spectrum (excitation wavelength 365 nm). (B) Heart: (1) Lipofuscin excitation spectrum (emission wavelength 430 nm); (2) lipofuscin fluorescence emission spec- trum (excitation wavelength 365 nm ). (C ) Liver: ( 1 ) Lipofuscin fluorescence excitation spectrum (emission wavelength 430 rmi); (2) lipofuscin emission spectrum (excitation wavelength 365 nm) before UV radiation; (3) Same as (2) after UV radiation (254 nm). significantly more fluorescent pigment, than did either the Tappel (1975) or the Bligh and Dyer (1959) method which extracted 66% and 46% of maximum extraction, respectively). Thus all three methods are useful for quantitative estima- tion of extractable lipofuscin in fish tissue but 403 comparisons of amount of concentrations between tissues or species cannot be made if different ex- traction methods are used. Lipofuscin Concentration The total concentration of extracted lipofus- cin in Dover sole brain tissue was positively correlated with fish length (Fig. 2A) {Y = -15.4 + 0.516X,r2 = 0.43;6 ^0,P<0.01, t - 5.1, df=34). Furthermore, total extracted lipofuscin content in the brain was positively cor- related with age in organisms estimated by annu- 5 300 >- cr Q LlI o CO UJ cc o 200 100 -A o o o - (7 ^ - o / CTTL o 5 o o - ^ o o O - o ~ Y = -l54+05 X 1 1 1 1 1 1 0 08 B o o 0 [ct - ^X <" X 006 - oo yU UJ ^ 004 Q < m 002 - ^ o 0 7 o CD O y / o %3 oo o - Y = -19 1 10"' +0.2 10"^ X 0 1 1 1 lar increments in otoliths to be from 2 to 15 years old (Fig. 3A) (Y = -2.5 + 1.0 X, r'^ = 0.75; 6^0, P < 0.01, t - 5.5, df = 10). On the other hand, in older Dover sole (estimated age from 20 to 45 years; n = 26) the total extracted lipofuscin in the brain did not increase with their estimated age. The concentration of lipofuscin in the brain (ex- tracted lipofuscin concentration per unit of g dry weight) followed a similar pattern (Fig. 3B). The concentration in the brain increased linearly with age for fish 2-15 years old (7 = 30.39 + 14.85 X, r2 = 0.79, b i=0, P < 0.01, t = 6.66, df = 10) but fishes older than 15 years did not show an in- crease in pigment concentration with age. The total extracted lipofuscin content extracted ft-om rainbow trout brains increased with age (Fig. 4A) (Y = 0.14 + 0.18 X, r^ = 0.37; 6^0, P < 0.01, t = 3.75, df = 23), as did the content of the heart (Fig. 4C) (Y = -0.54 + 0.82 X, r2 = 0.67; 6 ^ 0, P < 0.01, t = 7.36, df = 23), and liver (Fig. 4E) (Y = -2.86 + 5.81 X, r^ - 0.62; b i=0,P < 0.01, t = 7.54, df = 34) from 3 months to 3 years. The concentration of extracted lipo- fuscin per unit wet weight of heart tissue did not change with age (Fig. 4D) whereas that of brain and liver tissue decreased with age (Fig. 4B, F). < cc CD CC UJ CL UJ O 111 o (O UJ cc o Z) 10 200 400 LENGTH (mm) 600 oo o Figure 2. — Dover sole, Microstomas paciftcus, brain: (A) Extracted lijwfuscin per unit of dry weight as a function of fish length (mm) (n = 36); (B) brain dry weight (g) as a func- tion of fish length (mm). Fluorescence units: the intensity of fluorescence normalized to a standard solution of quinine sul- fate (1 mg L-i in IN sulphuric acid). Figure 3. — Lipofuscin in Dover sole. Microstomas pacificas, brain tissue as a function of age in = 36). (A) Total lipofuscin content per brain; (B) weight- specific (g dry weight) lipofuscin in brain. I a UJ >- cc Q to I- UJ to UJ cc o 300 200 100 o B o o° Yn., cr> Ld 2- o cr CL 0 100 1 80 o > k. n 60 CP :L 40 LU O LU CJ LD Ld a: o 20 15 - 10 5 - 0 SEABASS (5). 'OfHALIBUT 8 - 6 - 2 - 0 I 1 0 10 20 30 DAYS AFTER HATCHING Figure 7. — Lipofuscin fluorescence per larva vs. age since hatching for laboratory-raised larval grunion (top), white sea- bass (center), and California halibut (bottom). 200 160 120 80 40 0 100 75 50 25 0 GRUNION Y=-0.99X + 55 SEABASS . Y=-0.48X + 13 HALIBUT Y = -I.7X + 67 0 10 20 30 DAYS AFTER HATCHING Figure 8. — Lipofuscin fluorescence per unit dry weight vs. age since hatching for laboratory-reared larval grunion (top), white seabsiss (center), and California halibut (bottom). Linear fits Eire for comparative purposes only; relations for seabass and halibut look curvilinear. 414 and is common in culture of larval fishes. Such variability may be reduced in nature, where runts may be subject to intense predation. Lipofuscin analysis may be usefijl only when applied to postmitotic tissue, such as nervous tis- sue in mature fish, or to whole organisms whose mitotic growth has essentially ceased, such as adult copepods or insects (since the methods ap- pear usefiil in arthropods — Ettershank et al. 1983; Sohal and Donato 1978). In these organ- isms, the vagaries of growth are reduced, and the accumulation of lipofuscin during starvation or exercise might show that lipofuscin concentration is interpretable as a measure of physiological age, habitat quality, and net growth efficiency. Lipofuscin is known to be a polytypic sub- stance, probably variable in composition among different organisms. We have assumed that a con- stant proportion of the same substance is ex- tracted. This may not be true (Vernet et al. 1988), and considerable work remains to be done on the basic biochemistry of the component substance(s). Though the extraction and fluorometric measure- ment is tantalizingly simple, it may well be that the microscopical method used to quantify "ceroid bodies" is the best approach. Fluorescent tech- niques used in histochemical research (Brizzee and Jirge 1981), combined with automatic imag- ing procedures, might decrease the tedium of staining and visual microscopy. Acknowledgments We thank Refik Orhun and Steve Cadell for the larval white seabass and halibut, and Maria Ver- net and two referees for helpful comments. Re- search was supported by California Department of Fish and Game project C-921 and National Sci- ence Foundation grant OCE86-00742. Literature Cited Bentle, L a.. S Dutta, and J Metcoff. 1981. The sequential enzymatic determination of DNA and RNA. Anal. Biochem. 116:5-16. Brizzee, K R , and S K Jirge. 1981. Fluorescent microscope techniques for the visual- ization and histological quantification of autofluorescent lipofuscin bodies in brain tissues. In J. E. Johnson (edi- tor). Current trends in morphological techniques, Vol. III. CRC Press. CSALLANY, A. S , AND K L AYAZ 1976. Quantitative determination of organic solvent solu- ble lipofuscin pigments in tissue. Lipids 11:412-417. DoRSEY, T E , P W McDonald, and O A Roels 1977. A heated Biuret-Folin protein assay which gives equal absorbance with different proteins. Anal. Bio- chem. 78:156-164. DOWSON. J H 1982. The evaluation of autofluorescence emission spec- tra derived from neuronal lipopigment. J. Microsc. 128:261-262. Ettershank, G. 1984a. A new approach to the assessment of longevity in the Antarctic krill Euphausia superba. J. Crust. Biol. 4(Spec. No. l):295-305. 1984b. Methodology for age determination of Antarctic krill using the age pigment lipofuscin. Biomass Handb. No. 26. SCAR/SCOR/LABO/ACMRR. Ettershank, G., I MacDonnell, and R Croft. 1983. The accumulation of age pigment by the fleshfly Sarcophaga bullata Parker (Diptera: Sarcophagiidae). Aust. J. Zool. 31:131-138. Fletcher, B L . C J Dillard. and A L Tappel. 1973. Measurement of fluorescent lipid peroxidation products in biological systems and tissues. Anal. Bio- chem. 52:1-9. Frey. H W 1971. California's living marine resources and their uti- lization. Calif. Dep. Fish Game, Resour. Agency, 148 p. May, R C 1971. Effects of delayed initial feeding on larvae of the grunion Leuresthes tenuis (Ayres). Fish. Bull., U.S. 69:411-425. MiQUEL, J , J Org, K G. Bensch, and J E Johnson, Jr. 1977. Lipofuscin: fine-structural and biochemical studies. In W. A. Pryor (editor). Free radicals in biology, p. 133— 182. Acad. Press, N.Y. NICOL, S 1987. Some limitations on the use of the lipofuscin ageing technique. Mar. Biol. (Berl.) 93:609-614. Shimasaki, H., N Veta, and O. S. Privett 1980. Isolation and analysis of age-related fluorescent substances in rat testes. Lipids 15:236-241. Sohal, R. S , and H Donato. 1978. Effects of experimentally altered life spans on the accumulation of fluorescent age pigment in the housefly, Musca domestica. Exp. Gerontol. 13:335-341. Theilacker, G H , AND M F. McMaster. 1971. Mass culture of the rotifer Brachionus plicatilis and its evaluation as food for larval anchovies. Mar. Biol. (Berl.) 10:183-188. Thomas, J C 1968. Management of the white seabass (Cynoscion no- bilis ) in California waters. Calif Dep. Fish Game, Fish. Bull. 142, p. 1-34. Vernet, M , J R Hunter, and R. D. Vetter 1988. Accumulation of age pigments in two cold-water fishes. Fish. Bull., U.S. 86:401-407. M M. Mullin E. R. Brooks Institute of Marine Resources Scripps Institution of Oceanography University of California, San Diego La Jolla, CA 92093-0218 415 NOTICES NOAA Technical Reports NMFS published during last 6 months of 1987 55. Proximate composition, energy, fatty acid, sodium, and cholesterol content of fmfish, shellfish, and their products. By Judith Krzynowek and Jenny Mur- phy. July 1987, iii + 53 p., 2 tables. 56. Some aspects of the ecology of the leatherback turtle, Dermochelys coriacea, at Laguna Jalova, Costa Rica. By Harold F. Hirth and Larry H. Ogren. July 1987, iii + 14 p., 12 tables, 13 figures. 57. Food habits and dietary variability of pelagic nekton off Oregon and Washington, 1979-1984. By Richard D. Brodeur, Harriet V. Lorz, and William G. Pearcy. July 1987, iii + 32 p., 32 tables, 1 figure. 58. Stock assessment of the gulf menhaden, Brevoortia patronus, fishery. By Douglas S. Vaughan. September 1987, iii + 18 p., 19 figures, 14 tables. 59. Atlantic menhaden, Brevoortia tyrannus , purse seine fishery, 1972-84, with a brief discussion of age and size composition of the landings. By Joseph W. Smith, William R. Nicholson, Douglas S. Vaughan, Donnie L. Dudley, and Ethel A. Hall. September 1987, iii + 23 p., 3 figures, 12 tables, 14 appendix tables. 60. Gulf menhaden, Brevoortia patronus, purse seine fishery, 1974-85, with a brief discussion of age and size composition of the landings. By Joseph W. Smith, Eldon J. Levi, Douglas S. Vaughan, and Ethel A. Hall. December 1987, iii + 8 p., 1 figure, 8 ta- bles, 2 appendix tables. 61. Manual for starch gel electrophoresis: a method for the detection of genetic variation. By Paul B. Aebersold, Gary A. Winans, David J. Teel, George B. Milner, and Fred M. Utter. December 1987, iii + 19 p., 8 figures, 1 table, appendices. 62. Fishery publication index, 1980-85; Technical Memo- randum index, 1972-85. By Cynthia S. Martin, Shelley E. Arenas, Jacki A. Guffey, and Joni M. Packard. December 1987, iii + 149 p. Some NOAA publications are available by purchaise from the Superinten- dent of Documents, U.S. Government Printing Office, Washington, DC 20402. 416 ERRATA Fishery Bulletin: Vol. 86, NO. 1 Butler, John L., and Darlene Pickett, "Age-specific vulnerability of Pacific sardine, Sardinops sagax, larvae to predation by northern anchovy, Engraulis mor- dax," p. 163-167. Pages 165 and 166 have incorrect figures for the legends, please replace with the following: o z o z o a. Vi HI a t- z LLI o cc LJ Q. uu — J 80 - 1 ..- X 60 - ANC 1 HOVY ..' - V.-- '■'' .^ — J ' J 40 - 1 t , - ■ uf' - - -l "" '/ r^ ^^SARDINE 20 ~ < ■' « --^ K^ n 1 1 1 1 1 1 1 1 1 1 1 8 12 16 LENGTH (mm) 20 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). 100 O 80 z a O 60 z UJ o 40 20 ANCHOVY v.- SARDINE 8 12 16 LENGTH (mm) 24 Figure 3. — 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). 100 O z a z o a. (A HI o lij Q. 80 60 40 20 .1 ANCHOVY If SARDINE 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). 100 z UJ o 80 - z Q. < O 60 V) 40 - 20 - T - ANCHOVY , -'- - 1 V - ..' F .^ r' J L SARDINE _ •••"f*Y -T- .^ , — t <^ , t-^ ,1 F= 1 f^ 1 ] 1 1 10 20 AGE (days) 30 40 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). 303 j 2 .q 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), %, %c, 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 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 GPO 791-008 .<*<■': ">». BuUetin ^^AT£S O^ Marine Biological Laboratory LIBRARY JAN 2 0 1989 Vol. 86, No. 3 Woods Hole, Mass. -^ July 1988 BARLOW, JAY. Harbor porpoise, Phocoena phocoena, abundance estimation for Califor- nia, Oregon, and Washington: I. Ship surveys 417^ BARLOW, JAY, CHARLES W. OLIVER, TERRY D. JACKSON, and BARBARA L. TAYLOR. Harbor porpoise, Phocoena phocoena, abundance estimation for California, Oregon, and Washington: IL Aerial surveys 433-' SIGLER, MICHAEL R, and JEFFREY T FUJIOKA. Evaluation of variability in sable- fish, Anoplopoina fimbria, abundance indices in the Gulf of Alaska using the bootstrap method 445 BOEHLERT, GEORGE W, and TAKASHI SASAKI. Pelagic biogeography of the armor- head, Pseudopentaceros wheeleri, and recruitment to isolated seamounts in the North Pacific Ocean 453 SAVOY, THOMAS F, and VICTOR A. CRECCO. The timing and significance of density- dependent and density-independent mortality of American shad, Alosa sapidissima. . 467 HOSS, DONALD E., LINDA COSTON-CLEMENTS, DAVID S. PETERS, and PATRICK. A TESTER. Metabolic responses of spot, LeiosUymus xanthurus, and Atlantic croaker, Micropogonias undulatiis, larvae to cold temperatures encountered following recruitment to estuaries 483 WATANABE, YOSHIRO, JOHN L. BUTLER, and TSUKASA MORI. Growth of Pacific saury, CololaMs saira, in the northeastern Pacific Ocean 489 DAVIES, N. M., R. W GAULDIE, S. A. CRANE, and R. K THOMPSON. Otolith ultrastructure of smooth oreo, Pseudocyttits maculatus, and black oreo, Allocyttus sp., species 499 BECKER, D. SCOTT. Relationship between sediment character and sex segregation in English sole, Parophrys vetulus 517 JAMIESON, GLEN S., and ANTAN C. PHILLIPS. Occurrence of Cancer crab (C. magister and C. oregonensis) megalopae off the west coast of Vancouver Island, British Columbia 525 STONER, ALLAN W, and ROGER J. ZIMMERMAN. Food pathways associated with penaeid shrimps in a mangrove-fringed estuary 543^ WILSON, ELIZABETH A., ERIC N. POWELL, and SAMMY M. RAY The effect of the ectoparasitic pyramidellid snail, Boonea impressa, on the growth and health of oysters, Crassostrea virginica, under field conditions 553 BROUSSE AU, DIANE J., and JENNY A BAGLIVO. Life tables for two field populations of soft-shell clam, Mya arenaria, (Mollusca: Pelecypoda) from Long Island Sound . . . 567 (Contimied on back cover) Seattle, Washington U.S. DEPARTMENT OF COMMERCE C. William Verity, 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 Na 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 ntunbers 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 Tftma Commission Dr. George W. Boehlert National Marine Fisheries Service Dr. Bruce B. CoUette 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. Tim D. Smith National Marine Fisheries Service Mary S. Fukuyama, Managing Editor The Fishery BvMetin (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 Officci 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 of this periodical has been approved by the Director of the Office of Manage- ment and Budget Fishery Bulletin CONTENTS Vol. 86, No. 3 July 1988 BARLOW, JAY. Harbor porpoise, Phocoena phocoena, abundance estimation for Califor- nia, Oregon, and Washington: I. Ship surveys 417 BARLOW, JAY, CHARLES W OLIVER, TERRY D. JACKSON, and BARBARA L. TAYLOR. Harbor porpoise, Phocoena phocoena, abundance estimation for California, Oregon, and Washington: IL Aerial surveys 433 SIGLER, MICHAEL P., and JEFFREY T FUJIOKA. Evaluation of variability in sable- fish, Ancyplopoma fimbria, abundance indices in the Gulf of Alaska using the bootstrap method 445 BOEHLERT, GEORGE W, and TAKASHI SASAKI. Pelagic biogeography of the armor- head, Pseudopentaceros wheeleri, and recruitment to isolated seamounts in the North Pacific Ocean 453 SAVOY, THOMAS F, and VICTOR A. CRECCO. The timing and significance of density- dependent and density-independent mortality of American shad, Alosa sapidissima. . 467 HOSS, DONALD E., LINDA COSTON-CLEMENTS, DAVID S. PETERS, and PATRICIA A. TESTER. Metabolic responses of spot, Leiostomus xanthurus, and Atlantic croaker, Micr&pogonias undidatvs, larvae to cold temperatures encountered following recruitment to estuaries 483 WATANABE, YOSHIRO, JOHN L. BUTLER, and TSUKASA MORI. Growth of Pacific saury, Cololabis saira, in the northeastern Pacific Ocean 489 DAVIES, N. M., R. W GAULDIE, S. A. CRANE, and R. K. THOMPSON. Otolith ultrastructure of smooth oreo, Pseudocyttits maculatus, and black oreo, Allocyttus sp., species 499 BECKER, D. SCOTT. Relationship between sediment character and sex segregation in English sole, Parophrys vetulus 517 JAMIE SON, GLEN S., and ANTAN C. PHILLIPS. Occurrence of Cancer crab (C. magister and C. oreganensis) megalopae off the west coast of Vancouver Island, British Columbia 525 STONER, ALLAN W, and ROGER J. ZIMMERMAN. Food pathways associated with penaeid shrimps in a mangrove-fringed estuary 543 WILSON, ELIZABETH A., ERIC N. POWELL, and SAMMY M. RAY. The effect of the ectoparasitic pyramidellid snail, Boonea impressa, on the growth and health of oysters, Crassostrea virginica, under field conditions 553 BROUSSE AU, DIANE J., and JENNY A. BAGLIVO Life tables for two field populations of soft-shell clam, Mya arenaria, (MoUusca: Pelecypoda) from Long Island Sound . . . 567 {Contimied on next page) Seattle, Washington 1988 For sale by the Superintendent of Documents, U.S. Government Printing Offi DC 20402— Subscription price per year; $16.00 domestic and $20.00 foreign issue; $9.00 domestic and $11.25 foreign. Marine Biological Laboratory LIBRARY e, Washingti 'ost per si t^m 2 0 1989 li Woods Hole, Mass. j Contents— ConiiriMed FARLEY, C. AUSTIN, PETER H. WOLF, and RALPH A. ELSTON. A long-term study of "microcell" disease in oysters witii a description of a new genus, Mikrocytos (g. n.), and two new species, Mikrocytos mackini (sp. n.) and Mikroq^tos roughleyi (sp. n.) . 581 Notes WILLIAMS, AUSTIN B. Cojoined twin adult shrimp (Decapoda: Penaeidae) 595 GOULD, EDITH, DIANE RUSANOWSKY, and DONNA A. LUEDKE. Note on muscle glycogen as an indicator of spawning potential in the sea scallop, Placopecten magellanixrus 597 JAMIESON, GLEN S, and ELLEN K PIKITCH. Vertical distribution and mass mortality of prawns, Pandalvs platyceros, in Saanich Inlet, British Columbia 601 YANG, MEI-SUN. Morphological differences between two congeneric species of pleuro- nectid flatfishes: Arrowtooth flounder, Atheresthes stomias, and Kamchatka flounder, A. evermanni 608 McINTYRE, JOHN D, REGINALD R. REISENBICHLER, JOHN M. EMLEN, RICHARD L. WILMOT, and JAMES E. FINN. Predation of Karluk River sockeye salmon by coho salmon and char 611 The National Marine Fisheries Service (NMFS) does not approve, recommend or en- dorse any proprietary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS ap- proves, recommends or endorses any proprietary product or proprietary material mentioned herein, or which has as its purpose an intent to cause directly or indirect- ly the advertised product to be used or purchased because of this NMFS publication. HARBOR PORPOISE, PHOCOENA PHOCOENA, ABUNDANCE ESTIMATION FOR CALIFORNIA, OREGON, AND WASHINGTON: I. SHIP SURVEYS Jay Barlowi ABSTRACT The density and total population size of harbor porpoise along the coasts of California, Oregon, and Washington are estimated from ship surveys using line transect methods. Surveys were completed between September 1984 and May 1986 using teams of 3-5 observers. Data include 852 porpoise groups sighted during 6,590 km of transects. Sighting rates varied more due to effects of sea state than due to the presence of rain, fog, or sun glare. Experiments using additional observers indicate that approx- imately 22% of trackline groups were missed by a team of 5 observers. Harbor porpoise density is calculated from transects along the 18 m isobath and is extrapolated to other depth zones based on a model of porpoise abundance as a function of depth. Total population size is estimated as 45,713 (SE = 7,865) animals. Approximately 200-300 harbor porpoise are taken annually in central California set net fisheries (Dia- mond and Hanan^; Hanan, et al.^). Little is known about porpoise abundance in this area. Dohl et al.* estimated that 1,600-3,000 porpoise reside in cen- tral and northern California based on their aerial surveys of coastal cetaceans. However, because har- bor porpoise are frequently missed in aerial surveys (Kraus et al. 1983), this estimate is probably low. More information is needed on abundance, distribu- tion, and population structure to determine the significance of harbor porpoise mortality in set nets. Beginning in 1984, the National Marine Fisheries Service (NMFS) has conducted ship and aerial surveys of harbor porpoise abundance in California, Oregon, and Washington. This report presents results from four ship surveys. Results of the aerial surveys are presented by Barlow et al. (1988). 'Southwest Fisheries Center La JoUa Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. ^Diamond, S. L., and D. A. Hanan. 1986. An estimate of har- bor porpoise mortality in California set net fisheries: April 1, 1983 through May 31, 1984. Adm. Rep. SWR-86-15, 40 p. Available from National Marine Fisheries Service, Southwest Region, 300 S. Ferry Street, Terminal Island, CA 90731. ^Hanan, D. A., S. L. Diamond, and J. P. Scholl. 1986. An esti- mate of harbor porpoise mortality in California net fisheries April 1, 1984 through March 31, 1985. Adm. Rep. SWR-86-16, 38 p. Available from National Marine Fisheries Service, Southwest Region, 300 S. Ferry Street, Terminal Island, CA 90731. ^Dohl, T. P., R. C. Guess, M. L. Duman, R. C. Helm. 1983. Cetaceans of central and northern California, 1980-83: status, abundance, and distribution. Report prepared for U.S. Minerals Management Service, Contract #14-12-0001-29090, 284 p. Porpoise density is estimated from survey data using line transect models (Burnham et al. 1980). Total abundance is estimated by extrapolating from density observed along transect lines to the entire area inhabited. Abundance in offshore regions is based on a model of porpoise density as a function of water depth. In addition to abundance estimation, survey data are used to examine the effect of envi- ronmental conditions on sighting efficiency and the possibility of temporal changes in harbor porpoise distribution. METHODS Ship Survey Methods Surveys were conducted from two National Oceanic and Atmospheric Administration (NOAA) research vessels, the 52 m RV David Starr Jordan (Surveys 1, 3, and 4) and the 54 m RV McArthur (Survey 2)^. Both vessels were of similar design with viewing stations located on top of the pilothouse (viewing height was approximately 10 m above sea surface). Transect lines followed as close as possi- ble to the 18 m isobath (roughly 2-4 km from the coast), although the actual depth along the transect varied from approximately 15-45 m, depending on the presence of local navigational hazards. The areas Manuscript accepted May 1988. FISHERY BULLETIN: VOL. 86, NO. 3, 1988. ^Cruise reports available from the Southwest Fisheries Center, P.O. Box 271, La Jolla. CA 92038. 417 FISHERY BULLETIN: VOL. 86, NO. 3 surveyed are shown in Figure 1. Survey design va^ried among the four trips: Surveys 1 and 3 were in September 1984 and 1985, respectively. Both were designed to survey harbor porpoise density and abundance from Point Concep- tion, CA to Cape Flattery, WA. An attempt was made to survey the entire coastline on each of these cruises, but several sections of the coast were missed (Fig. 1) because of fog and heavy weather. Five observation positions were used on these two surveys. Survey 2 was from 24 January to 9 February 1985 and was primarily designed to examine seasonal changes in harbor porpoise distribution between Point Conception, CA and Cape Flattery, WA. Data from this cruise were not used for density or abun- dance estimation. 1 \ \ \ 1 n 49°N 48° - 47° 10 PORPOISE/KM I I II SURVEY 1 SURVEY 2 SURVEY 3 SEPTEMBER JANUARY/FEBRUARY SEPTEMBER 1984 1985 1985 SURVEY 4 APRIL/MAY 1986 - 46° - 45° - 44° 41° - 38° 37° 36° - 35° 34° - 33° 32° 123° 121° 119° 117°W Figure 1.— Relative sightings per kilometer based on 30-minute latitudinal strata. Lines parallel to the coast indicate areas that were surveyed. Histograms indicate relative numbers of harbor porpoise seen per kilometer of transect, with bars to the left indicating rela- tive numbers in calm seas (Beaufort 0, 1, and 2) and bars to the right indicating relative number in rough seas (Beaufort 3, 4, and 5). 418 BARLOW: SHIP SURVEYS OF HARBOR PORPOISE Survey ^ was from 24 April to 5 May 1986 and was designed to investigate factors which affect har- bor porpoise density estimation. The surveys con- centrated on several areas of high porpoise density in central California. The vessel's activities were coordinated with a helicopter to gather information on the avoidance of the ship by harbor porpoise. Ex- periments were also conducted on survey 4 to deter- mine whether an independent team of 3 observers would sight any porpoise that were missed by the primary team of 5 observers. Data from this survey were not used for density or abundance estimation. Typically, 8-10 observers were used on each survey, with a rested observer starting every half hour and rotating through 5 primary observation positions at half-hour intervals. The 5 positions con- sisted of port and starboard inboard observers, port and starboard outboard observers, and a recorder positioned amidship. The inboard observers searched with 7 power (7 x ) binoculars from straight ahead to 90° (survey 1) or to 45° (surveys 2, 3, and 4) on their respective sides of the vessel. On survey 1 the outboard observers searched with 25 x, pedestal-mounted binoculars. Although sightings could be made at great distances from the vessel using the 25 x binoculars, these distant sightings contributed little to the estimation of trackline den- sity, and use of 25 x binoculars was discontinued. On subsequent surveys, both the inboard and out- board observers used 7 x binoculars. The outboard observers searched from straight ahead to 90° on their respective sides of the vessel. The recorder searched in the immediate vicinity of the ship using unaided eyes and (intermittently) 7x binoculars. On survey 2, only 3 observation positions were used from Point Conception to Point Sur, CA and from Point Reyes, CA to Cape Flattery, WA. When effort was reduced to 3 observers, the inside obser- vation positions were eliminated. On survey 4, a second team of 3 observers was added to monitor the effectiveness of the principal team. This monitor team searched using unaided eyes and (intermittently) 7x binoculars from the pilothouse deck (viewing height approximately 7 m from sea surface). The principal team and the moni- tor team did not communicate sighting information, and independent records were kept. Data were noted by the recorder on data coding forms. Data on search effort included the beginning and ending times and positions for continuous legs of effort, the ship's heading and speed, personal identification codes for the observers, sea surface temperature, water depth, Beaufort sea state, sun position relative to the ship, and codes indicating the presence of rain or fog within 5 km. The ship position was determined from a Loran navigational system or by triangulation using coastal landmarks and dead reckoning. Ship speed was recorded direct- ly from the Omega/Loran system or was calculated based on time and distance traveled between suc- cessive position fixes along straight transect lines. Water depth was measured using a 38 kHz acoustic depth sounder. Data for sightings consisted of the above ele- ments, plus estimated group size, distance to shore, an estimate of the angle between the trackline of the ship and the group, and an estimate of the distance from the ship to the group. Group size refers to all the individuals associated with a sight- ing event. In most cases, groups were closely asso- ciated individuals that surfaced together (mean = 2.92, median = 2.0). In two cases (Point Arena and Monterey Bay, CA), groups consisted of 50-80 loose- ly associated individuals that were organized in subgroups of 4-10). Group size was estimated and recorded independently by each observer; the mean of these estimates was used in subsequent analyses. The angle from the trackline to the porpoise was estimated visually with the aid of a pelorus mounted in front of the observer stations, or, when 25 x binoculars were used, from a calibrated collar on the pedestal mount. On surveys 1 and 2, distances to harbor porpoise were estimated visually using the radar distance-to-shore as a reference, or, when 25 x binoculars were used, distances were estimated using calibrated reticles in the oculars. On surveys 3 and 4, distances were estimated using calibrated reticles in the oculars of 7 x binoculars. Data were also collected on the porpoises' direction of travel relative to the ship. The length of a transect was estimated as the product of ship speed and elapsed time. To stratify density estimates by sea state, rain, and fog, the effort record was divided into segments during which the sea state, rain, and fog codes did not change. In five areas, information was collected on varia- tion in harbor porpoise density with water depth. During survey 3, three sections of the coast were surveyed intensively (Fort Bragg to Cape Vizcaino, CA; Cape Blanco to Coquille Point, OR; and Cape Lookout to Tillamook Head, OR), with transect lines following the 18, 56, 92, and 185 m isobaths. On survey 4, the 18 and 46 m isobaths were surveyed in Monterey Bay, CA and in the vicinity of the Rus- sian River, CA. These data formed the basis of a model (below) to extrapolate porpoise density from 419 FISHERY BULLETIN: VOL. 86, NO. 3 the usual transect lines (along the 18 m isobath) to deeper waters. Helicopter Observations During survey 4, a Hughes 500-D helicopter was used to collect information on harbor porpoise behavior in response to the survey ship. The heli- copter flew approximately 10 km ahead of the vessel, and 3 observers in the helicopter looked for harbor porpoise. Once a group of harbor porpoise was sighted, the helicopter hovered at 200-300 m while observers made behavioral observations and periodically recorded the helicopter's position using an on-board Loran system. Fluorescein dye pack- ages were dropped in the water to allow the heli- copter to maintain its position when harbor porpoise were diving. Radio communication was maintained with personnel on the ship who also kept records of the helicopter position using radar distances and bearings based on returns from an X-band radar transponder in the helicopter. The ship changed course, when necessary, to ensure that it passed in close proximity to the porpoise that were being observed. Porpoise observers on the ship were not aware of the helicopter's activities and were not told of sightings made by the helicopter observers (al- though they were able to see dye patches in some cases). Behavioral observations from the helicopter included time spent at the surface, time spent div- ing, and direction of porpoise movement. Density Estimation Line transect methods were used to estimate the density of harbor porpoise from sightings. The assumptions of these methods are considered in detail in the discussion. The usual formula for estimating density (D) based on line transect surveys of small cetaceans is given by D = m ■ n- G 2 ■ L (1) where /(O) n G L = the probability density function for sightings evaluated at zero perpen- dicular distance, number of sightings of groups, average group size calculated as the total number of individuals in all groups divided by the number of groups (iN/n), and length of the transect. (Holt and Powers 1982; Hammond and Laake 1983; Holt in press). I did not use mean group size explicit- ly in abundance estimation, and density of harbor porpoise individuals, D, was estimated as D = /(O) ■ iRI2) where R = the number of individuals length of transect (XN/L). (2) seen per Equation (2) is functionally equivalent to Equation (1), but it simplifies variance estimation. Typically when using Equation (1), variances (and possibly covariances) must be estimated for/(0), G, and n. Using Equation (2), variances are needed only for /(O) and R, and covariance between mean group size and number of groups is handled implicitly. Sight- ing distributions appear to be independent of group size, G, (Results section), hence no adjustments were made to /(O) for group size bias. The parameter /(O) is, in effect, a measure of sight- ing efficiency and should not vary with porpoise abundance. Sighting efficiency is, however, likely to change with sighting conditions, such as Beau- fort sea state. Given these expectations and because relatively large sample sizes are needed to estimate /(O) accurately, values for/(0) were estimated for each survey by pooling all sightings within defined sea state categories. In order to estimate density on a finer scale, estimates of R were stratified by geographic region and multiplied by the pooled esti- mate of /(O). The sighting probability density function evalu- ated at zero distance, /(O), was determined empirically by fitting curves to the frequency distribution of sightings as a function of perpen- dicular distance from the trackline (Burnham and Anderson 1976). Differences in distributions of perpendicular distance were tested using the Kolmogorov-Smirnov 2-sample test. To avoid bias due to rounding error, angle and radial distance data were "smeared" (Butterworth 1982; Hammond and Laake 1983). Angles were smeared by adding a uniformly distributed random number between - 5° and -1-5° to angle estimates. Radial distances were smeared by adding a uniformly distributed random number between 0.2 and -(-0.2 times the estimated distance. These smearing levels were based on the degree of rounding that was apparent from the data (Barlow^). ^Barlow, J. 1987. Abundance estimation for harbor porpoise (Phocoena phocoena) based on ship surveys along the coasts of Cahfornia, Oregon, and Washington. Adm. Rep. LJ-87-05. 420 BARLOW: SHIP SURVEYS OF HARBOR PORPOISE Several models were investigated for estimating /(O) from sighting distributions. The FORTRAN pro- gram Transect (Laake et al. 1979) was used to fit 2-, 3-, 4-, and 5-parameter Fourier series and 2-parameter exponential power series models. The FORTRAN programs Hazard and Hermite (S. Buck- land'') were used to fit the 2-parameter hazard rate model (constrained such that parameter P > 2, Buckland 1985) and the 1-, 2-, 3-, and 4-parameter Hermite polynomial model (Buckland 1985). Of these models, the 2-parameter hazard rate model was selected based on its ability to fit the observed distributions and its lack of dependence on group- ing criteria (Buckland 1985). Perpendicular distances were grouped into strata, the size of which increased with perpendicular distance: 0-25 m, 25-50 m, 50-100 m, 100-200 m, 200-400 m, 400-800 m, 800-1,600 m, and 1,600- 3,200 m. Several alternative groupings were inves- tigated, and the choice of outpoints made very little difference in estimates of /(O). The above strata (increasing with distance) gave lower variances in /(O) than when each stratum was of equal size (possibly because the hazard rate model assumes a distinct shoulder in the sighting distribution, and that shoulder is lost if the first distance strata are large). No established criteria exist for choosing an appro- priate perpendicular distance at which to truncate sighting distributions. Burnham et al. (1980) recom- mend that no more than 1-3% of sightings be eliminated by truncation. Using this recommenda- tion, models were not able to adequately fit the observed sighting distributions. In this report, trun- cation distance was chosen in four ad hoc steps: 1) The hazard rate model was fit to perpendicular distance data truncated at distances of 400, 800, 1,600, and 3,200 m. 2) Truncation distances were identified which gave acceptable x^ values (P > 0.1). 3) Of the acceptable truncation distances, the standard error in/(0) was estimated empirically by randomly drawing 10 samples (of the same size as the original sample) from the observed distribution of perpendicular distances and by calculating the standard deviation of/(0) estimated from each ran- dom sample. 4) Truncation distances were chosen as those which gave the lowest coefficient of varia- tion in/(0). Variance in R, the number of porpoise seen per kilometer, was estimated using jackknife statistics (Efron 1982). Jackknife estimates were calculated by first estimating the value of i? using all data. The value, Rj^, was again estimated excluding the A;th segment of search effort. This process was repeated for each effort segment. To ensure that each kth seg- ment was of equivalent length, effort segments with the same sea state, rain, and fog codes were com- bined in a linear array and were then divided into 10 segments of approximately equal length. The variance in the estimate of R was calculated as 10 s2 = ~ • 1 (i?, - Rf 10 ^=1 (3) Avail, from Southwest Fisheries Center, P.O. Box 271, La Joila, CA 92038. ^S. Buckland, Inter-American Tropical Tuna Commission, P.O. Box 271, La Jolla, CA 92038, pers. commun. July 1986. The variance of D was estimated using the Good- man (1960) product variance formula (assuming no covariance) using this jackknife variance for R and the above Monte Carlo variance for/(0). Fraction of Missed Animals On survey 4, a second, independent team of 3 observers were used to estimate the fraction of harbor porpoise that are missed by the primary team of 5 observers. The fraction of missed animals in a sighting survey is analogous to the fraction of unmarked animals in a mark/recapture experi- ment (Pollock and Kendall 1987). This fraction was estimated using the Chapman (1951) modifica- tion of the Petersen (or Lincoln) index method (Pollock and Kendall 1987). Confidence limits were estimated using Adams' (1951) method, which assumes a binomial sampling distribution. Standard error was estimated using standard binomial formulas. Abundance Estimation A model was used to estimate the number of har- bor porpoise along the entire coastline based on the density that was observed along the 18 m isobath. In shallow areas, such as the Bering Sea and Georges Bank, harbor porpoise are found a con- siderable distance from land (Gaskin 1984), hence offshore distribution is better modelled as a func- tion of depth than as a function of distance from shore. (Although harbor porpoise are also found in very deep water in fjords and inland waterways of Alaska [Taylor and Dawson 1984], this represents a special case that is not applicable to coastal waters considered here.) The model used to estimate abun- dance was based on data collected on surveys 3 and 4 and on data from a ship surveys by La Barr and 421 FISHERY BULLETIN: VOL. 86, NO. 3 Ainley^ and Szczepaniak^ in central California. The number of harbor porpoise seen per kilometer of transect was taken as an index of relative density along each isobath. A simple descriptive model was then constructed to give relative density as a func- tion of water depth. Fifteen depth strata were used in abundance estimation: 0-10, 10-20, 20-30, . . . , and 140-150 m. The surface area within the strata was calculated from digitized bathymetric data. Kelp beds were assumed to be unsuitable as harbor porpoise habitat; hence, kelp bed area was subtracted from the total area within the 0-10 m stratum. Kelp bed areas for the entire west coast were taken from Crandall (1915). More recent estimates for limited areas in central California are in good agreement with these previous values (G. Van Blaricom^"). For each of 15 depth strata, the abundance of har- bor porpoise was estimated as the product of their density along the survey line (the 18 m isobath), the density in that depth strata relative to that along the survey line, the surface area included within that depth strata, and the inverse of the estimated frac- tion of trackline animals that were seen. Since survey effort and harbor porpoise density both varied geographically, abundance estimates were made for each of 8 geographic regions (Fig. 2). Areas within the depth strata were estimated from NOAA bathymetric data. The estimate of total abun- dance along the coast, A/'y, is therefore given by 15 ^- = ^ I «^ .?. <^' ■ ^^.'> (4) where Dj = density of individuals observed on the transect line in the jth geographic strata, 4 = ratio of density in depth strata k to that on transect line (see Figure 4), Aji^ = area in geographic region j' and depth strata k, and F = the estimated fraction of trackline animals seen by the usual team of 5 observers. 'LaBarr, M. S., and D. G. Ainley. 1985. Depth distribution of harbor porpoise off central California: A report of cruises in April and May-June 1985. Report to U.S. National Marine Fisheries Service, Northwest and Alaska Fishery Center, 7600 Sand Point Way N.E., Seattle, WA. Contract No. 41-USC252. 'Szczepaniak, I. D. 1987. Abundance and distribution of har- bor porpoise (Phocoena phocoena) in the Gulf of the Farallones National Marine Sanctuary. Contract report prepared for National Park Service, Point Reyes National Seashore, Point Reyes, CA 94956. '"G. Van Blaricom, U.S. Fish and Wildlife Service, University of California, Santa Cruz, CA 93106, pers. commun. August 1986. Equation (4) was applied independently to the dif- ferent surveys and, within surveys, to different sea state strata. When combining estimates from differ- ent sea states or different cruises, abundance was calculated as the mean of the densities in each of the stratum, weighted by the length of the transect line within that stratum. In estimating standard error for total abundance, variances of products were calculated using the Goodman (1960) product variance formula, and vari- ances of ratios were estimated using a Taylor approximation (Yates 1953, p. 198). Area was assumed to be known without error. Statistical error in the indices of abundance for the depth strata could not be estimated given the paucity of available in- formation. To account for uncertainty in the model of depth distribution, three versions of the model are proposed to span a range of possibilities. RESULTS On the four surveys, 852 groups of harbor por- poise were sighted (an estimated 1,818 individuals). A distance of 6,590 km was surveyed during 56 days. The number of sightings per kilometer sur- veyed varied geographically and these geographic patterns appeared to change appreciably between cruises (Fig. 1). Sighting Distributions The number of sightings on the inshore and off- shore sides of the vessels were approximately equiv- alent (383 and 392, respectively). The cumulative distributions of perpendicular sighting distances were not significantly different for these two sides (P = 0.06). Therefore, sighting distributions were assumed to be symmetrically distributed about the trackline, and the distributions of perpendicular sighting distances from both sides of the vessel were pooled for subsequent analyses. The distributions of perpendicular sighting dis- tances for the first three surveys were significant- ly different from one another (P < 0.01 for all). This was probably the result of the modifications in survey methods between these cruises. Surveys 3 and 4 used the same methods, and sighting distribu- tions were not significantly different (P = 0.39). Given that changes in methods result in differences in sighting distributions, all surveys were treated separately in subsequent analyses. Distributions of perpendicular distance were not significantly different between individuals sighted 422 BARLOW: SHIP SURVEYS OF HARBOR PORPOISE 48° 46° 45» 44* 43" 420 . 41* - 40° 39° 38* 37° 36° 35° 34° - 33° - 32°N 132°W 131° 130° 129° 128" 127" 126" 125° 124° 123° 122° 121" 120° 119° 118° 117»W Figure 2.— Geographic regions used as strata in abundance estimation. Broken line indicates the 200 m isobath and delineates likely harbor porpoise habitat. 423 FISHERY BULLETIN: VOL. 86, NO. 3 alone (group size of 1) and larger groups (group size of 3 and greater) (P = 0.56). Environmental Conditions Affecting Sightings Sighting efficiency was not significantly affected by rain, fog, or sun glare. Rain/fog conditions were considered "poor" if rain or fog were present within 5 km of the vessel and "good" if neither were pres- ent. The distributions of perpendicular sighting distances were not significantly different between these two strata (P = 0.32, 0.44, 0.78, and 0.64, respectively, for surveys 1, 2, 3, and 4), and the number of porpoise per kilometer surveyed was higher in the "poor" category for two of the surveys. Sun glare from the water's surface was considered to contribute to "poor" sighting conditions if the sun was within 45° of the trackline in front of the ship. Conditions were considered "good" when the sun was in other positions or was obscured by clouds. As with rain/fog conditions, the distributions of perpendicular sighting distances were not signifi- cantly different between these "good" and "poor" sun glare categories (P = 0.87, 0.47, 0.30, and 0.55, respectively, for surveys 1, 2, 3, and 4). The number of harbor porpoise per kilometer surveyed were slightly higher in the poor category for three of the surveys. In paired comparisons when glare was pres- ent on only one side of the bow, approximately equal numbers of sightings were made on the sides with and without glare (60 vs. 59, respectively). All categories of rain, fog, and glare are included in subsequent analyses. Sea state did have a significant effect on porpoise sightings. Sea state was categorized as calm (without white-caps, Beaufort sea states 0, 1, and 2) or rough (with white-caps, Beaufort sea states 3, 4, and 5) following the classification used by Holt and Cologne (1987). Distributions of perpendicular distances were not significantly different between these categories for any of the surveys (P > 0.05); however for all surveys combined, the number of harbor porpoise detected per kilometer was much lower during rough seas (0.32 km"^) than during calm seas (1.22 km"^). There were insufficient sightings to estimate density for rough seas separ- ately; therefore, rough sea data were excluded in subsequent analyses. For all three surveys, the numbers of harbor porpoise detected per kilometer was higher at Beaufort 0 & 1 than at Beaufort 2, and for survey 3, the distributions of perpendicular sighting distance were significantly different be- tween these categories {P = 0.03). Porpoise density is, therefore, estimated separately for Beaufort 0 & 1 and for Beaufort 2 conditions. (For comparison, harbor porpoise abundance was also estimated pool- ing Beaufort sea states 0, 1, and 2. Estimated abun- dance was approximately the same by both methods, but the variance was slightly lower using the stratified sea state categories. For this reason, only the stratified estimates are presented here.) Helicopter Observations Helicopter observation of the behavior of harbor porpoise in response to the survey ship were made on only 6 groups of animals. Plots of vessel tracks and movements of the groups are given in the cruise report (see footnote 5). Only in one case was a distinct behavioral change noted in response to the ship. In that case, when the vessel was within 800 m, the group moved rapidly, perpendicular to the path of the vessel and then parallel to and in the op- posite direction of the vessel. Observers on the ship saw this harbor porpoise group as they moved rapid- ly out of the path of the vessel. Observers on the ship also saw 2 of the other 5 groups. Although this sample of behavior is small, movement in response to the survey vessel appeared limited to within 1 km of the vessel and, when it occurred, animals did not travel far from their original positions. Porpoise Density The probability density distributions of perpen- dicular sighting distances are shown in Figure 3 for surveys 1 and 3 and for Beaufort sea states 0 & 1 and 2. The hazard rate model gave acceptable fits for all sighting distributions {P > 0.1) when the trun- cation criteria was set at 400 m (Table 1). For survey 1, the optimum truncation points were chosen as 400 m for Beaufort 0 & 1 and 800 m for Beaufort 2; for survey 3, this distance was 400 m for both Beaufort sea state categories. The fits of these models are shown in Figure 3. Estimates of density and stand- ard errors are given in Table 2. Depth Distribution Model The model of harbor porpoise depth distribution was based on the relative densities of harbor por- poise at different water depths. Ship survey data were pooled into five depth ranges: 18-37 m (10-20 fathoms), 37-55 m (20-30 fathoms), 55-73 m (30-40 fathoms), 73-91 m (40-50 fathoms), and 91-110 m (50-60 fathoms). Ship surveys are generally not practical inshore of the 18 m isobath, but estimates 424 BARLOW: SHIP SURVEYS OF HARBOR PORPOISE 10 0.0 0.1 0.2 PERPENDICULAR DISTANCE (km) 8 \ SURVEY 1, BEAUFORT 2 6 - \ V, 4 - \ \ 2 - ^ 0 - ■ 0.4 00 10 0.4 0.0 0.2 0.4 0.6 0.8 SURVEY 3, BEAUFORT 2 -I r 0.1 0.2 1 r- 0.3 0.4 PERPENDICULAR DISTANCE (km) Figure 3.— Probability density distributions for perpendicular sighting distances. Histograms indicate observed distributions, and solid lines indicate the best fit of the hazard rate model to these data. Table 1 .—Estimated values of the probability density functions evaluated at zero perpendicular distance, ^(0). Estimates are based on the hazard rate model and were made for truncation distances of 400, 800, 1,600, and 3,200 m. Estimates are given only if the model gave an acceptable fit to the data (P > 0.1). Asterisks in- dicate f(0) values with the lowest coefficient of variation (paren- theses). Survey Beaufort sea state Truncation distance 400 m 800 m 1 ,600 m 3,200 m 1 1 3 3 0 & 1 2 0 & 1 2 7.85 • (0.23) 10.48 (0.59) 4.51 • (0.22) 6.97 • (0.19) 5.31 (0.29) 8.15 * (0.21) 3.10 (0.66) 4.31 (0.24) 7.09 (0.33) 2.69 (0.31) 5.78 (0.51) from aerial surveys (Barlow et al. 1988) show roughly equal density at 0.61 and 1.85 km from the shore (the latter corresponding approximately to the 18 m isobath). Relative density from ship surveys was measured in the number of sightings per kilo- meter of searching effort. Relative densities at Table 2.— Density estimates, D, for harbor porpoise (km"^) along the 18 m isobath in each of eight geographic strata. Density was calculated per Equation (2), using estimates of f(0) (Table 1) which had the lowest coefficients of variation. Values are not adjusted for missed animals. Standard errors are in parentheses. Geo- graphic region 1 2 3 4 5 6 7 8 Total Survey 1 Survey 3 Beaufort 0 & 1 Beaufort 2 Beaufort 0 & 1 Beaufort 2 0.0 (0.0) 1.1 (1.0) 1.4 (0.8) 7.9 (3.4) 1.8 (0.8) 2.5 (1.1) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 2.5 (1.0) 2.1 (1.3) 2.0 (0.6) 0.8 (0.6) 0.0 (0.0) 1.3 (0.6) 0.0 (0.0) 0.6 (0.5) 0.1 (0.1) 0.0 (0.1) 0.8 (0.6) 2.5 (0.8) 2.5 (0.6) 0.8 (0.4) 0.1 (0.1) 0.1 (0.1) 0.0 (0.0) 6.7 (7.2) 0.6 (0.3) 2.6 (0.9) 1.3 (0.7) 1.2 (0.5) Pooled estimates 0.04 (0.02) 0.51 (0.30) 0.03 (0.03) 2.83 (1.69) 0.91 (0.32) 2.22 (0.40) 2.64 (0.78) 1.09 (0.45) 1.33 (0.30) 425 FISHERY BULLETIN: VOL. 86. NO. 3 18-37 m show no consistent relationship to those at 37-55 m or 55-73 m (Table 3), but on average these appear to be approximately equal. Relative densities at 18-37 m are, however, consistently higher than densities at 73-110 m in all areas (Table 3). A total of 236 km were searched in waters deeper than 110 m and no harbor porpoise were seen. Despite high variability in patterns of depth distribution and lack of ship coverage in shallow waters, some generalizations can be made about the depth distribution of harbor porpoise along the west coast. The relative abundance of harbor porpoise ap- pears to be roughly constant from shore to 55 m, to be markedly lower at 73-110 m, and to be very low in waters deeper than 110 m. Based on the above relationships, I propose the following preliminary model for the depth distribu- tion of harbor porpoise along the coasts of Califor- nia, Oregon, and Washington: constant abundance from the coast to the 80 m isobath, linearly decreas- ing abundance from the 80-120 m isobaths, and zero abundance in waters deeper than 120 m (Fig. 4a). Because considerable uncertainty exists in this model, I propose two alternative models (Fig. 4b, c). Alternative models b and c are less likely than the primary model given because both conflict with some of the available data. The alternative models do, however, encompass the likely range of relative density values and provide a means to evaluate the sensitivity of the abundance estimate to different models of depth distribution. Fraction of Missed Animals The experiment on survey 4 indicates that some trackline groups were seen by 1 group of observers and were missed by the other. A total of 103 sight- ings was made by both teams, 33 of which were estimated to be within 100 m perpendicular distance from the transect line. Of the 103 total sightings, 85 were detected only by the 5 principal observers, 6 were detected only by the 3 monitor observers, and 12 were detected by both teams. Of the 33 trackline sightings, 20 were detected only by the principal observers, 3 were detected only by the monitor observers, and 10 were detected by both teams. The Petersen estimate of the fraction of trackline porpoise seen by the primary team of 5 observers is thus 0.780 (SE = 0.117, 95% C.L. = 0.45-0.95). This indicates that approximately 22% of trackline sightings are missed by the principal teams of 5 observers. Porpoise Abundance Estimates of porpoise abundance in each of the eight geographic strata are given in Table 4 for the primary model of offshore distribution. Independent estimates are given for survey 1 and for survey 3 Table 3. — Relative harbor porpoise abundance observed within the specified depth ranges at a variety of study sites. Relative abundance is measured as number of porpoise sightings made per kilometer. Numbers in parentheses indicate kilometers surveyed. Depth range Location 18-37 m 37-55 m 55-73 m 73-91 m 91-110 m Central California' 0.02 (89) 0.08 (172) 0.03 (403) 0.03 (279) 0.01 (166) Gulf of the Farallones, CA^ 0.29 (181) 0.00 (159) 0.30 (133) 0.00 (7) 0.00 (7) Fort Bragg, CA 0.05 (41) — 0.05 (43) — 0.03 (35) Coquille Pt, OR 0.24 (43) — 0.35 (52) — 0.08 (50) Tillamook Head, OR 0.50 (57) — 0.12 (52) — 0.00 (17) Monterey Bay, CA 0.46 (220) 0.29 (76) — — — Russian River, CA 0.00 (26) 0.30 (33) — — — 'Data taken from LaBarr and Ainley (see text footnote 8) assuming an average survey speed of 9 knots. ^Data taken from Szczepaniak (see text footnote 9) assuming an average survey speed of 9 knots. 426 BARLOW: SHIP SURVEYS OF HARBOR PORPOISE DEPTH DISTRIBUTION MODELS 2 18 16 14 12 1 0.8 0.6 0 4 0.2 0 2 1.8 16 1.4 1.2 1 0.8 0 6 0.4 0.2 0 Primary Model 5 15 25 35 45 55 65 75 85 95 105 115 125 135 145 LU > LU CC CO z lU Q LJJ > I- < _J UJ CC CO z lU o LU > LU CC 5 15 25 35 45 55 65 75 85 95 105 115 125 135 145 WATER DEPTH (meters) Figure 4.— Proposed models for the depth distribution of harbor porpoise expressed as relative den- sities, 4, within 10 m depth intervals. Density at 20-30 m is assumed to be knowm and is given a relative value of 1. Figures represent a) a primary model of offshore distribution, b) a high estimate of offshore range, and c) a low estimate. - — High Model — ' F " ' 1 5 15 25 35 45 55 65 75 85 95 105 115 125 135 145 in each area. Both surveys show similar patterns, with higher abundances in the northern strata (4-8) and very low abundance in strata 1 and 3. Despite similar patterns, differences between the paired estimates are in some cases, large and statistically significant (^tests, P < 0.05). Because region 8 was not covered on the third survey, it is not possible to compare estimates of total abundance for the en- tire coast between surveys. The total abundances for regions 1-7 (Point Conception to the Columbia River) are 46,550 (SE = 10,932) animals and 32,029 (SE = 10,906) animals for surveys 1 and 3, respec- tively. The difference between these estimates is not statistically significant (^test, P > 0.05). Pooling the results of the two surveys, the estimate of harbor porpoise abundance between Point Conception and Cape Flattery in September of 1984 and 1985 is 45,713 (SE = 7,865) animals (Table 4). The same estimate using the alternate models of offshore distribution ranges from 28,769 to 78,019 (Table 5). 427 FISHERY BULLETIN: VOL. 86, NO, 3 Table 4.— Estimated abundance of harbor porpoise in each of the eight geographic strata based on the primary model of offshore distribution. Estimates for Beaufort 0 & 1 and for Beaufort 2 were computed separately and then averaged, weighting by transect length. Pooled estimates for the eight strata were obtained as an average of the two surveys, weighting by transect length. All esti- mates are adjusted for missed animals. Standard errors are in parentheses. Geographic region Survey 1 Survey 3 Pooled estimates 1 2 3 4 5 6 7 8 Totals Regions 1-3 Regions 1-7 Regions 1-8 2,401 (2,180) 0 6,909 (2,959) 11,245 (6,943) 9,061 (2,724) 16,934 (7,097) 9,808 (4,311) 2,401 (2,180) 46,550 (10,932) 56,358 (11,751) 126 (68) 932 (646) 153 (158) 9,096 (9,855) 3,296 (1,410) 12,786 (3,676) 5,641 (2,424) 1,210 (669) 32,029 (10,906) 96 (52) 1,459 (885) 112 (116) 7,909 (4,784) 4,806 (1,745) 11,107 (2,363) 10,416 (3,311) 9,808 (4,311) 1,667 (895) 35,904 (6,578) 45,713 (7,865) DISCUSSION Distribution Harbor porpoise are not uniformly distributed between Cape Flattery and Point Conception. Al- though there are no obvious discontinuities within this range, density varies geographically and tem- porally. The most dramatic temporal changes are between the two September surveys and the Janu- ary-February survey (Fig. 1). The coasts of Wash- ington and northern Oregon were found to have relatively high densities of harbor porpoise in Sep- tember, but, despite excellent sighting conditions, very few porpoise were seen there in January. High densities of harbor porpoise were also seen in Monterey Bay on both September cruises and on survey 4 in May. This area was intensively surveyed in February, and few harbor porpoise were seen. As can be seen in Figure 1, adjacent areas tended to have similar densities within a survey. Less consis- tency is found when the same areas are compared between different surveys. Table 5.— Estimated abundance of harbor porpoise in central California (regions 1-3) and along the entire coast (regions 1-8) based on two alternate models of offshore distribution. All estimates are adjusted for missed animals. Standard errors are in parentheses. Survey 1 Survey 3 Pooled estimates Alternate Model b Regions 1-3 Regions 1-8 Alternate Model c Regions 1-3 Regions 1-8 3,966 (3,602) 95,132 (19,515) 1,505 (1,367) 35,736 (7,550) 1,986 (1,104) 770 (421) 2,744 (1,478) 78,019 (13,356) 1,054 (561) 28,769 (4,995) The apparent changes in distribution could be caused by small changes in depth distributions. The majority of survey effort was along the 18 m iso- bath. A large fraction of animals could be missed if their depth distribution changed by 10 m or less. More information on depth distributions is needed before the apparent temporal changes in geographic distribution can be interpreted. Porpoise Density Estimates of harbor porpoise density ranged from 0.03 to 2.8 animals/km^ along transect lines in the eight geographic regions (pooled estimates, Table 2). In another study, Szczepaniak (fn. 9) estimated 0-1.9 porpoise/km^ in four study areas in the Gulf of the Farallones, CA. Taylor and Dawson (1984) found 1.2-5.9 porpoise/km- at study sites in Glacier Bay, AK. Flaherty and Stark^^ estimated 0.8-1.6 porpoise/km^ in Washington Sound. Densities in the present study are therefore within the range of densities found in other areas along the same coast. Harbor porpoise density was estimated for California, Oregon, and Washington based on aerial surveys that were concurrent with the present study (Barlow et al. 1988). The overall estimate of harbor porpoise density from that study (corrected for missed animals) was 1.79 porpoise/km^. The overall estimate from the ship survey (1.33 porpoise/km^) can be corrected for missed animals to yield an estimate of 1.73 porpoise/km^. Given that the coef- I'Flaherty, C, and S. Stark. 1982. Harbor porpoise {Phocoena phocoena) assessment in "Washington Sound". Final Report #80-ABA-3584 submitted to National Marine Mammal Laboratory, National Marine Fisheries Service, NOAA, 7600 Sand Point Way, NE, Seattle, WA 98115. 84 p. 428 BARLOW: SHIP SURVEYS OF HARBOR PORPOISE ficient of variation in the pooled ship estimates is nearly 25%, these estimates are in very close agree- ment. However, because the aerial estimates are based only on the small fraction of the coastline that was surveyed under optimal conditions, the ship estimates are probably a better representation of porpoise density for the entire coast. Of the areas surveyed, harbor porpoise density is highest in northern California and Oregon. The highest density was seen in northern Oregon (region 7) during survey 1. The second highest density was observed in northern California between Bodega Head and Cape Mendocino (region 4) on survey 3. Two areas in central California (regions 1 and 3) were found to have very low densities. Region 1 in- cludes the Big Sur coastline from Point Conception to Point Sur. This area is characterized by steep depth gradients and hence has little habitat that is suitable for harbor porpoise. Region 1 was relatively well covered, with 378 km of trackline surveyed at Beaufort sea states 0-2. In contrast, region 3 in- cludes the Gulf of the Farallons with its broad coastal shelf within the 100 m isobath. Based on surveys of 764 km, Szczepaniak (fn. 9) estimated 1,033 harbor porpoise are found in the Gulf of the Farallones alone. This is much greater than my estimate of 112 animals in region 3 based on only 175 km of survey effort. Because of his greater amount of search effort in this area, I believe that Szczepaniak' s estimates for region 3 are more ac- curate than mine. Although regions 1 and 3 were both identified as low density areas, more confidence can be placed on this conclusion for region 1 than for region 3. Abundance The size and behavioral characteristics of harbor porpoise make estimating their abundance difficult. Harbor porpoise are small, occur in groups of only a few individuals, and surface without conspicuous splashes; their distribution is extremely patchy. Even v^th 5 observers, the effective path width that can be searched from a ship is <1 km, and that path width decreases very rapidly in rougher sea states. All of these factors contribute to high variability in the abundance estimates presented here. Seasonal and year-to-year changes in the distribution of har- bor porpoise may also contribute to the variability seen within geographic strata. These are, however, the best (and, for some regions, the only) estimates of harbor porpoise abundance for the study area. Although there are no prior estimates for Oregon or Washington coasts, Dohl et al. (fn. 4) estimated harbor porpoise abundance in central and northern California. Their estimates range from 3,000 har- bor porpoise in autumn to 1,600 in summer, which correspond (approximately) to the pooled estimate of 11,457 for regions 1-4 based on the present study. There are, however, several problems with the ap- plication of their methods to the estimation of har- bor porpoise abundance. In a direct comparison with shore counts, Kraus et al. (1983) showed that observers on aircraft saw only 10-20% of harbor porpoise groups. Dohl et al. (fn. 4) did not apply a correction to account for harbor porpoise groups that are submerged at the time the aircraft passed. Also, Dohl et al. did not stratify estimates by distance from shore or depth. Although most of their harbor porpoise sightings were within 0.5 km (0.25 nmi) of shore, their density estimates were extrap- olated to an area extending 166 km from the coast. Estimates from the current study are based on better methodology than previous estimates. In addition to exposed coastal habitats, harbor porpoise are also found in bays along the coasts of California, Oregon, and Washington. Goetz (1983) reported that harbor porpoise are found throughout the year in Humboldt Bay, CA. Harbor porpoise have been seen in San Francisco Bay, but are described as rarely present^^. Abundance of harbor porpoise in inland waters may, however, vary seasonally (Taylor and Dawson 1984). No estimates exist for the total number of harbor porpoise in- habiting bays. Survey effort in the present study was limited to exposed coastal areas (including Monterey Bay, but excluding San Francisco Bay, Humboldt Bay, Coos Bay, Yaquina Bay, the mouth of the Columbia River, Willapa Bay, and Grays Har- bor). If harbor porpoise density in bays were the same as that which was observed along the 18 m isobath, population sizes presented here could be in- creased by approximately 3.1% to account for por- poise inhabiting 900 km"^ (the approximate com- bined area of Humboldt Bay, Coos Bay, Yaquina Bay, the mouth of the Columbia River, Willapa Bay, and Grays Harbor). Line Transect Assumptions Biases in abundance estimates can be an even greater problem than high variability. In the case of estimates presented here, biases could be intro- duced if the assumptions of line transect sampling '^Szczepaniak, I. D., and M. A. VV^ebber. 1985. Status of the harbor porpoise {Phocoena phocoena) in the eastern North Pacific, with an emphasis on California. Contract report to the Center for Environmental Education, Washington, D.C., 52 p. 429 FISHERY BULLETIN: VOL. 86, NO. 3 are not met (Burnham et al. 1980; Hammond and Laake 1983). Of these assumptions, the most rele- vant to this study are 1) the area must be sampled randomly or the animals must be randomly distrib- uted within the area; 2) all groups on the trackline must be detected; and 3) group size must be esti- mated without error. These assumptions will be addressed below. To address the first assumption (random distribu- tion), cruise tracks were chosen to systematically cover the coast from Point Conception to Cape Flat- tery. Because the surveys were designed to cover the entire longshore range of harbor porpoise in this area, randomly placed survey tracks were deemed unnecessary. vVlthough some areas of the coast were missed, these locations were determined by weather and were presumably not correlated with porpoise abundance. Surveys were, however, limited to a very narrow strip along the 18 m isobath. Initially, the choice of this survey track was based on the observation that, in aerial surveys, harbor porpoise were usually found within 0.5 km (0.25 nmi) of the shoreline in California (Dohl et al. fn. 4). The 18 m isobath was simply the shallowest reasonable work- ing depth for the NOAA survey ships. In the course of these surveys, it was found that harbor porpoise are commonly distributed much further from the coast than 0.5 km and that one survey track could not adequately cover their habitat. The offshore distribution of harbor porpoise is not random, but is related to water depth, distance from shore, or both. The model from which I extrapolated density at 18 m to density at other depths was based on a rather limited sample at a few locations along the coast. The assumption of random search in offshore areas was not met. Additional work is required to evaluate the effect of this. The second assumption is that 100% of the animals in the immediate vicinity of the trackline were detected. Animals near the trackline can be missed because they move away from the path of the ship, because they do not surface within the visual range of the observers, or because the observers fail to detect animals that do surface. Any of these would result in a negative bias and an underestimation of porpoise abundance using line transect methods. These three problems are con- sidered in more detail. West-coast harbor porpoise are commonly said to avoid vessels (Flaherty and Stark fn. 11; Szczepa- niak and Webber fn. 12) and may be missed or not counted in the proper perpendicular distance category for this reason. On the surveys, the major- ity of harbor porpoise were oriented roughly parallel to the ship at the time they were sighted and were swimming parallel to the ship and in the opposite direction (see footnote 5). This was also observed in one instance from the helicopter; however, in that case the group first moved perpendicular to the path of the ship. These observations indicate that harbor porpoise are reacting to the ship before they are seen by observers. Reaction to and avoidance of the ship does not necessarily mean that estimates of trackline density are biased if animals are detected before they travel an appreciable distance from the trackline. In several instances, harbor porpoise sur- faced within 50 m of the ship and directly in its path. These animals appeared startled and quickly moved to avoid the ship. In these cases, the rapid move- ment of the animals and splashes associated with that movement made the animals more visible to observers. Because avoidance behavior may make harbor porpoise more visible and because the distributions of perpendicular distance show only a single mode (at the origin), vessel avoidance prob- ably does not introduce a large bias in harbor por- poise abundance estimation. More work is needed in this area. Harbor porpoise near the trackline may also be missed if they either inadvertently or intentionally do not surface within the visual range of the ob- servers. Typical mean dive times for harbor porpoise have been measured as 1.5-2.3 minutes (Glacier Bay, AK; Taylor and Dawson 1984), 1.8 minutes (northern Oregon; B. Taylor^^), and 0.4-1.4 minutes (Bay of Fundy; Watson and Gaskin 1983). The ships' speed during surveys was approximately 18.5 km/h or 310 m/min; thus, in 2 minutes the ship would travel 620 m. The average distance at which animals were first seen was 704 m from the ship. If in- dividual dive times were appreciably longer than 2 minutes, some trackline individuals would not be detected by observers. In data collected in north- ern Oregon, 16% of dive times were greater than 2.5 minutes (B. Taylor fn. 13). In addition, harbor porpoise have been reported to increase dive times up to 7 minutes in the presence of boat traffic (Flaherty and Stark fn. 9). (This latter estimate is considerably longer than any other published esti- mate, and it is possible that those researchers missed one or more surfacings). Helicopter observations in Monterey Bay indicated that porpoise groups did not extend dive times in the presence of the survey vessel (see footnote 5). This area might not be repre- sentative, however, because harbor porpoise may ''B. Taylor, Department of Biology, University of California, San Diego, CA 92093, pers. commun. August 1986. 430 BARLOW: SHIP SURVEYS OF HARBOR PORPOISE be more accustomed to vessel traffic there than along the majority of the coast. It is likely that some harbor porpoise are missed because they do not sur- face near the vessel; however, it is not possible to quantify this source of bias without additional study. Trackline animals may be missed even if they do not avoid the ship and do surface within visual range of the observers if their surfacing is not detected. In another study comparing ship surveys to aerial and shore surveys, Kraus et al. (1983) found that observers on ships saw only about 50% of the har- bor porpoise in an area. In that study, however, ship observers stood only 2.5 m above the sea surface (versus 10 m in this study), and the estimate of 50% was based on all groups, not just on trackline animals. Based on the experiment using monitor observers in the present study, an estimated 22% of harbor porpoise that surface on the trackline are missed by the usual team of 5 observers. If this is underestimated by some percentage, population size would be underestimated by the same percentage. The third critical assumption is that group size is estimated without error. In the case of harbor por- poise, group size is small and estimates are typical- ly based on actual counts. For tropical dolphins, which school in groups of several hundreds, the problem of group size estimation is more acute (Holt and Powers 1982; Hammond and Laake 1983). Only in two instances did harbor porpoise group size exceed 20: in Monterey Bay and near Point Arena, both in California. Excluding these two sightings, mean group sizes are 2.05, 2.33, 2.03, and 1.59 for surveys 1, 2, 3, and 4 (respectively); including the two sightings, means are 2.30 and 2.26 for surveys 1 and 3. These values are comparable to other esti- mates of mean group size for coastal populations of harbor porpoise: 2.2 based on aerial surveys in California (Dohl et al. fn. 4), 2.6 based on ship surveys in the Gulf of the Farallons (Szczepaniak and Webber fn. 12), 2.3 based on shore surveys in northern Oregon (see footnote 5), and 2.75-3.23 based on aerial surveys along California, Oregon, and Washington (Barlow et al. 1988). The consis- tency of all these estimates from different platforms indicates that group size estimation from ships is not likely to be a major source of bias in abundance estimation. Variance Estimation Although the estimates of standard error for abundance and density are very high, these may still be underestimates because the choice of a trunca- tion criterion was based on minimizing variance and because all possible sources of sampling errors were not considered. The model upon which relative abun- dance in the various depth strata was based is too crude to allow reasonable estimates of its variabil- ity. Estimates based on alternate models of depth distribution indicate that abundance estimation is relatively sensitive to the choice of models. Addi- tional field work may help refine this model and allow estimation of variance for the parameters 4 in Equation (4). ACKNOWLEDGMENTS Surveys of this magnitude could not be executed without the help of many people. I thank the survey crews for their many hours of labor: S. Beavers, P. Boveng, S. Bragg, S. Chivers, S. Diamond, V. DoUarhide, J. Flanders, B. Goetz, S. Hawes, S. Heimhch-Boran, A. Hohn, S. Kruse, S. Mizroch, F. Mann, M. Newcomer, R. Rasmussen, A. Read, A. Robles, M. Scott, K. Sechiguchi, S. Sexton, G. Silber, I. Szczepaniak, B. Taylor, B. Troutman, M. Webber, J. Wexler, and K. Wynne. The heli- copter crew included R. Holt, C. Oliver, and B. Taylor. The initial survey was planned in col- laboration with D. DeMaster and T. Jackson. I thank S. Buckland, R. Holt, and S. Sexton for aid in analysis and interpretation of line transect data. B. Taylor provided unpublished data from shore observations of harbor porpoise in northern Oregon. Geographic areas were computed by K. Forney. R. Allen prepared the figures presented here. This manuscript was improved by the critical reviews of P. Boveng, R. Brownell, D. Chapman, D. DeMaster, S. Diamond, D. Goodman, D. Hanan, A. Hohn, R. Holt, J. Lecky, S. Sexton, G. Smith, and B. Taylor. LITERATURE CITED Adams, L. 1951. Confidence limits for the Petersen or Lincoln index used in animal population studies. J. Wildl. Manage. 15: 13-19. Barlow, J., C. Oliver, T. D. Jackson, and B. L. Taylor. 1988. Harbor porpoise, Phocoena phocoena, abundance esti- mation for California, Oregon, and Washington: II. Aerial surveys. Fish. Bull., U.S. 86:433-444. Buckland, S. T. 1985. Perpendicular distance models for line transect sam- pling. Biometrics 41:177-195. BURNHAM, K. P., AND D. R. ANDERSON. 1976. Mathematical models for nonparametric inferences from line transect data. Biometrics 32:325-336. BuRNHAM, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr. No. 72, 202 p. 431 FISHERY BULLETIN: VOL. 86, NO. 3 BUTTERWORTH, D. S. 1982. On the functional form used for g{y) for Minke whale sightings, and bias in its estimation due to measurement in- accuracies. Rep. int. Whal. Comm. 32:883-888. Chapman, D. G. 1951. Some properties of the hypergeometric distribution with applications to zoological censuses. Univ. Calif. Publ. Stat. 1:131-160. Crandall, W. C. 1915. II. The kelp beds from lower California to Puget Sound. In F. W. Cameron (editor), Potash from kelp, 122 p. U.S. Dep. Agric. Rep. No. 100. Efron, B. 1982. The jackknife, the bootstrap and other resampling plans. CBMS Regional Conference Series in Applied Mathemetics 38, 92 p. Society for Industrial and Applied Mathematics, Phila. Gaskin, D. E. 1984. The harbour porpoise (Phocoena phocoena L.): regional populations, status, and information on direct and indirect catches. Rep. int. Whal. Comm. 34:569-586. GOETZ, B. J. 1983. Harbor porpoise (Phocoena phocoena L.) movements in Humboldt Bay, California and adjacent ocean waters. MS Thesis, Humboldt State University, Humboldt, CA, 118 P- Goodman, L. A. 1960. On the exact variance of products. J. Am. Stat. Assoc. 55:708-713. Hammond, P. S., and J. L. Laake. 1983. Trends in estimates of abundance of dolphins (Stenella spp. and Delphinics delphis) involved in the purse-seine fishery for tunas in the eastern tropical Pacific Ocean, 1977-81. Rep. int. Whal. Comm. 33:565-588. Holt, R. S. In press. Estimation of abundance of dolphin stocks taken incidentally in the eastern tropical Pacific yellowfin tuna fishery. Mar. Mamm. Sci. Holt, R. S., and J. Cologne. 1987. Factors affecting line transect estimates of dolphin school density. J. Wildl. Manage. 51:836-843. Holt, R. S., and J. E. Powers. 1982. Abundance estimation of dolphin stocks involved in the eastern tropical Pacific yellowfin tuna fishery determined from aerial and ship surveys to 1979. U.S. Dep. Commer., NOAA Tech. Memo. NOAA-TM-NMFS-SWFC-23, 95 p. Kraus, S. D., J. R. Gilbert, and J. H. Prescott. 1983. A comparison of aerial, shipboard, and land-based survey methodology for the harbor porpoise, Phocoena pho- coena. Fish. Bull., U.S. 81:910-913. Laake, J. L., K. P. Burnham, and D. R. Anderson. 1979. User's manual for program TRANSECT. Utah State University Press, Logan, Utah, 26 p. Pollock, K. H., and W. L. Kendall. 1987. Visibility bias in aerial surveys: A review of estimation procedures. J. Wildl. Manage. 51:502-510. Taylor, B. L., and P. K. Dawson. 1984. Seasonal changes in density and behavior of harbor por- poise (Phocoena phocoena) affecting census methodology in Glacier Bay National Park, Alaska. Rep. int. Whal. Comm. 34:479-483. Watson, A. P., and D. E. Gaskin. 1983. Observations on the ventilation cycle of the harbour porpoise (Phocoena phocoena L.) in the coastal waters of the Bay of Fundy. Can. J. Zool. 61(1):126-132. Yates, F. 1953. Sampling methods for censuses and surveys. Hafner Publishing Co. Inc., N.Y., 401 p. 432 HARBOR PORPOISE, PHOCOENA PHOCOENA, ABUNDANCE ESTIMATION FOR CALIFORNIA, OREGON, AND WASHINGTON: II. AERIAL SURVEYS Jay Barlow,! Charles W. Oliver,^ Terry D. Jackson,^ and Barbara L. Taylor^ ABSTRACT We conducted aerial surveys in September 1984 and September and October 1985 to determine the abun- dance of harbor porpoise along the coasts of California, Oregon, and Washington. Two observers and a recorder searched along predetermined transect lines at 0.61 and 1.85 km offshore. Strip transect methods were used. A total of 366 groups of harbor porpoise were seen in the 9,500 linear kilometers that were surveyed. Apparent density was significantly affected by sea state and cloud cover. Using observations made during optimal conditions (clear skies and calm seas), apparent harbor porpoise den- sity averaged 0.56 animals km'". Behavioral observations from shore and from a helicopter indicated that porpoise are near the surface only 23.9% of the time. To account for this, porpoise density was multiplied by a factor of 3.2, resulting in an adjusted estimate of 1.79 animals km"". Only a small percent- age of the total area inhabited was surveyed under optimal sighting conditions, hence density estimates were not extrapolated to estimate total porpoise abundance. Harbor porpoise density showed similar patterns to those measured from ship surveys, and adjusted aerial estimates are approximately equal to ship estimates. Harbor porpoise, Phocoena phocoena, are subject to mortality in the halibut set net fishery in central California (NMFS^; Diamond and Hanan^). To evalu- ate the significance of this mortality, an estimate of population size is needed. Two aerial surveys and three ship surveys were conducted from 1984 to 1986 to gather information on harbor porpoise abun- dance along the coasts of California, Oregon, and Washington. Observations were also made from shore-based stations and from a helicopter to pro- vide ancillary information needed for population estimation. Results and population estimates from 'Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. ^National Marine Fisheries Service, Southwest Region, 300 South Ferry Street, Terminal Island, CA 90731; present address: Pacific Marine Center, NOAA, 1801 Fairview Avenue East, Seattle, WA 98102. ^National Marine Fisheries Service, National Marine Mammal Laboratory, 7600 Sand Point Way, N.E., Seattle, WA 98115; present address: University of California, San Diego, Department of Biologj', La Jolla, CA 92093. ^NMFS. 1980. A report based on the workshop on stock assessment and incidental take of marine mammals involved in commercial fishing operations. January 1980. Available from National Marine Fisheries Service, National Marine Mammal Laboratory, 7600 Sand Point Way NE, Seattle, WA 98115. ^Diamond, S. L., and D. A. Hanan. 1986. An estimate of har- bor porpoise mortality in California set net fisheries: April 1, 1983 through March 31, 1984. Adm. Rep. SWR-86-15, 40 p. Available from National Marine Fisheries Service, Southwest Region, 300 S. Ferry Street, Terminal Island, CA 90731. the ship surveys are reported by Barlow (1988). Preliminary results from the 1984 and 1985 aerial surveys were presented by Oliver and Jackson^ and Oliver^ respectively. Here we present population density estimates based on the aerial surveys and on shore and helicopter observations. The aerial surveys were flown in September of 1984 and in September and October of 1985 from Point Conception, CA to Cape Flattery, WA. Sur- veys were coordinated by the National Marine Fish- eries Service (NMFS) in collaboration with the California Department of Fish and Game, the Ore- gon Department of Fish and Wildlife, and the Washington Department of Wildlife. Survey design was based on information given by Dohl et al.^ regarding harbor porpoise distribution in California. They reported that harbor porpoise were usually ^Oliver, C. W., and T. D. Jackson. 1987. Occurrence and distribution of marine mammals at sea from aerial surveys con- ducted along the U.S. west coast between December 15, 1980 and December 17, 1985. Adm. Rep. LJ-87-19, 189 p. Available from National Marine Fisheries Service, Southwest Fisheries Center, P.O. Box 271, La Jolla, CA 92038. ^Oliver, C. W. 1986. Trip report: 1985 harbor porpoise aerial survey, September 9 to October 15, 1985. Adm. Rep. LJ-86-21, 29 p. Available from National Marine Fisheries Service, South- west Fisheries Center, P.O. Box 271, La Jolla, CA 92038. »Dohl, T. P., R. C. Guess, M. L. Duman, R. C. Helm. 1983. Cetaceans of central and northern California, 1980-83: status, abundance, and distribution. Report prepared for U.S. Minerals Management Service, contract #14-12-0001-29090. Manuscript accepted May 1988. , fishery BULLETIN: VOL. 86, NO. 3, 1988. V//' ^Yj 433 FISHERY BULLETIN: VOL. 86, NO. 3 found within 0.25 nautical miles (nmi) of the shore- line. We therefore designed our aerial surveys to cover a very narrow coastal band. Subsequent in- formation from ship surveys (Barlow 1988) has shown their distribution to extend considerably far- ther from the coast. Therefore, estimates of porpoise density from aerial surveys apply to a relatively small portion of harbor porpoise habitat. For this reason we do not estimate population size by ex- trapolating aerial density estimates to the entire area inhabited. The density estimates presented here are used to corroborate estimates based on ship surveys and to estimate density for areas that were too shallow to be surveyed by ship. Based on previous studies of dive times (Watson and Gaskin 1983; Taylor and Dawson 1984) we ex- pected a proportion of the harbor porpoise to be diving and therefore missed by aerial observers. Shore-based studies were conducted in September 1985 from cliffs in northern Oregon to determine average dive times for west-coast harbor porpoise. Helicopter observations were made in April and May 1986 in Monterey Bay and near Bodega Head, CA to gather dive time information and obtain a direct measure of the fraction of time that harbor porpoise groups are visible from the air. These two samples did not differ significantly from previous samples of harbor porpoise in Alaska, so all samples were pooled to adjust estimates of porpoise density from aerial surveys to account for the probability of miss- ing submerged animals. METHODS Aerial Survey Methods Strip transect methodology (Seber 1973) was used during the aerial surveys. This method assumes that all individuals within a transect strip are detected. Transect lines were flown parallel to the coast line at distances of 0.61 and 1.85 km (0.33 and 1.0 nmi) offshore. Transect strips of equal width were sur- veyed on both sides of the aircraft. The margins of the strips were denoted by tape marks or streamers on the wing struts. Strips were divided into inside and outside swaths of unequal width (Fig. 1) by a third tape mark or streamer between the other two. When porpoise were sighted within the transect strip, the pilot was directed to leave the transect line and circle over the porpoise to obtain an ac- curate count of the number within the original group that was sighted. If additional groups or individuals were sighted during this circling, they were ex- cluded from density estimates. Porpoise density, x, was calculated as the number of individuals sighted within a transect, n, divided by the product of the transect width, w, times the distance, d, that was flown: X = n/iW ■ d) (1) We used both single and twin-propeller, high- wing, 4-passenger aircraft in our surveys. The search team consisted of two observers seated in the right and left passenger seats. A data recorder sat in the copilot's seat and did not search. If the recorder sighted animals that were missed by the observers, these were noted but were not included in density estimates. The planes were flown at an altitude of 213 m (700 feet) and at an airspeed of 158-167 km/h (85-90 knots). The original survey plan called for all sections of the coast to be covered twice on each survey. This was accomplished in 1984, but poor weather in 1985 resulted in the Washington coast and part of the Oregon coast be- ing covered only once. The dates flown and areas covered are given in Table 1. Table 1.— Dates, areas covered, and observer teams during aerial surveys for harbor porpoise. Geographic regions refer to those shown in Figure 2. Observer team refers to a pair of individuals. Regions Observer Distance Date covered team to shore (km) 9/09/84 7,8 A 0.61 9/10/84 5,7 A 0.61 9/11/84 5,6 A 0.61 9/1 3/84 1,2,3,4,5 B 0.61 9/14/84 1,2,3,4,5 B 0.61 9/17/84 1,2,3,4 B 0.61 9/18/84 1,2,3,4 B 0.61 9/11/85 6,7 C 0.61 9/16/85 1,2,3 D 0.61 9/17/85 4 D 0.61 9/18/85 4,5 D 0.61,1.85 9/1 9/85 4,5 D 0.61,1.85 9/20/85 2,3 D 1.85 10/04/85 7,8 E 0.61 10/14/85 1,2,3 D 1.85 10/15/85 4,5 D 1.85 Data gathered on both the 1984 and 1985 surveys were similar in format. Recorded data on sighting conditions included Beaufort sea state, a measure of cloud cover, a code indicating the presence of haze or fog, sun position relative to the aircraft, and a subjective measure of the observers' ability to see into the water through turbidity, surface reflection and diffraction. The latter was called surface pene- tration and was recorded separately for each of the 434 BARLOW ET AL.: AERIAL SURVEYS OF HARBOR PORPOISE STRIP TRANSECT 300 400 DISTANCE (meters) 500 600 700 E CO o H \ Inside \ Swatt \. c 1 ^ )utsi( vath --. 1985 \ '-y/yy/y//. •I-;-;-;-!-;-!";'! Wk |------^-N ^ 1 100 200 300 400 DISTANCE (meters) 500 600 700 Figure 1.— Configuration of transect strip widths and distributions of perpendicular sighting distances for aerial surveys in 1984 and 1985. Angles are given as declinations from horizontal. Histograms indicate the relative number of porpoise seen in the given distance interval. four swaths (inside and outside on both sides of the aircraft). In severe sun glare conditions, searching was discontinued for one or both swaths on one side of the aircraft. Additional data included individual observer numbers, date, time, and position (meas- ured to lOths of minutes of latitude and longitude). All of these data were recorded at the beginning and end of continuous transects and whenever condi- tions changed or a sighting was made. Additional data were recorded for marine mammal sightings, including a code for the species of animal seen, an estimate of the number of individuals, a code in- dicating on which side of the aircraft the animals were seen, and a code indicating in which of the swaths the animals were found (inside, outside, or, if the animals were not within the designated strips, neither). An attempt was made to gather the above infor- mation in a similar manner for both surveys. Some differences in subjective measures of sighting con- ditions could, however, be expected because there was no overlap in observers between years. In addi- tion, there were some differences in design between the 1984 and 1985 surveys. In 1984, the inside and outside swaths were from 123 to 305 m and from 305 to 620 m, respectively (as measured from the midline of the transect). The margins of these swaths corresponded to declinations angles of 19°, 35°, and 60° (Fig. 1). In 1985, the swaths were 91-294 m and 294-503 m and corresponded to declination angles of 23°, 36°, and 67° (Fig. 1). In 1984, effort was concentrated on the inside swath and the outside swath was not intended for abun- dance estimation. The change in swath size was in- tended to reduce the total area being searched, thus potentially allowing the outer swath to be used for density estimation. In 1985, effort was divided equally between the two swaths. During the 1985 435 FISHERY BULLETIN: VOL. 86, NO. 3 surveys the declination angle to marine mammals was measured when the animals were perpendicular to the aircraft using hand-held inclinometers. Declination angles were not measured during the 1984 survey. Distance to the coast was monitored using declination angles (19° for 0.61 km and 6° for 1.85 km). On the 1984 survey, this distance was 0.61 km for the entire survey. During the 1985 survey, we surveyed at both 0.61 and 1.85 km from the coast. In 1984, the coast was taken to be the outer limit of the surf zone. In 1985, the coast was taken to be the outer limit of the surf zone or, if kelp beds were present, the outer margin of those beds. Shore Observation Methods Observations of harbor porpoise diving behavior were made from rocky headlands in northern Ore- gon (Tillamook Head, Neahkahnie Mountain, Cape Meares, and Cape Lookout) immediately before the second aerial survey (7-11 September 1985). Ob- servers were equipped with 7 x 50 binoculars with compasses and ocular reticles and a single 20 x 120 binocular. Ventilation data were collected whenever possible and included the number of animals at the surface and the length of time spent at the surface. Observations were recited aloud by the observer and were written down by a second person or were recorded onto magnetic tape. The ventilation cycle typically consisted of a period with several surfacing rolls and breaths (which we call a surfacing series) followed by a much longer period of submergence (which we call a dive). This dive cycle corresponds to ventilation pattern B as described by Watson and Gaskin (1983) for harbor porpoise in the Bay of Fundy and the pattern described by Taylor and Dawson (1984) for porpoise in Glacier Bay. Helicopter Observation Methods Behavioral observations were also made by three observers in a 4-passenger, jet-turbine helicopter. Upon locating a group of harbor porpoise, a fluores- cein dye marker was dropped and the helicopter hovered or circled slowly above the group at an altitude of approximately 300 m. The number of animals, the time they were visible at the surface, and the dive times were recorded, along with infor- mation on cloud cover, sea state, and water turbid- ity. Each behavioral session was given a subjective rating based on how well the observers could follow the group and obtain accurate dive times. Only ses- sions with good or excellent ratings were included in analyses. Probability of Missing Submerged Animals Given that a porpoise would be within the visual range of an observer, the probability that it will be at the surface during the passage of the aircraft is related to the average time it spends at the surface, s, the average time spent below the surface, d, and the window of time during which it is within the visual range of an observer, t. This probability was calculated as Pr (being visible) s -I- t s + d (2) The probability of missing a submerged animal is equal to the complement of this value. Density Estimation Density of harbor porpoise was estimated as the number of animals seen divided by the area searched (Equation (1)). This raw density estimate was ad- justed by dividing by the probability that an animal would be visible from the air at any given instant (Equation (2)). The area searched was estimated as the swath widths times the lengths of the transects. Transect lengths were calculated as the sum of the great circle distances between successive position fixes. Densities were calculated for each of the eight statistical regions used by Barlow (1988) (Fig. 2). The statistical difference in harbor porpoise density between different sighting conditions or different areas was tested using the raw density estimates. Density estimates for short transects were frequently zero, thus violating the parametric assumptions of normally distributed, homoscedastic error. Nonparametric tests were therefore chosen for density comparisons. In discussing statistical tests, a transect segment refers to the length of transect line between two successive position fixes and are typically <20 km. The measured variables relating to sighting conditions are constant within a segment, and because each sighting is accom- panied by a new position fix, a segment will contain at most, one sighting. Whenever applicable, the Wilcoxon paired-sample test (Wilcoxon 1945) was used to test one factor while controlling for as many other factors as possi- ble. Ten paired measures of density were created 436 BARLOW ET AL.: AERIAL SURVEYS OF HARBOR PORPOISE 49°N 48° 47° 46° 43° -T — \ 1 ^-y 40° 38° 37° 36° 35° 32^ -l_ REGION REGION REGION REGION REGION 4 REGION 3 _i_ UNITED STATES Bodega Head San Francisco REGION 2 REGION 1 Point Sur \ i Point Conception \?^pa _i_ 132°W 131° 130° 129° 128° 127° 126° 125° 124° 123° 122° 121° 120° 119° 118° 117°W Figure 2.— Geographic regions used as strata in density estimation. by adding all appropriate transect segments to a linear array and dividing that array into 10 equal parts. When paired tests were not applicable, nonpara- metric ANOVA models were used. For simple com- parisons, the Kruskal-Wallis single-factor analysis of variance was used (Kruskal and Wallis 1952). For two-way comparisons, we used a two-factor exten- sion of the Kruskal-Wallis test (Scheirer et al. 1976). For both tests, three replicate measures of density were created for each cell by randomly assigning transect segments as replicate 1, 2, or 3. 437 FISHERY BULLETIN: VOL. 86, NO. 3 A principal assumption of strip transect methods is that all individuals within the designated strip are counted. Assuming that the fraction of diving animals does not vary with sighting conditions, any effect of sighting conditions on apparent density is likely due to missed animals. In any instance where we show that poorer sighting conditions result in a significant decrease in apparent density (one-tailed for paired tests, tw^o-tailed for ANOVA tests), we eliminate the category of sighting conditions which resulted in that lower estimate. If it is not possible to predict which category would be worse a priori in paired tests (e.g., right swath vs. left), two-tailed probabilities are used. Although we cannot be sure of eliminating all biases using these methods, this pattern of data paring should avoid much of the bias due to missed animals. RESULTS In 1984, 247 groups of harbor porpoise (680 in- dividuals) were seen within transect strips which covered a linear distance of 5,763 km. In 1985, we saw 119 groups (384 individuals) in surveys of 3,715 km. Mean group sizes were 2.75 and 3.23 in- dividuals, respectively, for 1984 and 1985. For 1984, the relative frequencies of individuals seen within the inside and outside swaths are illustrated in Figure 1. For 1985, the perpendicular distances from the trackline to the animals were calculated from declination angles, and the relative distribu- tion of sightings is shown in Figure 1 as a function of perpendicular distance. Inside vs. Outside Swath For 1984 data, only the inside swaths were used, but for 1985 both inside and outside swaths were considered for density estimation. For 1985, we tested whether the density in the inside swaths was greater than the density of the outside swaths. We only considered cases when the water surface pene- tration codes were equal in both the inside and out- side swaths. Data for inside and outside were thus paired, with all other sighting factors equal. For 1985, the density in the inside was greater (0.09 vs. 0.06 porpoise/km^), but this difference was not significant (P > 0.10). Surface Penetration Observers used a subjective coding system to describe their ability to see through the sea surface. Cloud cover, haze, and water turbidity contributed to poor surface penetration. In 1984, observers used codes to indicate good and poor conditions; in 1985, observers used codes for excellent, good, and poor. There were frequent cases when observers recorded different codes for the inside swaths on opposite sides of the plane; hence, paired tests were again appropriate. For 1984, we tested whether density in the "good" category was higher than density in the "poor" category. When surface penetration was different in the inside swaths on opposite sides of the plane, mean density in the "good" category was higher than in the "poor" category (0.16 vs. 0.12 porpoise/km^), but this difference was not signifi- cant (P > 0.25). For 1985, no tests were necessary because the mean density in "excellent" category was lower than in the "good" category, and like- wise, density in the "good" was lower than in the "poor" category. All categories of water surface penetration were included in subsequent analyses. Effects Due to Observers and Side of the Plane Sightings were classified based on which observer made the sighting and on whether the sighting was on the inshore or offshore side of the aircraft. In fact, these two classifications were confounded in 1985 because the two principal observers were sit- uated on the same sides of the aircraft for most of this survey. Effects of these classifications on den- sity estimation were considered together. Survey teams were defined as pairs of observers who worked together. There were two such teams for 1984 and three for 1985 (Table 1). Only one of the teams in 1985 had sufficient numbers of sightings to be considered here. Statistical tests were based on paired cases during which both members of the sighting team were searching. For 1984, porpoise density on the offshore side of the airplanes was greater than on the inshore side for both team A and team B (Table 2). The differ- ence in density estimates between observers was less than the difference between inshore and off- Table 2.— Relative harbor porpoise densities (km"^) for teams of observers. Density estimates are stratified by inshore and offshore sides of the aircraft and by individual observers. Observer team Inshore Offshore Observer 1 Observer 2 A B D 0.275 0.168 0.281 0.425 0.244 0.164 0.320 0.232 0.300 0.380 0.179 0.146 438 BARLOW ET AL.: AERIAL SURVEYS OF HARBOR PORPOISE shore (Table 2), and neither was statistically signif- icant. For 1985, the opposite was seen; the density on the inshore side was greater than on the offshore side and this difference was less than the difference between observers (Table 2), but again, neither dif- ference was statistically significant. In 1984, both observers had previous experience doing aerial surveys, but in 1985, the observer with the lower density estimates had no previous experience in cetacean surveys. Experience may be a factor in density estimates from strip transects, but since the inexperienced observer was always on the same side of the plane, it was not possible to test this with a factorial design. Area and Sea State Analysis of research vessel data (Barlow 1988) in- dicated two geographic regions in California with low porpoise density (regions 1 and 3 in Figure 2). We tested whether density observed from aircraft are also lower in these areas. Because observation conditions may have differed in the two areas, we included sea state as a second factor in a two-way ANOVA. Mean values are presented in Table 3. For 1984, three categories of sea state were used: Beaufort 0 & 1, Beaufort 2, and Beaufort 3 and greater. The effect due to area was significant (P < 0.001), with the area that showed low density in the ship surveys also showing lower density in the aerial survey. For 1985, only the first two categories of sea state were used due to insufficient data at Beaufort 3 and greater. Again the effect of area was significant (P < 0.025), and the same trends were seen. In neither case were the effects of sea state Table 3. — Uncorrected harbor porpoise densities (km' ) for the two-way comparison of area and sea state. Low-density areas refer to two regions in California that were found to have much lower than average density in previous ship surveys (see text). High-density areas include all other regions. Only inside swaths were included. Numbers in parentheses refer to area (km^) surveyed under the given condition. Densities in brackets were excluded from two-way comparisons due to the small area covered in one cell. 1984 survey 1985 survey Beaufort sea state Low- density area High- density area Low- density area High- density area 0 & 1 0.027 (183) 0.579 (264) 0.089 (124) 0.762 (126) 2 0.081 (308) 0.312 (1012) 0.089 (214) 0.183 (425) 3-1- 0.102 (79) 0.191 (252) [0.000] (28) [0.049] (566) or the interaction effects significant. To eliminate area effects from confounding statistical results, only data for the larger, high-density area were in- cluded in subsequent tests. The low-density area was included in later estimates of overall harbor porpoise density. Sea State and Cloud Cover Both sea state and cloud cover can affect sighting conditions. Because both are affected by local weather, the effects of these are likely to be con- founded. These two factors were therefore tested simultaneously in a two-way ANOVA. We con- sidered only the inside swath and excluded the two low-density regions. We used the same sea state categories as above. The sky was categorized as clear if cloud cover was <25% and cloudy if >25%. Mean porpoise densities for each category are given in Table 4. It was necessary to exclude the Beaufort 0 & 1 category for 1985 because only 52 km were surveyed in cloudy conditions for these sea states. The effect due to cloud cover was signifi- cant for 1984 (P < 0.025) and 1985 (P < 0.05). The effect of sea state and the interaction effect of sea state and cloud cover were not significant for either survey (P > 0.10). Transect segments with >25% cloud cover were excluded from subsequent analyses. Table 4. — Uncorrected harbor porpoise densities (km"^) for the two-way comparison of sea state and cloud cover. Clear refers to <25% cloud cover, and cloudy refers to >25%. Data include only inside swaths in high-density areas. Numbers in parentheses refer to area (km^) sur- veyed under the given condition. Densities in brackets were excluded from two-way comparisons due to the small area covered in one cell. Beaufort sea state 1984 survey 1985 survey Clear Cloudy Clear Cloudy 0& 1 1.340 (81) 0.240 (183) [0.832] (115) [0.000] (11) 2 0.371 (572) 0.236 (440) 0.271 (232) 0.078 (193) 3 + 0.266 (162) 0.056 (90) 0.088 (272) 0.014 (294) Sea State The effect of sea state was tested alone using only the transect segments which occurred under clear skies (<25% cloud cover) and within the high-density areas. For 1984, only the inside swath was included. 439 FISHERY BULLETIN: VOL. 86, NO. 3 For 1985, both inside and outside swaths were in- cluded. As in previous stratifications, apparent den- sity clearly decreased with increasing sea state (Table 5). This effect was significant for surveys in both 1984 (P < 0.05) and 1985 (P < 0.005). Transect segments surveyed during sea states of 2 or greater were excluded in subsequent analyses. •Uncorrected harbor porpoise den- ^) for the stratification based on sea Table 5.- sities (km state. Data include only high-density areas sur- veyed when cloud cover was <25%. Data for 1984 include inside swaths only; data for 1985 include inside and outside swaths. Numbers in parentheses refer to area (km^) surveyed under the given condition. Beaufort sea state 1984 survey 1985 survey 0 & 1 1.340 0.807 (81) (234) 2 0.371 0.193 (572) (471) 3 + 0.266 0.058 (162) (553) Between Survey Differences We considered the 1984 survey and the 1985 surveys at 0.61 km and 1.85 km from shore as three independent estimates of harbor porpoise density. Because apparent density was shown to vary greatly with sighting conditions and because sighting con- ditions varied between surveys, it was necessary to compare these three under similar conditions. The highest (and presumably least biased) den- sities were obtained when sea state was Beaufort 0 & 1 and when cloud cover was <25%. Between survey comparisons imder these conditions are given in Table 6 for the eight geographic regions given in Figure 2. For 1985, there were no transect segments at 0.61 km from shore under the condi- tions Beaufort 0 & 1 and clear skies. For 1984, only three regions contained more than 10 km- of searching effort at 0.61 km from shore. For 1985, only four areas had any searching effort at 1.85 km from shore. The only direct density comparisons with reasonable sample sizes are for region 1 (0.000 vs. 0.048 porpoise/km^) and region 3 (0.111 vs. 0.110 porpoise/km^) (respectively for 1984 and 1985). The densities for all regions pooled (0.671 and 0.510 porpoise/km^) are similar, but because of the small sample size and geographic variation in sam- pling, a statistical test of this difference is mean- ingless. In comparing surveys, sample size and regional coverage improved slightly when Beaufort 2 was considered (still allowing a maximum of 25% cloud cover) (Table 7). For 1984, coverage was relatively complete in all regions. For 1985, coverage at 0.61 km from shore was limited to regions 2-4, and coverage at 1.85 km was limited to regions 1-5. Comparing the two surveys in 1985, distance from shore made little difference in overall density for all regions combined, and neither survey had con- sistently higher values than the other. Comparing the 1984 survey to the two 1985 surveys, the former had a higher overall density for all regions combined, but again this difference was not consistent among regions. Sample size and regional coverage were again too poor for meaningful statistical tests. Table 6.— Uncorrected harbor porpoise densities (km"^) in eight geographic regions surveyed during Beaufort 0 & 1 con- ditions. Data for 1984 are based on inside swaths of transects flown 0.61 km from the coast. Data for 1985 are based on in- side and outside swaths of transects flown at 0.61 and 1.85 km from the coast. Only those segments surveyed when cloud cover was <25% are included. Numbers in parentheses refer to area (km^) surveyed under the given condition. Geographic region 1984 survey density 0.61 km 1985 survey density 0.61 km 1985 survey density 1.85 km 1 0.000 0.048 (43) (0) (103) 2 0.000 — 0.609 (1) (0) (51) 3 0.111 — 0.110 (45) (0) (55) 4 0.953 — 0.862 (6) (0) (183) 5 1.120 — — (9) (0) (0) 6 1.374 — — (64) (0) (0) 7 5.745 — — (1) (0) (0) 8 (0) (0) (0) All regions 0.671 — 0.510 (170) (0) (392) Ventilation Patterns Harbor porpoise did not appear to react to the helicopter during aerial observations; they were visible throughout a surfacing series and were not visible during dives. Knowing this, we were able to use data on surfacing series and dive times to deter- mine the fraction of time harbor porpoise would be visible from the air. Data on ventilation patterns 440 BARLOW ET AL.: AERIAL SURVEYS OF HARBOR PORPOISE were available from the helicopter study in Califor- nia (13 groups were observed, mean group size was 2.7), the Oregon shore observations (11 groups, mean group size was 4.2), and a previous shore study in Glacier Bay, AK (28 solitary individuals). The mean times spent in surfacing series were 34.2, 24.6, and 30.0 seconds (respectively for the three studies). The corresponding mean dive times did not differ significantly between study sites (P > 0.05 using pairwise ^tests). Using the pooled data set (n = 52), the mean time spent in a surfacing series was 30.02 seconds (SE = 1.95) and the mean time spent in a dive was 95.81 seconds (SE = 5.32). The percent- age of time spent at the surface is 23.9%. Average and Adjusted Density Estimates Harbor porpoise densities under optimal condi- tions (Beaufort 0 & 1 and <25% cloud cover) were averaged for the two surveys, weighting by transect length (Table 8). Given an average survey speed of 160 km/h and assuming that the window for harbor porpoise observation is 400 m long, the time win- dow during which a point would be visible is 9.0 Table 7— Uncorrected harbor porpoise densities (km"^) in eight geographic regions surveyed during Beaufort 2 condi- tions. Data for 1984 are based on inside swaths of transects flown 0.61 km from the coast. Data for 1985 are based on in- side and outside swaths of transects flown at 0.61 and 1.85 km from the coast. Only those segments surveyed when cloud cover was <25% are included. Numbers in parentheses refer to area (km^) surveyed under the given condition. Geographic region 1984 survey density 0.61 km 1985 survey density 0.61 km 1985 survey density 1.85 km 1 2 3 4 5 6 7 8 All regions 0.000 (107) 0.092 (109) 0.196 (117) 0.562 (180) 0.489 (119) 0.439 (57) 0.114 (70) 0.268 (37) 0.295 (796) (0) 0.102 (59) 0.062 (48) 0.282 (110) 0.459 (6) (0) (0) (0) 0.192 (224) 0.184 (103) 0.103 (68) 0.287 (52) 0.070 (57) 0.234 (171) (0) (0) (0) 0.188 (452) seconds. Using this estimate and the surface and dive times estimated above, the probability that a porpoise will be seen is estimated as 0.310 from Equation (2). An instantaneous count would there- fore underestimate porpoise abundance by a factor of 3.2. Average values were therefore multiplied by this factor (Table 8). DISCUSSION Results indicate that sighting conditions must be very good in order to estimate harbor porpoise abun- dance from aerial strip transects. Both sea state and cloud cover had very large and significant effects on apparent density. Limiting observations to the best categories of sea state (Beaufort 0 & 1) and cloud cover (<25%) can be used to minimize the bias due to missed animals. These conditions are, how- ever, rare and only occurred during 5.3% and 10.3% of the transects in 1984 and 1985. The actual occur- rence of these conditions is even more rare if one considers flights that were cancelled due to bad weather. The effects on sea state and cloud cover on sight- ing conditions were predicted by observers before analysis of survey data was begun. Most harbor por- poise were first seen when submerged a small distance below the surface. Surfacings were rela- tively inconspicuous to aerial observers and were not an important cue in sighting porpoise. Both sea state and cloud cover affect the ability of observers to see through the water's surface and to spot submerged animals. Increasing sea state causes more refrac- tion of light at the water's surface, increases glit- Table 8.— Harbor porpoise densities (km"^) in eight geographic regions from a) a weighted average of uncorrected estimates from 1984 and 1985 aerial surveys, b) the same average adjusted by a factor of 3.2 to account for submerged porpoise that were missed by aerial observers, and c) a weighted average for the 1 984 and 1985 ship surveys. Aerial estimates are based on observations made under conditions of Beaufort 0 & 1 and with <25% cloud cover. Ship estimates include Beaufort sea states of 0, 1, and 2. Geographic region Area surveyed (km*) Aerial estimates Uncorrected a Research vessel Corrected estimates b c 1 146 0.03 2 1 — 3 100 0.11 4 189 0.86 5 9 1.12 6 64 1.37 7 1 — 8 0 — All regions 562 0.55! 0.10 0.05 — 0.66 0.35 0.04 2.75 3.68 3.58 1.18 4.38 2.88 — 3.43 — 1.42 1.79 1.73 441 FISHERY BULLETIN: VOL. 86, NO. 3 tery reflection of sunlight, and causes whitecaps which obscure subsurface observation. Cloud cover decreases penetration of sunlight into the sea and causes its surface to appear dark and glazed. It is therefore not surprising that calm seas and clear skies result in higher apparent densities. It is more surprising that apparent density did not vary with observer's subjective appraisal of surface penetration. This may have been because surface penetration was only tested in paired cases for which sea state and cloud cover were identical. In these cases, differences in surface penetration may have been due primarily to subjective differences in the way individual observers were coding it. If all cases are considered, surface penetration is very highly correlated with sea state and cloud cover and prob- ably could be used as an alternative measure of sighting conditions. We prefer, however, to use sea state and cloud cover because their measure is less subjective than surface penetration and could more easily be used by other researchers. Missed Animals A principal assumption of strip transect methods is that all individuals within the strip are counted. We cannot necessarily meet this assumption just by eliminating the categories of sighting conditions with significantly lower density. In fact, it is possi- ble that this method of selectively eliminating data could overestimate density by eliminating a category of sighting conditions which (by random chance) had a significantly lower density. We do not believe that this is likely in the cases of sea state or cloud cover because the trends were the same for both surveys and because the categories that were eliminated were judged a priori as being poorer sighting con- ditions. We believe that porpoise density is more likely underestimated due to missed animals. Kraus et al. (1983) found that observers in aircraft saw only 14% of the harbor porpoise groups known to be pres- ent based on shore-based observations. Missed animals may include some individuals that were near the surface and visible but were not seen, as well as others that were diving and were too deep to be seen. We infer that some near-surface animals were missed based on three reasons. First, apparent den- sity decreased with increasing sea state and cloud cover, hence near-surface animals must be missed in (at least) the poorer conditions. (An alternative explanation is that porpoise spend less time near the surface when sighting conditions are poor.) Second, of the two principal observers in 1985, the less ex- perienced observer may have missed more porpoise. Third, in five instances in 1984 and one instance in 1985, the data recorder saw harbor porpoise in the inside swath that were missed by the observer. In three of these cases, conditions were Beaufort 1 with <25% cloud cover. Recorders searched only occa- sionally as conditions permitted, so it is not possi- ble to use these data to quantify how many near- surface animals were missed under good conditions. Based on behavioral observations, we can also be certain that some harbor porpoise are missed be- cause they are too deep to be seen. Water visibility was typically only 2-5 m during the surveys. Aerial observers have seen harbor porpoise dive out of view during the passing of the plane or while circling. Harbor porpoise were not visible from the helicopter during dives, even in very calm water. Some frac- tion of porpoise must be missed because they are too deep to be seen. We have tried to account for the fraction of diving animals by dividing density estimates by the frac- tion of time harbor porpoise are known to be near the surface (i.e., within surfacing series). Uncor- rected estimates of harbor porpoise density are based on the assumption that porpoise are never too deep to be seen; this undoubtedly results in an underestimate of porpoise density. Our method of adjusted estimates of porpoise density assumes that, when diving, porpoise are always too deep to be seen; this has been corroborated by helicopter ob- servations. The latter estimate should therefore be closer to the true value of porpoise density (if biases due to other factors have been eliminated). Offshore Distribution of Harbor Porpoise In 1985, aerial transects were flown at 0.61 and 1.85 km from the shore. The latter value was chosen to correspond approximately to the ship transects along the 10-fathom (18.3 m) isobath. The intent was to directly compare aerial estimates of density to estimates made from ships at the same distance from shore (considered below) and to provide a means to extrapolate ship estimates to regions that were too shallow to survey by ship. In regard to the latter, we wish to know whether harbor porpoise density at 0.61 km from shore is different from that at 1.85 km. In 1985, surveys at 0.61 km were never flown under good sighting conditions (Beaufort 0 & 1 and <25% cloud cover), so direct comparisons between 0.61 and 1.85 km are not possible. Considering only the best category of sighting conditions, it is possi- 442 BARLOW ET AL.: AERIAL SURVEYS OF HARBOR PORPOISE ble to compare 1985 transects at 1.85 km with 1984 transects at 0.61 km in geographic regions 1 and 3 (Table 6). Sample sizes are very small, but den- sities are roughly comparable (0.070 vs. 0.057, respectively for 1.85 and 0.61 km from shore). Sam- ple sizes in 1985 can be increased if Beaufort 2 is considered instead of Beaufort 0 & 1 (still with <25% cloud cover). Based on geographic regions 1 through 5, porpoise densities are virtually identical at 1.85 and 0.61 km from shore (0.188 vs. 0.192, respective- ly, Table 7). In a preliminary model of harbor porpoise depth distribution. Barlow (1988) hypothesized that har- bor porpoise density is constant from the shore to the 75 m isobath. Ship data did not, however, in- clude any transects inshore of 18.3 m depth. In this shallow area, the model was not based on any data. Although sparse, data from aerial surveys show that densities are similar in areas that were surveyed by ship (1.85 km from shore) and in areas that were too shallow to be surveyed by ship (0.61 km from shore). These data are consistent with the model pro- posed by Barlow. Comparison of Ship and Aerial Density Estimates Estimates of harbor porpoise density from 1984 to 1985 ship surveys (Barlow 1988) can be compared with adjusted and unadjusted estimates from the aerial surveys (Table 8). The overall density for all regions is higher for the ship surveys than for the unadjusted aerial. Adjusted estimates from the aerial surveys are very close to the overall estimates from ship surveys. Previous comparisons have been made of sighting efficiency from aerial and surface vessel platforms (Kraus et al. 1983). They found that observers on boats saw 52% of the harbor porpoise groups seen by shore-based observers, whereas aerial observers saw only 14%. Based on this, density from aerial surveys might be expected to be only 27% of that from ship surveys. In the present study, unadjusted density based on aerial surveys is 32% of the den- sity from ship surveys. The two studies are not directly comparable, however. Weather conditions are not reported by Kraus et al. (1983) and may have included less than optimal sighting conditions. It should be noted that the fraction of harbor porpoise groups seen by aerial observers in their study (14%) is even lower than the fraction of harbor porpoise we assumed would be in surfacing series and hence near the surface (23.9%). It is possible that behavioral differences between harbor porpoise from the two coasts (such as travelling or behavior mode "A" noted by Watson and Gaskin (1983)) could account for some of the differences noted above. Estimates based on ship surveys cannot, of course, be considered the true density of harbor porpoise. The overall estimate based on ship surveys is rela- tively imprecise (C.V. = 49%) and may be biased (Barlow 1988). The ship survey estimate is, however, superior to current estimates from aircraft for sev- eral reasons. Line transect methods were used on the ship surveys, and the principal assumption of this method (that 100% of the animals in the imme- diate vicinity of the trackline are seen) is more easily met than the comparable strip transect assumption (that 100% of the animals within a strip are seen). Acceptable sighting conditions for ship surveys in- cluded Beaufort 0 & 1 , and 2 and were not restricted by cloud cover (Barlow 1988). This allowed more complete geographic coverage than did aerial surveys. Also, the ship travelled much slower than the aircraft (10 knots vs. 80-90 knots), thus the prob- ability of missing a diving individual was much less. Barlow (1988) calculated that diving animals located near the trackline would be missed by observers on ships only if dive times exceeded 2 minutes. Final- ly, estimates of the correction factor to account for submerged animals is relatively imprecise. Addi- tional observations on ventilation patterns may allow further refinements in density estimates based on aerial surveys. RECOMMENDATIONS The design of future surveys for harbor porpoise could be improved based on the results obtained from our aerial surveys. We found that sighting con- ditions deteriorated rapidly with both increasing cloud cover and rougher sea states. To the extent that is possible, aerial surveys for harbor porpoise should only occur on clear days with little wind. Ob- servations made by the data recorders indicate that some harbor porpoise will be missed even in good sighting conditions. If strip transects are used, ex- periments with two independent teams of observers searching at the same time could be used to quan- tify the fraction of animals that are missed by using just one team. Given that fewer harbor porpoise were seen between 400 and 500 m of the track line, we also suggest that, when surveying at 213 m altitude, the strip widths should be decreased to only include the area between 100 and 400 m. We believe, however, that the problem of miss- ing harbor porpoise could be reduced if line transect methods were used in place of strip transects. Line 443 FISHERY BULLETIN: VOL. 86, NO. 3 transect methods assume that 100% of all animals are seen directly along the trackline and use statistical techniques to estimate the number of animals that are missed as a function of the distance from this tracWine. Line transects would require use of an aircraft with unobstructed downward visibil- ity through a belly window and use of a third ob- server who could view animals directly under the aircraft. Harbor porpoise are now known to occur further from the shoreline than was believed at the begin- ning of this study (Barlow 1988). In future surveys, transect lines should be placed so as to cover a greater fraction of the harbor porpoise habitat. ACKNOWLEDGMENTS Support, personnel, and flight time were provided by the Southwest Fisheries Center (NMFS), the Southwest Regional Office (NMFS), the California Department of Fish and Game, the National Marine Mammal Laboratory (NMFS), the Washington State Department of Wildlife, and the Oregon Department of Fish and Wildlife. The observers and data recorders were R. Brown, S. Diamond, D. Hanan, R. Holt, S. Jeffries, W. Ferryman, and B. Troutman (also Jay Barlow, Terry D. Jackson, and Charles W. Oliver). The pilots were R. Anthes, J. Cain, L. Heitz, J. York, and B. Zorich. Shore observations were made by M. Herter, T. LoughHn, D. Merrick, R. Rowlett, D. Rugh, and B. Yerman (also Barbara L. Taylor). Unpublished porpoise data from Alaska were collected by P. Dawson (and Barbara L. Taylor). The method of adjusting for submerged animals was suggested by D. Chapman based on an unpublished report by R. A. Davis. Drafts of this paper were reviewed by R. Brown, D. DeMaster, K. Forney, D. Hanan, R. Holt, S. Jeffries, and two anonymous reviewers. LITERATURE CITED Barlow, J. 1988. Harbor porpoise Phocoena phocoena, abundance estima- tion for California, Oregon and Washington; \. Ship surveys. Fish. Bull., U.S. 86:417-432. Kraus, S. D., J. R. Gilbert, and J. H. Prescott. 1983. A comparison of aerial, shipboard, and land-based survey methodology for the harbor porpoise, Phocoena pho- coena. Fish. Bull., U.S. 81:910-913. Kruskal, W. H., and W. a. Wallis. 1952. Use of ranks in one-criterion analysis of variance. J. Am. Stat. Assoc. 47:583-621. Scheirer, C. J., W. S. Ray, and N. Hare. 1976. The analysis of ranked data derived from completely randomized factorial designs. Biometrics 32:429-434. Seber, G. A. F. 1973. The estimation of animal abundance and related param- eters. Hafner Press, N.Y., 506 p. Taylor, B. L., and P. K. Dawson. 1984. Seasonal changes in density and behavior of harbor por- poise (Phocoena phocoena) affecting census methodology in Glacier Bay National Park, Alaska. Rep. int. Whaling Comm. 34:479-483. Watson, A. P., and D. E. Gaskin. 1983. Observations on the ventilation cycle of the harbour porpoise Phocoena phocoena (L.) in the coastal waters of the Bay of Fundy. Can. J. Zool. 61(1):126-132. Wilcoxon, F. 1945. Individual comparisons by ranking methods. Biom. Bull. 1:80-83. 444 EVALUATION OF VARIABILITY IN SABLEFISH, ANOPLOPOMA FIMBRIA, ABUNDANCE INDICES IN THE GULF OF ALASKA USING THE BOOTSTRAP METHOD Michael F. Sigler and Jeffrey T. Fujioka^ ABSTRACT Relative population numbers (RPN's) and length compositions were computed for sablefish, Anoplopoma fimbria, in the Gulf of Alaska from the results of the Japan-United States cooperative longline survey from 1979 to 1986. A statistical evaluation of annual changes in the RPN's using the bootstrap method is demonstrated and showed that sablefish abundance increased significantly from 1979 to 1986, an in- crease likely due to recruitment of two strong year classes. The effect of missing data on the bootstrap calculations was examined and found to be negligible. Early in this century, United States and Canadian fishermen began harvesting sablefish, Anoplopoma fimbria, in nearshore waters from California north to southeastern Alaska, but sablefish were not heavily exploited until Japanese longline vessels began fishing in the Bering Sea in 1958. Japanese catches off both the U.S. and Canadian coasts rose dramatically in the following two decades. After passage of the Fishery Conservation and Manage- ment Act, foreign catches were reduced and the domestic allocation was increased. The domestic catch eventually increased nearly fivefold, thereby replacing the foreign fishery, and in 1985, sablefish in the Gulf of Alaska were harvested entirely by domestic fishermen. Before the reduction of the foreign fishery, infor- mation on sablefish abundance consisted of statistics on catch per unit effort from the Japanese longline fishery. In 1978, the Fisheries Agency of Japan and the U.S. National Marine Fisheries Service (NMFS) began a cooperative longline survey along the con- tinental slope of Alaska to assess the abundance of sablefish and Pacific cod, Gadus macrocephalus. The survey, conducted annually, has provided eight con- secutive years (1979-86) of data for the Gulf of Alaska, seven years (1980-86) of data for the Aleu- tian region, and five years (1982-86) of data for the eastern Bering Sea. The first year of the survey, 1978, was experimental. Relative population numbers (RPN's) and length compositions from the longline survey results from 'Northwest and Alaska Fisheries Center Auke Bay Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 210155, Auke Bay, AK 99821. 1979 to 1985 have been estimated previously by Sasaki (1986). In this study, the observed increase in the RPN's for sablefish in the Gulf of Alaska is evaluated statistically and an explanation of the probable source of the increase is discussed. Statistical analysis of the survey results is based on the bootstrap method (Efron 1982; Efron and Gong 1983). This method is a relatively new statis- tical procedure that has been little used in fisheries analysis. Thus, this paper also demonstrates an application of the bootstrap method to statistical evaluation of fishery survey data. SURVEY METHODS The Gulf of Alaska portion of the Japan-U.S. cooperative longline survey, conducted annually each summer from 1979 to 1986, covered five International North Pacific Fisheries Commission (INPFC) statistical areas: Shumagin, Chirikof, Kodiak, Yakutat, and Southeastern (Fig. 1). One of a total of 47 stations each ranging in depth from about 100 to 1,000 m was sampled daily by longline. The longline was 16 km long and consisted of 160 hachis (the Japanese word for "skate" or length of longline); each hachi was 100 m long and consisted of 45 hooks baited with squid. Soak time, the time between setting and retrieval, varied from 3 hours at the beginning of the longline gear to 7 or 8 hours at its end. The depth at which the fish were caught was estimated by measuring the depth of water under the vessel with an echo sounder every fifth hachi. The fish caught were tallied by species and hachi as the longline was brought aboard, then they were weighed and their length was measured. Most Manuscript accepted May 1988. FISHERY BULLETIN: VOL. 86, NO. 3, 1988. 445 FISHERY BULLETIN: VOL. 86, NO. 3 Figure 1.— International North Pacific Fisheries Commission statistical areas sampled during the Gulf of Alaska portion of the Japan-U.S. cooperative longline survey, 1979-86. sablefish were sexed, some were tagged and re- leased, and others were sampled for otoliths and scales. Detailed survey methods are described in Sasaki et al. (1983). STATISTICAL METHODS The catch data were stratified because of differ- ences in the catch rate by depth. Assignment of the 150 - U z o lU c 100- 0.5 5.5 10.5 15.5 CATCH PER HACHI 20.5 25.5 446 Figure 2.— Frequency distribution of sablefish catch per hachi (numbers) by year, station, and strata, 1979-86. SIGLER and FUJIOKA: VARIABILITY IN SABLEFISH catch data to a stratum was based on the recorded depth of every fifth hachi or the interpolated depth of each of the intervening four hachis. The number of strata totaled nine, with the first stratum repre- senting 101-200 m, the second representing 201- 300 m, and so on. A catch per hachi value was calculated for each stratum of a station (Fig. 2), and the resultant values within each statistical area were averaged. Not all strata were sampled at each station and as a result, sometimes no stations were sampled within certain strata (Table 1). Although not sam- pled one year, these strata generally were sampled the previous year, because of slight annual differ- ences in the depths sampled at each station. Catch per hachi values CPUE,j were estimated for sta- tions within the unsampled strata from the catch per hachi values for sampled stations from the previous year and the annual change in the catch per hachi values in the adjacent stratum. For each station: CPUE, J = CPUE,_i J * CPUE ,,j /CPUEi_i__,. where i = yed.r,j = stratum, a.ndj' = the adjacent stratum. Catch per hachi values were estimated only for stations within unsampled strata. If one or more stations within a stratum were sampled, then no catch per hachi values were estimated for the un- sampled stations. The effect of these missing values on the survey results vdll be discussed later. Relative population numbers were calculated from the catch per hachi values to index annual changes in sablefish abundance. As in Sasaki (1985), the aver- age catch per hachi was multiplied by the areal size of each stratum and statistical area to calculate an RPN for each stratum and statistical area. The areal sizes used to calculate the RPN's are for the con- tinental slope only (Table 2), which corresponds to the area covered by the survey. The resultant RPN's were summed across strata to calculate an RPN for each statistical area, and these RPN's were summed Table 1.— Number of stations sampled by year, INPFC area, and depth during the Japan-U.S. cooperative longline survey, 1979-86. Depth INPFC Statistical area Depth INPFC statistical area Shu- Chiri- South- Shu- Chiri- South- Year (m) magin kof Kodiak Yakutat east Year (m) magin kof Kodiak Yakutat east 1979 101-200 7 4 6 6 2 1983 101-200 9 5 7 8 3 201-300 7 5 7 10 5 201-300 10 6 9 11 9 301-400 7 5 8 6 301-400 10 5 9 11 8 401-500 7 5 9 6 401-500 10 6 9 11 8 501-600 6 4 9 6 501-600 10 5 9 11 8 601-700 6 4 8 5 601-700 9 5 9 11 7 701-800 6 4 7 5 701-800 8 4 8 11 7 801-900 5 1 4 9 4 801-900 5 4 4 10 7 901-1,000 4 1 2 4 1 901-1,000 2 0 1 5 2 1980 101-200 9 5 8 7 4 1984 101-200 10 7 7 8 2 201-300 10 6 8 10 6 201-300 10 7 9 11 9 301-400 10 6 8 10 8 301-400 10 6 9 11 8 401-500 10 6 9 10 8 401-500 10 6 9 11 8 501-600 10 5 9 10 7 501-600 10 5 9 11 8 601-700 9 5 9 9 7 601-700 8 5 8 11 8 701-800 6 4 8 9 7 701-800 7 5 6 10 8 801-900 6 2 3 7 4 801-900 3 0 3 7 5 901-1,000 5 1 2 6 1 901-1,000 0 0 0 6 1 1981 101-200 9 6 7 8 1 1985 101-200 9 6 6 7 1 201-300 10 6 9 11 9 201-300 10 7 8 11 9 301-400 10 6 9 11 9 301-400 10 6 9 11 8 401-500 10 6 9 11 9 401-500 10 6 9 11 7 501-600 10 6 9 11 9 501-600 10 6 9 11 8 601-700 9 4 9 11 9 601-700 9 5 9 11 8 701-800 7 2 7 11 9 701-800 7 4 9 10 8 801-900 6 1 5 7 7 801-900 4 4 4 11 6 901-1,000 2 0 1 2 1 901-1,000 0 0 0 5 2 1982 101-200 9 5 7 10 4 1986 101-200 9 5 7 7 2 201-300 9 6 9 11 9 201-300 10 6 9 11 9 301-400 10 5 9 11 8 301-400 10 6 9 11 8 401-500 10 6 9 11 8 401-500 10 6 9 11 8 501-600 10 5 9 11 8 501-600 10 6 9 11 8 601-700 10 5 9 11 8 601-700 10 4 9 11 8 701-800 9 4 8 11 7 701-800 10 3 8 11 8 801-900 7 4 6 9 6 801-900 5 2 7 11 8 901-1,000 3 3 5 8 3 901-1,000 1 1 3 5 1 447 FISHERY BULLETIN: VOL. 86, NO. 3 Table 2.— Slope area sizes (km^. 2\1 Depth (m) Slope area (km'') Shumagin Chlrikof Kodiak Yakutat Southeast 201-400 4,001 2,350 3,106 2,988 1,781 401- -600 2,269 1,766 2,255 1,666 822 601- -800 1,629 1,955 1,923 1,470 1,006 801- -1,000 1,248 2,012 2,296 1,489 1,165 201- -1.000 9,147 8,083 9,580 7,613 4,774 'Shumagin, Chirikof, and Kodiak areas and Yakutat area from long. 147-I440W; data from E. Brown (NWAFC Seattle Laboratory, NMFS. NOAA, 7600 Sand Point Way NE. Seattle, WA 98115, pers. commun. December 1985). Yakutat area from long. 144-137°W. and Souttieastern area; data from R. Haight (NWAFC Auke Bay Laboratory, NMFS, NOAA, P.O. Box 210155, Auke Bay, AK 99821, pers. commun. 1986). across statistical areas to calculate an RPN for the Gulf of Alaska. The bootstrap method (Efron 1982; Efron and Gong 1983) then was applied to the re- sultant RPN's to test the statistical significance of annual changes in the RPN's. The bootstrap method is a nonparametric statis- tical procedure based on Monte Carlo methods (see Shreider [1966] for a description of Monte Carlo methods). The bootstrap method is a new technique not common in the fisheries and ecology literature, but examples of its application to survey design and biomass estimation can be found in Kimura and Balsiger (1985) and Haslett and Wear (1985), re- spectively. In addition, Rao and Wu (1984) proved the applicability of the bootstrap method to strati- fied sampling, which is the sampling method used in the longline survey. The bootstrap method is useful when parametric assumptions are difficult to justify; no parametric estimate is readily available for the accuracy of a statistic, e.g., a sample median; or the procedure to compute the statistic of interest is complicated. Simply described, the bootstrap method works as follows: Given the observed data set , the sample is drawn by independent random sampling with replacement from the observed data set and the desired statistic (e.g., a median) is computed from the sample. The resultant statistic is termed the bootstrap replicate. In the next step, the sam- ple is drawn and the bootstrap replicate is computed some large number B times. The resultant B boot- strap replicates form the bootstrap distribution. An estimate of the accuracy of the median, a standard error, then can be calculated from the bootstrap distribution by standard methods. In this study, we used the bootstrap method to test the null hypothesis that the difference RPN^j;. - RPN^jr. = 0, for any RPNjjt , where i = year (1979-86), i' = any subsequent year, and k = sta- tistical area (Shumagin, Chirikof, Kodiak, Yakutat, and Southeastern). The bootstrap method was used because parametric assumptions are difficult to justify for the longline survey data and the pro- cedure to compute the statistic of interest, the variance of the RPN, is tedious and error prone. In our application of the bootstrap method, stations were randomly sampled with replacement within each area. A value denoted RPN^ ;^* was computed from the catch per hachi values of the sampled sta- tions by the method of RPN calculation described previously. Stations then were sampled with re- placement from year i' within each area, a second value denoted RPNj ;^* was computed, and the difference RPNj j^* - RPN^j^* was found. Sampl- ing with replacement from the 2 years and the com- putation of the difference were repeated 1,000 times producing a bootstrap distribution of 1,000 differences. Efron and Tibshirani (1986) outlined three methods for setting an approximate confidence interval from a bootstrap distribution for a statistic of interest, here the difference RPNj /^ - RPN^ /£. Use of the simplest method, the percentile method, is considered correct if the bootstrap distribution of the statistic of interest is described by a normal distribution (Efron and Tibshirani 1986). The nor- mality of the bootstrap distribution for the dif- ference was tested using the D'Agostino D Test (D' Agostino and Stephens 1986) and found to be nor- mal, thus justifying the use of the percentile method. The statistical significance of the difference RPNj /£ - RPNj /£ then was evaluated by the following criteria. If the 95% confidence interval for the difference did not include zero, then the null hypothesis was rejected, the annual change in the RPN was considered statistically significant, and the change in sablefish abundance was considered real. Conversely, if the 95% confidence interval for the difference included zero, the null hypothesis was ac- cepted and the change in sablefish abundance was not considered significant. RESULTS The RPN for the Gulf of Alaska increased 111% from 1979 to 1986 (Fig. 3). The 95% confidence interval for this increase did not include zero and therefore was judged statistically significant (alpha = 0.05; Table 3), showing that the difference was not due to random error in the survey and that sablefish abundance in the Gulf of Alaska has in- creased markedly since 1979. The difference con- sists of significant increases from 1980 to 1981, 1981 448 - KODIAK A. - / / - / / I 1 1 1 1 1 30 1979 80 81 82 83 84 85 86 1979 80 81 82 83 YEAR 84 85 86 Figure 3.— Relative population number (RPN) for the Gulf of Alaska and the Shumagin, Chirikof, Kodiak, Yakutat, and Southeastern areas, 1979-86. Dashed lines (— -) signify that the annual change was statistically significant. to 1982, and 1984 to 1985 (Fig. 3). Differences be- tween other years were not significant. The RPN for each of the statistical areas of the Gulf of Alaska generally increased from 1979 to 1986 (Fig. 3); the differences between 1979 and 1986 were statistically significant for all areas (Table 3), showing that sablefish abundance has increased throughout the Gulf of Alaska. The sharp jump in the RPN for the Gulf of Alaska from 1980 to 1982 was caused by significant increases in four areas, Shumagin, Chirikof, Kodiak, and Yakutat. The sharp jump from 1984 to 1985 was caused by signif- icant increases in two areas, Kodiak and Yakutat. DISCUSSION As noted earlier, sometimes not all strata were sampled at a station (Table 1). This shortcoming was a consequence of the length of the sampling gear and the topography at a station. At stations where the bottom gradient was slight for all or part of the station, the 16 km of longline gear sometimes was not long enough to sample all strata. As a result, sometimes no stations were sampled within certain strata. For these strata, catch per hachi values were estimated using combined data from the current and previous year's data. The resultant values are esti- 449 FISHERY BULLETIN: VOL. 86, NO. 3 Table 3.— Statistical significance of annual changes in relative population number of sablefish, Gulf of Alaska, Japan-U.S. cooperative longline survey, 1979-86. The sym- bols used are defined as follows; + signifies a significant increase in RPN; - signifies a significant decrease in RPN; O signifies no significant change. Area Year 80 81 82 83 84 85 86 Gulf of Alaska 1979 O + + + + + + 1980 + + + + + + 1981 + + + + + 1982 O o + + 1983 o + + 1984 + + 1985 o Shumagin 1979 O o + + + + + 1980 o + + + + + 1981 + + + + + 1982 o o + + 1983 o o o 1984 o o 1985 o Chirikof 1979 o o + + + + + 1980 o + + + + + 1981 + + + + + 1982 o o o + 1983 o o + 1984 o o 1985 o Kodiak 1979 - o + + + + + 1980 o + + + + + 1981 + + + + + 1982 o o + + 1983 o + + 1984 + o 1985 o Yakutat 1979 o + + + + + + 1980 + + o + + + 1981 + o o + + 1982 - - o + 1983 o + + 1984 + + 1985 + Southeastern 1979 o + + o + + + 1980 o o o o + + 1981 o o o o + 1982 o o o + 1983 o o + 1984 o + 1985 o mates of catch per hachi as if the strata had been sampled. Because the estimated values mimic the observed data, the probability of rejecting the null hypothesis is somewhat higher than the nominal value. In general an inaccurate estimate should have a small total effect because catch per hachi values were missing for only 9 of the 320 strata available during the survey (8 years x 5 areas x 9 strata). The only comparison where the estimated catch per hachi values might have caused an incorrect conclu- sion is the 1984-85 comparison where 7 strata were unsampled. This comparison was retested using only the observed catch per hachi values. The annual change, like the comparison using the estimated catch per hachi values, was found to be statistically significant. Another potential effect of incomplete sampling at a station is bias in the bootstrap calculation. Each bootstrap replicate RPN^j^* was computed from a bootstrap sample selected by sampling with replace- ment from the stations within an area. If one or more stations were missing an observation for a strata, then it was possible for the bootstrap sam- ple to have all missing values for the strata. In this case, no catch per hachi values were available to calculate a value for the strata and the value for the strata was treated as zero. This treatment may negatively bias the resultant RPN, ;,*. The extent of the negative bias was tested by comparing the RPN to a bootstrap estimate of the RPN, denoted RPNfc, where RPN^ is the mean of the 1,000 RPNijt* from an area and year. If RPN^ generally was smaller than RPN, then the bootstrap replicates were negatively biased. The percentage difference between the resultant RPNj's compared to the RPN's ranged from - 0.7 to + 0.5 and averaged only -1-0.03, showing that the negative bias had little effect on the RPN,^*. Other fishery and survey data substantiate the significance of the overall increase in the RPN and also the marked increases from 1981 to 1982 and from 1984 to 1985. The CPUE in the Japanese long- line fishery showed a similar pattern to the RPN's for the years the data overlapped, from 1979 to 1983; the fishery CPUE increased from the late 1970's, with the largest increase from 1981 to 1982, and decreased from 1982 to 1983 (Fig. 4). Examination of length compositions for depths 101-200 m indicate that the strong 1977 year class (Sasaki 1982; McFarlane and Beamish 1983; Funk and Bracken 1984) was responsible for the RPN in- crease from 1980 to 1982 and that the RPN increase from 1984 to 1985 was due to a strong 1980 year class. Sablefish recruiting to the survey area first appear at depths 101-200 m, and strong year classes are more distinguishable in the length compositions at these depths. The length compositions for depths 101-200 m show the initial appearance and subse- quent increase in length of fish of the strong 1977 year class (Fig. 5). The mode at 47 cm FL in 1979 indicates the first year that the 1977 year class was available to the survey gear. The rightward move- ment of the mode in succeeding years illustrates the increase in fish length for the 1977 year class. The movement of the mode slowed in 1982 which we in- terpret to be due mainly to the movement of larger 450 SIGLER and FUJIOKA: VARIABILITY IN SABLEFISH 460 420 ■o 380 a (0 §340 JC 300 260 220 Q. 180 - \ \ FISHERY CPUE / v.. J 1977 78 Figure 4.— Sablefish catch per hachi (kg/10 hachis) for the Japanese longline fishery, 1977-83 (Fujioka 1986) and relative population number (RPN), 1979-86. Figure 5.— Length-frequency distributions of sablefish in the Gulf of Alaska for depths 101-200 m, 1979-86. The shaded bands illustrate the progression of the modes of two strong year classes, 1977 and 1980. T r 55 63 LENGTH (cm) r 71 T 79 451 FISHERY BULLETIN: VOL. 86, NO. 3 fish to depths >200 m. The timing of this movement corresponds to the year that the RPN significantly increased. The pattern of length compositions from 1983 to 1985 parallels those due to the strong 1977 year class and suggests that the 1980 year class is also strong (Fig. 5). The shoulder at 49-51 cm FL in 1983, not present in 1982, indicates the first year that the 1980 year class was available to the survey gear. This shoulder is due to distinct modes at 47-51 cm FL in the Chirikof, Yakutat, and Southeastern areas. The rightward movement of the mode in suc- ceeding years illustrates the increase in fish length for the 1980 year class. The mode at 53 cm FL in 1984 is similar to the modes at 50-57 cm FL found in the Gulf of Alaska for depths 101-200 m during a trawl survey conducted in 1984 by the Northwest and Alaska Fisheries Center (Brown 1986) and is further evidence for a strong 1980 year class. The rightward movement of the mode slowed in 1985, presumably due to movement of larger fish to depths >200 m, and again corresponds to the year of a significant increase in RPN. ACKNOWLEDGMENTS Thanks to Evan Haynes and Leslie Williams for thoroughly editing the manuscript, Jerry Fella for statistical advice, and Takashi Sasaki for courteously providing cooperative survey data. LITERATURE CITED Brown, E. 1986. Preliminary results of the 1984 U.S. -Japan cooperative bottom trawl survey of the central and western Gulf of Alaska. In R. L. Major (editor), Condition of groundfish resources of the Gulf of Alaska region as assessed in 1984, p. 259-296. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-106. D'Agostino, R. B., and M. A. Stephens. 1986. Goodness-of-fit techniques. Marcel Dekker, N.Y. Efron, B. 1982. The jackknLfe, the bootstrap and other resampling plans. Appl. Math. 38, 92 p. Soc. Ind. Appl. Math., Phila. Efron, B., and G. Gong. 1983. A leisurely look at the bootstrap, the jackknife, and cross-validation. Am. Stat. 37:36-48. Efron, B., and R. Tibshirani. 1986. Bootstrap methods for standard errors, confidence in- tervals, and other measures of statistical accuracy. Stat. Sci. 1:54-77. FUJIOKA, J. 1986. Sablefish. In R. L. Major (editor). Condition of groundfish resources of the Gulf of Alaska region as assessed in 1985, p. 79-108. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-106. Funk, F., and B. E. Bracken. 1984. Status of the Gulf of Alaska sablefish (Ancyplopoma fim- bria) resource in 1983. Alaska Dep. Fish Game, Info. Leafl. 235, 55 p. Haslett, S. J., AND R. G. Wear. 1985. Biomass estimation of Artemia at Lake Grassmere, Marlborough, New Zealand. Aust. J. Mar. Freshwater Res. 36:537-557. KiMURA, D. K., AND J. W. BALSIGER. 1985. Bootstrap methods for evaluating sablefish pot index surveys. N. Am. J. Fish. Manage. 5:47-56. McFarlane, G. a., AND R. J. Beamish. 1983. Overview of the fishery and management strategy for sablefish (Ancrpkrpoma fimbria) off the west coast of Canada. In Proceedings of the International Sablefish Symposium, p. 13-35. Alaska Sea Grant Rep. 83-8, Univ. Alaska, Fairbanks. Rao, J. N. K., and C. F. J. Wu. 1984. Bootstrap inference for sample surveys, /n J. E. Gentle (editor). Proceedings of the section on survey research methods, p. 106-112. Am. Stat. Assoc. Sasaki, T. 1982. Condition of sablefish stocks in the North Pacific. Far Seas Fish. Lab. Fish. Agency Jpn., Shimizu, 17 p. (Document submitted to the International North Pacific Fisheries Commission.) 1985. Studies on the sablefish resources in the North Pacific Ocean. Far Seas Fish. Res. Lab., Bull. 22, 107 p. Fisheries Agency of Japan, Shimizu. 1986. Stock assessment of sablefish in the eastern Bering Sea, Aleutian Islands region, and the Gulf of Alaska. Far Seas Fish. Res. Lab., Fish. Agency Jpn., Shimizu, 33 p. Sasaki, T., D. Rodman, and K. Funato. 1983. Preliminary report on Japan-U.S. joint longline survey by Ryusho maru No. 15 in the eastern Bering Sea, Aleutian region and Gulf of Alaska, 1982. Far Seas Fish. Res. Lab., Fish. Agency Jpn., Shimizu, 116 p. Shreider, Y. a. 1966. The Monte Carlo method. Pergammon Press, N.Y. 452 PELAGIC BIOGEOGRAPHY OF THE ARMORHEAD, PSEUDOPENTACEROS WHEELERI, AND RECRUITMENT TO ISOLATED SEAMOUNTS IN THE NORTH PACIFIC OCEAN George W. Boehlerti and Takashi Sasaki^ ABSTRACT The pelagic armorhead, Pseudopentaceros wheeleri, occurs widely in the North Pacific Ocean. Benthic specimens have been taken from Japan, the Hawaiian Archipelago, and the west coast of North America, but the main reproductive populations are located on southern Emperor-northern Hawaiian Ridge sea- mounts between lat. 29° and 35°N. The period between spawning and recruitment to the seamounts is apparently between 1.5 and 2.5 years, suggesting an extended pelagic existence. We describe the distributional patterns in the North Pacific based upon over 30 years of published and unpublished records. The majority of pelagic specimens are captured in the subarctic water mass in the northeast Pacific. Based upon the distributional patterns and the oceanography of the North Pacific, we propose migratory paths for both the main population and for the individuals that occur rarely in other locations. The long pelagic period and variability in ocean conditions may play an important role in recruitment to seamounts and the variability in year-class strength for this species. The pelagic armorhead, Pseudopentaceros wheeleri, is a member of the boarfish family Pentacerotidae. Until recently, considerable confusion existed about its taxonomy and distribution. Originally described as Pentaceros richardsoni Smith 1844, it was thought to be distributed virtually worldwide, with centers of abundance in the North Pacific and in the South Atlantic near South Africa (Fujii 1986). Oc- currences were typically sporadic; most records were from pelagic captures, hence the name pelagic armorhead. Several authors, including Welander et al. (1957), Clemens and Wilby (1961), and Wagner and Bond (1961), used the name Pseudopentaceros richardsoni, but more recent studies used the original binomial following Follett and Dempster (1963). The 1967 discovery of large concentrations of armorhead in the mid-Pacific, over southern Emperor-northern Hawaiian Ridge (SE-NHR) seamounts (from lat. 29° to 35°N) (Sasaki 1974; Uchida and Tagami 1984) stimulated increased interest in this species. Over the next 10 years, nearly 1 million metric tons (t) were taken by Soviet and Japanese trawlers (Borets 1975; Takahashi and Sasaki 1977). The increased availability of specimens allowed Hardy (1983) to revise the family. He first 'Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 96822-2396. ^Far Seas Fisheries Research Laboratory, 5-7-1 Orido, Shimizu, Shizuoka 424, Japan. separated the Pentaceros richardsoni complex from other members of the genus by referring it to Pseudopentaceros Bleeker, based on morphological considerations. Second, he separated the complex into three species: P. richardsoni confined to the Southern Hemisphere, and P. wheeleri and P. pectoralis in the North Pacific. The features used to distinguish P. wheeleri and P. pectoralis were largely morphological. Pseudopentaceros pectoralis is more robust and deeper bodied and its dis- tribution is typically pelagic; P. wheeleri was known only from benthic specimens on sea- mounts. Hardy (1983), however, lacked transitional specimens; subsequent work based on morpho- logical and electrophoretic grounds revealed that P. wheeleri was a morphological derivative of P. pectoralis, with changes occurring after settle- ment. The deeper bodied pelagic form ceases growth in length after recruiting to the seamounts and instead transforms body shape to a leaner form. "Fat" and "lean" specimens of the same length thus differ significantly in appearance (Humphreys et al. in press). Pseudopentaceros wheeleri has an unusual life history. Spawning occurs at the benthic population centers at the SE-NHR seamounts between Novem- ber and March (Sasaki 1974; Borets 1976; Bilim et al. 1978). Larvae and juveniles are pelagic (Honma and Mizusawa 1969; Fedosova and Komrokov 1975; Borets and Sokolovsky 1978) and widely distributed in the North Pacific. Recruitment to the seamounts Manuscript accepted April 1988. Fishery Bulletin; Vol. 86, No. 3, 1988. 453 FISHERY BULLETIN: VOL. 86, NO. 3 occurs between ages 1.5 and 2.5 years, but in the major spawning population centers, over 95% of the fish are <2 years (Uchiyama and Sampaga^). This lack of older specimens and presence of fish in poor physiological condition has led to suggestions that this species is semelparous, with death after a single spawning (Humphreys and Tagami 1986; Hum- phreys et al. in press). The long interval between spawning and recruitment for this species suggests that oceanographic conditions in the North Pacific regulate its distribution (Boehlert 1986). Published records of the pelagic distribution of P. wheeleri are too few and varied to understand its biogeography. In different years, specimens have been caught in the Gulf of Alaska, near the Aleu- tian Islands, and off the coasts of Japan, Oregon, California, and British Columbia. In this paper, we document the pelagic and benthic occurrences from 'Uchiyama, J. H., and J. Sampaga. In review. Age and growth of the pelagic armorhead, Pseudopentaceros wheeleri, at Hancock Seamounts. Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 96822-2396. both published and unpublished accounts, consider this distributional pattern in light of general ocean- ographic conditions in the North Pacific, propose hypotheses about the migration and recruitment of armorhead to the seamounts of the central North Pacific, and assess interannual variation in abun- dance and how it may relate to oceanographic variability. MATERIALS The data sources used include published and un- published collection records from throughout the North Pacific Ocean. Published records on occur- rences in Japan (Abe 1957, 1969; Zama et al. 1977; Okamura et al. 1982), the eastern North Pacific (Welander et al. 1957; Larkins 1964; Honma and Mizusawa 1969; Chikuni 1970; Ignell et al. 1986), and the west coast of North America (Wagner and Bond 1961; Follett and Dempster 1963; Smith 1965) were typically of single or a few specimens. Larger numbers of unpublished collection records were in data from large-scale surveys (Table 1). These data Table 1 .—Data bases searched for pelagic armorhead, Pseudopentaceros wheeleri, captures excluding single collections and literature reports. Areas and seasons are averages over the years covered. Years, season effective Area Data source Lat. N Long. Salmon drift gill-net surveys^ 1972-86, spring-fall 38°-65<' 140°E-166°W University of Washington (FRI) purse seining, longlining^ Canada salmon longlining^ 1956-82, spring-summer 42°-60° 165°E-125°W 1962-67, spring-summer 42°-54° 120°-145°W Bureau of Commercial Fisheries salmon gill net^'^ 1955-71, spring-summer 40°-60° 165'»E-125°W Oshoro Maru salmon 1963-66, 1968, 1969, 1971, gill-net-longline surveys'* 1974-86, spring-summer 36°-66° 163°E-124°W Hokusei Maru squid and salmon gill-net surveys'* 1973-86, summer 18°-58° 141°E-156°W Japan Marine Fishery Resource Research Center pomfret surveys* 1979-82, all months 22°-49° 143°E-124°W Japanese commercial whaling reports* 1952-79, summer 40°-58° Pacific Biological Station 1983, 1985-86, gill-net surveys'^ spring-summer 47°-54° 138«'-130<'W Taiwan Fisheries Research Institute squid surveys® 1986, July-August 38°-46° 153°E-175°W National Marine Fisheries Service Auke Bay cooperative squid gill-net surveys^ 1985-86 43°-55° 145°-175°W Japan Fisheries Agency squid gill-net surveys 1986-87, July-August 36°-47° 150°-165°W 'Salmon research data file. North Pacific Ocean, 1972-86. Salmon Division, Far Seas Fisheries Research Latxjratory, Shimizu, Japan. 2t^acy et al 1978 sLarklns 1964. ♦Data record of oceanographic observations and exploratory fishing. No. 1 (1957)-No. 28 (1985) Faculty of Fisheries, Hokkaido University. sjapan Marine Fishery Resourse Research Center 1980, 1983a, 1983b, 1985 ^Biological data file of whales In the North Pacific Ocean, 1952-79. Whale Section, Far Seas Fisheries Research Laboratory, Shimizu. Japan. See also ChikunI (1970). 'Sloan 1983; Robinson and Jamieson 1984. 'Z. Shyu, Taiwan Fisheries Research Institute, Keelung 20220, Taiwan, pers. commun. May 1987. ^Ignell et al. 1986. 454 BOEHLERT and SASAKI: PELAGIC BIOGEOGRAPHY OF THE ARMORHEAD are seasonally limited to primarily spring, summer, and early fall months, but cover 1952-87 and pro- vide broad geographic coverage between lat. 18° and 66 °N. Greater detail and a summary of many of these surveys are provided by Macy et al. (1978). Since armorhead were an incidental catch, it is dif- ficult to assess the relative value of the different surveys with respect to sampling effort. The most typical sampling gears used were long- lines, gill nets, and purse seines targeting salmon, squid, or pomfret; rarer collections were made with plankton nets, dip nets, or hook and line. Thus, sampling was typically in surface waters, most likely in the upper 50 m; the true depth distribution of the pelagic animals, however, is unknown. The only other "gear" was the sei whale, Balaenoptera bore- alis, from which stomach contents were studied for about 27 years; Chikuni (1970) reported armorhead in the stomachs. RESULTS AND DISCUSSION Distributional Patterns The known center of distribution and spawning for benthic P. wheeleri is in the SE-NHR seamount region, bounded by lat. 29°-35°N and long. 171°E- 179°W (Takahashi and Sasaki 1977; Humphreys and Tagami 1986). These reproductive fish are typical- ly between 23.0 and 28.5 standard length (SL) (Sasaki 1986) and range in age from 1.5 to 2.5 years (Uchiyama and Sampaga fn. 3). They are found at depths between 200 and 500 m (Takahashi and Sasaki 1977). Other benthic occurrences have been recorded but are sporadic (Fig. 1). Several speci- mens were captured off Japan in 1957 (Abe 1957), 1969, 1971-73, 1976 (Zama et al. 1977), and 1979 (Okamura et al. 1982). These fish were typically the same size as those captured on the seamounts, but no mention of reproductive status was made. Fish of similar length occur rarely on the west coast of North America, but only between lat. 37°00' and 44°25'N (Wagner and Bond 1961; Follett and Demp- ster 1963; Smith 1965). Again, reproductive status is unknown for these specimens. No benthic occur- rences off British Columbia or Alaska have been reported, despite extensive trawling surveys which included seamounts (Hughes 1981; Alton 1986). Large specimens have been taken rarely in the Hawaiian Archipelago (Fig. 1; Randall 1980; Humphreys et al. in press) and are typically 4-5 years old and appear to be in reproductive condi- 120' 150 Figure 1.— Known distribution of benthic occurrences of pelagic armorhead, Pseudopentaceros wheeleri. Squares represent the main population centers at central North Pacific seamounts. All others are rare or sporadic occurrences. 455 FISHERY BULLETIN: VOL. 86, NO. 3 tion. Their relationship with the SE-NHR seamount populations is unknown. The most common pelagic occurrences (1955-87) were fish ranging from 18 to 26 cm SL; smaller sizes are poorly represented, probably because of the nature of the sampling gear. Smaller specimens were generally captured with plankton nets (Fedo- sova and Komrakov 1975; Borets and Sokolovsky 1978), dip nets (Honma and Mizusawa 1969; Ran- dall 1980; Fujii'*), and in one case, in a whale stomach (Kawamura 1982). The sea-surface temperatures at which the specimens were captured ranged from 8.6° to 15.0°C (average 12.1°C). For all years, pelagic captures were restricted to the eastern North Pacific, with the exceptions of larvae or early juveniles captured near the spawning centers at the seamounts (Komrakov 1970; Table 2), over 20 specimens in a Bryde's whale stomach in 1979 (Kato^, Fig. 2B), and two individuals captured in 1986 at long. 155°E (Fig. 2). This is despite exten- sive sampling in the western North Pacific, par- ticularly by the Japan Marine Fishery Resource Research Center (JAMARC) and the Hokkaido University, and in the Bering or Okhotsk Seas (Table 1). Pelagic captures in the 1950's were sporadic, but several fish were taken north of lat. 50°N, especially in 1958 (Fig. 2 A). Relatively few were captured in the 1960's, with the exception of July through early August 1969, when the armorhead was an impor- *E. Fujii, Nippon Luther Shingaku Daigaku, Tokyo, Japan, pers. commun. July 1987. ^H. Kato, Whales Research Institute, Tokyo, Japan, pers. commun. December 1987. Table 2. — Pelagic specimens of armorhead, Pseudopentaceros wheeleri, for which sizes were available. The growth curve from Uchiyama and Sampaga (text footnote 3), based upon enumeration of daily growth increments, was used to deduce age from length. For the youngest fish (bottom), average readings from daily growth increments, rather than the growth curve, were used. Data are from various published sources and collections listed in Table 1 . (SST = sea-surface temperature.) Standard Lat. length Year Date N Long. SST (mm) Age (yr) No. Gear 1955 9/27 45°41' 165°5'W 108-122 0.38-0.42 2 Gill net 1955 9/27 45°35' 165°5'W — 200 0.79 Dip net 1956 7/31 45°49' 160°3'W 12.2 245 1.29 Gill net 1956 9/15 45°00' 155°00'W 12.1 245 1.29 2 Gill net 1956 9/17 49°00' 150°00'W 13.2 245 1.29 Gill net 1956 9/17 51°00' 150°00'W 11.0 245 1.29 Gill net 1958 8/14 44°39' 174°48'W — 64-91 0.26-0.33 2 Dip net 1958 8/25 49°43' 146°10'W 11.8 256 1.50 Gill net 1963 8/24 49°0' 162°0W 12.9 249 1.40 Gill net 1967 6/15 54°38' 150°1'W 8.8 121-134 0.42-0.46 4 Dip net 1968 7/ 41°30' 165°30'W — 285 >2.00 1 Handline 1969 4/1 29°48' 179°5'E — 8-23 0.08-0.25 ? Plankton 1969 4/1 31 °6' 176°0'E — 8-23 0.08-0.25 — Plankton 1969 4/1 32°6' 173°0'E — 8-23 0.08-0.25 — Plankton 1972 5/21 38°30' 175°3W 14.3 42-50 0.21-0.23 5 Whale stomach 1984 7/25 45°30' 155°0'W 13.0 262-265 1.81-2.02 3 Gill net 1984 7/25 45°26' 154°58'W 12.0 278 >2.00 2 Longline 1985 7/9 46°57' 129°4W — 272 >2.00 1 Gill net 1985 7/10 47°30' 128°36W — 272-298 >2.00 2 Gill net 1985 7/11 47°59' 129°19'W — 277 >2.00 1 Gill net 1985 7/13 46°23' 131°29W — 188-219 0.71-0.94 5 Gill net 1985 7/14 46°40' 131°8W — 183-233 0.68-1.1 9 Gill net 1985 7/15 46°35' 131°5W — 206 0.83 1 Gill net 1985 7/19 46°54' 131°28'W — 190-217 0.72-0.92 2 Gill net 1985 7/21 47°18' 130°23'W — 191-220 0.73-0.95 3 Gill net 1985 7/22 47°35' 130°48'W — 190-201 0.72-0.79 2 Gill net 1985 7/23 47°37' 130°3rW — 198-200 0.77-0.79 2 Gill net 1985 7/12 47°0' 155°1'W 11.6 263-265 1.87-2.02 2 Longline 1985 7/13 45°28' 155°5W 11.8 238-262 1.17-1.81 12 Longline 1985 7/11 47'='0' 155°0'W 11.6 248-262 1.35-1.81 5 Gill net 1985 7/12 45°30' 155°0W 11.8 235-240 1.12-1.20 4 Gill net 1985 7/13 44°0' 155°0'W 12.8 248-258 1.35-1.63 4 Gill net 1985 2/23 30° 16' 181°18W — 8-15 0.08-0.16 15 Neuston 1985 2/24 29°27' 180°56'W — 5-15 0.04-0.16 15 Neuston 1986 8/3 43°51' 164°54'W 13.0 50-100 0.23-0.35 5 Dip net 456 BOEHLERT and SASAKI: PELAGIC BIOGEOGRAPHY OF THE ARMORHEAD 150° 120' Figure 2.— Records of pelagic occurrences of armorhead, Pseudopentaceros wheeleri, in the North Pacific. Most data were obtained fronn either published records or unpublished data described in Table 1: Each symbol represents a collection taking armorhead, rather than single specimens. (A) 1955-69. Here, note that the area in the box contains the locations of 195 sei whales captured in July-August 1969 that had fed extensively upon armorhead. The samples from the central Pacific were larvae and early juveniles (Fedosova and Komrakov 1975). (B) 1970-87. 457 FISHERY BULLETIN: VOL. 86, NO. 3 tant diet component of sei whales taken by the Japanese commercial fleet. This species normally feeds on zooplankton and micronekton (Kawamura 1982), but in this year, 195 whales were recorded to have fed extensively on armorhead. Although data records for whale stomach contents extend from 1952 to 1979, 1969 was the only year when large numbers fed on armorhead, with the excep- tion of two occurrences each in 1971 and 1972. Based on the known feeding mode of sei whales, Chikuni (1970) suggested that the armorhead ap- parently schools in surface waters. In the 1970's, catches were relatively small, but in the 1980's, specimens have been captured in each year, with the greatest catches in 1985. In that year, fish occurred in more easterly areas as compared to other years. Relationship of Distribution Patterns to North Pacific Oceanography The SE-NHR seamounts are located in the mid- Pacific transition zone, between subarctic and sub- tropical water masses (Roden 1970). The spawning season of armorhead typically lasts from late November through March (Sasaki 1974; Bilim et al. 1978), and larvae are neustonic, at least through the first few months of life (G. W. Boehlert unpubl. data). Near surface drift in this region is largely wind driven (McNally 1981); as shown by Lagran- gian drifters, mean surface flow in the SE-NHR sea- mount region during winter months is typically eastward (Kirwan et al. 1978; Emery et al. 1985) or southeastward (McNally et al. 1983), and long- term mean Ekman transport has a southeastward component (Favorite et al. 1976). Based on mean ship-drift data, progressive vector diagrams over a 3-mo period, with starting points at four SE-NHR seamounts (Bakun^), show this typical eastward flow with a southward component (Fig. 3). If pelagic armorhead followed such a pattern, individuals would remain in the subtropical gyre, continuing eastward and then turning south, entering the California Current system. The scarcity of speci- mens in this region (Figs. 1, 2) suggests instead that most fish move northeastward in some manner, entering subarctic waters and residing within the Alaska gyre; surface drift in winter is northeastward in anomalous years (McNally 1981), and average 'Compilation by the NMFS Pacific Fisheries Environmental Group from historic ship drift files assembled by the U.S. Naval Oceanographic Office (A. Bakun, Pacific Fisheries Environmen- tal Group, National Marine Fisheries Service, NOAA, P.O. Box 831, Monterey, CA 93942, pers. commun. April 1987. 36°N 35°N 30*N 25°N 17rE 175°E 180' 178°W Figure 3.— Progressive vector diagrams of mean surface currents (starting date, 15 December; duration, 90 days), from long-term mean ship drift data from 1° squares, with distances calculated on a daily basis. Triangles indicate starting locations at four seamounts with armorhead spawning populations. Each mark on the vector between the origin and arrow represents a 5-d period. See text footnote 6. 458 BOEHLERT and SASAKI; PELAGIC BIOGEOGRAPHY OF THE ARMORHEAD surface flow is in this direction in summer months (McNally et al. 1983). The two available studies on larval distribution of armorhead are conflicting and suggest that patterns of movement may differ from year to year. Komra- kov (1970) conducted surveys in March-April 1969 and observed larvae only at northern seamounts and in the open ocean several hundred kilometers north- east of the SE-NHR seamounts. Borets (1979), how- ever, observed 5-20 mm larvae remaining mostly around the seamounts, with highest abundance south of lat. 33°N in 1976. Because spawning oc- curs at seamounts northwest of this area, this pat- tern is consistent with southeastward drift during the first few months of life (Fig. 3). If larvae remain in the seamount region, they could be transported to the northeast in summer months, or they may actively migrate northeastward after reaching the juvenile stage. A similar migration apparently occurs for the pomfret, Bramajaponica, which prior to spawning moves into the North Pacific current region in a broad latitudinal band between the sub- tropical convergence and the subarctic boundary; immature fish and spent adults then move north- ward into the subarctic region to feed during April- May (Shimazaki and Nakamura 1981). The pomfret differs from the armorhead, however, in that it remains pelagic throughout its life and also is abun- dant in the western North Pacific. Larval and juvenile Brama sp. do occur in the same area as lar- val armorhead (Borets and Sokolovsky 1978; G. W. Boehlert unpubl. data). A possible scenario of movements after the lar- val stage can be deduced from the ages of individuals captured in different regions. Based upon the growth of pelagic armorhead (Uchiyama and Sam- paga fn. 3), most of the larger specimens indicated in Figure 2 are from 1 to 2 years old, and the pelagic duration for this species is most often from 1.5 to 2.5 years. Using daily growth increments for lar- val and small juveniles and fitted growth curves for larger juveniles and adults, one can convert lengths to ages. The pelagic specimens of known length range in estimated age from 0.04 to >2 years (Table 2). The youngest animals are typically found in the region of the seamounts (Komrakov 1970; Fedosova and Komrakov 1975; Borets and Sokolovsky 1978), but intermediate-sized fish are found north and east of the seamounts. Different age groups seem to be distributed in different areas in the eastern North Pacific (Fig. 4). Armorhead spawn in the seamount region (as area A in Figure 4) during November to March, and juveniles 5-25 mm long are found in the SE-NHR seamounts area from February to early April (Fedosova and Komrakov 1975; Borets 1979). Subsequent occurrences of fish older than 0.25 year (Table 2) are found only in the region northeast of the seamounts and in the Gulf of Alaska (area B in Figure 4). Northeastward movements are contrary to mean ocean currents, but Roden et al. (1982) described intense northward flow along the axis of the southern Emperor Seamounts that could con- ceivably play a role in transport of fish at the young- est stages. Recruitment of later stage armorhead to the sea- mounts requires movement to the west. The only significant westward circulation in the North Pacific is found in the Alaskan Stream (Favorite 1967). This narrow current occurs north of lat. 51 °N and ex- tends westerly to near long. 170 °E, where it splits to two branches, one turning northerly into the Bering Sea and the other southerly under the influ- ence of the northern Emperor Seamounts or Kor- mandorskie Ridge (Favorite 1967; dashed line in Figure 4). Damitsky et al. (1984) described southerly movement of this water to at least lat. 40 °N and suggested that it played a role in transport of planktonic food for armorhead on the seamounts. Unless returning armorhead are deep in the water column and therefore not sampled, distributional data (Fig. 2) do not support this route for westerly movement. Fish near recruitment size and age are instead captured in the region marked C on Figure 4, south of the area where younger fish are captured. These fish apparently migrate back to the region of the seamounts by some unknown, but probably active, mechanism. Two less important, but plausible, movement patterns may explain the rarer occurrences in south- ern Japan, off Oregon and California, and in the Hawaiian Archipelago (Fig. 4). Chelton (1984) dis- cussed possible interannual changes in the lati- tudinal position of the West Wind Drift which may lead to differing magnitudes of transport in the Alaska and California Currents. The rare benthic occurrences of the pelagic armorhead off the west coast of North America (area D in Figure 4) may come from specimens that drift southward in the California Current in years of greater southern transport. Some of these may not settle out and re- main within the subtropical gyre, possibly recruit- ing to the Hawaiian chain far to the south of the normal reproductive population. This route, because of its distance, apparently takes considerably longer than that in the subarctic gyre; based upon satellite drifters, McNally et al. (1983) suggested that a full circuit of the subtropical gyre takes 4.5 years. The armorhead collected in the Northwestern Hawaiian 459 FISHERY BULLETIN: VOL. 86, NO. 3 Figure 4.— Distributions of benthic and pelagic specimens of armorhead, Pseudopentaceros wheeleri, by age group with possible patterns of movement. Area A represents the major spawning area, and area E the region inhabited by the large, older specimens. Area B is the region in which specimens younger than 1 year have been captured, and area C where fish from 1 to 2 years old have been captured. The arrow through this region represents the main population movement pattern. Other arrows represent possible movements for stray fish. Areas marked D and F represent locations of benthic collections from California and Japan, respectively. Islands (area E in Figure 4) are typically larger than those in the normal part of the range and are corre- spondingly older, with ages estimated at 4 or 5 years (Humphreys et al. in press), similar to the time sug- gested for a circuit of the gyre by McNally et al. (1983). The other plausible movement pattern can be used to explain the existence of specimens in Japan. Ben- thic specimens of armorhead are rarely captured in Japan (Fig. 1); a report of 3,000 t landed in 1969 by Soviet fishermen in this region (Abe 1969) may be in error, since the Soviet fishery on the SE-NHR seamounts developed heavily that year (Borets 1975) and SE-NHR fish were likely sold in Japan. The reproductive condition of these animals is not known, so the source of the spawning population supplying them is at question. The large, reproduc- tive individuals in area E (Fig. 4) have been found as far south as lat. 23°43'N (Humphreys et al. in press). It is possible that larvae and juveniles are transported to Japan in the northern part of the North Equatorial Current (Uda and Hasunuma 1969). This is consistent with the capture of over 20 fish in a Bryde's whale stomach in September 1979 at lat. 26°N (H. Kato fn. 5; Fig. 2B). Subse- quent captures of this species have not been made in Japanese waters, where their presence may re- quire a rare recruitment event from an upstream population source, as has been suggested for sea- mount populations of lobsters (Lutjeharms and Heydorn 1981). These animals may also remain within a gyral circulation, however, as suggested by the presence of two specimens captured in the Kuroshio in 1986 (Fig. 2B). Recruitment to the Seamounts The seasonal timing of the recruitment of the pelagic armorhead to the SE-NHR seamounts is unknown. That no records of larger pelagic fish cap- tured near the seamounts (Fig. 2) exist may be an artifact of sampling in inappropriate seasons or lack of sampling in deeper water. The youngest benthic specimens from the seamounts were slightly in ex- 460 BOEHLERT and SASAKI: PELAGIC BIOGEOGRAPHY OF THE ARMORHEAD cess of 1.5 years in early summer (Uchiyama and Sampaga fn. 3) and 2 years in winter, suggesting that some recruitment occurs in spring. In most fisheries, temporal recruitment patterns can be discerned from length-frequency analyses, because the smaller recruits represent a larger pro- portion of the population during the season of recruitment. For the armorhead, no increase in smaller size classes is apparent in monthly length- frequency samples. This may be due to cessation of somatic growth after recruitment to the seamount (Humphreys and Tagami 1986; Uchiyama and Sam- paga fn. 3). Early recruits, with their deeper bodies and greater fat content, differ morphologically from longer seamount residents. The transition from so-called "fat" to "lean" morphotypes occurs in association with development of gonads (Humphreys et al. in press). Thus, an index of morphological change can be used to detect recruitment patterns, much the same as length frequencies can be used for other species. Condition factor is frequently used to assess "fatness" of fish; it is normally expressed as weight divided by length to a power (typically 3) multiplied by some scaling factor (Ricker 1975). For armorhead, there is a marked change in condition factor with morphological change; newly recruiting fish have high condition factors relative to inter- mediate or lean fish. By considering the monthly proportion of fish above an arbitrary value of con- dition factor, we can estimate the influx of new recruits to the seamounts. During May 1972- December 1973, the major influx of new recruits with a high condition factor occurred in April-May 1973 (Fig. 5). The data for 1972 are incomplete, but some recruitment apparently occurred in August and September (Fig. 5). Differences in seasonality of recruitment between 1972 and 1973 may be indicative of interannual variability in temporal recruitment patterns. An influx of fish to the seamounts may also be reflected in the catch per unit effort (CPUE), an index of stock abundance. Monthly averages of CPUE (in metric tons per hour) by Japanese trawlers for armorhead at the SE-NHR seamounts were highest in March and April (Sasaki 1986). Overall armorhead CPUE decreased continuously from 54.1 t/h in 1972 to only 0.3 t/h in 1982 (Sasaki 1986) and further decreased to 0.06-0.07 t/h in 1983-85 (T. Sasaki unpubl. data). The estimated CPUE in 1986, however, was 0.31 t/h, an increase of four or five times that in the previous year. The CPUE of trawlers operating in May-Septem.ber 1986 was highest in May, decreased in June and July, but increased slightly in August and Septem- 40 35- 30- H 25 z HI O 20- oc UJ °- 15H 10 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 1972 MONTH 1973 Figure 5.— Seasonality of the proportion of armorhead, Pseudopentaceros wheeleri, with condition factors >2.0 from four seamounts of the northern Hawaiian Ridge (Southeast Hancock, Northwest Hancock, C-H, and Colahan) for 1972-73. Squares, males; triangles, females. Data are based upon 2,104 males and 1,704 females. 461 FISHERY BULLETIN: VOL. 86, NO. 3 ber. Based on all the information, the major recruit- ment may occur in spring with some additional recruitment in August and September (Fig. 5); data are not sufficient to determine the recruitment pattern in detail. A broad geographic separation exists between the location of pelagic captures of larger armorhead and the location of the spawning populations (Figs. 2, 4). If recruitment to the seamounts occurs predom- inantly in spring (as suggested for 1973 in Figure 5), then temporal sampling patterns may have missed these fish, although JAMARC pomfret surveys covered this area in some seasons (Table 1). Locating the seamounts, which have small (2-5 km) summits must be a formidable task given the wide ocean areas over which armorhead are distributed. A similar situation exists for rock lobster, Jasus tristani, in the South Atlantic, that recruits from an upstream population some 2,000 km away and in sufficient numbers to support a fishery on Vema Seamount in some years (Lutjeharms and Heydorn 1981). Open-ocean migrations of fishes may depend upon many potential cues, including electric fields (McCleave and Power 1978), magnetic fields (Walker 1984), gyres (Williams 1972), and phero- mones (Nordeng 1977). Certain characteristics of these isolated, open-ocean seamounts may promote their detection by armorhead. First, current-topog- raphy interactions may create significant signals in physical and biological features. The region of the SE-NHR seamounts is active in front development (Roden and Paskausky 1978); upwelling, eddies, and other aspects of flow complexity also occur around these seamounts (Roden et al. 1982) and down- stream from them (Royer 1978). The biological sig- nals may include increased chlorophyll in response to upwelling or doming of isotherms (Genin and Boehlert 1985), or aggregations of various organ- isms and the larger animals which prey upon them around seamounts (see review in Boehlert and Genin 1987). Gravity anomalies associated with seamounts may also play a role; positive gravity anomalies exist at the summit and slopes, and negative anomalies are seen in the surrounding "moat" regions (Wedge- worth and Kellogg 1987). Seamounts often have strong magnetic dipoles associated with them, and the dipole might serve as a landmark for magnetic orientation by fish (Klimley^). While fish have been shown to have magnetoreceptors (Walker et al. 1985), their use of magnetic maps remains specula- T. Klimley, Scripps Institution of Oceanography, La JoUa, CA 92038, pers. commun. 30 June 1987. tive but possible (Gould 1985). Although we cannot postulate the mechanism that armorhead use for recruitment, it is clear that the effects of seamounts may be detected at distances greater than their area alone would suggest. Interannual Variations in Recruitment Strength The year-class strength of armorhead recruiting to seamounts appears to be independent of the parent stock size (Wetherall and Yong 1986). As an example, Borets (1975) estimated that on the SE- NHR seamounts from 1968 to 1973, the stock size varied by a factor of <1.8 while recruitment varied by >5.5 times. The relative abundance of armorhead at the SE-NHR seamounts area increased in 1986 after a long period at a very low level. This increase probably corresponds to the high abundance of pelagic specimens captured in the northeastern Pacific in 1985 (Fig. 2B; Table 2) that consisted of two age groups. The increased recruitment at the SE-NHR seamounts in 1986 suggests that environ- mental conditions were favorable to the survival of young armorhead in the 1984 and 1985 winter seasons. A wide variety of factors, both biotic and physical, can affect survival and ultimate year-class strength in fishes (Lasker 1978). During the 2 years between spawning and recruitment for armorhead, an ex- tended migration through varied pelagic environ- ments occurs (Fig. 4). Feeding conditions for larval and juvenile stages are characterized by interannual variability; Fedosova (1980) suggested that warm years were more productive for zooplankton prey and, thus, favorable to the survival of young armor- head. Interannual variation in atmospheric systems (Seckel 1988) or large-scale ocean currents of the kind described by Mysak et al. (1982) may also play a role in armorhead recruitment strength. Changes in the position of the Alaska gyre by up to 700 km southwest of its normal position may have occurred from 1981 to 1985, with an associated increase in seawater temperature (Royer and Emery 1987). Large-scale atmospheric phenomena, such as the longitudinal position of the Aleutian Low, may create definite interannual variations in winter wind systems that may be seen in surface current pat- terns (Seckel 1988). These patterns may, in turn, be related to the latitudinal position of the subtrop- ical front, which varies interannually between lat. 28° and 32°N (Roden 1970). Variability in these features influences surface drift (McNally 1981), which in turn affects the neustonic young of armor- 462 BOEHLERT and SASAKI: PELAGIC BIOGEOGRAPHY OF THE ARMORHEAD head. If flow in certain years results in transport of armorhead to other regions (such as the western Pacific or the southern part of the northeast Pacific; Fig. 4), it is possible that large-scale mortalities of armorhead occur later in life, resulting in weak year classes at the SE-NHR seamounts. ACKNOWLEDGMENTS We appreciate assistance from many individuals who provided access to different data sources used in this paper, including Akito Kawamura, Colin Har- ris, Sigeiti Hayasi, Genji Kobayashi, and Glen Jamie- son. We also thank Ted Pietsch, Gunter Seckel, and Peter Klimley and an anonymous reviewer for com- ments on an earlier draft of the manuscript. LITERATURE CITED Abe, T. 1957. New, rare or uncommon fishes from Japanese [waters]. VI. Notes on the rare fishes of the family Histiopteridae. Jpn. J. Ichthyol. 6:35-39, 71-74. 1969. Notes on some edible marine fishes collected between the Bonin Islands and the mouth of Sagami Bay— III. Bull. Tokai Reg. Fish. Res. Lab. 60:5-8. Alton, M. S. 1986. Fish and crab populations of Gulf of Alaska seamounts. In R. N. Uchida, S. Hayasi, and G. W. Boehlert (editors), Environment and resources of seamounts in the North Pacific, p. 45-51. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 43. BiLIM, L. A., L. A. BORETS, AND L. K. Platoshina. 1978. Characteristics of ovogenesis and spawning of the boar- fish in the region of the Hawaiian Islands. In Fisheries oceanography, hydrobiology, biology of fishes and other denizens of the Pacific Ocean, p. 51-57. Izv. Pac. Ocean Sci. Res. Inst. Fish. Oceanogr. (TINRO), Vladivostok, 102. (Engl, transl. by W. G. Van Campen, 1986, 9 p., Transl. No. 106; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396.) Boehlert, G. W. 1986. Productivity and population maintenance of seamount resources and future research directions. In R. N. Uchida, S. Hayasi, and G. W. Boehlert (editors). Environment and resources of seamounts in the North Pacific, p. 95-101. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 43. Boehlert, G. W., and A. Genin. 1987. A review of the effects of seamounts on biological processes. In B. Keating, P. Fryer, R. Batiza, and G. W. Boehlert (editors), Seamounts, islands, and atolls, p. 319- 334. Geophys. Monogr. 43. BORETS, L. A. 1975. Some results of studies on the biology of the boarfish {Pentaceros richardsoni Smith). Invest. Biol. Fishes Fish. Oceanogr., TINRO, Vladivostok 6:82-90. (Engl, transl. by W. G. Van Campen, 1984, 9 p., Transl. No. 97; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396.) 1976. The state of the boarfish, Pentaceros richardsoni (Smith) stock in the Hawaii submersive range areas. Invest. Biol. Fishes Fish. Oceanogr., TINRO, Vladivostok 7:58-62. (Engl, transl. by W. G. Van Campen, 1987, 5 p., Transl. No. 112; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822- 2396.) 1979. The population structure of the boarfish, Pentaceros richardsoni, from the Emperor Seamounts and the Hawaiian Ridge. J. Ichthyol. 19:15-20. BORETS, L. A., AND A. S. SOKOLOVSKY. 1978. Species composition of the ichthyoplankton of the Hawaiian submarine ridge and the Emperor Seamounts. In Fisheries oceanography, hydrobiology, biology of fishes and other denizens of the Pacific Ocean, p. 43-50. Izv. Pac. Ocean Sci. Res. Inst. Fish. Oceanogr. (TINRO), Vladivostok, Vol. 102. (Engl, transl. by W. G. Van Campen, 1986, 10 p., Transl. No. 105; available Southwest Fish. Cent. Hono- lulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396.) Chelton, D. B. 1984. Commentary: Short-term climatic variability in the northeast Pacific Ocean. In W. G. Pearcy (editor). The in- fluence of ocean conditions on the production of salmonids in the North Pacific, p. 87-99. Oreg. State Univ. Sea Grant, ORESU-W-83-001. Chikuni, S. 1970. On gregarious fish, Pseudopentaceros richardsoni (Histiophoridae). Enyo (Far Seas) Fish. Res. Lab. News 3:1-4. (Engl, transl., The "phantom fish," "kusakari tsu- bodai"— an outline, by J. H. Shohara, 7 p.; available Natl. Mar. Fish. Serv., NOAA, Terminal Island, CA 90731.) Clemens, W. A., and G. V. Wilby. 1961. Fishes of the Pacific coast of Canada. 2d ed. Bull. Fish. Res. Board Can. 68, 443 p. Darnitsky, V. B., B. L. Boldyrev, and A. F. Volkov. 1984. Environmental conditions and some ecological char- acteristics of fishes from the central North Pacific sea- mounts. In P. A. Moiseev (editor). Proceedings, conditions of formation of commercial fish concentrations, p. 64-77. Minist. Fish., U.S.S.R., All-Union Res. Inst. Mar. Fish. Oceanogr. VNIRO. (Engl, transl. by W. G. Van Campen, 1987, 12 p., Transl. No. 114; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396.) Emery, W. J., T. C. Royer, and R. W. Reynolds. 1985. The anomalous tracks of North Pacific drifting buoys 1981 to 1983. Deep-Sea Res. 32:315-347. Favorite, F. 1967. The Alaskan Stream. Int. N. Pac. Fish. Comm. Bull. 21, 20 p. Favorite, F., A. J. Dodimead, and K. Nasu. 1976. Oceanography of the subarctic Pacific region, 1960- 1971. Int. N. Pac. Fish. Comm. Bull. 33, 187 p. Fedosova, R. a. 1980. Inter-year changes in the mesoplankton of the waters of the submarine banks of the Hawaiian Ridge (Northwest- ern Pacific Ocean). [In Russ.] Izv. Pac. Ocean Sci. Res. Inst. Fish. Oceanogr. (TINRO), Vladivostok 104:84-87. Fedosova, R. A., and 0. E. Komrakov. 1975. Feeding of Pentaceros richardsoni frys in the Hawaiian region. [In Russ.] Invest. Biol. Fishes Fish. Oceanogr., TINRO, Vladivostok 6:52-55. (Engl, transl. by W. G. Van Campen, 1985, 4 p., Transl. No. 100; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv.. NOAA, Honolulu, HI 96822-2396.) FOLLETT, W. I., AND L. J. DEMPSTER. 1963. Relationships of the percoid fish Pentaceros richard- soni Smith, with description of a specimen from the coast 463 FISHERY BULLETIN: VOL. 86, NO. 3 of California. Proc. Calif. Acad. Sci., 4th Ser. 32(10):315- 338. FUJII, E. 1986. Zoogeographical features of fishes in the vicinity of sea- mounts. In R. N. Uchida, S. Hayasi, and G. W. Boehlert (editors). Environment and resources of seamounts in the North Pacific, p. 67-70. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 43. Genin, a., and G. W. Boehlert. 1985. Dynamics of temperature and chlorophyll structures above a seamount: An oceanic experiment. J. Mar. Res. 43:907-924. Gould, J. L. 1985. Are animal maps magnetic? In J. L. Kirschvink, D. S. Jones, and B. J. MacFadden (editors), Magnetite biomineral- ization and magnetoreception in organisms, p. 257-268. Plenum Press, N.Y. Hardy. G. S. 1983. A revision of the fishes of the family Pentacerotidae (Perciformes). N.Z. J. Zool. 10:177-220. HONMA, Y., AND R. MiZUSAWA. 1969. A record of the young of boar fish, Pseudopentaceros richardsoni, from North Pacific. Jpn. J. Ichthyol. 15:134- 136. Hughes, S. E. 1981. Initial U.S. exploration of nine Gulf of Alaska sea- mounts and their associated fish and shellfish resources. Mar. Fish. Rev. 43(l):26-33. Humphreys, R. L., Jr., and D. T. Tagaml 1986. Review and current status of research on the biology and ecology of the genus Pseudopentaceros. In R. N. Uchida, S. Hayasi, and G. W. Boehlert (editors). Environ- ment and resources of seamounts in the North Pacific, p. 55-62. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 43. Humphreys, R. L., Jr., G. Winans, and D. T. Tagaml In press. Synonomy and proposed life history of the North Pacific pelagic armorhead, Pseudopentaceros wheekri Hardy (Pisces: Pentacerotidae). Copeia 1988. Ignell, S., J. Bailey, and J. Joyce. 1986. Observations on high-seas squid gill-net fisheries, North Pacific Ocean, 1985. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-105, 52 p. Japan Marine Fishery Resource Research Center. 1980. Report of exploratory fishing of Pacific pomfret. Jpn. Mar. Fish. Resour. Res. Cent. Rep. 53-16. 1983a. Report of exploratory fishing of Pacific pomfret. Jpn. Mar. Fish. Resour. Res. Cent. Rep. 54-18. 1983b. Report of exploratory fishing of Pacific pomfret. Jpn. Mar. Fish. Resour. Res. Cent. Rep. 55-16. 1985. Report of exploratory fishing of Pacific pomfret. Jpn. Mar. Fish. Resour. Res. Cent. Rep. 57-12. Kawamura, a. 1982. Food habits and prey distributions of three rorqual species in the North Pacific Ocean. Sci. Rep. Whales Res. Inst. 34:59-91. KiRWAN, A. D., Jr., G. H. McNally, E. Reyna, and W. J. Merrell, Jr. 1978. The near-surface circulation of the eastern North Pacific. J. Phys. Oceanogr. 8:937-945. KOMRAKOV, 0. E. 1970. Distribution and fishery of the boarfish (Pentaceros richardsoni Smith) in the Hawaiian region. In the collection: The present state of biological productivity and the volume of biological resources of the world ocean and prospects for their utilization. Kaliningrad, p. 155-163. (Engl, transl. by W. G. Van Campen, 1987, 9 p., Transl. No. 117; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396.) Larkins, H. a. 1964. Some epipelagic fishes of the North Pacific Ocean, Bering Sea, and Gulf of Alaska. Trans. Am. Fish. Soc. 93:286-290. Lasker, R. 1978. The relation between oceanographic conditions and lar- val anchovy food in the California Current: Identification of factors contributing to recruitment failure. Rapp. P. -v. R6un. Cons. int. Explor. Mer 173:212-230. Lutjeharms, J. R. E., and A. E. F. Heydorn. 1981. Recruitment of rock lobster on Vema Seamount from the islands of Tristan da Cunha. Deep-Sea Res. 28A: 1 237. Macy, p. Y., J. M. Wall, N. D. Lampsakis, and J. E. Mason. 1978. Resources of non-salmonid pelagic fishes of the Gulf of Alaska and eastern Bering Sea. Parts 1-3, Final Rep. BLM/OCSEAP Contracts R7120811 and R7120812, North- west and Alaska Fish. Cent., NMFS, NOAA, Seattle, Wash. McCleave, J. D., AND J. H. Power. 1978. Influence of weak electric and magnetic fields on turn- ing behavior in elvers of American eel Anguilla rostrata. Mar. Biol. (Berl.) 46:29-34. McNally, G. J. 1981. Satellite-tracked drift buoy observations of the near- surface flow in the eastern mid-latitude North Pacific. J. Geophys. Res. 86:8022-8030. McNally, G. J., W. C. Patzert, A. D. Kirwan, Jr., and A. C. Vastano. 1983. The near-surface circulation of the North Pacific using satellite tracked drifting buoys. J. Geophys. Res. 88:7505- 7518. Mysak, L. a., W. W. Hsieh, and T. R. Parsons. 1982. On the relationship between interannual baroclinic waves and fish populations in the northeast Pacific. Biol. Oceanogr. 2:63-103. Nordeng, H. 1977. A pheromone hypothesis for homeward migration in anadromous salmonids. Oikos 28:155-159. Okamura, I., K. Amaoka, and F. Mitani (editors). 1982. Fishes of the Kyushu-Palau Ridge and Tosa Bay. Jpn. Fish. Res. Conserv. Assoc, Tokyo, 435 p. Randall, J. E. 1980. New records of fishes from the Hawaiian Islands. Pac. Sci. 34:211-232. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Fish. Res. Board Can. Bull. 191, 382 p. Robinson, S. M. C, and G. S. Jamieson. 1984. Report on a Canadian commercial fishery for flying squid using drifting gill nets off the coast of British Colum- bia. Can. Ind. Rep. Fish. Aquat. Sci. 150, 25 p. Roden, G. I. 1970. Aspects of the mid-Pacific transition zone. J. Geophys. Res. 75:1097-1109. Roden, G. I., and D. F. Paskausky. 1978. Estimation of rates of frontogenesis and frontolysis in the North Pacific Ocean using satellite and surface meteor- ological data from January 1977. J. Geophys. Res. 83:4545- 4550. Roden, G. I., B. A. Taft, and C. C. Ebbesmeyer. 1982. Oceanographic aspects of the Emperor Seamounts region. J. Geophys. Res. 87:9537-9552. ROYER, T. C. 1978. Ocean eddies generated by seamounts in the North Pacific. Science (Wash., D.C.) 199:1063-1064. 464 BOEHLERT and SASAKI: PELAGIC BIOGEOGRAPHY OF THE ARMORHEAD RoYER, T. C. AND W. J. Emery. 1987. Circulation in the Gulf of Alaska, 1981. Deep-Sea Res. 34:1361-1377. Sasaki, T. 1974. Kita Taiheiyo no kusakari tsubodai (The pelagic armor- head, Pentaceros richardsoni Smith, in the North Pacific). Bull. Jpn. Soc. Fish. Oceanogr. 24:156-165. (Engl, transl. by T. Otsu, 1977, 13 p., Transl. No. 16; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396.) 1986. Development and present status of Japanese trawl fisheries in the vicinity of seamounts. In R. N. Uchida, S. Hayasi, and G. W. Boehlert (editors). Environment and resources of seamounts in the North Pacific, p. 21-30. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 43. Seckel, G. R. 1988. Indices for mid-latitude North Pacific winter wind systems; an exploratory investigation. GeoJournal 16:97- 111. Shimazaki, K., and S. Nakamura. 1981. Ecological studies of the pomfret (Brama japonica Hilgendorf). I. The seasonal distribution pattern and eco- logical considerations. Res. Inst. N. Pac. Fish., Spec. Vol., p. 91-103. Sloan, N. A. 1983. Canadian-Japanese experimental fishery for oceanic squid off British Columbia, summer 1983. Can. Ind. Rep. Fish. Aquat. Sci. 152, 42 p. Smith, J. G. 1965. A second record of the pelagic armorhead, Pentaceros richardsoni Smith, 1849, in California waters. Calif. Fish Game 51:213-214. Takahashi, Y., and T. Sasaki. 1977. Kita Taiheiyo chubu kaizan ni okeru tororu gyogyo (Trawl fishery in the central North Pacific seamounts). Northern Waters Groundfish Resour., Far Seas Fish. Res. Lab., 45 p. (Engl. Transl. by T. Otsu, 1977, 49 p.; Transl. No. 22; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396.) Uchida, R. N., and D. T. Tagami. 1984. Groundfish fisheries and research in the vicinity of sea- mounts in the North Pacific Ocean. Mar. Fish. Rev. 46(2): 1-17. Uda, M., and K. Hasunuma. 1969. The eastward subtropical countercurrent in the west- ern North Pacific Ocean. J. Oceanogr. Soc. Jpn. 25:201- 210. Wagner, E. J., and C. E. Bond. 1961. The percoid fish Pseudopentaceros richardsoni from Oregon waters. Fish. Comm. Oreg., Res. Briefs 8(l):71-73. Walker, M. M. 1984. Magnetic sensitivity and its possible physical basis in the yellowfin tuna, Thunniis albacares. In J. D. McCleave, G. P. Arnold, J. J. Dodson, and W. H. Neill (editors). Mechanisms of migration in fishes, p. 125-142. Plenum Press, N.Y. Walker, M. M., J. L. Kirschvink, and A. E. Dizon. 1985. Magnetoreception and biomineralization of magnetite. Fish. In J. L. Kirschvink, D. S. Jones, and B. J. MacFad- den (editors). Magnetite biomineralization and magneto- reception in organisms, p. 417-437. Plenum Press, N.Y. Wedgeworth, B., and J. Kellogg. 1987. A 3-D gravity-tectonic study of Ita Mai Tai Guyot: an uncompensated seamount in the East Mariana Basin. In B. Keating, P. Frj-er, R. Batiza, and G. W. Boehlert (editors), Seamounts, islands, and atolls, p. 73-84. Geophys. Monogr. 43. Welander, a. D., R. C. Johnson, and R. A. Hajny. 1957. Occurrence of the boarfish, Pseudopentaceros richard- soni, and the zeid, Allocyttus verrucosus, in the North Pacific. Copeia 1957:244-246. Wetherall, J. A., and M. Y. Y. Yong. 1986. Problems in assessing the pelagic armorhead stock on the central North Pacific Seamounts. In R. N. Uchida, S. Hayasi, and G. W. Boehlert (editors), Environment and resources of seamounts in the North Pacific, p. 73-85. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 43. Williams, F. 1972. Consideration of three proposed models of the migra- tion of young skipjack tuna {Katsuwoniis pelamis) into the eastern Pacific Ocean. Fish. Bull., U.S. 70:741-762. Zama, a., M. Asai, and F. Yasuda. 1977. Records of the pelagic armorhead, Pentaceros richard- soni from Hachijo Island and the Ogasawara Islands. Jpn. J. Ichthyol. 24:57-60. 465 THE TIMING AND SIGNIFICANCE OF DENSITY-DEPENDENT AND DENSITY-INDEPENDENT MORTALITY OF AMERICAN SHAD, ALOSA SAPIDISSIMA Thomas F. Savoy and Victor A. Creccoi ABSTRACT We used stock-recruitment, pre- and postrecruitment mortality data for American shad, Alosa sapidissima, in the Connecticut River to estimate density-dependent and density-independent mortality rates at the prejuvenile (age 1-100 days) and postjuvenile (age 101 days to 5 years) phases. Total post- juvenile mortality rates from 1967 through 1982 were dominated by density-independent mortality, with only 6% (^opost = 0.30) of the mean total (Z = 4.85) being ascribed to density-dependent processes. By contrast, 23% {Z^ = 1.13) of the total prejuvenile mortality was compensatory, of which nearly all took place during the embryonic and early larval periods. Egg and early larval mortality rates from 1979 through 1987 were positively correlated to June river flows, and inversely related to mean June temperature, whereas mortality rates of all other life stages showed no such relationships. Daily egg and early larval mortality rates from 1979 through 1987 were usually higher and more variable than mortality rates during later stages. There were significant negative correlations between egg and early larval mortality rates and the strength of 1979-87 year classes in the adult stock, whereas mortality rates of late larvae and juveniles were independent of year-class strength. Density-dependent mortality during the egg and early larval stages comprised over 40% of the total mortality at those stages, resulting in the number of midlarvae and juveniles being positively correlated to adult recruitment. These data suggest that year-class strength of American shad in the Connecticut River is established after the egg and larval stages. It is generally accepted that year-class strength of most fishes is established prior to the juvenile stage from density-independent (climatic factors) and density-dependent (competition, predation, cannibal- ism) processes (Gushing 1974; May 1974; Goodyear 1980). Although density-dependent mortality is believed to be the mechanism that keeps fish popu- lations stable under low to moderate exploitation (McFadden 1977; Gushing 1980), density-dependent processes such as inter- and intraspecific competi- tion, predation, and cannibalism are difficult to measure. This is particularly so for many highly fecund fishes, whose population sizes undergo wide fluctuations that often cannot be predicted by con- ventional stock-recruitment models (Parrish and MacGall 1978; Bakun 1984). Despite the acknowl- edged importance of density-dependent mortality to stock-recruitment theory (Ricker 1975; Ware 1980), few studies have been able to quantify density- dependent mortality, or determine the life stages at which compensation occurs. Density-dependent mortality may be confined to the prejuvenile stage (Gushing 1974, 1980) or the 'Connecticut Department of Environmental Protection, Marine Fisheries Office, Waterford, CT 06385. postjuvenile oceanic stage (Peterman 1978, 1982), or may occur throughout the prerecruitment period (Gulland 1965). The testing of these hypotheses has proceeded slowly because of environmental noise surrounding many stock-recruitment relationships, measurement errors associated with recruitment estimates (Ludwig and Walters 1981), and the lack of long-term mortality data on eggs and larvae. One species, for which accurate and long-term stock-recruitment and prerecruitment mortality data exist, is the American shad, Alosa sapidissima, an anadromous clupeid that spawns in many Atlan- tic coast rivers (Walburg and Nichols 1967). Pre- vious studies on American shad in the Gonnecticut River have demonstrated that larval and juvenile mortality rates decline with age (Grecco et al. 1983, 1986) and that growth and survival rates among discrete larval cohorts are significantly affected by short-term hydrographic and meteorological events (Grecco and Savoy 1985b, 1987a). Although pre- vious stock-recruitment studies of Gonnecticut River shad (Leggett 1977; Lorda and Grecco 1987) found evidence of density-dependent mortality before the juvenile stage, no studies have attempted to esti- mate density-dependent mortality during the egg and larval stages, or quantify density-dependent and Manuscript accepted March 1988. FISHERY BULLETIN: VOL. 86, NO. 3, 1988. 467 FISHERY BULLETIN: VOL. 86, NO. 3 density-independent mortality after the juvenile stage. In this study we used stock-recruitment, pre- and postrecruitment mortality data for American shad in the Connecticut River from 1967 through 1987 to 1) estimate egg mortality rates for the 1979-87 year classes; 2) measure the contribution of density- dependent and density-independent mortality before and after the juvenile stage; and 3) determine the life stage(s) at which most of the density-dependent mortality takes place. METHODS Data Source Estimates of adult recruitment and parent stock size (± SE) from 1967 to 1982 (Table 1) were based on annual population estimates derived by earlier mark-recapture studies (Leggett 1976; Crecco and Savoy 1987^) and by the annual mean number of American shad lifted over the Holyoke Dam (Crecco and Savoy 1985a). The parent stock size (PAR J from 1967 through 1987 was the annual population estimate of female shad minus that year's commer- cial catch of female shad. Parent stock (PAR^) was separated into the parent stock that spawned above the Holyoke Dam (PJ (Moffitt et al. 1982; O'Leary and Booke 1986^, 1987^) and the spawning stock that spawned below the dam (P,, = PAR, - P„). Since female American shad mature between ages four and six, female shad recruitment from the 1967-82 year classes was the sum of virgin 4-, 5-, and 6-yr-old female shad in the 1970-87 runs based on the age-class structure from previous studies (Jones et al. 1976; Leggett 1976; Crecco et al. 1984^). Since the sex ratio of mature progeny from each year class is close to 1:1 (Leggett 1976), total recruitment (Rf) from the 1967-82 year classes was estimated by doubling the female shad recruitment estimates. Direct estimates of male shad recruit- ment are biased by gill net selectivity and differen- tial culling practices of commercial fishermen (Crecco et al. 1984^). ^Crecco, V. A., and T. F. Savoy. 1987b. Fishery Management Plan for American shad in the Connecticut River. Unpubl. manuscr., 117 p. Connecticut Department of Environmental Pro- tection, Hartford, CT 06106. 'O'Leary, J., and H. E. Booke. 1986. Connecticut River anad- romous fish investigations. Mass. Coop. Fish. Res. Unit Proj. Per- formance Rep., F-45-R-2, 37 p. "O'Leary, J., and H. E. Booke. 1987. Connecticut River anad- romous fish investigations. Mass. Coop. Fish. Res. Unit Proj. Per- formance Rep., 32 p. ^Crecco, V. A., T. Savoy, and L. Gunn. 1984. Population dynamics studies of American shad in the Connecticut River. CT Dep. Environ. Prot. Final Rep. AFC 13, 76 p. •^Crecco, V. A., L. Gunn, and T. Savoy. 1981. Connecticut River shad study, a progress report. Unpubl. manuscr., 87 p. Connecticut Department of Environmental Protection, Hartford, CT 06106. Table 1.— Estimates of total female American shad parent stock, female spawn- ing stock above and below the Holyoke Dam, mean June flows from 1967 to 1987, and total number of adult recruits from the 1967-82 year classes. SE = stand- ard errors about the estimates x 10^. Total Spawning Spawning spawning stock stock Recruit- June stock above below ment flow Year X 10^ SE X 10^ SE X 10^ SE X 10^ SE m^/s 1967 167 35 4 0.8 163 34 444 112 437 1968 202 38 5 0.9 197 37 236 59 603 1969 384 70 10 1.8 374 68 490 116 375 1970 413 95 12 2.8 401 92 550 126 243 1971 424 111 14 3.7 410 107 982 205 203 1972 167 40 8 1.9 159 38 430 87 616 1973 111 30 3 0.8 108 29 308 70 534 1974 306 62 10 2.0 296 60 652 139 334 1975 247 65 16 4.2 231 61 560 101 379 1976 435 80 166 31.0 269 49 650 105 286 1977 207 39 112 21.0 95 18 1,240 191 250 1978 210 54 45 12.0 165 42 714 110 488 1979 248 51 87 18.0 161 33 882 127 445 1980 341 58 196 33.0 145 25 1,256 196 201 1981 293 50 143 24.0 150 26 758 132 316 1982 501 67 109 15.0 392 52 282 49 643 1983 423 77 185 34.0 238 43 386 1984 610 67 245 27.0 365 40 661 1985 555 96 219 38.0 336 58 311 1986 392 71 175 32.0 217 39 439 1987 235 37 136 14.0 146 23 250 468 SAVOY AND CRECCO: MORTALITY OF AMERICAN SHAD American shad larvae (10-30 mm TL) were sam- pled during daylight hours in the Connecticut River with a 6.1 m plankton bag seine (2.4 m deep, wing and bag mesh of 0.505 mm) and 30 m lead ropes from 15 May to 20 July 1979-84, 1986, and 1987. No larval sampling was conducted in 1985. One seine haul was taken weekly at 8-12 fixed stations located throughout the major spawning areas (Fig. 1). Fur- ther details on sampling and methods of estimating larval abundance from net samples are contained in Crecco et al. (1983). Juvenile American shad (40-90 mm TL) were col- lected weekly from 20 July through 15 October 1967-87 at 7-14 fixed stations located above and below the Holyoke Dam (Fig. 1) based on weekly and biweekly seine surveys (Scherer 1974; Foote 1976; Marcy 1976; Crecco and Savoy 1984). The annual juvenile index of relative abundance (IND() from 1967 to 1987 was the mean juvenile catch per seine haul ( ± SE) from all stations and collection dates (Table 2). The Timing of Density-Dependent Mortality Analysis One of the primary objectives of this study was to determine the magnitude and timing of density- dependent mortality for American shad. Peterman (1978, 1982) found that density-dependent mortality for some stocks of the anadromous sockeye salmon, Oiworhynchus nerka, was confined mainly to the 4-5 yr oceanic postjuvenile phase. Given that the Ameri- STATIONS X - L A RVA L 0-JUVENILE km 1 90 SUN D ERLU N D HATFIELD HO LYOK E km 140 X AG AW AM ENFIELD Figure 1.— Location of larval and juvenile American shad sampling stations on the Con- necticut River. MASS CONN WILSON O GLASTON BURY km 25 O SALMON RIVER DEEP RIVER ESSEX k m 1 2 469 FISHERY BULLETIN: VOL. 86, NO, 3 Table 2.— Mean juvenile indices, scaled juvenile abundance from Equation (2) and total egg production for American shad in the Con- necticut River from 1967 through 1987. SE = standard errors about the estimates. Juvenile Total egg Juvenile abundance SE production SE Year index 20.2 SE 9.4 X 10^ X 10^ X 10« X 10^ 1967 71,070 33,072 334 70 1968 11.1 1.7 39,053 5,981 404 76 1969 19.0 3.8 68,848 13,770 768 140 1970 27.8 8.2 97,809 28,850 826 190 1971 65.7 14.1 231,151 49,608 848 222 1972 15.3 2.9 53,830 10,203 334 80 1973 12.7 3.4 44,682 11,962 222 60 1974 21.4 6.3 75,292 22,165 612 124 1975 23.7 5.7 83,384 20,054 494 130 1976 22.4 5.9 78,810 20,760 870 160 1977 --NOT AVAILABLE- - 414 78 1978 27.2 5.9 95,698 20,758 420 108 1979 19.6 3.2 68,959 11,259 496 102 1980 42.7 11.0 150,231 38,701 682 116 1981 16.0 2.7 56,293 9,499 586 100 1982 4.7 1.1 16,536 3,870 1,002 134 1983 26.3 6.6 92,531 23,221 846 154 1984 13.0 2.1 45.738 7,388 1,220 134 1985 17.8 2.7 62,626 9,499 1,110 192 1986 17.0 2.8 59,811 9,851 784 142 1987 44.7 15.1 157,268 53,126 470 74 can shad has a similar 4-6 yr postjuvenile phase as sockeye salmon, we tested Peterman's hypothesis for American shad with key factor analysis (Bellows 1981; Rosenberg and Doyle 1986). We fixed the postjuvenile period for American shad in the Con- necticut River between 101 days and 5 years, corre- sponding to the average age (101 days) at which juvenile shad leave the river (Crecco and Savoy 1985b) and the average age (5 years) when they return to the river as mature adults. The total post- juvenile (age 101 days to 5 years) mortality rates (ZA() for the 1967-82 year classes were related to the scaled juvenile indices (J<) for those year classes (Table 2) in a linear model: ZA, = a + b{Jt), (1) where ZAf = -logg(i2(/J<). If significant density- dependent mortality is present during the post- juvenile stage, the slope (6) of Equation (1) would be positive and differ significantly (P < 0.05) from zero. The juvenile indices were scaled to thousands of fish (Jf) so that the y-axis intercept in Equation (1) directly estimates the mean density-independent mortality (Zj^g^i), and the slope (b) times the geo- metric mean juvenile abundance (GM) from 1967 through 1987 is the mean density-dependent mor- tality rate (Z^post). Total juvenile abundance (J() was estimated by multiplying the juvenile indices (IND,) by a scalar (SC): SC = [GR exp{-EZA) ■ GM] = 3,518.3, (2) where GR is the geometric mean total adult recruitment for 1967 through 1982; GM is the geo- metric mean juvenile index from 1967 through 1987; EZA (4.85) is the mean total instantaneous mortality among postjuveniles from 1967 through 1982. We estimated EZA as the sum of mortality during the late juvenile period (age 101-365 days), and the subadult stage (age 1-5 years). The mean total in- stantaneous mortality rate during the late juvenile stage was estimated as 2.65 (0.01 • 265 days), using a mean daily mortality rate of 0.01 (SE = 0.002) extrapolated from the 1979-84 larval and juvenile survivorship curves (Crecco and Savoy 1985b). The mean total instantaneous mortality rate of subadult shad was 2.2, based on an annual instantaneous natural mortality rate of 0.45 from the method of Pauly (1980) plus 0.10 to reflect oceanic fishing mor- tality (2.2 = 4(0.45 -I- 0.10)). We estimated the an- nual natural mortality rate (0.45) by substituting the ;^(0.25, SE = 0.03) and L (55 cm FL, SE = 3 cm) parameters of the von Bertalanffy equation for male and female shad combined and preferred ocean tem- perature (14 °C) of American shad (Leggett and Whitney 1972) into Pauly' s multiple regression model. The oceanic fishing mortality estimate (0.10) was based on tagging studies in Delaware Bay (White et al. 1969; Zarbock 1969) and off the New York-New Jersey coast (Nichols 1958). To determine if density-dependent mortality takes place during the egg and larval stages as Cushing (1980) hypothesized, we related total prejuvenile mortality rates {ZEJt) to annual egg production (EggSj ), and to both egg production (Table 2) and mean June river flow (JFLOW) from 1967 to 1987 (Table 1) in linear regression models: ZEJt = a -H 6 (Eggs,) (3a) and ZEJt = a + ^(Eggs,) -H c(JFLOW,), (3b) where ZEJt = -log^(J( /Eggs,). Mean June river flows (m^/s) were included in Equation (3b) because previous studies (Crecco and Savoy 1985b, 1987a) have shown that high June flows reduced prejuvenile survival rates, leading to a significant inverse cor- relation (r = - 0.74, P < 0.01) between the juvenile indices of abundance from 1967 through 1980 and 470 SAVOY AND CRECCO: MORTALITY OF AMERICAN SHAD mean June river flow (Crecco and Savoy 1984, 1985b). Mean June flows (m^/s) were measured within the major spawning areas (Leggett 1977) by the United States Geological Survey (U.S. Geo- logical Survey 1967-84). We estimated total egg production (Eggs,) in Equations (3a) and (3b) as the product of the mean fecundity of a female American shad times that year's parent stock (PARf) of female shad (Table 1). The average fecundity was reduced from 269,000 ova (Leggett 1969) to 200,000 to reflect the average rates of egg retention and in- complete fertilization (Watson 1970; Reed and Russo 1976^). Since Leggett (1969) showed that the aver- age fecundity of American shad varied by less than 10% from 1966 through 1973, we were justified in using an average fecundity for all years. The mean density-independent mortality rate (Z/pre) present during the prejuvenile stage was the y-axis intercept (a) in Equation (3b) plus the slope (c) times the overall geometric mean June flow 'Reed, R. J., and A. Russo. 1976. American shad research Connecticut River, Massachusetts, 1976. 1. Fecundity, egg reten- tion, sex ratio, and age class composition. Unpubl. manuscr., 16 p. Massachusetts Cooperative Fisheries Research Unit, Univer- sity of Massachusetts, Amherst, MA 01003. (GJFLOW) from 1967 through 1987: Zipre = a + c(GJFLOW), (4) The mean density-dependent mortality rate (Zq^^^) was the slope (6) of Equations (3a) and (3b) times the geometric mean egg production (GEgg) from 1967 through 1987. A positive and statistically sig- nificant slope (6) of prejuvenile mortality on egg abundance would support the Gushing hypothesis that density-dependent mortality takes place before the juvenile stage. If the b estimates of Equations (1), (3a), and (3b) were all positive and significant, our results would support Gulland's (1965) hypoth- esis that density-dependent mortality occurs over the entire prerecruitment phase. Results Although total juvenile abundance varied 14-fold from 1967 through 1982 (Table 2), the total post- juvenile mortality rates {ZA() exhibited relatively low variability (95% G.I.: 4.70-5.00) about the esti- mated mean {EZA = 4.85) (Fig. 2). The slope of the linear regression between total postjuvenile mortal- 6.6 5.4 5.2 5.0 < N 4.8 4.6 4.4 4.2 4.0 50 1 1 1 100 160 200 Juvenie Abundance (thousands) 260 300 Figure 2.— Relationship between American shad postjuvenile mortality (ZA,) and relative abundance of juveniles (J() from 1967 through 1982. 471 FISHERY BULLETIN: VOL. 86, NO. 3 ity rates (ZAt) and scaled juvenile abundance (J,) was positive and statistically significant (P < 0.04) (Table 3). Since the average density-dependent mortality (Z^post = 0.30) present during the post- juvenile stage comprised only 6% of the mean total postjuvenile mortality rate {ZA = 4.85), density- dependent mortality is not large enough to alter the significant linear correlation (r = 0.78, P < 0.01) betv^^een the juvenile indices (J^ ) and adult recruit- ment (Rf) for the 1967 through 1982 year classes (Fig. 3). The prejuvenile mortality rates (ZEJ,) from 1967 through 1987 v^ere positively correlated (r = 0.70, P < 0.0004) to total egg production (EggS() (Fig. 4). Egg production alone explained 49% of the varia- tion in prejuvenile mortality (Table 3) and the slope Table 3.— Estimates of density-dependent (Zp), density-independent mortality (Z,) and the fraction of density- dependence (ZqIZi^Zq) during the postjuvenile (1967-82) and prejuvenile (1967-87) mortality phases for American shad in the Connecticut River. SE = standard error, and numbers in parentheses = 95% C.I. for Zq,ZEJ, = prejuvenile mortality, Z4, = postjuvenile mortality. Model Parameters SE Z, Zol{Z,,Z^) Prejuvenile ZEJ, = a + b(Eggs,) ZEJt = a + b(Eggs,) + c (J Flow,) Postjuvenile ZA, = a + b(J,) a = 5.466 b = 1.88 X 10"^ r^ = 0.49 a = 4.287 b = 1.97 X 10"^ c = 0.0028 r^ = 0.80 a = 4.53 b = 3.49 X 10" r^ = 0.275 0.313 4.43 X 10"^ 0.303 2.87 X 10"^ 0.0005 0.15 1.51 X 10"^ 1.13 5.47 0.17 (0.862-1.392) (0.136-0.203) 1.18 5.34 0.18 (1.007-1.351) (0.158-0.202) 0.30 4.53 0.062 (0.035-0.560) (0.008-0.110) 1300 « 1025 c <0 (0 3 o I 750 u 3 < 475 200 — n 10 —J- 20 — I— 30 — r- 40 — I— 50 — I— 60 70 Juvenle Index Figure 3.— Relationship between total adult recruitment (R,) and the juvenile abundance indices (/ATZ), ) of American shad from 1967 through 1982. 472 SAVOY AND CRECCO: MORTALITY OF AMERICAN SHAD of the regression differed significantly from zero, indicating the presence of significant density- dependent mortaHty during the prejuvenile stage. When prejuvenile mortality rates were related to both egg production (Eggs,) and mean June river flow, (JFLOW,), the multiple regression model ac- counted for 80% of the variability in prejuvenile mortality (Table 3), and the slope estimates for egg production (6) and June flow (c) were positive and highly significant. Note that the standard error (SE) about the slope estimate (6) was reduced by 60% when June flow effects were considered. These results suggest that prejuvenile mortality rates are affected by a combination of density-dependent (egg production) and density-independent (June river flow) factors. The mean density-dependent mortality rate dur- ing the prejuvenile phase (Z^pj.^ = 1.18) from the multiple regression model was four times greater than the mean Z^post value (0.30) for postjuveniles (Table 3), suggesting that 80% (1.18/1.18 + 0.30) of the total compensatory reserve for American shad occurs before the juvenile stage. Whereas compen- satory density-dependent mortality may play a sig- nificant role in regulating egg and larval abundance during years of high egg production, it is clear that most of the variability (82%) in egg and larval abun- dance is ascribed to density-independent factors. Egg, Larval, and Juvenile Mortality, 1979-87 Analysis To examine how larval and juvenile mortality rates varied with year-class strength, survivorship curves for 19,000-180,000 American shad larvae and 800-3,500 juveniles were developed annually from 1979 to 1987 following the techniques of Lough (1976) and Hewitt et al. (1985). Mortality rates could not be estimated directly for prolarvae because American shad yolk-sac larvae (7-9 mm TL) remain in deep water (Marcy 1976) and were only partially susceptible to the plankton seine. All larvae and juveniles collected annually were separated into four length intervals: 1) 10-13 mm TL, reflecting first- feeding larvae with undeveloped pelvic fins (Wiggins et al. 1984); 2) 14-19 mm TL, associated with the onset of pelvic fin development (Lippson and Moran 1974); 3) 20-28 mm TL, repre- senting larvae approaching metamorphosis, char- acterized by invagination of the gut (Maxfield 1953); 9.0 ao W 7.0 6.0 6.0 i T- 20 — r- 30 — r- 40 I 60 — I— 60 — r- 70 I 80 — I— 90 100 no 120 130 Total Eggs (bMons) Figure 4.— Relationship between American shad prejuvenile mortality (ZEJ^) and the number of eggs produced (Eggs,) from 1967 through 1987. 473 FISHERY BULLETIN: VOL. 86, NO. 3 and 4) 29-80 mm TL, corresponding to the juvenile stage and the onset of scale development (Marcy 1976). These four groups are hereafter referred to as early larval, midlarval, late larval, and juvenile stages, respectively. We used otolith increment counts (Savoy and Crecco 1987) to age a subsample of 100-400 Ameri- can shad larvae and 80-200 juveniles annually from 1979 to 1987. The age-length relationship for lar- vae and juveniles in all years was well described by the Gompertz growth equation (Crecco et al. 1983; Crecco and Savoy 1985b) so we used a pooled Gompertz equation to describe larval and juvenile growth. The average age (t) of each of the four length groups was estimated by rearranging the Gompertz equation: t = logAKIK - log,(L,(Lo))]/a (5) where L( = total length (mm); Lq = length at hatching (8.0 mm); K = growth rate at the inflec- tion point; and a = rate of exponential decay. The daily instantaneous mortality rates (d) for each stage and their standard errors from 1979 through 1987 were estimated by an exponential model, Nt = A exp{-dt), (6) that related abundance (A^^) and age (t) by non- linear regression methods (SAS 1982). The total instantaneous mortality rates of early (ZELf), mid (ZMLf), and late (ZLLt) larvae and juveniles (ZJf) were estimated by multiplying the corresponding daily mortality rate (d) by the average duration (Ds = days) within each stage. Previous studies (Crecco and Savoy 1985a; Savoy and Crecco 1987) found that the growth rates of early (10-13 mm) and mid- larvae (14-19 mm) were positively correlated with June water temperatures, whereas the growth of late larvae and juveniles were independent of river temperatures. As a result, we adjusted the stage duration {Ds ) of early and midlarvae with the mean June temperatures (U.S. Geological Survey 1979- 87) from 1979 through 1987 and the larval growth- subtracting the sum of larval {ZEL, , ZMLf , ZLL,), juvenile (ZJ^) and postjuvenile (ZA^) mortality rates from the total prerecruitment mortality rate (ZTotal,): ZEt = ZTotalt - [ZELi + ZML, + ZLLt + ZJt + ZAt), (7) where ZTotal^ = -logp(/2,/Eggs,). However, since the 1983 through 1987 year classes have not been fully recruited to the spawning population, we esti- mated total adult recruitment {Rp^ ) for those year classes by the following environment-dependent stock-recruitment model: Rpt = 24.29 (PARt) exp(- 0.0052 • P,,) exp(- 0.0032 • PJ exp(- 0.0025 • JFLOW() (8) Adult recruitment was estimated independently of the juvenile indices by substituting each year's parent stock size (PAR^), mean June flow (e/FLOW), female parent stock lifted over the Hol- yoke Dam (P„ ), and female parent stock below the Holyoke Dam (P^) (Table 1) into the model. This nonlinear model was shown (Lorda and Crecco 1987) to be a good predictor (r^ = 0.81, P < 0.001) of adult recruitment (P^) for the 1966 through 1982 year classes. Moreover, the predicted recruitment levels (P() of the 1983 through 1987 year classes from Equation (8) were closely correlated (r = 0.92, P < 0.01) with the corresponding juvenile abundance (J() for those years (Table 4) which is consistent with the positive correlation (r = 0.78, P < 0.01) between adult recruitment and the 1967-82 juvenile indices (Fig. 2). This justifies the use of Equation (8) to predict adult recruitment, total mortality, and postjuvenile mortality rates for the 1983-87 year classes. The standard errors about the egg mortality rates (ZEt) were derived as the sum of the variances of all other terms (Cochran 1965): SEze' = ^SEzel' + SE^ml' + S^zll' + ZJ^ + SE^/ + SE^xotai' 0) temperature equations (Crecco and Savoy 1985b: table 7). We estimated total egg and prolarval mortality rates (ZE^) indirectly from 1979 through 1987 by The standard errors about the total prerecruitment (ZTotal,) and postjuvenile (ZA^) mortality rates (Table 5) were based on the same principle as Equa- tion (9) (App. 1). 474 SAVOY AND CRECCO: MORTALITY OF AMERICAN SHAD Table 4— Predicted total adult recruitment of American shad from the environmental dependent stock-recruitment model (Equation (8)), 95% con- fidence limits about the recruitment values and juvenile abundance for the 1983-87 year classes. Predicted adult Juvenile recruitment abundance Year X 10^ 95% C.I. X 10^ 1983 634 362-906 92,531 1984 196 112-280 45.738 1985 555 317-793 62,626 1986 587 335-839 59,811 1987 955 547-1,363 157,268 Table 5.— Estimates of the total (ZTotal,) and postjuvenile (Z4,) instantaneous mortality rates and their standard errors (SE) for American shad from 1979 through 1987. Year ZTotal, SE ZA, SE 1979 10.94 0.26 4.36 0.23 1980 10.90 0.24 4.78 0.32 1981 11.26 0.25 4.31 0.25 1982 13.00 0.27 4.77 0.34 1983 11.80 0.30 4.98 0.36 1984 13.34 0.26 5.45 0.28 1985 12.21 0.29 4.72 0.28 1986 11.80 0.30 4.62 0.29 1987 10.80 0.28 5.10 0.47 Given that ZE represents total mortality through- out the egg and prolarval period, daily egg mortal- ity rates were determined by dividing ZE by the average duration (Ds) of the egg and prolarval stages of American shad in the Connecticut River. Watson (1968) reported an inverse relationship be- tv^^een the incubation period (D ) of shad eggs and water temperature (T) by the expression: D = 120.95 exp(-0.154 • T). (10) To determine the duration {Ds) of the egg stage from 1979 through 1987 we substituted the mean June temperatures for the Connecticut River (U.S.G.S. Annual Water Year Reports 1979-87) into Equation (10). The total number of prolarvae (age 2 days) and early larvae (age 10 days) for the 1979 through 1987 year classes was estimated by NE^ = Eggs, • exp(-Z£;,) (11) and NFt = Eggs, • exipi-ZEt - ZELi) (12) respectively, where NEt is the estimated number of prolarvae larvae and NFt is the number of early larvae. The abundance of older larvae and juveniles from each year class was estimated by adding their respective total mortality rates to Equation (12). To determine the life stage(s) at which year-class strength is established, we related the stage-specific total mortality rates and abundance estimates to the number of adult recruits (i?, or Rpt) from the 1979-87 year classes in several linear models. If year-class strength is established early, there should be a significant positive correlation between the number of prolarvae (NE) and early larvae (NF) and adult recruitment, and total mortality rates during these early stages {ZE and ZEL ) should be inverse- ly related to adult recruitment. To determine the extent of density-independent mortality at the egg, larval, and juvenile stages, we correlated the stage-specific total mortality rates from 1979 through 1987 to mean May and June river flows (m^/s) and water temperatures (°C) in sev- eral linear models. May and June hydrographic and meteorological parameters were used because they coincide with egg, larval, and juvenile develop- ment in the Connecticut River (Leggett 1977) and were the only monthly abiotic variables that were significantly linked to adult shad recruitment from 1966 to 1980 (Crecco and Savoy 1984, 1987c). May and June water flows and temperatures were re- corded by the U.S. Geological survey (U.S. Geo- logical Survey 1966-1980) within the major spawn- ing areas (river km 89) of American shad (Leggett 1977). Results The mean daily mortality rates from 1979 through 1987 declined by an order of magnitude from the egg through the juvenile stages (Table 6). Total egg mortality rates were relatively high (mean ZE = 2.584, cv = 18.5%) and were inversely correlated (r = - 0.76, P < 0.03) with adult recruitment (Rt or Rpt) from those year classes (Table 7). These data indicate that shad eggs and prolarvae (age 1-2 days) from 1979 through 1987 experienced high (85-96%) mortality that was directly linked to year-class success. The total mortality rates (ZEL) among early lar- vae were slightly lower and more variable (mean ZEL = 1.608, cv = 31.5%) than the egg mortality rates (Table 6), and were also inversely related (r = - 0.83, P < 0.01) to adult recruitment from 1979 through 1987. By contrast, the total mortality rates of older larvae {ZML and ZLL) and juveniles (ZJ) 475 FISHERY BULLETIN: VOL. 86, NO. 3 Table 6.— Stage-specific total instantaneous mortality rates of egg, larval, and juvenile American sfiad from 1979 to 1987. SE = stage-specific standard errors; cv = coeffi- cient of variation, Days = duration of each stage. Instantaneous total mortality Egg + pro- Early Mid- Late larvae larvae larvae larvae Juvenile Year or (day 2) (day 3-9) (day 10-18) (day 19-29) (day 30 1-100) statistic mean SE mean SE mean SE mean SE mean SE 1979 2.481 0.554 1.323 0.245 1.116 0.189 0.330 0.110 1.330 0.280 1980 1.853 0.819 1.218 0.186 1.096 0.336 0.693 0.099 1.260 0.280 1981 2.580 0.684 1.836 0.108 0.408 0.328 0.726 0.154 1.400 0.210 1982 2.940 0.501 2.144 0.152 1.160 0.210 0.726 0.187 1.260 0.210 1983 3.012 0.785 1.400 0.546 0.522 0.234 0.836 0.143 1.050 0.140 1984 2.706 0.601 2.520 0.352 0.640 0.330 0.484 0.242 1.540 0.280 1985 1986 N 0 0.196 DATA 1.206 1.330 1.120 0 pao 3.158 0.652 1.323 0.261 0.374 0.066 0.280 1987 1.944 0.852 1.098 0.180 0.864 0.128 0.464 0.088 1.330 0.210 2 2.584 1.608 0.877 0.579 1.286 SE 0.169 0.179 0.111 0.066 0.051 cv 0.185 0.315 0.359 0.325 0.120 Days 7.2 6.9 8.9 11.0 70.0 Daily Z 0.359 0.233 0.099 0.053 0.018 Table 7.— Correlations between stage-specific total mortalities and adult recruitment and several abiotic factors from 1979 to 1987 for American shad. P = probability levels. Stage Adult recruit- ment May flow May tempera- ture June flow June tempera- ture Egg p -0.76 0.03 0.38 0.35 -0.08 0.85 0.66 0.08 -0.59 0.13 Early P -0.83 0.01 0.42 0.29 -0.21 0.62 0.81 0.01 -0.64 0.09 Mid P 0.16 0.71 -0.41 0.32 0.49 0.22 0.09 0.83 -0.34 0.42 Late P -0.02 0.96 0.20 0.64 -0.51 0.20 -0.13 0.75 0.18 0.68 Juv P -0.19 0.65 0.05 0.91 0.02 0.97 0.26 0.53 -0.08 0.86 were highly variable, but showed no relationship to adult recruitment. Given that high (94-99%) egg and early larval mortality rates were inversely related to year-class strength, the abundance of first-feeding larvae (age 10 days) was closely correlated (r = 0.84, P < 0.01) with adult recruitment (Rpt) from 1979 through 1987 (Table 8). Because year-class strength from 1967 through 1987 was independent (r = - 0.15, P < 0.46) of egg production (EggS(), our results sug- gest that year-class strength of American shad is determined by the number of shad that survive the embryonic and early larval stages. Early larval mortality rates from 1979 to 1987 were positively correlated with mean June river flows. No such relationships were evident for any other stage-specific mortalities (Table 7). The high egg and early larval mortality rates (Table 6) and relative failure of the 1982 and 1984 year classes coincided vdth major storm events in mid- June 1982 and late May 1984 (Crecco and Savoy 1984) which increased river flows and kept water temperatures below 17°C until late June. The high river flows in 1982 and 1984 were also coupled with the highest parent stocks on record since 1967 (Table 1). By con- trast, the relatively low egg and early larval mor- tality rates for the dominant 1980 and 1987 year classes were associated with low June river flows, a steady rise in June water temperatures (Crecco and Savoy 1984) and low to moderate size parent stocks (Table 1). These results suggest that domi- nant year classes of American shad are most likely to occur when lower than normal June flows are coupled with relatively small spawning stocks. Prejuvenile Density-Dependent Mortality: A Closer Look Analysis Having estimated that 82% of the density- dependent mortality for American shad takes place before the juvenile stage (Table 3), we attempted to estimate the relative contribution of density- dependent mortality during the egg, early, mid-, and late larval and juvenile stages from 1979 through 1987. We related egg mortality rates (ZEi) from 476 SAVOY AND CRECCO: MORTALITY OF AMERICAN SHAD Table 8.— Relationship between the number (x 10®) of American shad at each life history stage and adult recruitment from each year class, r = correlation coefficient and P = probability levels from 1979 to 1987. Stage Year- Pro- Early Mid- Late class Egg larvae larvae larvae larvae Juvenile Adult^ 1979 496 41.5 11.0 3.6 2.6 0.7 882 1980 682 106.9 31.6 10.6 5.3 1.5 1,256 1981 586 44.4 7.1 4.7 2.3 0.6 758 1982 1,002 53.0 6.2 1.9 0.9 0.3 282 1983 846 41.6 10.3 6.1 2.6 0.9 634 1984 1,220 81.5 6.6 3.5 2.1 0.5 198 1985 1986 1,110 784 NO DATA- 555 587 33.3 8.9 2.7 1.8 0.6 1987 470 67.3 22.4 9.5 5.9 1.6 955 r = -0.15^ 0.33 0.84 0.80 0.79 0.82 P = 0.46 0.43 0.01 0.02 0.02 0.01 'Predicted or observed recruitment in thousands. ^Correlation coefficient for egg abundance and recruitment was based on the 1967-87 data (Tables 1.2). 1979 to 1987 to total egg production (Eggs^) in a linear model: ZEt = a + 6(Eggs,). (13) As before, significant density-dependence would be shown if the slope (6) in Equation (13) was positive and differed significantly from zero. The cumulative amount of density-dependent mor- tality during the egg and early larval stages com- bined was estimated by summing the instantaneous total mortality rates of early larvae (ZELf) and eggs (ZEt) and then regressing the total {ZEf + ZELf) against egg production as in Equation (13). This procedure was repeated for each subsequent stage by adding their respective instantaneous total mortality rates. We then estimated the relative magnitude of density-dependent mortality (Z^) at each stage by multiplying the slope (6) of each regression equation by the geometric mean egg pro- duction (GEgg) from 1967 through 1987. The density-independent mortality rate {Zj) at each stage was expressed by the y-a.xis intercept (a) of each regression equation. The percentage contribu- tion of density-dependent mortality {%Zf)) for each period was the ratio of Z^ to total mortality {Z^ + Zj) times 100. Results Our results showed that significant density- dependent mortality first occurs during the early larval stage and persists for all stages thereafter (Table 9). The percentage contribution of density- dependent mortality i%Z^) rose from 24% of the total during the egg stage to 41% during the early larval stage, and then declined during the mid- and late larval stages as the magnitude of density- independent mortality increased. The mean density- dependent mortality rate for the egg and early larval Table 9.— Relationship between the cumulative mortality rates and parent stock sizes (Eggs,). The cumulative mortality rate was egg mortality (ZE) plus early larval mortality (Z2), midlarval mortality (Z3), late larval mortality (Z4), and juvenile mortality (Z5). Zg = density-dependent mortality rate, Z, = density-independent mortality rate o/oZo = percentage density-dependent mortality, %Z, = percentage density-independent mortality, t = student /-statistic. Model b parameter SE(b) t Zo Zi o/oZo o/oZ, ZE = a + b(Eggs,) 9.89 X 10-3 6.40 X 10" -9 1.55 0.59 1.83 24 76 Z2 = a + fa (Eggs,) 2.57 X io-» 7.76 X 10" -a ^3.31 1.54 2.24 41 59 Z3 = a + b(Eggs,) 2.44 X io-« 8.16 X 10" -y '2.99 1.46 3.21 31 69 Z4 = a + 6(Eggs,) 2.60 X io-« 7.74 X 10" -y '3.36 1.55 3.67 30 70 Z5 = a + b(Eggs,) 2.69 X 10"® 7.50 X 10" -y '3.59 1.61 4.88 25 75 'Significant student f-statistic at the P < 0.01 level and SE(b) = standard error of b. 477 stages combined (Zq = 1.54) between 1979 and 1987 was not significantly different (Table 9) from the total density-dependent mortality rate (Zq^j.^ = 1.18) estimated during the prejuvenile period from 1967 to 1982 (Table 3), suggesting that nearly all of the density-dependent mortality for prejuvenile American shad occurs before the midlarval stage. DISCUSSION Although American shad eggs and early larvae experience high mortality (15-40%/day) in the Con- necticut River, the average density-dependent mor- tality rate (Z^p^g = 1.18) during those stages com- prised a relatively small percentage (18%) of the total prerecruitment mortality. This suggests that most of the annual variability in American shad recruitment is explained by density-independent fac- tors, which is consistent with the significant positive correlation between mean June flow and egg and early larval mortality rates, and with the significant inverse correlation (r = - 0.74, P < 0.001) between mean June river flow and adult recruitment from the 1967 through 1982 year classes (Crecco and Savoy 1984, 1987b). Whereas these data illustrate that density-dependent mortality plays a minor role in governing annual variability in American shad recruitment, this does not mean that compensatory processes are trivial. As pointed out by Ricker (1954) and others (Gushing 1974; Garrod and Horwood 1984), only a small amount of density-dependent mortality is required to stabilize the growth poten- tial of fish populations because density-dependent effects become progressively more effective at higher egg and larval densities (Shepherd and Gushing 1980; Murray 1982). Although our results on American shad support the Gulland (1965) hypothesis that density-depend- ent mortality persists throughout the prerecruit- ment period, most (82%) of the density-dependent mortality occurs during the egg and early larval stages. Our average estimate of density-dependent mortality {Zq = 1.54) during the early larval periods does not differ significantly from the mean Zd value (1.34) estimated among the 5-d larval cohorts in 1983 and 1984 (Grecco and Savoy 1987a), or from the mean density-dependent mortality rate Zj) (1.21) estimated by stock-recruitment methods for the entire prerecruitment period (Lorda and Grecco 1987). That year-class strength is established early in the ontogeny of American shad is supported further by the significant positive correlation (r = 0.84, P < 0.009) between the relative abundance of early larvae and all subsequent stages from 1979 FISHERY BULLETIN: VOL. 86, NO. 3 through 1987 and adult recruitment from those year classes (Table 8). The main causes of density-dependent mortality are thought to be predation, competition, and canni- balism (Ricker 1954). Since adult American shad are not thought to feed in freshwater (Walburg and Nichols 1967), we can probably eliminate cannibal- ism as a mechanism for significant density-depend- ent mortality during the egg and early larval stages. Therefore, density-dependent mortality among early American shad larvae is most likely caused by intra- specific competition for food or space and predation. The exact underlying density-dependent mortality mechanisms are difficult to quantify because the out- come of competition and predation may depend on June flow effects shown here (Table 3) and else- where (Grecco and Savoy 1987a, b) to be the prin- cipal density-independent factor. High June river flows have been shown to reduce June river tem- peratures (Grecco and Savoy 1984) and the growth rates of shad eggs (Watson 1968) and early larvae (Grecco and Savoy 1985b). Because slower growing larvae may be susceptible to predation for a longer period of time (Hunter 1976), periods of high flow may indirectly enhance egg and larval predation. Additionally, since high flows reduce the spatial patchiness (Grecco and Savoy 1987a), abundance and availability of river zooplankton (Whitton 1980; Threlkeld 1986), high flows may result in increased levels of competition among American shad larvae for available prey, especially if shad larvae are capable of depleting local aggregations of edible zooplankton. Periods of high runoff that coincide with peak larval production, such as in June 1982 and 1984, may advect larvae and their zooplankton prey from eddies and backwaters where they are normally found (Gave 1978) to areas of high pred- ator abundance. In light of the many ways in which June flows potentially mediate larval mortality, it is unlikely that a single compensatory mechanism is responsible for the relatively high density-depend- ent mortality rate (Z^pre) among early larvae. Since egg and prolarval shad have endogenous food reserves, density-dependent mortality of these stages is likely due to predation and competition for sites among spawning adults. As spawning stocks reach high densities, such as in 1982-84, crowding of adult fish on the spawning grounds may result in a reduction in the number of eggs released ((jood- year 1980). In addition, since the amount of spawn- ing habitat in the Connecticut River can be con- sidered fixed from year to year, larger spawning stocks are more apt to deposit an increasing per- centage of eggs in unfavorable areas. Layzer (1974) 478 SAVOY AND CRECCO: MORTALITY OF AMERICAN SHAD noted that the highest survival rates of American shad eggs occurred over gravel and rubble sub- strates. Since these areas generally have higher water velocities, eggs lodged within the interstices of the gravel and rubble are less likely to be eaten by predators or covered by silt. Dense aggregations of eggs may also be highly susceptible to fungal agents (Leach 1925), which could then act as effi- cient density-dependent predators. However, since predators of American shad eggs and larvae in the Connecticut River have not yet been identified, density-dependent losses are difficult to quantify. Also, predation mortalities in general may be either density-dependent or density-independent, depend- ing on the functional response between the predator and the prey (Hassell 1978), and whether predators actively search for discrete aggregations of eggs and larvae or prey on randomly encountered eggs and larvae (Gulland 1987). In the absence of direct egg mortality estimates of American shad, the accuracy of our egg mortal- ity estimates {ZEt) from Equation (7) are difficult to evaluate, particularly since the standard errors about the estimates are highly variable (Table 6). To provide an independent estimate of daily egg mor- tality, we used the multiple regression method of McGurk (1987) to estimate the average daily egg mortality rate. This method requires an estimate of the patchiness and the mean dry weight of a shad egg. We derived Lloyd's patchiness index (x = 7.26, SE = 0.376) for American shad eggs in the Connecticut River from the 1974-75 egg surveys in the Holyoke Dam impoundment (NUSCo 1977). Sub- stituting the mean weight of a shad egg (0.00025 g, Leggett 1969) and the mean patchiness index into McGurk's equation 4, yields an instantaneous daily egg mortality rate of 0.34 which closely approx- imates our average estimate (0.36) from 1979 to 1987 (Table 6). Lastly, since it is presently unknown if younger (smaller) virgin female American shad produce smaller, less viable eggs than older shad as was reported for cod, Gadus morhua, (Knutsen and Tilseth 1985), the contribution of the age structure to the parent progeny relationship (Rosenberg and Doyle 1986) should be considered. This issue and an examination of whether egg deposition rates are density-dependent should be addressed by monitor- ing egg retention rates among postspawning Ameri- can shad. ACKNOWLEDGMENTS We would like to thank Rita Lorenzetti-Langan, Mary Payette, Thomas Stanford, and all the other people who have contributed field and laboratory time or helpful advice. We also thank the two anonymous reviewers for their comments on the manuscript. LITERATURE CITED Bakun, a. 1984. Report of the working group on environmental studies and monitoring. In J. Csirlce and G. Sharp (editors), Reports of the expert consultation to examine changes in abundance and species composition of neritic fish resources, p. 41-54. FAO Fish. Rep. 291 Vol. 1. Bellows, T. S., Jr. 1981. The descriptive properties of some models for density- dependence. J. Anim. Ecol. 50:139-156. Cave, J. R. 1978. American shad, Alosa sapidissima, larval distribution, relative abundance and movement in the Holyoke Pool, Con- necticut River, Mass. M.S. Thesis, Univ. Massachusetts, Amherst, MA, 65 p. Cochran, W. G. 1965. Sampling Techniques. 2d ed. John Wiley & Sons, Inc., N.Y., 413 p. Crecco, V. A., AND T. Savoy. 1984. Effects of fluctuations in hydrographic conditions on year-class strength of American shad {Alosa sapidissima) in the Connecticut River. Can. J. Fish. Aquat. Sci. 41: 1216-1223. 1985a. Density-dependent catchability and its potential causes and consequences on Connecticut River shad, Alosa sapidissima. Can. J. Fish. Aquat. Sci. 42:1649-1657. 1985b. Effects of biotic and abiotic factors on growth and relative survival of young American shad in the Connecticut River. Can. J. Fish. Aquat. Sci. 42:1640-1648. 1987a. Effects of climatic and density-dependent factors on intra-annual mortality of larval American shad. Am. Fish. Soc. Symp. 2:69-81. 1987b. Recruitment mechanisms of the American shad, Alosa sapidissima: test of the critical period and match-mismatch hypothesis. Am. Fish. Soc. Symp. 1:455-468. Crecco, V. A., T. Savoy, and L. Gunn. 1983. Daily mortality rates of larval and juvenile American shad {Alosa sapidissima) in the Connecticut River with changes in year-class strength. Can. J. Fish. Aquat. Sci. 40:1719-1728. Crecco, V., T. Savoy, and W. Whitworth. 1986. Effects of density-dependent and climatic factors on American shad, Alosa sapidissima, recruitment: a predic- tive approach. Can. J. Fish. Aquat. Sci. 43:457-463. Gushing, D. H. 1974. The possible density dependence of larval mortality and adult mortality in fishes. In J. H. S. Blaxter (editor), The early life history of fish, p. 21-38. Springer- Verlag, Berlin. 1980. The decline of the herring stocks and the gadoid out- burst. J. Cons. Int. Explor. Mer 39:74-85. FOOTE, P. S. 1976. Blood lactic acid levels and age structure of American shad {Alosa sapidissima, Wilson) utilizing the Holyoke Dam fish lift, MA, 1974 and 1975. M.S. Thesis, Univ. Massachusetts, Amherst, MA, 97 p. GaRROD, D. J., AND J. W. Horwood. 1984. Reproductive strategies and the response to exploita- 479 FISHERY BULLETIN: VOL. 86, NO. 3 TION. In G. W. Potts and R. J. Wooton (editors), Fish reproduction: strategies and tactics, p. 367-384. Acad. Press, London. Goodyear, P. C. 1980. Compensation in fish populations. In C. H. Howatt and J. R. Stauffer, Jr. (editors). Biological monitoring of fish, p. 253-280. Lexington Books, Lexington, MA. GULLAND, J. A. 1965. Survival of the youngest stages of fish, and its rela- tion to year-class strength. Spec. Publ. ICNAF 6, p. 363-373. 1987. Natural mortality and size. Mar. Ecol. Prog. Ser. 39, p. 197-199. Hassell, M. p. 1978. The dynamics of arthropod predator-prey systems. Princeton Univ. Press, Princeton, NJ, 237 p. Hewitt, R. P., G. H. Theilacker, and N. C. H. Lo. 1985. Causes of mortality in young jack mackerel. Mar. Ecol. Prog. Ser. 26, p. 1-10. Hunter, J. R. 1976. Culture and growth of northern anchovy, Engraulis mordax, larvae. Fish. Bull., U.S. 74:81-88. Jones, R. A., P. Minta, and V. A. Crecco. 1976. A review of American shad studies in the Connecticut River. In Proceedings Workshop on American Shad, p. 135-164. U.S. Natl. Mar. Fish. Ser. Knutsen, G. M., and S. Tilseth. 1985. Growth, development, and feeding success of Atlantic cod larvae Gadiis rrwrhua related to egg size. Trans. Am. Fish. Soc. 114:507-511. Layzer, J. B. 1974. Spawning sites and behavior of American shad, Alosa sapidissima, (Wilson), in the Connecticut River between Holyoke and Turners Falls, MA, 1972. M.S. Thesis, Univ. Massachusetts, Amherst, MA, 46 p. Leach, G. C. 1925. Artificial propagation of shad. Rep. U.S. Comm. Fish., 1925, app. VnL459-486. (Doc. 981.) Leggett, W. C. 1969. Studies on the reproductive biology of the American shad (Alosa sapidissima). A comparison of populations from four rivers of the Atlantic seaboard. Ph.D. Thesis, McGill Univ., 125 p. 1976. The American shad, Alosa sapidissima, with special reference to its migrations and population dynamics in the Connecticut River. In D. Merriman and L. M. Thorpe (editors). The Connecticut River ecological study: the impact of a nuclear power plant, p. 169-225. Am. Fish. Soc. Monogr. 1, 252 p. 1977. Density-dependent, density-independent, and recruit- ment in the American shad (Alosa sapidissima) population of the Connecticut River. In W. Van Winkle (editor). Pro- ceedings of the conference on assessing the effects of jx)wer- plant-induced mortality on fish populations, p. 3-17. Pergamon Press, N.Y. Leggett, W. C, and R. R. Whitney. 1972. Water temperature and the migrations of American shad. Fish. Bull., U.S. 70:659-670. Lippson, A. J., and R. L. Moran. 1974. Manual for the identification of early developmental stages of fishes of the Potomac River estuary. MD Dep. Natl. Resour., PPSR-MP-13, 282 p. LoRDA, E., AND V. A. Crecco. 1987. Stock-recruitment relationship and compensatory mor- tality of American shad in the Connecticut River. Am. Fish. Soc. Symp. 1:469-482. Lough, R. G. 1976. Mortality and growth of Georges Bank - Nantucket Shoals herring larvae during the 1975-1976 winter period. Int. Comm. Northwest At!. Fish. Resident Doc. 76/6/123, Ser. No. 4004, 30 p. LuDWiG, D., AND C. Walters. 1981. Measurement errors and uncertainty in parameters estimates for stock and recruitment. Can. J. Fish. Aquat. Sci. 38:711-720. Marcy, B. C, Jr. 1976. Early life history studies of American shad in the lower Connecticut River and the effects of the Connecticut Yankee plant. In D. Merriman and L. M. Thorpe (editors). The Con- necticut River ecological study: the impact of a nuclear power plant, p. 141-168. Am. Fish. Soc. Monogr. I. 252 p. May, R. C. 1974. Larval mortality in marine fishes and critical period concept. In J. H. S. Blaxter (editor). The early life history of fish, p. 3-19. Springer- Veriag, N.Y. Maxfield, G. H. 1953. The food habits of hatchery-produced pond-cultured shad Alosa sapidissima, reared to a length of two inches. Chesapeake Biol. Lab. Publ. No. 98, 38 p. McFadden, J. T. 1977. An argument supporting the reality of compensation in fish populations and a plea to let them exercise it. In W. Van Winkle (editor), Proceedings of the conference on assessing the effects of power-plant-induced mortality on fish populations, p. 153-183. Pergamon Press, N.Y. McGuRK, M. D. 1987. Natural mortality and spatial patchiness: reply to Gulland. Mar. Ecol. Prog. Ser. 39:201-206. Moffitt, C. M., B. Kynard, and S. G. Rideout. 1982. Fish passage facilities and anadromous fish restoration in the Connecticut River. Fisheries 7:2-11. Murray, B. G., Jr. 1982. On the meaning of density dependence. Oecologia (Beri.) 53:370-373. Nichols, P. R. 1958. Effect of New Jersey-New York pound-net catches on shad runs of Hudson and Connecticut Rivers. U.S. Fish Wildl. Serv., Fish Bull. 58:491-500. NUSCo (Northeast Utilities Service Company). 1977. Entrainment analyses of shad eggs in the Connecticut River. Northeast Util. Serv. Co., 167 p. Parrish, R. H., and a. D. MacCall. 1978. Climatic variation and exploitation in the Pacific mackerel fishery. Calif. Dep. Fish. Game Fish Bull. 167, 110 p. Pauly, D. 1980. On the relationships between natural mortality, growth parameters, and mean environmental temperature in 175 fish stocks. J. Cons. Explor. Mer 39:175-192. Peterman, R. M. 1978. Testing for density-dependent marine survival in Pacific salmonids. J. Fish. Res. Board Can. 35:1434-1450. 1982. Nonlinear relation between smolts and adults in Babine Lake sockeye salmon (Onx^orhynchus nerka) and implications for other salmon populations. Can. J. Fish. Aquat. Sci. 39:904-913. RiCKER, W. E. 1954. Stock and recruitment. J. Fish. Res. Board Can. 11: 559-623. 1975. Handbook of computations for biological statistics of fish populations. Bull. Fish. Res. Board Can. 119, 382 P- 480 SAVOY AND CRECCO: MORTALITY OF AMERICAN SHAD ROSENBURG, A. A., AND R. W. DOYLE. 1.986. Analysing the effect of age structure on stock-recruit- ment relationships in herring (Clupea harengus). Can. J. Fish. Aquat. Sci. 43:674-679. Savoy, T. F., and V. A. Crecco. 1987. Daily increments on the otoliths of larval American shad and their potential use in population dynamics studies, /n R. C. Summerfelt (editor), Age and growth of fish, p. 413-431. Iowa State Univ. Press, Ames. SAS. 1982. Statistical analysis system user guide. SAS Institute Inc. Gary, NC, 584 p. SCHERER, M. D. 1974. Analysis of factors affecting passage of American shad {Alosa sapidissima; Wilson) at Holyoke Dam, MA, and assessment of juvenile growth and distribution above the dam. Ph.D. Thesis, Univ. Massachusetts, Amherst, MA, 224 p. Shepherd, J. G., and D. H. Gushing. 1980. A mechanism for density-dependent survival of larval fish as the basis of a stock-recruitment relationship. J. Gons. Int. Explor. Mer 39:160-167. Threlkeld, S. T. 1986. Lifetable reponses and population dynamics of four cladoceran zooplankton during a reservoir flood. J. Plank- ton Res. 8:639-647. U.S. Geological Survey. 1966-87. Water resources data, Connecticut, water year reports. U.S. Geol. Surv., var. pagination. Walburg, G. H., and P. R. Nichols. 1967. Biology and management of the American shad and status of the fisheries. Atlantic coast of the United States, 1960. U.S. Fish Wildl. Serv. Spec. Sci. Rep. 550, 105 P- Ware, D. M. 1980. Bioenergetics of stock and recruitment. Gan. J. Fish. Aquat. Sci. 37:1012-1024. Watson, J. F. 1968. The early life history of the American shad, Alosa sapidissima (Wilson), in the Connecticut River above Hol- yoke, Massachusetts. M.S. Thesis, Univ. Massachusetts, Amherst, MA, 55 p. 1970. Distribution and population dynamics of American shad, Alosa sapidissima (Wilson), in the Connecticut River above Holyoke Dam, Massachusetts. Ph.D. Thesis, Univ. Massachusetts, Amherst, MA, 105 p. White, R. L., J. T. Lane, and P. E. Hamer. 1969. Population and migration study of major anadromous fish. Final Project Report AFSC-1-1, 1-2. NJ Dep. Gonserv. Econ. Dev. Misc. Rep. No. 3M, 6 p. Whitton, B. a. 1980. River Ecology. Univ. Calif. Press, Berkeley, 711 p. Wiggins, T. A., T. R. Bender, V. A. Mudrak, and J. A. Coll. 1984. The development, feeding, growth and survival of cultured American shad larvae through the transition from endogenous to exogenous nutrition. Spec. Publ. Penn. Fish. Comm., Benner Spring Fish Research Station, 34 p. Zarbock, W. M. 1969. Annual progress report, Delaware River Basin Anad- romous Fishery Project, AFS-2-2. U.S. Bur. Sport Fish. Wildl., 20 p. 481 FISHERY BULLETIN; VOL. 86, NO. 3 APPENDIX 1 The total prerecruitment (ZTotal,) and postjuvenile (ZAi) mortality rates from 1979 through 1987 were estimated by ZTotal^ = -\og,iPart * 200,000) - \og,{Rt) and ZAt = log,(J,) - \og,iR,), respectively. The standard errors about ZTotal^ and ZAi were estimated as the sum of the variances (Cochran 1965): ■'ZTotal, - C.2 SE^Tntal. = V + nt n^ and respectively. The variance estimates (5|^, Spavf) of adult recruitment and spawning stock size were calculated from the log transformed estimates from the daily lift rates at the Holyoke Dam, where Uf is the total number of days in which 99% of the American shad were lifted. The variance (Sj) about the juvenile abundance estimates (J^) was based on the log transformed catches per seine haul from all stations and col- lection dates, where rij is the total number of seine hauls made in that year. 482 METABOLIC RESPONSES OF SPOT, LEIOSTOMUS XANTHURUS, AND ATLANTIC CROAKER, MICROPOGONIAS UNDULATUS, LARVAE TO COLD TEMPERATURES ENCOUNTERED FOLLOWING RECRUITMENT TO ESTUARIES Donald E. Hoss, Linda Coston-Clements, David S. Peters, AND Patricia A. Tester^ ABSTRACT The larvae of marine fishes that spawn during fall-winter in coastal North Carolina waters experience a decrease in temperature as they enter estuarine nursery areas. To determine the effect of changes in temperature on larval metabolism, the oxygen consumption of spot, Leiostomus xanthurus, and Atlantic croaker, Micropogonias undulatus, was measured and their QO2 and Q^g values were determined. Atlan- tic croaker respiration decreased with temperature at rates that would be expected if no compensation or stress were involved. Spot showed unexpectedly high respiration rates at low temperature. The in- creased respiration is apparently due to stress. Based on laboratory feeding and growth data, we con- cluded that spot are subject to an energy deficit at <10°C. We infer the timing of larval immigration corresponds with environmental temperatures reaching tolerable levels. Atlantic croaker larvae immigrate earlier in the winter and are exposed to cold water for longer periods than spot larvae. Our conclusion is that stress and energy loss experienced by early immigrating spot larvae may result in increased mortality. The larvae of fishes that spawn during fall and winter in offshore North Carolina waters experience a decrease in both temperature and salinity as they enter estuarine nursery areas (Fig. 1). The spot, Leiostomus xanthurus, and Atlantic croaker, Micro- pogonias undulatus, two sympatric species of Sciae- nidae, are representative of winter spawning species off the North Carolina coast. Previously, we have examined the effects of in- creased temperature on the oxygen and food con- sumption of the postlarval stages of these two species (Hoss et al. 1971, 1974; Peters and Kjelson 1975). In this paper we continue our research on the early life history of these species and evaluate how decreasing water temperature, encountered follow- ing recruitment into estuarine waters, might affect oxygen consumption, food consumption, and ulti- mately survival. Crawshaw et al. (1981) stated that young fish typically select warm shallow water because 1) it permits more rapid growth owing to higher metab- olism, given an adequate food supply, and 2) the predation by larger fish is less in shallow water. For some fish this explanation is plausible, but this does not apply to the larvae of many winter-spawning 'Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Service, NOAA, Beaufort, NC 28516. marine fishes which begin life in relatively warm coastal waters and then enter colder estuarine waters. For these species we expect that metabolism and growth of the estuarine immigrants should be reduced (Brett 1956). The specific objective of this paper is to describe how decreasing temperature affects the metabolism of larval fish as they are moved from warm to cold water by a combination of passive and active transport mechanisms that are not, as yet, completely understood. Oxygen con- sumption is a common method of estimating meta- bolic activity, which frequently changes in response to environmental conditions (O'Hara 1968). In this study we measured routine oxygen consumption which is the amount used by fish whose only move- ments are spontaneous. STUDY AREA AND METHODS Spot and Atlantic croaker spawn off the North Carolina coast inshore of the Gulf Stream over the continental shelf (Hildebrand and Cable 1930; Dawson 1958; Powles and Stender 1978). Here, spot spawn from October to February, but principally from December to January while Atlantic croaker spawn from September to May but principally be- tween October and December (Lewis and Judy 1983). After between 30 and 60 days in coastal Manuscript accepted May 1988. FISHERY BULLETIN: VOL. 86. NO. 3, 1988. 483 North Carol ^^6%o 10 km 21°C HOSS ET AL.: METABOLIC RESPONSES OF SPOT AND ATLANTIC CROAKER waters (Warlen 1982; Warlen and Chester 1985), the larvae enter estuaries where they develop into juveniles. In the spawning area, water temperatures are usually between 18° and 25°C (Fahay 1975; Hettler and Powell 1981). The fish encounter de- creasing temperatures as they move inshore to the estuarine nursery areas. In the lower Newport River estuary, for example, mean water temperatures be- tween November and March may range from 14° to 6°C with the highest temperatures during this period occurring in November and the lowest in January (Hoss 1974). Larvae of Atlantic croaker and spot were obtained from both field collections and eggs spawned in the laboratory. Older larvae were captured in a bridge net (Hettler 1979) and held in the laboratory for no more than a week prior to use. First feeding larvae were obtained from spawned fish, reproduced by the methods of Hettler and Powell (1981), and then were reared at experimental temperatures. Oxygen consumption was measured with a differ- ential respirometer (Umbreit et al. 1964), following procedures used by Hoss, Hettler, and Coston (1974). Fish were transferred to 15 mL respiration flasks and, following a 2-h acclimation period, their oxygen consumption was measured. Numbers of lar- vae per flask varied between 1 and 30, depending on the size of the larvae. Acclimation temperatures were 10°, 15°, and 20°C. Notochord or standard lengths and dry weights were obtained for individual fish. The metabolic equation Q = aW^, was used to describe the relation between oxygen consumption and dry weight of fish acclimated at 10°, 15°, and 20°C. In this equation, W is the weight of the fish, and a and k are constants for the species obtained from least-squares regression of the log of oxygen consumption on the log of weight (Winberg 1956). A k value of 0.67 implies that oxygen consumption varies in proportion to surface area, whereas a value of 1 indicates that respiration varies in proportion to weight. We used the metabolic equation to estimate oxy- gen consumption of larvae of equal weight at differ- ent temperatures. We compared larvae of 4 mg dry weight because this is the realistic estimate of their weight as they are transported from coastal to estu- arine waters (Warlen 1982; Warlen and Chester 1985). Growth and feeding rates of small spot (^20 mm SL) collected from the Newport River were calcu- lated from data collected in the laboratory at several temperatures. Wet weights (^15-30 mg) were re- corded to the nearest milligram, and 10 fish were randomly assigned to 4 L test and control con- tainers. Control fish were dried to determine the dry/wet weight ratio, which was then used to esti- mate initial dry weight of experimental fish. One experiment was conducted at 6°, 8°, 10°, 12°, and 16°C, and two experiments were conducted at 18°C. In all cases fish were fed newly hatched brine shrimp several times a day to assure an ad libitum food supply. After 4-6 days all food was removed; lar- vae were allowed time to clear their guts and then were dried and weighed. Growth and feeding rates were expressed as per- cent of body weight per day. Growth rate was calculated from the expression: Growth rate = 100 [(WwAVo) l/n 1] where Wn = dry weight of all fish in a tank at day n Wo = estimated original dry weight of fish n = number of days fed. Calculation of feeding rates required the assump- tion of constant growth rates. Using original dry weights and calculated growth rates we determined the total dry weight of fish in each container at the beginning of each day. Dry weights of brine shrimp eaten each day divided by the calculated dry weights of fish gives proportion of body weight ingested. These proportions were then summarized as aver- age daily percent of dry body weight ingested. In order to compare metabolic parameters, i.e., oxygen consumption, feeding, and growth rates, the following conversion factors were used: 1.0 mg dry wt = 5.0 cal (Thayer et al. 1973; Paffenhoffer 1967); 1.0 mg O2 = 3.38 cal (Phillipson 1966) and 0.7 mg O2 = 1 mL O2 at STP. One tenth calorie per fish per day was added to all the rates so that measured zeros could be shown on a log scale. RESULTS AND DISCUSSION The regression equations relating oxygen con- sumption to weight at several temperatures are shown in Table 1. Higher coefficient of determina- tion (R^) values were found at higher temperatures, a trend best explained by differences in the size range of fish measured at different temperatures. Values for k, were generally comparable to values reported by other investigators for fish of a similar size and at comparable temperatures— Hoss (1974), pinfish; Houde and Schekter (1983), bay anchovy, sea bream, and lined sole; Almatar (1984), herring; and Laurence (1978), cod and haddock. 485 FISHERY BULLETIN: VOL. 86, NO. 3 Table 1.— Metabolic equations relating oxygen consumption to body size. Q = aW* where Q = oxygen consumption (piL Oj ■ h"'), a and k are regression coefficients, W is dry wt in mg, N is tfie number of observations, T the temperature, and Rj the coef- ficient of determination. Species T N W a k R2 Atlantic 10 52 1.50 - 6.43 1.80 0.66 0.42 croaker 15 65 1.90 -11.71 2.00 1.02 0.65 20 81 0.013- 9.31 3.59 0.86 0.94 Spot 10 37 1.82 - 8.01 3.63 0.70 0.54 15 48 2.18 -10.27 3.15 0.64 0.63 20 79 0.014- 8.91 4.07 0.92 0.97 Comparing between measured oxygen consump- tion rates at 10° and 15°C and those predicted from Van't Hoff's equation (Vernberg and Vernberg 1972), we conclude that Atlantic croaker show no sign of regulating their oxygen consumption as water temperature is decreased. The difference be- tween oxygen consumption rates based on Q^q values of 2 and 3 (Fig. 2) is an expected range. For every 10°C change in temperature, the rate of a chemical reaction typically changes by a factor of 2 to 3. A Qio value of appreciably <2 or more than a- z O h- o. ID CO z o o X o \. ^ *^^ -^r ^9^ K^"^ yur *H \, ,?■ -■ 18jiii3e.lkU 1.31E3 6717/99 88676 d a» 10jiii30.1kU 1.62E3 6749/99 SS014B Figure 3c. —Porous crystals on the medial surface of the regular lobe. Scale bar = 10 tixn. d.— Honeycomblike crystal structure on the medial surface of the irregular lobe. Scale bar = 10 fim. 505 FISHERY BULLETIN: VOL. 86, NO. 3 Figure 3e.— A laminar pattern of crystalline growth on the medial face of the irregular lobe. Scale bar = 10 fxm. edge of this lobe, a laminar crystalline growth pat- tern develops, adding to the overall variation of crystal structures on the medial surface (Fig. 3e). Other studies of otolith ultrastructure present the otolith as homogeneous in crystal form, composed almost entirely of monoclinal aragonite crystals (Degens et al. 1969). The complex crystallinity of the oreo otoliths resembles that of the mollusc shell, v^hich, although aragonitic, often has a pattern of complex variation in crystal habit (Carriker et al. 1980). Within the broken otolith, the nucleus lies at the center of a spherical primordium (Fig. 4a). Crystals grow outwards from the primordium and epitaxial (Degens 1976) growth patterns exist (Fig. 4b). Com- plex leaf-shaped crystals occur in areas directly beneath the lateral surface of the fractured irregular lobe (Fig. 4c). Beneath the medial surface, a remark- able series of hexogonal crystals of calcite occur as large rectangular blocks embedded within the otolith (Fig. 4d). Calcitic prisms have been described in molluscs as resulting from the regeneration of broken shells (Watabe 1983). It is difficult to imagine otoliths being broken and regenerated in situ. Major and minor check rings similar to those described (Gaul die 1987) for otoliths from the orange roughy, a deepwater species from the same habitat, occur (Fig. 4e). When polished and etched, the check rings become clearly visible with deep etched checks occurring between the less deeply etched checks analagous to microscopic growth increments in other species (Fig. 4f). At a higher magnification the finer increments become obscured by the coarseness of the underlying crystal type (Fig. 4g). Those fine increments that are visible occur irregularly and have varying widths. Large-scale rings analogous to opaque/hyaline an- nual zones were observed in the regular lobe. The mean width of these zones, measured using transmitted light (Fig. lb), was 0.34 mm (±0.06). The concentric ridges observed by SEM on the lateral surface of the regular lobe (Fig. Ic) have a mean width of 0.27 mm (±0.06). These two struc- tures, opaque/hyaline zones and surface ridges, have about the same width with no statistically signifi- cant differences between them. Examination of otolith cross sections reveals widely spaced, large rings with many finer rings in between (Fig. 4h). When examined in greater detail, the demarcation between the large and finer rings 506 DA VIES ET AL.: OTOLITH ULTRASTRUCTURE OF SMOOTH AND BLACK OREO a Figure 4a. —Broken smooth oreo otolith reveals the spherical primordium (PM) and nucleus (N) and the surround- ing radial crystal growth. Scale bar = 0.1 mm. b.— Epitaxial crystal development in the broken otolith. Scale bar = 0.1 mm. 507 FISHERY BULLETIN: VOL. 86, NO. 3 10iim30.lkU 1.49E3 6828/99 SSOl Figure 4c.— Leaf-shaped crystals beneath the lateral surface of the broken otolith. Scale bar = 10 ^m. d.— Calcitic prism crystals beneath the medial surface of the broken irregular lobe. Scale bar = lOji. 508 DAVIES ET AL.: OTOLITH ULTRASTRUCTURE OF SMOOTH AND BLACK OREO e - Figure 4e.— Discontinuous uniform crystal growth forming rings in the broken otolith. Scale bar = 0.1 mm. f. —Fine and deep (arrowed) increments on the polished and etched otolith surface. Scale bar = 0.1 mm. 509 FISHERY BULLETIN: VOL. 86, NO. 3 g ' i . V pit 10iiin30.1kU 1.31E3 6725/99 SSeSB Figure 4g.— Fine increments obscured by coarse crystallinity on the polished and etched otolith surface. Scale bar = 10 jim. h.— Fine increments (arrowed) found between widely spaced large rings in the otolith cross-section. Magnification = 72x . 510 DAVIES ET AL.: OTOLITH ULTRASTRUCTURE OF SMOOTH AND BLACK OREO j r , - ' • . • . v- .4, o:^^"^^i*i ^.^ X \,* •^-.^r ^ ^:j-j 5^- ,-* « ^ ■^- Figure 4i. —Variation of increment widths (arrowed) in the otolith cross-section. Magnification = 160 x. j.— Cellulose-acetate peel showing the intermittency of fine increment sequences (arrowed), often obscured by a coarse underly- ing crystallinity. Magnification = 1 60 x . 511 FISHERY BULLETIN: VOL. 86, NO. 3 becomes difficult to determine because of the inter- mittent nature of the growth increments and the variety of width-sizes (Fig. 4i). In large areas of the otolith, increments appear to be absent or indeter- minate, with a coarse underlying crystal structure (Fig. 4j) making accurate determination of incre- ment sequences difficult. However, the fine incre- ments of the oreo otolith are 3 to 5 ^m wide, which is within the range of daily growth increments described for other species (Jones 1986; Gauldie in press). Black Oreo Otolith The black oreo otolith is almost identical to that found in the smooth oreo in overall shape, propor- tion, structure, topography, surface and internal crystallinity, and increment pattern. Some minor differences do, however, exist. In the medial sulcus, the prominent knobs found in the smooth oreo otolith are smaller than in the black oreo. On the surface of the sulcus, large leaf- like crystals having various orientations occur (Fig. 5a). Also present in the sulcus are porous, sponge- like crystals adjacent to membranous structures (Fig. 5b). Smooth patches, where crystals appear ab- sent, occur on the lateral surface of the otherwise coarsely crystalline irregular lobe (Fig. 5c). At higher magnification the smooth patches are seen to be smaller growth forms of the larger adjacent crystals. DISCUSSION Despite some minor differences in topography and crystallinity, the sagittae of both species are essen- tially identical. The otoliths are structurally complex v^ath a great variety of crystalline forms. The coarse Figure 5a. —Split-screen SEM of large leaflike crystals in the black oreo otolith sulcus. Magnification = 356 x and 979 x . b.— Porous, sponge-like crystals in the otolith sulcus. Scale bar = 10 ^^m. c. —Smooth patches on the central lateral surface of the otolith. Scale bar = 0.1 mm. 512 DA VIES ET AL.: OTOLITH ULTRASTRUCTURE OF SMOOTH AND BLACK OREO 513 FISHERY BULLETIN: VOL. 86, NO. 3 crystals on the central lateral surface are com- parable with those in the oyster shell described as individual laths (Carriker et al. 1980). The leaflike crystals in the sulcus of the black oreo otolith are similar to the chalky crystal forms in the oyster shell (Carriker et al. 1980). Such a variety of crystalline forms is uncommon in teleost otoliths. The low legi- bility of structures of various kinds in the otolith may reflect this complex crystallinity. However, the com- plex crystallinity of the mollusc shell is thought to reflect changes in both the external and internal milieu of the organism (Wilbur and Saleuddin 1983). Thus the difficulties of reading the oreo otolith in the conventional sense may be offset by the life history record (albeit difficult to translate) provided by its complex crystallinity. The broken sections of the otolith reveal the in- ternal structure organization, and development of crystals. Epitaxial crystal grov^h in the oreo otolith results in columnar, monoclinal crystals of ara- gonite. However, the presence of calcite-like prisms has not been described for other otoliths. Calcite occurs on the antisulcal surface of some otoliths apparently by simple crystallization out of the fluid of the endolymphatic sac (Morales-Nin 1985), but calcite has never been described from within an aragonite otolith (Carlstrom 1963). In molluscs, calcite replacement of aragonite results in an orderly alignment of calcite crystals following the alignment of the original aragonite crystals. The disorderly appearance of the calcite-like hexagons in the smooth oreo otolith may be due to a diagenetic transformation of aragonite to calcite with depth. The compensation depth for the aragonite/calcite transformation is about 3,000 m (Fyfe and Bischoff 1965) well beyond the known range of the smooth oreo which has a maximum recorded depth of 1,300 m. However, there may be enough variation in either the kind or amount of stabilizing protein in the smooth oreo otolith to allow crystal changes to occur at shallower depths than 3,000 m. The complex, and often coarse, crystal structure of the oreo otolith obscures the sequences of incre- ments when they do occur. As a result, large rings observed at low magnification become indistinct at higher magnification when many finer increments appear. The large rings could be assumed to be annual check rings, but the difficulties in differen- tiating between the fine and large rings create am- bivalence in one's interpretation. The finer micro- scopic growth increments, analagous to daily growth rings, have no uniform width and occur intermit- tently making accurate counting impossible. The suitability of an otolith for determining the age of a fish depends on the pattern of both annual and daily check rings inferred from the structure of the otolith. The hyaline/opaque zones observed in the regular lobe (using transmitted light) had a similar mean width to the concentric ridges found by SEM on the lateral surface. The ^statistic we obtained accepts the null hypothesis that no sig- nificant difference exists between the two means. Mel'nikov (1981) regarded these opaque/hyaline zones as annual check rings in the otolith of Allo- cyttus verrucosus. However, because no evidence ex- ists for a relationship between the surface ridges and fish age, it is possible that Mel'nikov's (1981) ages are incorrect. Furthermore, the width of the opaque/ hyaline zones (0.34 mm) would indicate daily growth increments less than 1 \xm wide. There are no reports in the literature of validated daily growth rings of such small size. In addition, the microscopic growth increments which we have observed in the oreo otolith are 3 to 5 ^m wide, which is a size range commonly observed in other species. With the techniques available we have been unable to use either annual- or daily-type structures to develop a technique for age estimation for Pseudo- cyttus maculatus and Allocyttu^ sp. The reasons for these difficulties may lie in the crystal morphology of the otoliths which are more complex than any so far described in the literature. ACKNOWLEDGMENTS All otoliths were supplied by Peter McMillan (Fish- eries Research Centre). All 35 mm photographs were processed by Alan Blacklock (Fisheries Re- search Centre). SEM photographs were taken at the SEM Unit, Zoology Department, Victoria Univer- sity of Wellington. LITERATURE CITED Carlstrom, D. 1963. A crystallographic study of vertebrate otoliths. Biol. Bull. (Woods Hole) 125:441-463. Carriker, M. R., R. E. Palmer, and R. S. Prezant. 1980. Functional ultra-morphology of the dissoconch values of the oyster Crassostrea virginica. Proc. Natl. Shellfish. Assoc. 70:139-183. Degens, E. T. 1976. Molecular mechanisms of carbonate, phosphate and silica deposition in the living cell. Top. Curr. Chem. 64: 1-112. Degens, E. T., W. G. Deuser, and R. L. Haedrich. 1969. Molecular structure and composition of fish otoliths. Mar. Biol. 2:105-113. Fyfe, W. S., and J. L. Bischoff. 1965. The caJcite-aragonite problem. In L. C. Pray and R. C. Murray (editors), Dolomitization and limestone diagenesis, 514 DA VIES ET AL.: OTOLITH ULTRASTRUCTURE OF SMOOTH AND BLACK OREO p. 3-13. Symp. Soc. Econ. Minist. Spec. Publ. 13. Gauldie, R. W. 1987. The fine structure of check rings in the otolith of the New Zealand orange roughy. NZ J. Mar. Freshwater Res. 21:267-274. In press. A study of otolith daily growth rings in the orange roughy {Hoplostethus atlanticus) aimed at resolving prob- lems in age, growth, recruitment and otolith architecture. NZ Fish. Res. Bull. Gauldie, R. W., D. Dunlop, and J. Tse. 1986. The remarkable lungfish otolith. NZ J. Mar. Fresh- water Res. 20:81-92. Jones, C. 1986. Determining age of larval fish with the otolith incre- ment technique. Fish. Bull., U.S. 84:91-103. Last, P. R., E. 0. G. Scott, and F. H. Talbot. 1983. Fishes of Tasmania. Hobart, Tasmanian Fisheries Development Authority. Mel'nikov, Y. S. 1981. Size-age composition and growth pattern oiAUocyttus verrucosus (Oreosomatidae). Ichthyol. (Engl, transl. Vopr. Ikhtiol.) 21:178-184. Morales-Nin, B. 1985. Caracteristicas de los otolitos cristalinos de Genypterus capensis, (Smith, 1847) (Pisces: Ophidiidae). Invest. Pesq. 49:379-386. McMillan, P. J. 1985. Black and smooth oreo dories, /n J. A. Colman, J. L. McKoy, and G. G. Baird (editors), Background papers for the 1985 total allowable catch recommendations. (Report held in the Fisheries Research Centre Library, P.O. Box 297, Wellington, New Zealand.) Nelson, J. S. 1976. Fishes of the world. John Wiley and Sons, N.Y., 416 p. Watabe, N. 1983. Shell repair. In A. S. M. Saleuddin and K. M. Wilbur (editors). The mollusca, No/4, p 289-316. Acad. Press, N.Y. Wilbur, K. M., and A. S. M. Saleuddin. 1983. Shell formation. In A. S. M. Saleuddin and K. M. Wilbur (editors), The mollusca, No/4, p 236-288. Acad. Press, N.Y. 515 RELATIONSHIP BETWEEN SEDIMENT CHARACTER AND SEX SEGREGATION IN ENGLISH SOLE, PAROPHRYS VETULUS' D. Scott Becker^ ABSTRACT English sole, Parophrys vetulus, were collected by otter trawl in the nearshore zone of Puget Sound, Washington, during two surveys conducted in 1981-82 and 1984. Stations in both surveys were distributed across a broad range of sedimentary environments. Sex segregation by English sole was strongly associated with the grain-size characteristics of bottom sediments. This association was persistent across a variety of sampling conditions, including different years, seasons, embayments, and depths. Fish age did not appear to influence the observed association. Results of this study provide the first documenta- tion of the influence of sediment character on sex segregation by a pleuronectid, and suggest that this relationship should be considered in future studies of English sole and, perhaps, other pleuronectids as well. Sex segregation by fishes of the family Pleuronec- tidae (i.e., righteye flounders) has been documented for numerous species from the United States (Alver- son and Chatwin 1957; Fadeev 1970; Alton 1972), Europe (Rae 1965), and Asia (Moiseev 1953; Kovtsova 1982). These patterns have been related to such factors as geographic location, depth, season, and age of fish. However, because most of these observations have been made while fishes were being collected for other purposes, the under- lying reasons for the observed patterns are poorly understood. Several studies have documented sex segregation in English sole, Parophrys vetulus, a pleuronectid found in nearshore areas along the west coasts of the United States and Canada (Hart 1973). In de- scribing the catch of the English sole fishery in Hecate Strait, British Columbia, Ketchen (1956) noted that considerable sex segregation takes place, particularly during the summer months. In a study of the population characteristics of this species in Puget Sound, Holland (1969) found a difference of 18% between the percentages of males in the pop- ulations from two locations. Because both of these accounts were largely anecdotal, no evaluations of the observed patterns were made. One factor that has yet to be evaluated with respect to sex segregation by pleuronectids is the character of bottom sediments. Because pleuronec- tids live in close association with the sea floor, it 'Contribution No. 759, School of Fisheries, University of Wash- ington, Seattle, WA 98195. ^School of Fisheries, University of Washington, Seattle, WA 98195; present address: PTI Environmental Services, 3625 132nd Ave. S.E., Bellevue, WA 98006. might be surmised that the characteristics of that environment exert a differential influence on the distributions of the two sexes. Additional support for this hypothesis comes from the fact that the influence of sediment character on the overall distribution (i.e., both sexes pooled) of several pleuronectids has been well documented (Ketchen 1956; Alverson et al. 1964; Feder et al. 1974; Pearcy 1978; Scott 1982; Becker 1984). In this paper, I evaluate patterns of sex segrega- tion by English sole in relation to the character of bottom sediments in nearshore areas of Puget Sound, WA. This study provides the first determina- tion of how sediment character influences sex segre- gation in a pleuronectid. In relating sediment character to sex segregation in English sole, I also evaluate the potential influences of depth, season, and fish age on the observed relationships. METHODS English sole were sampled during two independ- ent surveys conducted in several embayments of Puget Sound in 1981-82 and 1984. The relationship between sex segregation and sediment character was evaluated separately for each survey. Although the two surveys differed with respect to such vari- ables as station locations, sampling times, and sam- ple sizes, the basic collection and analytical methods used in both surveys were similar. In both surveys, the sex ratio of the English sole captured at each station was represented by the pro- portion of the sample comprised of males (i.e., male proportion). Sediment character at each station was represented by the percentage (by weight) of the Manuscript accepted April 1988. FISHERY BULLETIN: VOL. 86, NO. 3, 1988. 517 FISHERY BULLETIN: VOL. 86, NO. 3 grain-size distribution comprised of fine-grained material (i.e., <0.063 mm in size). This variable was used as an index of sediment character because many physical, chemical, and biological character- istics of sedimentary environments covary with the grain-size distribution (Buchanan 1984). During 1981-82, English sole were collected from 15 May to 6 July 1981 at 12 transects in Commence- ment Bay, Elliott Bay, and Sinclair Inlet (Fig. 1). Fishes also were collected during three additional time periods (23-24 November 1981; 8-9 March 1982; 30-31 August 1982) to evaluate seasonal patterns at three stations in Commencement Bay (CB-1, CB-4, and CB-5; Fig. 1). These stations were selected to represent a gradient of sediment grain size. Although the seasonal sampling occurred over a 2-yr period, it was assumed that interannual varia- tion was not sufficiently large to obscure major seasonal patterns. During the May- July sampling, English sole were collected at each transect during morning (07:00- 09:30), midday (10:00-13:30), evening (20:30-23:30), and midnight (00:30-04:00). During the remaining sampling periods, fish were collected only during midday and evening. For each transect, results ob- tained for different periods of the diel cycle were pooled before male proportions were calculated. Sex determinations and evaluations of spawning condi- tion were made for all English sole larger than 160 mm total length (TL). Age was not determined. During 1984, English sole were collected from 4 to 9 June 1984 at 15 transects in Commencement Bay and at 2 transects in Carr Inlet (Fig. 1). Sam- pling was conducted during daylight hours (06:30- 17:00) at all transects. Sex determinations and evaluations of spawning condition were made for all individuals larger than 225 mm TL. Age was deter- mined by otolith analysis for all fish. During both surveys, English sole were collected using a 7.6 m (headrope) Marinovich otter trawl having a body mesh of 3.2 cm (stretched) and a cod end liner mesh size of 0.8 cm (stretched). Trawling was conducted along isobaths at a constant vessel speed of approximately 2.5 knots. During both surveys, bottom sediments were sam- pled using a modified van Veen bottom grab and the grain-size distribution of the top 2 cm of sediment was determined using standard sieve and pipette techniques (Folk 1968). During 1981-82, sediments were collected at five sampling points located at approximately equal distances along each transect. During 1984, sediments were collected at a variable number of stations located within 300 m from each transect. The association between percent fine-grained sediment and male proportion of English sole was tested using Spearman's rank correlation coefficient {Vg). The association between station depth and male proportion was tested in the same manner. Differences in male proportion among age-classes of fish and among the seasonal sampling transects were tested using the G-test (Sokal and Rohlf 1981). RESULTS Station depth, percent fine-grained sediment, sample size, and male proportion at each transect from the 1981-82 and 1984 surveys are presented in Tables 1 and 2, respectively. During 1981-82, 4,430 English sole were sampled from stations ranging in depth from 8 to 36 m. Percent fine- grained sediment ranged from 5.8 to 92.1% and male proportion ranged from 0.20 to 0.91. During 1984, 1,007 English sole were collected from stations ranging in depth from 5 to 20 m. Per- cent fine-grained sediment ranged from 8.6 to 87.5% and male proportion ranged from 0.13 to 0.98. The correlation between percent fine-grained sedi- ment and male proportion of English sole was significant (P < 0.01; r, = 0.73) for the May- July sampling of the 1981-82 survey (Fig. 2). This cor- relation was also significant (P < 0.001; r^ = 0.80) for the 1984 survey (Fig. 2). Correlations between depth and male proportion were not significant {P > 0.05) for either the 1981-82 (r, = -0.10) or 1984 (Vg = -0.04) survey. For the stations sampled seasonally in 1981-82 (CB-1, CB-4, and CB-5), heterogeneity of male pro- portion among stations was significant {P < 0.05) during all four sampling periods (Fig. 3). In all cases, male proportion was lowest at CB-5, highest at CB-4, and intermediate in magnitude at CB-1. This gradient of increasing male proportion was consis- tent with the gradient of increasing percent fine- grained sediment at the three stations (Table 1). The mean abundance of English sole at the three sta- tions rose from a minimum level in March to a peak in May-July, and then declined through August to a level in November that was similar to that in March (Fig. 3). For both surveys combined, 64 (1.2%) English sole were found to be in spawning condition. Twenty- four of those individuals were collected during March, to represent a spawning component of 14.8% during that month. The remaining 42 individuals were sampled between May and July to represent 518 BECKER: SEDIMENT AND SEX SEGREGATION IN ENGLISH SOLE r^ ^^ Elliott Bay ^^EB-1 /^ '^^EB-2 SEATTLE I- EB-4 LEGEND CR-Carr Inlet E B - Elliott Bay SN - Sinclair Inlet 10 ii NALmCAL MLES n POINT DEFIANCE LEGEND HY Hyielxjs Waterway BL Blair Waterway SI Silcum Waterway Ml Milwaukee Waterway SP St Paul Waterway MD Middle Waterway CI City Waterway SS Souttiwest Shoreline CB Commencement Bay KILOMETERS Figure 1.— Locations of transects sampled in 1981-82 (denoted by CB, EB, and SN) and 1984 (all remaining transects). 519 FISHERY BULLETIN: VOL. 86, NO. 3 Table 1. — Characteristics of transects and samples of English sole during 1981-82.' Mean percent fine-grained SampI e size^ Male proportion Depth May- Novem- May- Novem- Station^ (m) sediment^ July August ber March July August ber March CB-1 10 27.1 655 249 65 90 0.83 0.74 0.58 0.57 CB-2 33 47.0 554 0.78 CB-3 36 44.2 160 0.79 CB-4 14 71.6 589 146 174 38 0.91 0.82 0.77 0,87 CB-5 13 7.6 181 67 68 35 0.55 0.64 0.44 0.20 CB-6 35 11.2 128 0.66 EB-1 14 56.0 82 0.68 EB-2 31 76.2 17 0.76 EB-3 16 5.8 90 0.53 EB-4 32 8.8 49 0.49 SN-1 8 92.1 683 0.84 SN-2 10 13.7 310 0.30 'Sampling in May-July and November was conducted in 1981 , whereas sampling in August and March was conducted in 1982. ^Locations of transects are presented in Figure 1 . 'Each value Is based on five replicate samples. 0.05). Nineteen fish aged 2 or >10 years old were not included in this analysis because sample sizes by age-class were too small. OC o Q. o DC Q. LU 0.8 -1 0.6- 0.4 0.2- 126 217 261 209 97 40 24 14 Figure 4.— Comparison of male proportion among age- classes of English sole collected in 1984. Sample size is presented above each bar. T 1 \ \ \ 1 I r 3456789 10 FISH AGE (yr) 521 FISHERY BULLETIN: VOL. 86, NO. 3 DISCUSSION Results of this study demonstrate that sex segre- gation by English sole in nearshore areas of Puget Sound is strongly associated with the grain-size distribution of bottom sediments. In addition to the results of statistical analyses, the strength of this association is demonstrated by its persistence across a variety of sampling conditions. For example, the association was found in two independent surveys conducted 2-3 years apart. Furthermore, the 1981- 82 survey represented an analysis of large-scale patterns across three embayments, whereas the 1984 survey focused on small-scale patterns within a single embayment. The association between sex segregation and sediment grain size also persisted across seasons and across a range of depth. Although a significant association was found be- tween sediment grain size and male proportion of English sole, a variety of factors that covary with sediment grain size may have been partly or totally responsible for this association (Lagler et al. 1977; Moyle and Cech 1982). Examples of physical/chem- ical variables that often covary with grain size and may influence fish distributions include depth, cur- rent speed, turbidity, and dissolved oxygen. Ex- amples of biological factors include the composition of prey assemblages and the distribution of pred- ators. Although results of this study demonstrated that depth did not influence the observed associa- tion between sediment grain size and male propor- tion, further experimentation is required to identify the factor or suite of factors that directly influences sex segregation in this species. Despite the fact that depth did not influence the patterns of sex segregation identified for English sole, the range of depth considered in this study (5-36 m) is a small fraction of the total range occupied by this species (0-550 m; Hart 1973). It therefore is possible that depth-related sex segrega- tion in this species may be observed if a wider range of depths were examined. For example, considerable sex segregation by depth was found for Dover sole, Microstomus pacificus, off the west coast of the United States when a large depth range (110-440 m) was evaluated (Hagerman 1952; Alton 1972). Based on results of the 1984 survey, no relation- ship between age and male proportion in English sole was evident. It therefore is unlikely that age influenced the observed relationship between male proportion and sediment grain size. A decline in male proportion with increasing age has been found for several pleuronectids (Rae 1965; Kovtsova 1982). Because most of the fish examined in this study were relatively young (<8 years old) compared with the maximum age of 15-17 years sometimes reached by individuals of this species (Ketchen 1956; Holland 1969; Van Cleve and El Sayed 1969), consideration of older fish may alter the patterns observed in this study. Although sample sizes for fish aged 9-10 years were relatively small in this study, male pro- portion in those age groups (0.50) was lower than that observed in all younger age groups (0.57-0.66). The season of sampling did not alter the observed relationship between male proportion and sediment grain size. This relationship remained intact despite the fact that the mean abundance of English sole changed substantially throughout the seasonal cycle. The observed seasonal changes in mean abundance are consistent with the seasonal pattern of migra- tion exhibited by adults of this species and most other pleuronectids. That pattern includes a migra- tion to deeper water in the fall for overwintering and spawning, and a return migration to shallower water in spring for feeding throughout the summer (Ketchen 1956; Alverson et al. 1964; Roff 1982). It is unlikely that spawning substantially influenced the observed seasonal patterns of sex segregation in English sole, because the largest percentage of fish in spawning condition during any season was <15%. However, sex segregation in relation to spawning in pleuronectids has been documented for several species (Moiseev 1953; Alverson and Chat- win 1957), and may influence sex segregation in English sole when a substantial percentage of the population is in spawning condition. Results of this study suggest that the influence of sex segregation should be considered when characteristics of English sole populations are com- pared among habitats that differ with respect to sediment character. Patterns based on population characteristics capable of exhibiting a dependence on sex (growth, condition, disease prevalence, tissue contamination) could be strongly biased if differ- ences in the sex ratio among stations are not con- sidered. This concern should probably be extended to studies of other pleuronectids, because the rela- tionship between sex segregation and sediment character is unknown for most species. Although the reasons for the observed patterns of sex segregation in English sole are unknown, a potential explanation is that different energetic needs induce male and female fish to utilize differ- ent habitats. Female English sole grow to a larger size than males and mature at an older age (Ketchen 1956; Holland 1969; Van Cleve and El-Sayed 1969; Hart 1973). In addition, fecundity in females is pro- portionate to individual size (Ketchen 1956; Hart 522 BECKER: SEDIMENT AND SEX SEGREGATION IN ENGLISH SOLE 1973). These patterns are found for numerous pleuronectids and are thought to allow females to outgrow predators and maximize fecundity (Roff 1982). Female English sole may therefore select different habitats than males to support their differ- ential reproductive requirements. Because many characteristics of benthic macroinvertebrate assem- blages (i.e., the primary prey of English sole) are related to sediment character (Gray 1974; Rhoads 1974; Pearson and Rosenberg 1978), selection of habitats that differ in sediment grain size provides females with different prey spectra and energetic potentials than are available to males. In addition, differential habitat selection would reduce the poten- tial of competition for food between sexes. In summary, sex segregation by English sole was strongly associated with the grain-size character- istics of bottom sediments. This association was persistent across a variety of sampling conditions, including different years, seasons, embayments, and depths. Fish age did not appear to influence the observed association. Results of this study provide the first documentation of the influence of sediment character on sex segregation in a pleuronectid, and suggest that this relationship should be considered in future studies of English sole and, perhaps, other pleuronectids as well. ACKNOWLEDGMENTS The 1981-82 survey was supported by the National Oceanic and Atmospheric Administration (Contract NA80RAD00050), and was part of a dissertation submitted to the School of Fisheries of the University of Washington (Seattle) in partial fulfillment of a Ph.D. degree for the author; A. J. Mearns was the Project Officer. The 1984 survey was supported under Cooperative Agreement No. CX810926-01-0 between the U.S. Environmental Protection Agency and the State of Washington Department of Ecology; J. D. Krull of Ecology was the Project Manager. I thank R. A. Pastorok and two anonymous reviewers for their helpful sugges- tions on the manuscript. LITERATURE CITED Alton, M. S. 1972. Characteristics of the demersal fish fauna inhabiting the outer continental shelf and slope off the northern Oregon coast. In A. T. Pruter and D. L. Alverson (editors), The Columbia River estuary and adjacent ocean waters, p. 586- 634. Univ. Wash. Press, Seattle. Alverson, D. L., and B. M. Chatwin. 1957. Results from tagging experiments on a spawning stock of petrale sole, Eopsetta jordani (Lockington). J. Fish. Res. Board Can. 14:953-974. Alverson, D. L., A. T. Pruter, and L. L. Ronholt. 1964. A study of demersal fishes and fisheries of the north- western Pacific Ocean. H. R. MacMillan Lect. Fish., Inst. Fish., Univ. B.C., Vancouver, 190 p. Becker, D. S. 1984. Resource partitioning by small-mouthed pleuronectids in Puget Sound, Washington. Ph.D. Thesis, Univ. Wash- ington, Seattle, 120 p. + appendices. Buchanan, J. B. 1984. Sediment analysis. In N. A. Holme and A. D. Mclntyre (editors), Methods for the study of marine benthos, p. 41-65. Blackwell Scientific Publications, Boston, MA, 387 p. Fadeev, N. S. 1970. The fishery and biological characteristics of yellowfin soles in the eastern part of the Bering Sea. Transl. in Soviet Fisheries Investigations in the northeastern Pacific (Israel Program Sci. Transl), Part V, p. 332-396. Feder, H. M., C. H. Turner, and C. Limbaugh. 1974. Observations on fishes associated with kelp beds in southern California. Calif. Fish Game, Fish Bull. 160, 138 p. Folk, R. L. 1968. Petrology of sedimentary rocks. Univ. Texas, Austin, 170 p. Gray, J. S. 1974. Animal-sediment relationships. Oceanogr. Mar. Biol. Annu. Rev. 12:223-261. Hagerman, F. B. 1952. The biology of the Dover sole, Microstomus pacificvs (Lockington). Calif. Fish Game, Fish Bull. 85, 48 p. Hart, J. L. 1973. Pacific fishes of Canada. Fish. Res. Board Can., Bull. 180, 740 p. Holland, G. A. 1969. Age, growth, and mortality of races of English sole {Parophrys vetulus) in Puget Sound, Washington. Pac. Mar. Fish. Comm., Bull. 7:36-50. Ketchen, K. S. 1956. Factors influencing the survival of the lemon sole (Parophrys vetulus) in Hecate Strait, British Columbia. J. Fish. Res. Board Can. 13:647-694. KOVTSOVA, M. V. 1982. Size-age structure and sex ratio of the Barents Sea population of the plaice, Pleuronectes platessa (Pleuronec- tidae). J. Ichthyol. 22:68-79. Lagler, K. F., J. E. Bardach, R. R. Miller, and D. R. M. Passino. 1977. Ichthyology. John Wiley and Sons, N.Y., 506 p. MOISEEV, P. A. 1953. Cod and flounder of the Far Eastern Seas. Fish. Res. Board Can. Transl. Ser., No. 119, 576 p. MOYLE, P. B., AND J. J. CeCH. 1982. Fishes: an introduction to ichthyology. Prentice-Hall, Englewood Cliffs, NJ, 593 p. Pearcy, W. G. 1978. Distribution and abundance of small flatfishes and other demersal fishes in a region of diverse sediments and bathym- etry off Oregon. Fish. Bull., U.S. 76:629-640. Pearson, T. H., and R. Rosenberg. 1978. Macrobenthic succession in relation to organic enrich- ment and pollution of the marine environment. Oceanogr. Mar. Biol. Annu. Rev. 16:229-311. Rae, B. B. 1965. The lemon sole. Fishing News Ltd., Lond., 106 p. 523 FISHERY BULLETIN: VOL. 86, NO, 3 Rhoads, D. C. shelf. Can. J. Fish. Aquat. Sci. 39:943-947. 1974. Organism-sediment relations on the muddy sea floor. SoKAL, R. R., AND F. J. Rohlf. Oceanogr. Mar. Biol. Annu. Rev. 12:263-300. 1981. Biometry. W. H. Freeman and Co., San Francisco, ROFF, D. A. CA, 859 p. 1982. Reproductive strategies in flatfish: a first synthesis. Van Cleve, R., and S. Z. El Saved. Can. J. Fish. Aquat. Sci. 39:1686-1698. 1969. Age, growth, and productivity of an English sole Scott, J. S. {Parophrys vetuliis) population in Puget Sound, Washing- 1982. Selection of bottom type of groundfishes of the Scotian ton. Pac. Mar. Fish. Comm., Bull. 7:51-71. 524 OCCURRENCE OF CANCER CRAB (C MAGISTER AND C. OREGONENSIS) MEGALOPAE OFF THE WEST COAST OF VANCOUVER ISLAND, BRITISH COLUMBIA Glen S. Jamieson and Antan C. Phillips^ ABSTRACT The seasonal and cross-shelf occurrences of Cancer crab (C. magister and C. oregonensis) megalopae in 1985 along a transect line perpendicular to the coast off Tofino, British Columbia, are presented. Megalopae of both species were generally absent from surface waters during daylight hours. The two species may have slight temporal differences in nocturnal surface abundance, with C. magister occur- ring later in the evening and earlier in the morning than C. oregonensis. Their relative occurrence at the surface during the night was used to calibrate cross-shelf megalopal abundance data. Cross-shelf megalopal intermolt stage proportions were calculated, relating degree of megalopal development to proximity of nearshore habitat required for successful settlement of larvae. Cancer magister megalopae were present from April to August, with peak abundance in May and June. Megalopae were abundant in a broad band 37-148 km from shore, with peak abundance (2,871 10 m"^) 56 km offshore in June. In May, some late stage megalopae were collected in coastal inlets but settlement appeared low in coastal study areas. Megalopal abundance decreased abruptly shoreward of 28 km from the coast. Cancer oregonensis megalopae were also present from April to August, with their pattern of cross- shelf abundance basically similar to that of C. magister. However, in contrast to C. magister, abundance of late stage megalopae in coastal inlets was relatively high (313 10 m"^) in June, indicating that a signifi- cant settlement of megalopae of this species could have occurred. Evidence for cross-shelf movement of Cancer megalopae is discussed. Dungeness crab, Cancer magister, range from the Aleutian Islands to northern Mexico in the eastern Pacific (Hart 1982) and are commercially exploited from northern California to Kodiak Island, AK. As part of an ongoing study of Dungeness crab recruit- ment off the west coast of Vancouver Island, the abundance and distribution of larvae off Tofino, British Columbia, are being studied to determine how variability in annual recruitment is affected by larval settlement. Dungeness crab larvae are planktonic and pass through five zoeal stages and one megalopal stage before settling to the sea bottom. Studies of larvae prior to the 1970s primarily involved descriptions of larval morphology (Mir 1961; Poole 1966), and it was not until Reed (1969) developed laboratory culture methods that larval environmental require- ments were first described. Optimal ranges of tem- perature and salinity for laboratory-cultured zoeae were 10.0°-13.9°C and 25-30%o respectively, but their survival was not significantly affected by the temperature and salinity ranges occurring in the waters off Oregon, where Reed's study was con- ducted, at the time of year when larvae are com- monly found. Lough (1976) suggested offshore lar- val movement would allow larvae to avoid lower nearshore salinities, and that normal oceanic salinity levels probably favor survival over the long term. The temporal occurrence of larvae in open coast oceanic waters varies somewhat according to lati- tude, with larvae present earliest in the season in the southern part of the species' range. Seasonal occurrence has not been well documented in the in- shore waters of Georgia Strait, Puget Sound, and Juan de Fuca Strait, but appears to differ signifi- cantly from that in open coast waters. Larval settle- ment, which typically occurs in May and June off the outer coast of Washington, can occur as late as mid-September in northern Puget Sound (D. Arm- strong2). Temporal and spatial distributions of crab larvae have been documented for the years 1975-80 in the Gulf of Farallones and the San Francisco-San Pablo- Suisun Bay complex in central California (Reilly 'Department of Fisheries and Oceans, Fisheries Research Branch, Pacific Biological Station, Nanaimo, B.C., Canada V9R 5K6. ^D. Armstrong, University of Washington, Seattle, WA 98195, pers. commun. December 1986. Manuscript accepted March 1988. FISHERY BULLETIN: VOL. 86, NO. 3, 1988. 525 FISHERY BULLETIN: VOL. 86, NO. 3 1983) and for the years 1969-70 along a transect off Newport, OR (Lough 1975, 1976). These studies suggest that while larvae are hatched in the near- shore, shallow-water environment preferred by adult Dungeness crab, they subsequently move alongshore and offshore, and then inshore, often over considerable distances. Larvae hatched off California and Oregon in late winter have been hypothesized to move northward with the Davidson Current before its reversal in March- April, and off- shore as a result of upwelling, Ekman transport, estuarine runoff, and geostrophic flow, depending on location and season (Lough 1976; Wild and Tasto 1983). Later stage zoeae are typically found pro- gressively further offshore, and it seems to be the megalopal stage that returns inshore (Lough 1976; Reilly 1983). The megalopal stage is the strongest swimming stage (Jacoby 1982), but there is no direct evidence to indicate that their inshore movement is active. There are five species of crabs of the genus Cancer in British Columbia (Hart 1982), and larvae of C. oregonensis also occur in abundance in offshore waters along with C. magister (Lough 1975). Cancer oregonensis has no commercial importance, but since it occurs with C. magister, data on both species of crabs are included in this study. Comparisons of occurrence between the two species may provide insight into possible environmental mechanisms or processes that influence transport from offshore areas to the inshore juvenile habitats, which the lar- vae of both species must reach. This study establishes the seasonal and spatial occurrences of Cancer megalopae along a transect over and beyond the continental shelf off the west coast of Vancouver Island (Fig. 1). The geograph- ical area is of particular interest because of the loca- tion of a major regional crab fishery near Tofino (Jamieson 1985; Noakes and Jamieson 1986) and the resulting importance of understanding factors influ- encing the magnitude of local larval crab settlement. The oceanography off Vancouver Island has been relatively well studied (Freeland et al. 1984; Thom- son 1981), and because of the intrusion of Juan de Fuca Strait waters into the general longshore ocean- ographic regime and the relative increase in topo- graphical complexity of the continental shelf, it is considerably more complicated than that found Figure 1.— The survey transect line, with stations, over which this study was conducted. 526 JAMIESON and PHILLIPS: OCCURRENCE OF CANCER CRAB MEGALOPAE south of Cape Flattery (Hickey 1979). Along the eastern Pacific coast, Cape Flattery marks a sig- nificant change in nearshore oceanography, and this study to investigate larval crab distribution in open coast waters is the first one north of this landmark. GENERAL OCEANOGRAPHY OFF VANCOUVER ISLAND A seasonal surface current moving northward from about lat. 32°N to 51 °N (northern Vancouver Island) typically exists in nearshore waters from October to March (Fig. 2). South of Cape Flattery, this is called the Davidson Current (Hickey 1979) whereas off British Columbia, it appears to consist of two components. Nearshore, there is the year- round, northward flowing Vancouver Island Coastal Current, about 20-30 km wide, and further offshore, there is the seasonal Shelf-Break Current, perhaps a continuation of the Davidson Current, which reverses direction in response to changes in the large-scale wind field (Freeland et al. 1984). The Davidson Current off Oregon has an average northward flow of 50 cm/s for 30% of the time (Boisvert 1969), sufficient to transport larvae sig- nificant distances northward during their develop- mental period if they remained continuously in the current. The California Current is a seasonal, south- ward moving, surface current of similar magnitude and location to the Davidson Current. It occurs off Vancouver Island (Hickey 1979) in the spring and summer, along with the seasonal, southward flow- ing, Shelf-Break Current on the outer continental shelf (Freeland et al. 1984). These currents could transport larvae located off Vancouver Island dur- ing the spring and summer southwards (Fig. 2). North of Vancouver Island, outer continental shelf currents are poorly described, but off Cape St. James (southern tip of the Queen Charlotte Islands), the surface flow is strongly southward almost year- round, turning northward only in March and April (Freeland et al. 1984). The Vancouver Island Coastal Current originates at the mouth of Juan de Fuca Strait, and so could transport larvae out of the Puget Sound-Georgia Strait complex. MATERIALS AND METHODS General Methodology Sampling was largely done along an offshore transect line perpendicular to the coast off Tofino, British Columbia (Fig. 1), with stations located at 0, 9.3, 18.5, 27.8, 37, 46, 55.6, 74.1, 92.3, 111.2, 148.2, and 185.3 km from shore. Stations were more closely spaced nearshore where larval abundance was expected to be greatest, but the transect ex- tended well beyond the shelf break to determine what the seaward distribution of megalopae might be. Six offshore samplings were made along the tran- sect from February through August 1985, but only three extended the full 185 km. Dates of sampling were 27 February-4 March, 17 and 18 April, 30 and 31 May, 14 and 15 June, 6 and 7 July, and 24 and 25 August. February-March, June, and July sam- plings were terminated 74 km, 130 km, and 111 km offshore, respectively, because of poor weather and time constraints. At each station, a neuston tow and an oblique bongo tow to 250 m, or to within 5 m of bottom, were made, and a temperature profile was recorded by means of an expendable bathythermograph. On all cruises, bongo tows were done during daylight hours and neuston tows at night; late-stage crab larvae congregate near the water surface at night (Booth et al. 1985). On 17 June 1985, an hourly, noc- turnal series of neuston tows was made at one loca- tion, 9.3 km offshore, to identify patterns of night- time abundance of megalopae in surface waters. To supplement offshore sampling along the tran- sect, neuston tows were made after dusk in the in- lets around Tofino biweekly from early May through August 1985 to monitor presence and movement of larvae. In addition, a beam trawl and epibenthic sled were used in inshore waters in July 1985 to sample for newly settled larvae in an effort to establish time and magnitude of larval settlement. The RV G. B. Reed was used for all offshore sampling except for the June sample, which was collected from a 27 m charter vessel. Inshore sam- pling was conducted from a 7 m aluminum herring skiff except for mid-June, when the charter vessel was again used. Gear The neuston sampler was a modified otter surface sampler (Mason and Phillips 1986) with a square mouth opening 45 cm on each side; under calm sea conditions, it sampled the top 35 cm of the water column. A General Oceanics flowmeter^ in the mouth of the net was used to establish volume of water filtered. Netting was black, 500 pt Nitex, and ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 527 FISHERY BULLETIN: VOL. 86, NO. 3 WINTER 54° 51' 48° CONFUSED AND ^;i^;S^ VARIABLE CURRENTS 54° 51° 48° Figure 2.— Prevailing surface circulation off the British Columbia-Washington coast in winter and summer. Broken arrows indicate uncertain currents. Numbers give speeds (cm s"^) (modified from Thomson 1981). 528 JAMIESON and PHILLIPS: OCCURRENCE OF CANCER CRAB MEGALOPAE was the same size as that on the bongo gear. The neuston sampler was towed at 4 kn approximately 10 m from the side of the vessel to reduce hull avoidance by the larvae. Tow duration was usually 15 minutes but was shortened when crab larvae occasionally became very abundant at dawn and dusk. Bongo gear was a modified SCOR design (Mason et al. 1984), having a mouth opening of 25 cm (each side) with the outboard (left) net of 500 ^a Nitex and the inboard (right) net of 230 ^ Nitex. A General Oceanics flowmeter was mounted in the mouth of each net. Sampling procedure followed that de- scribed by Smith and Richardson (1977*). At stations <100 m water depth, the bongo was fished in an un- dulating fashion from the bottom to the surface in order to filter a standard volume of water (about 300 m^). The epibenthic sled (Phillips and Mason 1986) used in inshore waters had a 60 cm mouth opening and 1 mm Nitex netting; it was towed for 10 minutes at 2-3 kn. Nine tows were made over bottoms of unknown characteristics. The modified beam trawl (Gunderson and Ellis 1986), also used in inshore waters, had a 3 m mouth opening, a 7 mm mesh net, and a 3 mm mesh cod end; it was towed for 10 minutes at about the same speed as the sled. For both gear types, distance towed was calculated by radar triangulation to reference points on land. Data Analysis All plankton samples were preserved initially in a 4% formaldehyde solution of saltwater. In the laboratory, settled volume was determined, and general composition of the plankton noted. Samples were then sieved, and Cancer larvae picked by hand and returned to a 2% formaldehyde solution before their identification. Cancer magister megalops are readily identifiable by their larger size (Trask 1970; Lough 1975), but the currently used key (Lough 1975) was not always effective in separating C. ore- gonensis and C. jyroductus. Presence of lateral spines is a subjective criterion, and morphological dimen- sions and counts of setae were two variable to distinguish species. These smaller megalops were finally concluded to be C. oregonensis after rearing hundreds of larvae to the juvenile stage and finding no C. productus. With beam trawl samples, the catch was sorted *Smith, P. R., and S. L. Richardson. 1977. Manual of methods for fisheries resource survey and appraisal. Part 4. Standard tech- niques for pelagic fish egg and larval surveys. Southwest Fish. Cent., Natl. Mar. Fish. Serv., NOAA, Adm. Rep. No. 77-11. on deck to remove juvenile crabs, and the megalopae were preserved as above. Species composition of the catch was noted. For the bongo tows, the volume of water filtered was used to calculate a haul factor (Mason et al. 1984), which accounted for tow depth and allowed expression of the data as the integrated number of organisms beneath 10 m^ of sea surface. For neuston tows, relative abundance was ex- pressed as area swept rather than as volume filtered and again expressed as number of animals under 10 m^ of sea surface. Calculations of larval abundance are necessarily conservative, and direct comparisons between dif- ferent gear types are not presently possible, given our current understanding of gear efficiency and lar- val catchability, which varies with both depth and time of day. Numbers of larvae reported here are thus directly comparable only within each gear type used. For bottom gear, numbers of crabs were cal- culated with no consideration of gear efficiency. Molt Staging State of development within the megalopal stage was determined for all Cancer megalopae collected at a station, or 25 randomly selected individuals of each species if the number collected exceeded 25. The sequence of epidermal changes occurring dur- ing this intermolt period has been described for Dungeness crab by Hatfield (1983), and her proce- dures and staging criteria were applied for both species. Whole megalopae were stained with 0.25% Fast Green dye in either water or polyvinyl lacto- phenol for 2-18 hours, and then the second maxil- lipeds were removed and mounted. Megalopae were identified to 1 of the 13 intermolt stages recognized by Hatfield (1983); these were then combined to form 3 groups: early (stages 1-4), middle (stages 5-8), and late (stages 9-13) megalopae. Temporal durations of these three groups in laboratory studies were 5.8, 15.4, and 6.3 days, respectively (Hatfield 1983). Corresponding durations for C. oregonensis are unknown. Stages were grouped into the three categories described for simplification of analysis; many of Hatfield's stages were of <48-h duration, and this was considered too fine a resolution for our purposes. RESULTS Water Temperature Water temperature at the surface ranged from 529 FISHERY BULLETIN: VOL. 86, NO. 3 8.4° to 14.4°C, and from 7.6° to 8.8°C at 50 m, dur- ing the period 17 April to 7 July (Table 1). Warmest surface temperatures were in May and July, and at 50 m, were warmest in June. Species Occurrence Three species of crab larvae predominated in the plankton collected. The porcellanid crab, Petrolisthes cinctipes, predominated in waters within 8 km of shore, while C. oregonensis and C. magister domi- nated in more offshore waters. Other crab species were present but at much lower abundance than these three species. Megalopae of C. oregonensis were of comparable abundance to C. magister at most stations sampled. Surface Abundance Presence of crab megalopae at the surface was dependent on time of day, with slight differences in timing of maximal abundance occurring between the two Cancer species (Fig. 3). At 9.3 km offshore in June, C. oregonensis was several orders of mag- nitude more abundant and also seemed to peak in abundance a little earlier in the evening and later in the morning than did C. magister. Relative abun- dance of both species declined in the middle of the night. Data in Figure 3 was smoothed by eye (Fig. 4) to allow megalopal abundance data from the transect to be weighted by time of capture at night. Multiplier values were determined, and these were used to adjust actual abundance data by sampling time for the May and June sampling periods (Tables 2, 3). Only these data were weighted, and since day length changes with time of year, ad- justment of data collected more than two weeks on either side of 14 and 15 June was considered inappropriate. In hindsight, some data (indicated by asterisks in Tables 2, 3) was found to have been collected at a time when megalopae were just beginning to reach, or had just left, the surface. The calibration slope (Fig. 4) is very steep at both these times and it was impossible to estimate accurately a meaningful scal- ing multiplier for these data. Our estimated mega- lopal abundance at these stations should be viewed with particular caution; a zero value may not in- dicate that megalopae were absent, but only that they were not at the surface. Bongo tows during the day poorly sampled the abundance of C. magister megalopae (Table 4) shown to be present by nighttime neuston tows (Tables 2, 3). This was possibly due to sampling the total column below only a few square meters of sea surface, because of integration of the results, or to avoiding the net by the megalopae. It was unlikely because of failure to sample deeply enough, since on the continental shelf, tow depth approached bot- tom depth. Megalopae (maximum per tow was 4) were primarily collected by bongo gear at stations where they were abundant in neuston samples at night, but relative abundance estimates obtained from bongo nets are considered only useful in a general qualitative sense because of the low absolute numbers caught. Temporal Occurrence (Time of Year) No Cancer megalopae were found in samples col- lected shortly after dusk on 27 February-4 March. Cancer megalopae were first observed on the next Table 1 .- -Watei ' temperatures (°C) by depth (m) at the stations sampled off Tofino, B.C. on 17 April, 31 May, 14 June, and 6 and 7 July 1985. Depth (m) Distance offshore (km) April May June July 0 20 50 0 20 50 0 20 50 0 20 50 0 9 19 28 37 56 74 93 111 130 148 185 8.8 8.8 8.7 8.8 8.6 8.7 8.7 8.5 8.5 8.5 8.4 8.8 8.7 8.1 8.8 8.6 8.7 8.6 8.5 8.4 8.5 8.6 7.9 8.2 8.3 8.2 8.2 8.2 8.0 8.0 8.4 12.2 13.4 13.4 13.1 13.8 13.3 11.8 11.9 11.5 11.0 11.4 9.2 9.5 8.6 9.6 9.3 9.8 10.0 10.3 10.5 10.0 10.0 7.6 7.9 7.9 8.2 8.2 8.2 8.1 8.7 8.4 12.2 12.2 12.3 12.8 12.7 12.9 12.9 12.8 12.7 12.7 11.4 9.0 10.4 12.3 9.4 10.5 12.0 12.2 12.2 12.4 8.0 8.0 8.3 8.1 8.7 8.5 8.5 8.6 13.5 14.0 13.3 13.3 14.4 14.4 11.0 9.0 10.0 10.7 13.3 13.3 7.6 7.7 7.7 8.6 8.6 530 JAMIESON and PHILLIPS: OCCURRENCE OF CANCER CRAB MEGALOPAE gsijung- jq6i|iM J I lAio - — )L|6l|IM| |D3||nDU in O to g ID <1> 00 00 mB|||Mj |DO!jnDu - jgBjiiMj |iA|0 - jasuns- p— I — I — r O O o [III — I — I r O O -r-[ — I — I — r B o O o •q (O S o to () a n -4-> rf JS O * bo < C Q c o o o ^ (M bb o o ^ ^ C\J "3. V rt V) M o -) 5? o n c OJ u, f\j bserved du .C. ^^ O pQ H 2 - 5 o to 9- c Q ° iG CL 0! O 8SI Jung — mbi |iMt I IMO -mbl|IM| |D3UnDU (q6i|iM( I Dounou |L|&||IM4 |IAI3 i8suns- I I I o T — I — r o o o o to z i § o (- o ^ ( ) z < . 33 o O <=> o s^ «^ o »5 o CM O) o b o o CVI ^ o CO a> o 3 c o Q c« (M -13 CJ C 1 o o o 'S o o CJ CO OS D o (aoD^jns oas j.ojoI-N) 33NVaNnaV 531 Figure 4.— Smoothed abundance of Cancer magister and C. oregonensis megalopae ob- served during sampling through one night at a station 9 km off the coast off Tofino, B.C. 1000 o o Z! c/) O 0) 'E O LlI O z < Q 3 QD < ^1 Q UJ I t- O o CO FISHERY BULLETIN: VOL. 86, NO. 3 10 JUNE 1985 o in O 10 100 10- 'E O UJ u < Q CD < 6> 51 ^•| Q LU I f- o o CO 1800 2000 2200 2400 0200 0400 0600 TIME (POST) Table 2.— Scaling values used to standardize the actual number of megalopae caught (unsealed) to the estimated number which would have been caught at the times of peak nocturnal abundance, 22:40 and 23:40 hours, for Cancer oregonensis and C. magister. respec- tively, on 30 and 31 May 1985. * = Value uncertain because of time of sampling. Values rounded to one decimal place. c. magister C. oregonensis Distance (km) (A/ 10 m-2) (N 10 m-2) Time Multiplier Unsealed Scaled Multiplier Unsealed Scaled 18:12 0 oo 2.3 ?* oo 0 0* 19:02 9 36.7 0.2 7.2* CO 4 ?' 19:43 19 26.8 0.0 1.1* 00 0 0* 20:26 28 20.0 0 0 oo 0 0* 21:09 37 11.0 0.4 4.7 oo 0 0* 21:52 46 3.9 14.4 56.4 5.2 2.2 11.5 22:35 56 1.6 29.9 47.0 1.0 0.1 0.1 23:37 74 1.0 17.0 17.0 — 0 0 00:47 93 3.3 36.0 119.9 24.8 1.0 25.7 01:55 111 5.0 2.6 13.3 40.0 1.8 72.1 03:56 148 110.0 1.1 118.4* 2.5 0.1 0.2 05:57 185 OO 0 0* oo 0 0* 532 JAMIESON and PHILLIPS: OCCURRENCE OF CANCER CRAB MEGALOPAE sampling date, 16 and 17 April, but maximum abun- dance (unweighted as to diel time of sampling) occurred in late May and the second week of June (Fig. 5). Cancer magister megalopae were collected as late as 24 August, the last sampling date, but relative abundance along the transect was low after the beginning of July. Cancer oregonensis megalopae were caught in April but occurred in abundance only in the late May and mid- June surveys, with peak abundance in June. 1000 100 120 140 160 180 200 220 240 260 CALENDAR DAYS Figure 5.— Neustonic abundance of megalopae observed along the transect line during April to August 1985 surveys. Table 3.— Scaling values used to standardize the actual number of megalopae caught (unsealed) to the estimated number which would have been caught at the times of peak nocturnal abundance, 22:40 and 23:40 hours, for Cancer oregonensis and C. magister, respec- tively, on 14 and 15 June 1985. * = Value uncertain because of time of sampling. Values rounded to one decimal place. C. magister C. oregonensis Distance {N 10m-2) (N 10 m"^) Time (km) Multiplier Unsealed Sealed Multiplier Unsealed Sealed 01:42 0 0 0 37.1 8.8 327 00:58 9 3.6 0.1 0.4 27.4 170.5 4,671 00:09 19 1.2 0.1 0.1 16.3 0.1 2.3 23:19 28 1.0 0.2 0.2 2.7 7.6 20.5 05:36 37 OO 0.1 ?* OO 0 0* 03:52 56 55.0 52.2 2,871* 2.6 15.7 40.9 02:02 74 5.2 3.7 19.2 40.0 4.8 192 00:25 93 2.4 8.9 21.6 20.8 1.5 30.4 22:45 111 1.2 8.4 10.3 1.0 12.5 12.5 21:10 130 — 0 0 — 0 0 533 FISHERY BULLETIN: VOL. 86, NO. 3 Table 4.— Tow characteristics and number of Cancer magister megalopae (10 m^^ of sea surface) caught by 0.25 m bongo tows off Tofino, B.C., on 31 May and 14 June 1985. May June Distance Water Max. Water Max. offshore Time volume depth No. Megalopae Time volume depth No. Megalopae (km) (h) (m3) (m) 33 megalops (A/ 10 m-'') (h) K) (m) megalops (A/ 10 m"^) 0 19:27 210 1 1.6 08:00 240 36 3 4.5 9 18:52 182 55 1 3.0 08:55 220 55 0 0 19 18:16 114 72 0 0 09:51 222 75 0 0 28 17:37 166 120 0 0 10:50 209 115 0 0 37 16:55 199 151 0 0 11:47 284 146 1 5.1 56 15:30 352 250 4 28.4 13:32 381 192 0 0 74 14:07 327 250 1 7.6 15:21 338 200 1 5.9 93 12:42 326 250 3 23.0 17:04 296 200 2 13.5 111 11:17 360 250 1 6.9 18:56 269 200 0 0 130 21:12 327 200 2 12.2 148 08:55 400 250 0 0 — 185 06:18 319 250 0 0 — Cross-Shelf Spatial Distribution With both species, no clear pattern in cross-shelf distribution was evident from the April data, per- haps because of relatively low overall megalopal abundance. In late May (Table 2), scaled abundance (weighted as to time of night) of C. magister was highest at 93 km offshore, but megalopae were generally abundant (>10 10 m"^ sea surface) from 46 to 148 km offshore. In mid-June, the basic pat- tern observed in late May was still evident, although megalopal abundance near shore had declined (Table 3). Highest estimated C. magister abundance was at 56 km offshore. With C. oregonensis, scaled abundance in late May peaked (>20 10 m' 2) 93-111 km offshore (Table 2). An exceptionally large number of megalopae (4,671 10 m"^) was found 9 km offshore in June (Table 3), while from 56 to 111 km offshore, abundance re- mained high (>20 10 m- 2). Sampling of inlet waters around Tofino with neuston gear showed that a few Cancer megalopae were present but that no substantial (average was <1 10 m~2) numbers occurred. Maximum C. ma- gister megalopal abundance was 2.9 megalopae 10 m"^ on 5 June, with megalopae observed only be- tween 29 May and 19 June. Maximum C. oregonen- sis abundance was 2.3 megalopae 10 m"^ on June 19, with megalopae observed only between 4 and 19 June. Intermolt Stage All C. magister megalopae collected in April were of early developmental stage (Fig. 6), whereas all those collected in August were mid-stage mega- lopae. In May, June, and July, the general pattern of offshore distribution by molt stage was for late stage larvae to be nearest inshore and early stage larvae to be furthest offshore. Cancer oregonensis megalopae had a similar developmental distribution pattern to those of C magister (Fig. 7); late molt megalopae were most abundant closest to shore. However, in late May, most megalopae were early stage, in contrast to C. magister, whereas by mid-June, all three intermolt groups were present. When the scaled megalopal abundance at distance offshore (Tables 2, 3) is multiplied by the percent- age at each molt stage at a specific location (Figures 6, 7, for each species, respectively), the actual abun- dance by molt group with distance offshore is deter- mined (Figures 8, 9, respectively). In late May, late stage C. magister megalopae were relatively abun- dant 9 km offshore, indicating that some Dungeness crab settlement may have occurred. However, lar- val sampling in the inlets and bays showed few megalopae present. In contrast to previous years, no recently settled juvenile crabs (0 age-class crabs) were found by local fishermen in intertidal areas or on floating objects, confirming that the magnitude of megalopal settlement in 1985 was relatively small. In mid- June, there were few C. magister megalopae of any intermolt stage present within 40 km of the coast and little evidence of late stage megalopae off- shore. Early and mid-stage megalopae were most abundant from 56 to HI km offshore. For C. oregonensis, the opposite was observed (Fig. 9). Few late stage megalopae were present in late May, and unlike C. magister, megalopae were not concentrated in nearshore waters. However, megalopae in all three intermolt stages were abun- dant inshore in mid-June, with late stage megalopae dominating at the coast. The timing of occurrence 534 JAMIESON and PHILLIPS: OCCURRENCE OF CANCER CRAB MEGALOPAE Figure 6.— Percentage of the three Cancer magister megalopal intermolt groups observed at each station during the April to August 1985 surveys. LlI o UJ CL LU O CC LlI CL (a) APRIL 16/17, 1985 2 33 2 (- z UJ o tr LU CL (b) may 30/31 100 80- 60- 40 20 0 3- ^25 25 231 I )!^ 1 '1 f X '■'i 25 25 25 o— o.. (c) JUNE 16 ® 5 4 25 24 17 25 25 (e) august 23/26 100 -, 80 60 40 20 0 — t \ 1 1 0 40 80 120 140 DISTANCE (Km) JULY 6/7 63 3 2 0 40 80 120 DISTANCE (Km) C. magjster INTERMOULT STAGES o o ■ a 1-4 5-8 9-13 n= 10 unless otherwise specified above graph of late stage megalopae at the coast thus differed between the two species. Inshore Benthic Sampling Survey of commercial crab habitat with beam trawls and an epibenthic sled yielded few newly settled megalopae or juvenile C magister (Table 5). A maximum of 83 crab/10,000 m^ was collected, well below densities which might be expected if settlement had been substantial. No C. oregonensis were caught, perhaps because of the habitat being sampled, and no larval settlement of either species was observed when intertidal mud flat areas were searched by foot. In a concurrent study, no signifi- cant abundance of 0 age-class C. magister was observed during monthly sampling throughout the remainder of the year. DISCUSSION Megalopal Source Temporal duration of C. magister larval stages has been estimated both by laboratory rearing (Poole 1966; Reed 1969; Brugman 1972; Gaumer 1973; Ebert et al. 1983) and the first appearances of lar- val stages in field studies (Poole 1966; Lough 1976; Reilly 1983). The length of the total larval period has been estimated as 105-125 days by Reilly (1983), 535 FISHERY BULLETIN: VOL. 86, NO. 3 (a) may 30,31 LlI (J cr UJ Q- 100-1 80- 60 40 20 0 3 C, oregonensis INTERMOULT STAGES • . 1-4 o <, 5-8 n = 25 unless otherwise specified above graph ^■,..,r,.£.,,---,-°— °- 1 — f r»— , 0 20 40 60 80 100 120 140 160 (b) JUNE 13-17 II C) JULY 6,7 15 15 0 20 40 60 80 100 120 100-1 80- 60 40 20 0 0 20 40 60 80 100 DISTANCE OFFSHORE (km) Figure 7.— Percentage of the three Cancer oregonensis megalopal intermolt groups observed at each station during the May to July 1985 surveys. Table 5. — Beam trawl and epibenthic sled catches from nearshore waters around Tofino, 16-18 June 1985. Area Depth (m) Gear type Catch of crabs No. 10" m-2 Area swept (m ) Juveniles Megalops Juveniles Megalops Templar Channel 7 trawl 2,650 sled 375 Chesterman Beach 5 trawl — sled 300 Cox Bay 8 trawl 2,500 7 sled 400 13 trawl 2,600 15 sled 420 30 trawl 2,260 23 sled 360 Lennard Island 25 trawl sled 350 Long Beach 7 trawl 2,700 sled 400 13 trawl 2,650 sled 400 23 trawl sled 220 0 0 0 0 0 0 0 0 3 0 0 0 1 1 26 0 0 4 75 27 48 0 0 0 18 50 23 50 0 0 0 0 0 0 0 0 83 0 0 0 4 25 45 536 JAMIESON and PHILLIPS: OCCURRENCE OF CANCER CRAB MEGALOPAE 5? o CO CD \a 00 <3> i-H 8 s ^ •-s o 1 bo c 'C C o JS o a) -a > O bo c 6 o e 3 o ■kJ .5 ■— +J ^g 0) c C3 3 o a> ^■^ bo "S I": §> I o V o c -a c 3 -3 o o (\i X 1-. Si CU 6 U) X! oo > < o o (VI O in O o — r~ O in o CO o O O O - O O CD O ID O o CO Q T^O §■ ° o > s -^ c (U 3 ft 2 O ^ bd > P II Cn o O 2 km from the swamp contained mangrove-derived carbon. To this date, there is little evidence to suggest that natant species such as fishes and decapod Crustacea can use detritus as a primary food source even if microbially enriched (Boesch and Turner 1984). Localities where mangroves grow often support a variety of other primary producers, and variation in the relative significance of detrital and algal car- bon sources may be associated with tidal amplitude and flushing. Laguna Joyuda, with microtidal influ- ences, appears to be fueled by algal carbon where- as shrimp-producing mangrove areas in Ecuador have meso- and macrotidal regimes and no apparent algal growth (Zimmerman, pers. observ.). That man- groves provide critical substratum and protective cover for a large number of organisms is undisputed; however, the assumed significance of mangrove- derived detritus should be examined further. ACKNOWLEDGMENTS This research was supported by the Office of Sea Grant, NOAA (Grant no. R/A-01-2), a grant from the National Science Foundation (No. R-II-8610677), and funding from the University of Puerto Rico to the Center for Energy and Environment Research. D. Corales participated in the field work, along v^ath G. Owen and 1. Sanders. B. A. Buchanan identified the shrimps collected and assisted in the prepara- tion of figures. L. L. Cruz conducted all of the gut analyses and T. Robles assisted in manuscript prep- aration. W. J. Richards and B. A. Buchanan pro- vided helpful criticisms of the manuscript. LITERATURE CITED Boesch, D. F., and R. E. Turner. 1984. Dependence of fishery species on salt marshes: the role of food and refuge. Estuaries 7:460-468. Bray, J. R., and J. T. Curtis. 1957. An ordination of the upland forest communities of southern Wisconsin. Ecol. Monogr. 27:325-349. Carr, W. E. S., and C. a. Adams. 1972. Food habits of juvenile marine fishes: evidence of the cleaning habit in the leatherjack, Oligoplites saurus, and the spottail pinfish, DiplodxLS holbrooki. Fish. Bull., U.S. 70: 1111-1120. Chong, V. C, AND A. Sasekumar. 1981. Food and feeding habits of the white prawn Penaeiis merguiensis. Mar. Ecol. Prog. Ser. 5, p. 185-191. Craig, H. 1957. Isotopic standards for carbon and oxygen and correc- tion factors for mass-spectrometric analysis of carbon diox- ide. Geochim. Cosmochim. Acta 12:133-149. Dall, W. 1968. Food and feeding of some Australian penaeid shrimp. FAO Fish. Rep. 2:251-258. Darnell, R. M. 1958. Food habits of fishes and larger invertebrates of Lake Ponchartrain, Louisiana. Publ. Inst. Mar. Sci., Univ. Texas. 5:353-416. Edwards, R. R. C. 1978. The fishery and fisheries biology of penaeid shrimp on the Pacific coast of Mexico. Oceanogr. Mar. Biol. Annu. Rev., p. 145-180. Field, J. G., and G. McFarlane. 1968. Numerical methods in marine ecology. I. Quantitative "simUarity" analysis of rocky shore samples in False Bay, South Africa. Zool. Air. 3:119-138. Fry, B., R. K. Anderson, L. Entzeroth, J. L. Bird, and P. L. Parker. 1984. '^C enrichment and oceanic food web structure in the northwestern Gulf of Mexico. Contrib. Mar. Sci. 27:49-63. Fry, B., R. Lutes, M. Northam, and P. L. Parker. 1982. A '^C/'^C comparison of food webs in Caribbean seagrass meadows and coral reefs. Aquat. Bot. 14:389-398. Fry, B., and P. L. Parker. 1979. Animal diet in Texas seagrass meadows: d^^C evidence for the importance of benthic plants. Estuarine Coastal Mar. Sci. 8:499-509. Fry, B., and E. B. Sherr. 1984. d^^C measurements as indicators of carbon flow in marine and freshwater ecosystems. Contrib. Mar. Sci. 27: 13-47. Garcia, S. 1974. Biologie de Penaeus duorarum notialis en Cote D'lvoire, IV. Relations entre la repartition et les conditions du milieu - etude des variations du sex-ratio. Doc. Sci. Cent. Rech. Oceanogr. Abidjan. 5(3-4): 1-39. Garcl\, S., M. Lemoine, and E. Lebrun. 1985. Seasonal and long-term variability of recruitment in French Guiana shrimp fishery on Penaevs suhtilis. FAO Fish. Rep. No. 327, SuppL, p. 242-250. Gleason, D. F. 1986. Utilization of salt marsh plants by postlarval brown shrimp: carbon assimilation rates and food preferences. Mar. Ecol. Prog. Ser. 31:151-158. Gleason, D. F., and R. J. Zimmerman. 1984. Herbivory potential of postlarval brown shrimp associ- ated with salt marshes. J. Exp. Mar. Biol. Ecol. 84:235- 246. Hughes, E. H., and E. B. Sherr. 1983. Subtidal food webs in a Georgia estuary: d^^C anal- ysis. J. Exp. Mar. Biol. Ecol. 67:227-242. Laughlin, R. a. 1982. Feeding habits of the blue crab, Callinectes sapidus Rathbun, in the Apalachicola Estuary, Florida. Bull. Mar. Sci. 32:807-822. 550 STONER and ZIMMERMAN: FOOD PATHWAY ASSOCIATED WITH PENAEID SHRIMPS Leber, K. M.. III. 1983. Feeding ecology of decapod crustaceans and the in- fluence of vegetation on foraging success in a subtropical seagrass meadow. Ph.D. Thesis, Florida State Univ., Tallahassee, 166 p. 1985. The influence of predatory decapods, refuge, and microhabitat selection on seagrass communities. Ecology 66:1951-1964. Levine, E. a. 1981. Nutrient cycling by the red mangrove, Rhizophora mangle L., in Joyuda Lagoon on the west coast of Puerto Rico. M.S. Thesis, Univ. Puerto Rico, Mayaguez, 103 p. Lindner, M. J. 1971. Shrimp resources of the Caribbean Sea and adjacent regions. FAO Fish. Rep. 2:149-156. Lindner, M. J., and H. L. Cook. 1970. Synopsis of biological data on the white shrimp Penaeus setiferus (Linnaeus) 1767. FAO Fish. Rep. 4:1439-1469. Livingston, R. J. 1984. Trophic response of fishes to habitat variability in coastal seagrass systems. Ecology 65:1258-1275. Marte, D. L. 1980. The food and feeding habit of Penaeus monodon Fabricus collected from Makato River, Aklan, Philippines (Decapoda, Natantia). Crustaceana 38:225-236. MORIARTY, D. J. W. 1976. Quantitative studies on bacteria and algae in the food of the mullet Mugil cephal-us L. and the prawn Metapenaeus bennettae (Racek & Dall). J. Exp. Mar. Biol. Ecol. 22:131- 143. 1977. Quantification of carbon, nitrogen and bacterial biomass in the food of some penaeid prawns. Aust. J. Mar. Freshwater Res. 28:113-118. MORIARTY, D. J. W., AND M. C. BARCLAY. 1981. Carbon and nitrogen content in food and assimilation efficiencies of penaeid prawns in the Gulf of Carpentaria. Aust. J. Mar. Freshwater Res. 32:245-251. Neiva, G. deS. 1969. Observations on the shrimp fisheries of the central and southern coast of Brazil. FAO Fish. Rep. 3:847-858. NiKOLic, M., and M. E. Ruiz de Quevedo. 1969. Notas biologico-pesqueras sobre camaron bianco Penaeus schmitti Burkenroad 1936. FAO Fish. Rep. 3: 1107-1117. Odum, W. E., and E. J. Heald. 1972. Trophic analyses of an estuarine mangrove community. Bull. Mar. Sci. 22:671-738. 1975. The detritus-based food web of an estuarine mangrove community. In L. E. Cronin (editor), Estuarine research Vol. L p. 265-286. Acad. Press, N.Y. RoDELLi, M. R., J. N. Gearing, P. J. Gearing, N. Marshall, AND A. SaSEKUMAR. 1984. Stable isotope ratio as a tracer of mangrove carbon in Malaysian ecosystems. Oecologia 61:326-333. Roman, M. R. 1984. Utilization of detritus by the copepod, Acartia txmsa. Limnol. Oceanogr. 29:949-959. Sheridan, P. F. 1979. Trophic resource utilization by three species of sciaenid fishes in a northwest Florida estuary. Northeast Gulf Sci. 3:1-15. Staples, D. J. 1980. Ecology of juvenile and adolescent banana prawns, Penaeus merguiensis, in a mangrove estuary and adjacent offshore area of the Gulf of Carpentaria. L Immigration and settlement of postlarvae. Aust. J. Mar. Freshwater Res. 31:635-652. Stofer, Z. 1980. Preparation of carbon dioxide for stable carbon isotope analysis of petroleum fractions. Anal. Chem. 52:1389-1391. Stoner, a. W. 1980. The feeding ecology of Lagodon rhovnhoides (Pisces: Sparidae): Variation and functional responses. Fish. Bull., U.S. 78:337-352. 1986. Community structure of the demersal fish species of Laguna Joyuda, Puerto Rico. Estuaries 9:142-152. 1988. A nursery ground for four tropical Penaeus species: Laguna Joyuda, Puerto Rico. Mar. Ecol. Prog. Ser. 42:133-141. Williams, A. B. 1955. A survey of the North Carolina shrimp nursery grounds. J. EHsha Mitchell Sci. Soc. 71:200-207. 551 THE EFFECT OF THE ECTOPARASITIC PYRAMIDELLID SNAIL, BOONEA IMPRESSA, ON THE GROWTH AND HEALTH OF OYSTERS, CRASSOSTREA VIRGINICA, UNDER FIELD CONDITIONS Elizabeth A. Wilson,^ Eric N. Powell/ and Sammy M. Ray^ ABSTRACT Boonea (= Odostomia) impressa are contagiously distributed on oyster reefs so that some oysters are parasitized more than others. The parasite's mobihty and the abihty of oysters to recover from snail parasitism may be important in assessing the impact of parasitism on oyster populations. During a 4-week exposure period in the field, B. impressa reduced American oyster, Crassostrea virginica, growth rate and increased the intensity of infection by the protozoan, Perkinsus ( = Dermocystidium) marinus, but produced few changes in the oyster's biochemical composition because, although net productivity was reduced, the oysters retained a net positive energy balance (assimilation > respiration). During a 4-week recovery period, growth rate returned to normal (control) levels, but infection by P. ynarinv^ continued to intensify in previously parasitized oysters kept B. impressa-tree. Most changes in biochemical com- position during recovery, including increased lipid and glycogen contents, could be attributed to the con- tinuing increase in infection intensity of P. marinus. Consequently, the temporal stability and size of snail patches, particularly as they regulate infection by P. marinus, may be the most important factors influencing the impact of B. impressa on oyster reefs. Parasitism can be an important factor affecting the population dynamics (Wickham 1986; Brown and Brown 1986; Rabat 1986) and health (Brockelman 1978; Mohamed and Ishak 1981; Ford 1986) of host species. Three parasites are known to be especially important in oysters. Perkinsus (= Dermocysti- dium) marinus, Haplosporidium nelsoni (MSX), and Boonea ( = Odostomia) impressa detrimentally af- fect oyster growth, health, and biochemical composi- tion (Mengebier and Wood 1969; Feng et al. 1970; Soniat and Koenig 1982; White et al. 1984; Ford 1986; Ward and Langdon 1986; White et al. 1988, in press). The pyramidellid gastropod, Boonea impressa, is one of a widely distributed group of parasitic, marine opisthobranchs (Fretter and Graham 1949; Fretter 1951; Allen 1958). Boonea impressa removes nutri- ents directly from its host by piercing the flesh with a hollow stylet and sucking the body fluids using a buccal pump (Fretter and Graham 1949; Fretter 1951; Allen 1958). The most common host of 5. im- pressa is Crassostrea virginica (Hopkins 1956; Allen 1958; Wells 1959) but, like other odostomians, it is not entirely host specific (Wells 1959; Robertson 1978; Robertson and Mau-Lastovicka 1979). Found 'Department of Oceanography, Texas A&M University, College Station, TX 77843. ^Department of Marine Biology, Texas A&M University at Galveston, Galveston, TX 77550. Manuscript accepted March 1988. FISHERY BULLETIN: VOL. 86, NO. 3, 1988. abundantly on oyster reefs from Massachusetts to the Gulf of Mexico, B. impressa has been reported in numbers as high as 100 per oyster (Hopkins 1956). Under laboratory conditions, B. impressa reduced oyster growth rates as the result of both direct removal of assimilated carbon from the oyster and direct interference with the oyster's ability to feed (White et al. 1984, 1988; Ward and Langdon 1986). Not surprisingly, parasitism by B. impressa produc- ed changes in the biochemical composition of oyster tissue. Parasitism by 15 snails, a relatively high field density found in dense snail patches decreased car- bohydrate and free amino acid content and increased lipid content of mantle tissue. Reproduction was af- fected as well (White et al. in press). Perkinsus marinus is an important cause of mor- tality of oysters in the Gulf of Mexico (Mackin 1962; Ray 1966a; Hofstetter 1977). Sublethal effects in- clude reduced growth (Menzel and Hopkins 1955; Ray et al. 1953) and changes in biochemical com- position (Soniat and Koenig 1982; White et al. in press). Perkinsus marinus can be transmitted from one oyster to another through the water (Ray 1954; Mackin 1962; Andrews 1965) or by B. impressa feeding (White et al. 1987). The intensity of infec- tion by P. marinus is also increased in B. impressa- parasitized oysters (White et al. 1987). Usually stress affects organisms by altering nor- mal metabolic activity. Many organisms may be able 553 FISHERY BULLETIN: VOL. 86, NO. 3 to return to normal metabolic condition after the stress is removed (Portner et al. 1979; Pickering et al. 1982; Kendall et al. 1984). This ability, termed recovery, is an important adaptation to changes in the natural environment. Recovery is not always complete nor does it occur over relatively short time scales in all cases. The time required for recovery varies depending on the parameter being studied, the stress that was applied, and the species involved (Kendall et al. 1984; Haux et al. 1985; Neff et al. 1985). Apparently deleterious effects occurring after exposure, during the so-called recovery period, are well described (e.g. Kendall et al. 1984; Powell et al. 1984). White et al. (1984) showed that C. virginica could attain normal growth rates within one week after B. impressa were removed. However, growth rates frequently return to normal more rapidly than other metabolic parameters (Ken- dall et al. 1984). Boonea impressa are extremely mobile and are more contagiously distributed than their hosts, so that some oysters are highly para- sitized while others remain parasite free (Powell et al. 1987). Individual B. impressa change hosts often but typically move between existing aggregates (Wilson et al. in press). Therefore, refugia from parasitism may exist and recovery may be impor- tant in assessing the overall impact of B. impressa on oysters. Stress, produced by laboratory conditions, fre- quently accompanies laboratory experimentation (e.g., Koenig et al. 1981; Powell et al. 1984; Kukal and Kevan 1987). The effect of B. impressa on C. virginica has been assessed primarily through laboratory experimentation. Consequently, we ex- amined the effect of B. impressa on C. virginica under field conditions and assessed the ability of oysters to recover normal growth rates and bio- chemical composition once snail parasitism ceased. MATERIALS AND METHODS Field Study Oysters and snails used in this study were col- lected at Goose Island State Recreation Area near Rockport, TX. The oysters were weighed on a Mettler balance using the underwater m.ethod of Andrews (1961). Fifteen oysters (precontrols) were sacrificed to define the biochemical composition of the oysters and Perkinsus marinus levels that ex- isted naturally at the collection site. The rest of the oysters were placed in semi-enclosed plexiglass domes (see figure 2 in Kendall et al. 1984 for descrip- tion) in a tidal creek near the Aransas Pass Lighthouse on Lydia Ann Channel, near Port Aran- sas, TX. The domes allowed water to circulate over the oysters, while excluding such large predators as oyster drills and crabs. Boonea im- pressa, however, could readily move into or out of the domes. Two of the four domes (20 oysters per dome) con- tained oysters exposed to B. impressa at a concen- tration of 10 snails per oyster, a level of parasitism commonly observed on reefs in the collection area (White et al. 1984). The domes were positioned so that the probability of snails moving from the ex- posure domes with parasitized oysters to the con- trol domes was minimized (Fig. 1). Snails on each experimental oyster were counted twice weekly for 4 weeks. No B. impressa were ever found on con- trol oysters. Boonea impressa did emigrate from the exposure domes, however, so snails were added as needed to maintain the 10:1, snaihoyster ratio. Con- trol oysters were handled the same as experimentals each week in an effort to minimize differential ef- fects caused by handling stress (see Pickering et al. 1982; Andrews and Hewatt [1957] were unable to find any effect of handling on P. marinus infection in oysters). At the end of 4 weeks, the oysters were reweighed and half from each dome were sacrificed. The re- maining oysters were replaced in the field for a 4-wk recovery period without B. impressa. All B. im- pressa were removed from the previously para- sitized oysters by hand prior to replacement. The domes were visited twice weekly and each oyster handled as before. No B. impressa were found on the oysters during the recovery period. Laboratory Analysis Perkinsus marinus infection was measured in each oyster by incubating a small piece of mantle tissue in thioglycollate medium by the method of Ray (1966b). Intensity of infection was based on a semi- quantitative 0 to 5 rating assigned during micro- scopic inspection of the tissue after treatment with Lugol's solution (Mackin 1962). Small pieces of gonadal tissue were preserved in Bouin's fixative, sectioned and stained in 0.5% toluidine blue (Preece 1972) for gonadal analysis. Mantle cavity volume was determined by filling the shells with silicon caulking. Oyster growth, as expressed by shell de- position, was measured using the underwater weigh- ing method. Condition index was derived by divid- ing the total lyophilized dry weight by the mantle cavity volume (Lawrence and Scott 1982). The mantle and adductor muscle from each oyster 554 WILSON ET AL.: EFFECT OF SNAIL ON OYSTERS TREATMENTS: 4 WEEK EXPERIMENTAL PERIOD Unparasitized Controls Parasitized 4 WEEK RECOVERY PERIOD Recovery Control Previously Parasitized I 1 Figure 1.— Position of plexiglass domes during 4-wk experimental period and during 4-wk recovery period. were frozen immediately on dry ice and kept at -40°C until lyophilized. Prior to biochemical anal- ysis, lyophilized tissues were weighed and homog- enized on ice. Total lipids were isolated from a por- tion of the homogenate using the water:methanol: chloroform method of Folch et al. (1957) with a modified ratio of 0.8:2.1. This isolation produced a two phase system and an insoluble pellet after cen- trifugation in the cold. The pellet was used for pro- tein and glycogen assays. The water phase was used for amino acid analysis. The organic phase was divided and dried under Ng gas at 40°C. One half was used to determine lipid phosphate concentra- tion by the spectrophotometric method of White et al. (1979). The other half was dissolved in chloro- form to which nonadecanoic acid was added as an internal standard. The redissolved lipids were frac- tionated on a silicic acid (Unisil,^ 100-200 mesh) column. Neutral lipids were recovered by eluting with 10 times the column volume of chloroform. The chloroform was removed under N2 at 40°C ((rehron and White 1982). The resulting chloroform fraction, containing fatty acid methyl esters, was analyzed by gas chromatography using a capillary, nonpolar, methyl silicone high performance column and flame ionization detector. The amino acids were analyzed on a Dionex 3000 amino acid analyzer. Because residual chloroform interfered with the analysis, a second extraction was performed on the sample in the cold using a chloro- form:water ratio of 1:3; a-amino-n-butyric acid (2.5 fimole • mL~^ sample) was used as an internal standard. The amino acids accounting for the bulk of the free amino acid (FAA) pool, taurine, hypo- 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. taurine, aspartic acid, serine, threonine, glutamine, glutamate, glycine, and alanine, were separated us- ing a lithium citrate buffer and measured using o-phthalaldehyde as the detecting compound. Total soluble protein was estimated by Peterson's (1977) modification of the Lowry method. Protein was precipitated using a final concentration of 10% trichloroacetic acid in the cold. After centrifugation at 4°C the resultant pellets were resuspended in 50:50, 10% SDS:0.8N NaOH for 30 minutes. Replicate samples were centrifuged and analyzed spectrophotometrically for protein concentration. Bovine serum albumin (Sigma) was used as the standard. Glycogen was degraded to glucose enzymatically with amyloglucosidase (Carr and Neff 1984). After preincubation with amyloglucosidase, glucose was measured using the glucose oxidase-peroxidase pro- cedure (Sigma glucose kit) (Roehrig and Allred 1974). Rat liver glycogen (Sigma) was used as a standard. Statistical Analysis To assess the effect of Boonea impressa on the oysters during the 4-wk exposure period, the level of snail parasitism for each oyster had to be deter- mined. Because the number of snails on each oyster was counted only every third day, the total number of snails that parasitized each oyster was estimated by assuming that the same number of snails were present on days between counts as found on the previous visit. The total number of snail-days per oyster, the sum of the number of snails present on each day during the 4-wk exposure, will be referred to as the snail scale. For example, during the first week of the exposure period, one oyster had 15 555 FISHERY BULLETIN: VOL. 86, NO. 3 snails on Monday, 10 on Wednesday, and 4 on Satur- day; therefore Tuesday was assigned 15 snails, Thursday and Friday 10 snails, and the total for the week was 64 snail-days. Except where noted in the text, results were analyzed by multiple analysis of covariance (MANCOVA) using ranked data followed by Dun- can's multiple range tests (a = 0.05) to locate sig- nificant differences within the MANCOVA. Mantle cavity volume was used as the covariate for com- paring levels of P. marinus infection intensity, dry weight, and shell weight gain between treatment groups (e.g., parasitized and control oysters). Similar tests comparing biochemical parameters between treatment groups included the intensity of P. marinus infection and tissue dry weight as co- variates. When the requirement of parallelism was not met, the MANCOVA was modified as described by Smith and Coull (1987). Perkinsus marinus in- fection intensity was used as both a dependent and independent variable because, although B. impressa can influence the intensity of P. marinus infection (White et al. 1987), P. marinus itself can affect the biochemical composition of oysters (Soniat and Koenig 1982; White et al. in press). Only parasitized oysters (or those recovering from parasitism) were used in analyses of the effect of snail scale (intensity of parasitism). The FAA pool was analyzed three ways: 1) the entire pool, 2) the pool minus taurine and hypo- taurine because these are the only amino acids not found in protein, and 3) after exclusion of the major components (taurine, hypotaurine, glycine, and ala- nine) so that changes in the lesser constituents of the pool could be examined. All analyses for the 4-wk exposure period used MANCOVA analyses with nested variables, which took into account which of the 4 domes the oysters were in during the exposure period. Overall, signif- icant differences in biochemical composition (i.e., amino acid content, glycogen, protein, etc.) between the 2 nonparasitized domes or between the 2 snail parasitized domes did not occur more frequently than expected by chance (Binomial Test, a = 0.05). Therefore, the two equivalently treated domes of each treatment group (Fig. 1) were lumped together for comparison with precontrol and recovery oysters. Nevertheless, the experimental design represents a case of pseudoreplication; hence the caveats of Hurlbert (1984) should be considered when reviewing the statistical analysis. RESULTS Growth, Perkinsus marinus Intensity, and Reproduction Boonea impressa affected oyster growth rate, the intensity of infection by P. marinus, and reproduc- tive state. The mean initial shell weights of the precontrol, control, and parasitized groups were not significantly different from each other. By the end of the 4-wk exposure period, both parasitized and unparasitized oysters had gained weight, but oysters parasitized by B. impressa gained significantly less weight than unparasitized oysters (Table 1). Oysters with more snails typically gained less weight than oysters with fewer snails. The relationship between the intensity of B. impressa parasitism (snail scale) and weight gain among parasitized oysters was significant (Spearman's rank, P = 0.03; weight gain normalized to initial weight, P = 0.001, Fig. 2). Mantle cavity volume did not vary among any of the treatment or recovery groups. Condition index was not significantly different between treatment groups or between recovery groups. Condition in- dex was significantly higher in the two recovery Table 1 .—Mean and standard deviations for initial shell weight (g) and average shell weight gain per oyster during the 4-wk exposure period and 4-wk recovery period. Significance levels (sig.) from Duncan's multiple range test (o = 0.05). A test was restricted to a single column. Groups having the same letter within a column are not significantly different. Initial weight Weight gained during 4-wk exposure Mean ± SD Sig. Weight gained during 4-wk recovery Group Mean ± SD Sig. Mean ± SD Sig. Precontrol (n = 15) Treatment Control (n = 40) Parasitized (n = 40) Recovery Control (n = 6) Previously parasitized (n = 15) 22.82 ± 12.16 A 17.16 ± 8.48 A 16.84 ± 7.47 A 3.49 ±1.21 A 2.54 ±1.31 B 2.16 ± 0.73 B 2.89 ±1.05 A 556 WILSON ET AL.: EFFECT OF SNAIL ON OYSTERS Parasitized U.4J) " • 0.40 - 0.35 - O) 0.30 • S • c CO O 0.25 • 0.20 • • o 0.15 • • • • • • 0.10 • • • • • 0.05 • • I 1 I 1 ' Previously Parasitized 0.45 • • 0.40 0.35 • O) 0.30 • • 2 c CO (3 *-* SI 0.25 • • (f) o 0.20 0.15 • 0.10 • • • • 0.05 • 1 1 1 1 1 100 200 300 Snail Scale 400 500 100 200 300 Snail Scale 400 500 Figure 2.— Left: Weight gain (g) per g initial weight of individual snail-parasitized oysters after the 4-wk experimental period as a func- tion of the level of Boonea impressa parasitism, or snail scale. Right: Weight gain (g) per g initial weight of individual oysters after the 4-wk recovery period as a function of the level of B. impressa (snail scale) during the 4-wk treatment period. groups than in the treatment and precontrol groups (Table 2). During the recovery period, when all oysters were parasite-free, the previously parasitized oysters (those with snails during the treatment period) gained significantly more weight than the recovery controls (Table 1). Oysters which were previously parasitized by fewer B. impressa during the treat- ment period gained more weight during the recovery period than those that were previously parasitized by more snails (Spearman's Rank corre- lation, P = 0.07, Fig. 2), but the relationship was considerably weaker than during the treatment period. The intensity of infection by P. marinus increased throughout the experiment in parasitized oysters so that, after recovery, previously parasitized oysters had higher prevalences and intensities of infection than they did after 4 weeks of parasitism; these values in turn were higher than the precontrol values at the experiment's inception (Table 3). Exactly the opposite trend occurred in control oysters. Recovery controls had the lowest values of prevalence and intensity. Consequently, after the recovery period, previously parasitized oysters had significantly higher intensities of infection than recovery controls. The proportion of infected in- dividuals (33% of the controls, 93% of the previously parasitized oysters) was significantly higher as well (x^, P < 0.05). Within the parasitized oysters, the intensity of parasitism (snail scale) did not correlate with the increase in intensity of P. marinus infec- tion during either the treatment or recovery period (Fig. 3, Spearman's Rank, P > 0.05; the interaction 557 FISHERY BULLETIN: VOL. 86, NO. 3 Table 2.— Means and standard deviations for volume (mL), condition index (g mL"') and dry weight (g). Significance levels (sig.) from Duncan's multiple range test (o = 0.05). A test was restricted to a single column. Groups having the same letter within a column are not significantly different, n, see Table 1 . Volume Condition index Dry weight Group Mean ± SD Sig. Mean ± SD Sig. Mean ± SD Sig. Precontrol 10.4 ± 4.1 A 3.4 ± 0.8 C 0.35 ± 0.14 AS Treatment Control 11.2 ± 3.6 A 5.4 ± 1.0 B 0.34 ± 0.11 AB Parasitized 9.2 ± 1.7 A 5.5 ± 1.0 B 0.28 ± 0.12 B Recovery Control 9.5 ± 4.6 A 7.1 ± 1.5 A 0.32 ± 0.14 B Previously parasitized 10.8 ± 3.8 A 7.7 ± 1.5 A 0.51 ± 0.24 A Table 3.— Mean and standard deviations for Perkinsus marinus infection intensity among groups. Significance levels from Dun- can's multiple range test (o = 0.05) apply to the infection intensity data. Groups with the same letter are not significantly different. Percent incidence is the number of oysters infected divided by the number of oysters in the sample. Infection intensity was calculated using the 5-point scale of Mackin (1962). n, see Table 1. Table 4. — Results of histological examination of gonadal samples for each group. Significance levels (sig.) from Duncan's multiple range test (a = 0.05) are for differences In the number of eggs present per female in a histological section. Percent of females spawning is the number of females with eggs observed in the gonoducts divided by the number of females in the sample. Perkinsus marinus Percent of females Infection Intensity Signifi- Preva- Sex Eggs present Group Mean ± SD cance lence Group Female Male Mean ± SD Sig. spawning Precontrol 1.47 ± 1.45 AB 66.6 Precontrol 7 7 31.1 ± 4.4 A 86.0 Treatment Treatment Control 1.10 ± 1.06 AB 65.0 Control 6 4 36.6 ±11.0 A 100.0 Parasitized 1.63 ± 1.32 AB 79.0 Parasitized 4 6 21.6 ± 3.9 B 100.0 Recovery Recovery Control 0.83 ± 1.60 B 33.3 Control 5 1 37.6 ± 6.0 A 80.0 Previously Previously parasitized 2.21 ± 1.38 A 92.9 parasitized 11 4 42.9 ± 9.8 A 81.0 term in the MANCOVA was also nonsignificant). The number of eggs per microscopic field esti- mated from histological sections, was significantly lower in parasitized oysters after the 4-wk treatment period (Table 4). No differences between controls and previously parasitized oysters were present after the 4-wk recovery period. The proportion of oysters ready to spawn (female oysters which had eggs visible in the gonoducts) was not significantly different in parasitized and unparasitized oysters during treatment or recovery (x^, a = 0.05). Biochemical Composition Mean levels of the various biochemical compo- nents measured in adductor muscle and mantle tissue are given in Tables 5 and 6. A comparison of precontrol, treatment control, and recovery control oysters documents the changes in biochemical com- position produced by handling stress and natural environmental changes that occurred during the experiment. Glycogen increased significantly in the mantle tissue during the 4-wk treatment period, then decreased during the recovery period (Table 7). Glycogen in the adductor muscle also decreased during the recovery period (Table 7). Hypotaurine increased in both tissues during the treatment period (Table 8). Most other FAA and the total pool dropped in concentration in the mantle tissue in the first 4 weeks, but then stabilized. In contrast, in the adductor muscle, significant increases in glycine and hypotaurine during the treatment period were off- set by a significant decrease in alanine, so the total pool changed little. Again, the FAA pool stabilized during the first 4 weeks. Hence, there was little difference in treatment and recovery controls in any measured biochemical component. Snail parasitized and control oysters did not differ significantly in the concentration of any of the biochemical components in either the mantle or adductor muscle (Tables 7, 8). Only one biochemical parameter differed between the recovery control 558 WILSON ET AL.: EFFECT OF SNAIL ON OYSTERS Parasitized 4.5 r • 4.0 3.5 B 3.0 _c c o CO 3 § 2.0 CD CO ^ 1.5 0) 1.0 0.5 • • • 100 -••- 200 300 Snail Scale 400 500 Previously Parasitized 4.5 4.0 3.5 CO r 0 3.0 ■*-* c c o .^^ o ? h 0) c CO 3 c ?0 l~ CO E to 3 eo C 1.5 * >- 0) Q. •• 1.0 0.5 • • 100 200 300 Snail Scale 400 500 Figure 3.— Left: Intensity oiPerkinsus marinus infection in snail -parasitized oysters as a function of the level of snail parasitism after the 4-wk treatment period. Right: Intensity of P. marinus infection in previously snail-parasitized oysters as a function of the level of snail parasitism after the 4-wk recovery period. and previously parasitized oysters. Recovery con- trols had significantly less glycogen in the mantle tissue (Table 7). No biochemical component was significantly correlated with the intensity of snail parasitism (snail scale) in parasitized oysters dur- ing the 4-wk treatment period except taurine in the adductor muscle (Table 9). A similar comparison, using the previously parasitized oysters after the 4-wk recovery yielded only two significant correla- tions with snail scale; soluble protein and hypo- taurine content. In addition, lipid phosphate in the mantle tissue and fatty acid content in the adductor muscle were significantly correlated with the inten- sity of Perkinsus marinus infection after the recovery period (all recovery oysters, controls and previously parasitized, were included in the analysis. Table 9). DISCUSSION Oysters averaged as few as about 3 and as many as 14 snails per day during the 4-wk treatment period; however, only 11% of the oysters had 10 or more snails per day and 76% had fewer than 9 per day. This range of parasitism is typical for many reefs in the experimental area (Copano Bay- Aransas Bay, TX-White et al. 1984; Powell et al. 1987) and corresponds to the lower parasite levels used by White et al. (1988, in press) in laboratory studies on these animals. The effects of snail parasitism were minor on a biochemical level, but substantial on an organismal level. Growth, reproductive capacity, and health, as measured by Perkinsus marinus infection, were significantly affected. Other odostomians also re- 559 FISHERY BULLETIN: VOL. 86, NO. 3 Table 5.— Means and standard deviations for all biochemical components in mantle tissue. FAA = total of the 9 free amino acids measured; FAA- = FAA less taurine and hypotaurine; FAA- = FAA less taurine, hypotaurine, glycine, and alanine; amino acids in ^moles mL" \ other tissue components in mg • g dry wt'\ Treatment Recovery Previously Precontrol Control Parasitized Control parasitized (n = 15) {n = 10) (n = 10) (n = 6) (n = 15) X ± SD X ± SD X ± SD X ± SD X ± SD FAA 351 ± 94 240 ± 39 280 ± 38 294 ± 64 270 ± 64 FAA- 220 ± 67 116 ± 14 121 ± 15 134 ± 50 117 ± 43 FAA- 57 ± 18 42 ± 6 46 ± 7 49 ± 26 42 ± 13 Taurine 102 ± 33 95 ± 27 119 + 24 108 ± 15 99 ± 21 Hypotaurine 27 ± 10 39 ± 10 50 ± 23 49 ± 10 54 ± 20 Aspartic acid 22 ± 5 15 ± 2 19 ± 3 22 ± 13 18 ± 5 Serine 8 ± 5 5 ± 2 6 ± 2 5 ± 3 5 ± 3 Threonine 3 ± 1 1 ± 0.2 1 ± 0.7 1 ± 0.4 1 ± 0.5 Glutamine 8 ± 5 2 ± 1 3 ± 0.8 4 ± 3 5 ± 3 Glutamic acid 14 ± 5 16 ± 3 18 ± 5 14 ± 5 12 ± 4 Glycine 71 ± 23 33 ± 5 38 ± 10 44 ± 31 38 ± 17 Alanine 87 ± 33 36 ± 8 38 ± 10 31 ± 10 34 ± 16 Protein 62 ± 45 28 ± 39 60 ± 57 45 ± 40 23 ± 16 Glycogen 28 ± 22 111 ± 59 66 ± 28 37 ± 13 88 ± 48 Lipid phosphate (as PO4-3) 0.62 ± 0.41 0.48 ± 0.27 0.82 ± 0.39 0.5 ± 0.2 0.87 ± 0.54 Fatty acids 20.7 ± 33.1 33.2 ± 49.1 48.7 ± 58.1 9.4 ± 10.5 6.3 ± 3.2 Table 6.— Means and standard deviations for biochemical components of adductor muscle tissue. See Table 5 for additional information. Treatment Recovery Previously Precontrol Control Parasitized Control parasitized {n = 11) {n = 10) (n = 10) (" = 6) {n = 15) X ± SD X ± SD X ± SD X ± SD X ± SD FAA 328 ± 119 306 ± 56 289 ± 47 277 ± 110 347 ± 81 FAA- 272 ± 121 227 ± 54 207 ± 35 210 ± 95 266 ± 78 FAA- 57 ± 16 57 ± 8 44 ± 5 75 ± 25 84 ± 24 Taurine 52 ± 17 53 ± 25 66 ± 25 42 ± 16 47 ± 13 Hypotaurine 8 ± 4 22 ± 7 23 ± 3 24 ± 10 33 ± 6 Aspartic acid 17 ± 7 18 ± 4 14 ± 4 18 ± 6 21 ± 8 Serine 5 ± 4 7 ± 3 5 ± 2 27 ± 21 28 ± 21 Threonine 3 ± 1 2 ± 0.9 1 ± 0.5 2 ± 1 3 ± 1 Glutamine 6 ± 2 4 ± 2 2 ± 1 6 ± 4 10 ± 8 Glutamic acid 20 ± 6 18 ± 4 21 ± 6 21 ± 4 20 ± 5 Glycine 54 ± 15 124 ± 37 119 ± 42 85 ± 54 121 ± 44 Alanine 142 ± 63 49 ± 19 39 ± 10 49 ± 30 60 ± 26 Protein 41 ± 13 49 ± 40 45 ± 56 45 ± 35 36 ± 22 Glycogen 6 ± 3 7 ± 3 6.8 ± 0.3 4 ± 2 4 ± 2 Lipid phosphate (as PO;3) 0.1 ± 0 0.2 ± 0 0.34 ± 0.3 0.25 ± 0.28 0.2 ± 0.1 Fatty acids 15.0 ± 10.0 17.5 ± 16.7 15.3 ± 24.1 6.9 ± 11.5 8.6 ± 16.9 duce host growth rate (Nishino et al. 1983). Oysters parasitized by Boonea impressa gained significant- ly less weight than nonparasitized oysters. Para- sitized female oysters had significantly fewer eggs than controls (the effect on males was not quan- tified). Reduced growth and impaired reproduction in the host are commonly associated with marine parasites (e.g., Menzel and Hopkins 1955; Cheng et al. 1983; Hawkes et al. 1986). Starvation produces a similar phenomenon (Fair and Sick 1982; Pipe 1985; Wright and Hetzel 1985; Devi et al. 1985). Several lines of evidence suggest that a reduction in net productiv- ity, but not a negative energy balance, produced the results observed here. White et al. (1988) developed an energy flow model for oysters and snails. Using that model, oysters of 4 to 7 cm long, the size we used, would not incur a negative energy balance un- 560 WILSON ET AL.: EFFECT OF SNAIL ON OYSTERS Table 7.— Results of Duncan's multiple range test for biochemical components of the mantle and adductor muscle. Each test considered data for one biochemical component (one vertical group of 5 means) for each tissue. Different letters within a vertical group indicate significant differences at a = 0.05 within that group. Lipid Fatty Lipid Fatty Protein Glycogen phosphate acids Protein Glycogen phosphate acids Mantle Adductor muscle Precontrol A B A A Precontrol A AB A A Treatment Treatment Control B A A A Control A A A A parasitized AB A A A Parasitized A AB A A Recovery Recovery Control AB B A A Control A B A A Previously Previously parasitized B A A A parasitized A B A A Table 8. — Results of Duncan's multiple range test for free amino acids in the mantle tissue and adductor muscle. Each test considered data for one biochemical component (one vertical group of 5 means) for each tissue. Different letters within a vertical group indicate significant differences at a = 0.05 within that group. Tau = taurine; Hyp = hypotaurine; Asp = aspartic acid; Ser = serine; Glu = glutamic acid; Gin = glutamine; Gly = glycine; Ala = alanine; FAA = total for the amino acids measured; FAA- = FAA less taurine and hypotaurine; FAA- = FAA less taurine, hypotaurine, glycine, and alanine. Tau Hyp Asp Ser Thr Gin Glu Gly Ala FAA FAA- FAA- Mantle Precontrol A B A A A A AB A A A A A Treatment Control A A B A B B AB B B B B A Parasitized A A AB A AB B A B B AB B A Recovery Control A A AB A B AB B B B AB B A Previously parasitized A A AB A B AB B B B B B A Adductor muscle Precontrol A C AB B A A A C A A A BC Treatment Control A B AB B AB A A AB B A A CD Parasitized A B B B B A A AB B A A D Recovery Control A AB AB A AB A A BC B A A AB Previously parasitized A A A A A A A AB B A A A Table 9.— Significant P-values obtained by MANCOVA from comparison of snail scale and Perkinsus mahnus infection intensity. Tests using snail scale considered only the parasitized oysters after the treatment period and the previously parasitized oysters after the recovery period. Tests using P. mahnus considered all oysters, controls and snail parasitized. Abbreviations same as Table 8. M = mantle tissue; A = adductor muscle; — = not significant in either tissue at o = 0.05. Lipid Gly- phosphate cogen (as PO4 ^) Fatty acids Tau Hyp Asp Thr Ser Glu Gly Ala FAA FAA- FAA- Protein Treatment Snail scale Perkinsus mahnus Recovery Snail scale Perkinsus mahnus — 0.02(A) — — — — ___— — — — — _ _ _ 0.004(A) — — ——— — — 0.03(M) 0.002(A) — 0.04(A) — — — — — ———— 0.004(M) 561 FISHERY BULLETIN: VOL. 86, NO. 3 til fed upon by at least 25 average-sized (1.75 mm maximum width) snails. Seven snails, a typical value in our experiments, would reduce net productivity by only 5 to 30%, on a daily basis. In addition, con- dition index and mantle glycogen levels increased during the treatment period in both control and parasitized oysters and the effect of snail parasitism on all biochemical components was small (oysters can regulate some biochemical components even during starvation, Swift and Ahmed 1983; but see Riley 1976). Consequently, both reduced growth and impaired reproductive capacity can be attributed to a reduc- tion in assimilated carbon available to the host, as a result of either a reduction in filtration rate result- ing in less energy being assimilated (Ward and Langdon 1986) or the direct removal of assimilated carbon by the snail. Neither effect was permanent. Growth rate resumed and reproductive state re- turned to control levels during recovery. In both cases compensatory adjustments occurred during the recovery period so that previously parasitized oysters gained weight and increased egg number faster than the controls. Loosanoff and Nomejko (1955) and Eagle and Chapman (1953) also noted compensatory shell growth in oysters. Perkinses marirms infection is an important cause of mortality in oysters (Mackin and Sparks 1962; Mackin 1962; Hofstetter 1977). Boonea impressa can transmit P. marinus from one oyster to another and can also increase the intensity of infection (White et al. 1987). Continued deterioration after a stress is removed occurs frequently in "recovery" experi- ments (Kendall et al. 1984; Powell et al. 1984) demonstrating the necessity of examining recovery capacity in acute (vs. chronic) stresses. Growth rate typically recovers more rapidly than most biochem- ical parameters. In snail-parasitized oysters, both prevalence and intensity of P. marinus infection in- creased during the recovery period. Hence, in con- trast to growth rate, no recovery from P. marinus actually occurred. By the end of the 8-wk period, infection intensity had increased by about 1 unit on Mackin's (1962) 5 point scale. Growth still occurred and reproductive capacity returned to control levels, however, during this period. These results contra- dict those of Menzel and Hopkins (1955) who showed that P. marinus retarded growth in proportion to the intensity of infection and Mackin (1953) who observed decreased fecundity in heavily infected oysters. Haven (1962) obtained results analogous to ours. Possibly, the oysters in our study were not infected heavily enough to retard growth and reproduction. Mean infection intensity in previous- ly parasitized oysters after the recovery period was only 2.2, a light to moderate infection. Changes observed at the biochemical level among the previously parasitized oysters after the recovery period, particularly in fatty acid and lipid phosphate content, were related to increased infection inten- sity of P. marinus (Table 9). Stein and Mackin (1955), Mackin (1962), and White et al. (in press) also noted changing lipid levels related to infection intensity. Lipid phosphate is predominantly a struc- tural component whereas fatty acids, usually as tri- glycerides, are storage materials in many marine invertebrates (Gabbott 1976; Trider and Castell 1980; Gehron and White 1982). The increased lipid phosphate content in mantle tissue, however, prob- ably indicates an increase in structural material, noted histologically by Stein and Mackin (1955, 1957) to occur in conjunction with P. marinus infections. Glycogen is the primary storage material in most bivalves (Beninger and Lucas 1984). Parasitism fre- quently affects carbohydrate metabolism (Cheng 1963; Mohamed and Ishak 1981; Thompson and Binder 1984; Thompson 1986). White et al. (in press) suggested that P. marinus alters oyster metabolism favoring gluconeogenesis. Our results support this hypothesis. Glycogen levels dropped only in recovery control oysters in which P. marinus infection inten- sity also declined. Changing fatty acid content might be similarly explained. In contrast. Stein and Mackin (1957) noted decreased glycogen levels in heavily in- fected oysters (3 to 5 on Mackin's scale). Few of our oysters were this heavily infected, however. An alternative explanation, that slower reproductive development in parasitized oysters was responsible for variation in glycogen and fatty acid content, can- not be completely excluded because P. marinus in- fection intensity did not correlate with glycogen levels in recovery oysters. Parasitized oysters had fewer eggs than control oysters, however. Addi- tionally, neither the number of eggs present nor the number of oysters spawning differed significantly between control and previously parasitized oysters during the recovery period. Results of previous workers suggest that signif- icant decreases in storage compounds, whether caused by B. impressa or P. marinus, only occur in heavily infected animals (e.g.. Stein and Mackin 1957; White et al. in press). This, too, is true for the amino acid pool where significantly decreased levels are usually associated with more severely stressed animals (e.g., Powell et al. 1982, 1984). The few significant effects on amino acids in this study, like all the other biochemical components measured. 562 WILSON ET AL.: EFFECT OF SNAIL ON OYSTERS were produced by increases in concentration. Soniat and Koenig (1982) observed significant changes in the free ammo acid pool due to P. marinus, par- ticularly in taurine concentration. We noted changes in taurine during the treatment period and hypo- taurine during recovery in the adductor muscle but these were related to snail parasitism. Biochemical components, though rarely signifi- cantly affected, were affected not just in the mantle tissue but also in the adductor muscle. One possi- bility, that the snail's effect is localized at the point of feeding, is not supported by the data. Snail para- sitism produces systemic effects. CONCLUSIONS Complementary results of White et al. (1984, 1988, in press) and this study permit a general description of the impact of snail parasitism on oysters at normal field levels. Both growth rate and reproductive development slow significantly, but recover rapidly once the snails are removed. Hence, the temporal stability of snail patches must deter- mine the cumulative effect on field populations of oysters. The prevalence and intensity of infection by Perkinsus marinus is significantly increased, but recovery does not occur. That is, Boonea impressa probably facilitates and encourages the normal spread and intensification of P. marinus from which oysters, if they recover, only do so the following winter when low temperatures typically reduce in- fection levels (Hewatt and Andrews 1956; Burrell et al. 1984; Soniat 1985). This effect, then, is long term. Most changes in biochemical components were due to infection by P. marinus. Snail feeding reduces net productivity but, at normal field levels, starvation is an unlikely result. Increased infection by P. marinus typically raises glycogen and lipid levels, at least in light to moderate infections. Of the free amino acids, taurine and hypotaurine have been shown to be affected by P. marinus and B. im- pressa. Little change in the remaining FAA or the total pool has been observed singly or in concert. Feng et al. (1970) noted increased taurine levels in oysters parasitized by Bucephalus sp. and Minchinia nelsoni. Hence, an increase in taurine and hypo- taurine levels apparently is a general response to parasitism in oysters. Yuill (1987) emphasized the importance of subtle effects produced by parasites on host populations. Perkinsus marinus is an important source of mor- tality in oyster populations (Mackin and Sparks 1962; Mackin 1962; Hofstetter 1977). Data suggest that one of the most important aspects of parasitism by B. impressa is to encourage this second parasitic organism. To this extent, over the year, B. impressa at normal field densities could be responsible for a substantial amount of mortality in oyster popula- tions. ACKNOWLEDGMENTS We would like to thank M. White, D. Davies, and L. Priest for laboratory and statistical assistance. T. J. McDonald graciously provided his expertise in GC analysis, and M. C. Kennicutt and A. Vastano ran the fatty acid analyses. Suggestions by J. Par- rack, T. Bright, and an anonymous reviewer im- proved the manuscript. We thank R. Covington for typing the manuscript and tables. R. Pratt, care- taker of the Aransas Pass Lighthouse, provided space for the field portion of the study. This research was supported by part of an institutional grant NA85AA-D-FG128 to Texas A&M University by the National Sea Grant Program, National Oceanic and Atmospheric Administration, U.S. Department of Commerce to E. N. Powell and S. M. Ray. We ap- preciate this support. LITERATURE CITED Allen, J. F. 1958. Feeding habits of two species of Odostomia. Nautilus 72:11-15. Andrews, J. D. 1961. Measurement of shell growth in oysters by weighing in water. Proc. Natl. Shellfish Assoc. 52:1-12. 1965. Infection experiments in nature with Dermocystidium marinum in Chesapeake Bay. Chesapeake Sci. 6:60-67. Andrews, J. D., and W. G. Hewatt. 1957. Oyster mortality studies in Virginia. II. The fungus disease caused by Dermocystidium Tnarinum in oysters of Chesapeake Bay. Ecol. Monogr. 27:1-25. Beninger, p. G., and a. Lucas. 1984. Seasonal variations in condition, reproductive activity, and gross biochemical composition of two species of adult clam reared in a common habitat: Tapes decussatus L. (Jeffreys) and Tapes phillippinarum (Adams & Reeve). J. Exp. Mar. Biol. Ecol. 79:19-37. Brockelman, C. R. 1978. Effects of parasitism and stress on hemolymph protein of the African giant snail, Achatinafulica. Z. Parasitenkd. 57:137-144. Brown, C. R., and M. B. Brown. 1986. Ectoparasitism as a cost of coloniality in cliff swallows (Hirundo pyrrhonota). Ecology 67:1206-1218. Burrell, V. G., M. Y. Bobo, and J. J. Manzl 1984. A comparison of seasonal incidence and intensity of Perkinsus marinus between subtidal and intertidal oyster populations in South Carolina. J. World Mariculture Soc. 15:301-309. Carr, R. S., and J. M. Neff. 1984. Quantitative semi-automated enzymatic assay for tissue 563 FISHERY BULLETIN: VOL. 86, NO. 3 glycogen. Comp. Biochem. Physiol. B Comp. Biochem. 77:447-449. Cheng, T. C. 1963. The effects of Eckinoparyphium larvae on the struc- ture of and glycogen deposition in the hepatopancreas of Helisoma trivolvis and glyconeogenesis in the parasite lar- vae. Malacologia 1:291-301. Cheng, T. C, J. T. Sullivan, K. H. Rowland, T. F. Jones, AND H. J. MORAN. 1983. Studies on parasitic castration: soft tissue and shell weights of Ilyanassa obsoleta (MoUusca) parasitized by larval trematodes. J. Invertebr. Pathol. 42:143-150. Devi, V. U., Y. P. Prabhakara Rao, and D. G. V. Prasada Rao. 1985. Body component indices of an intertidal gastropod Morula granulata (Daclos) subjected to starvation. Indian J. Mar. Sci. 14:44-45. Eagle, J. B., and C. R. Chapman. 1953. Oyster condition affected by attached mussels. Natl. Shellfish Assoc. Conv. Addendum 1951, p 70-78. Fair, P. H., and L. V. Sick. 1982. Serum amino acid concentrations during starvation in the prawn, Macrobrachium rosenbergii, as an indicator of metabolic requirements. Comp. Biochem. Physiol. B Comp. Biochem. 73:195-200. Feng, S. Y., E. A. Khairallah, and W. J. Canzonier. 1970. Hemolymph-free amino acids and related nitrogenous compounds of Crassostrea virginica infected with Bucepha- lus sp. and Minchinia nelsoni. Comp. Biochem. Physiol. 34:547-556. Folch, J., M. Lees, and G. H. S. Stanley. 1957. A simple method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem. 226:497- 509. Ford, S. E. 1986. Comparison of hemolymph proteins from resistant and susceptible oysters, Crassostrea virginica, exposed to the parasite Haplosporidium nelsoni (MSX). J. Invertebr. Pathol. 47:283-294. Fretter, V. 1951. Turbonilla elegantissima (Montagu), a parasitic opis- thobranch. J. Mar. Biol. Assoc. U.K. 30:37-47. Fretter, V., and A. Graham. 1949. The structure and mode of life of the Pyramidellidae, parasitic opisthobranchs. J. Mar. Biol. Assoc. U.K. 28: 493-532. Gabbott, P. A. 1976. Energy metabolism. In B. L. Bayne (editor). Marine mussels: their ecology and physiology, p. 293-356. Cam- bridge Univ. Press, N.Y. Gehron, M. J., AND D. C. White. 1982. Quantitative determination of the nutritional status of detrital microbiota and the grazing fauna by triglyceride glycerol analysis. J. Exp. Mar. Biol. Ecol. 64:145-158. Haux, C, M. Sjobeck, AND A. Larsson. 1985. Physiological stress responses in a wild fish population of perch (Perca fluviatilis) after capture and during subse- quent recovery. Mar. Environ. Res. 15:77-95. Haven, D. S. 1962. Seasonal cycle of condition index of oysters in the York and Rappahannock Rivers. Proc. Natl. Shellfish. Assoc. 51:42-66. Hawkes, C. R., T. R. Meyers, and T. C. Shirley. 1986. Length-weight relationships of blue, Paralithodes platypus, and golden, Lithodes aequispina, king crabs para- sitized by the rhizocephalan, Briurosaccus callosus Boschma. Fish. Bull., U.S. 84:327-332. Hewatt, W. G., and J. D. Andrews. 1956. Temperature control experiments on the fungus disease, Dermocystidium marinum, of oysters 1. Proc. Natl. Shellfish. Assoc. 46:129-133. HOFSTETTER, R. P. 1977. Trends in population levels of the American oyster, Crassostrea virginica Gmelin on public reefs in Galveston Bay, Texas. Tex. Parks Wildl. Dep. Tech. Ser. No. 24, 90 p. Hopkins, S. H. 1956. Odostomia impressa parasitizing southern oysters. Science (Wash. D.C.) 124:628-629. Hurlbert, S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monogr. 54:187-211. Kabat, a. R. 1986. Effects of trematode parasitism on reproductive out- put of the bivalve Transennella tantilla. Can. J. Zool. 64: 267-270. Kendall, Jr., J. J., E. N. Powell, S. J. Connor, T. J. Bright, AND C. E. Zastrow. 1984. The importance of monitoring metabolic recovery in the coral Acropora cervicomis after short-term exposure to drilling muds: calcification rate and protein concentration. Coral Reefs. 2:215-225. Koenig, M. L., E. N. Powell, and M. R. Kasschau. 1981. The effects of salinity change on the free amino acid pools of two nereid polychaetes, Neanthes succinea and Leonereis culvert. Comp. Biochem. Physiol. A Comp. Physiol. 70:631-637. KUKAL, 0., AND P. G. KEVAN. 1987. The influence of parasitism on the life history of a high arctic insect, Gynaephora groenlandica (Wocke) (Lepidop- tera:Lymontriidae). Can. J. Zool. 65:156-163. Lawrence, D. R., and G. I. Scott. 1982. The determination and use of condition index of oysters. Estuaries 5:23-27. LOOSANOFF, V. L., AND C. A. NOMEJKO. 1955. Growth of oysters with damaged shell-edges. Biol. Bull. (Woods Hole) 108:151-159. Mackin, J. G. 1953. Incidence of infection of oysters by Dermocystidium in the Barataria Bay area of Louisiana. Natl. Shellfish. Assoc. Conv. Addendum 1951, p 22-36. 1962. Oyster diseases caused by Dermocystidium marinum and other microorganisms in Louisiana. Publ. Inst. Mar. Sci., Univ. Tex. 7:132-229. Mackin, J. G., and A. K. Sparks. 1962. A study of the effect on oysters of crude oil loss from a wild well. Publ. Inst. Mar. Sci., Univ. Tex. 7:230-261. Mengebier, W. L., and L. Wood. 1969. The effects of Minchinia nelsoni infections on enzyme levels in Crassostrea virginica - II. serum phosphohexose isomerase. Comp. Biochem. Physiol. 29:265-270. Menzel, R. W., and S. H. Hopkins. 1955. The growth of oysters parasitized by the fungus Dermo- cystidium marinum and by the trematode Bucephalus cuculus. J. Parasitol. 41:333-342. Mohamed, a. M., and M. M. Ishak. 1981. Growth rate and changes in tissue carbohydrate dur- ing schistosome infection of the snail Biomphalaria alexan- drina. Hydrobiologia 76:17-21. Neff, J. M., P. D. BoEHM, AND W. E. Haensly. 1985. Petroleum contamination and biochemical alterations in oysters (Crassostrea gigas) and plaice (Pleuronectes platessa) from bays impacted by the Amoco Cadiz crude oil spill. Mar. Environ. Res. 17:281-283. 564 WILSON ET AL.: EFFECT OF SNAIL ON OYSTERS NiSHINO, T., S. NOJIMA, AND T. KiKUCHL 1983. Quantitative studies of the life history and interspecific relationship of two gastropod species, Odostomia sp. (ecto- parasite) and Umbonium (Suchium) moniliferum (Lamarck) (host). I. Life history and population dynamics oi Odostomia sp. Kyushu Univ., Publ. Amakusa Mar. Biol. Lab. 7:61- 79. Peterson, G. L. 1977. A simplification of the protein assay method of Lowry et al. which is more generally applicable. Anal. Biochem. 83:346-356. Pickering, A. D., T. G. Pottinger, and P. Christie. 1982. Recovery of the brown trout, Salmo trutta L., from acute handling stress: a time-course study. J. Fish Biol. 20:229-244. Pipe, R. K. 1985. Seasonal cycles in and effects of starvation on egg development in Mytilus edulis. Mar. Ecol. Prog. Ser. 24: 121-128. POrtner, H., B. Surholt, and M. Grieshaber. 1979. Recovery from anaerobiosis of the lug worm, Arenicola marina L.: changes of metabolite concentrations in the body-wall musculature. J. Comp. Physiol. 133:227-231. PovreLL, E. N., S. J. Connor, J. J. Kendall, Jr., C. E. Zastrow, AND T. J. Bright. 1984. Recovery by the coral Acropora cervicomis after drill- ing mud exposure. The free amino acid pool. Arch. Environ. Contam. Toxicol. 13:243-258. Powell, E. N., M. Kasschau, E. Chen, M. Koenig, and J. Pecon. 1982. Changes in the free amino acid pool during environmen- tal stress in the gill tissue of the oyster, Crassostrea vir- ginica. Comp. Biochem. Physiol. A Comp. Physiol. 71: 591-598. Powell, E. N., M. E. White, E. A. Wilson, and S. M. Ray. 1987. Small-scale spatial distribution of a pyramidellid snail ectoparasite, Boonea impressa, in relation to its host, Crasso- strea virginica, on oyster reefs. Pubbl. Stn. Zool. Napoli I Mar. Ecol. 8:107-130. Preece, a. 1972. A manual for histologic technicians. Little, Brown and Company, Boston, 428 p. Ray, S. M. 1954. Experimental studies on the transmission and patho- genicity of Dermocystidium marinum, a fungus parasite of oysters. J. Parasitol. 40:235. 1966a. Notes on the occurrence oi Dermocystidium marinum in the Gulf of Mexico coast during 1961 and 1962. Proc. Natl. Shellfish. Assoc. 54:45-54. 1966b. A review of the culture method for detecting Dermo- cystidium marinum, with suggested modifications and pre- cautions. Proc. Natl. Shellfish. Assoc. 54:55-69. Ray, S. M., J. G. Mackin, and J. L. Boswell. 1953. Quantitative measurement of the effect on oysters of disease caused by DerTnocystidium murinum. Bull. Mar. Sci. Gulf Caribb. 3:6-33. Riley, R. T. 1976. Changes in the total protein, lipid, carbohydrate, and extracellular body fluid free amino acids of the Pacific oyster, Crassostrea gigas, during starvation. Proc. Natl. Shellfish. Assoc. 65:84-90. Robertson, R. 1978. Spermatophores of six eastern North American pyra- midellid gastropods and their systematic significance (with the new genus Boonea). Biol. Bull. (Woods Hole) 155:360- 382. Robertson, R., and T. Mau-Lastovicka. 1979. The ectoparasitism of Boonea and Fargoa (Gastropoda: Pyramidellidae). Biol. Bull. (Woods Hole) 157:320- 333. Roehrig, K. L., and J. B. Allred. 1974. Direct enzymatic procedure for the determination of liver glycogen. Anal. Biochem. 58:414-421. Smith, L. 0., and B. C. Coull. 1987. Juvenile spot (Pisces) and grass shrimp predation on meiobenthos in muddy and sandy substrata. J. Exp. Mar. Biol. Ecol. 105:123-136. SONIAT, T. M. 1985. Changes in levels of infection of oysters by Perkinsus marinus, with special reference to the interaction of tem- perature and salinity upon parasitism. Northeast Gulf Sci. 7:171-174. SoNiAT, T. M., AND M. L. Koenig. 1982. The effects of parasitism by Perkinsus marinus on the free amino acid composition of Crassostrea virginica mantle tissue. J. Shellfish Res. 2:25-28. Stein, J. E., and J. G. Mackin. 1955. A study of the nature of pigment cells of oysters and the relation of their numbers to the fungus disease caused by DerTnocystidium marinum. Tex. J. Sci. 7:422-429. 1957. A histochemical study of the glycogen content of oysters infected by Dermocystidium marinum. Tex. A&M Res. Found., Tech. Rep. No. 23, Proj. No. 23, 5 p. Swift, M. L., and M. Ahmed. 1983. A study of glucose, Lowry-positive substances, and triacylglycerol levels in the hemolymph of Crassostrea vir- ginica (Gmelin). J. Shellfish Res. 3:45-50. Thompson, S. N. 1986. Effects of dietary carbohydrate on the nutritional physiology and blood sugar level of Trichoplusia ni para- sitized by the insect parasite, Hyposoter exiguae. Parasit- ology 92:25-30. Thompson, S. N., and B. F. Binder. 1984. Altered carbohydrate levels and gluconeogenic enzyme activity in Trichoplusia ni parasitized by the insect para- site, Hyposoter exiguae. J. Parasitol. 70:644-651. Trider, D. J., AND J. D. Castell. 1980. Influences of neutral lipid on seasonal variation of total lipid in oysters, Crassostrea virginica. Proc. Natl. Shell- fish. Assoc. 70:112-117. Ward, J. E., and C. J. Langdon. 1986. Effects of the ectoparasite Boonea ( = Odostomia) im- pressa (Say) (Gastropoda: Pyramidellidae) on the growth rate, filtration rate, and valve movements of the host Crassostrea virginica (Gmelin). J. Exp. Mar. Biol. Ecol. 99:163-180. Wells, H. 1959. iiotes on Odostomia impressa (Say). Nautilus 72: 140- 144. White, D. C, M. W. Davis, J. S. Nickels, J. D. King, and R. J. Bobbie. 1979. Determination of the sedimentary microbial biomass by extractible lipid phosphate. Oecologia (Berl.) 40:51-62. White, M. E., E. N. Powell, and C. L. Kitting. 1984. The ectoparasitic gastropod Boonea ( = Odastomia) im- pressa: Population ecology and the influence of parasitism on oyster growth rates. Pubbl. Stn. Zool. Napoli I Mar. Ecol. 5:283-299. White, M. E., E. N. Powell, and S. M. Ray. 1988. Effect of parasitism by the pyramidellid gastropod Boonea impressa on the net productivity of oysters (Crassostrea virginica). Estuarine Coastal Shelf Sci. 25: 359-377. 565 FISHERY BULLETIN: VOL. 86, NO. 3 White, M. E., E. N. Powell, S. M. Ray, and E. A. Wilson. 1987. Host-to-host transmission of Perkinsus marinus in oyster {Crassostrea virginica) populations by the ectopara- sitic snail Boonea impressa (Pyramidellidae). J. Shellfish Res. 6:1-5. White, M. E., E. N. Powell, S. M. Ray, E. A. Wilson, and C. E. Zastrow. In press. Metabolic changes induced in oysters {Crassostrea virginica) by the parasitism of Boonea impressa (Pyramidel- lidae:Gastropoda). Comp. Biochem. Physiol. WiCKHAM, D. E. 1986. Epizootic infestations by nemertean brood parasites on commercially important crustaceans. Can. J. Fish. Aquat. Sci. 11:2295-2302. Wilson, E. A., M. E. White, E. N. Powell, and S. M. Ray. In press. Patch formation by the ectoparasitic snail, Boonea impressa, on its oyster host, Crassostrea virginica. Veliger. Wright, D. A., and E. W. Hetzel. 1985. Use of RNA:DNA ratios as an indicator of nutritional stress in the American oyster Crassostrea virginica. Mar. Ecol. Prog. Ser. 25:199-206. YuiLL, T. M. 1987. Diseases as components of mammalian ecosystems: mayhem and subtlety. Can. J. Zool. 65:1061-1066. 566 LIFE TABLES FOR TWO FIELD POPULATIONS OF SOFT-SHELL CLAM, MYA ARENARIA, (MOLLUSCA : PELECYPODA) FROM LONG ISLAND SOUND Diane J. Brousseau' and Jenny A. Baglivo^ ABSTRACT Life tables were constructed for two populations oiMya arenaria from Long Island Sound, USA, based on schedules of age-specific fecundity and mortality determined under natural conditions. Mya arenaria shows a basic conservatism in general life history pattern. In both populations, fecundity increases with increasing female size; sexual maturity is attained at 1 year of age; a single annual breeding season occurs and survivorship curves approximate the type III of Deevey, which is characterized by extremely heavy mortality early in life followed by relatively constant mortality thereafter. Differences in the age-specific parameters for the two populations exist, however; for clams greater than 1 year of age, both age-specific fecundity and survivorship are significantly higher in the Stonington population. These differences in the structure and dynamics of the two populations may be due to environmental heterogeneity. Reduced body size due to slower growth in coarse substrate, as well as the increased maintenance demands resulting from burrowing and valve activity in large-grained sediment, may account for the lower egg production and lower survival rates found in the Westport population. The life history pattern of a species has been defined as the way in which that species partitions the limited resources of time or energy among the three basic biological processes of growth, maintenance, and reproduction. Efforts to generate empirical values for life history parameters, age-specific fecun- dity and survivorship, have only recently allowed the construction of life tables for field populations. Determining life history parameters for a commer- cially important species such as the soft-shell clam, Mya arenaria, is particularly useful since they can be used in theoretical models which are designed to analyze the effect of changing survival and fecun- dity values on the growth rate of the population. Life tables now exist for a number of benthic marine invertebrates: the ha,rna,c\e—Chathamulus stellatus (Connell 1961); the prosobranchs— Dicathais orbita (Phillips and Campbell 1974), Nucella (= Thais) lapillus (Frank 1969), Nodilit- torina tuber culata (Doran 1968), and Conns pen- naceus (Perron 1983); the coelenterates— Mwr-icea califomica (Grigg 1977), M.fruticosa (Grigg 1977), and Balanophyllia elegans (Fadlallah 1983); the bivalves— Mi/a arenaria (Brousseau 1978), Tapes phillipinarum (Yap 1977), and Gemma gemma 'Department of Biology, Fairfield University, Fairfield, CT 06430. ^Department of Mathematics and Computer Science, Fairfield, University, Fairfield, CT 06430; present address: Mathematics Department, Boston College, Chestnut Hill, MA 02167. (Weinberg 1985). These studies, however, examine life history parameters for a single species popula- tion or for successive cohorts within a population. The present study examines age-specific fecundity and survivorship in two geographically separated populations of the soft-shell clam, Mya arenaria, using identical methodology. Since differences in methodology can influence estimates of demo- graphic parameters (Fadlallah 1983) uniformity of approach is necessary for interpopulation com- parisons. This study is the first reported examina- tion of species- specific traits in two naturally occur- ring populations and was carried out as part of a broader study of the population dynamics of this species along the Connecticut shore of Long Island Sound. MATERIALS AND METHODS Field Study Areas Field studies were conducted at two intertidal sites in Long Island Sound, one located at Barn Island in Stonington, CT Oat. 41°20'N; long. 71°53'W) and the other in the Saugatuck River in Westport, CT Gat. 41°06'N; long. 73°23'W)(Fig. 1). The Stonington site is a narrow intertidal sandflat which extends approximately 10 m shoreward to a coarse sand beach. At low tide the Westport site ex- tends 30 m (at its widest point) shoreward to a Spar- Manuscript accepted April 1988. FISHERY BULLETIN: VOL. 86, NO. 3, 1988. 567 FISHERY BULLETIN: VOL. 86, NO. 3 Figure 1.— Map showing locations of the two study sites: Barn Island in Stonington, CT (A) and Saugatuck River in Westport, CT (B). tina altemijlora marsh. The substrate there is coarser than at the Stonington site; 50% by weight is >4 mm grain size diameter, much of that classi- fiable as either cobble or boulder (Brousseau and Baglivo 1987). Both study sites are closed to shellfish harvesting to high levels of bacterial contamination. Although annual salinity and temperature profiles for the two sites are not available, monthly surface and bottom salinity and temperature readings for the period March 1979- June 1981 and October 1985-August 1986 are available for four sites in western Long Island Sound: Norwalk, CT (lat. 41°02.5'N; long. 73°27.2'W), Bridgeport, CT Gat. 41°08.7'N; long 73°11.1'W), Stratford, CT (lat. 41°07.6'N; long. 73°07.6'W), and New Haven, CT (lat. 41°14.4'N; 72°54.2'W)(Tettlebachetal., 1984; Blogoslawski, pers. comm.^). All water sampling was done at spring low tide in approximately 20 ft of water. The annual average range of surface temper- atures for the four sites was 0.5°-24.2°C. The lowest surface temperature, 0.0 °C, was recorded at the New Haven station in February 1980; the high- est was 25.0°C recorded at the Stratford station in August 1986. Annual average range of bottom temperatures was 0.6°-23.4°C. Surface sahnities ranged from 16.9 to 33.3°/oo with a low of 15.97oo reported for Stratford and a high of 35.77oo re- corded at the Norwalk station. Mean bottom salin- ities during that period ranged between 17.0 and 34.4«/oo. Fecundity Oocyte production by female clams collected dur- ing the summer spawning seasons of 1984 and 1985 was estimated using a histological technique (Brous- seau 1976). One hundred twenty-five gravid females from the Westport population and 123 from the Stonington population were examined. Size-specific fecundity rates were converted to age-specific rates using age-size information for M. arenaria obtained from analysis of internal shell growth bands (Brous- seau and Baglivo 1987). Age-specific fecundity estimates for x-yr-old clams, m^., were calculated using the formula: Mx = 0.5 ^ Wj P(size-class i | x yr-old) (1) 'W. Blogoslawski, Northwest Fisheries Center Milford Labora- tory, National Marine Fisheries Service, NOAA, Roger Avenue, Milford, CT 06460, pers. commun. March 1987. where m^ is the mean fecundity for size-class i, P(size-class i \ x-yr-old) is the conditional probabil- ity that an x-yr-old clam is in size-class i, and the sum is taken over 10 mm size-classes i. The condi- tional probabilities P (size-class i \ x-yr-old) were derived empirically (see Appendix Tables 1, 2). Ten millimeter size-classes were used because of sparse- ness in the data. By convention, m^ represents the 568 BROUSSEAU and BAGLIVO: FIELD POPULATIONS OF MYA ARENARIA total number of eggs produced by a female of age X which may be considered to be "female eggs". Because the sex ratio of both populations is 1:1 (Brousseau 1987a), only half of the oocytes produced will eventually become females. Accordingly, each rrijc is one-half the total annual fecundity per female of age X. Mortality To determine size-specific l-yr survival rates of adult clams (>1 year of age), M. arenaria from Stonington and Westport were collected, individual- ly tagged, and returned to their original sites (Ston- ington: 31 May-1 June 1985; Westport (Plot A): 1 May 1984 and Westport (Plot B): 22 May 1984). Clams were measured to the nearest 0.1 mm antero- posteriorly and marked for identification using a method described previously (Brousseau 1979). Tagged clams were replanted in parallel furrows ex- cavated to a depth of 20-30 cm. All plots were 0.3 m apart and located at midtide level (-I-3.0 m). At the end of the test period, the clams were collected. All numbered individuals, both alive and dead, were returned to the laboratory for measurement. A total of 1,049 clams from the Stonington population were tagged, of which 78% were recovered. One-year recovery rate at the Westport site was 40% of the 1,977 clams initially planted. Mortality rates were determined on the basis of the number of clams recovered. This method gives a more accurate estimate of mortality since it does not consider clams that were not recaptured in these estimates. If, on recapture, a dead clam showed no evidence of growth, either death had occurred naturally in a slow-growing individual or premature death had occurred as a result of trauma due to the marking procedure. In order to correct for premature mor- tality caused by handling, the proportion of live clams which showed an increase in shell length over the year was calculated (Table 1). For the Westport population all live clams <40 mm shell length grew, while 94.7% of live clams over 40 mm grew. For the Stonington population all live clams <50 mm shell length had grown during the year, while 91.9% of Table 1 . — Distribution of Mya arenaria recovered alive showing growth and no growth during the test period. Westport (clams >40 mm) Stonington (clams >50 mm) live clams over 50 mm grew. In order to estimate size-specific survivorship then, the following rules were applied: Westport: Classes <40 mm: P(surviving one year) (2) = NJiN^ + Nog) Classes >40 mm: F(surviving one year) = NJ{N^ + Nog/OMI) Stonington: Classes <50 mm: ^(surviving one year) (3) = NJiN^ + Nog) Classes >50 mm: P(surviving one year) = NJiN^ + Ar^G/0.919) where A^^ is the number alive and Nog is the num- ber dead with growth. In the smaller classes (West- port, <40 mm; Stonington, <50 mm), the sum A^^ + Nog represents all recovered individuals, while in the larger classes, the adjustment factor allows us to add a number of dead without growth back into the total recovered in the same proportion as the live with no growth. Age-specific survival rates, P_j, were based on the same age-size information used to determine age- specific fecundity rates. These rates were calculated using the formula: Px = -t.pj P(size-class i \ a;-yr-old) (4) Number Percent Growth No growth Growth No growth 196 11 440 39 94.7 5.3 91.9 8.1 where p^ is the probability that a clam in size-class i survives one year, P(size-class i \ x-yr-old) is the conditional probability that an a;-yr-old clam is in size-class i, and the sum is taken over 10 mm size classes i. The probability of surviving from settlement to one year (a period of about 10 months) is calculated from the estimates of the density of the population in 2 consecutive years. Both populations were sam- pled during periods of maximum settlement in 1985. This period occurred approximately 3 weeks earlier in the Westport population. At each sampling, 18 to 35 samples (0.1 m^ x 30 cm deep) were taken along transects running from mean low water shore- ward to the mean high water mark. Samples were wet-seived in the laboratory (2 mm mesh) and the shell lengths of the clams retained by the seive were measured. Clams of size 2-20 mm in the first year of sampling are assumed to be the spat; those of size 20-50 mm the following year are assumed to have come proportionately from the settlement of the previous year in the same ratios as the empirical probability distributions (Appendix Tables 1, 2). 569 FISHERY BULLETIN: VOL. 86, NO. 3 RESULTS Fecundity Fecundity of Mya arenaria increased with in- creasing female body size (Fig. 2). Size-specific fecundity schedules, however, differed at the two study sites. Comparison of regression lines for log(fecundity) vs. shell length by analysis of covari- ance indicated that the relationships were signifi- cantly different {P = 0.004). Size-specific oocyte pro- duction in the Stonington clams was larger than in clams from Westport. The smallest gravid females observed were 34 and 27 mm in shell length in the Stonington and Westport populations, respective- ly. Age at first reproduction is 1 year in clams from both sites. Sex ratios in both populations did not differ significantly from 1:1 (Brousseau 1987a). Recruitment and Mortality Both temporal and spatial variations in annual recruitment were observed during the study period. A substantial settlement occurred at the Westport site in 1985, as evidenced by the high densities of 2-20 mm clams present in late July of that year (Table 2). In contrast, spat densities at the Stoning- ton site were low in early August of 1985, when settlement occurred. Nonetheless, persistence of the 1985 year class at Westport was poor, whereas ap- proximately one-half of those set in the Stonington population during that year were alive one year later. The coarse substrate at the Westport site may have aided attachment of the byssal stage juveniles, resulting in the higher recruitment rate during the summer of 1985. Direct estimates of M. arenaria survivorship during the postsettlement to 1-yr 1 7 30 35 46 52 Stn 57 63 68 - + + - 79 16 D O 16 LLI « « e « « • « * « « a * « • * * • * • * * « * * « « 2 * 3* »••««•« »«*«*«« « «* ' * 22 * • • 2 • * • 2 •2** O ,5 o 15 13 27 33 38 - + + - 49 55 60 66 82 SHELL LENGTH (mm) Figure 2.— Log of fecundity versus shell length for Mya arenaria from 570 BROUSSEAU and BAGLIVO: FIELD POPULATIONS OF AfTA ARENARIA Table 2.— Densities per 0.1 m^ of Mya arenaria in size-classes 2-50 mm in the Westporl and 2-60 mm in the Stonington popula- tions in 1985 and 1986. Location Size-class and date Samples (mm) /V/O.l m^ (range) Westport, Connecticut 23 July 1985 28 2-20 23.36 (0-84) 09 June 1986 18 10-20 0.11 (0-1) 20-30 0.44 (0-5) 30-40 0.89 (0-4) 40-50 0.50 (0-3) Stonington, Connecticut 15 Aug. 1985 35 2-20 0.37 (0-7) 24 June 1986 24 20-30 0.04 (0-1) 30-40 0.13 (0-1) 40-50 0.33 (0-2) 50-60 0.75 (0-3) period (10 months) obtained by dividing the number of 1-yr-olds alive in the summer 1986 by the esti- mated cohort size in the previous year were 0.0283 and 0.5869 for the Westport and Stonington popu- lations, respectively. Size-specific survival rates of adult clams {>1 year) at both sites are shown in Tables 3 and 4. These represent empirical estimates from mark and recap- ture studies with survival adjustments in the larger classes made according to Equations (2) and (3). Statistical comparison of the survival distributions of adult clams, however, was limited to those re- covered live or recovered showing some growth. At the Westport site survival was constant across the 10 mm size classes: 20-29 through 50-59 and 60 -i- . At Stonington, however, the size-specific survival for classes 30-39 through 70-79 and 80 -i- increased slightly with size (P < 0.05, x^ test). The probability 18 30 35 Wp 1 46 52 57 63 68 74 79 17 16 O 16 LU U- O 15 O « • » « * « « 4 • * * • * * * * V ■« • • • * * • * •3 ;• 2* *»2* *♦ **? * * 2 • * * « « * ««* 22*** • • 2 • * * * 2 * * • •*223* • 15 13 27 33 38 44 49 55 6C 66 7 1 82 SHELL LENGTH (mm) Figure 2.— Continued— the Stonington (STN) and Westport (WPl) populations. 571 FISHERY BULLETIN: VOL. 86, NO. 3 Table 3.— Size-specific survival rates for marked Mya arenaria from the Stonington population. P = probability of an individual from size-class / sur- viving one year; N^ = number of live clams; Nqq = number of dead clams sfiowing growth. Size-class Size (mm) Na Nog P 2 20-30 3 30-40 23 3 0.8846 4 40-50 106 18 0.8548 5 50-60 134 7 0.9462 6 60-70 161 11 0.9308 7 70-80 159 9 0.9420 8 eo+ 25 0 1 .0000 Table 4.— Size-specific survival rates for marked Mya arenaria from the Westport population. P = probability of an individual from size-class / surviv- ing one year; Na = number of live clams; N, DG - number of dead clams showing growth. Size-class Size (mm) N^ N. DG 2 20-30 15 5 0.7500 3 30-40 87 56 0.6084 4 40-50 150 76 0.6515 5 50-60 48 19 0.7052 6 60-70 9 5 0.6303 7 70-80 0 0 0.0000 8 80-1- of surviving one year for adult clams at Stonington (30 -1- mm in shell length) is significantly higher than at Westport (20+ mm in shell length) (P < 0.001, X^ test). Age-specific survivorship estimates (Table 5) were calculated by assuming that 20-30 mm clams at Stonington survivied with the same probability as 30-40 mm clams and that 90-100 mm clams at Ston- ington survived with the same probability as 80-90 mm clams and application of Equation (4) using the empirical distributions in Appendix Tables 1 and 2. Table 5. — Age-specific survival rates for marked Mya arenaria from the Westport and Stonington, CT populations. P^ = probability of surviving from age-class x to age-class x + ^. Age-class (yr) Westport 1 2 3 4 5 6 7 8 9 0.6390 0.6442 0.6625 0.6680 0.6595 0.6422 0.6128 0.5991 0.5021 Stonington P. 0.8843 0.9307 0.9390 0.9550 0.9661 0.9746 0.9746 0.9746 0.9746 Population Dynamics The age-specific survivorship and fecundity sched- ules for M. arenaria are combined in Hfe table form in Tables 6 and 7. Since rates of larval survivorship are difficult to measure for species with planktonic larvae, simplifying assumptions are necessary for estimating survivorship probabilities of larvae (age 0). In order to complete the life tables the equilib- rium settlement rates (rj were calculated using the method of Brousseau et al. (1982). Our choice of r^ implies that the net reproductive rate, Rq, which is defined by Rq = ^l^m^, were Ij = sur- vivorship and m_c = fecundity, equals one for both populations. This is not meant to imply however, that the populations studied here are considered to be in equilibrium. Rather, it is used simply as a theoretical construct in which to examine possible consequences of differing fecundity and mortality schedules on the two populations of M. arenaria. Table 6.— Life table for the Westport, CT Mya arenaria population, assuming equilibrium conditions, i^ = survivorship to beginning of age interval x, or /^_, x Pj, . , when X > 1 ; m^ = fecundity during age interval x. Age (yr) 'x m^ 'x^x 0 1.0 1 ^1.0599 X 10-' 1485615 0.1575 2 6.7728 X 10"^ 2989512 0.2025 3 4.3630 X io-« 3867846 0.1688 4 2.8905 X io-« 5192053 0.1501 5 1.9309 X 10-8 5892286 0.1138 6 1.2734 X 10-8 6464925 0.0823 7 8.1780 X lo-'^ 6478050 0.0530 8 5.0115 X 10-*^ 6842903 0.0343 9 3.0024 X io-« 8390977 0.0252 10 1.5075 X 10-*^ 8390977 0.0126 1 /, = r, X (probability of surviving from 2 months to 1 year). Table 7.— Life table for the Stonington, CT Mya are- naria population, assuming equilibrium conditions. 1^ = survivorship to beginning of age interval x, or /,_, X P^_, when x > 1; m^ = fecundity during age inter- val X. Age (yr) 'x m. 'x^x 0 1.0 1 '2.014 X 10-8 1109952 0.0224 2 1.7811 X 10-8 3712746 0.0661 3 1.6577 X 10-8 4947569 0.0820 4 1.5566 X 10-8 7125027 0.1109 5 1.4866 X 10-8 8287563 0.1232 6 1.4362 X 10-8 8724158 0.1253 7 1.3997 X 10-8 8724158 0.1221 8 1.3642 X 10-8 8724158 0.1190 9 1.3296 X 10-8 8724158 0.1160 10 1.2958 X 10-8 8724158 0.1130 ' /, = r^ X (probability of surviving from 2 months to 1 year). 572 BROUSSEAU and BAGLIVO: FIELD POPULATIONS OF MYA ARENARIA The equilibrium settlement rate (r,) for the West- port population is 3.7453 x 10'^ and that for the Stonington population is 3.4318 x 10"^. The equi- librium first year survival rates (i.e., the probability of surviving from egg to year 1 necessary to main- tain a population at equilibrium) then, become 1.0599 X 10-^ and 2.014 x lO'^ for the Westport and Stonington populations, respectively. Under the assumption of equal rates of larval im- port and export, it is possible to calculate empirical settlement rates by estimating the total number of eggs released per unit area by spawning females in 1985 and dividing this value into the densities of 2-20 mm spat settled per unit area 2 months later. Size-frequency data for the populations was used to determine the number of clams in each 10 mm size class (Fig. 3). The number of animals in each class (assuming that one-half the population is female) was multiplied by the mean size-specific annual fecun- dity value for that size class. Summing over all size classes of reproducing females gives an estimate of the 1985 oocyte production per unit area. Empirical settlement rates of 8.4589 x 10"*^ and 7.1953 x 10"^ were calculated for the Westport and Stoning- ton populations. These represent values of 2.2585 and 2.0967 times the estimated equilibrium settle- ment rate of these populations. Multiplying the em- pirical settlement rate by the probability of surviv- ing the remainder of the year give actual first year actual first year survival rates of 2.3939 x 10"^ and 4.2229 x 10"^ for Westport and Stonington, respectively. The heavy mortality evident by the low juvenile survivorship rates {l{) for M. arenaria take into ac- count the losses incurred during fertilization, meta- morphosis and recruitment, and subsequent survival through the first year of life. Table 8 compares the first year survival rates of representative marine invertebrate species with planktonic and non- planktonic modes of development. Early survival in species lacking planktonic larvae is on average, three orders of magnitude higher than that of species which pass through a planktonic larval stage. DISCUSSION General comparisons of life history traits are use- ful but quantitative comparisons are possible only from the more detailed information found in life tables (age-specific fecundity and survivorship). The difficulty in generating such information, especial- ly for marine bivalves, many of which have plank- tonic larval stages during their life cycles, has resulted in the construction of few complete life tables. Moreover, no reported field study has ex- amined life history traits for more than one popu- lation of a species simultaneously. Consequently, the extent to which quantitative differences in life history parameters are characteristic of the life history of a single species is unknown. Table 8.— Empirical estimates of first year survival rates of marine invertebrate species with and without planktonic larvae (adapted from Perron 1983). Class Species Survivorship Developmental mode References Anthozoa Muricea californica 2.6 X 10-^ planktonic Grigg (1977) Muricea fruticosa 3.7 X io-« planktonic Grigg (1977) Balanophyllia elegans 5.0 X 10-2 nonplanktonic Fadlallah (1983) Bivalvia Gemma gemma Cohorts 1978-1981 2.3 1.2 2.0 1.1 X X X X 10-2 10-^ 10-^ 10-^ nonplanktonic Weinberg (1985) Mya arenaria Westport population 2.4 = 10-^ planktonic Present study Stonington population 4.2 X 10-8 Tapes phillipinarum 1.8 X io-« planktonic Yap (1977) Gastropoda Thais lamellosa ^9.0 X 10-3 nonplanktonic Spight (1975) Conus pennaceus 7.0 X 10-" nonplanktonic Perron (1983) Lacuna pallidula 2.0 X 10-* nonplanktonic Smith (1973) Aplysia Juliana 28.4 X 10-^ planktonic Sarver(1979) Crustacea Balanus glandula 6.0 X 10-5 planktonic Connell (1970) Barnacles (3 spp.) 2.0 X 10-5 planktonic Mines (1979) 'Does not include prehatching mortality. ^Survival through 30-d planktonic period only. 573 FISHERY BULLETIN: VOL. 86, NO. 3 WP1 1985 7.7- a.3 - 2.2- I I i-^ I -I 1 k * 1 1 1- 2 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 SHELL LENGTH (MM) STN 1985 11.0 B.a-- 7.7-- in 5.5-- 9.S- 2.2- I II i i a a ^ a a ^ ll I i i ^ i . -t — ♦ — —I 2 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 SHELL LENGTH (MM) Figure 3.— Size-frequency distributions of Mya arenaria from the Stonington (STN) 574 BROUSSEAU and BAGLIVO: FIELD POPULATIONS OF MYA ARENARIA WP1 1986 11.0 9.9-- a.s 7.7- 3.3 - 2.2 - 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 SHELL LENGTH (MM) STN 1986 11.0 a.s 7.7 - CO 3.3 - 2.2- ii ll ^ a^ »ii » i ^ II ^ 1 1 I ^ t^ i U I Im 2 5 1 0 1 5 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 SHELL LENGTH (MM) Figure Z— Continued— a.nd. Westport (WPl) populations in 1985 and 1986. 575 FISHERY BULLETIN: VOL. 86, NO. 3 The two populations of M. arenaria from Long Island Sound appear to show a basic conservatism in general life history pattern. In both populations, fecundity increases rapidly in young females with individuals reaching sexual maturity in time to reproduce at the beginning of their second year of life. Similarly, larval and adult survivorship sched- ules follow the type III survivorship curve of Deevey (1947). Extremely heavy mortality early in life is followed by roughly constant mortality thereafter. This pattern is similar to that described for a popu- lation of M. arenaria from Gloucester, MA (Brous- seau 1978) except that in the latter, age of first reproduction occurs at the end of the second year. It is interesting to note, however, that some of the major life history features of M. arenaria show con- siderable latitudinal variation within the species. Frequencies of spawnings during the year increases (for review see Ropes and Stickney 1965; Brousseau 1987a) and length of life and body size show a ten- dency to decrease with decreasing latitude (Belding 1930; Newcombe 1935). Details of the ways in which such variations affect the life history parameters, however, remain to be studied. The possibility of gene flow between populations of animals with planktonic larval stages always exists. Nevertheless, the amount of genetic overlap is effectively reduced as the geographical distance between the populations increases. The significant quantitative differences in the age-specific demo- graphic parameters for the two populations studied here suggest that v^ithin the framework of a general life history strategy, a response to the biotic and abiotic components of the immediate environment is possible. Evidence from this study indicates that environmental conditions at the Westport site may be less optimal for the growth and maintenance of M. arenaria than are those at the Stonington site. The higher equilibrium settlement rate calculated for the Westport population indicates that on aver- age, a larger annual spatfall is needed to maintain that population than is required at the Stonington site. The biotic factors most often cited as agents capa- ble of altering the survival and fecundity of in- dividuals are predation, competition, disease, and parasitism. It is unlikely that predation was a major source of adult mortality at either site for reasons previously discussed. Moreover, the effects of preda- tion would be limited to the small, surface-dwelling clams, since 1) crabs, fish, and birds are unable to capture deep burrowing adults and 2) it has been demonstrated that M. arenaria exhibits a "size refuge" from naticid predators (Edwards and Hueb- ner 1977). Hancock (1973) has suggested that com- petition for food or space between the spat and adults may contribute to lowered survival in newly settled clams. If this were the case, one would ex- pect lower juvenile survivorship rates in the Stoning- ton population where population densities were greater (Fig. 3). Trematode infestations have been demonstrated to cause castration and high mortalities in a popula- tion of venerid clams, Transenella tantilla, in California (Obrebski 1968). Although both trematode infections (Stunkard 1938; Uzmann 1951) and fungal parasites have been reported in M. arenaria (Andrews 1954), no evidence of either was observed in the 2,826 histologically prepared clams (1,583 from Stonington; 1,243 from Westport) examined in a study of the reproductive cycle in the two popu- lations (Brousseau 1987a). Sarcomatous neoplasia, a proliferative disorder characterized by increased number of "leukemia-like" cells in the tissues and organs (Farley 1969; Brown et al. 1977; Cooper et al. 1982), has been identified in samples of clams from both Stonington and Westport. Prevalence of the disorder in the Stonington population ranges from 0 to 46%, whereas 0-69% of the M. arenaria from Westport were neoplastic, depending on the collection date (Brousseau 1987b). At present it is not known if neoplasia is a source of mortality in field populations of clams. It seems reasonable to assume that it is, however, since neoplastic cells are invasive and at times cause the destruction of organs and tissues in infected animals (Yevich and Barszcz 1977). If sarcomatous neoplasms prove to be malig- nant, this disease could be responsible to some degree for the higher mortality rates reported in the Westport population. The abiotic factors with the greatest effect on the biotic potential of estuarine organisms are temper- ature, salinity, substrate, and food availability. Adult M. arenaria typically inhabit the intertidal zone and are adapted to a vvide range of fluctuations in water temperature and salinity. In addition, sediments tend to buffer temperature and salinity fluctuations (Sanders et al. 1965; Johnson 1965, 1967). There- fore, infaunal organisms like the soft-shell clam are subject to less extreme environmental fluctuations than are exposed organisms living on or attached to the surface. Belding (1930) noted that M. arenaria withstands extreme variation in salinity, being able to adjust to changing tides every six hours. Shaw and Hamons (1974) found that lethal conditions for burrowed clams were met only when temperatures persisted in the high 20 °C range and salinities were 2°/oo or lower. In the laboratory. 576 BROUSSEAU and BAGLIVO: FIELD POPULATIONS OF MYA ARENARIA adult survival is not altered by salinities of 2.5°/oo (Chanley 1957; Pfitzenmeyer and Drobeck 1963; Castagna and Chanley 1973) to 35"/oo (Castagna and Chanley 1973). Annual water temperature and salinity patterns in Long Island Sound fall well within the range of conditions tolerated by M. arenaria (see Materials and Methods). If tempera- ture and salinity differences did occur, their effects would be minimal. Substrate differences at the Stonington and West- port sites appear to be the most immediate cause of the observed differences in age-specific fecundity and survivorship. Several investigators have found that sediment type is important in controlling growth rate and shell allometry in M. arenaria (Belding 1930; Swan 1952; Newell and Hidu 1982). Clams grown on coarse sediments (gravel, cobble), such as that at the Westport site are slow-growing and more globose in shape than clams from sand or mud environments. Although differences in growth rate could be due to differences in food availability at the two sites, the allometric variations (Brous- seau and Baglivo 1987) strongly support the hypoth- esis that substrate effect is the factor controlling growth. Reduced body size in Westport clams may have an indirect effect on fecundity by restricting egg production. In addition, both Glude (1954) and Pfitzenmeyer and Drobeck (1967) demonstrated that M. arenaria burrows fastest in fine-grained sedi- ments since physical resistance to burrowing in- creases with increasing particle size (Trueman 1954). Hence, the high daily maintenance requirements of Westport clams may result in less energy available for reproduction, long-term maintenance, and sur- vival. The Westport population may be an example of a population inhabiting a marginal environment. The underlying pattern of life history (reproduc- tive effort, patterns of recruitment, survival profile, and growth schedule) for both populations of M. arenaria is very similar. This is not surprising since population parameters are viewed as evolved, species-specific traits. Nevertheless, our evidence for local differentiation in different habitats demon- strates the degree to which individual populations of a widespread species can respond in different ways to their immediate environments; environmen- tal heterogeneity can be reflected in the structures and dynamics of local populations. ACKNOWLEDGMENTS We wish to thank K. Schellinkhout, J. Smeriglio, J. Trautman, and J. Wachter for technical assist- ance in the field. Working space and running sea- water raceways for marking and holding animals were made available at the National Marine Fish- eries Laboratory, Milford, CT (J. E. Hanks and A. Calabrese, Directors). Temperature and salinity data for Long Island Sound was provided by W. Blogos- lawski. Financial support for this study was provided under grant number NA82AA-D-00018 of the Con- necticut Sea Grant Program. LITERATURE CITED Andrews, J. D. 1954. Notes on fungal parasites of bivalve molluscs in Chesa- peake Bay. Proc. Natl. Shellfish. Assoc. 45:157-163. Belding, D. L. 1930. The soft-shelled clam fishery of Massachusetts. Dep. Conserv., Div. Fish. Game, Commonw. Mass. Mar. Fish. Ser. 1, 65 p. Brousseau, D. J. 1976. Spawning cycle, fecundity and recruitment in a popula- tion of Mya arenaria (soft-shell clam) from Cape Ann, Massachusetts. Fish. Bull., U.S. 76:155-166. 1978. Population dynamics of the soft-shell clam, Mya arenaria. Mar. Biol. (Berl.) 50:63-71. 1979. Analysis of growth rate in Mya arenaria using von Bertalanffy equation. Mar. Biol. (Berl.) 51:221- 227. 1987a. A comparative study of the reproductive cycle of the soft-shell clam, Mya arenaria in Long Island Sound. J. Shellfish Res. 6:7-15. 1987b. Seasonal aspects of sarcomatous neoplasia in Mya arenaria (Soft-shell clam) from Long Island Sound. J. Invertebr. Pathol. 50:269-276. Brousseau, D. J., and J. Baglivo. 1987. A comparative study of age and growth in Mya arenaria (soft-shell clam) from three populations in Long Island Sound. J. Shellfish Res. 6:17-24. Brousseau, D. J., J. Baglivo, and G. E. Lang. 1982. Estimation of equilibrium settlement rates for benthic marine invertebrates: its application to Mya arenaria (Mollusca:Pelecypoda). Fish. Bull., U.S. 80:642-644. Brown, R. S., R. E. Wolke, S. B. Saila, and C. W. Brown. 1977. Prevalence of neoplasia in 10 New England populations of the soft-shell clam (Mya arenaria). Ann. N. Y. Acad. Sci. 298:522-534. Castagna, M., and p. Chanley. 1973. Salinity tolerance of some marine bivalves from estuarine environments in Virginia waters on the western Mid-Atlantic coast. Malacologia 12:47-96. Chanley, P. E. 1957. Survival of some bivalves in water of low salinity. Proc. Natl. Shellfish. Assoc. 48:52-65. CONNELL, J. H. 1961. The effects of competition, predation by Thais lapillus and other factors on natural populations of the barnacle Balanus balanoides. Ecol. Monogr. 31:61-104. 1970. A predator-prey system in the marine intertidal region I. Balanus glandula and several predatory species of Thais. Ecol. Monogr. 40:49-78. Cooper, K. R., R. S. Brown, and P. W. Chang. 1982. The course and mortality of a hematopoietic neoplasm in the soft-shell clam, Mya arenaria. J. Invertebr. Pathol. 39:149-157. 577 FISHERY BULLETIN: VOL. 86. NO. 3 Deevey, E. S. 1947. Life tables for natural populations of animals. Q. Rev. Biol. 22:283-314. DORAN, L. D. 1968. Hermit crabs and gastropod survivorship. Can. J. Zool. 46:719-722. Edwards, D. C, and J. D. Heubner. 1977. Feeding and growth rates of Polinices duplicatns prey- ing on Mya arenaria at Barnstable Harbor, Massachusetts. Ecology 58:1218-1236. Fadlallah, Y. H. 1983. Population dynamics and life history of a solitary coral, Balanophyllia elegans from Central California. Oecologia 58:200-207. Farley, C. A. 1969. Sarcomatoid proliferative disease in a wild population of blue mussel (Mytilus edulis). J. Natl. Cancer Inst. 43:509-516. Frank, P. W. 1969. Growth rates and longevity of some gastropod molluscs on the coral reef at Heron Island. Oecologia (Berl.) 2, p. 232-250. Glude, J. B. 1954. Survival of soft-shell clams, Mya arenaria buried at various depths. Maine Dep. Sea Shore Fish. Res. Bull. 22, 260 p. Grigg, R. W. 1977. Population dynamics of two gorgonia corals. Ecology 58:278-290. Hancock, D. A. 1973. The relationship between stock and recruitment in ex- ploited invertebrates. Rapp. P. -v. R6un. Cons. Perm. int. Explor. Mer 164:113-131. HiNES, A. H. 1979. The comparative reproductive ecology of 3 species of intertidal barnacles. In S. E. Stancyk (editor), Reproduc- tive ecology of marine invertebrates, p. 213-234. Belle W. Baruch Library in Marine Science, Vol. 9. Univ. South Carolina Press, Columbia, SC. Johnson, R. G. 1965. Temperature variation in the infaunal organisms of a sandflat. Limnol. Oceanogr. 10:114-120. 1967. Salinity of intertidal water in a sandy beach. Limnol. Oceanogr. 12:1-7. Newcombe, C. L. 1935. Growth of Mya arenaria in the Bay of Fundy region. Can. J. Res. 13:97-137. Newell, C. R., and H. Hidu. 1982. The effects of sediment type on growth rate and shell allometry in the soft-shell clam Mya arenaria. L. J. Exp. Mar. Biol. Ecol. 65:285-195. Obrebski, S. 1968. On the population ecology of two intertidal in- vertebrates and the paleoecological significance of size- frequency distributions of living and dead shells of the bivalve Transenella tantilla. Ph.D. Thesis, Univ. Chicago, 102 p. Perron, F. E. 1983. Growth, fecundity and mortality of Cornis pennaceus in Hawaii. Ecology 64:53-62. Phillips, B. F., and N. A. Campbell. 1974. Mortality and longevity in the whelk, Dicathais orbita (Gmelin). Aust. J. Mar. Freshwater Res. 25:25-33. Pfitzenmeyer, H. T., and K. G. Drobeck. 1963. Benthic survey for populations of soft-shelled clams, Mya arenaria in the lower Potomac River, Maryland. Ches- apeake Sci. 4:67-74. 1967. Some factors influencing reburrowing activity of soft- shell clam Mya arenaria. Chesapeake Sci. 8:193-199. Ropes, J. W., and A. P. Stickney. 1965. The reproductive cycle of Mya arenaria in New England. Biol. Bull. (Woods Hole) 128:315-327. Sanders, H. L., P. C. Mangelsdorf, Jr., and G. R. Hampson. 1965. Salinity and faunal distribution in the Pocasset River, Massachusetts. Limnol. Oceanogr. 10 (Suppl.), p. R216- R229. Sarver, D. J. 1979. Recruitment and juvenile survival in the sea hare ylpii/sra JM^iana (Gastropoda:Opisthobranchia). Mar. Biol. (Berl.) 54:353-361. Shaw, W. N., and F. Hamons. 1974. The present status of the soft-shell clam in Maryland. Proc. Natl. Shellfish. Assoc. 64:38-44. Smith, D. A. S. 1973. The population biology of Lacuna pallidula (DaCosta) and Lacuna vincta (Montagu) in Northeast England. J. Mar. Biol. Assoc. U.K. 53:493-520. Spight, T. M. 1974. Sizes of populations of marine snail. Ecology 55:712- 729. Stunkard, H. W. 1938. The morphology and life cycle of the trematode Hinms- thla quissentensis (Miller and Northrop, 1926). Biol. Bull. (Woods Hole) 75:145-164. Swan, E. F. 1952. The growth of the clam, Mya arenaria, as affected by the substratum. Ecology 33:530-534. Tettlebach, S., L. Potti, and W. Blogoslawski. 1984. Survey of Vibrio associated with a New Haven shell- fish bed, emphasizing recovery of larval oyster pathogens. In R. Colwell (editor). Vibrios in the environment. John Wiley and Sons, Inc., N.Y. Trueman, E. R. 1954. Observations on mechanisms of the opening of the valves of a burrowing lamellibranch, Mya arenaria. J. Exp. Biol. 31:291-305. UZMANN, J. R. 1951. Cercaria myae sp. nov., a fork-tailed larva from the marine bivalve, Mya arenaria. J. Parasitol. 38(2):161-164. Weinberg, J. R. 1985. Factors regulating population dynamics of the marine bivalve Gemma gemma: intraspecific competition and salin- ity. Mar. Biol. (Berl.) 86:173-182. Yap, W. G. 1977. Population biology of the Japanese little-neck clam. Tapes phillipinaT^m in Kaneohe Bay, Oahu, Hawaiian Islands. Pac. Sci. 31:223-244. Yevich, p. p., and C. a. Barszcz. 1977. Neoplasia in soft-shell clams (Mya arenaria) collected from oil-impacted sites. Ann. N.Y. Acad. Sci. 298:409-426. 578 BROUSSEAU and BAGLIVO: FIELD POPULATIONS OF MYA ARENARIA APPENDIX Appendix Table 1.— Empirical distributions for age-size relationships for the Westporl, CT Mya arenaria population. Percentage in each size category per age are indicated in parentheses. Size (mm) Age (yr) 1 2 3 4 5 6 7 8 9 + 10- -20 1 (3.8) 20- -30 9 (34.6) 3 (2.0) 30- -40 13 (50.0) 55 (37.2) 19 (12.7) 2 (2.9) 40- -50 3 (11.5) 70 (47.3) 65 (43.3) 15 (21.4) 7 (15.6) 4 (8.7) 3 (8.3) 3 (7.9) 50- -60 19 (12.8) 60 (40.0) 40 (57.1) 24 (53.3) 23 (50.0) 16 (44.4) 17 (44.7) 6 (23.1) 60- -70 1 (0.7) 5 (3.3) 12 (17.1) 13 (28.9) 17 (37.0) 14 (38.9) 14 (36.8) 14 (53.8) 70- -80 1 (1.4) 1 (2.2) 2 (4.3) 2 (5.6) 4 (10.5) 6 (23.1) 80- -90 1 (2.8) Appendix Table 2.— Empirical distributions for age-size relation- ships in the Stonington, CT Mya arenaria population. Percentage in each size category per age are indicated in parentheses. Size (mm) Age (yr) 1 2 3 4 5 6 + 20-30 15 (26.8) 30-40 22 (39.3) 4 (5.1) 40-50 13 (23.2) 5 (6.4) 2 (1.1) 50-60 6 (10.7) 31 (39.7) 37 (20.2) 5 (3.8) 1 (7.1) 60-70 31 (39.7) 94 (51.4) 26 (19.8) 2 (7.1) 1 (7.1) 70-80 7 (9.0) 41 (22.4) 66 (50.4) 14 (50.0) 4 (28.6) 80-90 9 (4.9) 31 (23.7) 11 (39.3) 6 (42.9) 90-100 3 (2.3) 1 (3.6) 2 (14.3) 579 A LONG-TERM STUDY OF "MICROCELL" DISEASE IN OYSTERS WITH A DESCRIPTION OF A NEW GENUS, MIKROCYTOS (G. N.), AND TWO NEW SPECIES, MIKROCYTOS MACKINI (SR N.) AND MIKROCYTOS ROUGHLEYI (SR N.) C. Austin Farley,' Peter H. Wolf,^ and Ralph A. Elston^ ABSTRACT Continuing long-term studies of oyster disease problems have been carried out over the past 26 years using field monitoring, gross, histologic, and ultrastructural pathologic methods. A microorganism of uncertain taxonomy was discovered in 1963 by J. G. Mackin in association with lesions and mortalities of Japanese oysters, Crassostrea gigas, from Denman Island, British Columbia, Canada. Mackin coined the term "microcell" for this organism and described the parasite as 1-3 fim cells with small nuclei which occurred within vesicular connective tissue cells adjacent to characteristic abscesses. We are describing this organism as Mikrocytos mackini sp. n. in his honor. Similar appearing organisms were seen by the senior author in flat oysters, Ostrea edulis, from Milford, Connecticut, on three different occasions: 1) in oysters transferred from Milford, Connecticut, to Chincoteague Bay, Virginia; 2) in oysters transferred from Milford to Elkhom Slough, California; and 3) in oysters trans- ferred from Milford to Oxford, Maryland, and held in recirculated sea water. The causative organism in these three episodes has been shown by electron microscopy to be Bonamia ostreae, the parasite that was implicated in recent mortalities in flat oysters in Europe. Similar organisms have also been seen in Olympia oysters, Ostrea lurida, from Oregon and in the Sydney rock oyster, Saccostrea commercialis, from Australia. Presence of the organism in the latter species is associated with the winter mortalities originally described by T. C. Roughley, and the pathogen is here described as Mikrocytos roughleyi (sp. n.) in his honor. "Microcell" type parasites of oysters are associated with a complex of diseases that occur in Japanese oyster, Crassostrea gigas; Sydney rock oyster, Sac- costrea commercialis; flat oyster, Ostrea edulis; and Olympia oyster, 0. lurida, in North America, Europe, and Australia. Severity of disease varies from an acute, highly lethal form to a chronic, seasonally recurring disease that does not produce massive mortalities. The etiologic agents are small, morphologically simple, and very difficult to com- pare and characterize taxonomically at light micro- scope levels of resolution. Associated lesions vary according to species affected and provide some of the differences that may be used to distinguish the agents involved. The complexity of this group and the difficulties involved in achieving an understand- ing regarding whether we are dealing with one or a group of organisms and how they were transferred to new locations, the long time span involved in 'Northeast Fisheries Center Oxford Laboratory, National Marine Fisheries Service, NOAA, Oxford, MD 21654. ^62 MacKenzie Street, Bondi Junction, New South Wales, Australia 2022. 'Center for Marine Disease Control, Battelle/Marine Research Laboratory, Sequim, WA 93282. answering these questions, and the continuing dis- semination of unpublished privileged information shared in informal workshop gatherings of scientists with common interests, make it necessary to use un- published anecdotal information in order to provide as complete a story as possible. The first oyster mortality known to be associated with "microcell" disease was reported in C. gigas from Denman Island, British Columbia, Canada by Quayle (1961). Quayle's report documents the epi- zootic aspects from 1956 to 1960 and demonstrates the gross appearance of the disease in the Pacific oyster. A causative agent was not identified until several years later, when the late J. G. Mackin"* (unpubl. data) discovered a small intracellular organism intimately associated histologically with tissue abscesses in diseased oysters (C gigas) from Denman Island and called this organism "micro- cell". He demonstrated this material at the 1963 Shellfish Mortality Conference held at Oxford, MD. Mackin's demonstration provided us with the in- sight to identify similar organisms in histologic sec- ^J. G. Mackin, deceased, Texas A&M University, College Station, TX. Manuscript accepted May 1988. FISHERY BULLETIN: VOL. 86, NO. 3, 1988. 581 FISHf:RY BULLETIN: VOL. 86, NO. 3 tions of Denman Island oysters in our laboratory collection and further led to the discovery of micro- cell disease in 0. edulis, 0. lurida, and 5. commer- cialis from Australia ("Australian winter disease") described by Roughley (1926). In 1979, the char- acteristics of microcell disease were demonstrated at a microscopic diagnosis workshop held at the Ministry of Agriculture and Fisheries Laboratory in Weymouth, England. This workshop was attended by molluscan pathologists from several European countries including France. In the late 1970s, serious mortalities of 0. edulis in France, associated with microcell infections, were described in a paper by Pichot et al. (1979), in which the organism was named Bonamia ostreae. Reference to the earlier work on microcell by Mackin and others (Katkansky et al. 1969) was not included in their report. Since the taxonomic relationship and status of these similar parasites have not been described, it is the purpose of this paper to present a complete background on microcell related epizootics and mor- phological information which, in North America, preceded the French report (Pichot et al. 1979). Fur- thermore, microscopic and ultrastructural compari- sons of the microorganisms are provided and a new genus, Mikrocytos g. n., and two new species, Mikro- cytos mackini sp. n. and Mikrocytos roughleyi sp. n., are described. MATERIALS AND METHODS General Procedures Oyster tissues were collected from a variety of sources as follows: Code WWC were C. gigas collected from Henry Bay, Denman Island, British Columbia, Canada on a periodic basis from April 1968 to June 1969 by D. B. Quayle (spring 1969 samples were collected by N. Bourne). Live oysters were sent by air freight to the Oxford Laboratory where clinical and gross features were recorded and they were processed for histological and, in some cases, ultrastructural studies. Code S-124-A were C. gigas from Hawaii collected in September 1972. Code S-41 were S. commercialis collected by Peter Wolf from 22 July to 23 July 1965 from the Georges River, Woolooware Bay, New South Wales and ship- ped to the Oxford Laboratory for processing. Code FK were progeny of 0. edulis from Boothbay Harbor, ME that were spawned at Milford, CT in April 1961. Seed oysters were transplanted to Chincoteague Bay, VA in May 1961 and processed at subsequent intervals (FK-1-1 to FK-2-5, August 1961; FK-3-1, August 1961; FK-4-1 and FK-4-2, February 1962). Code WAC were 0. edulis bred in the Milford, CT hatchery from 1963 to 1965, and introduced into California bays as follows: WAC-1, Milford 1963 seed oysters planted in Morro Bay, CA in December 1964 and sampled on 7 December 1965 during a heavy mortality. Sample WAC-21-28 consisted of 38 oysters from a Milford 1963 stock shipped to California in 1964 and held at Pigeon Point Laboratory, Pigeon Point, CA until heavy mortality occurred and sampled on 11 May 1966. WAC-3-1-10 were 1963 Milford oysters placed in Morro Bay in 1964. Heavy mortality was noted and samples were taken. WAC-3-11-15 were 1963 Milford stock placed in Morro Bay in 1964. Mortality was low. Oysters were necropsied and fixed on 1 May 1966. WAC-3-16-26 were Milford 1962 stock placed in Morro Bay in 1963. They experienced 40% mortality and were examined and fixed on 1 May 1966. Oysters WAC-4-1 through 4-5 were from Milford 1963 stock placed in Tomales Bay, CA in 1964; low mortality was observed. They were ex- amined and fixed on 1 May 1966. Code FMT were 0. edulis used in an experimental holding study at the Oxford Laboratory. Ten 2-yr- old 0. edulis from the Milford hatchery were placed in each of three tanks receiving 0.45 /um membrane filtered 26°/oo seawater on 23 February 1968; pH, temperature, salinity, and mortality were monitored daily until 10 October 1968. Crassostrea virginica from the Mispillian River, Delaware Bay, were placed in each tank on 28 March. Tanks were desig- nated A (control), B (fed tissues of moribund 0. edulis from Pigeon Point, CA), and C (fed tissues from Denman Island C. gigas infected with Denman Island disease). Seawater in each tank was recir- culated through a glass wool, charcoal, calcium flow- through filter via an airlift system moving from C to B to A, respectively. Oyster codes were FMT-A- 1-20, FMT-B-1-20, FMT-C-1-20, numbered as they were fixed (species were designated at time of fixation). Code WAO-A were 50 specimen samples of native oysters (0. lurida) from Yaquina Bay, OR, sampled 582 FARLEY ET AL.: MICROCELL DISEASE IN OYSTERS monthly from February 1969 to January 1971. Samples were coded from WA0-7A to WAO-22A. Tissues were fixed in Zenker's acetic, Davidson's fluid (Shaw and Battle 1957), or McDowell's fixative (McDowell and Trump 1976) as modified by Farley et al. (1986) (1% glutaraldehyde/4% formaldehyde; pH 7.2-7.4 in one-half ambient seawater). Six /im sections were stained with Harris' hematoxylin- eosin, Ziehl's fuchsin, periodic acid Schiff reagent (PAS) with malt diastase digestion, Feulgen picro- methyl blue (Farley 1969), or Giemsa (Howard and Smith 1983). Electron Microscopy Procedures Lesions from C. gigas from Denman Island were fixed in 2% glutaraldehyde in pH 7.2 seawater, post- fixed in 1% osmium tetroxide in phosphate buffer, pH 7.2, and embedded in Epon-Araldite (Feng et al. 1971). Fifty to 100 nm sections, selected on the basis of interference color (silver), were cut and stained with lead citrate and uranyl acetate. Ostrea edulis sections from the WAG material (WAC-2-19) were deparaffinized, postfixed in 2% glutaraldehyde and 1% osmium tetroxide, plastic embedded, ultrasec- tioned, and stained in the same fashion as the C. gigas materials. Sections were examined in a Zeiss^ EM 9 electron microscope. RESULTS condition index remained high (most oysters in medium to fat condition). Mantle recession occurred most commonly from April through June and was most prevalent in June. Pale digestive gland was present in up to 24% of the oysters in the spring and was also seen in up to 16% of the oysters in fall samples. Shell pustules (Fig. 1), abscesses, and ulcers as described by Quayle (1961) (Fig. 2) were present from April through June. Microscopical examination (Table 2) revealed granular hemocyte infiltration of vesicular connec- tive tissue (VCT) in most samples with high prev- alences of microcell infections occurring sporadically throughout the years. Abscesses (Fig. 3) in the VCT of the mantle and gonad consisted of apparently viable granular hemocytes at the periphery with phagocytosis of moribund cells deeper in the lesion and coagulative necrosis in the center. Microcell organisms were associated with these abscesses. The parasite (Fig. 4) was 1-3 p1 ^m that was spherical with bipolar or eccentric nucleolar structures. The size and cytozoic location of these organisms suggest a strong similarity to other microcell type parasites seen in other species of oysters. Four of the six females had gonads in a state of resorption and digestive diverticular epithe- lium was slightly metaplastic in two of the seven oysters. Microcell Disease Study in Ostrea edulis Episode 1 Microcell disease in 0. edulis from progeny from Boothbay Harbor, ME brood stock spawned at Milford, CT and transferred to Chincoteague Bay, MD (Code FK). Two of the 13 oysters fixed between August 1961 and July 1962 had heavy infection of microcells (Fig. 13) which were 1-3 ^m in diameter with a Feulgen- positive nucleus (1 yim. diameter) and found intra- cellularly within hemocytes. In three of the other oysters, moderate infiltrations of hemocytes were observed but without parasites. 585 FISHERY BULLETIN: VOL. 86, NO. 3 r If* 1 * ♦ t f ^ j^ % * ♦ V* •J, •'• » 8 • •♦v. ."» .» jnl.. >'i'-. 1, 100 j: # ^* idir * JiiJf ft i . • «^. tl* ♦ • \'^' m 9 « 10 10 10 Figure 7.— Ulceration in the mantle of Saccostrea commercialis with Australian winter disease. Figure 8.— Large abscess-type lesion in the connective tissue-gonad region of Saccostrea commercialis (code S-41-7) from Australia with Australian winter disease. 100 x . HHE stain (specimen S-41-7). Figure 9. -Higher magnification photomicrograph from lesion in Figure 8. Note intracellular microcell parasites (Mikrocytos roughleyi sp. n.) with hemocytes. 1,000 x. HHE stain (specimen S-41-7). Figure 10.— Heavy infection of Australian winter disease; microcell parasites in the gonad of S-41-4. 1,000 x . HHE stain. 586 FARLEY ET AL.: MICROCELL DISEASE IN OYSTERS Episode 2 Microcell disease in 0. edulis transferred from Milford, CT to Elkhorn Slough, CA (Code WAC). Epizootiological gross and histopathologic data are presented in Table 3. Clear association is ap- parent between high mortality, emaciation (watery condition), mantle recession, hemocytic infiltration, and heavy microcell infection. Infections were particularly intense in gill and GI tract epithelia which contained dense infiltrations of infected hemocytes (Fig. 11). Infections were characterized by the intracellular (in hemocytes) and extracellular presence of "microcell" organisms (Fig. 12) similar to those seen in the Chincoteague Bay (FK) samples. Episode 3 Microcell disease in 0. edulis transferred from Milford, CT to aquaria in Oxford, MD (Code FMT). The experiment ran from 20 February 1968 until 9 October 1968. No transmission was noted. How- ever, tank C drained accidentally on 19 March and the filtered seawater was replaced with unfiltered seawater from Chincoteague Bay, VA. A combined total of 9 of 11 animals died between 22 March and 27 March and the remaining 2 live oysters were fixed on 28 March. The first six that died were from tank C and had moderate to heavy cases of microcell disease with associated hemocytic infiltration. Histologic exam- ination of both species, either after mortalities, or when they were sacrificed in October, failed to reveal any more infections. The parasites and disease characteristics were identical to the pre- viously described cases (Fig. 14). Microcell Disease Study in Ostrea lurida Microcell disease in 0. lurida from Yaquina Bay, OR (Codes WA0-7-A and WA0-12-A). Monthly to bimonthly samples of 50 specimens were examined grossly and for histopathology for a 2-yr period. A 24% prevalence of microcell disease was seen in WA0-7-A samples collected in February 1969, and 12% prevalence was seen in the WAO- 12-A samples collected in February 1970. No micro- cell infections were seen in any of the other samples. However, neoplasms and Mytilicola orientalis in- fections were seen commonly throughout the study. Microcell infections tended to be less intense than in 0. edulis, but organisms were similar in size and appearance to the organisms seen in flat oysters (Fig. 15). Intracellular infections were in VCT cells and not hemocytes. Ultrastructural Studies Electron microscopy revealed 0. edulis parasites that were usually intracellular with up to three organisms per cell. Parasites were 1-3 f^m in diam- eter, and contained an eukaryotic nucleus about 1 /im in diameter. A crescent-shaped, peripheral nucleolus was evident in most organisms (Fig. 16). The cytoplasm contained numerous ribosome-like organelles and spherical dense bodies, presumably membrane bound, which were 90-130 nm in diam- eter, and generally resembled the "haplosporo- somes" (Fig. 17) described in other haplosporidan Table 3. — Epizootiology of microcell disease in O. edulis from California. Stock origin Sample location Date fixed Mortality N Condition Mantle reces- sion Hemocyte infil- tration Microcell infection Code Fat Med Watery WAC-1 (1-20) [Gilford 1964 Morro Bay 7 Dec. 1965 Heavy 20 no gross data 18 18 WAC-2 (1-25) Milford 1964 Pigeon Pt. 1 May 1966 Heavy 38 1 8 19 22 25 25 WAC-3 (1-10) Milford 1964 Morro Bay 1 May 1966 Heavy 10 2 6 2 6 8 8 WAC-3 (11-15) Milford 1963 Morro Bay 1 May 1966 Low 5 4 1 0 0 0 0 WAC-3 (16-26) Milford 1963 Morro Bay 1 May 1966 40% 10 6 4 0 1 6 3 WAC-3 (27-31) Milford 1965 Morro Bay 1 May 1966 0 5 5 0 0 0 0 0 WAC-4 (1-5) Milford 1964 Tomales Bay 1 May 1966 Low 5 5 0 0 1 0 0 587 FISHERY BULLETIN: VOL. 86, NO. 3 Figure 1 1.— Low magnification view of Ostrea edulis from California (WAC-2-29). Note diffuse inflammatory infiltrate in vesicular connective tissue. 100 x. HHE stain. Figure 12.— Microcell infection in Ostrea edulis (specimen WAC-2-19) from population of oysters introduced into Califor- nia waters from Milford, CT. 1,000 X. HHE stain. Figure 13.— Microcell infection in hemocytes in Ostrea edulis transported from Milford, CT to Chincoteague Bay, VA in 1962 (first diagnosed case oi Bonamia ostreae). 1,000 x. HHE stain (specimen FK-5-1). Figure 14.— Microcell infection in Ostrea edulis. Transferred from Milford, CT to Oxford, MD recirculated seawater aquarium. Cells of the vesicular connective tissue are infected. 100 x . HHE stain (specimen FMT-B-1-3). Figure 15.— Microcell infection in vesicular connective tissue and hemocytes of Ostrea lurida from Yaquina Bay, OR. 1,000 X. FPM stain (specimen WAO-7A-41). 588 FARLEY ET AL.: MICROCELL DISEASE IN OYSTERS parasites (Perkins 1979). Larger "membrane" bound structures (500 nm) consistent with the ap- pearance of mitochondria were also present in the cytoplasm; however, cristi could not be distin- guished. Ultrastructural studies were also performed on positively diagnosed oysters (C. gigas) from Denman Island, British Columbia (sample WWC-18, collected 9 June 1980). Microcells were always associated with focal abscesses, but their occurrence was restricted to the periphery. Parasites were found as cytozoic organisms primarily in VCT cells (Fig. 18). Micro- cells were 3-4 p