^^i-^'°'Co. Fishery Bulletin r Vol. 85, No. 1 Marine Biological Laboratory LIBRARY APR 2 7 1987 Woods Hole, Mass. January 1987 WAPLES, ROBIN S., and RICHARD H. ROSENBLATT. Patterns of larval drift in southern California marine shore fishes inferred from allozyme data 1 UTTER, FRED, DAVID TEEL, GEORGE MILNER, and DONALD McISAAC. Genetic estimates of stock compositions of 1983 chinook salmon, Oncorhynchits tshawytscha, harvests off the Washington coast and the Columbia River 13 BRILL, RICHARD W. On the standard metabolic rates of tropical tunas, including the effect of body size and acute temperature change 25 LUTZ, PETER L., and ANN DUNBAR-COOPER. Variations in the blood chemistry of the loggerhead sea turtle, Caretta caretta 37 VERMEER, GREGORY K. Effects of air exposure on desiccation rate, hemolymph chemistry, and escape behavior of the spiny lobster, Panulinis argus 45 WILSON, K A, K L. HECK, Jr., and K W. ABLE. Juvenile blue crab, CaUinectes sapidus, survival: an evaluation of eelgrass, Zostera marina, as refuge 53 MINELLO, THOMAS J., ROGER J. ZIMMERMAN, and EDUARDO X. MARTINEZ. Fish predation on juvenile brown shrimp, Penaetcs aztecus Ives: effects of turbidity and substratum on predation rates 59 GROVER, JILL J., and BORI L. OLLA. Effects of an El Nino event on the food habits of larval sablefish, Anoplopoma fimbria, off Oregon and Washington 71 GARTNER, JOHN V, Jr., THOMAS L. HOPKINS, RONALD C. BAIRD, and DEAN M. MILLIKEN. The lantemfishes (Pisces: Myctophidae) of the eastern Gulf of Mexico 81 LOVE, MILTON S., BRITA AXELL, PAMELA MORRIS, ROBSON COLLINS, and ANDREW BROOKS. Life history and fishery of the California scorpionfish, Scorpaena guttata, within the Southern California Bight 99 HORWOOD, J. W Bias and variance in Allen's recruitment rate method 117 CLARKE, THOMAS A. Fecundity and spawning frequency of the Hawaiian anchovy or nehu, Encrasicholina purpurea 127 Notes SILVERT, W, and D. PAULY. On the compatibility of a new expression for gross conver- sion efficiency with the von Bertalanffy growth equation 139 McCABE, GEORGE T, Jr., ROBERT L. EMMETT, TRAVIS C. COLEY, and ROBERT J. McCONNELL. Effect of a river-dominated estuary on the prevalence of Carcino- nemertes errans, an egg predator of the Dungeness crab, Cancer rrwigister 140 (Continued on back cover) / / Seattle, Washington U.S. DEPARTMENT OF COMMERCE Malcolm Baldrige, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Anthony J. Calio, Administrator NATIONAL MARINE FISHERIES SERVICE William E. Evans, Assistant Administrator 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 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. Bruce B. Collette 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 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. njar Fishery BulletinPRnm/ Marine Biological Laborat LIBRARY CONTENTS Woods Hole, Mass. Vol. 85, No. 1 January 1987 WAPLES, ROBIN S., and RICHARD H. ROSENBLATT. Patterns of larval drift in southern California marine shore fishes inferred from allozyme data 1 UTTER, FRED, DAVID TEEL, GEORGE MILNER, and DONALD McISAAC. Genetic estimates of stock compositions of 1983 chinook salmon, Oncorhynchus tshawytscha, harvests off the Washington coast and the Columbia River 13 BRILL, RICHARD W. On the standard metabolic rates of tropical tunas, including the effect of body size and acute temperature change 25 LUTZ, PETER L., and ANN DUNBAR-COOPER. Variations in the blood chemistry of the loggerhead sea turtle, Caretta caretta 37 VERMEER, GREGORY K. Effects of air exposure on desiccation rate, hemolymph chemistry, and escape behavior of the spiny lobster, Panulints argus 45 WILSON, K. A, K L. HECK, Jr., and K W ABLE. Juvenile blue crab, CaUinedes sajndus, survival: an evaluation of eelgrass, Zostera marina, as refuge 53 MINELLO, THOMAS J., ROGER J. ZIMMERMAN, and EDUARDO X. MARTINEZ. Fish predation on juvenile brown shrimp, Penaeus azteeus Ives: effects of turbidity and substratum on predation rates 59 GROVER, JILL J., and BORI L. OLLA. Effects of an El Nino event on the food habits of larval sablefish, Anoplopoma fimbria, off Oregon and Washington 71 GARTNER, JOHN V, Jr., THOMAS L. HOPKINS, RONALD C. BAIRD, and DEAN M. MILLIKEN. The lantemfishes (Pisces: Myctophidae) of the eastern Gulf of Mexico 81 LOVE, MILTON S., BRITA AXELL, PAMELA MORRIS, ROBSON COLLINS, and ANDREW BROOKS. Life history and fishery of the California scorpionfish, Scorpaena guttata, within the Southern California Bight 99 HORWOOD, J. W. Bias and variance in Allen's recruitment rate method 117 CLARKE, THOMAS A. Fecundity and spawning frequency of the Hawaiian anchovy or nehu, Encrasicholina purpurea 127 Notes SILVERT, W, and D. PAULY. On the compatibility of a new expression for gross conver- sion efficiency with the von Bertalanffy growth equation 139 McCABE, GEORGE T, Jr., ROBERT L. EMMETT, TRAVIS C. COLEY, and ROBERT J. McCONNELL. Effect of a river-dominated estuary on the prevalence of Carcino- nemertes errans, an egg predator of the Dungeness crab, Cancer magister 140 {Continued on next page) Seattle, Washington 1987 For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington DC 20402— Subscription price per year: $21.00 domestic and $26.25 foreign. Cost per single issue: $6.50 domestic and $8.15 foreign. Contents— Contrnwed KYNARD, BOYD, and JOHN P. WARNER. Spring and summer movements of subadult striped bass, Morone saxatilis, in the Connecticut River 143 WALSH, S. J. Habitat partitioning by size in witch flounder, Glyptocephaltts cynoglossus: a reevaluation with additional data and adjustments for gear selectivity 147 MATHEWS, S. B., and M. LaRIVIERE. Movement of tagged lingcod, Ophiodon elongatiis, in the Pacific Northwest 153 VONDRACEK, BRUCE. Digestion rates and gastric evacuation times in relation to temperature of the Sacramento squawfish, Ptychocheilus grandis 159 HANLON, ROGER T., PHILIP E. TURK, PHILLIP G. LEE, and WON TACK YANG. Laboratory rearing of the squid Loligo pealei to the juvenile stage: growth comparisons with fishery data 163 CHAPMAN, ROBERT W. Changes in the population structure of male striped bass, Morone saxatilis, spawning in the three areas of the Chesapeake Bay from 1984 to 1986 . . . 167 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. ^v ffiuje AWARDS ^V ffiUJe The Publications Advisory Committee of tlie National Marine Fisheries sen/ice has announced the best publications au- thored by the NMFS scientists and pub- lished in the Fishery Bulletin and the Marine Fisheries Review for 1984 and 1985. Only effective and interpretive ar- ticles which significantly contribute to the understanding and l3,000 m) water. It is remote enough (275 km west of the central Baja California coast) that genetic differentiation of shore fishes might be expected. Collections in the vicinity of Pun- ta Eugenia were made to serve as controls for evaluating the extent of differentiation at Guada- lupe and to estimate the relative importance of east- west larval drift in this area. Well-developed oceanic currents serve as poten- tial transport mechanisms for pelagic larvae in the study area. The California Current brings relative- ly cold, low salinity water from high latitudes toward the Equator; its principal characteristics have been known for some time (Reid et al. 1958; Hickey 1979). The California Current is most strongly developed north of Point Conception; further south, nearshore flow becomes somewhat variable because of the eastward jut of the coastline and the complicating effects of the Channel Islands (Fig. 1). Between about lat. 30° and 33 °N, the current shifts toward the east, and a portion of the water is deflected WAPLES and ROSENBLATT; LARVAL DRIFT IN SOUTHERN CALIFORNIA 125 35' 30" 25' 125° 120° 115° 110° Figure 1.— Schematic representation of mean flow patterns in the study area, based on data from Wyllie (1966) and Hickey (1979) and modified from Cowen (1985). Consistent flow directions are shown with soHd arrows; dashed arrows indicate more variable features. Study sites are also indicated: La Jolla (L); San Nicolas (N) and Catalina (C), California Channel Islands; Isla de Guadalupe (G); Islas de San Benito (B); Cabo Thurloe (T). northward along the Southern California Bight, forming the Southern California Eddy (Schwartzlose 1963). This eddy can be found throughout the year except during periods of peak southward flow (gen- erally January to May). The 10 shore fish species used in the analysis (Table 1) were generally those that could be collected in adequate numbers during brief visits to remote localities. However, attempts were made to include species with widely varying life history strategies and, hence, different dispersal capabilities. The life history and larval capture data summarized in Table 1 were taken from personal observations, unpub- lished data from the California Cooperative Fish- eries Investigations (CalCOFI) and the Ichthyo- plankton Coastal and Harbor Studies (ICHS), and from the literature; see Waples (1986) for discus- sion and references. Sample sizes of about N = 50 individuals per species were collected at each of the four areas [ranges of mean_sample sizes: for species, iV = 36 (bjacksmith) to AT = 63 (sheephead); for localities, A/" = 46 (Punta Eugenia) toN = 55 (Gua- dalupe)]. Electrophoresis and Data Analysis Whole fish or tissue samples were frozen in the field, transported to Scripps Institution of Ocean- ography, and stored at -25°C to -35°C. Proce- dures of horizontal starch gel electrophoresis and FISHERY BULLETIN: VOL. 85, NO. 1 Table 1 .—Summary of life history information for the 10 shore fish species used in the analysis. Batch Length of Larval Dispersal Family/species Common name fecundity larval life catches capability Embiotocidae Embiotoca jacksoni black perch 10 none (viviparous) " nil Cottidae Clinocottus analis wooly sculpin 10^-10^ few weeks? only near rocky shores low Clinidae Alloclinus holder! island kelpfish 103 brief? inshore? limited Gobiidae Lythrypnus dalli bluebanded goby 10^-10^ two or more months inshore moderate Malacanthidae Caulolatllus princeps ocean whitefish 10^ few months? inshore/offshore high Pomacentridae Chromis punctipinnis blacksmith 10^ few months? inshore/offshore high Kyphosidae G/>e//a nigricans opaleye 10^ few months? mostly inshore high Serranidae Paralabrax clathratus kelp bass 10^ few months? mostly inshore high Labridae Semicossyphus pulcher sheephead 10^ 2-4 months Inshore/offshore high Kyphosidae Medialuna californiensis halfmoon 10^ few months offshore very high histochemical staining have been described else- where (Waples 1986). The 26 enzymes and proteins surveyed were acid phosphatase, aconitate hydra- tase, adenosine deaminase, adenylate kinase, alcho- hol dehydrogenase, aspartate aminotransferase, creatine kinase, esterase (a-naphthyl acetate), fumarate hydratase, glucose-6-phosphate dehydro- genase, glucosephosphate isomerase, glutamate dehydrogenase, glyceraldehyde-phosphate dehydro- genase, glycerol-3-phosphate dehydrogenase, L-iditol dehydrogenase, isocitrate dehydrogenase, lactate dehydrogenase, malate dehydrogenase, man- nosephosphate isomerase, phosphoglucomutase, phosphogluconate dehydrogenase, peptidase (leucyl- tyrosine; leucylglycyl-glycine), superoxide dismu- tase, umbelliferyl esterase, xanthine dehydrogenase, and general muscle proteins. Presumptive gene loci for which any variant alleles were detected were surveyed in all individuals. Loci for which only a single allele had been identified after sampling at least 20 individuals in each population were con- sidered to be monomorphic and were not surveyed further in that species. Wright's Fgx was computed for each polymorphic locus in each species by the method of Weir and Cockerham (1984). Fgr values (0 < ^5,^ < 1) in- dicate the proportion of total variance in allele fre- quencies attributable to differences between (as opposed to within) populations. Workman and Nis- wander's (1970) test was used to identify Fg^ values significantly larger than zero. Data for all presumptive gene loci resolved in each species were combined in an index of overall genetic differentia- tion (Nei's [1972] genetic distance (Z))), which pro- vides a direct means of comparing levels of genetic divergence between pairs of populations. D is the negative natural logarithm of genetic similarity (/), which is essentially the proportion of genes shared by two populations. To determine whether similar patterns of popula- tion structure occur in several species, D values for each pair of localities (or the mean D values for each locality) were ranked within each species. The resulting matrix of rankings was evaluated for re- curring patterns (departure from randomness) by Friedman's method for randomized blocks (Sokal and Rohlf 1981), which computes a statistic that is a chi-square variate with 6-1 degrees of freedom: 6 o 4_,^ = [il2l{ab[b +1])) I (Ii2y)2] - 3a(fe + 1) (1) where a = number of rows (species, in this case), b = number of columns (localities, or pairs of local- ities), and Rij is the ranking of the i^^ locality (or pair of localities) for the j*^ species. To identify species that exhibit anomalous pat- terns of genetic differentiation, a jackknife pro- cedure was used, the rankings of localities (or pairs of localities) for each species being compared with the overall ranking computed for all the other species combined. Spearman's rank-order correla- tion coefficient (rj was used to determine the WAPLES and ROSENBLATT: LARVAL DRIFT IN SOUTHERN CALIFORNIA strength of the agreement (or disagreement) be- tween these two sets of rankings: 1 - (6 Z rff/[n(w2 - 1)]) (2) where n is the number of items ranked (in this case, 4 locahties or 6 pairs of localities) and d, is the dif- ference in rankings of the i* locality (or pair of localities). RESULTS The electrophoretic analysis provided information regarding variation at 32-42 presumptive gene loci in the 10 species. The genetic interpretation of band- ing patterns was guided by comparisons of observed and expected number of bands exhibited by pre- sumed heterozygotes, by tissue specificity of iso- zyme expression, and by quality and consistency of resolution. A detailed discussion of results for each enzyme can be found in Waples (1986). Except for Semicossyphus pulcher (discussed below), no over- all departures of heterozygote frequencies from those expected under conditions of Hardy- Weinberg equilibrium were found (Waples 1986, in press). Table 2 summarizes the allozyme data. Average heterozygosities for the 10 species (mean H = 0.031; range = 0.009-0.087) are somewhat lower than the mean value of 0.055 reported for over 100 marine fishes by Smith and Fujio (1982), but at least 5 loci {Embiotoca jacksoni) and as many as 19 loci (Ly- thrypnus dalli) were found to be polymorphic in each species. Space does not permit reporting here the allele frequencies for all of these variable loci; these data appear in Waples (1986), or can be obtained from the first author. Interpopulational genetic distance values (Table 2) were generally fairly small: for half of the species {Alloclinus holderi, Chromis punctipinnis, Girella nigricans, Medialuna calif omiensis, Paralabrax clathratus) all possible pairwise comparisons of populations yielded D values <0.001. Even the largest observed D value (0.029 for the Guadalupe- Punta Eugenia comparison in E. jacksoni) is well within the range of values typically found between conspecific populations of fish species (Shaklee et al. 1982; Thorpe 1983). Nevertheless, it is apparent that populations of most of these shore fishes do not behave as a single panmictic unit. For 8 of the 10 species, significantly nonzero single-locus Fgj values indicate heterogeneity of allele frequencies among populations (Table 2; see also Waples 1986). Furthermore, the statistically significant tendency for species that are better dispersers to have lower mean D values (Waples in press) suggests that the relatively small D values reported here for most species contain valid information relating to popula- tion structure. Our interest here is primarily to identify recur- ring patterns (across species) of genetic similarity between areas. One way to approach this topic is to compute, for each species, a mean of all the pair- wise D values involving each locality. In Table 3 these mean D values have been ranked within each species, thus providing an indication of which populations are most similar (or dissimilar) gene- tically to the other populations as a whole. Two species (A. holderi, L. dalli) that could be collected from only three of the four areas have been deleted from this analysis. The two southern populations, Guadalupe and Punta Eugenia (total of rankings for each = 15), are consistently more divergent than are La Jolla (24.5) and San Nicolas (25.5). Substitution of these totals and values for a (8 species) and b (4 localities) into Equation (1) yields a x^ value of 7.54 with 3 df. This Table 2.— Summary of electrophoretic results. Number of loci surveyed (T), number polymor- phic (P), and number with significantly nonzero Fgj values (F) are indicated. H = average heterozygosity; L = La Jolla; C = Channel Islands; E = Punta Eugenia; G = Islade Guadalupe. Nu T mber of loci P F H Genetic distance ( x 10^) Species L-C L-G L-E C-G C-E G-E A. holderi 32 10 1 0.009 0.063 0.042 0.023 Ca. princeps 35 14 1 0.049 0.183 0.146 0.205 0.062 0.144 0.050 Ch. punctipinnis 40 10 1 0.009 0.007 0.021 '0.009 0.023 '0.012 '0.027 CI. analis 36 17 10 0.046 0.158 0.293 0.237 0.269 0.158 0.362 E. jacksoni 40 5 4 0.015 0.665 0.457 1.55 1.45 0.338 2.87 G. nigricans 42 17 1 0.025 0.043 0.022 0.021 0.046 0.073 0.052 L dalli 35 19 3 0.087 0.094 0.218 — 0.171 — — M. californiensis 38 18 — 0.025 0.022 0.019 0.024 0.018 0.037 0.054 P. clathratus 41 12 — 0.012 0.011 0.020 0.015 0.007 0.028 0.032 S. pulcher 38 14 3 0.033 0.009 0.178 '0.098 0.155 '0.070 '0.083 'Mean of comparisons involving Cabo Thurloe and Islas de San Benitos. FISHERY BULLETIN: VOL. 85, NO. 1 Table 3.— Chi-square test of homogeneity of ranking of areas by decreasing mean D values and correlation of ranking of areas in each species with overall ranking of all other species. Statistics computed for 8 species collected at all four areas and for the remaining 7 species after data for Ca. princeps were omitted. L = La Jolla; C = Channel Islands; E = Punta Eugenia; G = Isia de Guadalupe. Rankings by area Correlation {r^) with other species Species L C E G 8 spp 7 spp Ca. princeps Ch. punctipinnis CI. analis E. jacksoni G. nigricans M. califomiensis P. ciathratus S. pulcher 1 4 3 3 4 4 3.5 2 3 3 4 4 1 3 3.5 4 2 2 2 2 2 1 1 3 4 1 1 1 3 2 2 1 -0.80 - 0.60 1.0 0.75 0.80 0.75 0.80 -0.23 0.20 0.60 0.80 0.75 0.75 0 0.40 Totals CSQ Signif. (3 df) level 8 spp 7 spp 24.5 23.5 25.5 22.5 15 13 15 11 7.54 NS 10.59 P < 0.05 value is not quite significant (0.1 > P > 0.05; critical value 7.81). Although the pattern of differentiation over all eight species cannot be shown to depart significantly from randomness by this nonpara- metric test, it is instructive to continue the analysis to see whether anomalous results in one or two species may be obscuring an underlying pattern in the others. Aberrant species can be identified by measuring the correlation (rj of rankings for each species with the overall rankings for all other species combined. To do this, rankings for the localities were computed as each species in turn was deleted from the analysis. These rankings were then compared with those for the species deleted. The r^ values for this analysis clearly indicate a core group of five species (Chromis punctipinnis, Clinocottus analis, E. jacksoni, M. califomiensis, P. ciathratus), rank- ings for each of which are highly correlated with those of all other species (Table 3). At the other ex- treme, rankings of Caulolatilu^ princeps are essen- tially the opposite of those of the other species (r^ = - 0.80). Thus C. princeps is the only species for which La Jolla was ranked the most divergent local- ity, as it is the only species for which Guadalupe is the locality with the highest overall genetic similar- ity to the other populations. In order to evaluate the influence of Caulolatilus princeps on the overall analysis, Friedman's test was repeated after data for C. princeps had been deleted. The resulting chi-square value for seven species (10.59; 3 df) is significant at the 0.05 level. After omitting C. princeps, the correlation (r^) for each species with all other species was again com- puted (Table 3). It is apparent that the remaining species form a more coherent group with C. prin- ceps omitted, values for each species being positively correlated with those from all other species. More detail regarding possible pathways of lar- val drift can be obtained by considering the relative degree of divergence of each pair of populations. D values for the six possible pairwise comparisons of the four study areas have been ranked within each species in Table 4. An analysis similar to the pre- ceding indicates that the two northern populations (La Jolla and the Channel Islands) are consistently the most similar genetically, and the two southern populations (Punta Eugenia and Guadalupe) are the most divergent. There are no consistent differences in rankings of the four other comparisons, each of Table 4.— Chi-square test of homogeneity of ranking of pairs of localities by decreasing mean D values and correlation of ranking in each species with overall ranking of all other species. Statistics computed for 8 species collected at all four localities and for the remaining 7 species after data for Ca. princeps were omitted. Abbreviations as in Table 3. Ranki ings of Correlation (rj pairs of localities with othi 8 spp 3r species Species L-C L-E L-G C-E C-G E-G 7 spp Ca. princeps 2 1 3 4 5 6 -0.77 — Ch. punctipinnis 6 5 3 4 2 1 0.49 0.94 CI. analis 5.5 4 2 5.5 3 1 0.46 0.93 E. jacksoni 4 2 5 6 3 1 0.14 0.26 G. nigricans 4 6 5 1 3 2 -0.16 0.31 M. califomiensis 4 3 5 2 6 1 0.09 0.14 P. ciathratus 5 4 3 2 6 1 0.43 0.54 S. pulcher 6 3 1 5 2 4 -0.03 0.54 CSQ Signif. Totals (5 df ) 7.09 level 8 spp 36.5 28 27 29.5 30 17 NS 7 spp 34.5 27 24 25.5 25 11 11.84 P < 0.05 WAPLES and ROSENBLATT: LARVAL DRIFT IN SOUTHERN CALIFORNIA which involves one northern and one southern population. The same five species identified in the previous analysis (Chromis punctipinnis, Clino- cotttis analis, P. clathratus, E. jacksoni, M. califor- niensis) have the highest correlation with rankings of the other species, although only for the former three is r^ > 0.40. Again, results for Caulolatilus princeps (r^ = -0.77) are strongly negatively cor- related with those of the other species. The chi- square value testing the equality of rankings for pairs of localities (7.09; 5 df) is not statistically significant (critical value = 11.07). In light of the results obtained above, the analysis was repeated after deletion of Caulolatilus princeps. When this was done, the r^ values for each of the other species increased, to as high as 0.93 and 0.94 for Clinocottus analis and Chromis punctipinnis, respectively. The chi-square value (11.84) indicates that for the remaining species the rankings of pairs of localities are significantly heterogeneous. With data for Caulolatilus princeps omitted, it is even more apparent that the Guadalupe-Punta Eugenia comparison is the most divergent, and La Jolla- Channel Islands remains the most similar pair of localities (Table 4). DISCUSSION Two major points emerge from the analysis of pat- terns of genetic similarity between areas. First, large-scale patterns of larval dispersal for most species appear to be affected in a similar way by the local current regime. The recurrent patterns can be summarized as follows: 1) La Jolla and the Chan- nel Islands are the two areas with the greatest (and Punta Eugenia and Guadalupe the two areas with the lowest) overall genetic affinity with other popu- lations; 2) the two northern populations share similar allele frequencies, while the two southern populations have much stronger genetic affinities with the northern populations than with each other. That the southern populations are relatively iso- lated genetically is not surprising, since they are at the periphery of the distributional range for most of the species. However, it was not expected that the Punta Eugenia populations would show nearly the same degree of genetic isolation as do those from Guadalupe, an oceanic island with a substantial endemic component in its marine flora and fauna (Briggs 1974). The nature of genetic differentiation of Guadalupe shore fishes is discussed more fully in Waples (1986). That many marine species with northern affinities are found along the coast of Baja California, Mexico only in localized upwelling areas (Dawson 1945; Hubbs 1960) may be responsible for the observed divergence of Punta Eugenia popula- tions. These upwelling populations, isolated from other shore fish populations by areas with water temperatures up to 10°C warmer, may represent largely independent reproductive units. One aspect of the population structure that seems clear from the results of this study is that the southern popu- lations studied exchange genes much more frequent- ly with northern populations than with each other. Such a finding would be difficult to predict on the basis of known current patterns, which are quite variable and complex off the coast of Baja Califor- nia (Fig. 1). Because the southerly flowing California Current is the dominant hydrological feature in the study area, it is of interest to examine the possibility that the link between northern and southern populations is due primarily to one-way gene flow from the north. This possibility can be evaluated in terms of the presence or absence of rare alleles. If gene flow were unidirectional (north to south), one would ex- pect most alleles present in the northern populations also to appear in samples from the south. Alleles originating in the southern populations, on the other hand, would have no tendency to spread to the north. For the 10 species combined, 50 alleles are found in one or more northern populations but are absent from all southern populations, while only 36 alleles are restricted to southern populations (Waples 1986). These data thus do not provide evidence for gene flow only from the north, as such a model would predict more alleles restricted to southern popula- tions. Furthermore, the average frequency of alleles restricted to the southern populations (0.0098) is slightly higher than the frequency of those restricted to the north (0.0092); this is the opposite of the result expected if unidirectional migration were "swamp- ing" alleles restricted to the south. It is possible that the episodic northward advection of water from the south is an important source of genetic exchange among populations. Such movement is known to occur even in years not associated with El Nino events, and organisms with southern affinities that apparently have been transported into southern and central California are reported on a fairly regular basis (Hubbs and Schultz 1929; Hubbs 1948; Rado- vich 1961; Brinton 1981). The data for restricted alleles are consistent with the hypothesis that such processes may be important in the overall genetic structure of these shore fishes. Two factors, how- ever, argue for caution regarding this conclusion: 1) the pattern of occurrence of restricted alleles is FISHERY BULLETIN; VOL. 85, NO. 1 quite variable among species, and four species have more alleles restricted to southern localities; 2) relatively few restricted alleles are found in these shore fishes, further increasing the already large sampling variation in the number and frequency of restricted alleles (Waples 1986, in press; M. Slatkin^). That the Channel Islands populations are no more genetically isolated than those at La Jolla was some- what unexpected, as La Jolla is part of the major mainland metapopulation that includes much of the distributional range of these shore fishes. It was therefore thought that La Jolla samples would show the greatest overall genetic affinity with other localities. Such a pattern was reported by Haldor- son (1980), who found allele frequencies in the surf- perch Damalichthys vacca to be similar in a series of mainland populations but distinctive at Catalina. Furthermore, Tegner and Butler's (1985) study of drift bottles released at the Channel Islands in- dicated at most 5-10% reach the mainland, sug- gesting that the amount of genetic exchange may likewise be low. However, these findings are not inconsistent with the results of the present study when two factors are considered. First, Tegner and Butler's (1985) study was designed to estimate the numerical im- pact on local green abalone, Haliotis fulgens , popu- lations of larvae derived from the Channel Islands. Because relatively few H. fulgens larvae appear like- ly to cross from the Channel Islands to the main- land, it was concluded that the Channel Islands populations cannot be expected to reseed those on the mainland that are locally depleted through over- fishing, pollution, destruction of habitat, etc. Al- though a small percentage (say 5%) of larval exchange may not exert a significant numerical im- pact on a population, migration at that rate is very high from the perspective of maintaining similar fre- quencies of neutral alleles. In fact, the exchange of only a few breeding individuals per generation is suf- ficient to prevent substantial genetic divergence between populations (Spieth 1974). Second, the Channel Islands populations might well have proved to be relatively more divergent in the present study if additional mainland populations had been included, as was the case in Haldorson's study. Nevertheless, it is noteworthy that Channel Islands populations do not appear to be genetically isolated to any substantial degree. They may thus play a more significant role in the population struc- 'M. Slatkin, Department of Zoology, University of California, Berkeley, CA 94720. ture of marine species in this area than had been believed. The consistently strong affinity between Channel Islands and La Jolla populations suggests that the Southern California Eddy may be effective as a means of larval transport between mainland and island localities. The second major point to emerge from this study is that the population genetic structure of Caulola- tilus princeps is very different from that of any of the other species. In fact, the pattern of genetic af- finity between populations of the ocean whitefish is almost exactly the opposite of the pattern typical of the remaining shore fishes. This result was puz- zling at first, as the life history features of this species are not particularly unusual. However, through the aid of H. Geoffrey Moser (National Marine Fisheries Service, La Jolla, CA), we obtained unpublished larval capture data that shed consider- able light on this problem. Figure 2 is a plot of these data, collected by CalCOFI sampling programs dur- ing 1955-59. In this 5-yr period no C. princeps lar- vae were collected north of central Baja California, Mexico (lat. 30 °N). In this respect, the larval distri- bution of the ocean whitefish is similar to that observed by Kramer and Smith (1973) for the California yellowtail, Seriola dorsalis{= S. lalandi). In contrast, larvae of the other species in this study for which data are available were frequently taken in the Southern California Bight during 1955-59 (percentage of positive collection localities north of lat. 30°N: Chromis punctipinnis, 28%; G. nigricans, 44%; M. calif omiensis, 54%; S. pulcher, 39%; Waples 1986). In a more extensive survey of larval catches, Moser et al. (1986) confirmed the unusual pattern for the ocean whitefish for years 1954-81 (only 4 of 163 larvae taken north of 30°N, and none taken in the Southern California Bight), and sug- gested some possible explanations for the southward shift observed in this species. Thus while the southern populations are near the periphery of the range for most study species, it is the northern populations that are far removed from the apparent sources of ocean whitefish larvae. As we have seen, a significantly nonrandom pat- tern of genetic affinity among areas or pairs of areas was found when data for Caulolatihis princeps were omitted. This result is not entirely unexpected, as removing the most aberrant data in an analysis of this nature will generally result in an improved significance level of the test statistic. On the other hand, such an approach seems justified in this case, as the objectives of this study were to search for generalized patterns of genetic differentiation and to attempt to explain data for anomalous species in 8 WAPLES and ROSENBLATT: LARVAL DRIFT IN SOUTHERN CALIFORNIA 35' 30' 25« T 1 1 r 125° 120° 116° 110° Figure 2.— Location of positive collections of Caulolatilus princeps larvae taken in CalCOFI sam- pling program, 1955-59. terms of life history features. Given the larval cap- ture data discussed above, it is not difficult to under- stand v^hy the inclusion of data for C. princeps tends to obscure patterns of genetic differentiation shared by the other species. Two other species are exceptions (albeit not as dramatic exceptions as C. princeps) to the recurring patterns discussed above. Girella nigricans is the only species for which the Channel Islands was found to be the most genetically divergent locality (Table 3), and S. pulcher is the only species apart from C. princeps for which a strong Punta Eugenia- Guadalupe connection was observed (Table 4). The pattern in S. pulcher is due to loci for which consis- tent heterozygote deficiencies were found (Waples 1986) and thus may provide information that is un- related to actual levels of gene flow. Girella nigri- cans was the only species to be collected primarily as juveniles; these samples largely comprise a single year class, the allele frequencies for which might be prone to short-term variations. Sampling of juve- niles might thus have been expected to yield rela- tively high levels of genetic divergence, but there was no a priori reason to expect the particular pat- tern of D values found in this species. Whether the results for G. nigricans are due to as-yet-undetected processes of larval transport or merely random noise in our analysis is thus unclear at present. We face a similar difficulty in explain- ing the heterogeneity (even among the "core" species) in patterns of genetic affinities between the two northern and two southern populations (Table 4). The decision to include a large number of species in this study mandated a geographically restricted FISHERY BULLETIN: VOL. 85, NO. 1 sampling program, and the resulting analysis pro- vides only a basic outline of these species' popula- tion genetic structure. More extensive sampling would no doubt reveal more variations on the pat- terns identified here. It is likely that such variations would be significantly affected by differences be- tween species in location and timing of spawning. At present, there are neither sufficient inshore hydrographic data nor extensive life history infor- mation over the geographic range of most of these species to allow more specific predictions concern- ing the dynamics of larval drift. Our understanding of the process of larval transport in shallow-water marine organisms can thus be enhanced by more comprehensive sampling programs, involving both genetic and life history analyses. Nevertheless, it is significant that no major differ- ences in patterns of genetic differentiation could be attributed to dispersal capability per se. Thus of the five "core" species with the most strongly correlated sets of rankings, E. jacksoni is a livebearer; Clino- cottus analis spawns intertidally and has a brief lar- val life; P. clathratus and Chromis punctipinnis have high fecundity and a lengthy larval life; and M. californiensis has pelagic juveniles, commonly occurs far offshore with drifting kelp, and thus has the highest dispersal capability of all. This result suggests that the multispecies approach used here may provide information of general use for study- ing the population biology of other marine organ- isms (fishes and invertebrates) with pelagic larvae. ACKNOWLEDGMENTS We would like to thank all those who helped with collections of the nearly 2,000 specimens used in this study. We are also very grateful to H. Geoffrey Moser and Gary Brewer for providing us access to unpublished records of larvae captured by CalCOFI and ICHS, respectively. We thank an anonymous referee for a number of constructive suggestions. Funds for this research were provided in part by grants from the Marine Life Research Group and the Chancellor's Associates, University of Califor- nia, San Diego. Robert S. Waples was supported by an NSF Graduate Fellowship during a portion of this research. LITERATURE CITED Briggs, J. C. 1974. Marine zoogeography. McGraw Hill, N.Y. Brinton, E. 1981 . Euphausid distributions in the California Current dur- ing the warm winter-spring of 1977-78, in the context of a 1949-1966 time series. CalCOFI Rep. 22:135-54. COWEN, R. K. 1985. Large scale pattern of recruitment by the labrid, Semi- cossyphus pulcher: causes and implications. J. Mar. Res. 43:719-742. Dawson, E. Y. 1945. Marine algae associated with upwelling along the north- western coast of Baja California, Mexico. Bull. South. Calif. Acad. Sci. 44:57-71. Haldorson, L. 1980. Genetic isolation of channel islands fish populations: evidence from two embiotocid species. In D. M. Power (editor), The California Islands: Proceedings of a Multi- disciplinary Symposium, p. 433-42. Santa Barbara Mus. Nat. Hist., Santa Barbara, CA. HiCKEY, B. M. 1979. The California Current System - hypotheses and facts. Prog. Oceanogr. 8:191-279. HUBBS, C. L. 1948. Changes in the fish fauna of western North America correlated with changes in ocean temperature. J. Mar. Res. 7:459-82. 1960. The marine invertebrates of the outer coast. Syst. Zool. 9:134-47. HUBBS, C. L., AND L. P. SCHULTZ. 1929. The northward occurrence of southern forms along the Pacific Coast in 1926. Calif. Fish Game 15:234-40. Kramer, D., and P. E. Smith. 1973. Seasonal and geographic characteristics of fishery resources. IX. Inshore sportfishes. Mar. Fish Rev. 35(5-6): 15-18. Moser, H. G., B. Y. Sumida, D. A. Ambrose, E. M. Sandknop, and E. G. Stevens. 1986. Development and distribution of larvae and pelagic juveniles of ocean whitefish, Caulolatilus princeps, in the CalCOFI survey region. CalCOFI Rep. 27:162-169. Nei, M. 1972. Genetic distance between populations. Am. Nat. 106: 283-292. Radovich, J. 1961. Relationships of some marine organisms of the north- east Pacific to water temperatures, particularly during 1957. Calif Dep. Fish Game, Fish Bull. 112, 62 p. Reid, J. L., Jr., G. I. Roden, and J. G. Wyllie. 1958. Studies of the California Current System, /w CalCOFI progress report, 7-1-56 to 1-1-58, p. 27-56. Sacramento, CA. Schwartzlose, R. a. 1963. Nearshore currents of the western United States and Baja California as measured by drift bottles. In CalCOFI progress report, 7-1-60 to 6-3-62, p. 15-22. Sacramento, CA. Shaklee, J. B., C. S. Tamaru, and R. S. Waples. 1982. Speciation and evolution of marine fishes studied by the electrophoretic analysis of proteins. Pac. Sci. 36:141- 157. Smith, P. J., and Y. Fujio. 1982. Genetic variation in marine teleosts: high variability in habitat specialists and low variability in habitat general- ists. Mar. Biol. 69:7-20. SOKAL, R. R., AND F. J. ROHLF. 1981. Biometry. 2d ed. W. H. Freeman, San Franc. Spieth, p. T. 1974. Gene flow and genetic differentiation. Genetics 78: 961-965. Tegner, M. J., AND R. A. Butler. 1985. A drift tube study of the dispersal potential of green 10 WAPLES and ROSENBLATT: LARVAL DRIFT IN SOUTHERN CALIFORNIA abalone (Haliotis fulgens) larvae in the Southern California Bight: implications for recovery of depleted populations. Mar. Ecol. Prog. Ser. 26:73-84. Thorpe, J. P. 1983. Enzyme variation, genetic distance and evolutionary divergence in relation to levels of taxonomic separation. In G. S. Oxford and D. Rollinson (editors), Protein poly- morphism: adaptive and taxonomic significance, p. 131-152. Acad. Press, Lond. Waples, R. S. 1986. A multispecies approach to the analysis of gene flow in marine shore fishes. Ph.D. Thesis, Univ. California, San Diego, 334 p. In press. A multispecies approach to the analysis of gene flow in marine shore fishes. Evolution 41(2). Weir, B. S., and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38:1358-70. Workman, P. L., and J. D. Niswander. 1970. Population studies on southwestern Indian tribes. II. Local genetic differentiation in the Papago. Am. J. Hum. Genet. 22:24-29. Wyllie, J. G. 1966. Geostrophic flow of the California Current at the sur- face and at 200 meters. CalCOFI Atlas No. 4. State of California, Mar. Res. Comm., La Jolla, CA. 11 GENETIC ESTIMATES OF STOCK COMPOSITIONS OF 1983 CHINOOK SALMON, ONCORHYNCHUS TSHAWYTSCHA, HARVESTS OFF THE WASHINGTON COAST AND THE COLUMBIA RIVER Fred Utter,' David Teel,' George Milner,' and Donald McIsaac^ ABSTRACT Allele frequency data for 17 polymorphic protein coding loci from 88 populations of chinook salmon between British Columbia, Canada and California, U.S.A. were used to obtain maximum likelihood estimates of contributing populations to fisheries off the coast of Washington, U.S.A. Data were available for the commercial troll fishery of May 1982 and for commercial, Indian, and sport fisheries during spring and summer 1983. The estimated contributions of fall run fish returning to areas of the lower Columbia River (collectively called "tules") to the May troll fisheries were 76.5% in 1982 and 54.9% in 1983. In contrast, the estimated proportion of fall run fish destined for areas of the upper Columbia River (collec- tively called "upriver brights") was less than 5% in both years, although these runs are known to make substantial contributions to more northern fisheries of Canada and Alaska. A considerable difference for each year occurred in the estimated proportion of California fish (2.8% in 1982 and 18.7% in 1983). Differences occurred among the fisheries and areas sampled in 1983. Larger estimates for Canadian and Puget Sound (Washington) fish occurred in fisheries of northern areas; the largest was 41% for the Indian fishery in the Strait of Juan de Fuca. A greater proportion of California fish in any particular area was taken in sport fisheries. The subset of tule populations returning to the Kalama and Cowlitz river drainages was harvested at a higher rate in sport than commercial fisheries. This study demonstrates the capabilities of the involved procedures for generating timely and reliable estimates of stock composi- tion, and serves as a starting point for more detailed understandings of the oceanic distribution of chinook salmon populations. Chinook salmon, Oncorhynchus tshawytscha, runs returning to Pacific drainages of the western United States are a major biological, recreational, and eco- nomic resource. Their importance persists in spite of the often excessive harvests, disruptions of habi- tats, and blockages of migratory routes that have occurred during the past century. The vitality of these runs continues to fluctuate under the influence of many factors. Conflicting demands of multiple user groups, including recreational, commercial, native American, and international fishing interests, tend to stress the overall resource. Water require- ments for energy, irrigation, and human consump- tion often conflict v^ith even minimal conditions for fish rearing, passage, and reproduction. Instabilities of nature in freshwater and marine environments also contribute substantially to fluctuations in growth, migration, and survival. The management of this resource is further com- 'Northwest and Alaska Fisheries Center Montlake Laboratory, National Marine Fisheries Service, NOAA, 2725 Montlake Boule- vard East, Seattle, WA 98112. ^Columbia River Laboratory, Washington Department of Fish- eries, P.O. Box 999, Battle Ground, WA 98604. plicated by the ecological and genetic diversity of its individual populations. For instance, fish har- vested off the Washington coast represent a com- plex and continually changing mixture of stocks destined for many areas (Fig. 1; see also Miller et al. 1983). Runs returning to the Columbia River illus- trate this diversity; here freshwater entry extends from February through October, and upstream migration distances range from virtually nothing to many hundreds of miles. The largest numbers of Columbia River chinook salmon return in the fall and consist of two distinct types. Fish of that segment of the run commonly called "brights" retain the silver color of ocean- caught salmon for extended periods following their freshwater entry and return primarily to areas above The Dalles Dam. Brights are largely main- tained by natural reproduction with hatchery supple- mentation of some segments. Fish of the largest seg- ment of the fall run are referred to as "tules"; they approach spawning condition rapidly as soon as, and often before, they enter fresh water. Tules return to areas below The Dalles Dam and are perpetuated almost entirely by hatcheries. Although both tules Manuscript accepted October 1986. FISHERY BULLETIN: VOL. 85, NO. 1, 1987. 13 125 OOW FISHERY BULLETIN; VOL. 85, NO. 1 115 OOW 48 OON 43 OON 38 OON - 48 OON 43 OON 38 OON 125 OOW 115 OOW Figure 1.— Areas of baseline and mixed stock sampling. 14 UTTER ET AL.: STOCK COMPOSITION OF 1983 CHINOOK SALMON and brights contribute to oceanic fisheries (Pacific Fishery Management Council 1981^) the persisting prime condition of brights makes them highly favored in river fisheries. An ideal program for harvest management of Chinook salmon would include the capability of iden- tifying the abundance and distribution of distinct breeding groups (such as component stocks of the tule and bright runs of the Columbia River) in a par- ticular fishery. This capability would permit adjust- ments of regulations to permit both protection of weaker stocks and more optimal harvest of abun- dant stocks, depending on their proportions in a fishery. Current information based primarily on data from coded wire tags provides a broad and general overview of hatchery stocks, but lacks details to im- pose differential harvest regulations adequately and does not yield information on wild populations. In addition, a sufficient number of tags must accum- ulate in the fishery or in terminal areas before any quantitative interpretation can be made concerning stock distribution. This requirement coupled with the lag time between field collection and tag de- coding has precluded in-season regulatory adjust- ments based on relative stock strengths. The ability to estimate component stocks in stock mixtures based on genetic profiles of contributing groups has recently been developed and applied (Grant et al. 1980; Fournier et al. 1984; Beacham et al. 1985; Pella and Milner 1987). Numerous esti- mates of stock mixtures of chinook salmon have been made using a genetic stock identification (GSI) procedure described by Milner et al. (1983). These applications (Miller et al. 1983; Milner et al. 1985) have substantially increased the ability to manage stock mixtures of chinook salmon. The genetic procedures provide estimates of stock composition with greater detail and precision than has previously been possible when the following two conditions are met. First, known genetic differences (presently identifiable by electrophoretic methods, among other techniques) must exist among popula- tions contributing to a particular stock mixture. Sec- ond, a data base of calculated genotypic frequencies (based on a sufficient number of genetic systems) must be developed for those populations that are likely to compose a fishery. The GSI procedure obtains maximum likelihood estimates of stock composition using the genotypic 'Pacific Fishery Management Council. 1981. Proposed plan for managing the 1981 salmon fisheries off the coast of Califor- nia, Oregon, and Washington. Pacific Fishery Management Coun- cil, 526 S.W. Mill St., Portland, OR 97201. frequencies of the data base and of the stock mix- ture. The GSI analysis of the May 1982 troll fish- ery off the Washington coast using a data base for California through British Columbia provided the most detailed analysis of an oceanic salmon fishery to date (Miller et al. 1983). This paper follows a general description of the GSI and its application to stock mixtures of salmonids provided in Milner et al. (1985). Estimates of stock composition were obtained from samples collected from fisheries off the Washington coast during the spring and summer of 1983. A particular focus was given to the fall runs of the Columbia River because of the major contributions these runs have histori- cally made to oceanic fisheries. This information is intended to provide managers and biologists with better insights into the life histories of chinook salm- on populations in this area of intermingling, and to initiate a continuing record of this species' oceanic distribution and relative abundance. MATERIALS AND METHODS The procedures used in this study are outlined below. Many of the details required for specific ap- plication are necessarily omitted, but are available in the referenced sources. Baseline Populations Data were obtained from 88 collections taken from British Columbia through California and repre- sented distinct breeding units in most cases (Table 1). Intact juveniles or samples of tissues (eye and liver were the tissues of interest in the present study) from adult fish were taken in the field and transported frozen (usually on dry ice) to the labora- tory for further processing prior to electrophoresis. Methods used for detection of electrophotetic vari- ants followed procedures outlined in Utter et al. (1974) and May et al. (1979). The three buffer sys- tems used included: 1) A Tris-boric acid-EDTA gel and tray buffer, pH 8.5 (Markert and Faulhaber 1965). 2) An amine citric acid gel and tray buffer, pH 6.5 (Clayton and Tretiak 1972). 3) A Tris-citric acid-lithium hydroxide-boric acid gel buffer, and a lithium hydroxide-boric acid tray buffer, pH 8.5 (Ridgway et al. 1970). A system of nomenclature for locus and allelic designations followed Allendorf and Utter (1979). 15 FISHERY BULLETIN: VOL. 85, NO. 1 Table 1.— Geographical area and stock group of the 88 baseline populations used in estimating composition of mixed stock fisheries. Major geographical district, type of Stock Total no. Major geographical district, type of Stock Total no. stock, and baseline population ^'^ group examined stock, and baseline population^ '^ group examined Columbia River Basin Oregon and Washington coas\— Continued Tule (lower river fall run) 1 150 Oregon, fall run— Continued Spring Creek-Little White Coquille Estuary 115 Salmon R.-Washougal R. la — Siuslaw Bay 82 Cow/litz R.-Kalama R. lb 144 Salmon R. 99 Upriver brights (fall run) 2 Nestucca R.-Alsea Bay-Siletz R.- Mid river Fall Ck. 346 Deschutes R. 2a 49 Cedar Ck.* 100 Priest Rapids-Hanford Reach 2b 249 Trask R. -Tillamook R. 188 Snake River Nehalem Estuary 141 Ice Harbor 2c 200 Oregon, spring run Other stocks 3 — Cole R.-Hoot Owl Ck. 163 Lower river, fall run Rock Ck. 100 Lewis R., brights 50 Cedar Ck. 99 Lower river, spring run Trask R. 100 Cowlitz R.,-Kalama R. 100 Washington, fall run Lewis R. 50 Naselle R. 99 Willamette R., spring run Humptulips R. (early) 50 Eagle Creek-McKenzie R. 88 Quinault R. 100 Hatcheries using stocks of upper Queets R. 120 river origin, (spring run) Hoh R. 100 Little White Salmon '-Carson *- Soleduck 50 Leavenworth* 148 Washington, summer run Mid-Columbia River, spring run Soleduck R.* 100 Klickitat R. 50 Washington, spring run Deschutes R., spring run Soleduck R. 100 Warmsprings*-Round Butte* 109 Puget Sound and British Columbia 6 Upper Columbia River Puget Sound, fall run Winthrop, spring run 129 Elwha R. 100 Wells Dam, summer run 50 Hood Canal* 98 Snake River, spring run Deschutes R. 150 Rapid R.-Valley Ck. Green R.-Samish R. 149 Sawtooth Hatchery-Red R. 165 Puget Sound, summer run Snake River, summer run Skykomish R. 100 McCall Hatchery *-Johnson Ck.* 106 Skagit R. 100 California 4 Puget Sound, spring run Sacramento River, fall run N. Nooksack R. 50 Coleman (Battle CK.)-Nimbus- S. Nooksack R. 50 Upper Sacramento (late) 300 British Columbia, tali run Feather River-Mokelumne 200 Qualicum* 85 Sacramento River, spring run Puntledge* 100 Feather River 50 Quinsam* 97 Klamath River, fall run Robertson Ck. 100 Iron Gate 99 Capilano* 99 Trinity R. 100 San Juan R. 50 Klamath River, spring run British Columbia, summer run Trinity R. 50 Fraser River Oregon and Washington coast 5 Tete Jaune 38 Oregon, fall run Clearwater 45 Chetco R. 100 Chiiko R.* 49 Lobster Ck. 50 Stuart R.* 50 Elk R. 100 Nechako R. 55 Sixes Estuary* 100 Babine R.* 39 'Populations joined by hyphens were not distinguishable based on significant differences of allelic frequencies, and were analyzed as single units. ^Asterisks (') indicate stocks or stock groups excluded from final estimates based on preliminary estimates indicating a contribution of less than 30 fish to total catch of all fisheries sampled (see text). Multiple loci for the same class of protein are num- bered sequentially starting with the locus having the most cathodal activity. Alleles are designated num- erically as the percentage of the mobility of the homomeric band of a protein encoded by a variant allele relative to the mobility of the homomeric pro- tein band encoded by the common allele (which is designated 100). Estimates of allele frequencies were obtained for 17 polymorphic loci from each population sampled. These loci, the conditions for their electrophoretic detection, and the numbers and relative electro- phoretic mobilities of their variant allelic forms are outlined in Table 2. The allelic frequencies of these collections for these loci will be presented in a com- panion publication describing the genetic population 16 UTTER ET AL.: STOCK COMPOSITION OF 1983 CHINOOK SALMON Table 2.— Polymorphic enzymes providing genetic information for baseline popula- tions and stock mixtures. Tissue used were eye (E), and liver (L). Explanations for locus and allele designations and for buffers are given in text. Enzyme Locus designation (enzyme commission no.) (variante alleles) Tissue Buffer Aconitate hydratase Ah-4(69,86,108,116) L 2 (4.2.1.3) Adenosine deaminase Ada-1(83) E 1 (3.5.4.4) Aspartate aminotransferase Aat-3(90) E 1 (2.6.1.1) Dipeptidase (glycyl-L-leucine) Dpep-1(90) E,L 1 (3.4.13.11) Glucose-6-phiosphate isomerase Gpi-3 (93,105) E.L 3 (5.3.1.9) Glutathiione reductase Gr-1(85) E,L 1 (1.6.4.2) Hydroxyacylglutathione Hagh(140) L 1 fiydrolase (3.1.2.6) Isocitrate dehydrogenase ldh-3,4(50,74,127,142) E,L 2 (1.1.1.42) Lactate dehydrogenase Ldh-4(71,134) E,L 3 (1.1.1.27) Ldh-5(70,90) E 1 Ivlalate dehydrogenase Mdh-1, 2(27, -45,146) E,L 2 (1.1.1.37) Mdh-3,4(70,83,121) E,L 2 Mannose-6-phosphate isomerase Mpi(95, 109,1 13) E,L 1 (5.3.1.8) Phosphoglucomutase Pgm-1(-70,-84) E.L 2 (5.4.2.2) Phosphoglycerate kinase Pgk-2(90) E.L 2 (2.7.2.3) Superoxide dismutase Sod-1(- 260,560,1250) L 1 (1.15.1.1) Tripeptide aminopeptidase (L-leucylglycylglycine) Tapep(45,130) E.L 3 (3.4.11.4) structure of North American chinook salmon stocks. The general locations of baseline samples and mixed fisheries are outlined in Figure 1. The data base of six major groupings used to ana- lyze the ocean fisheries was derived from the data of the 88 collections as follows: 1) Contingency tests were used to combine data for populations lacking significant allelic differences, thus reducing the number of groups to 65. 2) GSI estimates were made from weighted (by catch) samples of genotypic data from each fishery; based on this information, those groups estimated by maximum likelihood (see below and Milner et al. 1983^) to contribute less than 0.034% (30 fish) to the total catch of all fish- eries sampled were eliminated. This brought the number of groups down to 50. 3) Estimates made for each of the 50 groups were combined into the six major groupings of Table 1 to permit a particular ^Milner, G. B., D. J. Teel, and F. M. Utter. 1983. Genetic stock identification study. Unpubl. rep., 66p. Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 2725 Montlake Blvd. E., Seattle, WA 98112. (Final Report of Research to Bonneville Power Administration, Agreement DE-A179- 82BP28044M001.) focus on tule and upriver bright stocks in the Colum- bia River. Estimates of the composition of the 1 September gill net fishery in the Columbia River were obtained for each of the 11 Columbia River fall run collec- tions and combined as indicated in Table 1. Population Mixtures Almost all of the ocean sampling was done at port of landing. Eye fluid, the only tissue, was collected in tubes placed on chipped ice, stored in various freezers, and shipped weekly in a portable freezer to the laboratory where storage was at - 90°C until preparation for electrophoresis. The September gill net fishery was sampled for livers only. Samples ob- tained by Washington Department of Fisheries (WDF) personnel from fish buyers in Ilwaco and Chinook, WA, and Astoria, OR, were collected and shipped on dry ice, and electrophoresis was carried out immediately following their arrival. Electrophoretic data were collected only for those polymorphic loci that were expressed in the collected tissues: 17 FISHERY BULLETIN: VOL. 85, NO. 1 eye - Aat-3; Ada-1; Dpep-1; Gpi-3; Gr-1; Idh-3,4; Ldh-4; Ldh-5; Tapep; Mdh-1,2; Mdh-3,4; Pgm-1; Pgk-2; Mpi. liver - Ah-4; Dpep-1; Hagh; Gpi-3; Gr-1; Idh-3,4; Tapep; Ldh-4; Mdh-1,2; Mdh-3,4; Pgm-1; Pgk-2; Mpi; Sod-1 (Table 2). Mixed Fishery Analysis Maximum likelihood estimates of proportionate contributions of baseline populations to different population mixtures were obtained by the pro- cedures described in Milner et al. (fn. 4). Through an iterative procedure (the EM algorithm, Demp- ster et al. 1977), the estimates are obtained using the frequencies of genotypes in the mixtures and in the baseline populations. Standard deviations of indiviual and pooled estimates were based on an asjonptotic variance as described in Milner et al. (fn. 4). These variances were found to be consistently higher than empirically derived variances within the sample sizes of the present study (Milner et al. fn. 4). The geographic range of mixed fisheries sampled in this study was from the Strait of Juan de Fuca southward through the mouth of the Columbia River to the northern coast of Oregon (Fig. 1). Table 3.— Estimated proportions of stock groups and subgroups in fisheries of 1983. Fishery class' area^ month^ Stock group (Estimated contribution and in parenth eses standard deviation) Number in sample [fishery] la lb 2a 2b 2c 3 4 5 6 C 1 5 0.350 0.130 0.009 0.010 0.008 0.147 0.049 0.092 0.202 1,243 2 (0.022) 0.393 (0.020) 0.228 (0.042) 0.010 (0.036) 0.001 (0.038) 0.002 (0.034) 0.100 (0.052) 0.019 (0.079) 0.080 (0.037) 0.187 [10,870] 2,050 3 (0.044) 0.456 (0.047) 0.054 (0.052) 0.022 (0.021) 0.004 (0.069) 0.009 (0.041) 0.168 (0.101) 0.124 (0.101) 0.042 (0.055) 0.122 [23,780] 319 4 (0.183) 0.375 (0.187) 0.054 (0.488) 0.007 (0.282) 0.003 (0.523) 0.004 (0.314) 0.133 (0.471) 0.189 (0.873) 0.081 (0.449) 0.153 [1,600] 600 1-4 (0.076) 0.362 (0.111) 0.187 (0.212) 0.013 (0.112) 0.001 (0.255) 0.004 (0.098) 0.121 (0.385) 0.047 (0.430) 0.079 (0.226) 0.187 [4,062] 3,475 2 7 (0.002) 0.515 (0.003) 0.156 (0.005) 0.014 (0.007) 0.001 (0.009) 0.011 (0.007) 0.123 (0.010) 0.035 (0.007) 0.047 (0.004) 0.101 [40,314] 1,044 3 (0.068) 0.225 (0.073) 0.214 (0.180) 0.032 (0.120) 0.000 (0.195) 0.012 (0.090) 0.069 (0.296) 0.115 (0.263) 0.191 (0.114) 0.142 [4,965] 784 4 (0.060) 0.341 (0.104) 0.078 (0.161) 0.016 (0.105) 0.013 (0.166) 0.016 (0.140) 0.101 (0.109) 0.160 (0.287) 0.148 (0.162) 0.126 [3,511] 1,243 2-4 (0.045) 0.353 (0.066) 0.143 (0.111) 0.019 (0.071) 0.007 (0.123) 0.026 (0.075) 0.111 (0.093) 0.087 (0.195) 0.133 (0.099) 0.121 [5,739] 2,989 1-4 5-7 (0.032) 0.366 (0.039) 0.187 (0.074) 0.014 (0.037) 0.001 (0.069) 0.006 (0.047) 0.101 (0.046) 0.056 (0.104) 0.090 (0.060) 0.181 [14,214] 4,701 1 3-4 (0.001) 0.298 (0.001) 0.197 (0.004) 0.021 (0.003) 0.002 (0.005) 0.019 (0.003) 0.163 (0.005) 0.126 (0.006) 0.061 (0.003) 0.114 [54,527] 462 4a (0.113) 0.357 (0.129) 0.054 (0.391) 0.003 (0.246) 0.002 (0.304) 0.003 (0.149) 0.076 (0.161) 0.409 (0.622) 0.022 (0.261) 0.082 [3,923] 428 3-4a (0.095) 0.330 (0.125) 0.115 (0.477) 0.012 (0.226) 0.003 (0.277) 0.008 (0.166) 0.171 (0.426) 0.221 (0.663) 0.032 (0.358) 0.109 [2,283] 731 S 1-2 6 (0.064) 0.218 (0.082) 0.158 (0.259) 0.011 (0.130) 0.000 (0.243) 0.015 (0.098) 0.085 (0.448) 0.014 (0.409) 0.099 (0.202) 0.399 [6,206] 1,633 1 6-7 (0.045) 0.202 (0.060) 0.266 (0.128) 0.017 (0.076) 0.001 (0.124) 0.010 (0.093) 0.125 (0.147) 0.029 (0.177) 0.079 (0.092) 0.271 [13,232] 1,530 2 (0.007) 0.294 (0.010) 0.215 (0.028) 0.005 (0.016) 0.000 (0.024) 0.006 (0.018) 0.099 (0.032) 0.037 (0.063) 0.114 (0.024) 0.228 [9,581] 1,760 1-2 (0.008) 0.247 (0.017) 0.263 (0.031) 0.008 (0.013) 0.000 (0.033) 0.009 (0.024) 0.122 (0.044) 0.046 (0.047) 0.107 (0.015) 0.198 [15,522] 2,846 7-8 (0.006) 0.326 (0.007) 0.217 (0.014) 0.014 (0.007) 0.000 (0.016) 0.011 (0.010) 0.048 (0.018) 0.069 (0.023) 0.070 (0.007) 0.232 [25,103] 368 8-9 (0.165) 0.240 (0.238) 0.228 (0.849) 0.009 (0.780) 0.047 (0.563) 0.040 (0.548) 0.057 (0.981) 0.088 (1.58) 0.137 (0.747) 0.155 [2,483] 739 5-9 (0.076) 0.251 (0.110) 0.211 (0.304) 0.007 (0.205) 0.001 (0.251) 0.013 (0.146) 0.118 (0.435) 0.044 (0.517) 0.096 (0.206) 0.259 [6,066] 5,315 C CR 9 (0.001) 0.780 (0.002) 0.146 (0.004) 0.011 (0.003) 0.052 (0.005) 0.007 (0.003) 0.004 (0.008) (0.007) (0.003) [46,884] 2,040 (<0.001) (0.001) (0.001) (<0.001) (<0.001) (<0.001) [15,668] 'C = commercial, I = Indian, S = Sport. ^Areas are as indicated in Figure 1. ^In sport fishery, dates for month are 6 = 5-29 9 = 9-1 only. 6-17, 7 = 6-18 - 7-29, 9 = 8-16 - 9-11; in Columbia River commercial fishery 18 UTTER ET AL.: STOCK COMPOSITION OF 1983 CHINOOK SALMON DISTRIBUTIONS OF STOCK GROUPS IN OCEAN FISHERIES Estimated contributions to different fisheries by the two tule and three upriver bright subgroups, and by the four other geographic groupings, are Hsted in Table 3. Some of the major features of Table 3 are graphically projected in Figure 2. A stock struc- ture that varies with regard to both time and area is evident. Some consistency over time is seen in comparisons of the May and July commercial troll catches in sam- pling areas 2 and 4. Notable features in sampling area 2 include the overall predominance of the tule stocks and a minimal contribution of Puget Sound and Canadian fish. Sampling area 4 has a smaller tule contribution and a substantially larger propor- tion of Puget Sound and Canadian fish. Although comparisons of the sport and commer- cial fisheries are limited by somewhat different sampling areas, a greater proportion of California fish is taken in the sport fisheries. This is seen par- ticularly in the early fishery. More intense sam- pling in area 1 may account for some of the early differences, but the data of Table 3 suggest a per- sistent trend even in common times and sampling areas. The Indian troll fishery provided the only infor- mation from sampling area 4B. The most distinc- tive feature of this fishery was the high proportion (41%) of the Puget Sound and Canadian group. This figure was more than double the estimated contribu- tion of this group in any other fishery. The ocean fisheries off the Washington coast are notable for the usually negligible representation of upriver bright stocks. The highest estimated contri- bution (9.6%) occurred in the sport fishery of 16 August- 11 September which was the largest fishery sampled. The timing and distribution of upriver bright fish will be considered in greater detail below. Estimated distributions for the May 1982 and 1983 troll fisheries were compared (Table 4), reveal- California Washington-Oregon coast Puget Sound British Columbia Other Columbia River stocks Upriver brights Tule-Kaiama, Cowlitz Tule-Spring Creek group Number in fishery (thousands) ////////////// Commercial troll 1' )' r t'l'yj'/'r^ '/y///////// ' ////////// Indian troll 23 4.1 5.7 2.3 Catch area Date May 4B July May- July /////////////////////// ////////////////////// ////////////////////// Sport fishery I r ~i 1 1 1 1 1 1 r r- 10 20 30 40 50 1 1 1 1 1 1 1 1— 60 70 80 90 — t 100 n 1 9 May 28- '■^ ^-^ June 17 25 1-2 June 18 July 29 2.5 1-2 June 30 Aug 15 6.1 1-2 Aug 16- Sept 11 Percent contribution to fishery Figure 2.— Estimated contributions of seven stock groups in different ocean fisheries. 19 FISHERY BULLETIN: VOL. 85, NO. 1 Table 4.— Comparisons of estimated percentage stock group contributions to May troll fisheries of 1982 and 1983 within sampling areas 1 and 4 and 1 through 4. Percentage catch estimated from sampling area 1 4 1 through 4 Stock group 1982 (1,414)2 1983 (1,243) 1982 (448) 1983 (600) 1982' (2,504) 1983 (3,475) Columbia River Tule Upriver bright All groups California Oregon-Washington coast Puget Sound-Canada 78.2 4.0 88.9 5.3 3.8 1.9 48.2 2.7 65.6 20.2 9.2 4.9 45.6 8.8 56.8 4.4 9.7 29.2 42.9 1.0 57.6 15.3 8.1 18.9 76.5 4.3 90.9 2.8 2.9 3.4 54.9 1.8 69.5 18.3 7.9 4.7 'Data from Miller et al. 1983. ^Values In parentheses designate number of fish sampled. ing considerable dissimilarity as well as some consistency. A much larger proportion of tules and a correspondingly smaller contribution of Califor- nia fish was seen in 1982 in sampling areas 1 through 4. The comparisons of estimates within sampling areas 1 and 4 differed between both sampling areas and years. For each year, estimates within sampling area 1 were similar to the estimates based on the total sample. Estimates from sampling area 4 were consistent with those of the total sample but with a larger proportion of California fish estimated in 1983 and substantially larger Puget Sound-Canadian stock group and smaller tule estimates. These obser- vations are consistent with the location of sampling area 4 at the southern point of entry for most of the populations destined for Puget Sound and British Columbia, and are reinforced by the high propor- tion of fish estimated for this stock group in sam- pling area 4B. The much higher total harvest of the 1982 May troll fishery (73,196; Miller et al. 1983) than in the same fishery for 1983 (40,312) accentuates the dif- ference in numbers of tules taken (approximately 56,000 vs. 22,700). The numbers of tule group fish returning to the mouth of the Columbia River and to hatcheries (i.e., spawning escapement) were also much lower in 1983 (Washington Department of Fisheries 1984'^) and were insufficient to fulfill hatchery requirements. This contrast was attributed to a climatological phenomenon termed "El Niiio" that affected the oceanic distribution and survival of many species beginning in 1983 (Mysak 1986). ^Washington Department of Fisheries. 1984. Status of fall Chinook stocks in the northern Oregon through Vancouver Island ocean fishing areas. Unpubl. rep., 35 p. Department of Salmon Harvest Management, Washington Department of Fisheries, 115 General Administration Building, Olympia, WA 98504. DISTRIBUTION AND RELATIVE CONTRIBUTIONS OF TULES AND UPRIVER BRIGHTS The actual and potential value of tule and upriver bright runs in the sampling areas of this study war- rant a more detailed focus on the abundance of these stocks and their subgroups. The great value of tule stocks in ocean fisheries off the Washington coast has been demonstrated from these and other data (e.g., Miller et al. 1983). Although a similar value for upriver brights in either oceanic or river harvests is not yet apparent, it is premature to assign a lesser value to these runs because of geographic and tem- poral limitations of sampling. Indeed, data from coded vdre tags (Table 5) indicate a distinctly differ- ent oceanic distribution of tules and upriver brights. Over half of the recoveries of the tagged fish from the tule stock (Spring Creek) were harvested off the Washington coast. However, only about 5% of the tagged fish from upriver bright stocks were re- covered in this area, with over 90% harvested in waters of Alaska and British Columbia. The substantially increased contribution of upriver brights in the late sport fishery (Table 3, Fig. 2) is consistent with a late migratory surge of these fish. Clearly, based on distributions indicated through tag data in Table 5, the upriver brights contribute heavily to fisheries in areas north of those sampled in this study. However, they were estimated at sizable numbers only very late in this study presum- ably enroute through the areas sampled to their spawning grounds. The tules and upriver brights have been con- sidered as unit populations to this point. The sub- group data indicate considerable heterogeneity within both groups with regard to time, area, and fishery. Comparisons of the two tule subgroups (Table 3, Fig. 2) indicate a considerable difference 20 UTTER ET AL.: STOCK COMPOSITION OF 1983 CHINOOK SALMON Table 5. — Summary of distribution of oceanic coded wire tag recoveries (N) of 1975 brood year fall chinook salmon from the Snake River, and Priest Rapids and Spring Creek hatcheries. Sample Recovery area Source and Canada Washing- Total type of stock size Alaska (B.C.) ton Oregon no. fish Snake River^ N 176 272 21 11 480 (upriver bright) % 36.7 56.7 4.3 2.3 Priest Rapids^ N 1,314 1,597 171 13 3,095 (upriver bright) % 42.4 51.9 5.4 0.3 Spring Creek^ N 0 984 1,319 147 2,450 (tule) o/o 0 40.2 53.8 6.0 ' Data from L. Gilbreath, Northwest and Alaska Fisheries Center, Seattle WA 981 1 2, pars, commun September 1984 ^Data obtained from Pacific Marine Fisheries Commission in 1983. in their relative frequencies. In both May and July, the proportion of the Cowlitz-Kalama subgroup (group lb) to the overall tule contribution in the com- mercial fisheries was considerably higher in sam- pling area 2 (average 30%) than in sampling area 4 (average 16%). The proportion of this subgroup was highest in the sport fisheries, approaching equality (46%) with the Spring Creek subgroup (la) in the overall data set and predominating in the June-July fisheries (52%). The Spring Creek sub- group strongly predominated in the tule catch of the river fishery of 1 September (84%, Table 3). The relative contributions of the three upriver bright subgroups vary considerably in the ocean fisheries (Table 3, Fig. 3). The most notable feature is the absence or negligible contribution of the Priest Rapids subgroup (2b) in all but the last ocean fishery that was sampled, where this subgroup contributes a substantial proportion (49%) of the total estimated upriver bright harvest. This finding was unexpected because this subgroup is by far the largest contrib- utor to the overall upriver bright production (Pat- tillo and Mclsaac 1982). The data of the 1 September river fishery are more consistent with expectations, with 83% of the upriver bright catch estimated to be from the Priest Rapids subgroup. The low esti- mated contribution of this subgroup to ocean fish- eries cannot be explained by the large standard deviations accompanying most estimates because the more precise pooled estimates (Table 3) also in- dicate a small Priest Rapids contribution. MANAGEMENT CONSIDERATIONS The difference in relative proportions of the two tule groups, based largely on class of fishery and Commercial troll Sport fishery ^ Snake River Priest Rapids Deschutes Estimated number in fishery Catch (hundreds) area Date 7.3 3.4 4.3 5.8 1-4 1-2 1-2 May 11.5 1-4 July May 28- June 17 June 18- June 29 1 ^ Aug 16- '"-^ Sept 11 I I I I I I I I 1 I r~i I I I I I 1 I I I 0 10 20 30 40 50 60 70 80 90 100 Percent of upriver bright contribution to fishery Figure 3.— Estimated proportions of three upriver bright stocks to different ocean fisheries. 21 FISHERY BULLETIN: VOL. 85, NO. 1 area, has implications for management. The higher representation of the CowHtz-Kalama subgroup in the sport fisheries than in the troll fisheries of com- mon times and areas suggests a greater suscep- tibility of this subgroup to sport harvests. In addi- tion, the relative abundance of the Cowlitz-Kalama subgroup compared with the Spring Creek subgroup was higher in more southern areas for both commer- cial and sport fisheries. If these trends continue to be observed, different management strategies could be applied for these groups when warranted. The low estimates of the Priest Rapids subgroup of upriver brights relative to the two less abundant subgroups suggest different oceanic distributions of these subgroups. However, the coded wire tagging data (Table 5) indicate that at least the Snake River and Priest Rapids subgroups are harvested much more intensely in areas to the north of those sam- pled in this study (no tagging data were available for the Deschutes subgroup). Any attempts to iden- tify and protect the weaker subgroups within the sampling areas of this study would be futile unless similar efforts could be applied to these much larger catches in more northern areas. A general occurrence of larger proportions of Puget Sound and Canadian fish in the northern sam- pling areas is suggested by the similar observations for 2 consecutive years and by the particularly high estimates for these fish in area 4B. Since 1983, more detailed GSI estimates from area 4B have, in fact, been used by the WDF to monitor and regulate Chinook salmon fisheries in the Strait of Juan de Fuca and Puget Sound areas. Preliminary results from the September gill net fishery in the lower Columbia River (based on a sub- sampling of 500 fish) were available on the day following the collection of the samples. This poten- tial for rapid turnaround time increases the value of the GSI as a management tool by permitting in- season regulatory adjustments. Such information would allow greater harvest of a healthy stock while continuing to provide for maximum protection of a depressed stock. For example, in years when bright fish are expected to return in great abundance and tules in low abundance, the GSI method could be used to monitor extended fall gill net fisheries to time the entry of tules. When ratios of tules to brights became unfavorable, fisheries could be curtailed. It is important to emphasize the arbitrary nature of many of this study's groupings, which were neces- sary to provide a manageable basis for reporting. A focus on the tule and upriver bright contributions was appropriate because of the extensive baseline data from the Columbia River drainage, the domi- nance of the tule runs in ocean fisheries, and the distinct oceanic distributions of the tule and upriver bright groups. However, a similar focus on other groupings (e.g., Columbia River spring runs or wild and hatchery stocks of the Oregon coast) is equally feasible, and could easily provide a basis for more detailed information on the distributions of in- dividual populations within such groups. The completeness and the reliability of the sets of baseline data that are used affects the accuracy of GSI estimates. This study's focus on the contribu- tion of Columbia River populations to stock mixtures in ocean areas adjacent to the mouth of the Colum- bia River was appropriate for the sets of baseline data that were used. Most estimates were obtained through a data base that included most of the major contributing groups within the Columbia River and allele frequency data from 17 polymorphic loci. These same baseline data can be used over succes- sive years, providing the allele frequencies remain stable among year classes and over succeeding generations. Such stability has been observed for some loci and populations of anadromous salmonids (e.g., Utter et al. 1980; Grant et al. 1980; Altukhov 1981). This temptation to regard the present baseline data as a static entity should nevertheless be re- sisted for a number of reasons. Gene flow, genetic drift, and selection could modify allelic frequencies over extended time periods; thus, periodic updating of previously sampled populations is desirable. Tem- poral changes in allele frequencies of chinook salm- on have been reported (Carl and Healey 1984; Kris- tiansson and Mclntyre 1976). The extensive stock transplantations of chinook salmon within the Columbia River make the possibility of gene flow particularly likely for the focal populations of this study. Hatchery populations perpetuated by limited numbers of breeders are particularly susceptible to allele frequency changes through genetic drift (Allendorf and Ryman 1987). Previously unsampled baseline populations should be added, particularly in areas where limited sampling has occurred, to in- crease the accuracy and broaden the usable range of analyses. The discriminatory powers of GSI analyses are substantially increased as new variable loci are added (see Milner et al. fn. 4). The con- tinuing search for additional markers requires col- lection of electrophoretic data from previously sampled populations for each new variable locus that is found. Increasing application of procedures used in this study seems virtually inevitable in view of the per- 22 UTTER ET AL.: STOCK COMPOSITION OF 1983 CHINOOK SALMON sistent need to understand the composition of stock mixtures of salmonids (and other structured groups) better. The obvious management potential of such uses is matched by increased understanding of popu- lation structuring and of migratory behavior that will emerge as information accumulates. ACKNOWLEDGMENTS This study could not have been carried out with- out the assistance in all phases of sampling of Richard Lincoln, Mark Miller, and Patrick Pattillo of Washington Department of Fisheries. LITERATURE CITED Allendorf, F. W., and N. Ryman. 1987. Genetic management of hatchery stocks. In N. Ryman and F. Utter (editors), Population genetics and fishery man- agement, p. 141-160. Univ. Washington Press, Seattle. Allendorf, F. W., and F. M. Utter. 1979. Population genetics. In W. S. Hoar, D. J. Randall, and J. R. Brett (editors). Fish physiology. Vol. 8, Bioenergetics and growth, p. 407-454. Acad. Press, Inc. N.Y. Altukhov, Yu. p. 1981. The stock concept from the viewpoint of population genetics. Can. J. Fish. Aquat. Sci. 38:1523-1538. Beacham, T., R. Withler, and A. Gould. 1985. Biochemical genetic stock identification of chum salmon (Oncorhynchus keta) in southern British Columbia. Can. J. Fish. Aquat. Sci. 42:437-448. Carl, L. M., and M. C. Healey. 1984. Differences in enzyme frequency and body morphology among three juvenile life history types of chinook salmon {Oncorhynchus tshawytscha) in the Nanaimo River, B.C. Can. J. Fish. Aquat. Sci. 41:1070-1077. Clayton, J. W., and D. N. Tretiak. 1972. Amine-citrate buffers for pH control in starch gel elec- trophoresis. J. Fish. Res. Board Can. 29:1169-1172. Dempster, A. P., N. M. Laird, and D. B. Rubin. 1977. Maximum likelihood from incomplete data via the EM algorithm. J. R. Stat. Soc. B39:l-38. FouRNiER, D., T. Beacham, B. Riddell, and C. Busack. 1984. Estimating stock composition in mixed stock fisheries using morphometric, meristic, and electrophoretic character- istics. Can. J. Fish. Aquat. Sci. 41:400-408. Grant, S., G. Milner, P. Kranowski, and F. Utter. 1980. Use of biochemical genetic variants for identification of sockeye salmon (Oncorhynchics nerka) stocks in Cook In- let, Alaska. J. Fish. Res. Board Can. 37:1236-1247. Kristiansson, a. C, and J. D. McIntyre. 1976. Genetic variation in chinook salmon {Oncorhynchus tshawytscha) from the Columbia River and three Oregon coastal rivers. Trans. Am. Fish. Soc. 105:620-623. Markert, C. L., and I. Faulhaber. 1965. Lactate dehydrogense isozyme patterns of fish. J. Exp. Zool. 156:319-332. May, B., J. E. Wright, and M. Stoneking. 1979. Joint segragation of biochemical loci in salmonidae: Results from experiments with Salvelimis and review of the literature on other species. J. Fish. Res. Board Can. 36: 1114-1128. Miller, M. P., P. Pattillo, G. B. Milner, and D. J. Teel. 1983. Analysis of chinook stock composition in the May, 1982 troll fishery off the Washington coast: an application of the genetic stock identification method. Wash. Dep. Fish., Tech. Rep. 74, 27 p. Milner, G. B., D. J. Teel, F. M. Utter, and G. A. Winans. 1985. A genetic method of stock identification in mixed popu- lations of Pacific salmon. Mar. Fish. Rev. 47(1); 1-8. Mysak, L. a. 1986. El Nino, interannual variability and fisheries in the northeast Pacific Ocean. Can. J. Fish. Aquat. Sci. 43:464- 497. Pattillo, P., and D. McIsaac. 1982. Unexplained loss of adult fall chinook in the Columbia River between Bonneville and McNary Dams, 1977-1981. Wash. Dep. Fish., Progress Rep. 162, 28 p. Pella, J., and G. Milner. 1987. Genetic marks and stock identification. /nN. Ryman and F. Utter (editors). Population genetics and fisheries management, p. 247-276. Univ. Washington Press, Seattle. RiDGWAY, G. J., S. W. Sherburne, and R. D. Lewis. 1970. Polymorphism in the esterases of Atlantic herring. Trans. Am. Fish. Soc. 99:147-151. Utter, F. M., H. 0. Hodgins, and F. W. Allendorf. 1974. Biochemical genetic studies of fishes: potentialities and limitations. In D. C. Malins and J. R. Sargent (editors). Biochemical and biophysical perspectives in marine biology. Vol. 1, p. 213-237. Acad. Press, Lond. Utter, F. M., D. Campton, S. Grant, G. Milner, J. Seed, and L. Wishard. 1980. Population structures of indigenous salmonid species of the Pacific Northwest. In W. J. McNeil and D. C. Hims- worth (editors), Salmonid ecosystems of the North Pacific, p. 285-304. Oregon State Univ. Press, Corvallis. 23 ON THE STANDARD METABOLIC RATES OF TROPICAL TUNAS, INCLUDING THE EFFECT OF BODY SIZE AND ACUTE TEMPERATURE CHANGE Richard W. Brilli ABSTRACT The standard metabolic rates (SMR's) of fishes and the effect of body weight on SMR's are important input parameters to energetics, growth, and population models. This study was undertaken to obtain these data for the tropical tuna species, yellowfin tuna, Thunnus albacares, and kawakawa, Euthynnus affinis. These data compliment similar SMR measurements from skipjack tuna, Katsuwonus pelamis, previously published. The effect of acute temperature change on the SMR of all three species was also determined. The SMR was estimated by directly measuring the oxygen uptake rate of animals paralyzed with a neuromuscular blocking drug, rather than by the more commonly used method of extrapolation of swim- ming speed-metabolic rate curves back to zero swimming speed. To test the adequacy of this technique, the SMR's of aholehole, Kuhlia sandvicensis. and rainbow trout, Salmo gairdnerii, were determined using similar methodology. The SMR's measured in this way were not significantly different from the published SMR's of these species determined by extrapolation of swimming speed-metabolic rate curves back to zero swimming speed. All three tuna species have very high SMR's, over five times higher than other active teleost species such as salmon and trout. The effect of body size on the SMR is similar in all three tuna species, but the weight specific SMR of tuna decreases more rapidly with increasing body size than in other fishes. Based on SMR's measured at 20° and 25°C, the Qjo's were 3.16, 2.31, and 2.44 for yellowfin tuna, kawakawa, and skipjack tuna, respectively. These are similar to Qig values found for the SMR's of other teleosts. Tunas can achieve exceptionally high maximum aerobic metabolic rates. This ability requires a com- plete set of anatomical, physiological, and biochemical adaptations. I hypothesize that one of these adap- tations, large gill surface areas, causes tunas to have exceptionally high energy demands even at rest. Tunas' high SMR's are an inevitable consequence of their ability to achieve exceptionally high maximum aerobic metabolic rates. The standard metabolic rate (SMR) (the metabohc rate of a postabsorptive animal completely at rest) and the effect of body size on SMR are important input parameters to growth, energetics, and popula- tion models (e.g., Sharp and Francis 1976; Kitchell et al. 1978). This study was therefore undertaken to obtain these data for yellowfin tuna, Thunnus albacares, and kawakawa, Euthynnus affinis. These measurements were designed to directly compli- ment the SMR measurements for skipjack tuna, Katsuwonus pelamis, that had been previously pub- lished (Brill 1979). The effect of acute temperature change on the SMR of skipjack tuna, yellowfin tuna, and kawakawa was also determined. The effect of acute temperature change, as opposed to the effect of temperature adaptation, is relevant to tuna be- cause of the 5° to 15°C water temperature changes 'Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 96822-2396. Manuscript accepted November 1986. FISHERY BULLETIN: VOL. 85, NO. 1, 1987. these species normally experience during the daily vertical movements which are a constant feature of their behavior in the open ocean (Dizon et al. 1978; Carey and Olson 1982; Yonemori 1982). In other teleosts, SMR's have been determined by extrapolating metabolic rate-swimming speed curves back to zero swimming speed (e.g., Brett 1965). Although Graham and Laurs (1982) have suc- cessfully measured the metabolic rate of albacore, T. alalunga, (a temperate tuna species) swimming in a water tunnel, this methodology is presently not possible with tropical tunas (skipjack tuna, yellow- fin tuna, and kawakawa). Attempts to get these species to swim in several prototype water tunnel designs have shown that they will do so for only very short periods (Brill and Dizon 1979 and unpublished observations). As a result, measuring the SMR's of tropical tunas directly in animals paralyzed with a neuromuscular blocking agent is currently the only method available to obtain these data. To validate this technique, the SMR's of rainbow 25 FISHERY BULLETIN: VOL. 85, NO. 1 trout, Salmo gairdneri, and aholehole, Kuhlia sand- vicensis, were also measured using paralyzed ani- mals. These two species were chosen because they are available in Hawaii and because there are pub- lished data on their SMR's based on extrapolation of swimming speed-metabolic rate curves back to zero swimming speed (Muir et al. 1965; Bushnell et al. 1984). The SMR of a 1 kg skipjack tuna (412 mg Og/h, Brill 1979), is almost five times greater than that of a 1 kg sockeye salmon, Oncorhynchus nerka (83 mg 02/h, Brett and Glass 1973). The former mea- surements were made at 25 °C and the latter at 20°C, because 25°C is the upper lethal temperature for salmon (Brett 1972). However, a 5°C tempera- ture difference could not account for this SMR dif- ference because the Qio's for the SMR's of fishes are generally about 2 (Robinson et al. 1983). The maximum sustainable aerobic metabolic rate (MMR, the metabolic rate at the maximum swimming speed sustainable for at least 1 h) of a 1 kg sockeye salmon at 20°C is 796 mg 02/(kg-h), whereas 1.8-2.2 kg skipjack tuna at 24°C have been shown to be able to achieve active metabolic rates over 2,000 mg 02/(kg-h) (Gooding et al. 1981). Although there are no metabolic rate measurements available for tunas at their maximum sustainable swimming speeds, two conclusions are still obvious: 1) skipjack tuna have very high SMR's even when compared with other active equal sized teleosts and 2) skipjack tuna are capable of very high aerobic metabolic rates. I hypothesize that the high SMR's of tunas are primarily a result of their large gill surface areas (Hughes 1979). In other words, adaptations that per- mit high maximum sustainable rates of oxygen up- take (i.e., high MMR's) obligate tunas to have high SMR's. Analogous arguments with respect to the resting and maximal metabolic rates of terrestrial vertebrates have been presented by Bennett and Ruben (1979). MATERIALS AND METHODS SMR Measurements-Tuna Live skipjack tuna, yellowfin tuna, and kawakawa were purchased from local fishermen and main- tained at the Kewalo Research Facility (Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA). Animal procure- ment, handling, and maintenance procedures at this facility are described by Nakamura (1972), Queenth and Brill,^ and Chang et al.^ Fishes were main- tained in outdoor tanks for a few days to over 1 yr before use. Temperature of the seawater supplied to the holding tanks was 25°C ( + 2). Food was pre- sented daily; however, individuals were not fed for at least 20 h prior to use in an experiment. This allowed sufficient time for gut clearance and for blood glucose level to return to prefeeding levels (Magnuson 1969). Each experimental animal was removed from its holding tank by dip net and injected intramuscular- ly with 1-3 mg/kg of the neuromuscular blocking agent Flaxedil"* (gallamine triethiodide). The animal was quickly returned to its holding tank, and when it could no longer swim, it was immediately rushed into the laboratory and placed in a Plexiglas flow- through box respirometer similar to that used by Stevens (1972). The respirometer was equipped with a movable partition which was placed immediately behind the fish to reduce the respirometer' s volume and, thus, reduce the lag time between actual and measured changes in metabolic rate to only minutes (Niimi 1978). Water flow through the respirometer was maintained at 3-7 L/(kg- min) and was measured every 30-60 min by recording the time to fill a 1 L graduated cylinder. Water temperature was con- trolled by a chiller and freshwater heat exchanger and by a quartz heater mounted in the inflow sea- water line. Temperature control was ±0.3°C. Unlike the previous study on the SMR of skipjack tuna (Brill 1979), the spinal cord was not cut to stop all overt muscular activity. Rather, an 18-gauge hypodermic needle was placed intramuscularly and connected to the outside of the respirometer via a short length of polyethylene tubing. Through this tube, 0.1-0.3 mL doses of Flaxedil were adminis- tered when the fish began to show any slight tail movements. To monitor heart rate, electrocardio- gram leads were mounted subcutaneously on the ventral body surface. Heart rate was determined by timing the interval between successive beats with a Hewlett-Packard (HP) 5308A frequency counter. Thermistors were used to measure fish muscle and water temperatures. With the aid of an 18-gauge hypodermic needle, a thermistor bead mounted in 0.9 mm diameter polyethylene tubing was inserted 2Queenth, M. K. K., and R. W. Brill. 1983. Operations and pro- cedures manual for visiting scientists at the Kewalo Research Facility. Southwest Fisheries Center Honolulu Laboratory, Na- tional Marine Fisheries Service, NOAA, Honolulu, HI 96822-2396, Administrative Report H-83-7, 16 p. ^Chang, R. K. C, R. W. Brill, and H. 0. Yoshida. 1983. The Kewalo Research Facility, 1958 to 1983—25 years of progress. Southwest Fisheries Center Honolulu Laboratory, National Mar- ine Fisheries Service, NOAA, Honolulu, Hawaii 96822-2396, Ad- ministrative Report H-83-14, 28 p. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 26 BRILL: STANDARD METABOLIC RATES OF TROPICAL TUNAS into the red muscle immediately adjacent to the spinal column. Thermistor probes were also mounted in the incoming seawater line and in the respirometer box itself. Red muscle and water tem- peratures were determined by measuring the resis- tance of the various thermistors with an HP 3456A digital multimeter. Oxygen concentration (milligrams per liter) of the water upstream and downstream of the fish was determined with a dissolved oxygen meter (Yellow- springs Instrument, model 51A) equipped with a Clark-type polarographic electrode oxygen-temper- ature probe. The probe was normally in the outflow seawater line, but was moved to the inflow seawater line to determine inflow seawater oxygen levels every 30-60 min. The analog output of the oxygen meter was also measured with the HP digital multi- meter. An HP 9825A computer was used to control an HP 5930A six-channel relay actuator which per- mitted the digital multimeter to determine sequen- tially the resistances of various thermistors and the analog output of the oxygen meter. Seawater oxy- gen level, red muscle and water temperatures, meta- bolic rate, and heart rate were calculated and printed by the computer at 5-min intervals. After being sealed, the respirometer box was covered with black plastic to minimize disturbance to the fish. Temperature of the seawater supplied to the respirometer was maintained at 21° -22 °C for the first 1-2 h because reduced water temperature has been shown to help tuna survive after handling (Barrett and Connor 1964). Seawater temperature was then changed to either 20° or 25°C, and the fish maintained at the test temperature until its metabolic rate remained relatively stable for at least 1 h. The SMR was estimated by averaging the last 5-12 metabolic rate measurements. The standard deviations of the metabolic rate measurements used to estimate SMR were <11% of the mean (i.e., SMR) in all cases, and in 70% of the cases, the standard deviations were <5% of the mean. To determine the SMR at a second temperature, the water temperature was changed to either 20°, 25°, or 30° C, and metabolic rate measurements con- tinued again until the fish's metabolic rate remained stable for 1 h. SMR Measurements-Aholehole and Rainbow Trout Aholehole were obtained from Sea Life Park (Wai- manalo, HI) and rainbow trout from a commercial fish farmer (through the University of Hawaii, Hilo). The former were maintained in an outdoor tank with running seawater at 25°C ( ± 2) and the latter, in an indoor tank with running freshwater at 15°C (± 2). Both species were fed daily, but individuals were not fed for at least 20 h prior to use in an experiment. The respirometer used for aholehole was essen- tially identical to that used by Davis and Cameron (1971) and Jones and Schwarzfeld (1974) to measure water flow and gas exchange across the gills of rain- bow trout. The aholehole were anesthetized in 1: 10,000-1:30,000 MS222 (Tricaine methanosulfo- nate). A thin, rubber membrane was sutured around the fish's mouth and sealed with a small amount of tissue glue (Histoacryl, B. Braun Melsungen AG, West Germany). The fish was then placed in a black Plexiglas box that was open at both ends. This box was then placed in a larger tank that was divided into two chambers by a partition with a hole through it. The membrane sealed around the fish's jaws was attached to the edge of the hole and sealed in place with a Plexiglas plate held with stainless steel wing nuts. This system allowed separation of the inspired and expired water, yet allowed the fish to make nor- mal respiratory movements. Water level in the two chambers was maintained by standpipes (constant level drains). Ventilation volume was determined by measuring the water flow rate from the standpipe in the chamber containing the fish. By lowering this standpipe, the fish could be force-ventilated. Water samples were drawn from the anterior chamber, and from the black Plexiglas box contain- ing the fish, approximately every 15 to 20 min. Water oxygen level was determined with a water- jacketed oxygen electrode (Radiometer, Copen- hagen) maintained at 25 °C. Metabolic rate was calculated using the oxygen content difference be- tween inspired and expired water and the ventila- tion volume. Aholehole were given 2 h to recover from the anesthesia before metabolic rate measurements were begun. A series of metabolic rate measure- ments were made with the water level in the two chambers even and the fish actively pumping water over its gills, until the its metabolic rate remained relatively stable for at least 1 h. The water level in the chamber containing the fish was then lowered and measurements taken while the animal was being force- ventilated, continuing again until the meta- bolic rate stablized. Finally, the fish was given 0.1- 0.3 mL Flaxedil (intramuscularly) and metabolic rate measurements continued while the animal was para- lyzed and force-ventilated. In two cases, the fish was left in the respirometer overnight on forced venti- lation to allow the effects of Flaxedil to wear off. Metabolic rate measurements were made again 27 FISHERY BULLETIN: VOL. 85, NO. 1 before and after Flaxedil injection the next day. The SMR was calculated as the mean of the last four to six metabolic rate measurements. Water tempera- ture was maintained at 25°C (±0.3) throughout the experiment. The SMR of rainbow trout was directly deter- mined in the same respirometry box as that used for tunas, using essentially identical methodology, except freshwater was used, inspired and expired water were sampled, and oxygen levels were mea- sured with a water-jacketed Radiometer oxygen electrode. RESULTS Effects of Body Size on SMR The SMR's of 21 kawakawa (0.540-2.153 kg) and 13 yellowfin tuna (0.585-3.890 kg) were determined at 25°C. Regression lines of SMR versus body weight were fitted by Gauss-Newton iteration (Bio- medical Computer Programs, Program BMDP 3R), rather than a log-log transformation of the data (Fig. 1). The advantages of the former and disadvantages of the latter method are discussed by Zar (1968) and Glass (1969). The best fitting allometric equations are 1) Kawakawa: SMR = 392.5 (±32.3) pro.496(±o.i45) n = 21 2) Yellowfin tuna: SMR = 286.8 (±26.9) l^-573(±o.ii6) n = 13. For comparison, the allometric equation relating 1000 900 800 700 600 500 400 300 o> 200 E DC s (0 100 90 80 70 60 50 1 — I — I — rnr-r T — I — r SKIPJACK TUNA' FROM BRILL (1979) KAWAKAWA A • / / / / / A ^'^ ^YELLOWRN TUNA / SKIPJACK TUNA FROM GOODING et al., (1981) -SALMON FROM BRETT & GLASS (1973) • YELLOWFIN A KAWAKAWA 40 0.3 _L J ' I I I I _L J I L 0.4 0.5 0.6 0.8 1.0 2.0 3.0 4.0 5.0 6.0 8.0 BODY WEIGHT (kg) Figure 1.— A double logarithmic plot of the standard metabolic rates (SMR) of 13 yellowfin tuna and 21 kawakawa. The lines represent the allometric equations: SMR = 286.8 W^°^^^ SMR = 392.5 W°^^^ and SMR = 412.0 VF"^^^ for yellowfin tuna, kawakawa, and skipjack tuna, respectively, where the SMR is mg Og/h and W is body weight in kilograms. The line for skipjack tuna is from Brill (1979). For com- parison, the regression lines based on swimming skipjack tuna (Gooding et al. 1981) and for salmon at 20°C (Brett and Glass 1973) are also shown. All tuna data is from fish at 23°-25°C. 28 BRILL: STANDARD METABOLIC RATES OF TROPICAL TUNAS SMR and body weight in skipjack tuna is (Brill 1979) SMR = 412.0 (±27.1) W"0.563(±o.07) n = 33. The SMR is in mg O^/h and W is body weight in kilograms. The values in parentheses are the stan- dard errors of the parameters. Effects of Acute Temperature Change on SMR, Heart Rate, and Excess Red Muscle Temperature A total of 8 kawakawa, 12 yellowfin tuna, and 5 skipjack tuna were subjected to 5°C temperature changes. Most temperature changes were made between 20°C and 25°C, which all fish survived. Ten fish were exposed to 25° and 30°C, but only four survived long enough at 30°C to provide usable data. Because of the expense and difficulty in obtaining live tunas, the latter treatment was not pursued. The SMR's and mean heart rates at 20°, 25°, and 30°C are given in Table 1. The Qio's of SMR for water temperatures changes from 20° to 25°C were variable and ranged from 5.82 to 1.39. The mean Qio's (±95% confidence intervals) were 2.44 ± 0.97, 2.31 ± 0.51, and 3.16 ± 0.93 for skipjack tuna, yellowfin tuna, and kawakawa, respectively. The range of mean excess red muscle tempera- tures are given in Table 2. These excess muscle temperatures are lower than those measured in free swimming yellowfin and skipjack tunas (Dizon and Brill 1979). This is as expected because in paralyzed tunas, most of the heat production (i.e., energy con- sumption) most likely occurs at the heart and gills where the heat would not be retained by the vascular countercurrent heat exchangers. The SMR of Aholehole and Rainbow Trout Aholehole, unlike rainbow trout, will not sit quietly in a darkened respirometer box nor stop breathing movements when force-ventilated. Because Flaxedil Table 1. — Effect of temperature on the standard metabolic rate and heart rate of yellowfin tuna, kawakawa, and skipjack tuna. Species Weight SMR (mg Og h) Q 10 Heart rate (min -') Q 10 20°C 2e i°C 30°C 20°-25°C 25'='-30°C 20°C 25°C 30°C 20°-25°C 25°-30°C Yellowfin tuna 2.215 470 + 47 625 + 53 — 1.77 — 89 + 1 133 + 2 — 2.21 — 1.438 258 ± 8 386 + 21 605 ± 13 2.24 2.46 98 + 1 146 + 2 176 ± 1 2.22 1.45 1.635 161 ± 17 330 ± 17 — 4.20 — 81 + 2 138 ± 2 — 2.89 — 3.890 311 + 12 501 + 32 — 2.60 — 57 + 2 90 + 2 — 2.49 — 0.704 112 + 6 184 + 6 — 2.70 — 99 + 4 149 + 1 — 2.26 — 0.877 150 + 6 193 + 6 — 1.66 — 77 + 1 121 ± 3 — 2.45 — 0.599 103 + 3 166 + 6 — 2.60 — 73 ± 6 138 ± 8 — 3.55 — 0.595 153 ± 6 187 ± 6 — 1.49 — 74 ± 6 137 ± 5 — 3.41 — 0.585 154 + 9 199 + 3 266 + 10 1.67 1.73 118 ± 1 159 ± 3 200 ± 2 1.82 1.58 1.290 333 + 4 493 ± 17 — 2.19 — 106 ± 1 160 ± 2 — 2.28 — Mean: 2.31 2.10 2.56 1.59 Standard deviation: 0.80 0.52 0.56 0.09 Kawakawa 1.439 258 + 12 431 + 19 — 2.79 — 128 + 1 205 ± 8 — 2.57 — 1.713 331 + 12 497 + 11 — 2.25 — 147 + 2 213 ± 2 — 2.10 — 0.870 170 + 14 410 ± 27 — 5.82 — 147 ± 4 217 + 6 — 2.18 — 1.283 379 + 9 598 ± 25 — 2.49 — 146 ± 14 206 ± 23 — 1.99 — 1.623 363 + 27 640 ± 23 — 3.11 — 118 + 4 175 ± 3 — 2.20 — 1.377 — 543 ± 35 761 ± 23 — 1.96 — 200 ± 14 272 ± 27 — 1.85 1.653 309 ± 7 560 ± 34 — 3.28 — 155 + 1 221 ± 4 — 2.03 — 0.700 195 ± 10 300 ± 11 — 2.37 — 183 + 4 253 ± 15 — 1.91 — Mean: 3.16 1.96 2.14 1.85 Standard deviation: 1.23 — 0.22 — Skipjack tuna 1.069 140 ± 7 263 + 22 — 3.53 — 78 + 28 202 ± 4 — 6.71 — 0.582 214 + 8 282 + 6 386 ± 7 1.87 1.87 148 + 2 237 + 9 275 + 20 2.56 1.35 0.425 173 ± 6 226 + 11 — 1.71 — 122 + 6 191 + 5 — 2.45 — 0.448 179 + 9 211 + 8 — 1.39 — 134 + 8 197 + 3 — 2.16 — 0.629 113 ± 6 217 + 7 — 3.71 — 145 + 3 212 + 3 — 2.14 — Mean: 2.44 1.87 3.74 1.35 Standard error: 1.09 — 2.11 — 29 FISHERY BULLETIN: VOL. 85, NO. 1 Table 2. — Range of mean (±SD) excess muscle temperatures in paralyzed tuna. Tuna Kawakawa Yellowfin Skipjack 20°C 25°C 30°C Mean SD Mean SD Mean SD 0.7(±0.2)-1.9(±0.3) 0.3(±0.2)-1.4(±0.1) 0.2(±0.1)-0.6(±0.1) 0.5(±0.2)-1.5(±0.3) 0.0(±0.1)-0.7(±0.2) 0.0(±0.2)-1.0(±0.2) ^1.0(0.1) 0.2(±0.1)-0.4(±0.1) ^0.6(±0.2) 'Only one fish survived long enough to provide useful data. stops all movements, all fish showed a decrease in metabolic rate after injection. The decrease ranged from 10 to 52% (mean 36%). The directly measured SMR's from four aholehole and four rainbow trout paralyzed with Flaxedil are given in Table 3. DISCUSSION Adequacy of Directly Measured SMR Muir et al. (1965) provided a regression equation for SMR versus weight for aholehole adapted to 23 °C freshwater, based on extrapolation of swim- ming speed-metabolic rate curves back to zero swim- ming speed. The predicted freshwater SMR's based on their regression equation was increased by 75% to account for the higher osmoregulatory costs of seawater adapted animals (Nordlie and Leffler 1975). No correction was made for temperature. As shown in Table 3, in all cases but one, the directly measured SMR's are close to the SMR's based Muir et al.'s data when corrected for seawater adapted animals. With respect to rainbow trout, in all cases but one, directly measured SMR's are within one standard deviation of the SMR's obtained by extra- polation to zero swimming speed for rainbow trout at 15°C obtained by Bushnell et al. (1984). There- fore, directly measuring SMR's in Flaxedil-para- lyzed aholehole and rainbow trout yields data that are similar to data obtained by the more widely used method of determining SMR by extrapolation of a swimming speed-metabolic rate curve back to zero swimming speed. Tropical tuna species such as yellowfin, skipjack, and kawakawa will survive in a swimming tunnel for only short periods of time. Although other methods to control svdmming speed (such as weight- ing and fin clipping, Dizon and Brill 1979; Boggs 1984) have been tried, they have met with only hmited success. Therefore, direct measurement of SMR's using Flaxedil-paralyzed animals is for now the only way to obtain these data for tropical tuna species. As the data from aholehole and rainbow trout show, direct measure of SMR's using para- lyzed animals yields results similar to that obtained by the more commonly used method of extrapolating swimming speed-metabolic rate curves back to zero swimming speed. The heart rates ( -i- 1 SE , at 25°C) observed in this study were 230 + 20, 206 + 36, 132 + 17/min for skip- jack tuna, kawakawa, and yellowfin tuna, respec- tively. These heart rates are higher than those observed for lightly anesthetized skipjack tuna (Stevens 1972), and are 60 and 39% higher than heart rates measured in skipjack and yellowfin tunas (respectively) that have been immobilized by spinal blockade with lidocaine and are force ventilated (un- publ. obs.). The heart rates measured in Flaxedil- paralyzed animals are, however, within the range exhibited by free-swimming skipjack tuna (80-240 beats/min, Kan wisher et al. 1974). The higher heart Table 3. — Standard metabolic rate of aholehole and rainbow trout. Oxygen consumption (mg 0, kg-' h -') ± SD Aholehole Rainbow trout Weight (g) Measured Predicted SMR SMR' Weight (g) Measured SMR Predicted SMR2 65.5 80.9 91.2 108.5 264(±12.2) 118 146(+16) 113 ^1 35/1 27( + 9.9/8.8) 111 ^114/162(+ 11.1/12.0) 106 289 401 403 568 53.3( + 3.8) 78.8( + 7.0) 60.5(±10.9) 55.5(±9.9) 82.5 + (27.4) 82.5 + (27.4) 82.5 ±(27.4) 82.5 ±(27.4) 'Based on Muiretal. (1965) and corrected for saltwater adapted fish based on Nordlie and Leffler (1975). 2From Bushnell et al. (1984), for 250-350 g fish adapted to 15°C. No corrections for the vtreight dependence of Sf^R were provided. 3SK/IR determinations made approximately 20 h apart. 30 BRILL; STANDARD METABOLIC RATES OF TROPICAL TUNAS rates observed in paralyzed tunas may be due to the vagolytic action of Flaxedil (Grollman and Grollman 1970). However, as the data from aholehole and rain- bow trout show, estimating SMR using animals paralyzed with Flaxedil and by extrapolation of swimming speed-metabolic rate curves back to zero swimming speed yield similar results. Effect of Body Size and Acute Temperature Change on SMR In Figure 1, it appears that the SMR's of yellow- fin tuna are lower than those of skipjack tuna and kawakawa. However, based on the 95% confidence intervals, the heights of the regression lines (at mean body weights) are not significantly different from each other. Based on the 95% confidence inter- vals, the weight exponents of the regression equa- tions for kawakawa, yellowfin tuna, and skipjack tuna also are not significantly different over the size ranges tested (Fig. 1). In other words, the effect of body weight on the SMR is not significantly differ- ent among the three tuna species. The exponent in the allometric equation describing the effect of body size on the SMR of other teleosts ranges from ap- proximately 0.65 to >1 (Winberg 1956; Fry 1957; Beamish 1964; Beamish and Mookherjii 1964; Glass 1969; Brett 1972). The lower values of the exponents for tunas indicate that the weight specific SMR^ (i.e., mg 02/(g-h)) of tunas decreases more rapidly as body size increases than it does for other teleosts. Gooding et al. (1981) also estimated the SMR of skipjack tuna. When converted to the same units used in this study (SMR in mg 02/h and W in kg), the relationship they found for the effect of body weight on SMR was SMR = 234 1^119. The exponent greater than one means that they predict the weight specific SMR to increase with in- creasing body size. As shown in Figure 1, Gooding et al.'s predicted SMR's are lower than mine for small fish, but exceed my estimates above approx- imately 2.5 kg body weight because of the large weight exponent. To estimate SMR, Gooding et al. (1981) used a multiple linear regression equation of the logarithm ^If the allometric equation to describe the effect of body size on whole body standard metabolic rate (SMR) is SMR = aW^, then the corresponding equation to describe weight-specific SMR ver- sus body weight is SMRIW = aW^-IW or SMR' = aW'-i; where SMR' = weight-specific SMR, W = body weight, and a and 6 are fitted parameters. of metabolic rate versus swimming speed and the logarithm of body weight, and then extrapolated back to zero swimming speed. Their data and extra- polations were based on several groups of different- sized fish swimming at voluntary speeds in a tank respirometer. This methodology is not equivalent to the more conventional one of estimating SMR based on swimming speed-metabolic rate curves that are constructed by forcing one fish, swimming in a tun- nel respirometer, to undergo stepwise increases in swimming speed during which the fish remains for at least 1 h at each speed (Brett 1972). Furthermore, Gooding et al. (1981) expressed swimming speeds in body lengths per second. Boggs (1984) has shown that this will cause appreciable bias when fitting multiple linear regression equations because the effect of the body size on active metabolic rate is different at different swimming speeds. The Effect of Acute Temperature Change on SMR and Heart Rate As shown in Table 1, the Qio's (effect of tempera- ture) for the SMR's of skipjack tuna, yellowfin tuna, and kawakawa are the same. They are also close to the Qio's for SMR's of other teleost species sub- jected to acute temperature change (Qio = 2.16, Moffitt and Crawshaw 1983; Qio = 2.10, Boehlert 1978), and for the effect of temperature on SMR where fish were acclimated to each test tempera- ture (Qio = 2.48, Ott et al. 1980; Qio = 1.82-2.83, Duthie 1982). This result was not expected since studies on the effect of temperature change on the metabolic rate of isolated red and white muscle samples (Gordon 1968, 1972a, 1972b), volitional swimming speed (Dizon et al. 1978), and prehminary work on active metabolic rate of skipjack tuna showed all three to be unaffected by temperature. Comparing the metabolic rate (1,052 mg 02/h, from Gooding et al. 1981) of a 2.0 kg skipjack tuna at its minimum swimming speed (1.4 body lengths/ s) to its directly measured SMR (608 mg 02/h, from Brill 1979), shows that the SMR constitutes 58% of the minimum swimming metabolic rate. Because skipjack tuna's SMR constitutes a large fraction of their metabolic rate at minimum swimming speeds and increases as temperature increases, whereas swimming metabolic rate and volitional swimming speed do not, increases in muscle efficiency (i.e., in- creases in thrust developed by the caudal propeller per unit of O2 uptake), reductions in hydrodynamic drag (perhaps due to reduction in water viscosity), or unknown physiological adjustments must occur 31 FISHERY BULLETIN: VOL. 85, NO. 1 when ambient temperature increases to keep active metabolic rate temperature independent. The effect of water temperature (20°-25°C) on heart rate was variable (Qio's ranged from 6.71 to 1.82). The mean values (±95% confidence intervals) of 3.74 (±1.9), 2.56 (±0.35), 2.14 (±0.17) for skip- jack tuna, yellowfin tuna, and kawakawa, respec- tively, are not significantly different from each other and are close to the Q^q (2-3) found for the effect of temperature on the heart rate of lingcod, Ophio- don elongatus, (Stevens et al. 1972). Why Are The SMR's of Tunas So High? Also shown in Figure 1 is the SMR-body weight relationship for sockeye salmon at 20°C, taken from Brett and Glass (1973). Even with the differences in the slopes of the lines, it is still apparent that tunas have remarkably high SMR's. In the follow- ing paragraphs, I argue (as did Stevens and Neill 1978; Stevens and Dizon 1982) that tunas are "energy speculators", gambling high rates of energy expenditure against high rates of energy return. I also hypothesize that tunas' physiology and anatomy have evolved to increase maximum sustainable (i.e., aerobic) metabolic rates (MMR's) and that high SMR's are an inevitable consequence of this ability. In other words, high SMR's are a result of anatom- ical and physiological adaptations (primarily large gill surface areas) associated with high MMR's. Tunas have high MMR's and high SMR's, whereas sluggish bottom-dwelling flatfish (e.g., Platichthys Jleusus) have low MMR's and low SMR's (Duthie 1982). Active fish like salmon have MMR's and SMR's intermediate between these two extremes (Brett 1972). Advantages of High Maximum Metabolic Rates Tunas live in the open ocean, an environment which provides no shelter and where patches of forage are widely scattered (Sund et al. 1981). In this environment, high sustainable swimming speeds (i.e., high MMR's) enable tunas to travel quickly be- tween food patches and to search large volumes of water in the least amount of time. Also, tunas have been shown to have very high rates of digestion (Magnuson 1969), which is advantageous for species that must be able to fully exploit a food patch when- ever one is found. Since digestion is an energy con- suming process, high rates of oxygen delivery and blood flow are required for high rates of digestion. Because the pelagic environment provides tuna no place to hide and rest while repaying an oxygen debt, the ability to quickly metabolize lactate is also advantageous. High MMR's therefore allow tuna to rapidly repay an oxygen debt when one is accum- ulated. Tuna's only defense against predators such as blue marlin, Makaira nigricans, is presumably a burst of maximum (i.e., anaerobic) swimming. Prey capture by tunas also must involve some high speed swimming. Coulson (1979) has argued that the ability to achieve high rates of anaerobic glycolysis allows vertebrate ectotherms to successfully com- pete with vertebrate endotherms, which are capable of much higher rates of aerobic metabolism. How- ever, most vertebrate ectotherms, whether terres- trial or aquatic, must spend long quiescent periods to metabolize lactate (Coulson et al. 1977). Yet tunas have the ability to metabolize some of the highest muscle lactate levels ever recorded in vertebrates in only a few hours (Barrett and Connor 1964; Hochachka et al. 1978). Other teleosts may take as long as 24 h to recover from severe exercise even though they accumulate lower white muscle lactate concentrations (Black et al. 1961; Wardle 1978). Tunas' vascular heat exchangers appear to also aid the rapid movement of lactate from the white muscle where it is produced to the red muscle where it is presumably metabolized (Stevens and Carey 1981). Although using different terminology, McNab (1980) citing terrestrial vertebrates and Pauly (1981) citing fishes, both argue that given certain con- straints, high MMR's are advantageous because rates of somatic and gonadal growth are dependent upon rates of delivery of oxygen and substrate to the tissues. Indeed, Pauly (1981) has shown that the growth rates of fishes are proportional to, and per- haps controlled by, gill surface area. Furthermore, he suggests that it is maximum rate of oxygen delivery to the tissues, rather than food supply, that limits growth rates and that species like tunas, which have the largest gill surface areas, have the highest growth rates. Koch and Wieser (1983) have shown that fish reduce activity levels during periods of gonadal growth. Tunas cannot make this trade off. For tunas, it is probably necessary to maintain a high rate of activity during gonadal synthesis which, in turn, requires respiratory and cardiovas- cular systems capable of delivering oxygen and metabolic substrates to the tissues at high rates. Adaptations of Tunas For Achieving High Maximum Metabolic Rates In a series of studies on the MMR's in land mam- mals (see Taylor and Weibel 1981, and the papers 32 BRILL: STANDARD METABOLIC RATES OF TROPICAL TUNAS that follow), pulmonary diffusing capacity, mito- chondrial volume and capillary density in muscles were shown to be limiting factors in achieving high MMR's. From these studies Weibel et al. (1981) pro- posed that, at maximum rates of aerobic metabo- lism, there is no excess capacity at any level in the respiratory chain. In other words, to achieve high MMR's, a complete series of anatomical/physiologi- calAjiochemical adaptations must be present. And, as shown in Table 4, these adaptations are present in tunas. Table 4.— Adaptations of tunas for high nnaximum metabolic rates. Large gill surface areas Thin secondary lamella in the gills High hematocrit, high hemoglobin levels (i.e., high blood O2 carrying capacity) High maximum cardiac output Elevated muscle temperatures High muscle myoglobin levels High muscle mitochondrial density High muscle capillary density High muscle aerobic enzyme activity levels Muir and Hughes 1969 Muir and Brown 1971 Klawe et al. 1963; Jones et al. 1986 Poupa and Lindstrom 1983 Stevens and Neill 1978 Stevens 1982 George and Stevens 1978 Stevens and Carey 1981 George and Stevens 1978 Hulbert et al. 1979 Hulbert et al. 1979 Guppy et al. 1979 One of tunas' adaptations for high MMR's are gills with large respiratory surface areas. However, high rates of oxygen uptake are inexorably linked with high osmoregulatory costs, since gills that permit high rates of oxygen uptake must also permit high rates of water and ion movements. This is especially true in marine fishes like tunas where seawater and blood osmolality are approximately 1,000 and 400 mosm, respectively (Bourke 1983). Rao (1968), Farmer and Beamish (1969), Nordhe and Leffler (1975), and Furspan et al. (1984) estimated that the cost of osmoregulation can account for 27 to 50% of the SMR. The gills are a main osmoregulatory effector organ (Evans 1979), and Daxboeck et al. (1982) found that gill tissue respiration alone can account for 27% of the SMR in trout. The SMR, therefore, is obviously strongly influenced by osmo- regulatory cost, which in turn is strongly influenced by gill surface area. Ultsch (1973, 1976) came to a similar conclusion after finding that the SMR's of aquatic (i.e., gill breathing) salamanders were con- trolled by respiratory (i.e., gill) surface area. Muir and Hughes (1969) measured the total sec- ondary lamellar gill surface (i.e., respiratory) area in skipjack tuna, yellowfin tuna, and bluefin tuna, Thunnus thunnus. They found total secondary lamellar areas for 1 kg tunas to be an order of magnitude or more larger than 1 kg bass or roach. Also, they found gill areas were proportional to body weight and the exponent to be 0.85 for the combined data from the three tuna species. This exponent is significantly different from the exponents I found for the effect of body weight on SMR's. It appears that in tunas, the SMR is not strictly determined by secondary lamellar surface area, although high osmoregulatory costs are most likely the main cause of tunas' high SMR's. Also, the difference between the effct of body size on SMR and gill respiratory area implies that larger tunas have greater scope of activity than smaller fishes, as has been shown to occur in other teleosts (Hughes 1984). ACKNOWLEDGMENTS I wish to thank Christofer H. Boggs, Peter G. Bushnell, Terry Foreman, Carol Hopper, Robert Olson, and E. Don Stevens for reviewing this paper and providing useful criticisms and suggestions; Robert E. Bourke for obtaining aholehole and rain- bow trout; and David Jones for lending the respir- ometer for aholehole. LITERATURE CITED Barrett, I., and A. R. Connor. 1964. Muscle glycogen and blood lactate in yellowfin tuna, Thunnus albacares, and skipjack, Katsuwonus pelamis, following capture and tagging. Inter-Am. Trop. Tuna Comm. Bull. 9:219-268. Beamish, F. W. H. 1964. Respiration of fishes with special emphasis on standard oxygen consumption. II. Influence of weight and tempera- ture on respiration of several species. Can. J. Zool. 42: 177-188. Beamish, F. W. H., and P. S. Mookherjii. 1964. Respiration of fishes with special emphasis on standard oxygen consumption. I. Influence of weight and tempera- ture on respiration of goldfish, Carassius auratus L. Can. J. Zool. 42:161-175. Bennett, A. F., and J. A. Ruben. 1979. Endothermy and activity in vertebrates. Science (Wash., D.C.) 206:649-654. Black, E. C, A. C. Robertson, and R. R. Parker. 1961. Some aspects of carbohydrate metabolism in fish. In A. W. Martin (editor), Comparative physiology of carbohy- drate metabolism in heterothermic animals, p. 89-124. Univ. Washington Press, Seattle. BOEHLERT, G. W. 1978. Changes in the oxygen consumption of prejuvenile rockfish, Sebastes diplaproa, prior to migration from the sur- face to deep water. Physiol. Zool. 51:56-67. BOGGS, C. H. 1984. Tuna bioenergetics and hydrodynamics. Ph.D. Thesis, Univ. Wisconsin-Madison, Madison, 198 p. Bourke, R. E. 1983. Causes of mortalities in skipjack tuna (Katsuwonus pelamis) following capture stress. M.S. Thesis, Univ. 33 FISHERY BULLETIN: VOL. 85, NO. 1 Hawaii, Honolulu, 82 p. Brett, J. R. 1965. The relation of size to rate of oxygen consumption and sustained swimming speed of sockeye salmon (Oncorhynchus nerka). J. Fish. Res. Board Can. 22:1491-1501. 1972. The metabolic demand for oxygen in fish, particularly salmonids, and a comparison with other vertebrates. Respir. Physiol. 14:151-170. Brett, J. R., and N. R. Glass. 1973. Metabolic rates and critical swimming speeds of sock- eye salmon {Oncorhynchus nerka) in relation to size and temperature. J. Fish. Res. Board Can. 30:379-387. Brill, R. W. 1979. The effect of body size on the standard metabolic rate of skipjack tuna, Katsuwonus pelamis. Fish. Bull., U.S. 77:494-498. Brill, R. W., and A. E. Dizon. 1979. Red and white muscle fibre activity in swimming skip- jack tuna, Katsuwonus pelamis (L.). J. Fish Biol. 15:679- 685. BusHNELL, P. G., J. F. Steffensen, and K. Johansen. 1984. Oxygen consumption and swimming performance in hypoxia-acclimated rainbow trout Salmo gairdneri. J. Exp. Biol. 113:225-235. Carey, F. G., and R. J. Olson. 1982. Sonic tracking experiments with tunas. Collective Volume of Scientific Papers, International Commission for the Conservation of Atlantic Tunas (SCRS-1981) 17:458-466. COULSON, R. A. 1979. Anaerobic glycolysis: The Smith and Wesson of the heterotherms. Perspect. Biol. Med. 22:465-479. CouLsoN, R. A., T. Hernandez, and J. D. Herbert. 1977. Review. Metabolic rate, enzyme kinetics in vivo. Comp. Biochem. Physiol. 56A:251-262. Davis, J. C, and J. N. Cameron. 1971. Water flow and gas exchange at the gills of rainbow trout, Salmo gairdneri. J. Exp. Biol. 54:1-18. Daxboeck, C, p. S. Davie, S. F. Perry, and D. J. Randall. 1982. Oxygen uptake in a spontaneously ventilating, blood- perfused trout preparation. J. Exp. Biol. 101:35-45. Dizon, A. E., and R. W. Brill. 1979. Thermoregulation in tunas. Am. Zool. 19(l):249-265. Dizon, A. E., R. W. Brill, and H. S. H. Yuen. 1978. III. Correlations between environment, physiology, and activity and the effects on thermoregulation in skipjack tuna. In G. D. Sharp and A. E. Dizon (editors). The physio- logical ecology of tunas, p. 233-259. Acad. Press, N.Y. Duthie, G. G. 1982. The respiratory metabolism of temperature-adapted flatfish at rest and during swimming activity and the use of anaerobic metabolism at moderate swimming speeds. J. Exp. Biol. 97:359-373. Evans, D. H. 1979. Osmotic and ionic regulation by freshwater and marine fishes. In M. A. Ali (editor). Environmental physiology of fishes, p. 93-122. NATO Advanced Study Institute, Series A: Life Sciences 35. Farmer, G. J., and F. W. H. Beamish. 1969. Oxygen consumption of Tilapia nilotica in relation to swimming speed and salinity. J. Fish. Res. Board Can. 26: 2807-2821. Fry, F. E. J. 1957. The aquatic respiration of fish. In M. E. Brown (editor). The physiology of fishes. Vol. 1, p. 1-63. Acad. Press, N.Y. Furspan, p., H. D. Prange, and L. Greenwald. 1984. Energetics and osmoregulation in the catfish, Ictalunxs nebulosu^ and /. punctatus. Comp. Biochem. Physiol. 77A: 773-778. George, J. C, and E. D. Stevens. 1978. Fine structure and metabolism adaptation of red and white muscles in tuna. Environ. Biol. Fishes 3:185-191. Glass, N. R. 1969. Discussion of calculation of power function with special reference to respiratory metabolism in fish. J. Fish. Res. Board Can. 26:2643-2650. Gooding, R. M., W. H. Neill, and A. E. Dizon. 1981. Respiration rates and low-oxygen tolerance limits in skipjack tuna, Katsuwonus pelamis. Fish. Bull., U.S. 79: 31-48. Gordon, M. S. 1968. Oxygen consumption of red and white muscles from tuna fishes. Science (Wash., D.C.) 159:87-90. 1972a. Comparative studies on the metabolism of shallow- water and deep-sea marine fishes. I. White-muscle meta- bolism in shallow-water fishes. Mar. Biol. (Berl.) 15:222- 237. 1972b. Comparative studies on the metabolism of shallow- water and deep-sea marine fishes. II. Red-muscle metabolism in shallow-water fishes. Mar. Biol. (Berl.) 15:246-250. Graham, J. B., and R. M. Laurs. 1982. Metabolic rate of the albacore tuna Thunnus alalunga. Mar. Biol. (Berl.) 72:1-6. Grollman, a., and E. F. Grollman. 1970. Pharmacology and therapeutics. Lea & Febiger, Philadelphia, PA. Guppy, M., W. C. Hulbert, and P. W. Hochachka. 1979. Metabolic sources of heat and power in tuna muscles. II. Enzyme and metabolite profiles. J. Exp. Biol. 82:303- 319. Hochachka, P. W., W. Hulbert, and M. Guppy. 1978. I. 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. Hughes, G. M. 1979. Morphometry of fish gas exchange organs in relation to their respiratory function. In M. A. Ali (editor). Envi- ronmental physiology of fishes p. 33-56. Plenum Press, N.Y. 1984. Scaling of respiratory areas in relation to oxygen con- sumption of vertebrates. Experientia (Basel) 40:519-524. Hulbert, W. C, M. Guppy, B. Murphy, and P. W. Hochachka. 1979. Metabolic sources of heat and power in tuna muscles. I. Muscle fine structure. J. Exp. Biol. 82:289-301. Jones, D. R., R. W. Brill, and D. C. Mense. 1986. The influence of blood gas properties on gas tensions and pH of ventral and dorsal aortic blood in free-swimming tuna, Euthynnus affinis. J. Exp. Biol. 120:201-213. Jones, D. R., and T. Schwarzfeld. 1974. The oxygen cost to the metabolism and efficiency of breathing in trout (Salmo gairdneri). Respir. Physiol. 21: 241-254. Kanwisher, J., K. Lawson, and G. Sundnes. 1974. Acoustic telemetry from fish. Fish. Bull., U.S. 72: 251-255. KiTCHELL, J. F., W. H. Neill, A. E. Dizon, and J. J. Magnuson. 1978. III. Bioenergetic spectra of skipjack and yellowfin tunas. In G. D. Sharp and A. E. Dizon (editors). The physiological ecology of tunas, p. 357-368. Acad. Press, N.Y. 34 BRILL: STANDARD METABOLIC RATES OF TROPICAL TUNAS Klawe, W. L., I. Barrett, and B. M. H. Klawe. 1963. Hemoglobin content of the blood of six species of scom- broid fishes. Nature 198:96. Koch, F., and W. Wieser. 1983. Partitioning of energy in fish: Can reduction of swim- ming activity compensate for the cost of production? J. Exp. Biol. 107:141-146. Magnuson, J. J. 1969. Digestion of food consumption by skipjack tuna {Kat- suwonus pelamis). Trans. Am. Fish. Soc. 98:379-392. McNab, B. K. 1980. Food habits, energetics, and the population biology of mammals. Am. Nat. 116:106-124. MoFFiTT, B. p., and L. I. Crawshaw. 1983. Effects of acute temperature changes on metabolism, heart rate, and ventilation frequency in carp Cyprimis car- pio L. Physiol. Zool. 56:397-403. Mum, B. S., and C. E. Brown. 1971. Effects of blood pathway on the blood-pressure drop in fish gills, with special reference to tunas. J. Fish. Res. Board Can. 28:947-955. MuiR, B. S., and G. M. Hughes. 1969. Gill dimensions for three species of tunny. J. Exp. Biol. 51:271-285. MuiR, B. S., G. J. Nelson, and K. W. Bridges. 1965. A method for measuring swimming speed in oxygen consumption studies on the aholehole Kuhlia sandvicensis. Trans. Am. Fish. Soc. 94:378-382. Nakamura, E. L. 1972. Development and uses of facilities for studying tuna behavior. In H. E. Winn and B. L. 011a (editors). Behavior of marine animals. Current perspectives in research 2: Verte- brates, p. 245-277. Plenum Publ. Corp. Niimi, a. J. 1978. Lag adjustment between estimated and actual physio- logical responses conducted in flow-through systems. J. Fish. Res. Board Can. 35:1265-1269. Nordlie, F. G., and C. W. Leffler. 1975. Ionic regulation and the energetics of osmoregulation in Mugil cephalus Lin. Comp. Biochem. Physiol. 51A:125- 131. Ott, M. E., N. Heisler, and G. R. Ultsch. 1980. A re-evaluation of the relationship between tempera- ture and the critical oxygen tension in fresh water fishes. Comp. Biochem. Physiol. 67A:337-340. Pauly, D. 1981. The relationships between gill surface area and growth performance in fish: a generalization of von Bertalanffy's theory of growth. Meeresforschung 28:251-282. PouPA, 0., and L. Lindstrom. 1983. Comparative and scaling aspects of heart and body weights with reference to blood supply of cardiac fibers. Comp. Biochem. Physiol. 76A:413-421. Rao, G. M. M. 1968. Oxygen-consumption of rainbow trout (Salmo gaird- neri) in relation to activity and salinity. Can. J. Zool. 46: 781-786. Robinson, W. R., R. H. Peters, and J. Zimmermann. 1983. The effects of body size and temperature on metabolic rate of organisms. Can. J. Zool. 61:281-288. Sharp, G. D., and R. C. Francis. 1976. An energetics model for the exploited yellowfin tuna, Thunnus albacares, population in the eastern Pacific Ocean. Fish. Bull., U.S. 74:36-51. Stevens, E. D. 1972. Some aspects of gas exchange in tuna. J. Exp. Biol. 56:809-823. 1982. The effect of temperature on facilitated oxygen diffu- sion and its relation to warm tuna muscle. Can. J. Zool. 60:1148-1152. Stevens, E. D., G. R. Bennion, D. J. Randall, and G. Shelton. 1972. Factors affecting arterial pressures and blood flow from the heart in intact, unrestrained lingcod, Ophiodon elongatus. Comp. Biochem. Physiol. 43A:681-695. Stevens, E. D., and F. G. Carey. 1981. One why of the warmth of warm-bodied fish. Am. J. Physiol. 240:R151-R155. Stevens, E. D., and A. E. Dizon. 1982. Energetics of locomotion in warm-bodied fish. Annu. Rev. Physiol. 44:121-131. Stevens, E. D., and W. H. Neill. 1978. Body temperature relations of tunas, especially skip- jack. In W. S. Hoar and D. J. Randall (editors), Fish physiology. Vol. VII, p. 316-360. Acad. Press, N.Y. SuND, P. N., M. Blackburn, and F. Williams. 1981. Tunas and their environment in the Pacific Ocean: A review. Oceanogr. Mar. Biol., Annu. Rev. 19:443-512. Taylor, C. R., and E. R. Weibel. 1981. Design of the mammalian respiratory system. I. Prob- lem and strategy. Respir. Physiol. 44:1-10. Ultsch, G. R. 1973. A theoretical and experimental investigation of the relationships between metabolic rate, body size, and oxygen exchange capacity. Respir. Physiol. 18:143-160. 1976. Respiratory surface area as a factor controlling the standard rate of 0, consumption of aquatic salamanders. Respir. Physiol. 26-'357-369. Wardle, C. S. 1978. Non-release of lactic acid from anaerobic swimming muscle of plaice Pleuronectes platessa L.: A stress reaction. J. Exp. Biol. 77:141-155. Weibel, E. R., C. R. Tracy, P. Gehr, H. Hoppeler, 0. Mathieu, and G. M. 0. Maloiy. 1981. Design of the mammalian respiratory system. IX Func- tional and structural limits for oxygen flow. Respir. Physiol. 44:151-164. WiNBERG, G. G. 1956. Rate of metabolism and food requirement of fishes. [In Russ.] Nauch. Tr. Beloruss. Gos. Univ. Minsk., 253 p. (Translated, 1960, Fish. Res. Board Can. Trans. Ser. 194, 202 p.) YONEMORI, T. 1982. Study of tuna behavior, particularly their swimming depths, by the use of sonic tags (Cho-onpa hoshiki-ho ni yoru maguro-rui no yuei suishin no kenkyu). Frovi Enyo (Far Seas) Fish. Res. Lab. (Shimizu) Newsl. 44:1-5. (Translated by Tamio Otsu, 1982, 7 p.. Translation No. 70; available Southwest Fish. Cent. Honolulu Lab., Natl. Mar. Fish. Serv., NOAA, Honolulu, HI 96822-2396). Zar, J. H. 1968. Calculation and miscalculation of the allometric equa- tion as a model in biological data. Bioscience 18:1118- 1120. 35 VARIATIONS IN THE BLOOD CHEMISTRY OF THE LOGGERHEAD SEA TURTLE, CARETTA CARETTA Peter L. Lutz and Ann Dunbar-Cooper^ ABSTRACT The natural blood chemistry profile of loggerhead sea turtles living in Cape Canaveral waters was deter- mined over a 3-year period. Overall plasma osmotic pressure, potassium, and magnesium values were similar to those reported for other reptiles, sodium and chloride was much less than for sea snakes. Plasma calcium and glucose values were among the lowest of any reptile. Osmotic pressure, sodium, and potassium values increased during the warmer months. Chloride and in particular magnesium, glucose, and hematocrit levels were comparatively constant. Calcium and urea values showed wide variations but no seasonal trend was apparent. Changes in urea concentrations closely tracked those of osmotic pressure. Blood lactate values from trawl-captured sea turtles were 10-80 times higher than those from quiescent sea turtles and calculations suggest that at least 20 hours is required for full recovery. The complex changes in blood chemistry observed reflect changes in the sea turtle physiology and biochemistry; signifi- cant changes from normal in plasma magnesium, potassium, and hematocrit could be useful indicators of hibernation in sea turtles. For any animal a knowledge of the normal pattern and changes in blood chemistry can be related to its physiological state and can also be used to identify chronic and pathological conditions. With the excep- tion of sea turtles, there are many studies and reviews on seasonal changes in the blood chemistry of reptiles (Dessauer 1970; Duguy 1970; Gilles- Baillien 1974; Minnich 1982). Since there is an urgent need to understand the ecological physiology of these endangered and threatened species, this lack of information on sea turtles is undoubtedly due to the logistical difficulties of long-term sampling of a wild marine population. The year-round presence of large numbers of loggerhead sea turtles, Caretta caretta, in and around the Port Canaveral ship channel provided a rare opportunity to study the monthly changes that occur in the biology of this little understood group of animals. Such a study was rendered all the more urgent by finding, in the winter of 1978, num- erous black stained and apparently torpid turtles lodged in the mud of the ship channel (Carr et al. 1980). It was suggested that the loggerhead sea turtle was able to survive prolonged exposure to cold seawater temperatures (less than 15°C) by partial- ly lodging in the mud at the bottom of the Port Canaveral ship channel and by going into a state of winter dormancy or apparent hibernation (Carr et al. 1980; Ogren and McVea 1982). If this hypothe- ^Rosenstiel School of Marine and Atmospheric Science, 4600 Rickenbacker Causeway, Miami, FL 33149. Manuscript accepted October 1986. FISHERY BULLETIN: VOL. 85. NO. 1. 1987. sis were correct, it would mean that the Cape Canaveral ship channel was serving as a hibernacu- lum for this endangered species and the identifica- tion of features that could confirm hibernation in these loggerhead sea turtles was of some practical importance. For this purpose, a study of blood chem- istry is particularly apt. There is abundant evidence of significant changes in certain blood constituents in hibernating mammals (Fisher and Manery 1967; Soivio and Kristoffersson 1974; Al-Badry and Taha 1983) and there are a few studies showing similar changes in some reptiles (e.g., freshwater turtles, Hutton and Goodnight 1957; lizards, Haggag et al. 1965). The purpose of this study was to establish the nor- mal seasonal changes in blood chemistry that occur throughout the year in the Cape Canaveral popula- tion of loggerhead sea turtles and from this base of data to identify, if found, those animals that are in a state of hibernation. METHODS Selected National Marine Fisheries Service (NMFS) shrimp trawl turtle surveys of the Port Canaveral ship channel were accompanied by the authors from December 1978 to August 1982. On board ship the activity levels of newly caught logger- head sea turtles were observed, and body tempera- ture, weight, and sex recorded. Blood samples were taken from freshly captured sea turtles, using a heparinised syringe, from a venous sinus on the 37 lateral dorsal region of the neck (Bentley and Dunbar-Cooper 1980). Some sea turtles were also resampled after 3, 4, and 5 h on deck. The hemato- crit was measured on board ship immediately after taking the sample. Blood cells were centrifuged and plasma stored on ice for transport to Miami. The plasma was frozen until used (-4°C). The items that were measured are as follows: Plasma osmotic pressure, using a Wescor vapor pressure osmometer; sodium and potassium concen- trations, by flame emission spectrophotometry; calcium and magnesium, by atomic absorption spec- trophotometry using appropriate standards (Lutz 1972); chloride, using an Aminco Chloridometer; and urea and glucose, using enzymatic kit techniques (Sigma). The blood chemistry values reported are means + SD. Statistical differences between groups were determined with Student's t test, and level of sig- nificance was set at P < 0.05 for all compari- sons. FISHERY BULLETIN: VOL. 85, NO. 1 RESULTS Hematocrit The hematocrit levels of the loggerhead sea turtle were remarkably constant and were not influenced by season (Table 1). The range of values was 28-48%, and the mean 35.4% is very similar to that found by Dessauer (1970) for the same species (32%). The sea turtles caught in December 1978 were a strik- ing exception to this uniformity with very much lower mean hematocrits (15%) and one individual having a value as low as 5%. Loggerhead sea turtles resampled 3-5 h after cap- ture showed some interesting changes. Four animals resampled after 3 h on deck showed an average in- crease in hematocrit of 10.4% (±14.89), for five animals after 4 h the average increase was 4.8% (±4.7), and for four animals after 5 h the hemato- crit had decreased on average 15.2 (± 13.3) from the initial value. The reason for this change is not clear. Table 1.— Blood chemistry values of loggerhead sea turtles trapped by shrimp trawl in the Port Canaveral of samples in parentheses. Groups that differ significantly Date °C Na K Ca Mg CI Dec. 21.0 150.2 ± 13.45 3.7 + 0.81 1.09 + 0.50 •1.7 -h 0.46 109.3 ± 11.36 1979 (5) (4) (5) (5) (4) Jan. 18.0 138.1 + 13.07 3.3 -1- 0.49 1.4 + 0.48 2.19 -1- 0.30 105.9 + 8.40 1980 (11) (11) (13) (13) (11) Feb. 16.0 129.2 + 12.7 3.05 + 0.63 1.18 + 0.86 1.84 ± 0.46 110.5 + 11.71 1980 (8) (8) (8) (8) (8) Mar. 18.0 142.2 + 8.95 3.32 + 0.63 1.04 + 0.38 2.2 + 0.56 108.6 + 4.36 1980 (6) (6) (7) (7) (5) Apr. 19.0 140.9 + 4.7 3.58 + 0.37 1.03 + 0.16 2.24 + 0.42 112 + 10.42 1980 (4) (4) (4) (4) (4) May 23.0 139.3 3.5 — — 112 1980 June 24.0 139.5 + 5.87 3.5 + 0.40 0.86 ± 0.50 2.29 + 0.41 103.5 + 9.02 1980 (6) (6) (5) (5) (5) July 25.0 143.1 + 5.55 3.9 + 0.83 1.08 + 0.77 2.38 + 0.39 •114.4 + 3.34 1980 (7) (7) (7) (7) (7) Aug. 28.0 139.5 — 1.58 1.75 121 1980 Sept. — 145.2 ± 10.95 3.8 + 0.65 1.4 + 0.3 2.23 + 0.46 ••121.8 + 16.4 1980 (5) (5) (5) (5) (7) Nov. 24.0 ** 162.2 ± 7.93 **4.18 ± 1.49 ••2.18 + 0.325 1.49 ± 0.77 107.0 + 3.25 1981 (5) (5) (9) (5) (9) Feb. 19.0 142.07 ± 21.42 4.145 ± 0.69 1.863 + 0.456 1.976 + 0.303 102.7 + 10.58 1982 (4) (4) (5) (5) (10) Mar. 18.0 152.1 ± 14.2 •*4.17 + 0.454 1.40 + 0.765 2.31 + 0.66 108.86 + 9.88 1982 (5) (5) (10) (5) (10) May 24.5 •165.5 ± 4.41 •4.08 ± 2.04 2.09 + 0.41 2.00 ± 0.62 110.3 + 5.62 1982 (3) (3) (8) (4) (9) June 27.0 ** 168.9 + 4.09 "4.6 -1- 0.48 1.78 + 0.93 1.93 + 0.55 108.2 + 13.0 1982 (3) (3) (10) (5) (9) Aug. — •159.3 ± 15.4 5.14 ± 0.83 1.69 ± 0.96 2.65 + 0.69 ••117.3 + 6.57 1982 (5) (5) (8) (5) (9) 38 LUTZ and DUNBAR-COOPER: BLOOD CHEMISTRY OF LOGGERHEAD SEA TURTLE It is too short a time lapse for an erythropoesis response particularly as sea turtle red cells can have life spans of 600-800 d (Altman and Brace 1962); but it is possible that the loggerhead sea turtle has a con- siderable ability to store and release blood cells on demand. fall (Table 1), but the individual range, 102.7-131.2 mM, was much narrower than that found for sodi- um. The narrow excursion suggests that chloride is under comparatively tight control. The population average, 107.2 ± 18.80 mM(n = 86) is very similar to that reported by Dessauer (1970) (110 mM). Sodium Plasma sodium increased as the year advanced with minimal values found in February of each year followed by a gradual rise to maximums in late sum- mer and early fall (Table 1). However, the range of values is very wide (Table 1); the lowest for an in- dividual was 105.5 mM, the highest 173.0 mM. The mean sodium concentration for the whole popula- tion was 145.03 ± 13.80 mM (n = 82). Chloride Like sodium the highest values were found in the Potassium The field data showed little change in the absolute potassium levels (Table 1). The population mean 3.82 ± 0.764 mM (n = 70) is considerably lower than that found for salt water adapted Malaclemys (8.8 mM, Dunson 1970). Minimal values were found in early spring (February) and a gradual rise was seen as summer advanced. Calcium Calcium values ranged quite widely over the sam- pling period and the results are fairly scattered and ship channel, December 1979 -August 1982. Unless otherwise stated from January 1980 group (* P < 0.05; ** P < 0.01). units are m/W. Mean + SD, number Date °C Lactate Glucose Urea Osmotic pressure mOsm Hematocrit o/o Dec. 1979 21.0 — — 10.9 -1- 6.64 (5) 315.2 -1- 15.2 (5) ••15.67 + 8.10 (4) Jan. 1980 18.0 — — 6.19 + 2.82 (11) 301.5 + 22.6 (11) 34.9 + 2.70 (9) Feb. 1980 16.0 — — 4.56 + 1.86 (8) 300.4 + 20.6 (7) 36.1 + 7.18 (7) Mar. 1980 18.0 — — 5.02 + 1.44 (5) •324.8 + 9.06 (6) 35.0 + 5.19 (8) Apr. 1980 May 1980 19.0 23.0 — — *'15.5 + 9.72 (4) 9.22 340.5 + 16.09 (4) 334 37.8 + 7.63 (4) 35.3 + 5.90 (4) June 1980 24.0 — — 2.28 + 2.46 (6) 305.3 + 19.1 (6) 34.3 + 4.35 (9) July 1980 25.0 — — 5.73 + 6.66 (7) •329.4 + 26.27 (7) 35.5 ± 6.87 (6) Aug. 1980 28.0 — — 6.8 327 32.9 -1- 4.8 (2) Sept. 1980 — — — 7.8 + 2.61 (5) •332.6 + 9.81 (5) 34.3 + 4.89 (7) Nov. 1981 24.0 — — 9.43 + 4.56 (9) "330.1 + 19.12 (9) 35.3 + 5.27 (9) Feb. 1982 Mar. 1982 19.0 18.0 3.51 + 0.27 (4) 3.42 -1- 1.39 (4) 1.17 + 0.367 (4) 0.98 + 0.468 (3) 6.78 + 2.04 (10) 5.55 + 2.51 (10) 309.0 + 28.8 (10) 309 + 9.36 (10) 33.7 -1- 5.85 (10) 36.1 ± 5.71 (10) May 1982 June 1982 27.0 24.5 3.58 + 0.07 (3) 1.31 4.41 + 3.8 (10) 4.45 + 2.11 (9) •329.4 + 20.3 (10) 314.8 + 10.9 (9) 31.8 + 3.60 34.08 + 5.46 (6) Aug. 1982 — **16.2 + 8.1 (3) 1.12 + 0.18 (3) 6.19 + 4.49 (9) •343.3 + 23.1 (9) 33.0 + 2.68 (7) 39 FISHERY BULLETIN: VOL. 85, NO. 1 no pattern is discernable, peaks being found in November and May (Table 1). The lowest plasma calcium for an individual was 0.19 mM and the high- est 4.90 mM. For the whole population the mean is 1.53 ± 0.76 mM{n = 115). Magnesium The population mean is 2.10 ± 0.542 mM(n = 88). The lowest and highest values for individuals were 0.96 and 3.80 mM respectively, a smaller excursion than that found for calcium. It appears that plasma magnesium levels are under comparatively tight control. Osmotic Pressure The osmotic pressure values showed the greatest absolute excursion, individuals ranging from 258 to 360 mOsm. The lowest monthly means were found from January to March of each year (Table 1). The average osmotic pressure for the whole population was 321.3 ± 24.10 mOsm (n = 117). Urea Plama urea values showed the greatest relative range in individuals, 0.4-23.8 mM. Interestingly, the pattern of changes is remarkably similar to that of the osmotic pressure (Fig. 1), suggesting strongly that both are linked in some way. The mean value for the population (6.57 + 5.82 mM, n = 101) is very similar to that reported for the same species (6.0 mM, Dessauer 1970). Glucose In the field blood glucose was remarkably steady at about 1 mM (Table 1), suggesting that blood glu- cose levels are highly regulated. This value is con- siderably lower than that reported earlier for the loggerhead sea turtle (3.3 mM, Dessauer 1970). Lactate For most loggerhead sea turtles the blood lactate concentrations ranged from 3 to 4 mM shortly after capture (Table 1). However, noticeably higher lac- tate values (8.8-16.2 mM) were obtained from sea turtles caught in a single trawl (August 1982). This was possibly the result of more severe trawl stress. Rates of recovery while on deck varied. For 6 in- dividuals, lactate had declined an average of 16.8% from the initial value after 3 h; for 4 sea turtles after 4 h, average lactate had declined 52.6%; and for 4 sea turtles lactate had declined 16.4% after 5 h. DISCUSSION This study examines, for the first time, the month- ly changes in the blood chemistry of a marine turtle. However it must be borne in mind that this is a field study without "controls" and alterations in body chemistry and metabolism could be due to intrinsic biological rhythms cued to extrinsic factors such as photoperiod or could be directly determined by en- vironmental changes in, for example, temperature. As turtles are ectotherms (with the possible excep- tion of leatherbacks) seasonal changes in tempera- ture will be accompanied by matching changes in body temperature. It was not possible, therefore, to distinguish between temperature effects per se and changes due to annual rhythms acting as Zeit- gebers. Temperature effects are the subject of a separate study (Lutz and Dunbar-Cooper 1984). The total sample number assembled over the course of this study for each blood constituent is very large, as far as we are aware the set is much Figure 1.— Seasonal changes in plasma urea (■) and osmotic pressure (♦) in the loggerhead sea turtle, December 1979 to September 1981 and November 1981 to August 1982. 40 LUTZ and DUNBAR-COOPER: BLOOD CHEMISTRY OF LOGGERHEAD SEA TURTLE larger than any previous study on reptiles, and allows some general comments on the composition of sea turtle blood to be made. The osmotic pressure found in this study, of 321 mOsm, is significantly lower than that found by Schoffeniels and Tercafs (1965) for the loggerhead sea turtle (465 mOsm), and the value 408 mOsm quoted by Dessauer (1970). It is, however, similar to that found for other reptiles including crocodiles and freshwater turtles (about 290 mOsm, Dessauer 1970). The observation, therefore, that marine turtles have relatively high osmotic pressures (Min- nich 1982) would appear unwarranted. Plasma sodi- um and chloride concentrations are so much less than those reported for the sea snake Pelamis plattis caught in the wild (Na = 210 mM, CI = 167 mM, Dunson and Elhart 1971) that phylogenetic con- siderations may be involved. Potassium values found in this study (3.8 mM) fall within the range charac- teristic of other reptiles (3-6 mM, Dessauer 1970) arguing against the observation that sea turtles have peculiarly high potassium concentrations (Dessauer 1970). Magnesium values are similar to those re- ported for other turtles, including sea turtles (Min- nich 1982) but calcium is rather low (1.5 mM this study, 3.1 mM quoted by Dessauer 1970). As men- tioned above, the hematocrit, glucose, and urea data agree with earlier estimations. The changes observed in this study are of con- siderably physiological significance if internal ionic concentrations are used to regulate the activity of ion sensitive metabolic pathways (Lutz 1975) par- ticularly if some salts, such as Na, K, and CI, have highly perturbing effects on enzyme function (Hochachka and Somero 1984). The contrast between the behaviour of sodium and chloride is of interest. Sodium shows a wide excur- sion in values throughout the year with several peaks and troughs but tends to rise as the year pro- gresses. Compared with sodium, chloride is rela- tively constant and the minor changes that do occur do not match in time with those of sodium. Although both ions account for most of the plasma osmotic pressure (78.5%), neither by themselves was sig- nificantly related to osmotic pressure. Changes in either sodium or chloride do not determine changes in osmotic pressure. Lance (1976) found likewise that plasma sodium showed a much wider excursion than plasma chloride in the cobra Naja naja, but in this species only a single summer sodium peak was seen. It is noteworthy that the lowest sodium values were found in the coldest month (February 1980, Table 1). A winter decrease in plasma sodium has been found for several freshwater turtle species, particularly those hibernating (Gilles-Baillien 1974). We found that plasma potassium increased as the summer progressed and laboratory data suggests that this may be a temperature related phenome- non (Lutz and Dunbar-Cooper 1984). A rise in plasma potassium during the warmer months has also been observed in the lizard Trachysaurus rugosus and the terrapin Malaclemys centreta (Gilles-Baillien 1973). However, the pattern is not constant; a fall has been seen in Varanus grisus (Haggag et al. 1965) and no change seen in Pseu- demys scripta (Hutton and Goodnight 1957). Although highly variable, calcium values are low. There are several peaks per year but no consistent pattern was seen. It is very likely, however, that the changes in blood calcium reflect changes in physi- ology. High values have been found in some reptiles during vitellogenesis (as high as 34 mM, Lance 1976) and calcium has also been found to rise to extra- ordinary high levels in cold torpoid freshwater turtles (Jackson et al. 1984). The seasonal changes in magnesium were much smaller over this study suggesting that wide excur- sions from this narrow range would be indicative of exceptional circumstances. One of the most remarkable findings of this study is the parallel sweeps in the patterns shown by blood urea and osmotic pressure. As far as we are aware such a phenomenon has not been reported before. It is not simply a matter of changes in urea concen- trations causing changes of osmotic pressure since the magnitude of the urea changes are much less than those of osmotic pressure. An integrated re- sponse is called for; possibly the perturbing effects of increasing osmotic pressure are compensated by heightened urea levels (Yancey et al. 1982). In loggerhead sea turtles, blood urea concentration would not appear to be diet determined since we observed that captured loggerhead sea turtles held at RSMAS, which were all fed the same food, had widely different urea values (range 3-21 mM). In- terestingly, the field group with outstandingly high urea levels (April 1980) were all males. The unchanging glucose levels demonstrate a high degree of conservatism. Seasonal changes in blood glucose have been observed in alligators with higher levels in the summer (Coulson and Hernandez 1980). In P. scripta, on the other hand, blood glucose in- creases during winter (Hutton and Goodnight 1957). The hematocrit was also remarkable in its con- stancy, contrasting with other reptiles where sea- sonal changes in hematocrit have been recorded; typically as an increase during winter (Duguy 1970; (Jilles-Baillien 1974). In contrast, the very low values 41 FISHERY BULLETIN: VOL. 85, NO. 1 for December 1978 stand out strongly as a set by themselves and indicate some special condition. The lactate values are of interest in that they give an index of the stress of capture in the trawl net. For quiescent loggerhead sea turtles kept in cap- tivity at RSMAS, blood lactate is very low (0.2-0.4 mM). The initial blood lactate values obtained on deck were, by contrast, 10-80 times higher (3.2-16.2 mM, Table 1). Down to at least 3-4 mM, the rate of lactate recovery for sea turtles held on board was clearly concentration dependent (Fig. 2,P < 0.01). If the rate did not further decline, then it would take about 20 h for full recovery of the least stressed sea turtles in this study (those with initial blood lactate values of 3-4 mM). If the rate of decline continued to be concentration dependent then the recovery time would be much greater. Unfortunately, since no lethargic loggerhead sea turtles were found during this study, one of its prin- ciple objectives, the identification of the state of hibernation in sea turtle, was not realizable. This occurred because South Florida has been blessed with warm winters since 1975 and water tempera- tures have not been lower than 15°C in the Cape Canaveral region. Nevertheless, the wealth of in- formation on the seasonal changes in blood chemistry we now have is sufficient to enable a clear diagnosis of hibernation in sea turtles if and when animals in this condition are found. Magnesium is a prime candidate for such a purpose, since this study identifies the normal range for plasma mag- nesium throughout the year. Substantial increases in blood magnesium have been seen in many hiber- nating animals, including mammals and reptiles (Haggag et al. 1965; Soivio and Kristoffersson 1974; Al-Badry et al. 1983). Significant changes in plasma sodium and potassium have also been associated with hibernation in reptiles (Gilles-Baillien 1974). The normal range of potassium is so narrow that extraordinarily high values should be easily de- tected. Substantial increases in blood lactate have been associated with cold torpor in several fresh- water turtles (Jackson et al. 1984); however, as we have seen elevated blood lactate can occur with stress. And finally hematocrit is of high interest since significant changes in hematocrit, both in- creases and decreases, have been widely reported in hibernating reptiles (Gilles-Baillien 1974). With a single exception, hematocrit was remarkably steady over the course of this survey, and perhaps significantly, the exception occurred in the coldest month encountered. Perhaps the very low hema- tocrits found in December 1979 were part of a prep- aratory condition for hibernation. ACKNOWLEDGMENTS This work was supported by National Marine Fish- eries Service contract FSE 49-3-12-40. We wish to thank Fred Berry and Larr Ogren for their encour- agement and assistance. LITERATURE CITED Al-Badry, K. S., and H. D. Taha. 1983. Hibernation-hypothermia and metabolism in hedgehogs. Changes in water and electrolytes. Comp. Biochem. Physiol. 74A:435-441. r E Q) 4-1 u o 30 1 25 20- 15- 10 5 ♦ -( 0^ ( D 1 Rote of lactate decline mM.hr -I Figure 2.— Rate of blood lactate decline compared to initial lactate con- centration for shrimp trawl trapped loggerhead sea turtles held on board ship for 3-5 h. For the lowest data point w = 11, SDs are illustrated. For other data points w = 1. 42 LUTZ and DUNBAR-COOPER: BLOOD CHEMISTRY OF LOGGERHEAD SEA TURTLE Altman, p. D., and K. C. Brace. 1962. Red cell life span in the turtle and toad. Am. J. Physiol. 203:1188-1190. Bentley, T. B., and a. Dunbar-Cooper. 1980. A blood sampling technique for sea turtles. Contract No. Na-80-GE-A-00082 for the National Marine Fisheries Service. September 1980. Carr, a., L. Ogren, and C. McVea. 1980. Apparent hibernation by the Atlantic loggerhead turtle off Cape Canaveral, Florida. Biol. Conserv. 19:7-14. CouLSON, R. A., and T. Hernandez. 1980. Alligator metabolism — Studies on chemical reactions in vivo. Permagon Press, Lond. Dessauer, H. C. 1970. Blood chemistry of reptiles: physiological and evolu- tionary aspects. In C. Cans and T. S. Parsons (editors), Biology of the reptilia, p. 1-72. Morphology 3. Acad. Press, N.Y. DUGUY, R. 1970. Le sang des reptile. Traite de Zoologie 14:474- 498. DUNSON, W. A. 1970. Some aspects of electrolyte and water balance in three estuarine reptiles, the diamondback terrapin, American and "salt water" crocodiles. Comp. Biochem. Physiol. 32:161- 174. DuNSON, W. A., AND G. W. Elhart. 1971. Effects of temperature, salinity and surface water flow on distribution of the sea snake, Pelamis. Limnol. Oceanogr. 16:845-853. Fisher, K. C, and J. F. Manery. 1967. Water and electrolyte metabolism in heterotherms. In K. C. Fisher, et al. (editors). Mammalian hibernation, Vol. Ill, p. 235-279. American Elsevier, N.Y. Gilles-Baillien, M. 1973. Hibernation and osmoregulation in the diamondback terrapin Malaclemys centrata centrata (Latreille). J. Exp. Biol. 59:45-51. 1974. Seasonal variations in reptiles. In M. Florkin and B. T. Scheer (editors), Chemical zoology. Vol. IX, p. 353-376. Acad. Press, N.Y. Haggag, G., K. a. Raheem, and R. Kahlil. 1965. Hibernation in reptiles. I. Changes in blood electrolytes. Comp. Biochem. Physiol. 16:457-465. Hochachka, p. W., and G. N. Somero. 1984. Biochemical adaptation. Princeton Univ. Press, Princeton, NJ. Hutton, K. E., and C. J. Goodnight. 1957. Variations in the blood chemistry of turtles under ac- tive and hibernating conditions. Physiol. Zooi. 30:198-207. Jackson, D. C, C. V. Herbert, and G. R. Ultsch. 1 984 . The comparative physiology of diving in North Ameri- can freshwater turtles. II. Plasma ion balance during pro- longed anoxia. Physiol. Zool. 57:632-640. Lance, V. 1976. Studies on the annua! reproductive cycle of the female cobra, Naja naja. Seasonal variation in plasma inorganic ions. Comp. Biochem. Physiol. 53A:285-289. LuTZ, P. L. 1972. Extracellular spaces and composition of various tissues of the perch. Comp. Biochem. Physiol. 41A:187-193. 1975. Adapative and evolutionary aspects of the ionic con- tent of fishes. Copeia 1975:369-373. LuTZ, P. L., and a. Dunbar-Cooper. 1984. Effect of forced submergence and low seawater tem- perature on the physiology and behavior of sea turtles. Final report to National Marine Fisheries Service, Contract FSE 81-125-60. MiNNICH, J. E. 1982. The use of water. /mC. Cans (editor), Biology of the reptilia. Vol. 12, p. 325-395. Acad. Press, N.Y. Ogren, L., and C. McVea, Jr. 1982. Apparent hibernation by sea turtles in North American waters. In K. A. Bjorndal (editor), Biology and conserva- tion of sea turtles. Smithson. Inst. Press Conserv., Wash. D.C., 583 p. Schoffeniels, E., and R. R. Tercafs. 1965. Adaptation d'un reptile marin. Caretta caretta L. a I'eau douce et d'un reptile d'eau douce, Clemmys leprosa L. a I'eau der mer. Ann. Soc. R. Zool. Belg. 96:1-8. SOIVIO, A., AND R. Kristoffersson. 1974. Changes in plasma main inorganic ion concentrations during the hibernation cycle in the hedgehog {Erinaceus europaeus L.). Ann. Zool. Fenn. 11:131-134. Yancey, P. H., M. E. Clark, S. C. Hand, R. D. Bowlus, and G. N. Somero. 1982. Living with water stress: Evolution of osmolyte systems. Science 217:1214-1222. 43 EFFECTS OF AIR EXPOSURE ON DESICCATION RATE, HEMOLYMPH CHEMISTRY, AND ESCAPE BEHAVIOR OF THE SPINY LOBSTER, PANULIRUS ARGUS Gregory K. Vermeer^ ABSTRACT Desiccation rates and hemolymph pH, lactic acid and ammonia concentrations of spiny lobsters, Panulirus argus, exposed in air for up to 2 hours were measured. Desiccation rates were faster in smaller lobsters. During a 2-hour exposure, hemolymph lactic acid levels increased more than 11 times, pH decreased more than one-half unit, and ammonia concentrations nearly doubled. Ex-posure-induced changes in hemo- IjTiiph parameters occurred most rapidly in the first 30 minutes and began to level off by 2 hours. Lobsters exposed for 2 hours, then reimmersed for 24 hours, survived and had normal hemolymph chemical values. However, 75% of the reimmersed spiny lobsters had a delayed or absent tail-flip escape response; most individuals also exhibited diminished antennal defensive motions. Results suggest that desiccation and hemolymph chemical changes, caused by exposure, do not directly cause mortality, but rather induce secondary physiological damage, manifested as aberrant defensive and escape behavior. The South Florida fishery for spiny lobster, Panu- lirus argus (Latreille, 1804), uses sublegal (<76 mm carapace length, CL) lobsters, locally called shorts, as living attractants in traps for legal-sized lobsters. Shorts used in this manner are customarily held in wooden boxes on deck until replaced in traps. Aerial exposure ranges from a few minutes to several hours but is typically about 1 h (Bill Moore-). Hunt et al. (1986) reported an average 26.3% mortality rate after 4 wk for lobsters that had been exposed between V2 and 4 h and estimated that 600,000 to 3.7 million shorts die annually as a result of handling and exposure. Because this mortality is incurred by sublegal lobsters which otherwise would soon con- tribute to legal harvest, economic loss to the fishery is considerable, perhaps as high as $9.0 million annually. This study examines desiccation rate, hemolymph chemistry, and escape behavior of spiny lobster to document physiological and behavioral changes in- duced by air exposure. The relationship between these changes and mortality is discussed. MATERIALS AND METHODS One hundred seventy intermolt spiny lobsters, averaging 80.2 mm CL (range, 56.7-120.7 mm), were collected from traps at the Atlantic reefs south 'Florida Department of Natural Resources, Bureau of Marine Research, 100 Eighth Avenue S.E., St. Petersburg, FL 33701. ^Bill Moore, lobster fisherman, pers. commun. December 1984. of Marathon, FL, in the Florida Keys. Approximate- ly 26 lobsters at one time were allowed to acclimate for a minimum of 2 d in a 800 L (179 x 76 x 60 cm) outdoor fiberglass tank. The tank was fully shaded by three plywood sections which could be removed individually, allowing easy access while minimizing disturbance. Flow-through water circula- tion was maintained by a pump drawing approx- imately 3,600 L/hour from a clean, well-oxygenated canal. Complete water exchange occurred every 15 min. Periodic canal water samples had oxygen con- centrations of 5-7 ppm and no detectable ammonia or lactic acid. There were resident spiny lobsters in the canal. Shelter inside the tank was provided by a double layer of two-hole cinder blocks (39.5 x 19.5 x 19.5 cm) centered and aligned parallel to the long axis of the tank. This arrangement of blocks allowed for easy removal of spiny lobsters by the antenna-tug technique, described later. In rare instances when a spiny lobster evaded capture on the first attempt, sampling of that animal was postponed for at least 24 h. This was necessary because repeated tail- flips depressed hemolymph pH (unpubl. data). Spiny lobsters were not fed during confinement or held longer than 10 d. Both sexes were used equally. Desiccation Rate Spiny lobsters were randomly selected from the acclimation tank at 10 min intervals, marked for in- Manuscript accepted September 1986. FISHERY BULLETIN: VOL. 8.5, NO. 1, 1987. 45 FISHERY BULLETIN: VOL. 85, NO. 1 dividual identification, and alternately assigned to either an exposure or control group. After marking, control spiny lobsters were weighed to the nearest 0.1 g and promptly placed inside a shaded, wood-slat fish box two-thirds sub- merged inside the acclimation tank. Weights were also recorded at 1 and 2 h. Excess water clinging to the exoskeleton and inside the branchial cham- bers was removed prior to each weighing by holding the spiny lobster around the carapace in a head down position and gently moving it through a short down- ward arc six times. Exposed spiny lobsters were marked and weighed as above, but were held in a fish box located in a fully shaded outdoor area. Evaporative water loss was indicated by weight decrease over time. During the period when desiccation experiments were performed (late March to early May 1984), relative humidity was 61-72%, air temperature 22-30 °C, wind speed 10 km/h or less, and cloud cover ranged from clear to lightly scattered or hazy. Ex- periments were not performed on very wet or windy days to avoid excessive variation in desiccation rates between experiments. Hemolymph Chemistry To assess effects of exposure on hemolymph chem- istry, spiny lobsters were air-exposed in fish boxes for ¥2, 1, or 2 h as previously described. Control spiny lobsters were removed directly from the ac- climation tank. Hemolymph sampling was via cardiac puncture. A 1.6 mm (yie-in) hole drilled through the dorsal carapace directly over the heart allowed easy hypo- dermic removal of 8-10 mL of hemolymph. There is no suitable chemical method to prevent hemo- lymph clotting (Young 1972). At ambient tempera- ture, spiny lobster hemolymph forms a tough rub- bery clot within seconds. Prompt cooling of the hemolymph by immersion of the syringe in an ice water bath (4°C, 60 s) inhibited clotting long enough to prepare subsamples for pH, ammonia, and lactic acid analysis. All hemolymph samples were collected between the hours of 10:00 and 16:00 and analyzed the same day. Intervals between netting and completion of hemolymph removal were 70 s or less, thus mini- mizing trauma associated with handling and cardiac puncture. Since net confinement reduced struggling, spiny lobsters were not removed from the net for hemolymph sampling unless access to the dorsal carapace was restricted. In preliminary experi- ments, repetitive handling and sampling of controls depressed hemolymph pH values. Consequently, each spiny lobster was sampled only once in ex- periments reported here. Hemolymph pH was deter- mined by a digital pH meter with a calomel micro- electrode. Hemolymph subsamples (2 mL) and a 7.0 buffer solution were chilled to 4 °C in a second ice water bath before recording pH. Blood pH at 4°C probably varies from in vivo pH at ambient tem- perature, but this was an essential concession to retard clot formation. Anaerobic, radiometer-type pH measurements were also impossible due to clot- ting. However, care was taken to minimize hemo- lymph air contact since changes in CO9 equilibrium can alter pH values. Truchot (1975) reported the pH of crustacean blood exposed to air without mixing varies little from anaerobically obtained samples. Serum was prepared by injecting the remaining 6-8 mL of chilled hemolymph into a 15 mL tissue grinder, then gently grinding for 1-2 min until the clotting hemolymph was liquified. The still cool serum was then refrigerated in capped test tubes for subsequent ammonia analysis. Ammonia was measured using the Conway micro- diffusion method (Conway and Byrne 1933) with modifications suggested by Seligson and Seligson (1951). With this method, ammonia from a 0.5 mL blood sample was diffused onto an acidified glass rod inserted inside a microdif fusion cell. Microdif- fusion cells were rotated for 50 min to facilitate diffusion, then the rods were washed off with 5 mL of Nessler's reagent. Intensity of color developed in Nessler's reagent, corresponding to ammonia con- centration, was measured in a colorimeter at 420 nm. All samples were done in duplicate as were blanks and two concentration standards (10 ;. 0.05). Predation rate varied among densities of eelgrass (Fig. 3). Medium density seagrass provided the best refuge from predation with only 9% eaten per day (A'' = 45). A mean of over 19% per day was eaten in low-density (N = 47) and high-density {N = 44) grass sites. A Dunn's Multiple Comparison test (Hollander and Wolfe 1973) was used to analyze the predation-vegetation density data from July through September, excluding October because no predation occurred in eelgrass during that month. Predation rates in low and high densities were found to be sig- nificantly greater (P < 0.05) than in medium-density eelgrass. Eelgrass biomass in low, medium, and high den- sity 0.062 m^ plots (Table 1) was found to be sig- nificantly different in a one-way analysis of variance (P < 0.001). Scheffe contrasts found that the mean dry weight of medium-density plots was significantly higher than low-density and significantly lower than high-density eelgrass plots. Table 1.— Mean dry weights (g/0.062 m^) of vegetation from experimental plots. ** Significantly different at the P < 0.01 level. Density Mean SD P Low Medium High 12.19 43.24 79.04 5.24 17.07 11.47 * * * • * * DISCUSSION These data confirm results from other experimen- tal studies of predation on decapod crustaceans (Heck and Thoman 1981; Orth and van Montfrans 54 WILSON ET AL.: JUVENILE BLUE CRAB SURVIVAL IN EELGRASS 3 4% BOt E 3 30 32 V. 28% 1 ^ a ^ m 17 % ^ 14% # ^ ^ 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-9091-100 Carapace Width (mm) Figure 2.— Blue crab body size (carapace width (CW)) and risk of predation. Hatched bars in- dicate number of individuals tethered at that size and open bars indicate number of tethered crabs eaten. Low Med VegeloUon Density High Figure 3.— The effect of eelgrass density on predation rates. Histograms are mean rates of predation from July through October on sand and at each eelgrass density. Vertical bars are + / - one standard error. 1982) and amphipods (Stoner 1982) that describe the importance of seagrasses as protective cover for prey. They clearly show that eelgrass provides refuge from predation and increased survival for juvenile blue crabs compared to that on adjacent unvegetated sand substrates. Rates of predation on blue crabs within the three densities of vegetation, however, did not conform to patterns previously established, where predation on crustacean epifauna is inversely proportional to vegetation biomass (Stoner 1982; Leber 1985). In this study, risk of predation was lowest in interme- diate densities rather than in high-density eelgrass. Savino and Stein (1982) found that attack rates by largemouth bass on bluegills dramatically de- clined with increasing density of artificial vegeta- tion, and capture rates by the predators were lower in vegetation than on bare substratum. Epifaunal amphipods and caridean shrimp also suffer lower rates of predation at high densities of vegetation (Nelson 1979; Stoner 1980; Coen et al. 1981; Leber 1985). These studies and others (Vince et al. 1976; Crowder and Cooper 1982; Minello and Zimmerman 1983) indicate that above-ground vegetation biomass reduces a visual predator's search and capture effi- ciencies and that vegetation may also provide a matching background in which epifaunal prey may hide (Endler 1978; Orth et al. 1984). The root and rhizome mat of seagrasses may also lower search and capture efficiency of predators (Orth 1977; Blundon and Kennedy 1982b; Peterson 1982), but in addition a high-density root mat may reduce the ability of hard-bodied prey to bury and hide in the substratum. For example, Brenchley (1982) found that the burrowing ability of decapods in dense eelgrass root mats was reduced or pre- vented, and Bertness and Miller (1984) found that fiddler crabs, TJca pugnax, preferred to construct burrows in intermediate densities of salt marsh roots. 55 FISHERY BULLETIN: VOL. 85, NO. 1 Juvenile blue crabs, unlike epifaunal caridean shrimp or amphipods, utilize below-ground refuges in seagrass beds. Our field and laboratory observa- tions suggest that their primary mode of predator avoidance is to bury in the substratum. Orth and van Montfrans (1982) also noted burying behavior of juvenile blue crabs in laboratory experiments that examined predation by adult blue crabs in three den- sities of artificial seagrass and root mat. Their data also suggested mortality of juveniles is lowest in intermediate densities of seagrass. We infer that at low seagrass densities the blue crabs are able to bury in the substratum, but the leaves and root mat of the grass do not reduce detec- tion and capture efficiency of the predators as do intermediate seagrass densities. Furthermore, we suggest that the dense root mat and shoots of high- density seagrass may reduce the ability of blue crabs to bury themselves and that high blade density may reduce the crabs' visual ability to detect predators. Based on our observations the dominant predators on blue crabs appear to be toadfish, Opsanus tau, the American eel, Anguilla rostrata, and other blue crabs. Toadfish are extremely common in the Mana- hawkin grass beds in the summer (June-September) and are known to readily consume brachyuran crabs, including blue crabs (Schwartz and Butcher 1963; McDermott 1965; Wilson et al. 1982; Gibbons and Castagna 1985). In this study, there were instances where, upon recovery of tethers after a predation trial, toadfish had swallowed both the crab and tether and remained on the line, providing confir- mation that toadfish are blue crab predators under field experimental conditions. Gut contents of American eels from the study area contained blue crabs (K. Able, pers. obs.) and Wenner and Musick (1975) found blue crabs to be a major part of the eel's diet. Predation intensity appears to be distributed even- ly over the size classes tested, although there is a trend of lower predation rates on the largest blue crabs (>71 mm CW). However, the sample size is small for these size classes {N = 17) so the estimate of predation on larger crabs may be inadequate. Escape in size has been observed in other inverte- brate prey (Blundon and Kennedy 1982a; Peterson 1982; Wilson 1985) and a similar pattern was ex- pected in this study because large adult blue crabs are found frequently on unvegetated substratum where risk of predation is highest (Heck and Thoman 1984). An additional large predator, the smooth dogfish, Mtcstelus canis, occurs in Barnegat Bay (Tatham et al. 1983) and we suspect it may feed on blue crabs in seagrass meadows at night (Casterlin and Reynolds 1979). Mustelus can grow to 1.5 m (Hildebrand and Schroeder 1928) and preys on blue crabs in eelgrass beds (Bigelow and Schroe- der 1953). Hence, predation by smooth dogfish may account for loss of larger crabs and also suggests that there may be a temporal as well as spatial pat- tern of predation. Researchers have suggested that the value of refuges for juvenile blue crabs and other inverte- brate macrofauna is dependent on the interaction of several factors including species of vegetation, vegetation density, water quality, and type of pre- dator (Heck and Thoman 1984; Orth et al. 1984). The data from these tethering experiments clearly in- dicate that eelgrass serves as protective cover and that eelgrass density is indeed an important factor in determining predation rates on juvenile blue crabs. The unexpected result that crabs in interme- diate densities of eelgrass suffered lower predation rates than those in high densities underscores the complexity of the interactions that determine sur- vival of juvenile blue crabs. ACKNOWLEDGMENTS Support for this study was provided by the New Jersey Department of Environmental Protection, Marine Fisheries Administration, and the Academy of Natural Sciences of Philadelphia. LITERATURE CITED Bertness, M. D., and T. Miller. 1984. The distribution and dynamics of Uca pugnax (Smith) burrows in a New England salt marsh. J. Exp. Mar. Biol. Ecol. 83:211-237. Bigelow, H. G., and W. C. Schroeder. 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv. Fish. Bull. 53:1-577. Blundon, J. A., and V. S. Kennedy. 1982a. Mechanical and behavioral aspects of the blue crab, Callinectes sapidus (Rathbun), predation on Chesapeake Bay bivalves. J. Exp. Mar. Biol. Ecol. 65:47-65. 1982b. Refuges for infaunal bivalves from blue crab, Calli- nectes sapidus (Rathbun), predation in Chesapeake Bay. J. Exp. Mar. Biol. Ecol. 65:67-81. Brenchley, G. a. 1982. Mechanisms of spatial competition in marine soft- bottom communities. J. Exp. Mar. Biol. Ecol. 60:17-33. Casterlin, M. E., and W. W. Reynolds. 1979. Diel activity patterns of the smooth dogfish, Mxistelus canis. Bull. Mar. Sci. 29:440-442. Coen, L. D., K. L. Heck, and L. G. Abele. 1981. Experiments on competition and predation among shrimps of seagrass meadows. Ecology 62:1484-1493. Crowder, L. B., and W. E. Cooper. 1982. Habitat structural complexity and the interaction be- tween bluegills and their prey. Ecology 63:1802-1813. 56 WILSON ET AL.: JUVENILE BLUE CRAB SURVIVAL IN EELGRASS Diaz, R. J., and T. J. Fredette. 1982. Secondary production of some dominant macroinverte- brate species inhabiting a bed of submerged vegetation in the lower Chesapeake Bay. In R. J. Orth and J. van Mont- frans (editors), Structural and functional aspects of the biology of submerged aquatic macrophyte communities in the lower Chesapeake Bay. Vol. Ill, p. 95-123. VIMS- SRAMSOE 267. Endler, J. A. 1978. A predator's view of animal color patterns. Evol. Biol. 12:319-364. EPIFANIO, C, AND A. I. DiTTEL. 1982. Comparison of dispersal of crab larvae in Delaware Bay, USA and the Gulf of Nicoya, Central America. In V. S. Kennedy (editor), Estuarine comparisons, p. 477-487. Acad. Press, N.Y. Epifanio. C. E., C. C. Valenti, and A. E. Pembroke. 1984. Dispersal and recruitment of blue crab larvae in Dela- ware Bay, USA. Estuarine Coastal Shelf Sci. 18:1-12. Gibbons, M. C, and M. Castagna. 1985. Biological control of predation by crabs in bottom cultures of hard clams using a combination of crushed stone aggregate, toadfish and cages. Aquaculture 47:101-104. Heck, K. L., Jr., and R. J. Orth. 1980. Seagrass habitat complexity, competition and preda- tion in structuring associated fish and motile macroinverte- brate assemblages. In V. S. Kennedy (editor), Estuarine perspectives, p. 449-464. Acad. Press, N.Y. Heck, K. L., Jr., and T. A. Thoman. 1981. Experiments on predator-prey interactions in vege- tated aquatic habitats. J. Exp. Mar. Biol. Ecol. 53:125-134. 1984. The nursery role of seagrass meadows in the upper and lower Chesapeake Bay. Estuaries 7:70-92. Hildebrand, S. F., and W. C. Schroeder. 1928. Fishes of Chesapeake Bay. Bull. U.S. Bur. Fish. 43: 1-388. Hollander, M., and D. A. Wolfe. 1973. Nonparametric statistical methods. John Wiley and Sons, N.Y., 503 p. Johnson, D. R., B. S. Hester, and J. R. McConaugha. 1984. Studies of a wind mechanism influencing the recruit- ment of blue crabs in the Middle Atlantic Bight. Continen- tal Shelf Res. 3:425-437. Kennish, M. J., D. J. Danila, and R. J. Hillman. 1982. Assessment of the blue crab, Callinectes sapidus Rath- bun, in Barnegat Bay. Bull. N.J. Acad. Sci. 27:59-71. Kennish, M. J., J. J. Vouglitois, D. J. Danila, and R. A. Lutz. 1984. Shellfish. In M. J. Kennish and R. A. Lutz (editors). Ecology of Barnegat Bay, p. 171-200. Springer-Verlag, N.Y. Leber, K. M. 1985. The influence of predatory decapods, refuge and micro- habitat selection on seagrass communities. Ecology 66: 1951-1964. LiPPSON, A. J. 1973. The Chesapeake Bay in Maryland - An atlas of natural resources. John Hopkins Univ. Press, Baltimore, MD, 55 P- Macomber, R. T., and D. S. Allen. 1979. The New Jersey submerged aquatic vegetation distribu- tion atlas final report. N.J. Dep., 25 p. McConaugha, J. R., D. F. Johnson, A. J. Provenzano, and R. C. Moris. 1983. Seasonal distribution of larvae of Callinectes sapidus (Crustacea:Decapoda) in the waters adjacent to Chesapeake Bay. J. Crustacean Biol. 3:582-591. McDermott, J. J. 1965. Food habits of the toadfish, Opsanus tau (L.) in New Jersey waters. Proc. Pa. Acad. Sci. 38:64-71. Minello, T. J., AND R. J. Zimmerman. 1983. Fish predation on juvenile brown shrimp, Penaeus aztecus Ives: the effect of simulated Spartina structure on predation rates. J. Exp. Mar. Biol. Ecol. 72:211-231. Nelson, W. G. 1979. Experimental studies of selective predation on amphi- pods: consequences for amphipod distribution and abun- dance. J. Exp. Mar. Biol. Ecol. 38:225-245. Orth, R. J. 1977. The importance of sediment stability in seagrass com- munities. In B. C. Coull (editor), Ecology of marine ben- thos, p. 281-300. Univ. South Carolina Press, Columbia, SC. Orth, R. J., K. L. Heck, Jr., and J. van Montfrans. 1984. Faunal communities in seagrass beds: a review of the influence of plant structure and prey characteristics on predator-prey relationships. Estuaries 7(4A):339-350. Orth, R. J., and J. van Montfrans. 1982. Predator-prey interactions in an eelgrass ecosystem in the lower Chesapeake Bay. In R. J. Orth, and J. van Montfrans (editors). Structure and functional aspects of the biology of submerged aquatic macrophyte communities in the lower Chesapeake Bay, Vol. Ill, p. 81-94. VIMS- SRAMSOE 267. Penry, D. L. 1982. Utilization of a Zostera marina and Ruppia maritima habitat by four decapods with emphasis on Callinectes sapidus. M.S. Thesis, College of William and Mary, Williamsburg, VA, 101 p. Peterson, C. H. 1982. Clam predation by whelks (Busycon spp.): experimen- tal tests of the importance of prey size, prey density and seagrass cover. Mar. Biol. (N.Y.) 66:159-170. Provenzano, A. J., Jr., J. R. McConaugha, K. B. Philips, D. F. Johnson, and J. Clark. 1983. Vertical distribution of the first stage larvae of the blue crab, Callinectes sapidus. at the mouth of the Chesapeake Bay. Estuarine Coastal Shelf Sci. 16:489-499. Saving, J., and R. A. Stein. 1982. Predator-prey interaction between large-mouth bass and bluegills as influenced by simulated, submerged vege- tation. Trans. Am. Fish. Soc. 111:255-266. Schwartz, F. J., and B. W. Dutcher. 1963. Age, growth, and food of the oyster toadfish near Solomons, Maryland. Trans. Am. Fish. Soc. 92:170-173. Stoner, a. W. 1980. The role of seagrass biomass in the organization of ben- thic macrofaunal assemblages. Bull. Mar. Sci. 30:537-551. 1982. The influence of benthic macrophytes on the foraging behavior of pinfish, Lagodon rhomboides (Linnaeus). J. Exp. Mar. Biol. Ecol. 58:271-284. SULKIN, S. D. 1984. Behavioral basis of depth regulation in the larvae of brachyuran crabs. Mar. Ecol. Prog. Ser. 15:181-205. SuLKiN, S. D., W. VAN Heukelem, P. Kelly, and L. van Heukelem. 1980. The behavioral basis of larval recruitment in the crab Callinectes sapidus Rathbun: a laboratory investigation of ontogenetic changes in geotaxis and barokinesis. Biol. Bull. 159:402-417. Tagatz, M. E. 1968. Biology of the blue crab, Callinectes sapidus Rathbun, in the St. Johns River, Florida. Fish. Bull., U.S. 67:17-33. 57 Tatham, T. R., D. L. Thomas, and D. J. Danila. 1983. Fishes of Barnegat Bay, New Jersey. In Michael J. Kennish and Richard A. Lutz (editors), Ecology of Barnegat Bay. Springer-Verlag, N.Y., 396 p. Van Engel, W. A. 1958. The blue crab and its fishery in Chesapeake Bay. Comnn. Fish. Rev. 20(6):6-17. ViNCE, S., I. Valiela, N. Backus, and J. M. Teal. 1976. Predation by the salt marsh killifish Fundulus hetero- clitus (L.) in relation to prey size and habitat structure: con- sequences for prey distribution and abundance. J. Exp. Mar. Biol. 23:255-266. Wenner, C. a., and J. A. MusiCK. 1975. Food habits and seasonal abundance of the American eel, Anguilla rostrata, from the lower Chesapeake Bay. Chesapeake Sci. 16:62-66. FISHERY BULLETIN: VOL. 85, NO. 1 Williams, A. B. 1984. Shrimps, lobsters and crabs of the Atlantic coast of the eastern United States, Maine to Florida. Smithson. Inst. Press, Wash., D.C., 550 p. Wilson, C. A., J. M. Dean, and R. Radtke. 1982. Age, growth rate and feeding habits of the oyster toad- fish, Opsanus tau (Linnaeus) in South Carolina. J. Exp. Mar. Biol. Ecol. 62:251-259. Wilson, K. A. 1985. Physical and biological interactions that influence habitat use of mangrove crabs. Ph.D. Thesis, Univ. Penn- sylvania, Philadelphia, 167 p. Zimmerman, R. J., and T. J. Minello. 1984. Densities of Penaeus aztecus, Penaeus setiferus, and other natant macrofauna in a Texas salt marsh. Estuaries 7(4A):421-433. 58 FISH PREDATION ON JUVENILE BROWN SHRIMP, PENAEUS AZTECUS IVES: EFFECTS OF TURBIDITY AND SUBSTRATUM ON PREDATION RATES Thomas J. Minello, Roger J. Zimmerman, and Eduardo X. Martinez! ABSTRACT Predation on juvenile brown shrimp, Penaeus aztecus, by three species of estuarine fishes was examined in a series of laboratory experiments to determine the effect of turbid water and the presence of a suitable substratum for burrowing. Regardless of the type of substratum, turbid water increased predation by southern flounder, Paralichthys lethostigrm, and decreased predation by Atlantic croaker, Micropogonias undulatus. In both clear and turbid water, the presence of sand, which allowed shrimp to burrow, decreased predation by southern flounder but had no significant effect on feeding rates of Atlantic croaker. There was a significant interaction between the effects of turbidity and substratum on predation by pin- fish, Lagodon rhomboides. Turbid water decreased predation in tanks with hard substrata but had no significant effect in tanks with sand. The presence of sand reduced predation only in clear-water tanks. Burrowing by brown shrimp was reduced in turbid water which may explain this interaction. Overall, the data indicate that both turbid water and a suitable substratum for burrowing may reduce predation on brown shrimp, but the value of these refugia is highly dependent upon the species of predator. Predation by fishes appears to be a major source of mortality of juvenile brown shrimp, Penaeus aztecus Ives, in estuarine nurseries. Brown shrimp spend several months as juveniles in estuaries, and anal- yses of the stomach contents of some estuarine fishes indicate a high incidence of predation on penaeid shrimp (see Minello and Zimmerman 1983 for review). The presence of salt marsh vegetation apparently offers shrimp protection from some of these predators (Minello and Zimmerman 1983; Zimmerman and Minello 1984), but other habitat characteristics that modify or control the extent of predator-related mortality have not been examined. Estuarine systems in the northern Gulf of Mexico are generally characterized by high turbidity and fine-grained sediments owing to an abundant supply of suspended sediment from rivers and a relatively low-energy environment (Chapman 1968; Linton 1968; Folger 1972). Production of penaeid shrimp in these estuaries is high, and the presence of tur- bid water together with suitable substrata for bur- rowing may contribute to productivity by reducing predation. The effect of turbidity on predator-prey inter- actions varies with the organisms examined. In laboratory experiments with the flounder, Platich- 'Southeast Fisheries Center Galveston Laboratory, National Marine Fisheries Service, NOAA, Galveston, TX 77550. thys flesus, Moore and Moore (1976) found that turbid water reduced the ability of the fish to see epibenthic prey and increased the ability of prey to avoid capture. The degree of this effect varied with prey species. Gardner (1981) also found that turbid- ity reduced predation by bluegill, Lepomis macro- chirus, on Daphnia in laboratory aquaria. Boehlert and Morgan (1985), however, found that predation rates of larval Pacific herring, Clupea harengus pallasi, apparently increased up to a point in turbid water. Other work in the laboratory and in fresh- water lakes and streams has shown that turbidity can interact with the activity, behavior, and distri- bution of both predators and prey (Heimstra et al. 1969; Swensen and Matson 1976; DeVore et al. 1980; Gradall and Swenson 1982; Matthews 1984; Sigler et al. 1984), and predation rates in turbid water may be reduced or enhanced (Swenson 1978). Burrowing by prey in the substratum may also af- fect predation rates, and burrowing by the crayfish, Orconectes propinquus, has been shown to reduce predation by smallmouth bass, Micropterus dolo- mieui (Stein and Magnuson 1976). Although experi- mental evidence is lacking, it has frequently been suggested that burrowing by penaeid shrimp func- tions in a similar manner (Williams 1958; Fuss and Ogren 1966; Hughes 1966, 1968a). Diel periodicity in the burrowing behavior of brown shrimp has been Manuscript accepted August 1986. FISHERY BULLETIN: VOL. 85, NO. 1, 1987. 59 FISHERY BULLETIN: VOL. 85. NO. 1 well documented (^^''ickham and Minkler 1975; Lakshmi et al. 1976; Minello and Zimmerman 1983), and this species generally remains beneath the sur- face of the substratum throughout the daylight hours, emerging to forage at night. The objective of this search vi^as to determine whether turbid water and a suitable substratum for burrowing affect predation rates on juvenile brown shrimp. Experiments were conducted in the labora- tory, and predatory fish were southern flounder, Paralichthys lethostigma Jordan and Gilbert, pin- fish, Lagodon rhomboides (Linnaeus), and Atlantic croaker, Micropogonias undulatus (Linnaeus). The effect of turbidity on burrowing by brown shrimp was also examined. METHODS AND MATERIALS Predation Experiments Collection and Handling of Experimental Animals Fish were collected with trawls and seines from Galveston Bay, TX, and held in clear-water tanks without a sand substratum. They were fed live shrimp daily and starved for 24 h before an experi- ment. Total lengths of fish were measured after each experiment, and specimens from a subsample in holding tanks were weighed and measured. A length-weight relationship was calculated and used to estimate weights of experimental fish. Shrimp were collected by trawling 2 to 3 d before each experiment. They were fed daily with pelleted shrimp food but not fed during experiments. Measurements of total length (tip of rostrum to tip of telson) were made on all shrimp placed into experimental tanks and all shrimp removed after an experiment. A length-weight relationship was calculated for each experiment from sub- samples of shrimp and used to estimate individual weights. Experimental Tanks Experiments were conducted in fiberglass tanks (1.75 m X 5.8 m X 0.5 m) located in a building with a white translucent roof which allowed the use of natural photoperiods. Each tank was divided in half by a wall of 1.5 mm mesh fiberglass forming two compartments (1.75 m x 2.9 m) of 5.07 m^ bottom area. A 5 cm layer of washed beach sand (well sorted with a graphic mean grain size of 2.95 "t*; analyzed according to Folk 1980) was placed in four tanks. In four other tanks, approximately 1 mm of sand was used to reduce the contrast between prey and the bottom of the tank. Tanks were filled to a depth of 26 cm with seawater (24-26%o) pumped from the beachfront off Galveston Island. During experi- ments, water temperatures varied among tanks by only 0.5°C, and diurnal ranges are listed in Table 1. Pulverized kaolinite was used to make the water turbid in four tanks (two with sand bottoms and two without sand). Particle size analysis (Folk 1980) in- dicated that the kaolinite was poorly sorted with a graphic mean grain size of 8.82 . A clay slurry was introduced into tanks through a 19 L settling bucket with an outlet hose (5 mm ID) located 5 cm from the bottom. This settling bucket served to remove some of the heavier particles and flocculated aggre- gates from the clay suspension. Each tank contained a small submersible pump (252 L/minute capacity) connected to a discharge pipe which extended along the length of the tank and sprayed water over the surface. This pump together with 12 airstones/tank provided some vertical mixing which helped keep clay particles suspended. Turbidity, light, and temperature were measured at 2-h intervals during each experiment. Turbidity Table 1.— Design and conditions for predator-prey experiments. Experiment Date (1984) Predator density' No. of repli- cates^ Predator size (mm TL) Prey size (mm) Turbidity^ (FTU) Light" (hE s-'m-2) Temperature (°C) Southern flounder 1 Southern May 11 1 2 84-126 30-40 46-30 152 21.0-23.0 flounder II May 15 1 2 82-111 30-40 54-37 73 22.0-24.5 Pinfish 1 May 18 3 4 62-80 30-40 53-36 48 23.0-24.0 Pinfish II Atlantic May 31 3 4 64-75 30-42 64-42 162 17.0-19.0 croaker June 6 3 4 98-117 30-40 58-37 132 26.0-27.5 ^Number of predators per compartment. ^Number of replicate compartments used per treatment combination ^Average initial and final turbidity in turbid tanks over experimental period. 'Average light levels measured in clear tanks over the first 5 h of the experimental period (n 16). 60 MINELLO ET AL.: PREDATION ON BROWN SHRIMP was measured with an HF Instruments DRT-15 turbidimeter- (calibrated with a Formazin stan- dard) and recorded as Formazin Turbidity Units (FTUs). A typical turbidity curve for acclimation and experimental periods is shown in Figure 1, and mean values from turbid tanks for each experiment are listed in Table 1. These turbidities were within the range of values measured over a 2-yr period in the Galveston Bay system (pers. obs.). Clear treatments ranged between 0.1 and 2.4 FTUs. Light levels in each tank were measured 13 cm below the surface of the water with a LI-COR integrating quantum meter (Model LI-188B) equipped with an under- water sensor. This sensor measures radiation in the 400 to 700 nm waveband, and light energy is ex- pressed in microeinsteins (^E s"^ m"^). Due to var- iability in the thickness of the roof over the experi- mental tanks, there were differences among the tanks in incident light reaching the surface of the water. During one experiment, light levels were measured at the water's surface, and these values were considered to be indicative of the differences among tanks during all experiments. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Experimental Design In all experiments, there were two replicate tanks (each divided into two compartments) for each of the four treatment combinations: clear water/no sand, clear water/sand, turbid/no sand, and turbid/ sand. Feeding by fish was restricted to daylight hours. Twenty-four hours before the initiation of an experiment, fish were placed in circular release cages (0.75 m diameter) within experimental com- partments, and clay was then added (200 mg/L) through the settling system to four of the tanks over a 3-h period. Twenty-five brown shrimp were placed in each compartment (4.9 shrimp/m^) approximate- ly 15 h before the start of an experiment. At 0600 h on the day of the experiment, turbidities were measured and additional clay was added to elevate the turbidity levels and reduce variability among the four tanks. The release cages were lifted at 0700 h, and fish were allowed to feed for 12 h. The tanks were drained at the end of the experimental period, and missing shrimp were assumed to be eaten. For each experiment, two control compartments (one turbid and one clear) were stocked only with shrimp to check survival and recovery of prey. The data were analyzed using the mean number PREDATORS PLACED IN RELEASE CAGES 70r ^ PREY PLACED IN TANKS PREDATORS RELEASED REMAINING PREY COUNTED EXPERIMENTAL PERIOD Ol L 0800 1200 1600 2000 2400 0400 0800 1200 1600 2000 TIME (HOURS) Figure 1.— Typical experimental sequence and turbidity (Formazin Turbidity Units) curve for predation experiments. 61 FISHERY BULLETIN: VOL. 85, NO. 1 of shrimp eaten by a predator in a tank over the ex- perimental period as the observation in a two-way analysis of variance (ANOVA). A second ANOVA was also performed on the weight of shrimp eaten per fish. In experiments with pinfish and Atlantic croaker, where both compartments within a tank contained predators, observations from the two compartments were considered to be within tank replicates or subsamples. With southern flounder, only one compartment was used in each tank. This experiment was repeated on a second day, and day was considered a blocking variable in the analysis. Because differences in incident light among tanks could potentially affect predation rates and increase within treatment variability, an analysis of covariance (ANCOVA) was also performed on the data from all experiments using incident light as the covariate. The size range of shrimp available to predators was kept as narrow as possible (Table 1) to avoid problems associated with size-selective predation. In addition, we attempted to keep the distribution of shrimp within this size range similar for all rep- licates. The size-frequency distributions of shrimp placed in the tanks and shrimp removed from the tanks after each experiment were compared to check for evidence of size-selective predation. Turbidity and Burrowing of Brown Shrimp The effect of turbidity on burrowing by juvenile brown shrimp was examined in eight rectangular tanks each with a bottom area of 0.92 m^. Water depth was maintained at 25 cm, and temperature and salinity were adjusted to 25 °C and 25°/oo, respectively. Light was provided through white translucent skylights. Lengths of PVC pipe were in- stalled along the walls on the bottom of each tank. The tanks were filled with washed beach sand to a depth of 5 cm, and the sand surface was approx- imately 5 mm below the top of the PVC pipe. The number of shrimp burrowed was determined using a net composed of fiberglass screen mounted on a wooden frame. The frame was the same width as the tanks and was pushed over the PVC runners along the bottom, passing just above the sand sur- face. Shrimp caught in the net were assumed to be in the water column or on the surface of the sub- stratum. Ten brown shrimp (50-100 mm) were placed in each tank on the day before an experiment. Before sunrise on the day of the experiment, kaolinite was added to four of the tanks through the settling bucket system. Airstones in all tanks provided enough turbulence to keep the clay in suspension. At 1100 h, turbidity and light levels were measured in the center of the water column in each tank, and nonburrowed shrimp were collected. The tanks were then drained, and the burrowed shrimp were recov- ered. The experiment was repeated with different shrimp on a second day, and an ANOVA, with day used as a blocking variable, was performed to test for an effect of turbidity. The percentage of shrimp burrowed in a tank was used as the observation after an arcsin transformation. The accuracy of our col- lecting technique was examined by comparing visual observations of the number of shrimp burrowed in the clear tanks with the catch in the net. All non- burrowed shrimp were captured in six out of seven trials, but one nonburrowed shrimp avoided capture. In one trial, a burrowed shrimp was collected. RESULTS Predation Experiments Data from the two control compartments (one turbid and one clear) used in each experiment in- dicated that mortality of prey was low. Only 1.6% of the 250 control shrimp were not recovered alive. This mortality was considered negligible, and all shrimp not recovered in predation experiments were assumed eaten by predators. The use of a relative- ly narrow size range of prey also appeared to eliminate problems associated with size-selective predation. Comparisons of size-frequency distribu- tions of shrimp introduced into experimental com- partments to those removed following the experi- mental period showed no apparent size-selective predation in any of the experiments. Southern Flounder Predation by southern flounder was highest in tanks with turbid water and without sand substrata (Table 2A). The interaction term in the ANOVA was not significant, and both main effects of turbidity and substratum were significant at the 0.05 level (Table 2B). Predation rates of these fish increased from a mean of 2.2 shrimp/fish in clear water to 4.4 shrimp/fish in turbid water. Predation rates were reduced in the presence of sand from a mean of 4.8 shrimp/fish in tanks without sand to a mean of 1.9 shrimp/fish in tanks with sand. An ANCOVA with incident light and an ANOVA using the weight of shrimp eaten as the observation gave similar results. The mean weight of shrimp eaten, expressed as a 62 MINELLO ET AL.: PREDATION ON BROWN SHRIMP percentage of body weight eaten by the fish over the experimental period, ranged from 6.1% in clear/sand tanks to 24.8% in turbid/no sand tanks. The feeding behavior of southern flounder (84-94 mm TL) on brown shrimp was also observed in aquaria. These fish exhibited a variety of feeding behaviors including active searching for prey on the bottom and in the water column as described by 011a et al. (1972) for summer flounder, Paralichthys den- tatus. Generally, however, the fish remained motion- less on the bottom and waited for potential prey to come within striking distance before attacking. Fish Table 2.— Predation on brown shrimp by southern flounder. A) Number of shrimp eaten per fish over the 1 2-h experimental period for treatment combinations of turbidity and substratum. B) ANOVA results using the number of shrimp eaten per fish as the obser- vation. Turbid Clear Date (1984) Sand No sand Sand No sand May 11 2 4 0 4 4 10 1 4 May 15 1 5 2 1 2 7 3 3 X 2.2 6.5 1.5 3.0 B Source of error df SS F P Turbidity 1 18.06 5.65 0.037 Substratum 1 33.06 10.34 0.008 Turbidity/ substratum 1 7.56 2.36 0.152 Day 1 1.56 0.49 0.499 Error 11 35.19 in the family Bothidae have been classified as primarily visual feeders by de Groot (1971). In our observations, all stalking activity by southern flounder was accompanied by active eye movements, tracking potential prey, which suggested the primary use of vision in prey detection. A study of diel feeding periodicity, similar to that conducted on red drum and Atlantic croaker by Minello and Zimmerman (1983), however, indicated that south- ern flounder could also feed at night even when tanks were enclosed in black plastic to completely eliminate light (unpubl. data). This finding suggests that sensory mechanisms, in addition to vision, can be used by these fish to detect prey. Pinfish In both experiments with pinfish, the largest num- ber of shrimp were eaten in tanks with clear water and without sand (Fig. 2). The ANOVA on the num- ber of shrimp eaten in the first experiment (pinfish I) indicated a significant interaction between turbid- ity and substratum (Table 3). The substratum ap- parently did not affect predation in tanks with turbid water, but in clear water the presence of sand sig- nificantly reduced predation rates (Fig. 2A). In a similar manner, turbidity did not significantly affect predation in tanks with sand substrata, but it did reduce predation rates in tanks without sand. An ANCOVA with incident light and an ANOVA using the weight of shrimp eaten (Table 3) did not alter X iZ I- < lU Q. S cc z tn u. o oc UJ 03 S Z 2 - \ NO SAND \ \ \ \ \ \ \ \ \ SAND \ ^4 V A. PINFISH I 4 _ OC ^ 1 X (O \ X (0 S NO SAND \^ 1 j: 3 - ^v z ^^ 1 UJ 1- < N s UJ • X 2 - (0 u. u. SAND o = 1 UJ ' - ID s 3 Z n B. PINFISH II 1 1 CLEAR TURBID CLEAR TURBID Figure 2.— Mean predation rates on brown shrimp by pinfish in treatment combinations of turbidity and substratum. Vertical Hnes, representing one half of Tukey's oj (Steel and Torrie 1960) on either side of the mean, can be used to compare means at the 0.05 significance level. 63 FISHERY BULLETIN: VOL. 85, NO. 1 Table 3.— ANOVA results from predation experiments with pinfish using the number of brown shrimp eaten per fish as the observa- tion. Probability values are also listed from an ANOVA using the weight of shrimp eaten as the observation. Turb/subs = Turbidity/substratum. Source of error df SS F P P (weight) Pinfish 1 Turbidity 1 23.39 55.2 0.002 0.003 Substratum 1 8.00 18.9 0.012 0.019 Turb/subs 1 6.70 15.8 0.016 0.033 Error 4 1.70 Pinfish II Turbidity 1 1.39 5.4 0.081 0.040 Substratum 1 4.00 15.5 0.017 0.040 Turb/subs 1 1.38 5.3 0.082 0.110 Error 4 1.03 Combined Turbidity 1 18.10 53.1 <0.001 Substratum 1 11.66 34.2 <0.001 Day 1 21.81 63.9 <0.001 Turb/subs 1 7.08 20.7 0.002 Turb/day 1 6.68 19.6 0.002 Subs/day 1 0.34 1.0 0.34 Turb/subs/day 1 1.00 2.9 0.12 Error 8 2.73 the results. Pinfish were voracious feeders eating between 19.8% (turbid/sand) and 57.1% (clear/no sand) of their body weight in shrimp over the 12-h experimental period. Predation rates were probably underestimated in clear-water treatments without sand, since in three out of four of these compart- ments the three pinfish ate all of the available shrimp. The duration of the pinfish II experiment was reduced to 6 h (0700-1300 h) to lower the overall number of shrimp eaten by the predators. Similar trends were apparent in the number of shrimp eaten for each treatment combination (Fig. 2B), but the interaction term (P = 0.082) in the ANOVA was not significant at the 0.05 level (Table 3). The size range of the prey in the second experiment was slightly larger than in pinfish I (Table 1), and variability in the size of shrimp available or small differences in size-selection may have affected our results. Using the weight of shrimp eaten as the observation should reduce this problem, and in this ANOVA (Table 3) both turbidity and substratum were significant ef- fects, but the F-test for interaction had a probabil- ity value of 0.110. To increase the error degrees of freedom and hence the power of the statistical test, the data from both pinfish experiments were combined and ana- lyzed. In one such ANOVA, day was considered to be a blocking variable (no interaction with other fac- tors), and the results on the number of shrimp eaten were similar to those from the pinfish I experiment, showing a significant interaction between turbidity and substratum (P = 0.021). We also analyzed the data in a completely randomized crossed design with day as a main effect (Table 3). In this ANOVA the turbidity/substratum interaction was highly signif- icant, but the turbidity/day interaction was also significant indicating that the effect of turbidity on predation was less during the second experiment. In addition to the shorter duration of pinfish II, overall light levels were higher during this second pinfish experiment (clear sunny day) compared with the first experiment (overcast day) (Table 1). Atlantic Croaker Mean predation rates for Atlantic croaker were highest in clear-water tanks without sand, and rates in all turbid tanks were low (Table 4A). The ANOVAs with both the number (Table 4B) and weight of Table 4.— Predation on brown shrimp by Atlantic croaker. A) Number of shrimp eaten per fish over the 12-h experimental period for treatment com- binations of turbidity and substratum. B) ANOVA results using the number of shrimp eaten per fish as the observation. Probability values from an ANCOVA using incident light as the covariate are also included. A Experimental tank Turbid d 1 Clear Sand No san Sand No sand Tank 1 compartment 1 0 0.7 1.0 4.7 compartment 2 0 0 0.3 0 Tank 2 compartment 1 0.3 0.3 1.0 3.3 compartment 2 0.7 0 0.3 1.0 X 0.2 0.2 0.7 2.2 B Source of error df SS F P P (ANCOVA) Turbidity 1 2.92 1.8 0.251 0.027 Substratum 1 1.26 0.8 0.42E ! 0.108 Turb/subs 1 1.26 0.8 0.42E ! 0.943 Error 4 6.49 64 MINELLO ET AL.: PREDATION ON BROWN SHRIMP shrimp eaten, however, showed no significant treat- ment effects. Overall, Atlantic croaker ate 2.7% of their weight in shrimp over the experimental period. Differences among tanks in incident light apparently increased the within treatment variability in the ex- periment. There was a significant negative linear correlation between incident light and the number of shrimp eaten in the 16 experimental compart- ments (r = -0.51, n= 16,P = 0.046). An ANCOVA using incident light as a covariate lowered the er- ror sum of squares, and the main effect of turbidity became significant (Table 4B). This was the only ex- periment in which variability in incident light among tanks had a major effect on our results. Atlantic croaker appeared to feed more actively at low light levels, but predation rates were higher in clear water than in turbid water. Turbidity therefore did not appear to affect predation by simply reducing the light in the water column. Turbidity and Burrowing of Brown Shrimp Burrowing by juvenile brown shrimp was mea- sured in both clear and turbid water to aid in the interpretation of significant interactions in the predation experiments. The percentage of shrimp burrowed was reduced from a mean of 85.7% in clear-water tanks to 46.9% in turbid tanks (Table 5). In the ANOVA the effect of turbidity was highly significant (P < 0.001). The effect of day was also significant (P = 0.041), and fewer shrimp burrowed on the second day of the experiment. Overall, light levels were lower on the second day, and a similar ANOVA on light measured 13 cm below the surface of the water also showed significant differences related to turbidity (P < 0.001) and day (P = 0.011). The turbidities used (30-47 FTUs) reduced the aver- age light level in the water by 29% compared with values in clear tanks. Burrowing did not appear to be related to shrimp size, and there was no signif- icant difference between the mean length of bur- rowed shrimp compared with nonburrowed shrimp (paired t-test, P > 0.40, 14 df). DISCUSSION Effect of Turbidity on Predation Turbidity reduces predation on prey possessing limited escape capabilities by reducing the visual reactive distance of the predator (Moore and Moore 1976; Vinyard and O'Brien 1976; Gardner 1981). Turbid water should have less of an effect on pre- dation if the predator-prey size ratio is large (Moore and Moore 1976; Vinyard and O'Brien 1976) or if the predator has the ability to use sensory mecha- nisms other than vision to detect prey. The signif- icant decrease in predation rates by pinfish in our experiments may be explained in part by the strict reliance of this predator on vision for prey detec- tion (Minello and Zimmerman 1983) and upon the relatively small predator:prey size ratio (Table 6). Turbidity appeared to have less of an effect on pre- dation by Atlantic croaker. This predator does not depend solely upon vision for prey detection, but also uses olfaction and touch (Chao and Musick 1977). The increased predation rates for southern flounder in turbid water may be related to the ambush feeding tactics of this predator and the effect of tur- bidity on prey behavior. The activity level of brown shrimp increased in turbid water as evidenced by a decrease in burrowing and the frequent observa- tion of actively swimming shrimp in turbid tanks. According to the model of Gerritsen and Strickler (1977), increased prey movement dramatically in- creases encounter rates with slow moving or sta- tionary predators. This effect of prey movement is Table 5.— The effect of turbidity on burrowing by juvenile brown shrimp. All measurements were taken at approximately 11 00 h. Percentages of shrimp burrowed in each tank (generally 10 shrimp/tank) are listed with turbidity levels and light levels in the water column. Turbid tanks Clear tanks > Burrowed Turbidity Light Burrowed Turbidity Light Date (1984) (0/0) (FTU) (mE :s-'m-^) (%) (FTU) (mE :s-'m-^) Aug. 15 56 49 51 90 5.6 89 50 30 66 100 3.2 93 80 42 54 80 5.9 87 50 47 59 Aug. 17 20 39 49 80 3.2 73 20 32 52 90 3.2 76 60 33 51 80 4.1 73 40 36 51 80 2.9 46 X 47.0 38.4 54.1 85.7 4.0 76.7 65 FISHERY BULLETIN: VOL. 85, NO. 1 Table 6. — Summary data on possible factors affecting predatlon rates for the species of predators examined. Predator searching Size ratio Effect on predatlon rates Mode of speed of Substratum Predator feeding (activity) predator:prey^ Turbidity (burrowing) Southern visual and flounder nonvisual low 3:1 increased decreased Atlantic visual and croaker nonvisual high 3:1 decreased no change Pinfish strictly visual high 2:1 decreased decreased 1 measured as total length. reduced as predator speed increases, and changes in prey activity should only have a negligible effect on encounter rates with more active predators such as pinfish and Atlantic croaker. Increased predation rates by fish in turbid water may also be related to the effect of turbidity on the reactive distance and escape behavior of prey. The ability of the predator to detect the prey before the prey detects the preda- tor is dependent upon differences in visual acuity, apparent size, and motion (Cerri 1983; Howick and O'Brien 1983). Although brown shrimp have the ability to visually detect predators and avoid attack, the acuity of the crustacean compound eye is much lower than that of the vertebrate eye (Waterman 1961; Goldsmith 1973), and shrimp do not respond to stationary predators. This last prey characteristic may explain why the southern flounder is a very ef- fective predator on brown shrimp. Effect of Substratum on Predation Rates Juvenile brown shrimp readily burrowed in experi- mental tanks with fine sand substrata, but they could not burrow in tanks without sand. Burrow- ing should reduce the apparent density and avail- ability of brown shrimp to visually feeding predators (Minello and Zimmerman 1984). Predators using olfactory or tactile mechanisms of prey detection, however, may have less difficulty detecting and feeding upon burrowed shrimp. Predation rates for pinfish and southern flounder, both visual feeders, were significantly reduced in tanks with sand sub- strata. Predation rates of Atlantic croaker were not affected by the presence of sand which suggests that burrowing does not protect brown shrimp from this predator. In other clear-water experiments con- ducted in our laboratory with Atlantic croaker (Albrecht et al. 1983^), we have been unable to detect any reduction in predation on brown shrimp related to the presence of sand substrata. This pred- ator does not depend solely on vision to detect prey (Minello and Zimmerman 1983), and Chao and Musick (1977) hypothesized that Atlantic croaker fed mostly by olfaction and touch. These fish also search through the upper layers of the substratum while foraging for food (Roelofs 1954; Chao and Musick 1977), and this behavior may reduce the number of burrowed shrimp. The presence of sand may also affect predation by altering the activity levels of both prey and pred- ator. Increased activity of brown shrimp in tanks without sand may have increased encounter rates with southern flounder in accordance with the model of Gerritsen and Strickler (1977). In addition, south- ern flounder periodically burrow in sand, and 011a et al. (1972) found that burrowed summer flounder did not respond to the presence of prey. Interactions Between Turbidity and Substratum Burrowing by juvenile brown shrimp is reduced in turbid water (Table 5), and in situations where burrowing protects shrimp from predators, an inter- action might be expected between the effects of turbidity and substratum on predation rates. This type of interaction was present in the pinfish ex- periments. Predation rates of pinfish were reduced in the presence of a sand substratum only in clear water; in turbid water predation was not significant- ly affected by substratum. Turbidity reduced pre- dation in tanks without sand, but in tanks with sand substrata the effect of turbid water on feeding by pinfish was apparently attenuated by a reduction in shrimp burrowing and an increase in the number of available prey. In experiments with southern ^Albrecht, C, T. J. Minello, and R. J. Zimmerman. 1983. The role of substrates in predation on brown shrimp (Penaeus aztecv^) by Atlantic croaker {Micropogonias undulatus). NOAA/NMFS Unpublished Report to Laboratory Director, SEFC, Galveston Laboratory, 18 p. 66 MINELLO ET AL.: PREDATION ON BROWN SHRIMP flounder, burrowing also appeared to reduce the number of shrimp eaten, but this reduction occurred in both turbid and clear water as evidenced by the nonsignificant tiirbidity/substratum interaction term (P = 0.152). In fact, the reduction in mean preda- tion rates associated with the presence of a sand substratum was greatest in turbid water, and the positive effect of turbidity on predation appeared greatest in tanks without sand (Table 2A). Further analysis of the effects of turbidity and substratum on the activity of brown shrimp and the feeding behavior of southern flounder would be needed to explain interactions between these variables. Bur- rowing does not appear to protect brown shrimp from predation by Atlantic croaker, and there was no experimental evidence for an interaction between turbidity and substratum. Experiments With Red Drum, Sciaenops ocellatus (Linnaeus) During the course of this study two experiments were also conducted on the effects of turbidity and substratum on predation rates of another fish pred- ator, the red drum (420-592 mm TL). An in depth analysis of these data was not included due to poor survival of shrimp in control compartments (18% of control shrimp died from unknown causes). Control mortalities, however, did not appear to be related to experimental variables, and data obtained in these experiments suggested that predation rates of red drum are not affected by turbid water or the pres- ence of sand substrata. That substratum has no significant effect on predation rates is supported by additional unpublished but well-controlled experi- ments in our laboratory, indicating that burrowing does not protect juvenile brown shrimp from pre- dation by red drum. Yokel (1966) described the feed- ing behavior of these fish which consists of search- ing along the bottom with the head down and lower jaw rubbing along the surface of the substratum. He concluded that this method of feeding would enable the fish to locate animals in shallow burrows. CONCLUSIONS The artificial nature of these laboratory experi- ments certainly must be considered when attempt- ing to interpret the data in relation to natural phenomena. One major advantage of the apparatus used in our experiments was the relatively large size of the experimental enclosures (5.07 m^ bottom area) which allowed the use of prey densities com- monly found in natural populations. The use of these large enclosures, however, made replicating treat- ment combinations more difficult, hence reducing the power of statistical tests. Despite this limitation, general conclusions about relationships between turbidity, substratum, and predation on brown shrimp can be made on the basis of our experimen- tal results. Under certain conditions, turbid estu- arine water should provide juvenile brown shrimp protection from fish predators such as pinfish and Atlantic croaker. Turbidity does not appear, how- ever, to reduce predation by southern flounder on juvenile brown shrimp. The effect of turbidity on predator-prey relationships apparently depends upon the feeding behavior and morphology of pred- ators and on the behavior of the prey. Burrowing into the substratum also appears to protect brown shrimp from some fish predators, and the ability of brown shrimp to burrow is affected by substratum characteristics (Williams 1958; Aldrich et al. 1968; Rulifson 1981). A change from hard shell botton to soft silty mud should enhance burrowing and reduce predation by estuarine fish such as pinfish, south- ern flounder, and perhaps spotted seatrout. Fishes such as Atlantic croaker and red drum, however, are apparently well adapted for feeding upon bur- rowed organisms, and differences in estuarine sedi- ments may not affect predation by these species. Because turbidity and substratum do not appear to alter predation of all fishes in a similar manner, the effects of these habitat characteristics on the mor- tality of juvenile brown shrimp should strongly de- pend upon the dominant fish predators present in an estuarine system. Comparisons of estuaries with regard to their pro- tective capacity for juvenile brown shrimp are com- plicated by interactions among habitat character- istics and their effects. In addition to the type of substratum, light levels (Wickham and Minkler 1975), temperature (Aldrich et al. 1968), and salin- ity (Lakshmi et al. 1976) have been shown to affect burrowing of brown shrimp. Starvation (Hughes 1968a), tidal movements (Hughes 1968b), shrimp size (Eldred et al. 1961; Hughes 1968a; Moctezuma and Blake 1981), and dissolved oxygen (Egusa and Yamamoto 1961) affect burrowing of other penaeids and may have a similar effect on brown shrimp. The presence of rhizomes and roots of estuarine vege- tation may also reduce burrowing by these animals. All of these factors, therefore, can potentially inter- act with predator-related mortality. In our experi- ments, burrowing by brown shrimp was reduced in turbid water, and this had a significant effect on predation rates of pinfish. Interactions that control the presence of protective habitat characteristics are 67 FISHERY BULLETIN: VOL. 85, NO. 1 also common. In low-energy areas, estuarine sys- tems with large amounts of suspended sediments and high turbidities frequently have fine sediments (Guilcher 1967; Folger 1972). Submerged vegeta- tion, shown to offer many crustaceans protection from predators (Nelson 1979; Stoner 1979; Coen et al. 1981; Heck and Thoman 1981), is associated with estuarine areas of low turbidity (Zieman 1982; Thayer et al. 1984), and these beds of submerged vegetation also reduce turbidity (Short and Short 1984) and alter sediment characteristics (Thayer et al. 1984). Determining the protective value of any suite of environmental characteristics, therefore, may be quite complex. Turbidity and sediment characteristics, however, appear to be important factors governing predation rates on juvenile brown shrimp, and anthropogenic modifications of estuarine systems that influence these characteristics may affect shrimp survival. Turbidity levels and patterns of sediment deposition in estuaries are mainly influenced by riverine inputs, tidal properties, and wave action (Postma 1967; Davis 1983), although biological processes are also important (Haven and Morales-Alamo 1972; Biggs and Howell 1984). Modifications of estuarine sys- tems through dredging, channelization, and altera- tion of freshwater inflows, therefore, can impact predator-prey relationships, and such effects should be addressed in evaluating these activities. ACKNOWLEDGMENTS We thank the Director E. Klima and staff of the National Marine Fisheries Service, Southeast Fish- eries Center Laboratory in Galveston for support- ing this research. In particular, C. Albrecht and M. de Diego are acknowledged for their experimental work on Atlantic croaker and red drum. K. N. Baxter provided needed field personnel and equip- ment; G. Zamora, Jr., D. Boss, and D. Emiliani aided in the collection of experimental organisms; and C. Caillouet, T. Williams, and W. Browning provided much of the experimental equipment. H. Wynn of Texas Industrial Minerals Company donated the kaolinite. K. Gilligan, W. Kittrell, and the staff of the Texas Parks and Wildlife Department and F. Schlicht and W. Baker of Houston Lighting and Power Co. aided in the collection of red drum. J. Matis and D. Colby offered statistical advice, and C. Coleman conducted the particle size analyses. Suggestions for improving the manuscript were pro- vided by Z. Zein-Eldin, E. Klima, S. Ray, D. Aldrich, G. Matthews, and N. Rabalais. B. Richardson and D. Patlan helped prepare the final draft. LITERATURE CITED Aldrich, D. V., C. E. Wood, and K. N. Baxter. 1968. An ecological interpretation of low temperature responses in Penaeus aztecus and P. setiferus postlarvae. Bull. Mar. Sci. 18:61-71. Biggs, R. B., and B. A. Howell. 1984. The estuary as a sediment trap: alternate approaches to estimating its filtering efficiency. In V. S. Kennedy (editor). The estuary as a filter, p. 107-129. Acad. Press, N.Y. BOEHLERT, G. W., AND J. B. MORGAN. 1985. Turbidity enhances feeding abilities of larval Pacific herring, Clupea harengus pallasi . Hydrobiologia 123:161- 170. Cerri, R. D. 1983. The effect of light intensity on predator and prey behavior in cyprinid fish: factors that influence prey risk. Anim. Behav. 31:736-742. Chao, L., AND J. A. MUSICK. 1977. Life history, feeding habits, and functional morphology of juvenile sciaenid fishes in the York River Estuary, Vir- ginia. Fish. Bull., U.S. 75:657-702. Chapman, C. 1968. Channelization and spoiling in Gulf coast and South Atlantic estuaries. In J. D. Newsom (editor). Proceedings of the marsh and estuary management symposium, Loui- siana State University, 1967, p. 93-106. Thos. J. Moran's Sons, Inc., Baton Rouge, LA. Coen, L. D., K. L. Heck, Jr., and L. G. Abele. 1981. Experiments on competition and predation among shrimps of seagrass meadows. Ecology 62:1484-1493. Davis, R. A., Jr. 1983. Depositional systems, a genetic approach to sedimen- tary geology. Prentice-Hall, Inc., NJ, 669 p. DeVore, p. W., L. T. Brooke, and W. A. Swenson. 1980. The effects of red clay turbidity and sedimentation on aquatic life in the Nemadji River system. In S. C. Andrews, R. G. Christensen, and C. D. Wilson (editors). Impact of non- point pollution control on western Lake Superior, Part II, p. 131-209. United States Environmental Protection Agency Red Clay Project, Final Report, Great Lakes Na- tional Program Office, Chicago, IL. Egusa, S., and T. Yamamoto. 1961. Studies on the respiration of the "Kuruma" prawn Penaevs japonicus Bate. I. Burrowing behavior, with special reference to its relation to environmental oxygen concen- tration. Bull. Jpn. Soc. Sci. Fish. 27:22-27. Eldred, B., R. M. Ingle, K. D. Woodburn, R. F. Hutton, and H. Jones. 1961. Biological observations on the commercial shrimp, Penaeus duorarum Burkenroad, in Florida waters. Fla. State Board Conserv., Prof. Pap. Ser. Mar. Lab. Fla., No. 3, 139 p. Folger, D. W. 1972. Characteristics of estuarine sediments of the United States. Geol. Surv. Prof. Pap., No. 742, 94 p. U.S. Gov. Print. Off., Wash., D.C. Folk, R. L. 1980. Petrology of sedimentary rocks. Hemphill Publishing Co., Austin, TX, 182 p. Fuss, C. M., Jr., and L. H. Ogren. 1966. Factors affecting activity and burrowing habits of the pink shrimp, Penaeus duorarum Burkenroad. Biol. Bull. (Woods Hole) 130:170-191. 68 MINELLO ET AL.: PREDATION ON BROWN SHRIMP Gardner, M. B. 1981 . Effects of turbidity on feeding rates and selectivity of bluegills. Trans. Am. Fish. Soc. 110:446-450. Gerritsen, J., AND J. R. Strickler. 1977. Encounter probabilities and community structure in zooplankton: a mathematical model. J. Fish. Res. Board Can. 34:73-82. Goldsmith, T. H. 1973. Photoreception and vision. In C. L. Prosser (editor), Comparative animal physiology, Vol. 1, Environmental physiology-, p. 577-632. W. B. Saunders Co., Phila. Gradall, K. S., and W. a. Svi^enson. 1982. Responses of brook trout and creek chubs to turbidity. Trans. Am. Fish. Soc. 111:392-395. Groot, S. J. de. 1971. On the interrelationships between morphology of the alimentary tract, food and feeding behavior in flatfishes (Pisces: Pleuronectiformes). Neth. J. Sea Res. 5:121-196. GUILCHER, A. 1967. Origin of sediments in estuaries. In G. H. Lauff (editor), Estuaries, p. 149-157. Am. Assoc. Adv. Sci. Publ. No. 83. Haven, D. S., and R. Morales-Alamo. 1972. Biodeposition as a factor in sedimentation of fine sus- pended solids in estuaries. In B. W. Nelson (editor). En- vironmental framework of coastal plain estuaries, p. 121- 130. Geological Society of America, Inc., Boulder, CO. Heck, K. L., Jr., and T. A. Thoman. 1981. Experiments on predator-prey interactions in vege- tated aquatic habitats. J. Exp. Mar. Biol. Ecol. 53:125-134. Heimstra, N. W., D. K. Damkot, and N. G. Benson. 1969. Some effects of silt turbidity on behavior of juvenile largemouth bass and green sunfish. U.S. Dep. Inter., Bur. Sport Fish. Wildl. Tech. Paper No. 20, 9 p. Howick, G. L., and W. J. O'Brien. 1983. Piscivorous feeding behavior of largemouth bass: an experimental analysis. Trans. Am. Fish. Soc. 112:508-516. Hughes, D. A. 1966. Investigations of the "nursery areas" and habitat pref- erences of juvenOe penaeid prawns in Mozambique. J. Appl. Ecol. 3:349-354. 1968a. Factors controlling emergence of pink shrimp (Penaeus duorarum) from the substrate. Biol. Bull. (Woods Hole) 134:48-59. 1968b. On the mechanisms underlying tide-associated move- ments oi Penaeus duorarum Burkenroad. FAO Fish. Rep. No. 57, p. 867-874. Lakshmi, G. J., A. Venkataramiah, and G. Gunter. 1976. Effects of salinity and photoperiod on the burying behavior of brown shrimp Penaev^ aztecus Ives. Aqua- culture 8:327-336. Linton, T. L. 1968. A description of the South Atlantic and Gulf coast marshes and estuaries. In J. D. Newsom (editor), Proceed- ings of the marsh and estuary management symposium, Louisiana State University, 1967, p. 15-25. Thos. J. Moran's Sons, Inc., Baton Rouge, LA. Matthews, W. J. 1984. Influence of turbid inflows on vertical distribution of larval shad and freshwater drum. Trans. Am. Fish. Soc. 113:192-198. Minello, T. J., AND R. J. Zimmerman. 1983. Fish predation on juvenile brown shrimp, Penaeus aztecus Ives: the effect of simulated Spartina structure on predation rates. J. Exp. Mar. Biol. Ecol. 72:211-231. 1984. Selection for brown shrimp, Penaeus aztecus, as prey by the spotted sea trout, Cynoscion nebulosus. Contrib. Mar. Sci. 27:159-167. MOCTEZUMA, M. A., AND B. F. Blake. 1981. Burrowing activity in Penaeus vannamei Boone from the Caimanero-Huizache lagoon system on the Pacific coast of Mexico. Bull. Mar. Sci. 31:312-317. Moore, J. W., and I. A. Moore. 1976. The basis of food selection in flounders, Platichthys Jlesus (L.)., in the Severn Estuary. J. Fish Biol. 9:139-156. Nelson, W. G. 1979. Experimental studies of selective predation on amphi- pods: Consequences of amphipod distribution and abundance. J. Exp. Mar. Biol. Ecol. 28:225-245. Olla, B. L., C. E. Samet, and A. L. Studholme. 1972. Activity and feeding behavior of the summer flounder, Paralichthys dentatus, under controlled laboratory condi- tions. Fish. Bull., U.S. 70:1127-1136. Postma, H. 1967. Sediment transport and sedimentation in the estuarine environment. In G. H. Lauff (editor). Estuaries, p. 158- 179. Am. Assoc. Adv. Sci. Pub. No. 83, Wash., D.C. Roelofs, E. W. 1954. Food studies of young sciaenid fishes Micropogon and Leiostomus, from North Carolina. Copeia 1954:151-153. Rulifson, R. a. 1981. Substrate preferences of juvenile penaeid shrimps in estuarine habitats. Contrib. Mar. Sci. 24:35-52. Short, F. T., and C. A. Short. 1984. The seagrass filter: purification of estuarine and coastal waters. In V. S. Kennedy (editor), The estuary as a filter, p. 395-413. Acad. Press, N.Y. SiGLER, J. W., T. C. BJORNN, AND F. H. EVEREST. 1984. Effects of chronic turbidity on density and growth of steelheads and coho salmon. Trans. Am. Fish. Soc. 113: 142-150. Steel, R., and J. Torrie. 1960. Principles and procedures of statistics. McGraw-Hill Book Co., Inc., N.Y., 481 p. Stein, R. A., and J. J. Magnuson. 1976. Behavioral response of crayfish to a fish predator. Ecology 57:751-761. Stoner, a. W. 1979. Species specific predation on amphipod Crustacea by pinfish (Lagodon rhomboides): mediation by macrophyte standing crop. Mar. Biol. (Berl.) 55:201-207. Swenson, W. a. 1978. Influence of turbidity on fish abundance in western Lake Superior. U.S. Environ. Prot. Agency, Ecol. Ser., EPA 600/3-78-067, 83 p. Swenson, W. A., and M. L. Matson. 1976. Influence of turbidity on survival, growth, and distribu- tion of larval lake herring (Coregonus artedii). Trans. Am. Fish. Soc. 105:541-545. Thayer, G. W., W. J. Kenworthy, and M. S. Fonseca. 1984. The ecology of eelgrass meadows along the Atlantic coast of North America: a community profile. U.S. Fish Wildl. Serv., Off. Biol. Serv., Wash., D.C, FWS/OBS-84- 02. Vinyard, G. L., and W. J. O'Brien. 1976. Effects of light and turbidity on the reactive distance of bluegill {Lepomis macrochirus). J. Fish. Res. Board Can. 33:2845-2849. Waterman, T. H. 1961. Light sensitivity and vision. In T. H. Waterman (editor), The physiology of Crustacea, Vol. II, Sense organs, integration, and behavior, p. 1-64. Acad. Press, N.Y. 69 WiCKHAM, D. A., AND F. C. MiNKLER III. 1975. Laboratory observations on daily patterns of burrow- ing and locomotor activity of pink shrimp, Penaeus duo- rarum, brown shrimp, Penaeus aztecus, and white shrimp, Penaeus setiferus. Contrib. Mar. Sci. 19:21-35. Williams, A. B. 1958. Substrates as a factor in shrimp distribution. Limnol. Oceanogr. 3:283-290. Yokel, B. J. 1966. A contribution to the biology and distribution of the red drum, Sciaenops ocellata. M.S. Thesis, Univ. Miami, 160 p. FISHERY BULLETIN: VOL, 85, NO. 1 ZlEMAN, J. C. 1982. The ecology of the seagrasses of south Florida: a com- munity profile. U.S. Fish Wildl. Serv., Off. Biol. Serv., Wash., D.C., FWS/OBS-82/25, 158 p. Zimmerman, R. J., and T. J. Minello. 1984. Fishery habitat requirements: Utilization of nursery habitats by juvenile penaeid shrimp in a Gulf of Mexico salt marsh. In B. J. Copeland, K. Hart, N. Davis, and S. Fri- day (editors), Research for managing the nation's estuaries, p. 371-383. UNC Sea Grant Pub. 84-08, University of North Carolina, Raleigh, NC. 70 EFFECTS OF AN EL NINO EVENT ON THE FOOD HABITS OF LARVAL SABLEFISH, ANOPLOPOMA FIMBRIA, OFF OREGON AND WASHINGTON Jill J. Grover^ and Bori L. Olla^ ABSTRACT The effect of El Nino conditions on the food habits of larval sablefish, Anoplapoma fimbria, was examined by comparing the diet of larvae collected off Oregon and Washington during the 1983 El Nino event and during 1980, a year in which conditions were not anomalous. While differential utilization of appen- dicularians, pteropods, and amphipods was seen in the 2 years, the most notable difference was that small copepods contributed significantly more to the diet in 1983 than in 1980. Dietary data for 1983 were generally supported by independent plankton observations, especially with respect to the pre- dominance of Paracalanus parvus, a small calanoid copepod. Because adult sablefish live and spawn in deep water, changes in the food habits of neustonic larvae may represent one of the principal effects of the El Niilo conditions on this species. Larvae represent a precarious stage in the life history of marine fishes as they are highly vulner- able to fluctuations in oceanographic conditions and food resources. Their survival is dependent on suc- cessful feeding, avoidance of predation, and favor- able transport (Sinclair et al. 1985). The relative importance of these factors is difficult to assess since each varies with species, developmental stage (Hewitt et al. 1985), and environmental conditions. Additionally, these sources of mortality are inter- active insofar as the transport of larvae into areas with suboptimal feeding conditions may result in starvation, and starving larvae are at greatest risk for predation. Because survival past the larval and early juvenile stages clearly depends on successful feeding, understanding the success of larval popula- tions requires a thorough knowledge of feeding ecology. As a result, in recent years a number of studies have provided detailed descriptions of the food habits of marine fish larvae (e.g., Laroche 1982; Cohen and Lough 1983; Govoni et al. 1983; Gadom- ski and Boehlert 1984; Brewer and Kleppel 1986), and a few studies have documented the occurrence of starvation under natural conditions (O'Connell 1980; Hewitt et al. 1985; Grover and 011a 1986; Theilacker 1986). The present study examined the food habits of larval sablefish, Anoplopoma fimbria, collected during 2 years of differing oceanographic conditions. As oceanographic conditions are manifested through changes in zooplankton assemblages, between-year comparisons of the diet of larval fishes can reflect differences in oceanic conditions. Such comparisons are of particular interest when they in- clude periods characterized by highly anomalous conditions. One such oceanographic anomaly, an El Nino event, occurred in the eastern Pacific Ocean off North America from the fall of 1982 through late summer 1983. The magnitude of El Nino-induced anomalies in physical conditions appears to be greatest in surface waters, with anomalous conditions having the great- est effect on those life stages of fishes that occupy the upper water column. Adult sablefish inhabit deep slope waters and spawn at depths in excess of 300 m (Mason et al. 1983), and therefore may be rela- tively insulated from El Nino conditions (Bailey and Incze 1985). Their eggs may also be insulated, as they hatch in water deeper than 400 m (Mason et al. 1983). However, larvae ascend to surface waters where they reside through early juvenile stages (Shenker and 011a 1986; J. M. Shenker^). During this neustonic phase they are most vulnerable to anom- alous oceanographic conditions. While El Nifio con- 'College of Oceanography, Oregon State University, Hatfield Marine Science Center, Newport, OR 97365. ^Cooperative Institute for Marine Resources Studies, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, Hatfield Marine Science Center, Newport, OR 97365. ^Shenker, J. M. Oceanographic associations of neustonic lar- val and juvenile fishes and Dungeness crab megalopae off Oregon. Manuscr. in prep. University of California, Bodega Marine Laboratory, Bodega Bay, CA 94923. Manuscript accepted August 1986. FISHERY BULLETIN: VOL. 85, NO. 1, 1987. 71 ditions may affect fish larvae in several ways, Bailey and Incze (1985) speculated that sablefish may be particularly sensitive to altered food production. The aim of the present work was to examine the effect of an El Nino event on the food habits of lar- val sablefish, through a comparison of their diet off Oregon and Washington during the 1983 El Nino event and 1980, a year in which oceanographic con- ditions were not anomalous^. Since an earlier study showed that prey size-selection was a function of lar- val size (Grover and 011a 1986), ontogenetic differ- ences in diet were also considered. METHODS Sablefish larvae were collected by the cooperative U.S.-U.S.S.R. ichthyoplankton survey off Oregon and Washington, during 1980^ and 1983^, using a 0.5 m neuston net (Sameoto and Jaroszynski 1969) with 0.505 mm mesh towed for 10 min. Collections were made from 22 April to 4 May 1980 by the RV Tikhookaenskiy and from 22 April to 30 April 1983 by the RV Ekvator (Fig. 1). Samples with a mini- mum of 10 specimens were examined. A total of 267 larvae collected from 10 stations in 1980 and 136 larvae from 6 stations in 1983 were examined. In each year the number of larvae that were examined represented more than 45% of the total number of sablefish that were collected. In conjunction with neuston sampling surface water temperatures were recorded along major transects. All larvae were preserved in 10% Formalin^ at sea. Upon sorting they were switched into 5% For- malin, where they remained until their examination. After the standard length (SL) of each larva was measured, the digestive tract was removed. Con- tents of the entire digestive tract were evaluated. Only larvae with all or a large portion of the gut in- tact were examined. Gut contents were teased out, and prey items were identified: invertebrate eggs, pteropods, copepod nauplii, copepods, amphipods, euphausid larvae, appendicularians, and other prey 130° W *Sea Surface Thermal Analysis. Dates of issue: 8 May 1980 - 6 May 1986. Northwest Ocean Service Center, NOAA, 7600 Sand Point Way N.E., BIN C15700, Seattle, WA 98115. ^Kendall, A. W., and J. Clark. 1982. Ichthyoplankton off Washington, Oregon and Northern California, April-May 1980. Processed Rep. 82-11, 48 p. Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98112. «Clark, J., and A. W. Kendall. 1985. Ichthyoplankton off Washington, Oregon, and Northern California, April-May 1983. Processed Rep. 85-10, 48 p. Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. FISHERY BULLETIN: VOL. 85, NO. 1 127° 124° -48°N Figure 1.— Map of the stations off Washington and Oregon where larval sablefish were collected. Dots represent 1980 collections and crosses represent 1983 collections. items. When possible copepods were identified to species. Lengths of all prey items were recorded. After the study was begun, it became clear that prey widths should also be measured. From that point on, both lengths and widths were recorded for all prey that were not dorsoventrally or laterally flattened. A conservative approach was taken with prey dimensions, i.e., prey widths excluded appendages and cephalothorax lengths were measured for cope- pods. Both measurements were recorded for a total of 7,508 prey items. Diet was analyzed in terms of numerical percent- age composition (%N), percent frequency of occur- rence (%F0), and volumetric percentage composi- tion (%VOL). Prey volumes were calculated from prey dimensions, assuming a spherical shape for invertebrate eggs, while all other prey were more appropriately described by a cylindrical shape. For a comparison, volumes were also calculated assum- ing a spheroidal shape, following Gadomski and Boehlert (1984). Regardless of whether a cylindrical or spheroidal shape was assumed, the relative contri- 72 GROVER and OLLA: EFFECTS OF EL NINO ON SABLEFISH FOOD HABITS bution of each prey type was essentially the same, except that invertebrate eggs contributed slightly more (in all cases <0.1%) to the total volume when the spheroidal shape was assumed. After the volume of each individual prey was calculated, a median volume for each prey type was determined. Median values were used because they are free from implicit assumptions regarding the normality of these data. Each median volume was then multiplied by the raw data from the numerical percentage composition analysis to produce corresponding volumetric per- cent composition data. Data were examined by size group and year. The three analyses (%N, %F0, and %VOL) were coalesced to yield a more comprehensive assessment of prey importance, the index of relative importance (IRI = (%N + %VOL) X %FO)(Pinkasetal. 1971). For this combined analysis copepods were classified by size. Copepods <1 mm in length (TL) included Corycaeus anglicus, Oithona similis, Oncaea sp., Paracalanus parvus, and copepodites. Broken cope- pods of an indeterminable size were classified as unidentified small copepods. Copepods 1-2 mm long included Pseudocalanus sp., Clausocalanus spp., Ctenocalanus vanus, Aedideus sp., Scolecithricella minor, and Acartia spp. Calanus spp. and Metri- dia lucens were the only copepods >2 mm that were ingested. RESULTS Composition of the Diet of Small Larvae (<12.5 mm SL) During 1980, copepod nauplii comprised over 80% of the prey items ingested by number, although copepods <1 mm long contributed slightly more to the diet in terms of volume (Table 1, Fig. 2). The index of relative importance, IRI, (Pinkas et al. 1971) indicated that nauplii were more important than copepods (Table 1, Fig. 3). Many of the cope- pods <1 mm that were ingested by small larvae were in pieces and impossible to identify. Of those that could be identified, Oithona similis was the domi- nant species, followed by Paracalanus parvus (Table 2). The only noncopepod prey of much consequence was pteropods (Figs. 2, 3), locally comprising as much as 45.7% of the diet, by number. In 1983, copepod nauplii were again the dominant prey in terms of number, but copepods represented more than twice as much in volume as they did in 1980 (Table 1, Fig. 2). While both were significant components of the diet, copepod nauplii had a slight- ly larger IRI value than copepods <1 mm (Fig. 3). Although the proportion of 0. similis in the diet was similar in both years, P. parvus contributed more to the diet during 1983 (Table 2). A comparison of the sizes of copepods ingested during 1980 and 1983 (Table 3) showed that more small copepods were eaten in 1983, both in terms of the number of cope- pods ingested (x^ = 8.46, 1 df, P < 0.005) and volume (x^ = 755.32, 1 df , P < 0.001). Composition of the Diet of Medium-Sized Larvae (12.6-20.5 mm SL) While copepods ranked first in all analyses in 1980, other major prey items varied with method of analysis (Table 1, Fig. 2). From the IRI it is clear that copepods <1 mm were the most important prey, with 0. similis the dominant species by number (Table 2), and P. parvus dominant by volume. In addition, sizable numbers of 1-2 mm copepods, especially Acartia longiremis, were ingested at several stations. Copepod nauplii ranked second in IRI value. Of noncopepod prey items pteropods, euphausid larvae, and appendicularians were all of consequence in the diet. In 1983, while copepods <1 mm were again the principal prey, they comprised a greater portion of the diet than in 1980, with P. parvus dominating both in terms of number (Table 2) and volume. Cope- pod nauplii ranked second by number, but accounted for little prey volume, while euphausid larvae, the only noncopepod prey to contribute substantially to the diet, ranked higher than nauplii both in terms of volume and IRI. Of the copepod species ingested, the most striking difference between the years was the increased contribution of P. parvus in 1983 (Table 2). Copepods >1 mm contributed more to the diet in 1980 than 1983 while small copepods comprised a greater portion of the diet during 1983 (Table 3), both in terms of number (x^ = 130.25, 1 df , P < 0.001) and volume (x^ = 9906.86, 1 df , P < 0.001). Composition of the Diet of Large Larvae (20.6-28.5 mm SL) While ranking second to small copepods by num- ber and second to large copepods by volume, appen- dicularians comprised more than 30% of the diet by number and volume during 1980, and ranked highest in IRI value, slightly ahead of copepods <1 mm (Table 1, Fig. 3). In contrast, euphausid larvae and amphipods were the most important noncopepod prey items in the diet during 1983, although each contributed <15% to the diet. In 1983 the diet 73 FISHERY BULLETIN: VOL. 85, NO. 1 Table 1 .—Composition of the diet of larval sablefish in terms of the Index of Relative Importance (IRI) and its components: numerical percent composition (%N), frequency of occurrence (%F0), and volumetric percent composition (%VOL), by size class and year, with n = sample size and %ZlRI = the contribution of each prey type to the total IRI. Prey type %N %F0 %VOL IRI %IlRI Prey type %N %F0 %VOL IRI %IlRI Larval size <12.5 mm Larval size 12.6-20.5 mm— Continued - 1980 Invertebrate eggs Pteropods Copepod nauplii Copepods <1 mm 1-2 mm >2 mm Amphipods Euphausid larvae Appendicularians Other prey n = 1983 Invertebrate eggs Pteropods Copepod nauplii Copepods <1 mm 1.2 mm >2 mm Amphipods Euphausid larvae Appendicularians Other prey n = 1980 Invertebrate eggs Pteropods Copepod nauplii Copepods <1 mm 1-2 mm >2 mm Amphipods Euphausid larvae Appendicularians Other prey n = 0.5 5.8 84.0 7.4 0.8 0.4 0.6 0.5 165 0.5 2.1 74.3 21.7 1.0 0.4 49 7.9 25.5 96.4 66.1 14.5 6.7 7.3 4.2 8.2 14.3 91.8 85.7 18.4 4.1 0.1 11.2 27.1 27.8 8.0 15.3 7.2 3.3 <0.1 2.7 15.9 66.2 5.8 9.3 4.7 433.5 10,710.0 2,326.7 127.6 105.2 56.9 15.6 4.9 68.6 8,280.4 7,533.0 125.1 39.8 0.3 9.4 25.8 48.2 7.9 0.2 2.4 5.4 0.4 79 Larval size 12.6-20.5 mm 5.1 38.0 70.9 93.7 49.4 3.8 25.3 17.7 3.8 <0.1 3.7 1.7 37.7 19.3 5.0 18.3 13.7 0.5 4.7 497.8 1,949.7 8,048.8 1,343.7 19.8 523.7 338.1 3.4 <0.1 3.1 77.7 16.9 0.9 0.8 0.4 0,1 <0.1 0.4 51.6 46.9 0.8 0.2 <0.1 3.9 15.3 63.2 10.5 0.2 4.1 2.7 <0.1 1983 Invertebrate eggs Pteropods Copepod nauplii Copepods <1 mm 1-2 mm >2 mm Amphipods Euphausid larvae Appendicularians Other prey n = 1980 invertebrate eggs Pteropods Copepod nauplii Copepods <1 mm 1-2 mm >2 mm Amphipods Euphausid larvae Appendicularians Other prey n = 1983 Invertebrate eggs Pteropods Copepod nauplii Copepods <1 mm 1-2 mm >2 mm Amphipods Euphausid larvae Appendicularians Other prey n = 0.2 0.3 9.9 83.3 3.7 <0.1 0.1 2.3 0.2 66 1.4 40.6 11.0 4.1 0.5 0.9 38.4 0.2 23 0.2 0.5 0.4 71.3 9.6 4.0 8.1 5.5 0.4 21 1.5 9.1 60.6 100.0 48.5 1.5 4.5 40.9 7.6 0.1 0.1 0.6 71.6 10.1 1.1 0.2 15.7 0.5 0.5 3.6 636.3 15,490.0 699,3 1.8 1.4 736.2 5.3 Larval size 20.6-28.5 mm 2.9 13.0 <0.1 30.4 95.7 87.0 69.6 17.4 17.4 73.9 8.7 9.5 14.3 19.0 95.2 76.2 66.7 47.6 61.9 14.3 <0.1 13.2 10.0 42.0 0.4 2.2 32.0 0.1 <0.1 <0.1 <0.1 24.1 12.3 41.6 7.2 14.3 0.4 39.0 45.6 5,148.7 1,827.0 3,208.6 15.7 53.9 5,202.6 2.6 2.8 8.6 9.5 9,082.1 1,668.8 3,041.5 728.3 1,225.6 11.4 0.1 0.1 3.6 88.3 3.8 <0.1 0.1 4.2 <0.1 0.2 0.3 33.1 11.8 20.6 0.1 0.3 33.5 <0.1 <0.1 <0.1 <0.1 57.6 10.6 19.3 4.6 7.8 <0.1 shifted towards smaller copepods (Table 3), espe- cially P. parvus, so that copepods <1 mm ranked highest in IRI value. Evidence for this shift came from numerical data (x^ = 27.30, 2 df , P < 0.001) as well as volumetric data (x^ = 1928.73, 2 df , P < 0.001). A comparison of the diet of medium-sized and large larvae showed that despite other dietary dif- ferences, the number of copepods in the diet of medium-sized larvae equalled the number of cope- pods ingested by large larvae in 1980 (x^ = 0.13, 1 df , P > 0.75). However, large larvae ingested a greater volume of copepods in 1980 than did medium-sized larvae (x^ = 148.09, 1 df, P < 0.001). In 1983 the numbers of copepods ingested by medium-sized and large larvae were again equiva- lent (x^ = 2.93, 1 df, P > 0.05), although copepods contributed more, volumetrically, to the diet of medium-sized larvae (x^ = 657.22, 1 df , P < 0.001). Comparison of Oceanographic Conditions Because 1983 was a year of anomalous El Nifio conditions, differences in diet between 1980 and 1983 may have been due to differing oceanographic conditions in the 2 years. Surface water tempera- tures taken during the ichthyoplankton surveys, were used as an indicator of oceanographic condi- tions. As the two surveys followed the same tran- sects, station coordinates corresponded for the 2 years. Temperatures recorded from 20 April to 5 May 1980 averaged 10.67°C (s = 0.783), while those from 23 April to 6 May 1983 were significantly higher (P = 0.001) averaging 11.78°C (s = 0.772). 74 GROVER and OLLA: EFFECTS OF EL NINO ON SABLEFISH FOOD HABITS 1980 FISH SIZE 12.6- 20.5 mm UJ o LxJ (T cr O o o U- o >- o UJ 3 o u cr - o UJ O UJ cr u_ 7, VOL 80 40 0 100- 200 300 400 500 7. N 40 80 'AidJB-' ' ' ' A- INVERTEBRATE EGGS B-PTEROPODS C-COPEPOD NAUPLIl L D-COPEPODS 2mm G-AMPHIPODS H-EUPHAUSID LARVAE I-APPENDICULARIANS J-OTHER 1 I H -I Figure 2.— Relative importance of prey items in the diet of larval sablefish, by size class and year, expressed as numerical percent composition {%N), volumetric percent composition (%VOL), and per- cent frequency of occurrence (%F0). The area of each block represents the Index of Relative Impor- tance (IRI) of a given prey (IRI = (%N + %VOL) x %F0). Sample sizes are as listed in Table 1. Comparisons with other years revealed that ther- mal patterns off Oregon and Washington were similar for 1980 through 1982 and 1984 through 1986 (fn. 4). In contrast, 1983 was markedly differ- ent, being the only year among those compared when a 13°C isotherm developed and when surface temperatures <10°C were not found between the coast and long. 130°W during the first week in May. 75 FISHERY BULLETIN: VOL. 85, NO. 1 Table 2.— Species composition of copepods consumed by sablefish larvae. Data are expressed as numerical percentages of all copepods that were ingested by each size class in each year. Larval size class (mm) Copepod species <12.5 12.6-20.5 20.6-28.5 1980 <1 mm Corycaeus anglicus 0.6 8.5 Oithona similis 27.4 38.7 16.8 Oncaea sp. 0.2 0.2 0.5 Paracalanus parvus 9.4 19.5 35.8 unidentified 38.5 19.5 8.0 copepodites 14.5 7.1 3.3 1-2 mm Pseudocalanus sp. 1.8 1.7 3.1 Clausocalanus spp. 1.2 2.6 2.4 Ctenocalanus vanus 5.1 1.1 4.7 Aetideus sp. Scolecithhcella minor 0.1 Acartia sp. 1.4 7.8 4.9 unidentified 0.5 0.8 4.7 >2 mm Calanus spp. 0.2 5.4 Metridia lucens 0.1 0.2 unidentified 1.7 1983 <1 mm Corycaeus anglicus 0.3 0.3 1.5 Oithona similis 24.8 28.7 17.1 Oncaea sp. Paracalanus parvus 39.2 47.5 45.3 unidentified 23.7 13.2 15.8 copepodites 7.5 5.9 4.1 1-2 mm Pseudocalanus sp. 0.8 0.7 1.4 Clausocalanus spp. 0.5 0.7 3.6 Ctenocalanus spp. 2.9 1.3 1.5 Aetideus sp. 0.1 Scolecithricella minor Acartia spp. 0.3 0.4 0.7 unidentified 1.1 4.2 >2 mm Calanus spp. <0.1 3.4 Metridia lucens unidentified 1.4 Table 3.— Size selection of copepods by sablefish larvae. Data are expressed as a percentage of all copepods that were consumed by each size class of larvae in each year. Larval size class (mm) 1980 1983 Copepod 12.6- 20.6- 12.6- 20.6- size <12.5 20.5 28.5 <12.5 20.5 28.5 A. By number <1 mm 90.0 85.6 72.9 95.5 95.6 83.8 1-2 mm 10.0 14.1 19.8 4.5 4.3 11.4 >2 mm 0.0 0.3 7.3 0.0 <0.1 4.8 B. By voiur Tie <1 mm 77.8 61.0 20.3 91.9 86.5 30.7 1-2 mm 22.2 31.0 15.5 8.1 12.1 15.9 >2 mm 0.0 8.0 64.2 0.0 1.4 53.4 DISCUSSION A comparison of the diet of sablefish larvae in 1980 and 1983 revealed several differences. Most notably, for larvae of all sizes, copepods <1 mm con- tributed significantly more to the diet in 1983 than in 1980. Appendicularians were the dominant prey for large larvae in 1980, but were negligible in the diet during 1983. Amphipods were only of conse- quence in the diet of large larvae in 1983. Pteropods comprised a substantial portion of the diet of small and medium- sized larvae in 1980, but made a trivial contribution to the diet in 1983 although ingested by larvae of all sizes. Although concurrent zooplankton data are lack- ing in this study, judging from the diet, prey popu- lations were probably quite different in 1980 and 1983. As a result of the anomalous conditions dur- ing 1983, a separate study extensively sampled zoo- plankton off the Oregon coast (Miller et al. 1985), where almost half (48%) of the sablefish larvae from 1983 were collected. In both 1980 and 1983, the timing of sablefish lar- vae collections corresponded to the spring transi- tional period off Oregon reported for previous years (Peterson and Miller 1976, 1977). During this period winds and currents shift, upwelling develops, and the zooplankton is transitive between a winter assemblage that is dominated by southern species of copepods and a summer assemblage that is domi- nated by copepods with northern affinities. During 1980, 8% of the copepods that were ingested were northern species, representative of the spring transi- tion, i.e., Pseudocalanus sp. and Acartia spp., especially Acartia longiremis. In contrast, during 1983 as a result of the El Niiio event manifested through increases in surface water temperatures, sea level, and poleward currents, reductions in salin- ities and coastal upwelling, and a depression of the thermocline (Fiedler 1984; Huyer and Smith 1985; McGowan 1985), only a partial spring transition occurred, with southern (winter) species, especial- ly P. parvus continuing to dominate the plankton through July (Miller et al. 1985). The diet during 1983 reflected this same trend with P. parvus being paramount in importance while northern species ac- counted for <1.5% of the ingested copepods. The fact that the pteropod Limacina helicina, a species with northern affinities which is the domi- nant pteropod species off Oregon, was not very abundant in 1983 (C. B. Miller^) correlates well with *C. B. Miller, College of Oceanography, Oregon State Univer- sity, Corvallis, OR 97331, pers. commun. December 1985. 76 GROVER and OLLA: EFFECTS OF EL NINO ON SABLEFISH FOOD HABITS 20,000 r Ijj 10.000 o o Q. 5000 1980 SL 1983 A < u en o X UJ a 1000 500 200 100 B o - o A Pteropods Copepod Naupili 2mm COPEPOOS Amphipods .A_ Euphausid Lorvoe Appendiculorians PREY TYPE Figure 3.— Indices of relative importance (IRI values) of dominant prey items in the diet of larval sablefish, by size class and year. Only IRI values >100 are included. In each case the sum of IRI values <100 accounted for <1% of the sum of all IRI values. our dietary observations which showed a marked decrease relative to 1980. Plankton data regarding the relative abundance of appendicularians and am- phipods are less correlative with diet. Preliminary analyses of collections made off Ore- gon in the spring and summer of 1983 (Miller et al. 1985) indicated that zooplankton density was re- duced, possibly as low as 30% of that found in a non- El Nino year. Indirectly lending support for this was data from satellite imagery which monitored phyto- plankton pigment images and which indicated that primary productivity was substantially reduced dur- ing 1983 (Fiedler 1984). Fulton and LeBrasseur (1985) suggested that while interannual fluctuations in zooplankton bio- mass affect planktivorous fish living in the open ocean off the west coast of North America, major shifts in the particle size of zooplankton may have a greater effect. In particular, the extreme north- ward shifting of the subarctic Pacific boundary, which occurred during the 1957-58 El Nino event, resulted in the replacement of large copepods with small copepods off North America from lat. 40 °N to 52 °N. They hypothesized that the absence of large copepods and decreased biomass in these waters during a warm El Nino year would result in reduced growth and perhaps reduced survival of juvenile salmonids. Ontogenetic changes in the diet of sablefish lar- vae included the diminution of small prey such as copepod nauplii and pteropods, and the increasing contribution of larger prey such as amphipods, euphausid larvae, and appendicularians as larvae grew. These observations parallel earlier findings on prey size-selection of larval sablefish (Grover and 011a 1986) and agree with trends seen in many other marine fish (Hunter 1981). All size classes of larvae showed some flexibility in the prey they ingested, from year to year as well as from station to station, with large larvae ingesting the widest range of prey items. The expansive range of prey ingested by large larvae may have enabled them to ingest large num- bers of small copepods during 1983, when larger prey of high caloric value may not have been readi- ly available. From all indications, 1983 was a year of reduced planktonic productivity off Oregon and Washington (Fiedler 1984; Miller et al. 1985). It was also a year when P. parvus, a small copepod, was dominant both in the plankton (Miller et al. 1985) and in the diet 77 FISHERY BULLETIN: VOL. 85, NO. 1 of larval sablefish. As low productivity and the pre- dominance of small copepods were also observed during a previous El Nifio year (Fulton and LeBras- seur 1985), these planktonic conditions may be fairly typical of El Nino events off Oregon and Washing- ton. A comparison of the diet in 1980 and 1983 suggests that a decrease in the size of copepods in- gested by sablefish larvae in 1983 may be one of the principal effects of El Nino conditions on the diet of this species. While it is possible that some ener- getic deficit may be imparted because of this dietary shift, the actual effects on the growth of larval sable- fish are at present unknown. Although the diet in 1983 was reflective of the plankton, corresponding plankton data were not available for 1980, thus it is unclear how closely the diet resembled the plank- ton in 1980. However, if we assume that the cope- pods that were ingested by larvae are indicative of the copepods that were available, as was the case in 1983, then we may infer that larger copepods were more readily available during 1980, in the absence of anomalous conditions. While a reduction in zooplankton biomass might affect sablefish larvae of all sizes, a paucity of large copepods would likely have the greatest effect on larvae >20.6 mm for two reasons: 1) large larvae repeatedly ingested the largest prey (Grover and 011a 1986) and 2) they were the only size class that ingested a substantial volume of large copepods. ACKNOWLEDGMENTS We wish to thank Art Kendall for valuable discus- sions, encouragement, and continued interest in this project. We would also like to thank two anonymous reviewers for their comments on an earlier draft of this manuscript. This work was supported by the Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA Contract Nos. 83-ABC-00045, 84- ABC-00127, and 85-ABH-00025. LITERATURE CITED Bailey, K. M., and L. S. Incze. 1985. El Nino and the early life history and recruitment of fishes in temperate marine waters. In W. S. Wooster and D. L. Fluharty (editors), El Nino north: Nino effects in the eastern subarctic Pacific Ocean, p. 143-165. Washington Sea Grant Program, Univ. Wash., Seattle. 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. Cohen, R. E., and R. G. Lough. 1983. Prey field of larval herring Clupea harengiis on a con- tinental shelf spawning area. Mar. Ecol. Prog. Ser. 10:211- 222. Fiedler, P. C. 1984. Satellite observations of the 1982-1983 El Nino along the U.S. Pacific Coast. Science 224:1251-1254. Fulton, J. D., and R. J. LeBrasseur. 1985. Interannual shifting of the subarctic boundary and some of the biotic effects on juvenile salmonids. In W. S. Wooster and D. L. Fluharty (editors), El Nino north: Nino effects in the eastern subarctic Pacific Ocean, p. 237-247. Washington Sea Grant Program, Univ. Wash., Seattle. 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, Leio- stomiis xanthurus, and Micropogonias undulatits. Mar. Ecol. Prog. Ser. 13:189-199. Grover, J. J., and B. L. Olla. 1986. Morphological evidence for starvation and prey size selection of sea-caught larval sablefish, Anoplopoma fimbria. Fish. Bull., U.S. 84:484-489. Hewitt, R. P., G. H. Theilacker, and N. C. H. Lo. 1985. Causesof mortality in young jack mackerel. Mar. Biol. Prog. Ser. 26:1-10. Hunter, J. R. 1981. Feeding ecology and predation of marine fish larvae. In R. Lasker (editor). Marine fish larvae: morphology, ecology, and relation to fisheries, p. 34-77. Washington Sea Grant Program, Univ. Wash., Seattle. Huyer, a., and R. L. Smith. 1985. The signature of El Nino off Oregon, 1982-1983. J. Geophys. Res. Sec. C(Oceans) 90:7133-7142. Laroche, J. L. 1982. Trophic patterns among larvae of five species of sculpins (family: Cottidae) in a Maine estuary. Fish. Bull., U.S. 80:827-840. Mason, J. C, R. J. Beamish, and G. A. McFarlane. 1983. Sexual maturity, fecundity, spawning, and early life history of sablefish {Anoplopoma fimbria) off the Pacific coast of Canada. Can. J. Fish. Aquat. Sci. 40:2126-2134. McGowan, J. A. 1985. El Nino 1983 in the Southern California Bight. In W. S. Wooster and D. L. Fluharty, (editors). El Nino north: Nino effects in the eastern subarctic Pacific Ocean, p. 166-184. Washington Sea Grant Program, Univ. Wash., Seattle. Miller, C. B., H. P. Batchelder, R. D. Brodeur, and W. G. Pearcy. 1985. Response of the zooplankton and ichthyoplankton off Oregon to the El Nino event of 1983. In W, S. Wooster and D. L. Fluharty (editors). El Nino north: Nino effects in the eastern subarctic Pacific Ocean, p. 185-187. Washington Sea Grant Program, Univ. Wash., Seattle. O'CONNELL, C. P. 1980. Percentage of starving northern anchovy, Engraulis mordax, larvae in the sea as estimated by histological methods. Fish. Bull., U.S. 78:475-489. Peterson, W. T., and C. B. Miller. 1976. Zooplankton along the continental shelf off Newport, Oregon, 1969-1972: distribution, abundance, seasonal cycle, and year-to-year variations. Oregon State Univ. Sea Grant College Program Publ. No. ORESU-T-76-002, 111 p. 1977 Seasonal cycle of zooplankton abundance and species composition along the central Oregon coast. Fish. Bull., 78 GROVER and OLLA: EFFECTS OF EL NINO ON SABLEFISH FOOD HABITS U.S. 75:717-724. Anoplopoma fimbria. Can. J. Fish. Aquat. Sci. 43:930-937. PiNKAS, L., M. S. Oliphant, and I. L. K. Iverson. Sinclair, M., M. J. Tremblay, and P. Bernal. 1971. Food habits of albacore, bluefin tuna, and bonito in 1985. El Nino events and variabiUty in a Pacific mackerel California waters. Calif. Dep. Fish Game Fish. Bull. 152: (Scomber japonicus) survival index: support for Hjort's 1-105. second hypothesis. Can. J. Fish. Aquat. Sci. 42:602-608. Sameoto, D. D., and L. 0. Jaroszynski. Theilacker, G. H. 1969. Otter surface sampler: a new neuston net. J. Fish. 1986. Starvation-induced mortality of young sea-caught jack Res. Board Can. 26:2240-2244. mackerel, Trachurus symmetricus, determined with histo- Shenker, J. M., and B. L. Olla. logical and morphological methods. Fish. Bull., U.S. 84:1- 1986. Laboratory feeding and growth of juvenile sablefish, 17. 79 THE LANTERNFISHES (PISCES: MYCTOPHIDAE) OF THE EASTERN GULF OF MEXICO John V. Gartner, Jr./ Thomas L. Hopkins,' Ronald C. Baird,^ AND Dean M. Milliken^ ABSTRACT Forty-nine species from 17 genera of Myctophidae were taken in mid water trawl samples from the eastern Gulf of Mexico during March through October between 1970 and 1977. Seven abundant species (Ceratoscopelus warmingii, Notolychnus valdiviae, Lepidophanes guentheri, Lampanyctus alatus, Diaphus dumerilii, Myctophum affine, and Benthosema suborbitale) comprised 74.4% of the total number (13,369) of myctophids captured. Of the remainder, 10 species were common, 26 were uncommon, and 6 were rarely collected. Diel vertical profiles showed that all species except Taaningichthys vertically migrated. Daytime vertical ranges for the entire assemblage were between 300 and 900 m, while at night myc- tophids were most abundant between the surface and 150 m. A deep group remained below 600 m at night and was composed of mostly juvenile nonmigratory individuals of 19 species and Taaningichthys bathyphilus. Five daytime and five nighttime groups of associated species were defined based on ver- tical ranges, minimum depths of occurrence and zones of abundance. Species of tropical and tropical- subtropical zoogeographic affinities comprised the largest percentage of the total number of specimens and were about equal in their percentage contributions. The presence of many tropical species in the collections may have been due to the transport of the Florida Loop Current. Comparison of the species list with those reported for other myctophid assemblages from tropical-subtropical latitudes shows panoceanic distribution of 10 species. Myctophid fishes are one of the dominant compo- nents of oceanic mesopelagic ecosystems (McGinnis 1974; Maynard et al. 1975; Badcock and Merrett 1976; Nafpaktitis et al. 1977; Hulley 1981; Hopkins and Lancraft 1984; Hulley and Krefft 1985). With the exception of Clarke's (1973) work in Hawaiian waters, there have been no comprehensive studies on faunal structure and ecology of this family in subtropical-tropical oligotrophic regions where myc- tophids are exceptionally diverse (Backus et al. 1977). The Gulf of Mexico is one such regime. Backus et al. (1977) noted that although there are no endemic myctophid species in the Gulf of Mexico, it is zoo- geographically unique and faunistically separable from other regions of the western North Atlantic. Unlike the adjacent Caribbean Sea, which it hydro- graphically resembles (Nowlin and McLellan 1967), the Gulf of Mexico undergoes a marked change in surface water temperatures over an annual cycle (Jones 1973). In addition, circulation patterns are 'Department of Marine Science, University of South Florida, 140 Seventh Avenue, S.E., St. Petersburg, FL 33701. ^Director of Corporate Relations, Worcester Polytechnic In- stitute, Worcester, MA 01609. ^Florida Institute of Oceanography, 830 First Street S., St. Petersburg, FL 33701. strongly influenced by the Florida Loop Current, whose penetration into the Gulf of Mexico is both geographically and seasonally variable. The central Gulf of Mexico, despite seasonal variability, has many characteristics typical of low latitude oligo- trophic gyre systems. This paper details the taxonomic composition, zoo- geographic affinities, and vertical structure of the mesopelagic (sensu Marshall 1971) myctophid fauna in the eastern Gulf of Mexico (hereafter as Gulf) dur- ing the warm months of late March through early October. The results are based primarily on collec- tions with opening-closing midwater trawls made from 1970 to 1977 in the eastern central Gulf in the vicinity of lat. 27°N, long. 86°W. Additional data from other stations in adjacent northeastern and southeastern Gulf areas are included. MATERIALS AND METHODS The data are from 526 stations occupied during 12 cruises made between 1970 and 1977 (Fig. 1). The majority of samples were taken within a 20 nmi diameter circle centered around lat. 27°N, long. 86° W in the eastern central Gulf of Mexico, an area referred to as the "Standard Station". Samples were also taken from the northeastern and south- Manuscript accepted August 1986. FISHERY BULLETIN: VOL. 85, NO. 1, 1987. 81 FISHERY BULLETIN: VOL. 85, NO. 1 Figure 1.— Collection locations and numbers of tows per location in the eastern Gulf of Mexico. Sampling regions: NE - northeast; EC - eastern central; SE - southeast. Circle in EC region delimits 20 nmi diameter sampling area around lat. 27°N, long. 86°W. eastern Gulf, and although there is a large discrep- ancy in the relative number of samples taken and in day-night depth coverage, collections were assigned to three latitudinal areas: northeast (NE), eastern central (EC), and southeastern (SE). The northernmost Gulf is known to support populations of cooler water myctophid species (Backus et al. 1977), and our hydrographic data showed that the southeastern stations were the only collections that were directly impacted by the tropical Loop Current waters during sampling, hence this regional parti- tioning. Most of the cruises were made during the sum- mer months (June through early September), and although sampling extended from late March through early October, all cruises were made while summer hydrographic conditions prevailed in the study area. With respect to lunar cycles, 9 of the 12 cruises occurred during waxing or full moon period, 2 during first quarter, and 1 during new moon. Table 1 lists general data for the 12 cruises. Samples were taken primarily in the upper 1,000 m of the water column, although a few extended as deep as 1,500 m. The nets used were 3.2 m^ or 6.5 m- mouth area opening-closing modified Tucker trawls (Hopkins et al. 1973) which incorporated 1.1 cm stretch mesh in the body and 505 ^ mesh in the cod end. Opening and closing of the nets were accomplished mechanically using double release mechanisms and messengers. Fishing depth was 82 GARTNER ET AL.: LANTERNFISHES monitored on all cruises by mechanical time-depth recorders (TDR), meters of wire out and wire angle. Additionally, on three RV Columbus Iselin cruises (Table 1) depth was monitored by electronic deck readout via a depth transducer and conducting cable. Volume of water filtered per tow was calcu- lated from flowmeters mounted in the mouth of the plankton net and on the main trawl frame. Incli- nometer measurements and underwater observa- tions of the wire angle of nets towed just beneath the surface indicated the mouth of the net fished at an angle of about 30° from the vertical, which was used as a standard angle when filtration volumes were calculated. Trawl speed was 1.5 to 3 kn. Nets were fished obliquely over a depth range, or, as on Columbus Iselin cruises, at specific depth horizons. Vertical depth control during horizontal tows was -I-/- 10 m from 0-300 m (except for the surface and 5 m depth strata), -i- / - 25 m from 300- 700 m, -H/-50 m from 700-1,000 m. The trawl catch was initially fixed in 10% v:v sea- water Formalin" and subsequently transferred to 50% isopropanol. Myctophids were measured to the nearest millimeter standard length (mm SL) and identified to species using Nafpaktitis et al. (1977). Diel vertical distributions for all species were calculated using data pooled from all cruises. Sam- ples which were taken within 1 hour before or after sunrise and sunset were excluded from analysis, because vertical migration is pronounced at these times. Excluding the sunrise-sunset samples, a total of 155 samples (82 days, 73 nights) collected dur- ing the Columbus Iselin cruises from 0 to 1,000 m were used to construct vertical profiles for the abun- dant species (Table 2). Abundance was expressed as individuals per 10** m^. ••Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Table 1.— Sampling data. Samplin g regions: NE - northeast; EC - eastern central; SE - southeastern. Research vessel Date Location/Region No. tows (stations) Lunar Phase VOate Loop Current impingement? Joie de Vivre 2-5 Oct. 1970 Standard Station (EC) 12 New - 30 Sept. 1/4X - 8 Oct. No Dan Braman 25-31 July 1971 Standard Station (EC) 46 New - 22 July 1/4X - 30 July No data Mizar cruise 1 7-14 June 1971 25°50N, 85°30 W (SE) 20 Full - 9 June 3/4W - 16 June Yes Mizar cruise II 24-28 Mar. 1972 25°50N, 85°30W (SE) 5 3/4X - 22 Mar. Full - 29 Mar. No Bellows cruise 1 2-9 Aug. 1972 Standard Station (EC) 25 1/4W - 2 Aug. New - 9 Aug. No Mizar cruise III 13-23 Aug. 1973 Standard Station (EC) 27°36N, 88°40W (NE) 29°19'N, 87°01'W (NE) 28°28'N, 88°56'W (NE) 7 8 8 9 Full - 14 Aug. 3/4W - 21 Aug. No No No No Bellows cruise II 9-11 Oct. 1973 24°30'N, 85°20W (SE) 24°40N, 85°10'W (SE) 24°38N, 85°10'W (SE) 4 5 2 Full - 12 Oct. Yes Yes Yes Bellows cruise III 28-31 Aug. 1974 Standard Station (EC) 15 Full - 1 Sept. No Columbus Iselin cruise 1 15-22 June 1975 Standard Station (EC) 79 3/4X - 16 June Full - 23 June No Columbus Iselin cruise II 6-17 June 1976 Standard Station (EC) 133 3/4X - 5 June Full - 12 June 3/4W- 19 June No Bellows cruise IV 27 May- 2 June 1977 27°11N, 84°50'W(EC) 27°10N, 85°07W (EC) 27°09N, 85°04W (EC) 27°07N, 85°16W (EC) 27°06N, 85°19'W (EC) 27°04N, 85°40'W (EC) 27°14'N, 84°33.5W (EC) Standard Station (EC) • 7 8 8 6 5 10 4 4 3/4X - 27 May Full - 1 June No No No No No No No No Columbus Iselin cruise III 19 Sept. - 6 Oct. 1977 Standard Station (EC) 96 3/4X - 20 Sept. Full - 27 Sept. 3/4W - 5 Oct. No 'Lunar phases: X - waxing. W - waning. 83 FISHERY BULLETIN; VOL. 85, NO. 1 Table 2.— Depth horizon data for discrete-depth samples collected during Columbus Iselin cruises I, II, and III. Day Night No. of Total volume No. Of Total volume Depth stations filtered stations filtered (m) sampled (10'' m^) sampled (10'' m^) 0-5 16 18.4 20 24.8 10 9 5.9 3 2.0 15 5 3.4 8 2.7 30 4 3.0 3 2.0 50 6 4.5 3 3.7 75 0 — 2 3.0 105 0 — 3 4.9 125 0 — 3 3.9 155 2 2,9 2 2.6 210 1 1.5 3 3.7 300 4 6.8 3 4.6 350 4 5.1 2 2.3 400 2 2.4 2 2.8 450 5 8.0 2 2.5 500 5 4.7 3 3.7 550 3 3.0 4 5.2 600 4 11.9 1 2.5 650 2 2.6 0 — 700 3 6.6 1 2.1 800 4 7.4 0 — 900 2 4.3 2 3.7 1,000 1 2.2 3 5.9 Totals 82 104.6 73 88.6 Groups of associated species were determined for both day and night and were defined based on the minimum depth of occurrence (the shallowest discrete depth capture of a species) and zone of max- imum abundance, using the Columbus Iselin data. These associations included the abundant and common species, as well as those uncommon species whose sample size was sufficient to determine depth range or which had a very narrow range of capture. Although our data were limited, effects of lunar phase on vertical distribution of abundant species were examined by comparing the minimum depth of captures between new and full moon phases. Three cruises were defined as new moon (Joie de Vivre, Dan Braman, Bellows cruise I, Table 1); the remainder as full. RESULTS Hydrography The circulation of the eastern Gulf of Mexico is dominated by the Loop Current (Leipper 1970; NowHn 1971; Jones 1973; Molinari and Mayer 1980). This current, of Caribbean origin, enters the Gulf through the Yucatan Straits and moves anti- cyclonically, exiting through the Florida Straits. The extent of penetration into the Gulf is latitudinally variable and seasonally unpredictable. The Loop Current can be identified by the depth of the 22 °C or 20 °C isotherms which occur at 100 m and 150 m, respectively (Leipper 1970; Maul 1977; Sturges and Evans 1983). Meanders of the Loop Current often pinch off to form cold core eddies which spin cyclonically and drift southward along the eastern edge of the Loop Current off the West Florida Shelf (Vukovich and Maul 1985). Some of these cyclonic eddies have been tracked through our eastern sam- pling areas (EC and SE). All of the southeastern (SE) collections were in waters covered by the Loop Current at the time of sampling, whereas the more northerly samples (EC and NE) were from what is termed Loop Transition Water. The characteristics of Loop Transition Water during the collection period (summer) were as follows: a mixed layer of variable depth, usually extending 25 to 50 m with surface temperatures of 27° to 30° C; a sharp thermocline from the base of the mixed layer to approximately 150 m depth where the temperature was 15° to 18°C; a gradual temperature decline from 150 m to about 4°C at 1,000 m. Figure 2 illustrates a typical profile of the Loop Transition Water during the summer months. Productivity measurements within the Loop Cur- rent and in Loop Transition Waters indicate an TYPICAL WARM MONTH THERMAL PROFILE e GOM E X H Q. liJ O 600 1 r~r-i — I 1 I I — r-| — i i i i | i i i — i—j — i i i i — |— i — n — r 0 5 10 15 20 25 30 TEMPERATURE "C Figure 2.— Typical warm month thermal profile for the eastern Gulf of Mexico (eGOM). 84 GARTNER ET AL.; LANTERNFISHES oligotrophic regime in the eastern Gulf, with an annual primary production of <50 to 75 g C yr"^ (El-Sayed 1972; Hopkins 1982, unpub. data). Species Data A total of 13,369 myctophids were examined from all stations, with all but 77 (0.6%) identifiable to species. Table 3 lists all taxa, along with the number of individuals captured from each sampling region, their distribution pattern (from Backus et al. 1977), diel distribution ranges, and overall size ranges. The identified material comprised 17 genera and 49 species. The distribution ranges and estimates of abun- dance and standing stock of the abundant species in the following species accounts were limited to data collected during the three Columbus Iselin cruises to Standard Station (EC). Too few collections were made in the other two areas (NE, SE) to allow for comparable analyses. Based on frequency of capture and total number of specimens from EC, seven species were con- sidered abundant, i.e., dominant, (>500 specimens total captured) in the eastern Gulf. In decreasing order of abundance these were Ceratoscopelus warmingii, Notolychnus valdiviae, Lepidophanes guentheri, Lampanyctus alatus, Diaphus dumerilii, Benthosema suborbitale, and Myctophum affine. Together they comprised approximately 75% of the total number of specimens collected from EC (74.4%). Table 4 Usts the dominant species along with their percentage composition among the domi- nant species and overall. From all collections, 42 additional species were divided into the following abundance categories: common (101-500 specimens; 10 species); uncommon (10-100 specimens; 26 species), and rare (<10 speci- mens; 6 species). Abundant Species Ceratoscopelus warmingii: N = 2,267, 14-65 mm SL, Juvenile-Mature Adult The vertical distribution profile shows this species to be a strong migrator with a broad diel depth range (Fig. 3a). During the day, it occurred from 650 to 1,000 m (recent deep mesopelagic daytime tows in the Gulf of Mexico have taken C. warmingii below 1,000 m, J. V. Gartner unpub. data). Night captures were mainly between 75 and 125 m, with a maximum abundance of 95 individuals/lO'* m^ at 125 m. Abundance was bimodally distributed with some small juveniles apparently remaining near day- time depths at night (Fig. 3a, Table 5). Notolychnus valdiviae: N = 1,780, 9-22 mm SL, Postlarvae-Mature Adult Peak daytime distribution was mainly between 400 and 500 m (Fig. 3b). Some night captures were made as shallow as 50 m, but abundance maxima were found at 75 and 155 m, with the peak at 75 m (>72 individuals/lO"^ m^). The distribution pattern was discontinuous, with no specimens taken at 125 m. There was no evidence for a nonmigratory portion of the population as was found for C. warmingii. Analysis of the size vs. depth for N. valdiviae showed an increase in mean size with increasing depth at night (Table 5). No trend was apparent dur- ing the day because of small sample size. Lepidophanes guentheri: N = 1,610, 13-64 mm SL, Juvenile-Mature Adult This species was a moderate to strong vertical migrator. Although daytime captures were recorded as shallow as 400 m, this species was most abun- dant between 650 and 800 m and recent discrete depth hauls in the Gulf of Mexico have also taken L. guentheri from below 1,000 m (J. V. Gartner un- pub. data). Nighttime abundance was highest at 75 m (38.1 individuals/lO"* m^) and sharply decreased below this depth (Fig. 3c). Differences in day-night abundances were not as pronounced as in Cerato- scopelus or Notolychnus. At 650 m during the day, catch abundance was approximately the same as at 105 and 125 m at night. Lepidophanes guentheri was the only abundant species for which lunar influence on depth of cap- ture was apparent. During a new moon cruise (Dan Braman), L. guentheri measuring 27 to 37 mm SL were captured as shallow as 10 m. Nonmigration of members of the population was noted, mostly among juveniles measuring from 13 to 25 mm SL. Because of sample size, no other clear size-depth patterns were discernible. Lampanyctus alatus: N = 1,418, 15-48 mm SL, Juvenile-Mature Adult The daytime vertical profile extended from 350 to 900 m, with a peak at 650 m (Fig. 3d). This species was found between 75 and 155 m at night (maximum 24.2 individuals/lO"* m^ at 125 m). Some estimated daytime abundances for L. alatus were as large as 85 FISHERY BULLETIN: VOL. 85, NO. 1 N cn ■a c (0 o 0)2? 2 12 CO C8 t o> 5-^ oi— ^ CE CO Q CO CD .J. c c k- Q *- c g O) CD IT UJ CO o o Cb o. c/) en en > Z Z > Z Z o IT) CO CO ■>* en c\j CO CO in -^ CO Lo oj cb •<}■ cvj in oS cvj cj) o z CO CO en en z>>->->>- o999 o9 z z z z z z en CO > > Q Q „eneneoeocnQ_QenQQ co oeueDCDajajtJotrcDtctroa) z>>->>>zzz>-zzz> -.-cocjicocorj-cocor^iJ in'cx)c\jcDOOcoinc35incj>inin''-oocoooint^t^cor>-cD Tt'^'<:rcocMincoooininco'*cO';rooe3)coi^in'*cot^cO'*inm C\jeJ)00'r-OT--r-,- inoococ\jr^incjc35cj)c\j'<3-cD'-oc:5co T-T-C\l^,-,-C\JC\J,-r-l-T-1-l-T-1-T- in in (i> CM CM o i^ T- in in m in c\j r^ -C t^ CO in in ^ in (uooooininoo CM oinoin'.-cMCMcoo •»— cOt^-»— T-^— ■*— -^ '— CO ci'Hcidjcbcbincbincb (j),5oinootncDinr^i- o o o o o o o o o o o o o o o in o o in o o in N- T— in in t^ T— CD t— o 6 in in 6 in m o o o 6 in 6 6 o in 6 o o in in o in in in in o in r^ in in o h- o in CO T— ^— 1^ ▼— C\J CO -* CM ■ 6 in in o o 6 o CJ) g o o o o o in o in o in o •^ CO CO CO -"t o o oo 700 550 1,000 1,000 550 600 700 700 1,000 1,000 o o in o o o o in o in o in o CM o in 'a- ■* CO CO CO o o o o in o in o CO CO CO CO 8 CO CO CO h- I- I- H H H I- T- h~ CO in h- CO CO CO in CO CO t^ o T- CM i^ CM o CO in m CO CD r^ r^ ■* o CD »- ■«t CO C35 T- CO 00 CM CD 00 CO O CO O 'd- in CO r^ •<«■ OM CM CD ■•- CO CM CM ■■- CO t^ N- cj> en o CO o 00 CO f^ •>-•>- CD oo CO CM CM •* CD 00 h~ Cm' t-' -r-' T-' T-' CO (- CO CO CO I- co I- <5 CO H Q- co CO E CO CO CJ) ,COCM0Mt-CM-<-CMCO 1- CM '- 1- •.- ■* in-i-T-ocoooinoocDc35 ooo-^r^oooioocot '-'-1-CO CM COCMi- CDCOTj-CMinC0'*-'a-'>a-CD •.- T-r^^ CMCDCO'- OTri-^inr^i-cocMi- COOCOCOOOCMCDCOr^ T-T-T— Tjf— COl— ^CO^— § C3) .c ■P £ o -C3 CO CO E 0) *- CO CO £ ■o c C (13 < CO .fo eo :§^ OQ ~j eu is (0 c: •c *- o CJ 5 o eo CJ is C 0} o o s .O) £ CO o eo 3 o CO 3 o c O CD - CD O Ci C3 (0 w CL eo g) cici i3 o c CD ■Q 9- , c CO Q. E eo o o CM in CM o o ^- c^- o o o O O CO in CO CO CD 6 6 m in T- CM T- 6 CD o o CO in in T *> T C5 6 6 o -^ o in 1^ o (D o o in o o in CM 5 T T T "^ c) CD in O O in (D (D CD in in c\j o o o f— 1— T— eo eo CO o o o "5 ^ ^ 1- r^ CO CO CO in CM in in o in o o o o o o o in CJ) CO t^ a> o c^. 8 0000000000000000--00 c3C3C3cjt:)C3e:30ooooooooo --ooin ino'^oinoinomoinoooin'-inOLnincM T^coT-Ln'5f-cOT7T-cMcDinincococMc\ii-in.,-,-CM in CD CD iJo in CD in ei) cd cb cDcD-- t CO CO CO o in o in o CD O CM ■^ o o in 6 o oooooooooo oooininooooin cocoininincDcDcDcDin cboooocDocDcbcD ininoininooooo CMCOCOCOCOCOCOCO'^'S' o in CD CD o CD o o o o o o o ^. o o >i in o T- in o *-". CD r-~ A CD CI) T- CD CD CD C^- CD cb o in o o CD o in in in o CD 00 in o o e-l m „ o ^•m t^ 1 "^ 1 1 ^ o ^ 1 o 1 ' o o o o CO in CD CO I- COh I- I- CO Q-CO H CO I- H H t- CO CO I- E CO CO I- CO I- H CO CO COCO HHC0(-l-l-l-HHCLHHI-C0(-COCOI-l-HHHI-HI-m COCMCMCMCD'tO .COCO .CM'-i-inOM-'-in ,0 , .CMt-CO^ CM - I I - I - I I coi-i-ooO'i-oocooin'a-cMT-'-coc\jt^r^''-i-'53-i-incMT-oj'>-cD'- inCO->-OMCO' ?" -i = -CI is o a CD ^ < ■ i5 C a) £ • " «- * g « 5 CL E o , 2 e™-° •D 2 ra E ii £ E o T3 UJ O E IS « h- 3 4 5 J 1 M|i I I J $ ~T 1 1 r Ceratoscope/us warmlngll ABUNDANCE (NO./10'*m^) im 42.7 (78.8) ,27.4 (95.0) -Hh- 0. LU a 0-5. 15 30 50 75-1 105- 125 155- I I- Q. UJ 210- 300 350 400 450 500 550 600 650 700 750- 800 850 SOO' 950 1000- 0-5- 15 = 30 50- 75- 105 125 ^ 155 E 210^ 300 '- 350- 400 450-1 500 550 600 650 700 750 800 850 900 950 1000 DAY NIGHT 54321 01 2345 j— //" — ' — ^ ■ 1 1 • t N — ' — ^- • • _1__ J L/^ h • • u • ■ , . 39.3 (72.5) • , • •1 ! ". • r 1 • • • • • i • • ' ^A 1 r • 1 1 1 1 -1 1 rif Notolychnus yaldlvlae ABUNDANCE (NO./10'*m^) 5 -/>J- DAY 3 2 NIGHT 5 ^1 9.7 (24.2) (• -r/y Lepldophanes guentherl Lampanyctus alatus Figure 3.— Diel vertical profile of dominant myctophid species in the eastern Gulf of Mexico, a - Ceratoscopelns warmingii; b - Notolychnus valdiviae; c - Lepidophanes guentheri; d - Lampanyctus alatics. Numbers above bar indicate average abundance at depth, numbers in parenthesis below bar indicate maximum abundance at depth. 89 FISHERY BULLETIN: VOL. 85, NO. 1 ABUNDANCE (N0./10'*m^) DAY NIGHT 3 2 1 0 12 3 4 _1_ -\ 1 r DIaphus dumerllll ABUNDANCE (NO./10''m^) DAY NIGHT 3 2 1 0 1 2 3 t t u E X I- Q. UJ Q ~T 1 1 1 ■ 1 1 r Benthosema suborbltale E I H 0. UJ Q ABUNDANCE (NO./ lO'^m'^) DAY NIGHT 1 0 12 3 4 5 I Myctophum aftlne ABUNDANCE (NO./10'*m^) DAY NIGHT 0-5 15' 30 50 75 105 125 ^ 155 E 210 300 350 400 450 500 550 600 650 700 750 800 850 900 980 1000 5 5 diit= ■*^19.8 (102.0) 55.6(41.8) ^^17.3 a^ -^A 1 1 1 1 ' 1 1 1 1 r//~ DOMINANT SPECIES COMBINED Figure 4.— Diel vertical profile of dominant myctophid species in the eastern Gulf of Mexico, a - Diapkus dumerilii; b - Myctophum affine; c - Benthosema suborbitale; d - All dominant species (7) combined. Numbers above bar indicate average abundance at depth, numbers in parenthesis below bar indicate maximum abundance at depth. asperum, M. nitidulum, and M. obtusirostre were most abundant in the surface (0-5 m) strata. Hygo- phum macrochir, which also entered near surface waters, Diaphus perspicillatus, D. taaningi, and Myctophum selenops occurred primarily between 50 and 100 m. Seven species (Bolinichthys supralate- ralis, Diaphiisfragilis, D. garmani, D. termophilus, Lampadena luminosa, Lobianchia gemellarii, Noto- scopelus resplendens) were most abundant between 75 and 150 m. The remaining 11 species usually were found deeper than 100 m, although several {Diaphus brachycephalus, D. luetkeni, and Lampanyctus 90 GARTNER ET AL.: LANTERNFISHES nobilis) were captured as shallow as 50 m, and Hygo- phum taaningi was occasionally taken in surface (0-5 m) waters. Individuals of nine species were taken deep at night, between 450 and 950 m, indicating incomplete migration of the populations (Table 3). Six species were considered rare (<10 individuals). Both Taaningichthys species were taken below 600 m during the day while the other four species came primarily from night samples between 50 and 200 m (Table 3). NE and SE Sampling Areas A total of 1,943 specimens from 42 of the 49 east- ern Gulf myctophid species were captured in the NE sampling area, while 40 of the 49 species were re- corded from SE samples (813 specimens; Table 3). The seven most abundant species in both areas are the same as for EC with the exception ofMyctophum affine, which was infrequently captured because of a lack of surface night samples. Diaphus mollis was in the top seven species in the NE locale; Hygophum taaningi in the SE. The top seven species comprised 81.0% of the total number of specimens collected from NE samples and 75.2% from SE. DISCUSSION Vertical Distribution Despite spatial overlapping, the vertical profiles of the eastern Gulf of Mexico myctophids show dis- crete stacking of species groups, which presumably enhances partitioning of spatial and trophic re- sources. Based on minimum depths of occurrence (MDO), vertical ranges and zones of abundance, groups of clearly associated species can be defined (Table 6). Five day and five night groups were con- structed. Each day group consists of at least one abundant species plus one or more common species. However, because of differential migration ranges, only the three shallowest night groups contain abun- dant species. During the day, almost all species have an MDO within their zone of highest abundance. All Diaphus species are typically upper mesopelagic inhabitants, while Lampanyctus and Taaningichthys species primarily dwell in the lower mesopelagic zone. Hygo- phum and Myctophum species are divided among upper and middle zones. The daytime distribution of the abundant species Table 6.— Species groups based on minimum depth of occurrence (MDO) and zones of abundance by day (36 spp.) and night (43 spp.). Underline indicates abundant species. Species in parentheses are outside of zone of maximum abundance; number after name indicates zone. MDO (m) Species DAY MDO (m) Species NIGHT - - - 300 B. supralateralis D. problematicus Surface C. nigroocellatus A/f. affine D. dumerilii D. rafinesquii G. cocco M. asperum D. effulgens D. splendidus (H. benoiti; 50-125) /W. nitidulum D. luetkeni H. benoiti (H. macrochir; 50-100) M. obtusirostre D. mollis L gemellarii (H. reinhardtii; 100-150) D. perspicillatus (H. taaningi; 75-130) 350 D. brachycephalus D. atlantlcus 50 (8. photothorax; 90-125) D. perspicillatus D. fragilis H. taaningi S. supralateralis (D. problematicus; 80-150) D. garmani (L. alatus; >600) B. suborbitale D. splendidus 400 B. suborbitale C. nigroocellatus D. lucidus D. subtilis D. taaningi {L guentheri: >600) N. valdiviae (D. brachycephalus; 125-200) D. dumerilii (D. lucidus; 100-150) (D. luetkeni; 100-160) D. taaningi (L nobilis; 100-150) M. selenops (N. valdiviae; 75) 500 B. photothorax H. macrochir H. reinhardtii L ater M. affine' M. obtusirostre 75 D. mollis C. warmingii D. garmani D. termophilus (A/, resplendens; 75-125) L alatus L. tenuiformis L. luminosa >600 L. luminosa C. warmingii L cuprarius L. lineatus L nobilis T. bathyphilus >100 >600 D. atlantlcus L guentheri D. effulgens D. fragilis D. rafinesquii D. subtilis T. bathyphilus L gemellarii L ater L cuprarius L. lineatus 'Based on single capture; included because of additional unpublished data. 91 FISHERY BULLETIN: VOL. 85, NO. 1 are separated among these zones as well (Table 6; Fig. 4d). Diaphus dumerilii (300 m MDO) and Noto- lychnus valdiviae and Benthosema suborbitale (400 m MDO) are upper mesopelagic, while Myctophum affine appears to be middle mesopelagic (based also on recent unpublished data of J. V. Gartner). Cerato- scopelus warmingii inhabits lower mesopelagic depths during the day, and while both Lampanyc- tus alatus and Lepidophanes guentheri have upper mesopelagic MDO's, both are most abundant in the lower mesopelagic zone. Nighttime patterns show that many species fre- quently range shallower than their zones of max- imum abundance (Table 6) and that all but one species (excluding Taaningichthys minimus and T. paurolychnus) vertically migrate. Species from four genera (Myctophum, 4 spp.; Hygophum, 4 spp.; Cen- trobranchus nigroocellatus; Gonichthys cocco) are regularly found in the surface waters (0-5 m) at night. However, none of the Hygophum species have their highest abundances in this stratum. Diaphus species show several groupings at night though most occur below 100 m. The Lampanyctus species are all most abundant below 75 m, although individuals of L. nobilis are taken as shallow as 50 m. Taaningichthys bathyphilus, the only Taaning- ichthys species for which we have day and night data characteristically shows no evidence of vertical migration (see also Clarke 1973; Backus et al. 1977). This deep night group also includes nonmigratory individuals of many species. Of the seven abundant species, four have night- time MDO's at 75 m which coincide with their zones of maximum abundance (Table 6; Fig. 4d). A fifth species. A'', valdiviae, while also most abundant below 75 m, occurs as shallow as 50 m. Benthosema suborbitale is also captured as shallow as 50 m, but shows no particular zones of abundance at night. Myctophum affine is most abundant at the surface. Vertical partitioning and species associations have also been noted by Karnella (1983). Factorial anal- ysis of abundant species collected from the Bermuda "Ocean Acre" project resulted in from five to eight daytime and six to eight nighttime discrete species associations, depending on season. The Ocean Acre location is in a subtropical region (Backus et al. 1977) and 45 of the 63 species reported by Karnella (1983) also occur in the Gulf. Despite differences in species abundances and general faunal structure and the fact that factorial analysis was inapplicable to the present data set, there are similarities in species associations both day and night between our studies. Karnella (1983) also showed generally shallow day- time distributions for Diaphus species, Lobianchia gemellarii and Notolychnus valdiviae (the only species which is abundant for both studies), and deep (>700 m) daytime distributions for most Lampanyc- tus species and Ceratoscopelus warmingii. At night, his surface group included Myctophum and Hygo- phum species, Centrobranchus nigroocellatus, and Gonichthys cocco. Middle groups (30 and 50 m MDO) included Benthosema suborbitale, Ceratoscopelus warmingii, and Hygophum species, while most Lampanyctus species were in the deeper (>100 or 200 m, depending on season) groups. The abundant species also tended to be divided among several depth groups both day and night. Our findings are also in general agreement with the distribution and abundance ranges reported by other authors for the same species (Clarke 1973; Badcock and Merrett 1976; Nafpaktitis et al. 1977). Although vertical stacking of species groups is ap- parent, it is also obvious that the nighttime overlap of peak abundances among most of the abundant species is quite pronounced (Fig. 4d). It may be that vertical partitioning is on a much finer scale than the present data can resolve or that temporal par- titioning of the same depth stratum may occur. It does not seem that day-night MDO's are linked to the extent of vertical migration, since some species which live relatively deep by day (e.g., Ceratoscope- lus warmigii) are found at the same or shallower depths than those species inhabiting relatively shallow daytime depths (e.g., Diaphus dumerilii). A number of factors including light (e.g., Clarke and Backus 1956, 1964; Paxton 1967; Badcock 1970; Marshall 1980), temperature (Paxton 1967; Robison 1972), and feeding migrations (Marshall 1960) have been suggested as control mechanisms limiting the vertical distribution range of myctophids and other mesopelagic animals. The relationships between these factors and myctophid vertical distributions in the eastern Gulf, however, are not readily ap- parent. Clarke (1973) determined that lunar period at night was an important factor in limiting the up- ward extent of vertical migration for many species, finding that the upper depth of migration was de- pressed by 50 to 125 m during a full moon period for a number of myctophid species (although Hygo- phum species apparently did the opposite, migrating shallower during full moon). In the Gulf, only a single species, Lepidophanes guentheri, showed a markedly shallower depth of capture during a new moon period (10 m vs. 75 m during new and full moons, respectively). All other species tended to show the same upper depths of capture regardless of the lunar phase. In fact, individuals of the 10 Gulf species with nighttime surface captures were reg- 92 GARTNER ET AL.: LANTERNFISHES ularly taken at the surface under a full moon with the deck lights turned on. The effect of temperature in the eastern Gulf is also unclear since the night distributions of almost all species extend through the base of the thermocline and many species enter the mixed layer which is generally isothermal (Fig. 2). Thus, many species encounter the highest tem- peratures found in Gulf surface waters. If, as Marshall (1960) and later researchers sug- gested, nighttime vertical migrations of myctophids and other midwater animals are feeding migrations, a third possible control of depth range would be prey density. In this case, it would be reasonable to assume that zones of maximum potential prey and predator densities would be closely correlated. Anal- ysis of zooplankton catches taken concurrently with our fish trawls show that this is not the case in the eastern Gulf (see Hopkins 1982). Rather, maximum zooplankton biomass of potential forage size organ- isms occurs in the upper 30 m at night in the east- ern Gulf (Hopkins 1982), which is well above the MDO's of all species except surface dwelling Af^cto- phum species, C. nigroocellatus, G. cocco, and occa- sional individuals of Hygophum species and L. guentheri. Size Structure The trend of increasing body size or advancing ontogenetic stage with increasing depth has been demonstrated among myctophids by many workers (Badcock 1970; Gibbs et al. 1971; Clarke 1973; Bad- cock and Merrett 1976; Willis and Pearcy 1980; Hulley 1981; Robison et al.^). Our data, which are confined to the abundant Gulf species, are in general agreement with these earlier findings for all species at night and for most during the day as well. Many myctophid populations have individuals which do not migrate on a daily basis, and these non- migrators are usually small juveniles (Gibbs et al. 1971; Clarke 1973; Badcock and Merrett 1976; Willis and Pearcy 1980). Our data also show this in that at least 19 of the 49 Gulf species had individuals cap- tured at or below daytime depths (Table 3). In com- parison with published accounts of identical species, our data on nonmigratory individuals supports the findings of Clarke (1973) for Benthosema suborbitale, Bolinichthys supralateralis, Ceratoscopelus war- mingii, Lampadena luminosa, and Lampanyctus nobilis off Hawaii, and of Badcock and Merrett (1976) for B. suborbitale, Diaphus rafinesquii, and Hygophum reinhardtii in the eastern North Atlan- tic. Badcock and Merrett also captured C. war- mingii but did not observe nonmigration, possibly because they took no individuals of the deep non- migratory size range. Both Clarke (1973) and Bad- cock and Merrett (1976) reported that Notolychnus valdiviae had a significant nonmigratory fraction of the population, whereas in the Gulf the entire population apparently migrated. Comparison of the size ranges of our abundant species with published sizes of the same species from other tropical-subtropical areas (Clarke 1973; Hulley 1981) show distinctly smaller sizes of adult in- dividuals in the Gulf (Table 7). With a few excep- tions, none of the Gulf species approaches maximum recorded sizes. This may have to do with sampling mechanics (e.g., net mouth area, towing speed, and net avoidance); however, the fact that we have made many additional net hauls since 1977 (20 cruises, ca. 600 discrete depth and oblique samples from 0 to 1,000 m) with a variety of gear and have not sig- nificantly increased the upper size limit of the abun- dant species suggests that this is not the case. A second possibility, which is supported by research on a variety of inshore and offshore species, is that fish species in the Gulf tend to grow faster, with given developmental stages being smaller, and reach maturity at smaller sizes than the same species found outside the Gulf (e.g., Cynoscion nebulosus, Tabb 1961; Micropogonias cromis, White and Chit- tenden 1977; Mycteroperca microlepis, Manooch and Haimovici 1978; Mycteroperca phenax, Godcharles and Bullock 1984; adult Sciaenops ocellatus, Mur- Table 7.— Size range comparisons of dominant eastern Gulf myc- tophid species witii the same species from other tropical- subtropical regions. ^Robison, B. H., T. L. Hopkins, and J. J. Torres. Ecology, phys- iology and nutrient energy dynamics of the Southern Ocean myc- tophid £'/ecfro«a awiardi'ca. Manuscr. in prep. Marine Science Center, University of California at Santa Barbara, Santa Barbara, CA 93106. Clarke Hulley (1981) This (1973) Eastern and Species study Hawaii South Altantic Ceratoscopelus warmingii 14-65 11-79 25-80 Notolychnus valdiviae 9-22 9-25 '^9 Lepidophanes guentheri 13-64 29-76 Lampanyctus alatus 15-48 30-58 Diaphus dumerilii 12-53 25-85 Myctophum affine 12-58 ^28-47 Benthosema suborbitale 10-30 9-38 20-33 ^Based on 1 specimen. 2Based on 13 specimens. 93 FISHERY BULLETIN: VOL. 85, NO. 1 phy and Taylor MS in review^; larval and juvenile S. ocellatus, M. M. Leiby^; deep-sea benthic fishes, K. J. Sulak«). Species Composition, Zoogeographic Affinities, Hydrographic Influence Relatively few accounts have been published on Gulf of Mexico myctophids. Rass (1971) examined material from 5 years of deep otter travel collections and reported 20 species. Bekker et al. (1975) iden- tified 31 species from 19 stations in the southwest- ern Gulf, while Backus et al. (1977) collected 38 species from 7 midwater trawl stations extending from the north central to the southwestern Gulf. Nafpaktitis et al. (1977) recorded 52 species based on a variety of collections and earlier accounts from throughout the Gulf. Murdy et al. (1983) listed 39 species of myctophids taken from 35 Isaacs-Kidd Midwater Trawl stations also located throughout the Gulf. Most recently, Hopkins and Lancraft (1984) reported 34 species from 28 oblique hauls taken with an open net between 0 and 1,000 m at Standard Sta- tion. With the exception of Backus et al. (1977) and Nafpaktitis et al. (1977), these accounts are mainly annotated species lists. Of these earlier studies, only Bekker et al. (1975) and Nafpaktitis et al. (1977) reported myctophid species not found in the present study. Bekker et al. captured one specimen each of Lampadena ano- mala and Lampanyctus festivus. Nafpaktitis et al. listed five species {Diaphus adenormts, D. anderseni, D. metopoclampus, D. minax, Lepidophanes gaussi) from the Gulf which we have not collected. The records of D. anderseni and D. minax, each based on a single specimen from our University of South Florida collections, were found to be misidentifica- tions of D. brachycephalus and D. perspicillatus, respectively. Of the other five species, D. adenomus appears to be epibenthic (Clarke 1973; Hulley 1981) and as such cannot be considered part of the meso- pelagic myctophic assemblage. The captures of Lam- padena anomala, Lampanyctus festivus, and D. metopoclampus were well to the south and west of our study areas. Because of differences in circula- ^Murphy, M. D., and R. G. Taylor. Reproduction, growth and mortality of red drum, Sciaenops ocellatus in Florida. Manuscr. in prep. Department of Natural Resources, Bureau of Marine Research, 100 8th Avenue S.E., St. Petersburg, FL 33701. 'M. M. Leiby, Florida Department of Natural Resources, Bureau of Marine Research, 100 8th Avenue S.E., St. Petersburg, FL 33701, pers. commun. January 1986. *K. J. Sulak, Atlantic Reference Centre, Huntsman Marine Laboratory, St. Andrews, New Brunswick EOG 2X0, Canada, pers. commun. June 1986. tion patterns of western and eastern Gulf waters, these species may never occur as far east as our sampling areas. Thus, the only species that we can- not reconcile is Lepidophanes gaussi, which was reported from the eastern Gulf by Nafpaktitis et al. (1977), although at best this species appears to be an exceedingly rare visitor. Despite these records, we feel that with the data from the present study, the eastern Gulf of Mexico myctophid fauna has been defined. Of our 49 species, 42 were taken on the first three cruises and all 49 were collected by 1976. Despite an additional 20 cruises with approximately 600 mesopelagic trawl samples and over 8 x lO*' m^ water filtered, we have not added a single new species. We also con- clude that all 49 species, including the 6 species listed as rare, are typical components of the eastern Gulf myctophid assemblage during the warm months. In other studies of myctophid assemblages (Clarke 1973; Karnella 1983), the term "rare" is used to designate species whose centers of geo- graphic distribution lie outside the study area but which may occasionally be captured as strays, a definition which does not apply in the present study. With the exception of Hygophum hygomii, whose low numbers are attributed to geographic exclusion by its congener, H. benoiti (Nafpaktitis et al. 1977), the rare species in the eastern Gulf are everywhere rare or extremely uncommon. Data from the upper 1,000 m collected in the eastern Gulf since 1977 show all but the Taaningichthys species, which may occur below our normal fishing depth ranges, to be persistent low abundance members of the myctophid assemblage. Additional evidence of this is the cap- ture of the larvae of Symbolophorus rufinus and Notoscopelus caudispinosus in the eastern Gulf (Houde et al. 1979; W. J. Conley^). The number of myctophid species associated with a particular distribution pattern as defined by Backus et al. (1977) are listed in Table 8. Three of the 49 species captured in the Gulf (Diaphus taa- ningi, Taxiningichthys bathyphilus, T. paurolychnus) are omitted because they have indeterminate geographic distributions. Diaphus taaningi is a pseudooceanic species, associated primarily with land, while the two Taaningichthys species are bathypelagic and do not appear to conform to shallower mesopelagic zoogeographic patterns. Representatives of five of the nine Atlantic distri- bution patterns established by Backus et al. (1977) ^W. J. Conley, University of South Florida, Department of Marine Science, 140 7th Avenue S.E., St. Petersburg, FL 33701, pers. commun. September 1985. 94 GARTNER ET AL.: LANTERNFISHES Table 8.— Distribution patterns of eastern Gulf myctophids. NE - northeast, EC - eastern central, SE - southeastern. Total no. of spec ies No. of species by region Percent total Atlantic Gulf no. of specimens (Backus (Backus Gulf (percent total no . of Backus et al et al (This study) speuimens; et al. This 1977' study 1977) 1977) NE EC SE Temperate- Semisubtropical 8 3 3 2(3.4) 3(4.2) 2(1.5) 0.3 3.9 Subtropical 13 5 7 7(4.0) 6(5.2) 6(6.6) 2.1 5.1 Tropical- Subtropical 18 13 15 11(51.1) 14(40.9) 12(45.4) 64.8 42.4 Tropical- Semisubtropical 5 4 5 5(2.4) 5(4.9) 5(6.2) 1.4 4.6 Tropical 18 12 16 16(37.9) 16(44.6) 13(39.8) 31.0 43.5 'Collections from shallower than 200 m at night. are found in our collections. Species with tropical and tropical-subtropical affinities predictably form the largest component of the Gulf myctophid assem- blage during the summer, comprising almost 70% of the 46 species. The seven abundant species all belong to one of these two faunal associations: three (Ceratoscopelus warmingii, Notolychnus valdiviae, Benthosema suborbitale) are tropical-subtropical; the other four {Lepidophanes guentheri, Lampanyctus alatus, Diaphus dumerilii, Myctophum affine) are tropical. Species with subtropical and temperate- subtropical affinities, however, are poorly repre- sented in our collections. A comparison of species number by sampling locale (NE, EC, SE) within the Gulf reveals no particular pattern (Table 8). Ab- sences from an area are probably due to species rarity or inadequate depth coverage of samples rather than to geographic influence. Backus et al. (1977) characterized the mesopelagic Gulf of Mexico as a special zoogeographic region because of its unique physical and faunal character- istics. Although our collections captured a larger number of species than did theirs (49 vs. 38), the species composition patterns are very similar (Table 8) and indicate that the Gulf myctophid assemblage is overwhelmingly dominated by tropical-subtropical species. However, comparison of the percentage contribution of species within the two collections shows a much different composition (Table 8). Where the data of Backus et al. (1977) showed a 2:1 numerical predominance of tropical-subtropical myc- tophids over tropical species, our findings indicate that the two groups are roughly equal. This dis- crepancy may be due to the fact that Backus et al.'s data were based on collections from the western Gulf which has a different circulation pattern and is less directly influenced by the tropical Loop Cur- rent (Jones 1973). Their data were also from collec- tions <200 m at night (J. E. Craddock^") which could have affected species number and percentages. The largely tropical and tropical-subtropical com- position of the eastern Gulf, during the warmer months at least, is most probably due to the influ- ence of the Loop Current, which may entrain in- dividuals of many uncommon species from the Carib- bean. The size ranges of some of the species taken in our collections support such a hypothesis for tran- sient species. Among the 26 uncommon species, 15 are represented only by either newly metamor- phosed through juvenile or juvenile stages (Table 3). This suggests that the large, sexually mature adults may occur and spawn outside the Gulf, with occa- sional transport of eggs, larvae, and juveniles into the Gulf via the Loop Current. Four of the 15 species, however, are deep-dwelling Lampanyctus (L. ater, L. cuprarius, L. nobilis, L. tenuiformis) and the sexually mature individuals of these species may be present in the Gulf below our normal fish- ing depths, i.e., >1,000 m. Other evidence of Loop Current influence is the relative absence of subtropical species whose geo- graphic distributions usually place them at higher latitudes well to the north or south of the Caribbean Sea and Florida Straits. This is reinforced by the fact that of the six subtropical species we recorded from the Gulf, four {Hygophum reinhardtii, Lam- panyctus ater, L. lineatus, Taaningichthys mini- mus) have uncertain zoogeographic affinities, as they seem to occur often in lower latitudes, including the Caribbean and the Gulf (Backus et al. 1977). Although the Loop Current appears to play an im- portant role in the composition of the eastern Gulf myctophid fauna, the biomass transported appears to be low. Comparison of the 25 northern (NE) non- '"J. E. Craddock, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, pers. commun. August 1978 (reconfirmed July 1986). 95 FISHERY BULLETIN: VOL. 85, NO. 1 Loop Current-influenced stations with the 36 south- ern (SE) Loop Current samples showed that the samples were comparable in diel and vertical depth coverage and that species composition was also similar. However, the number of specimens collected at the NE sites were almost 2.5 times that of the SE total. This apparent impoverishment of mycto- phids in Loop Current waters is supported by the data of other workers showing low biomass of zoo- plankton in the Loop Current (Austin 1971; Jones 1973). Tropical-Subtropical Myctophid Faunal Structure The only other work which has extensively ana- lyzed the myctophid assemblage in a tropical-sub- tropical setting is that of Clarke (1973) in the waters off Hawaii (lat. 21 °N, long. 158°W). Estimates of micronekton standing stock off Hawaii (Maynard et al. 1975) and in the eastern Gulf (Hopkins and Lan- craft 1984) show that both ecosystems are oligo- trophic with roughly equivalent micronekton biomass in the upper 1,000 m. Because of these similarities, a direct comparison between the pres- ent study and the findings of Clarke (1973) is possible. High diversity was apparent in both myctophid assemblages, with 18 genera, 47 species off Hawaii (Clarke 1973) and 17 genera, 49 species in the east- ern Gulf. Of these taxa, the two systems shared 21 species. Excluding 2 bathypelagic species and 2 species of uncertain affiliation, 10 species are tropical-subtropical, 5 are tropical and 2 are tropical- semisubtropical according to Backus et al. (1977). Of the seven numerically dominant species of both studies, three {Benthosema suborbitale, Cerato- scopelus warmingii, Notolychnus valdiviae) are shared (Table 9). In fact, C. warmingii is the top ranked species on both lists. The other dominants on Clarke's list are Pacific congeners of Atlantic- Caribbean species, or are in very closely related genera (Paxton 1972). The percent of the total num- ber of individuals that the top seven species com- prise is strikingly similar (75.5%, Hawaii, 74.7% Gulf). Considering the difference in gear types and sampling strategy, estimates of numbers of individ- uals for the three shared species also agree well. Clarke's (1973) abundance ranges for B. suborbitale during warm months was 14 to 23 x 10^ m^, com- pared with 32 to 42 in the Gulf; C. warmingii ; 55 to 155 vs. 86 to 287; N. valdiviae; 23 to 104 vs. 27 to 128. Thus, the contribution of the abundant species in the myctophid faunas appears to remain relatively constant. Clarke's (1973) data on diel distributions of B. suborbitale, C. warmingii, and N. valdiviae are also similar to our own. Some differences, which may be the result of localized environmental variations, in- clude shallower MDO's, depression of MDO and zones of abundance during full moon and roughly equivalent day-night abundances off Hawaii. Addi- tionally, members of the Hawaiian A^. valdiviae population were found to have a significant non- migratory fraction, which has not been observed in the Gulf population. Clarke (1973) did note, however, that during several collection periods (March and June) no evidence of nonmigration was observed in Notolychnus. Comparison of our species list with those compiled from other studies encompassing tropical-sub- tropical latitudes (Nafpaktitis and Nafpaktitis 1969, Indian Ocean; Hulley 1972, SW Indian Ocean; Clarke 1973, central Pacific; Wisner 1976, eastern Pacific; Hulley 1981, eastern and South Atlantic) showed that 10 myctophid species (8 mesopelagic, 2 bathypelagic) were common to all regions and that three {Benthosema suborbitale, Ceratoscopelus war- Table 9.— Comparison of the top seven myctophid species and their percentage composition for the tropical-subtropical community off Hawaii (from Clarke, 1973) and the eastern Gulf of Mexico (the pres- ent study). Underline indicates shared species. Hawaii Eastern Gulf of Mexico Percent of Percent of total total Species No. specimens Species No. specimens Ceratoscopelus warmingii 3,911 20.7 Ceratoscopelus warmingii 2,267 17.0 Lampanyctus steinbecki 2,362 12.5 Notolychnus valdiviae 1,780 13.4 Tripiioturus nigrescens 2,120 11.2 Lepidophanes guentheri 1,610 12.1 Lampanyctus niger 1,946 10.3 Lampanyctus alatus 1,418 10.7 Boliniclithys longipes 1,458 7.7 Diaphus dumerilii 1,279 9.6 Notolychnus valdiviae 1,267 7.0 Myctophum affine 893 6.7 Benthosema suborbitale 1,157 6.1 Benthosema suborbitale 687 5.2 Total 75.5 Total 74.7 96 GARTNER ET AL.: LANTERNFISHES mingii, Notolychnns valdiviae) were relatively abun- dant everywhere. Distributional affinities of six of the eight mesopelagic species (the three preceding species plus Diogenichthys atlanticus, Lampadena luminosa, and Myctophum nitidulum) are tropical- subtropical; the other two {Lampanyctus nobilis and L. tenuiformis) are tropical (Backus et al. 1977). These species represent 5% to 17% of the total number of species from each region and between 9% and 38% of the total number of specimens ex- amined (Wisner 1976 excluded). Thus, based on our comparison with Clarke's (1973) Hawaiian mycto- phid assemblage and the species lists from various regions of the world ocean, our findings on the eastern Gulf of Mexico myctophid fauna provide additional support to the idea that oligotrophic low- latitude mesopelagic habitats show considerable structural and ecological uniformity allowing for circumglobal distribution of a number of species, a pattern similar to that demonstrated to an even greater degree by bathypelagic fish species (Mar- shall 1980). ACKNOWLEDGMENTS Shiptime was provided by the Florida Institute of Oceanography and UNOLS (University-National Oceanographic Laboratory System). We thank Donald Wilson of the Office of Naval Research, Naval Research Laboratory for arranging shiptime on the RV Mizar. The research was funded by Na- tional Science Foundation contracts DES 75-03845, OCE 75-03845, and OCE 84-10787. The senior author especially wishes to thank the Gulf Oceano- graphic Charitable Trust for providing financial assistance. We are grateful to the following Univer- sity of South Florida Marine Science (USFMS) per- sonnel: Denise Bombard for typing the tables, Deborah Walton for preparation of the figures, and David Williams for photographic layout. Finally, thanks are extended to Mark Leiby, Florida Depart- ment of Natural Resources, Bureau of Marine Re- search; Jose Torres and Tom Lancraft, USFMS; and the two anonymous reviewers for their critical review of the manuscript. LITERATURE CITED Austin, H. M. 1971. The characteristics and relationships between the calculated geostrophic current component and selected in- dicator organisms in the Gulf of Mexico Loop Current system. Ph.D. Thesis, Florida State University, Tallahasee, 369 p. Backus, R. H., J. E. Craddock, R. L. Haedrich, and B. H. ROBISON. 1977. Atlantic mesopelagic zoogeography. In R. H. Gibbs, Jr. (editor), Fishes of the western North Atlantic, p. 266- 287. Sears Found. Mar. Res., Yale Univ., Mem. 1, Pt. 7. Badcock, J. 1970. The vertical distribution of mesopelagic fishes collected on the SOND Cruise. J. Mar. Biol. Assoc. U.K. 50:1001- 1044. Badcock, J., and N. R. Merrett. 1976. Midwater fishes in the eastern North Atlantic-I. Ver- tical distribution and associated biology in 30°N, 23°W, with developmental notes on certain myctophids. Prog. Oceanogr. 7:3-58. Bekker, V. E., Y. N. Shcherbachev, and V. M. Tchuvasov. 1975. Deep-sea pelagic fishes of the Caribbean Sea, Gulf of Mexico and Puerto Rican trench. [In Russian] Tr. Inst. Okeanol. Akad. Nauk USSR 100:289-336. Clarke, G. L., and R. H. Backus. 1956. Measurements of light penetration in relation to ver- tical migration and records of luminescence of deep-sea animals. Deep-Sea Res. 4:1-14. 1964. Interrelations between the vertical migration of deep scattering layers, bioluminescence, and changes in daylight in the sea. Bull. Inst. Oceanogr. Monaco 64(1318): 1-36. Clarke, T. A. 1973. Some aspects of the ecology of lanternfishes (Mycto- phidae) in the Pacific Ocean near Hawaii. Fish. Bull., U.S. 71:401-434. El-Sayed, S. Z. 1972. Primary productivity and standing crop of phytoplank- ton. In V. C. Bushnell (editor). Chemistry, primary produc- tivity and benthic algae of the Gulf of Mexico. Serial Atlas of the Marine Environment, p. 8-13. Am. Geophys. Soc, Folio 22. Gibbs, R. H., Jr., R. J. Goodyear, M. J Keene, and D. W. Brown. 1971. Biological studies of the Bermuda Ocean Acre. II. Vertical distribution and ecology of the lanternfishes (fam- ily Myctophidae). Report to the U.S. Navy Underwater Systems Center, 141 p. Godcharles, M. F., and L. H. Bullock. 1984. Age, growth, mortality and reproduction of the scamp, Mycteroperca phenax (Pisces: Serranidae). FL. Dep. Nat. Res. Bur. Mar. Res. PL88-309 Annu. Rep. FY 82-83 to NMFS Jan. 1984, 30 p. Hopkins, T. L. 1982. The vertical distribution of zooplankton in the eastern Gulf of Mexico. Deep-Sea Res. 29:1069-1083. Hopkins, T. L., and T. M. Lancraft. 1984. The composition and standing stock of mesopelagic micronekton at 27°N, 86°W in the eastern Gulf of Mexico. Contrib. Mar. Sci. 27:143-158. Hopkins, T. L., R. C. Baird, and D. M. Milliken. 1973. A messenger-operated closing trawl. Limnol. Oceanogr. 18:488-490. Houde, E. D., J. C. Leak, C. E. Dowd, S. A. Berkeley, and W. J. Richards. 1979. Ichthyoplankton abundance and diversity in the eastern Gulf of Mexico. Report to the Bureau of Land Management, Contract No. AA550-CT7-28, 546 p. NTIS PB-299 839. HULLEY, P. A. 1972. A report on the mesopelagic fish collected during the deep-sea cruises of R.S. 'Africana IF, 1961-1966. Ann. S. Afr. Mus. 60(6):197-236. 1981. Results of the research cruises of F.R.V. "Walther Her- 97 FISHERY BULLETIN: VOL. 85, NO. 1 wig" to South America. LVIII. Family Myctophidae. (Oste- ichthyes, Myctophiformes). Arch. Fischereiwiss. 31:1-300. HULLEY, P. A., AND G. KREFFT. 1985. A zoogeographic analysis of the fishes of the family Myctophidae (Osteichthyes, Myctophiformes) from the 1979- Sargasso Sea expedition of R.V. "Anton Dohrn". Ann. S. Afr. Mus. 96(2): 19-53. Jones,, J. I. 1973. Physical oceanography of the northeast Gulf of Mex- ico and Florida continental shelf area. In State University System of Florida (Coordinator), A summary of knowledge of the eastern Gulf of Mexico. See IIB. FL. Inst. Oceanogr., St. Petersburg, FL, p. Ml. Karnella, C. 1983. The ecology of lanternfishes (Myctophidae) in the Ber- muda Ocean Acre. Ph.D. Thesis, George Washington Univ. , Washington, D.C., 499 p. Leipper, D. F. 1970. A sequence of current patterns in the Gulf of Mexico. J. Geophys. Res. 75:637-657. Manooch, C. S., Ill, AND M. Haimovici. 1978. Age and growth of the gag, Mycteroperca microlepis, and size-age composition of the recreational catch off the southeastern United States. Trans. Am. Fish. Soc. 107: 234-240. Marshall, N. B. 1960. Swimbladder structure of deep-sea fishes in relation to their systematics and biology. Discovery Rep. 31:1-122. 1971. Explorations in the lives of fishes. Harvard Univ. Press, Cambr., MA, 204 p. 1980. Deep-sea biology. Developments and perspectives. Garland STPM Press, N.Y., 566 p. Maul, G. A. 1977. The annual cycle of the Gulf Loop Current. Part I: Observations during a one-year time series. J. Mar. Res. 35:29-47. Maynard, S. D., F. V. RiGGS, AND J. F. Walters. 1975. Mesopelagic micronekton in Hawaiian waters: faunal composition, standing stock and diel vertical migration. Fish. Bull., U.S. 73:726-736. McGlNNis, R. F. 1974. Biogeography of lanternfishes (family Myctophidae) south of 30°S. Ph.D. Thesis, Univ. Southern California, Los Angeles. Molinari, R. L., and D. A. Mayer. 1980. Physical oceanographic conditions at a potential OTEC site in the Gulf of Mexico: 27.5°N, 85.5°W. NOAA Tech. Mem. ERL AOML-42, 82 p. Murdy, E. 0., R. E. Matheson, Jr., J. D. Fechhelm, and M. J. McCoiD. 1983. Mid water fishes of the Gulf of Mexico collected from the RA' Alaminos, 1965-1973. Tex. J. Sci. 35:109-127. Nafpaktitis, B. G. 1974. A new record and a new species of lanternfish, genus Diaphus (Family Myctophidae), from the North Atlantic Ocean. Los Ang. Cty. Nat. Hist. Mus. Contrib. Sci. No. 254, 6 p. Nafpaktitis, B. G., R. H. Backus, J. E. Craddock, R. L. Haedrich, B. H. Robison, and C. Karnella. 1977. Family Myctophidae. In R. H. Gibbs, Jr. (editor). Fishes of the western North Atlantic, p. 13-265. Sears Found. Mar. Res., Yale Univ., Mem. 1, Pt. 7. Nafpaktitis, B. G., and M. Nafpaktitis. 1969. Lanternfishes (Family Myctophidae) collected during Cruises 3 and 6 of the R/V Anton Bruun in the Indian Ocean. Bull. Los Ang. Cty. Nat. Hist. Mus. Sci. No. 5, 79 P- NowLiN, W. D., Jr. 1971. Water masses and general circulation of the Gulf of Mexico. Oceanol. Int. 6:28-33. NowLiN, W. D., Jr., and H. J. McLellan. 1967. A characterization of the Gulf of Mexico waters in winter. J. Mar. Res. 25:29-59. Paxton, J. R. 1967. Biological notes on southern California lanternfishes (family Myctophidae). Calif. Fish Game 53:214-217. 1972. Osteology and relationships of the lanternfishes (fam- ily Myctophidae). Bull. Los Ang. Cty. Mus. Nat. Hist. Sci. 13:1-81. Rass, T. S. 1971. Deep-sea fish in the Caribbean Sea and the Gulf of Mex- ico (The American Mediterranean Region). Symp. Invest. Res. Carib. Sea and Adj. Reg. UNESCO, Paris 1971:509- 526. Robison, B. H. 1972. Distribution of the midwater fishes of the Gulf of California. Copeia 1972:448-461. Sturges, W., and J. C. Evans. 1983. On the variability of the Loop Current in the Gulf of Mexico. J. Mar. Res. 41:639-653. Tabb, D. C. 1961. A contribution to the biology of the spotted seatrout, Cynoscion nebulosus (Cuvier) of east-central Florida. Fla. Board Conserv., Mar. Res. Lab., Tech. Ser. 35, 22 p. VuKOViCH, F. M., and G. a. Maul. 1985. Cyclonic eddies in the eastern Gulf of Mexico. J. Phys. Oceanogr. 15:105-117. White, M. L., and M. E. Chittenden, Jr. 1977. Age determination, reproduction and population dynamics of the Atlantic croaker, Micropogonias undulatus. Fish. Bull., U.S. 75:109-123. Willis, J. M., and W. G. Pearcy. 1980. Spatial and temporal variations in the population size structure of three lanternfishes (Myctophidae) off Oregon, USA. Mar. Biol. (Berl.) 57:181-191. Wisner, R. L. 1976. The taxonomy and distribution of lanternfishes (Family Myctophidae) of the eastern Pacific Ocean. U.S. Gov. Print. Off., Wash., D.C., 229 p. 98 LIFE HISTORY AND FISHERY OF THE CALIFORNIA SCORPIONFISH, SCORPAENA GUTTATA, WITHIN THE SOUTHERN CALIFORNIA BIGHT Milton S. Love,i Brita Axell,i Pamela Morris, ^ Robson Collins,^ AND Andrew Brooks' ABSTRACT We examined the life history of the California scorpionfish in the Southern California Bight. Based on sportfish creel census data, the species was most abundant in the southern part of the Bight, particular- ly around Catalina, San Clemente, and the Coronado Islands. Trawl studies from 1974 to 1984 indicated that California scorpionfish populations varied considerably in abundance, with numbers peaking in 1982. Though the species usually associates with hard substrata, it was abundant over mud about the Palos Verdes Peninsula, site of a major sewage outfall. We think that this anomalous abundance was due to the presence of large numbers of a prey species, the ridgeback prawn, Sicyonia ingentis, which was attracted to the nutrient-rich substrata. Female California scorpionfish lived to 21 years, males to 15. Females grew faster than males. Von Bertalanffy age-length parameters for females were L = 44.3, k = 0.13, Iq = -1.9, and for males L = 36.3, k = 0.12, ^Q = -3.86. Over 50% of both females and males were mature at 2 years of age. Males tended to mature at a slightly smaller size. Spawning occurred from May through August, peaking in July. California scorpionfish formed large offshore spawning aggregations in waters deeper than their off-season habitat. Tagging results indicated that fish return to the same spawning area annually. Crabs, primarily juvenile Cancer anthonyi, were the most important food item of fishes inhabiting soft substrata in shallow water. The family Scorpaenidae is represented by four genera in the northeastern Pacific— Scorpaena, Scorpaenodes, Sebastes, and Sebastolobus (Esch- meyer et al. 1983). One Scorpaena species, S. gut- tata, the California scorpionfish, is abundant as far north as southern California. The California scorpionfish is a medium-sized [to 43 cm TL (total length)], generally benthic species, found from central California into the Gulf of California between the intertidal and 183 m (Esch- meyer et al. 1983). It occurs on rocky reefs (often lodged in crevices), although in certain areas and seasons it aggregates over sandy or muddy sub- strata (Frey 1971; present paper). This species is oviparous, producing floating, gelatinous egg masses in which the eggs are embedded in a single layer (Orton 1955). Like others in the genus Scor- paena, California scorpionfish produce a toxin in their dorsal, anal, and pelvic spines, which produces intense, painful wounds. California scorpionfish comprise a minor part of the California sport and ^VANTUNA Research Group, Occidental College, Moore Laboratory of Zoology, Los Angeles, CA 90041. ^Marine Resource Branch, California Department of Fish and Game, 1301 West 12th, Long Beach, CA 90813. commercial fisheries (Wine and Hoban^, Wine*, Knaggs^, present paper). Perhaps because of this relatively small catch, the species has not been the subject of an in-depth life history study. Rather, much of what is known has been gleaned from larger ecological surveys (Table 1), in which the species played a minor role. How- ever, California scorpionfish have recently become important in pollution-related studies (Table 1), deriving from 1) its abundance about the Palos Verdes Peninsula (heavily polluted from the Whites Point sewage outfall which services Los Angeles), 2) its ease of capture by otter trawl and by hook and line, and 3) its ability to adapt to laboratory aquaria. This increased interest has given rise to questions regarding the species' growth rate, age at first maturity, and movements. Our paper details some Manuscript accepted September 1986. FISHERY BULLETIN: VOL. 85, NO. 1. 1987. ^Wine, v., and T. Hoban. 1976. Southern California indepen- dent sportfishing survey annual report, July 1, 1975-June 30, 1976 Calif. Dep. Fish Game, 109 p. ^Wine, V. 1979. Southern California independent sportfishing survey annual report, July 1, 1977-June 30, 1978. Calif. Dep. Fish Game, 100 p. ^E. Knaggs, California Department of Fish and Game, Long Beach, CA, pers. commun. May 1985. 99 FISHERY BULLETIN: VOL. 85, NO. 1 Table 1. — Previous studies involving Scorpaena guttata. Not in- cluded are geographical species lists. Systematics. —Gnard 1854, 1858; David 1943; Phillips 1957; Tsuyuki et al. 1968; Eschmeyer and Bailey 1970; Greenfield 1974. Anatomy and Physiology.— C\o\h\er 1950; Halstead 1951; Halstead et al. 1955; Saunders 1959; Halstead and Mitchell 1963; Taylor 1963; Munz 1964; Russell 1965, 1969; Carlson etal. 1971;Schaef- fer et al. 1971; Baines 1975; Sullivan and Somero 1980. Pollutant Levels and Effects.— MacGregor 1972; Young and Mearns 1978; Stout and Beezhold 1981; Brown et al. 1982, 1984a-c; Gossett et al. 1982a, b, 1984; Jenkins et al. 1982; Mearns 1982; Schafer et al. 1982; Szalay 1982; Gadbois and Maney 1983; Bay etal. 1984a, b; Perkins and Rosenthal 1984; Rosenthal etal. 1984; Cross et al. 1985. Life History, Distribution, and Behavior. — Jordan and Gilbert 1881 Holder 1900; Richardson 1905; Wilson 1908, 1935; Barnhart 1932 David 1939; Limbaugh 1955; Orton 1955; Montgonnery 1957 Causey 1960; Kunnenkeri and Martin 1963; Rosenblatt 1963 Taylor 1963; Aral and Koski 1964; Carlisle et al. 1964; Clarke et al. 1967; Quast 1968a-c; Carlisle 1969; Cressey 1969; Taylor and Chen 1969; Turner etal. 1969; Frey 1971; Hobson 1971; Ho 1972; Miller and Lea 1972; Varoujean 1972; Feder et al. 1974; Allen et al. 1976; Burreson 1977; Mearns 1979; Daileyetal. 1981; Helvey and Dorn 1981; Hobson et al. 1981; Stephens and Zerba 1981; Eschmeyer etal. 1983; Love and Moser 1983; Barnettetal. 1984; DeMartini and Allen 1984; Larson and DeMartini 1984; Thresher 1984; Love and Westphal 1985. F/s/7ery.— Phillips 1937; Daugherty 1949; Roedel 1953; Frey 1971. aspects of the growth, reproduction, food habits, movements, and fisheries of the California scorpion- fish. METHODS Distribution and Movements of Adults and Juveniles To estimate relative abundance of California scorpionfish over reefs and hard substrata, we used the California Department of Fish and Game creel census data, gathered from throughout the Southern California Bight from April 1975 to December 1978. In this study. Fish and Game personnel rode aboard randomly chosen commercial passenger vessels (hereafter referred to as "partyboats") and mea- sured and identified all fish captured. The sampler also noted numbers of anglers, fishing hours, and location and depth of each fished site. Catch per unit effort was used as our estimate of relative abundance, where effort was measured in angler hours (number of anglers x number of hours fished). For several reasons, data from this study could not give a completely unbiased estimate of Califor- nia scorpionfish abundance. First, virtually all fish- ing effort aboard partyboats occurs over reefs and hard substrata. Hence, this data base does not ef- fectively measure abundance over soft substrata. Second, most angling involved fishing with live bait (primarily northern anchovies, Engraulis mordax) or with lures simulating fishes. Thus the sample was biased away from very small individuals. However, California scorpionfish develop relatively large mouths and become mesocarnivores at relatively early sizes and our data indicates that most size classes were represented. As angling techniques were similar throughout the Southern California Bight, we believe this survey allows for an accept- able representation of abundance of all but the smallest size classes. To measure relative abundance of California scorpionfish living on soft substrates, we used trawl data collected by the Southern California Coastal Water Research Project (SCCWRP) and the Orange County Sanitation District. These data were based on 10-min tows of a 7.6 m headrope otter trawl fished on the bottom at about 23, 61, and 137 m off Palos Verdes and Huntington Beach (trawling sta- tions are illustrated in Cross [1985] and Orange County Sanitation District^). We analyzed data taken about Palos Verdes and Huntington Beach from January 1974 to December 1984 (except that no data was taken for Huntington Beach during 1975). Fishes in this survey were measured using standard length. We converted these measurements to total lengths using conversion factors based on measurements of 1,083 California scorpionfish. These factors are TL = (1.21)SL + 1.02; SL = (0.82)TL - 0.69. We also conducted a tagging program to give some insight into this species' movements. We tagged trawl-caught California scorpionfish with Floy'^ tags (orange FD-68BC) from an area between the southern part of Santa Monica Bay and Hunting- ton Beach. The tags (consisting of a plastic tube 50 mm long with a 10 mm cross bar) were injected into the dorsal musculature between the second and third dorsal spines, leaving the brightly colored end free. Most of the tagging effort was centered on Dago Bank, about 11 km southeast of Long Beach Harbor— an area we had identified as a spawning ground. A monthly otter trawl survey indicated that California scorpionfish were rare in this area be- tween October and April, with large numbers of ripe individuals occupying the habitat during late spring and summer. ^Orange County Sanitation District. 1984. Annual Report, 300 P- 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 100 LOVE ET AL.: LIFE HISTORY AND FISHERY OF CALIFORNIA SCORPIONFISH Juveniles and Adults Collection Individuals used in the analysis of age and length, length-weight, and reproduction were sampled monthly (2-10 samples/month) from May 1981 to June 1982 (and sporadically thereafter through May 1983). We used a 7.6 m (25 ft) or 4.9 m (16 ft) head- rope otter trawl in 7-90 m of water, between Ven- tura and San Onofre, CA. All specimens were frozen for later dissection. After thawing, all fish were measured (total length and standard length), weighed, and sexed, and the gonads were removed and weighed. Fish for food habit studies were taken by otter trawl over soft substrata in 6-16 m of water between Santa Monica Bay and San Onofre. These samples were frozen immediately after collection. In the laboratory, food items were identified to lowest possible taxa, then weighed and counted. Techniques for Aging Juveniles and Adults We attempted to age California scorpionfish by a variety of calcified structures (sagittae, scales, vertebrae, and pterygiophores) and found that cross sections of anal pterygiophores gave best results. The fused first and second anal pterygiophores (supporting the first and second anal spines) were removed from 613 specimens, cleaned, and stored in paper coin envelopes. Pterygiophores were placed on wood blocks and embedded in clear epoxy (Ciba 825 hardener and Ciba 6010 resin). Each block with its pterygiophore was placed on a Buehler Isomet low speed saw and an 0.05 cm wafer was cut through it, using two diamond-edge blades separated by a stainless steel shim. The cut was made near the pterygiophore' s site of articulation with an anal spine. Wafers were read under a compound micro- scope at a magnification of 100 x , with both reflected and transmitted light. All wafers were read twice, by M. S. Love, approximately 6 mo apart. When readings did not agree, they were read again. The value of two coincident readings was ac- cepted as the best estimate of age. We compared the age-length curves of males and females using an analysis of variance of regression coefficients over groups, testing the slopes of the two curves (Dixon 1981). Parenthetically, this was the same test used in comparing male and female length-weight curves. Back calculations of length on age were made using the techniques of Chen (1971). Procedures for Determining the Timing of Maturation and Reproduction We estimated length at first maturity by classify- ing gonads as immature or matured based on the techniques of Bagenal and Braum (1971). Smaller mature fish and fish just entering their first repro- ductive season become reproductive later in the year. Hence we estimated length at first maturity from just before spawning season (March) through its conclusion (August). A gonadosomatic index [(gonad weigh t)/( total body weight) x 100] was com- puted from frozen specimens to quantify changes in gonad size with season. (W - GW) We computed condition factor (100 x -, where W = body weight in grams, GW = gonad weight in grams, and L = total length in centi- meters), of mature California scorpionfish. Condi- tion factor was computed using body weight with gonad weight subtracted so as to minimize the ef- fects of seasonal changes in gonad size. We com- pared these values between seasons within sexes and between sexes, using the Mann- Whitney U-Test (Sokal and Rohlf 1969). Fishery To describe the California scorpionfish's role in the commercial passenger vessel (partyboat) sport fishery, we used the previously discussed Califor- nia Department of Fish and Game creel census data. We also examined the commercial fishery, inter- viewing fishermen and utilizing the fish landing data of the California Department of Fish and Game. RESULTS AND DISCUSSION Distribution and Movements Data from the Fish and Game creel census in- dicated differences in abundance between the north- ern and southern part of the Southern California Bight (Fig. 1). Catch rates were lowest near the city of Santa Barbara and generally increased to the south. Highest catch rates occurred off San Diego and around Catalina, San Clemente, and the Coro- nado Islands. Utilizing the same data base, we examined Califor- nia scorpionfish depth distribution (Fig. 2). Overall, California scorpionfish were taken from barely sub- tidal waters to 170 m. Depth distribution changed with season. We plotted catch per unit effort in 6 101 FISHERY BULLETIN: VOL. 85, NO. 1 1 1 1 1 1 1 1 1 1 1 1 1 \\ \ s M o o CO 0) o ^ £ ?i o ^ <= ^1 ^^ !- -S ^ c -CI — .^^ 3 ^ O) a. 0) I I 102 LOVE ET AL.: LIFE HISTORY AND FISHERY OF CALIFORNIA SCORPIONFISH y/. MAY -SEPT. OCT. -APRIL 10-3 10-^ 3 10-^ "io-« 10-7 - i 1 \ \ 1 0-30 31-60 61-90 DEPTH(M) 91-120 121-150 151-180 Figure 2.— Abundances (based on catch per unit effort in the partyboat sport fishery) of California scorpionfish tal A u -I « 1.3 10 .FEMALES • MALES -jia — rts — TKkn — jjpR — wAt — jdnn — jot? — wis — s^pt — jsir — Riv — tsfer MONTH 7.0 6.0 S.0 (0 111 4.0 < s ui 3.0 10 1.0 Figure 9.— Seasonal changes in the gonosomatic indices (GSI = gonad weight as a p*centage of total body weight) of female and male California scorpionfish (based on 396 females and 654 males). Vertical lines indicate 95% confidence intervals of the mean. We know little of the location of California scorpionfish larvae in the Southern California Bight. Over the past 30 -i- yr, few have been taken in off- shore waters despite considerable numbers of ich- thyoplankton surveys (Moser^). Moreover, only a few are known from ichthyoplankton surveys con- ducted in inshore waters (Barnett et al. 1984; McGowan^^). Particularly puzzling is the lack of lar- 'G. Moser, Southwest Fisheries Center La Jolla Laboratory, Na- tional Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038, pers. commun. September 1985. '"G. McGowan, Natural History Museum of Los Angeles County, vae taken in King Harbor, Redondo Beach. No larvae were caught during a 7-yr monthly survey, of both surface and bottom waters (Jordan ^i) despite the abundance of young-of-the-year and 1-yr-old fish in the Harbor. It appears that California scorpionfish utilize an "explosive breeding assemblage" reproductive mode 900 Exposition Blvd., Los Angeles, CA 90007, pers. commun. May 1985. "G. Jordan, VANTUNA Research Group, Occidental College, 1600 Campus Road, Los Angeles, CA 90041, pers. commun. May 1985. 108 LOVE ET AL.; Llf^E HISTORY AND FISHERY OF CALIFORNIA SCORPIONFISH (Emlen and Oring 1977), in which fish migrate to, and aggregate at a "traditional" spawning site for brief (though undefined) periods. Reproduction is polygamous and sexual selection is low. Thresher (1984) speculated such behavior may be the primary reproductive mode of larger pelagic spawning reef fishes— such as snappers, jacks, and barracudas. Smaller fishes would find a spawning migration dele- terious, owing to a higher predation risk while traveling. California scorpionfish do indeed migrate to "tra- ditional" spawning areas and are pelagic spawners. With 50% maturing at 17 cm, they are smaller than the usual explosive breeding assemblage species listed by Thresher (1984). However, it is likely that mature California scorpionfish are not heavily preyed upon (because of their toxin-carrying spines) and thus may be an exception to the rule. The Dago Bank spawning site is, for the most part, a sandy environment, usually inhabited by relatively few fish. Spawning in a deep-water, rela- tively depauperate area, the California scorpionfish may avoid some of the egg predation expected in the shallow reefs inhabited during fall-spring. More- over, by spawning well above the substrata, newly spawned eggs are kept away from benthos-dwelling predators. Many coral reef fishes exhibit the same behavior, which not only decreases egg predation but also places the fertilized eggs into surface cur- rents, increasing the chances of larval dispersal (Thresher 1984). We do not know how many spawning sites exist off southern California. Santa Monica Bay (Turner et al. 1969) and Dana Point (Cross fn. 6) are likely sites while Anacapa Island and the Coronado Islands (M. Love, unpubl. data) might also be utilized. We have no data on how many years these sites persist as spawning areas. Judging from other species (such as Clupea harengus—Cushing 1982), it is likely that scorpionfish spawning grounds are probably of long duration. For several reasons, this behavior is unusual among fishes in southern California. First, only a few species (notably kelp bass, Paralabrax clathra- tus; barred sand bass, P. nebulifer; sargo, Aniso- tremus davidsoni; kelp surfperch, BrachyistiiLS fre- natus; senorita, Oxyjulis califomica; and sheephead, Pimelometopon pulchrum, Feder et al. 1974) form relatively long-term (to a few months) spawning ag- gregations. It is noteworthy that, of these fishes, all except the barred sandbass are midwater, active, species particularly when compared with crevice- dwelling scorpionfish. Second, few reef associated species move off reefs to spawn. Barred sand bass are one of the few ex- ceptions. These form large spawning aggregations over low relief or flat substrata within the Southern California Bight (Turner et al. 1969). The vast ma- jority of reef dwelling fish are relatively sedentary. Many are either territorial or occupy home ranges. Virtually all stay within the reef vicinity. For these species, spawning takes place within their usual habitats. Lastly, the California scorpionfish does not have the morphology of a fish given to long movements. Such adaptations can be seen most graphically among the northeast Pacific rockfishes, genus Sebastes. Sedentary, territorial species, such as the gopher rockfish, 5. carnatus, and treefish, S. ser- riceps, are very spiny, squat, and deep-bodied forms. More active, midwater species, such as the yellow- tail rockfish, S. Jlavidus, and boccacio, S. pauci- spinis, are more streamlined, with reduced spines (particularly about the head). This trend culminates in the pelagic shortbelly rockfish, S. jordani, which resembles a mackerel or sardine. In contrast, California scorpionfish closely resemble the benthic rockfish. Yet, the species seems to move about considerably, even excluding movements to and from spawning grounds. Tagging data from studies of the California Department of Fish and Game show movements as much as 190 km (Hart- mann^^) Food Habits We sampled 24 California scorpionfish (TL = 21.2- 32.5 cm) with food in their stomachs. Though we captured many hundreds of scorpionfish throughout the Southern California Bight, individuals taken in water deeper than about 16 m regurgitated prey during capture. The 24 individuals with prey repre- sented 68.5% (24 of 35) of all scorpionfish taken in water <16 m. We have graphically represented prey importance (Fig. 10), using the Index of Relative Importance (Pinkas et al. 1971). Crabs were the most important food item. These were primarily juvenile Cancer an- thonyi, but we also found a few Loxyrhynchus sp., Randalia omata, and Pagurus sp. Fishes were sec- ond in importance. Recognizable species were the northern anchovy, Engraulis mordax, and the spotted cusk-eel, Chilara taylori. Octopi, isopods, shrimp (pumanly Alphaeus sp.), and small pebbles made up the rest of the diet. i^A. R. Hartmann, California Department of Fish and Game, Long Beach, CA 90802, pers. commun. June 1984. 109 FISHERY BULLETIN: VOL. 85, NO. 1 60 p UJ O'J CQ §40 z t 30 O20 10 ^ 10 20 30 40 z tu O ^50 60 IRI=5,706 CRABS <0 in IRI=1,697 E FISH IRI=596 MISC. -62.5 • -*-«-20.8* 8.3" PERCENT FREQUENCY -33.3- FlGURE 10.— Index of Relative Importance of prey found in stomachs of 24 California scorpionfish, captured in the South- ern California Bight. Turner et al. (1969) examined diets of California scorpionfish living on southern California artificial reefs. They found the species fed almost exclusive- ly upon juvenile Cancer crabs during fall and winter; at other times scorpionfish ate octopus and fish. This is similar to our findings, in which juvenile Cancer anthonyi were the most important prey. Thus, though the habitat we examined was different from that surveyed by Turner et al., California scorpion- fish may sometimes seek out juvenile Cancer crabs, regardless of whatever other potential prey are available. When juvenile crabs are not present, California scorpionfish prey on other forms, in- cluding octopus and fish. Limbaugh (1955), Quast (1968c), and Hobson et al. (1981), surveying California scorpionfish over natural rocky reefs, have all reported roughly equivalent food habits, with demersal crustaceans (particularly crabs and shrimps) of most importance, followed by fishes, octopi, and squids. Hobson et al. speculated the species captured most prey at night. Fishery Within the Southern California Bight, the Califor- nia scorpionfish is a relatively minor constituent of the partyboat sportfish catch (Table 7). The species ranked 15th in abundance, comprising about 1.5% of all fishes taken. As scorpionfish were less abun- dant in the northern part of the Bight, we deleted data from sites north of Pt. Mugu (shown in Figure 1). When species were reranked, scorpionfish moved up to 12th most abundant, forming 1.8% of the catch. Throughout the Bight, over the years 1975-78, the annual contribution of scorpionfish to the total partyboat catch, was fairly constant, hover- ing at about 1.5% (Fig. 11). Most of the scorpion- fish taken aboard partyboats were mature (Fig. 12). The importance of scorpionfish to the total party- boat fishery varied with season (Fig. 11). During the nearly 4 years of the creel census, scorpionfish con- tributed most heavily to the catch (as much as 3.0%) Table 7.— The twenty most commonly taken spe- cies aboard commercial passenger vessels in the Southern California Bight, April 1975-December 1978. A. Rankings for entire Bight, total number of fish sampled = 342,052. B. Southern Califor- nia Bight from Pt. Mugu south, total number of fish sampled = 278,664. Species No. % . . . A 1. Sebastes paucispinis 78,877 23.1 2. Paralabrax clathratus 38,315 11.2 3. Scomber japonicus 35,072 10.3 4. Sebastes goodei 27,218 8.0 5. Sebastes serranoides 19,455 5.7 6. Sarda chiliensis 16,295 4.8 7. Paralabrax nebulifer 13,987 4.1 8. Sebastes mystinus 13,646 4.0 9. Sphyraena argentea 8,391 2.5 10. Genyonemus lineatus 7,841 2.3 11. Sebastes miniatus 7,023 2.1 12. Sebastes chlorostictus 5,505 1.6 13. Sebastes hopkinsi 5,025 1.5 14. Caulolatilus princeps 4,990 1.5 15. Scorpaena guttata 4,976 1.5 16. Medialuna califomiensis 3,990 1.2 17. Sebastes entomelas 3,969 1.2 18. Sebastes rubrivinctus 2,859 0.8 19. Sebastes elongatus 2,568 0.8 20. Sebastes caurinus B Sebastes paucispinis 2,513 0.7 1. 61,962 22.2 2, Scomber japonicus 33,076 11,9 3. Paralabrax clathratus 29,655 10.6 4. Sebastes goodei 21,408 7.7 5. Sarda chiliensis 16,213 5.8 6. Sebastes serranoides 14,987 5.4 7. Paralabrax nebulifer 13,371 4.8 8. Sebastes mystinus 9,083 3.3 9. Sphyraena argentea 8,376 3.0 10. Genyonemus lineatus 7,257 2.6 11. Sebastes miniatus 5,109 1.8 12. Scorpaena guttata 4,880 1.8 13. Caulotilus princeps 4,778 1.7 14. Medialuna califomiensis 3,906 1.4 15. Sebastes hopkinsi 3,747 1.3 16. Sebastes chlorostictus 3,263 1.2 17. Sebastes rubrivinctus 2,381 0.9 18. Anoplopoma fimbria 2,279 0.8 19. Sebastes entomelas 2,101 0.8 20. Sebastes constellatus 2,019 0.7 110 LOVE ET AL.: LIFE HISTORY AND FISHERY OF CALIFORNIA SCORPIONFISH ^^ 'aT Ts 32 TOTAL LENGTH (CM) Figure 11.— Seasonal distribution of California scorpionfish catch in the southern California partyboat sport fishery, 1975-78. species is captured by hook and line, gill net, and, rarely, otter trawl. While hook-and-line catches pre- dominate, gill net landings are also important. In a 1984 study, Collins et al.^^ found that scorpion- fish were the 10th most abundant species in the California halibut gill net fishery. In recent years, the fishery has been almost entirely limited to the later spring and early summer months (Fig. 13), with catches between June and August accounting for about 80% of the total. Traditionally, the bulk of California scorpionfish have been caught by a few fishermen specializing in this species. From our observations, it seems like- ly that the number of specialists has declined markedly since the 1950's. A few vessels of the Newport dory fishery (Cross fn. 8) specialize in fish- 2.5r 2.0 X o O % 03 > ►- C < a. 1.5- 1.0 I- I- u. O oe a. oc UJ o IX. LJ s i - 3 w UJ i £ CO oc 3 C3 oc UJ 7 -i *- — ^ S SF oc UJ 0> O) 01 CO 0> T If) r~ O) 1975 1976 1977 1978 Figure 12.— Size distribution (with 100% maturity len^h) of California scorpionfish taken in the southern California sport fishery, 1975-78. during spring and summer. This is apparently due to some vessel operators targeting spawning aggre- gations. Figure 11 indicates the depths at which Califor- nia scorpionfish were taken over the years 1975-78. During the May- September spawning season, fish were most abundant in 61-90 and 121-150 m. Rela- tively few fish were taken between 61 and 150 m from October to April, with catches ranging from 10 to 10^ times as great in May to September. Similarly, October to April catches were highest in inshore waters. Historically, California scorpionfish were taken commercially by hook and line and, occasionally, round haul nets (Daugherty 1949). Currently, the ing for California scorpionfish and their techniques are illustrative. The fishermen concentrate their ac- tivities on the spawning grounds offshore of Long Beach— the same area we utilized in our tagging study. As the precise time of fish aggregation varies from year to year, occasional exploratory trips are made to the grounds beginning in May. Most catches begin in June and end in August. Using long lines, the fishermen deploy on the bottom 1,200-2,000 hooks (4/0-5/0 long shank) in 600-1,300 m (1,970- 4,265 ft) sets. The hooks (baited with anchovies, '^Collins, R. A., M. M. Vojkovich, and R. J. Reed. 1985. Pro- gress Report, Southern California nearshore gill and trammel net study 1984. Calif. Dep. Fish Game, 40 p. Ill FISHERY BULLETIN: VOL. 85, NO. 1 35 - .30 X o S X o o 25 20 0 15 u. O I- 5 10 o a HI OL JAN MAR MAY JULY SEPT NOV MONTH Figure 13.— Monthly percentile distribution of the com- mercial California scorpionfish catch. Based on 1970-84 California commercial landings. mackerel, or other fish) are usually set about 1 h before sunrise and pulled 1-2 h later. Fishermen report that the fish do not seem to feed well before or after this time. Traditionally, the Newport fishermen sell their catch to the public on the beach next to the New- port Pier. However, fishermen specializing in scorpionfish often sell their entire catch to fish pro- cessors in San Pedro, receiving a relatively high $1.98-$2.75/kg (90 i, CTi : total catch in year t, M : instantaneous natural mortality rate, Ff : instantaneous fishing mortality rate on fully recruited ages in year t, ages i > k, Zf : instantaneous total mortality rate, F" + M, on fully recruited ages, F-^p Z-^ : instantaneous fishing and total mortality rates on pre-fully recruited ages, ages i < k, Ff, Zf : instantaneous mortality rates if there are no differences between juvenile and adult rates, k : age of first full recruitment, U^t. = 1-0 i ^ k, Tf : net recruitment rate in year t. Let us define ^ = Fjil.O - exp(-Z^))/Zj and similarly for <})". By definition, the recruitment rate is rz = new recruits total exploitable A--1 ^ Ni,2 U,, - (exp(-Zi)) I iV,.i U,j (1) ^N,^2 U2 where x is the age of first partial recruitment. This can be rewritten as k-l k-l l=X l-X k-l fe-i I P,,2 + (f,/<|>^) 1 P,,2 i = A; (2) Compare this result with that given by Allen (1966, 1968). He gives A--1 r, = Pi,2 - 2: (1.0 - T,/B,^,yP,,,^„ 1=1 where 5,_i = P,+i,2-Qam /(Pu"Qa-+i,2) T, = exp(Z;' - Zi). Consequently, *: A--1 r, = Ip,.2 - Ti(Q,,i,2/Q^.i) 2: P,i. (3) l=X l=X To satisfy the above Equations (2) and (3) it is found that 21. The age subscripts have now been dropped from Z and F. Consequent- ly the numbers in the second year are given by A^i,2 = A -5,000. A^,>i,2 = N,AU,,expi-Z,) + (1.0 - [/,i)exp(-M)] ^21,2 = (A^2o.i + A^2i,i) exp(-Zi) Prj = {N,,t f/^)/? i^^,t Uu)- 21 I (5) 1=1 (4) If we wish to consider the effects of stochastic catches at age or problems in aging, then Pj , be- comes a variable, Pj,, and can be expressed in terms of the catch, so that 119 where C, , is determined as an independent random variable. For the expected catch at age, m > 50 Cjj is distributed as A''[m, „m, , (1 - P,t)] where Pit is calculated from Equation (5), and for m < 50 as Poisson [m^t]. m is obtained as and m,,, = F,(1.0 - exvi-Zt))N,, U^jlZ^, CT,= Iq,. If aging of the catch introduces bias or variance this can be investigated using a matrix A where the ele- ment ttj J is the probability that an animal of true age i will be called j. The new catch at allocated age can be given by C where C f = A Cf, and where C is the column vector of catch at age. From the validation model the true recruitment rate can be calculated as FISHERY BULLETIN: VOL. 85, NO. 1 (i) Fj ^ F, From Equation (6) it is evident that the value of Fo {F2 T^ 0) does not affect the recruitment rate and this is reflected in Equation (4) where only propor- tions in the catch each year are needed. The value does however affect the variance of r as shov^ni later. This was also noted by Ricker (1975). (ii) A 7^ 1.0 Equation (7) shows that r is a funtion of A, the rate of increase of the population given F-^. Equation (4) accurately gives the true value of r irrespective of A or F2. (iii) k - Incorrectly Chosen Let k ' be the selected age at first full recruitment. If A;' > A; then Equation (4) gives the same rate as Equation (6). As a confirmation of Ricker's (1975) findings it is easy to show that in the extreme case A--1 A^i,2f/i.2 + \ (iV,:,i,2f/,.i,2 - ^^^lJ^.^ exp(-Zi)) 1 = 1 21 ? iV,,2f/, (6) ! = 1 1,2 It can be easily shown that, for C/, j = t/, 2 -^(^ = 1) in a stationary age composition, this reduces to (A - exp(-Zi))/A. (7) In the following tests the results using Equations (6) and (4) will be compared when parameters are changed from year to year or when variability is in- troduced. In all tests F^ = 0.05 and M = 0.05. RESULTS The results of comparing the true recruitment rate from Equation (6) with those obtained from Equa- tion (4) are given below, for the cases when the fishing mortality is different in 2 adjacent years, for A # 1, and for the age at first full recruitment {k) incorrectly chosen when [/,; 2 = ^i, 1 (sections i - iii below). Section iv considers the effect of variability and biases in the age determination of the catch and section v considers stochastic effects, all with [/, 2 = t/; 1. Section vi considers the effects oi Ui^ i= of knife-edge recruitment k ' can he> k and that if k' is < k then no new recruitment is detected. For U, = OAi, i = 1 to 10 and U, = 1.0, otherwise Table 1 shows the reduction in r using Equation (4) as /c' is reduced from its true value at A; = 10. The reduction is substantial and the effect on the esti- mated net recruitment rate, r', is more so. r' is calculated as (Chapman 1983) r' = r - 1 + exp(-M). It can be seen that in this example, which is not un- like many examples in whale assessments, an error of 1 year would reduce r' by 20%. (It is worth noting that this equation for the net recruitment rate is ap- proximate and underestimates the true net rate by about the product of F and MorF and Z depending Table 1. — Reduction in recruitment rates (r) as k' Is incorrectly chosen where Ei^i) = 0 and var(£,) = (^, + 4?) Q-i,i. Consequently, A:-l r2 = [C,,2 + I (^,,1 - 0 exp(-Zi)) C,i + . I ^ £,]/ k E 21 22 [Cx,2 + 2! C,^i Qi + .2! £,]. For any given vector Ni and fixed Zi and Zg the above terms are independent and the approximate variance can be given by var(r2) = var(C.2)(9^/9C.2)^ 21 + Z var(C,i)(ar/aQi)^ 21 + Z var(£j)(9r/a£,;)2 i = x+ 1 = [^,2 (1 - r)^ /t-i + I {^,,i(l - r) - 0exp(-Zi)}2Cu k-l + I (1 - r)2(t,,i + eOC.i 21 + Z r^a,;,! + 2^1,) a, ]/{CT,r~.{S) i = k Comparison of the simulated variances, with the above pattern of recruitment and A; = 10, with those predicted from Equation (8) showed the analytical variance to be a very good approximation, averag- ing about 0.96 times the simulated variances. How- ever, with k = 12 the analytical variance was only 0.70 times the simulated variance. It is clear from Equation (8) that the variance will decrease as the square of the second year catch increases but the first year catches play a more linear role, except through the interactions of ^ and 6 vdth the first year catch. Equation (8) also shows there is a cost involved with increasing k. This is desirable to avoid any bias, but if too few age classes are considered to be fully recruited, then variance increases. In the example considered, raising k from 10 to 12 yr increased variance by 40% and the simulated variances show the increase may even be greater. Additional simulations also revealed that the use of an age-length key might reduce variance by smoothing out real differences in catch at age, but the reduction was nullified by the additional vari- ance due to increasing k. (Vi) f/,,2 ^ U., Equation (6) still allows a true recruitment rate to be calculated in this case and Allen's (1966) derivation allows U, o ¥" U,^. As k should not be 1,2 i,l- underestimated when used in Equation (4), then k can be defined as the larger of the two ages of first full recruitment in years 1 and 2. In this trial an initial stable age distribution was prescribed with A = 1.0, i^j = 0.05 and U^,^ = 0 = (i - 5) X 0.2 = 1.0 i < 5 i = 5 to 10 i> 10. A deterministic catch C^j was obtained given F^ and the population vector A^2 found. The catch and population vector in year 2 was then calculated with Fg = F^ and a changed Ui2, [/,- 2 = 0 i < k2 - 5 ^t,2 = {i + 5 - k2) X 0.2 i = A;2 - 5 to A;2 C/, 2 =1-0 ^ > h- With C/j3 = [/, 1 and F^ = F^ a, third catch was obtained. From this simulation two recruitment values can be obtained, r2 and r^, using Equation (4). The results are given in Table 3 and demonstrate 1) the effect of recruitment occurring earlier in the sec- ond year (r2 with k2 < 10, and r^ with k2 > 10); 2) the effect of recruitment occurring later in the sec- ond year (the converse); and 3) the effect of the age of recruitment fluctuating about an average k to ±\k - k2\. Under these conditions Equation (4) accurately gives the proportion of new recruits in the popula- tion and, as expected, if selection and recruitment 122 HORWOOD: BIAS IN ALLEN'S RECRUITMENT RATE METHOD occur at an earlier age in the second year then a large burst of new recruits will appear. If selection occurs much later, then even the recruits of the previous year will not be seen, giving the negative values. Such a feature was noted by Holt and de la Mare (1983). Horwood et al. (1985) fitted a selec- tion pattern with age that was constant over time for minke whales of the Southern Hemisphere and presented the residual differences. A substantial switching of effort on to different age classes was found over a period of years, and it was shown that this was reflected in the calculated recruitment rates. These residuals and recruitment rates are shown in Table 4 and clearly illustrate the character of the estimate. The problem is then not of calculation but of inter- pretation, in that we do not know selection has changed, and in using this technique it is assumed that the recruitment pattern is constant. A decreas- ing trend in recruitment rate will be interpreted as Table 3. — Recruitment rates calculated from Equation (4) for the model described in section vi. k is the age of first full recruitment used in Equation (4), Icj is the first full recruitment in year 2. r is the average of the two values and nr is the average approximate net recruitment rate. k. k ^2 '■3 r nr 6 10 0.367 -0.301 0.033 -0.015 7 10 0.311 -0.194 0.058 0.009 8 10 0.247 -0.093 0.077 0.028 9 10 0.176 0.004 0.090 0.041 10 10 0.095 0.095 0.095 0.046 11 11 0.004 0.181 0.092 0.043 12 12 -0.099 0.260 0.080 0.031 13 13 -0.213 0.332 0.059 0.010 14 14 -0.340 0.397 0.028 -0.020 a decline in the true rate and not as an increasing age at recruitment and vice versa. As Table 3 shows these rates differ greatly from the 0.095 for con- stant selection, being much higher or lower depend- ing on the trend in recruitment pattern. Consequent- ly a systematic change in recruitment to the fishery will cause substantial problems in interpretation of the recruitment rates. Table 3 also indicates what is likely to occur if the age of recruitment systematically fluctuates about a set pattern. It might be hoped that the r values would average to a useful measure of mean recruit- ment rate. For /cg = 6 the recruitment occurs much earlier in year 2, giving a high r-g, and returns to normal in year 3, giving a low r^. However, the average (0.033) is much smaller than the 0.095 and the approximate net recruitment rate is negative. A similar feature is seen for k2 = 14, but as \k2 - 10| tends to zero the discrepancy is less. If the system fluctuated so that we had a series k{t) = 10, 6, 10, 14, and 10, an approximate average value of r would be the average of the four values of r on Table 3 (0.367, -0.301, -0.340,0.397 = 0.031), and the ap- proximate symmetry gives a similar feature of low average recruitment rates. One way of using the recruitment rates would be to multiply the net recruitment rate by an estimated population size obtained over the same period to give a catch quota which should approximately stabilize the population. From the simulation the average of the recruited population in years 1,3, and 4 has been found. This is very near to the average of the 4 years if the basic recruitment pattern is assumed for the second year. A catch was then found which would make the recruited population in year 5 the same Table 4.— Direction of residuals after fitting a time constant selection at age to minke whale data showing switching of fishing selection across ages with time. Recruitment rate (r) values reflect this switching. (After Horwood et al. 1985.) Age 1974/75 1975/76 1976/77 1977/78 1978/79 1979/80 1980/81 1981/82 1 + + + _ + _ _ _ 2 + + + - + - - + 3 - + + - + + - + 4 - + + - + + - + 5 - - + + - - - + 6 - - + + - - - + 7 - - + + - - - - 8 - - + + - - - - 9 - - - + - + + - 10 - - - + - + + - 11 - - - + - + + - 12 + - - + - + + - 13 - - - + - + + - 14 + - - + - + + - 15 + + - + - + + - 16 - + - + - + + - r values 0-12 0-18 001 0-19 -008 -000 0-27 123 FISHERY BULLETIN: VOL. 85, NO, 1 size as this average. The ratio of this catch to the average population is the net recruitment rate that we would wish to use; these values varied from 0.04 (/cg = 6) to 0.05 {k2 = 14). This confirmed that the distortion of the age structure and population size by the change in selection had very little effect and that a value of r of 0.095, a net rate of 0.046, would be needed to calculate a stabilizing catch. The Table 3 averages are much smaller and we must conclude that if there is no trend in recruitment but a fluc- tuation of more than 1 year then the average esti- mated rates will be largely but undeterminably negatively biased, even if k is not underestimated. A 50-yr simulation confirmed this to be true. As can be gleaned from the above, selection plays an important role in determining r. However, this technique treats overlapping pairs of years as being independent and implies a selection pattern for a pair of years, say 1980 and 1981 and a different one for years 1981 and 1982; these assumptions may be inconsistent. The difference may be small or large but there is no criterion for acceptability. Some cur- rent techniques take arrays of catch-at-age data and obtain best fits to the overall pattern (Beddington and Cooke 1981; Pope and Shepherd 1982), and Pope and Shepherd reduced consideration to two param- eters. What is clear is that selection and recruitment or fishing rates are confounded, and these latter techniques make the assumptions clearly and would be expected to replace analyses of pairs of years. CONCLUSIONS If the recruitment pattern to the exploited popula- tion is constant then the following conclusions may be stated. 1. The "T" of Allen's technique is shown to be necessarily unity and this gives rise to Equa- tion (4) for the estimation of recruitment rate. 2. If the age of first full recruitment is selected correctly then calculated recruitment rates are unbiased for changing fishing efforts or for an increasing or decreasing population. 3. If the age of first full recruitment is overesti- mated then an unbiased recruitment rate is found. If it is underestimated then a negative bias ensues. Inspection of Equation (4) however would caution use of an assumed high value of k, such that a and (i were near unity, and this is reflected in the higher variances given by the approximate variance formula. 4. Aging bias and the use of age-length keys may spread the partially recruited age groups into allocated higher ages. The age of first full re- cruitment should be high enough to encompass this spreading. 5. No bias was detected in recruitment rates from a series of stochastic simulations although Allen (1981) found a small negative bias with low catches. As found by Allen (1981) coefficients of variation of the recruitment rates are high. 6. Equation (8) provides an approximate formula for the variance of the recruitment values given a fixed effort in the pairs of years. To use this the recruitment pattern needs to be estimated from the data as described by Allen (1966). If the recruitment pattern is not constant, serious biases follow: 7. If there is a trend to earlier recruitment over a period of years high recruitment values will be seen and vice versa. These are likely to be interpreted as true increases or decreases. 8. If the recruitment pattern fluctuates about a mean then the net or gross average recruitment rate will be negatively biased, the bias increas- ing with the amplitude of the fluctuations. It is likely that many of the very low rates found by the International Whaling Commission are due to this feature. 9. These last two points indicate that for the tech- nique to be useful it is necessary to establish that the recruitment pattern has been constant. This is likely to prove difficult and consequent- ly much of the value of this simple method is lost. 10. For groups of years of data, alternative tech- niques should be investigated. 11. Finally it appears that the Allen recruitment rate, as calculated in this study or through Allen's original equations with T = 1, should be used with great care. It is subject to large and undeterminable biases and large variances. Where possible other techniques should be used. ACKNOWLEDGMENT Thanks are extended to T. Featherstone, a stu- dent of Brunei University, for assistance in this study, and to D. C. Chapman and J. M. Breiwick for criticisms of the manuscript. LITERATURE CITED Allen, K. R. 1966. Some methods for estimating exploited populations. J. 124 HORWOOD: BIAS IN ALLEN'S RECRUITMENT RATE METHOD Fish. Res. Board Can. 23:1553-1574. 1968. Simplification of a method of computing recruitment rates. J. Fish. Res. Board Can. 25:2701-2702. 1973. Analysis of the stock-recruitment relation in Antarc- tic fin whales (Balaenoptera physalvs). Rapp. P. -v. R6un. Cons. Perm. int. Explor. Mer 164:132-137. 1981. Further notes on the calculation of rj, recruitment rates. Rep. Int. Whaling Comm. 31:597-599. 1982. Minke whales - Antarctic, r,, recruitment rates and mean ages at recruitment. Rep. Int. Whaling Comm. 32: 714. Beddington, J. R., AND J. G. Cooke. 1981. Development of an assessment technique for male sperm whales based on the use of length data from the catches, with special reference to the North-west Pacific stock. Rep. Int. WhaHng Comm. 31:747-760. Chapman, D. G. 1983. Some considerations on the status of stocks of south- ern hemisphere minke whales. Rep. Int. Whaling Comm. 33:311-314. GULLAND, J. A. 1977. Fish population dynamics. John Wiley and Sons, Lond., 372 p. Holt, S. J., and W. K. de la Mare. 1983. An analysis of recent r,] recruitment rates in the North West Pacific stock of sperm whales. Rep. Int. Whaling Comm. 33:279-281. HoRwooD, J. W., J. G. Shepherd, and J. L. Coleman. 1985. Age structure information in minke whales. Rep. Int. Whaling Comm. 35:227-229. Ohsumi, S. 1978. Estimation of natural mortality rate, recruitment rate and age at recruitment of southern hemisphere sei whales. Rep. Int. Whaling Comm. 28:437-448. Pope, J. G., and J. G. Shepherd, 1982. A simple method for consistent interpretation of catch- at-age data. J. Cons. Int. Explor. Mer 40:176-184. Ricker, W. E. 1975. Computation and interpretation of biological statistics offish populations. Fish. Res. Board Can. Bull. 191, 382 p. 125 FECUNDITY AND SPAWNING FREQUENCY OF THE HAWAIIAN ANCHOVY OR NEHU, ENCRASICHOLINA PURPUREA Thomas A. Clarke^ ABSTRACT Female nehu can begin spawning at 35 mm standard length; almost all fish over 40 mm SL from Kaneohe Bay were mature and in spawning condition. Mature females were found in all months of the year. Females from summer (May-October) had higher fecundity and relative cost per batch than fish from winter (November- April). In nehu and most other anchovies, fecundity appears to increase exponentially with weight. Nehu appear to be distinguished from other species by a higher exponent and consequently greater increase in relative fecundity over the reproductive size range. Nehu spawn during a short period 1 or 2 hours after sunset and begin hydrating ova only a few hours before spawning. Data on presence or absence of hydrated ova or postovulatory follicles along with differences in oocyte size in fish collected from throughout the diel cycle indicated that, after spawning, nehu can ripen a new batch of oocytes in 2 days and that most females spawn every other day. The estimated requirements for continued spawn- ing at this rate indicate that individual variation in recent feeding success or stress could be responsible for observed scatter about fecundity-weight relationships and deviation from the normal spawning frequency. The nehu, Encrasicholina 'purpurea, is a small an- chovy endemic to the Hawaiian Islands. It is one of the dominant planktivorous fishes in enclosed, semi- estuarine areas and is the major source of bait for the local skipjack tuna fishery. Nehu are short-lived; growth increments on otoliths indicate a maximum age of about 6 mo (Struhsaker and Uchiyama 1976). Leary et al. (1975) showed that nehu can reach maturity at 35 mm standard length (SL) and pre- sented fecundity data for 41 females. Leary et al. found very few females with hydrated ova and, on that basis, suggested that nehu spawn only once per lifetime. Reexamination of Leary et al.'s (1975) conclusions was prompted both by the great variability in their fecundity vs. weight relationship and by discovery in recent collections that female nehu with hydrated ova are not at all rare, but rather are found only at restricted times of the day. This paper presents results of more detailed investigations of fecundity and spawning frequency in nehu in order to com- pare and contrast aspects of reproductive output of a tropical anchovy with those of better studied tem- perate species. MATERIALS AND METHODS All nehu examined for this study were collected 'Department of Oceanography and Hawaii Institute of Marine Biology, University of Hawaii, Honolulu, HI 96822. from Kaneohe Bay, HI. Day samples were collected by beach seine or dip net in shallow water (1-2 m deep) or were taken from bait recently collected from similar areas by skipjack tuna vessels. Night samples were taken by blind sets with a ca. 67 m long by 13 m deep purse seine over deeper (12-14 m ) areas of the bay. Forty-four night samples and two day samples were taken in 1974-79, while 5 night samples and 18 day samples were taken in 1983-85. Samples with adult nehu were available from all months of the annual cycle and, for most months, from at least two different years. One or more samples with adults were available from all hours of the diel cycle except the period between midnight and dawn, when there were few samples and very few adults collected. In order to follow short-term oocyte development in the same group of fish, on two occasions a school of nehu was surrounded with a 60 m long beach seine in shallow water and sampled initially and twice later in the day. Samples were taken at the hours of 1300, 1500, and 1700 on 13 January 1984 and at 1000, 1300, and 1600 on 27 January 1984. Although the school was obviously disrupted by ini- tial surrounding and subsequent dipnetting of sam- ples, the fish held in the net appeared to resume nor- mal daytime behavior shortly after each disturbance and spent most of the time loosely schooled with other nehu on the outside of the net. The oocyte size- frequency data from these samples did not differ in any obvious manner from data taken from other Manuscript accepted November 1986. FISHERY BULLETIN: VOL. 85, NO. 1, 1987. 127 FISHERY BULLETIN: VOL. 85, NO. 1 samples at approximately the same times on other days; consequently, the data from these "enclose and hold" samples were pooled with the others for all analyses. Time of collection was recorded as the beginning of the set of the net; usually 15-30 min elapsed before the sample was actually preserved. For sam- ples taken from skipjack tuna vessels, the time of collection was often only known within ±15 min and the delay between collection and preservation of the sample was often somewhat longer than 30 min. For analyses of oocyte development rate and spawning frequency, collection time was adjusted to hours since the most recent spawning. Data on appearance of newly spawned eggs in the plankton (Clarke unpubl. data) indicate that spawning begins 1-2 h after sunset and is nearly over in about an hour; the delay after sunset is greatest during the summer. For samples considered here, spawning time was assumed to be 1 h after sunset for dates between mid-October and the end of April and 2 h after sunset for the remainder of the year. Given the frequent uncertainty in actual time of capture, this crude correction for spawning time was satis- factory for the purposes of the present study. All specimens were preserved and held in ca. 4% formaldehyde/seawater solution. The recently col- lected samples were held at least 1 wk before mea- surement and further analyses; by this time most shrinkage in length had occurred. Although the older samples had been in preservative for several years, there was no evidence that long-term storage had affected any parameters considered here, e.g., length-weight relationships were similar for both recent and older samples. For each sample, standard length (SL) of all or a subsample of ca. 100 specimens was measured to the nearest mm. Individuals for further examina- tion were selected from throughout the size range of nehu >35 mm SL in the sample. The selected in- dividuals were measured to the nearest 0.5 mm, opened, and the gonads examined under a dissect- ing microscope. Females were classed as immature —ovaries translucent and maximum oocyte length <0.40 mm; mature— ovaries mostly opaque, oocytes visibly yolked and over 0.40 mm; or hydrated— mature and at least some oocytes with translucent, globular yolk and the peri vitelline space visible. For mature females the length of the apparent largest oocyte was estimated to the nearest 0.1 mm using an ocular micrometer. To determine oocyte size frequency of mature females, a portion of the ovary was teased apart on a glass slide, placed under a compound microscope at 100 X, and the lengths of oocytes over 0.40 mm measured to the nearest 0.01 mm until 20-30 of the largest oocytes were measured. Spawned nehu eggs are ellipsoidal with the length about twice the width (Yamashita 1951). Oocytes >0.3-0.4 mm are also elongate but are more variable in shape. "Length" as used here refers to the maximum dimension. Ex- tremely elongate (length to width ca. 3 or more) and nearly round (length to width less than ca. 1.5) oocytes were noted as was the relative opacity of each oocyte measured. These observations were necessary in many cases to separate nearly round, heavily yolked oocytes that belonged to an advanced mode from very elongate, more nearly translucent oocytes of the same "length" that clearly belonged with a less developed mode. As reported by Leary et al. (1975), mature female nehu may carry 0-2 separate size-frequency modes of oocytes. If a distinct advanced mode of oocytes was evident from the measurements and associated notes, the maximum, minimum, and median lengths of oocytes in this "largest" mode were used for sub- sequent analyses. These parameters will be abbre- viated as LMX, LMN, and LMD, respectively. If all ova in the most advanced mode were hydrated, all lengths were arbitrarily assigned a value of 1 mm. If the largest mode was incompletely separated from smaller oocytes, only LMX and an estimate of LMD were recorded; if there was no separating mode evi- dent, LMX (the largest oocyte in the subsample) was the only datum recorded. If an advanced mode was present and a second or "next" mode was also sep- arated from yet smaller oocytes; the maximum, minimum, and median lengths of oocytes in the next mode will be abbreviated NMX, NMN, and NMD. In most females, however, the next mode was either only partially separated or not evident and, similarly to the case for unseparated advanced modes, only NMX and an estimate of NMD or only NMX, the largest oocyte not in the advanced mode, could be recorded. For 107 specimens for which size-frequency mea- surements were made from a sample of the right ovary, the left ovary was prepared, sectioned, and stained with eosin/hemotoxylin as described by Hunter and Goldberg (1980). The slides, identified by only a code number, were examined for presence of postovulatory follicles (POF). For determination of batch fecundity and dry weight, the fish was first rinsed with distilled water. The ovaries were removed and placed on a clean glass slide. Oocyte size frequency was determined as described above. If a distinct mode of advanced oocytes was present and oocytes in this mode could 128 CLARKE: FECUNDITY AND SPAWNING FREQUENCY OF NEHU be unequivocally discriminated under a dissecting microscope on the basis of size or opacity, the ovaries were teased apart and all ova in the ad- vanced mode counted. This technique eliminated any error in fecundity determination due to subsampling of the ovaries, but meant that very few determina- tions were based on specimens with oocytes smaller than ca. 0.65 mm. In most of the latter cases, even if an advanced mode was clearly evident from the size-frequency determinations, it could not be unequivocally discriminated for total counts under the dissecting scope. Females with hydrated ova free from the follicles and segregated from the smaller oocytes were not used for fecundity deter- minations. After the oocytes in the largest mode were counted, the entire ovaries were rinsed with distilled water from the slide into a preweighed aluminum pan. The stomach contents were removed from the fish and the body cavity was examined for parasites, specifically the presence of ca. 5 mm long nematodes around the liver and pyloric caeca. The fish was placed in a preweighed pan, and any tissue remain- ing on the slide was rinsed with distilled water into the same pan. The fish and gonads were dried at 60 °C for 24 h after which the pans were reweighed to the nearest 0.1 mg, and dry weights of the fish and gonads determined by subtraction. In all cases, fecundity and relative fecundity refer strictly to batch fecundity. Relative fecundity will be given as eggs per gram ovary-free dry weight, and gonad to somatic weight ratio (G/S) will be given as percent of dry weight values. Dry weights were used because of the difficulty in making consistent wet weight determinations on such small fish and even smaller ovaries. Careful wet-dry weight deter- minations on 10 females and gonads indicated that preserved nehu without gonads are about 73% water and that ovaries with yolked, but unhydrated, oocytes are about 60% water. To compare nehu fecundity data with those from other studies which had used wet weights, individual nehu dry weights were divided by 0.27, and relative fecundity and fecundity weight relationships were recalculated. This procedure admittedly ignored any variability in the wet-dry weight relationship. The G/S values given here can be multiphed by 0.675 (0.27/0.40) to make them roughly comparable to values based on wet weight from other studies. Unless otherwise noted, all regressions given below are Model II (or "functional"), GM regressions (Ricker 1973). Results of regressions using natural logarithms are expressed as power curves (antilog form). The 95% confidence limits for slopes of linear regressions and exponents of power curves ( = slopes of In-ln regressions) were calculated from for- mulae in Ricker (1975). For any previous studies which had given results from Model I regressions, original fecundity and weight data were used to calculate functional regressions. RESULTS Maturity and Oocyte Development The smallest mature females were 35 mm SL, the same minimum size reported by Leary et al. (1975), but in many of the samples most of the fish <40 mm SL were immature. Among the fish from the 36 samples from which more than cursory examina- tions were made, 30% of the 134 specimens <40 mm SL were immature. Only 8% of the 227 between 41 and 45 mm SL and <2% of the 284 over 45 mm were immature. Nehu oocytes begin to elongate at about 0.3 mm in length but remain relatively translucent with little visual evidence of vitellogenesis until about 0.4 mm long. Oocytes longer than 0.5 mm were almost always opaque, and those over about 0.6 mm were densely opaque and yellow to yellow-brown in color. The first signs of hydration appeared in oocytes about 0.75 mm long. The yolk became more trans- lucent and globular rather than granular in apparent texture, and the perivitelline space was evident at one or both ends. All ova longer than 0.8 mm were white in appearance and had an evident perivitelline space. At about this size or slightly larger, ova had left the follicles and begun moving to the main ovi- duct. Comparisons of fish from closely spaced purse seine samples taken just before and during spawn- ing indicated that migration of hydrated ova from the follicles to the oviduct occurred in <0.5 h. Only in a few fish with the ova segregated or partially spawned were one or two hydrated ova left in the follicles. Apparently once hydration begins, all ova in a batch are normally ovulated and spawned at one time. Separate batches of maturing oocytes become distinct from the numerous small oocytes between 0.45 and ca. 0.60 mm. In fish with LMX <0.45-0.50 mm there was little or no evidence of a separating size-frequency mode of oocytes. Variably separated modes with LMD at 0.45-0.55 mm were present in fish with the LMX at 0.55-0.65 mm. Usually modes centered at 0.60 mm or larger and with LMX over 0.65 mm were clearly separated from smaller and less opaque oocytes. There was no evidence from size-frequency data that, once oocytes reached ca. 129 FISHERY BULLETIN: VOL. 85, NO, 1 0.65-0.70 mm long, any were "left behind" and'not spawned with the ripening batch. For 248 fish which either had a clearly defined and separated advanced mode of unhydrated oocytes or carried hydrated ova with a clearly defined and separated next mode, the largest (LMX or NMX) and median-sized (LMD or NMD) oocyte in the mode were significantly correlated (r = 0.94, P < 0.01), and the slope of the Model II regression was nearly 1 (1.042). The correlation was essentially unchanged by addition of data from 51 more fish where the median size of an incompletely separated mode was only estimated. These results indicate that, even if a mode is incompletely separated, the estimated LMD is a useful parameter and, furthermore, that for purposes of comparing different fish, LMX is as appropriate an indicator of size of oocytes in a mode as LMD. Consequently, in subsequent analyses of LMD data both unequivocal and estimated values were used, and in other cases LMX was used to analyze change in oocyte size during ripening. Both decisions were made primarily to include data from specimens with small oocytes and without complete- ly or even partially separated modes. Fecundity Fecundity of 222 females 35-58 mm SL ranged from <100 to >1,600, and relative fecundity ranged from 432 to 4,098 eggs/gram. Although low relative fecundities were observed in samples from almost all months, most values over 2,000 were from fish taken in summer and fall (Fig. 1); consequently, the fecundity data from "winter" (November through April) and "summer" (May through October) fish were treated separately for all subsequent analyses. There were no significant differences in size com- position between the summer and winter specimens (Kolmogorov-Smirnov test, P > 0.20). The mean relative fecundity for winter (1,363, n = 93, range = 496-2,763) was significantly different (P < 0.01, i-test) from that for summer (2,097, n = 128, range = 433-4,099). Regressions between fecundity and length or weight (Table 1) also indicated that winter fish were less fecund than summer fish. When relative fecundity data for each season were partitioned according to LMD (<0.65 mm, 0.65-0.75 mm, and >0.75 mm), there were no significant dif- ferences between groups in the summer data (anal- ysis of variance, P > 0.05), but there were signifi- cant differences in the winter data (P < 0.001). Inspection of the data indicated that the latter was due mostly to low values for the fish with LMD <0.65 mm. This could result from incomplete re- cruitment of oocytes to modes barely separated from smaller oocytes. There were, however, only 12 fish in this category, and the small sample size plus the absence of similar evidence in summer fish indicates 7 4 o X U) 3- O) >- I— O 2 Z Z) > < LU OH 0 I • ;«• I i t j'f'm'a'm'j'j'a's'o'n'd' MONTH Figure 1.— Relative fecundity in thousands of eggs/g ovary-free dry weight vs. date of collection for 222 nehu from Kaneohe Bay, HI. Table 1 .—Summary of Model II regression statistics for relationships between length and w^eight and between fecundity and size based on data from 128 "summer" and 94 "winter" nehu plus relationship between gonad weight and bodily weight for 67 summer and 44 winter nehu with hydrated ova. Variables are standard length in mm (SL), ovary- free bodily dry weight (S) and total dry ovary weight (G) in g, and fecundity in numbers of ova in the most advanced mode (F). Results of regressions based on natural logarithms are given as power curves (antilog form). The 95% confidence limits are given for either the slopes of linear equations or the exponents of power curves. X.Y Summer (95% CL) Winter (95% CL) In SL, In S S = 8.868 x 10"' SL 3.25 n F SL, F In SL, S, F In S, In F In S, In G F F F F G = -2352 + 63.1 SL = 6.073 X 10"^ SL^^^ = -351 + 3787 S = 7094 S^^^ = 0.2339 S'^^ (3.11-3.40) 0.94 S = 3.696 x 10"^ SL 3.47 (56.3-70.6) (5.24-6.75) (3,420-4,194) (1.63-2.05) (1.65-2.14) 0.59 0.49 0.66 0.56 0.72 F F F F G -1465 + 38.9 SL 1.226 X 10'^ SL^^"* - 223 + 2444 8 4,538 S' ^° 0.1192 S'^' (3.30-3.65) 0.94 (33.6-45.0) (5.45-7.15) (2,119-2,819) (1.57-2.05) (1.32-1.95) 0.51 0.56 0.53 0.59 0.60 130 CLARKE: FECUNDITY AND SPAWNING FREQUENCY OF NEHU that the significant differences for winter may have resulted from chance alone. The regression statistics (Table 1) for data from each season indicate a great deal of variability about the functional relationships between fecundity and length or weight or between the logarithms of these. The correlation coefficients (r) for all regressions were significantly (P < 0.05) different from zero, but the coefficients of determination (r^) indicated that only about half the variance of fecundity or In fecun- dity was accounted for by the regression. The ex- ponents from the logarithmic regressions of fecun- dity on length are considerably higher than those of the weight-length relationships, and the expo- nents from the logarithmic regressions of fecundity on weight are significantly greater than one. Both indicate that fecundity is not linear vdth weight and that the appropriate expressions for the functional relationship with size are the power curves for fecundity vs. weight. The exponents of the curves for the two seasons were nearly the same, while the summer- winter ratio of the preexponential factors (antilogs of the regression intercepts), 1.56, was almost identical with the ratio of mean relative fecundities, 1.54. A small, but significant part of the variability in fecundity within seasons was related to variation in length- weight relationships of the fish. Using pre- dictions of weight and fecundity from Model I (least squares) logarithmic regressions on standard length, I tested for correlations between relative deviations (observed-predicted/predicted) of fecundity and weight. The relative deviations were positively and significantly correlated for both seasons (summer: r = 0.44, P < 0.01; winter : r = 0.24, P < 0.05). The coefficients of determination, however, indicate that the variation in relative deviation from pre- dicted weight accounted for small percentages of the variation in deviation from predicted fecundity. Maximum relative deviations in weight were ca. + 20% about the predicted value, while deviations in fecundity were much broader: ± 75% in summer and ± 60% in winter. Thus there was a tendency for relatively "fat" individuals to have higher fecundity, but this did not account for much of the scatter in the fecundity data. Nematodes were the only parasites noted fre- quently, and their presence had a minor and insig- nificant effect on fecundity. About half of the sum- mer fish and about a third of the winter fish had nematodes. For both seasons, the exponent from the logarithmic regression of fecundity on weight was higher for fish without nematodes than for those with them, but the 95% confidence limits over- lapped. G/S values ranged from under 2% to about 12% in summer fish and to about 7% in winter fish. For females with maturing oocytes, G/S is a function of both the number and size of oocytes. LeCluse (1979) showed for Sardinops ocellata that ovum dry weight does not increase once hydration begins, and my own preliminary data indicated that this was also true for nehu. Thus effects of variation in oocyte size could be eliminated by considering only fish with LMD >0.75 mm— the size at which hydration begins. The mean G/S for such fish from winter was 4.8% (n = 67: range: 2.4-7.1%) and from summer, 6.3% {n = 44; range: 2.1-12.0%). Among fish with LMD >0.75 mm, the exponents from logarithmic regres- sions of gonad weight on fish weight were signifi- cantly greater than one for both seasonal groups (Table 1). Postovulatory Follicle Deterioration Although the number of specimens examined for postovulatory follicles (POF) was limited (107 from 13 different samples), the results indicated that POF were a reliable indicator of recent spawning up to about 16 h after spawning. Among the 80 specimens from 9 samples taken 1-5 h after estimated spav/n- ing time, follicles were either present and obvious or completely absent. Only seven mature females were available from between midnight and dawn. There were no traces of POF in one specimen; in the others, POF were obvious but showed some signs of degradation similar to that described for northern anchovy, Engraulis mordax, by Hunter and Goldberg (1980). Among the 20 specimens from two samples taken 14-16 h after spawning, POF were further degraded but still distinguishable from other structures in half the fish, while the others showed no traces. Judged from descriptions of POF in E. mordax by Hunter and Goldberg, 14-16 h in nehu appears roughly equivalent to 24 h in E. mor- dax. Although controlled experiments such as those of Hunter and Goldberg were not conducted, it seems likely that, later in the day, POF cannot be distinguished reliably enough to indicate spawning the previous night. Since POF were either present and very obvious or totally absent in fish collected during the night, all traces of previous spawning are apparently gone after about 24 h. Spawning Examination of fish from purse seine samples 131 FISHERY BULLETIN: VOL. 85, NO. 1 taken over deep v^ater after sunset indicated that spawning began and ended during a relatively brief period near the predicted spawning time and that most females in the early night samples were spawners. Ten samples were taken within ± 40 min of the predicted spawning time. In four of these, 96-100% of the mature females in each (a total of 85 examined) carried hydrated ova. Most of those with hydrated ova appeared to have not yet started to spawn, i.e., the hydrated ova were not complete- ly separated from the ovarian tissue and smaller oocytes. In the other six samples, 0-75% of the females (total = 114) carried hydrated ova; most of these appeared to be either partially or nearly com- pletely spent. The largest oocytes in those with no hydrated ova were usually <0.65 mm— about the same size as the largest unhydrated oocytes in those with some hydrated ova present. Only 17 of the specimens without hydrated ova were examined histologically; POF were present in 13. Although it is not possible to separate spawners from non- spawners unequivocally on the basis of oocyte size (see below), the small size of the oocytes and the high fraction with POF among those examined indicate that most of the fish without hydrated ova from these samples had just finished spawning. Later in the night, the frequency of females with hydrated ova decreased, and nonspawning fish ap- peared to occur more frequently. In 16 samples taken between 40 min and 2 h after predicted spawning time the percentage of females that carried hydrated ova ranged from 0 to 80%. Most values were <25%, and only 24% of total of 332 ex- amined carried hydrated ova. Most of those with hydrated ova appeared at least partially spent; many carried only a few at the posterior end of the ovi- ducts. In 14 of these samples, most of the females without hydrated ova were probably recent spawners. The largest oocytes present were <0.65 mm long, and POF were present in 19 of 20 fish ex- amined histologically. In two other samples, how- ever, several of the females carried larger unhy- drated oocytes; POF were present in only 6 of 10 examined from one of these samples. Among the 20 samples taken later in the night (2-4 h after pre- dicted spawning time), only 5 of 254 mature females carried hydrated ova, and oocytes >0.65 mm were present in many of the others. POF were present in only 10 of 20 females examined from two of these samples. Spawning Frequency and Oocyte Development Rate Oocyte size-frequency data for 135 fish taken be- tween 0 and 3.25 h after spawning time indicated that spawners carried smaller oocytes than non- spawners. Few of these fish had clearly defined modes of unhydrated oocytes, so LMX (or NMX if hydrated ova were present) was used as a measure of oocyte development. The 20 specimens without POF carried significantly larger oocytes than those with either POF or some hydrated ova present (Table 2). Although there was some overlap in the size ranges of the two groups (Fig. 2), most of the other fish taken during this period but not examined histologically had relatively small oocytes and were probably recent spawners. Oocyte size-frequency data from fish taken inshore in the morning, 14-16 h after last spawning time. Table 2.— Means and ranges of largest oocyte in the most advanced (first) or next mode of oocytes for nefiu taken over spawning areas at night and in shallow areas during the morning and afternoon. For night fish which had tome hydrated ova present, the datum used was the largest unhydrated oocyte. Collection times were adjusted to hours since estimated time of the most recent spawning. For night and morning, "S" indicates fish that spawned the night of or the night before collection; for afternoon, "S" indicates fish about to spawn the next night. Similarly, "NS" indicates fish that had not spawned the night of or before collec- tion or were not about to spawn the next night. Probability values between three pairs are based on f -tests. Hours since spawning time Group N Mode Largest oocyte (mm) Time Mean (Range) Night 0-3.25 S 115 First or next 0.56 (0.46-0.71) 0-3.25 NS 20 First 0.66 (0.50-0.72) Morning 14-16 S 10 First 0.64 (0.60-0.70) 14-16 NS 10 First 0.72 (0.69-0.75) Afternoon 20-24 S 59 Next 0.52 (0.42-0.63) 20-24 NS 45 First 0.68 (0.60-0.75) P < 0,001 P < 0.001 P< 0.001 132 CLARKE: FECUNDITY AND SPAWNING FREQUENCY OF NEHU 1.0 0.9 MAN 0.8- LU M U^ 0.7 >- o O 0.6 0.5- 0.4 Ao o C300 [82&S-] •o o o o o o . < o • • * . t •w • • CXX3 o° • • ^ 0 b ooo o o • • • • • °°$8 o o o oo o 1 \ 1 1 \ 1 0 4 8 12 16 20 24 HOURS SINCE SPAWNING TIME Figure 2.— Maximum oocyte size vs. estimated hours since the most recent spawning time for 460 nehu collected at different times of the day from Kaneohe Bay, HI. Large open circles represent fish that either were spawning or had spawned at or near 0 hours (i.e., the night of or the night before collection) or were unlikely to spawn at 24 h (i.e., the next night). For such fish taken 0-3 h after estimated spawning time, the data points indicate the largest unhydrated oocyte. Triangles represent fish that had not spawned at 0 h or were apparently going to spawn at 24 h. Small solid circles represent fish which were not examined for postovulatory follicles and could not be assigned to the above groups on the basis of oocyte size alone. Solid and dashed lines indicate the mean and ± 1 stan- dard deviation of maximum oocyte size (horizontal lines) for fish represented by circles and triangles, respectively, for three dif- ferent time intervals (vertical lines) considered in Table 2. indicated that oocyte size of both spawners and non- spawners had increased considerably over nighttime values. The LMX of the 10 specimens without POF averaged significantly higher than that of the 10 with POF (Table 2). The mean LMX for spawners had increased to almost the same value observed for nonspawners in the night samples, while that for nonspawners was almost at the size at which hydra- tion begins. Other fish taken at the same times, but not examined histologically, already carried some hydrated ova (Fig. 2). Data from later in the day indicated that the pre- vious night's nonspawning fish do in fact begin hydrating ova and eventually become the spawners of the next night. Hydrated or hydrating ova oc- curred in about half the fish taken inshore between 16 and 24 h after last spawning, and most of the rest of the fish had considerably smaller oocjftes (Fig. 2). In 67 of 138 specimens taken inshore more than 20 h after the last spawning time (and thus 4 h before the next spawning time), the largest mode was clearly separated from smaller oocytes; LMX was >0.75 mm; and at least some ova were hy- drated. The NMX of 59 fish in this group averaged slightly but significantly (^-test, P > 0.001) smaller than the largest unhydrated oocytes of spawning fish from offshore early night samples (Table 2, line 5 vs. line 1). This indicates that during hydration and spawning of the current batch, which require no further increase in dry weight or energy content, the oocytes in the next batch were already starting to advance toward the next spawning. Although some of the remaining fish in the late afternoon ( -i- 20 h) samples could conceivably have begun hydrating oocytes and spawned by evening, most appeared to be spawning fish from the night before that were to become the nonspawners of the next night. LMX was >0.70 mm in only 9 of the 71 fish in this group, and about 25% of the 45 examined for oocyte size frequency did not have a clearly separated mode. The mean LMX of these 45 fish did not differ significantly {P > 0.05) from that of the nonspawners from the early night samples (Table 2, line 6 vs. line 2). The LMX did, however, aver- age significantly (P < 0.005) higher than that of the previous night's spawners in the morning samples (Table 2, line 6 vs. line 3), thus indicating continued oocyte growth between morning and late afternoon. In summary, the data indicate that most mature female nehu spawn every other day. The largest oocytes present just after spawning increase sub- stantially in size by the next morning and appear to reach the size found in nonspawning night fish by late afternoon. The largest oocytes in nonspawn- ing fish at night are almost at the point where hydra- tion begins by morning, and appear to begin hy- drating then or shortly afterwards such that the previous night's nonspawners are nearly ready to spawn by late afternoon. Alternative cycles are either impossible or diffi- cult to reconcile with the data. If spawning were more frequent, i.e., every day, there would be no nonspawners. This was essentially the case in most night samples taken over the spawning areas, but the day samples averaged about 50% spawners as would result from an every other day cycle. Less 133 FISHERY BULLETIN: VOL. 85, NO. 1 frequent spawning is not consonant with the ap- parent growth of oocytes in the 24 h after spawn- ing and the near absence of fish with LMX <0.60 mm in the late afternoon samples. Some individuals may, however, spawn more or less frequently than every other day. In the early night samples, some spawners carried larger oocytes than most non- spawners and some of the latter carried smaller oocytes than most of the former (Fig. 2). Thus a few spawners appeared to be capable of ripening the next batch in 24 h rather than 48 h, and the largest oocytes of some nonspawners appeared unlikely to be ready for spawning within 24 h. DISCUSSION Results of the present study indicate that the rate of oocyte development in nehu is much faster than in the northern anchovy, Engraulis mordax, or the Peruvian anchovy, E. ringens, the only other species for which comparable data are available. Hunter and Goldberg (1980) showed that oocytes of £". mordax which had spawned within 24 h averaged 0.46 mm long and, during the peak spawning season, grew to the size at which hydration begins in about 7 days. Alheit et al. (1984) indicated that about 6 d are re- quired in E. ringens. In nehu, oocytes in largest mode just after spawning averaged 0.52 mm (mean LMD of 54 spawners taken within 3 h after spawn- ing); these appear to advance to hydration stage in <48 h. Hunter and Goldberg's results also indicated that about 7% of the oocytes in the largest mode are not hydrated and spawned; whereas, in nehu it appears that once a batch of ooc}i:es is separated from smaller oocytes, oocytes in that batch are rare- ly left behind and not spawned with majority of the batch. Hydration, spawning, and degeneration of POF after spawning are also more rapid in nehu than in the Engraulis species. In E. mordax hydration begins in the morning about 12 h before spawning begins (Hunter and Macewicz 1980); Alheit et al.'s (1984) data indicated that E. ringens is similar. Both studies indicated that the Engraulis species spawn over a broad period after sunset with peak spawn- ing just before or near midnight. Nehu ova to be spawned on a given night begin hydrating only a few hours before spawning, and spawning occurs over a rather brief period shortly after sunset. Whereas POF are reliably identifiable up to 24 h after spawn- ing in E. mordax and even longer in E. ringens (Hunter and Goldberg 1980; Alheit et al. 1984), they appear to degenerate to a similar point in about 16 h in nehu. My estimates of spawning timing and duration conflict with those of Yamashita (1951) upon which Tester (1955) apparently based his statements that nehu spawn around midnight. As mentioned earlier, studies in progress on appearance of newly spawned eggs confirm the pattern indicated by presence of females with hydrated ova in purse seine samples after sunset. These studies further indicate that Yamashita was probably not sampling deep enough in the water column to collect newly spawned eggs and that his "freshest" eggs were actually one or more hours old. One of the broader implications of this study is that, when dealing with tropical species, the time scale of sampling must be on the order of hours rather than weeks or days. The latter may be appro- priate for investigation of species from higher lat- itudes, but would miss many events or stages in the reproductive cycle of nehu. Leary et al.'s (1975) con- clusion that nehu spawn only once per lifetime was in part based on the rarity of females with hydrated ova in their samples. This was almost certainly due to their not sampling during the short period be- tween late afternoon and shortly after sunset when hydrated ova are found in the current night's spawners. Leary et al. stated that all females with hydrated ova were captured between 2100 and 2300 h, i.e., well after the peak of spawning even in summer. Both of the above studies of Engraulis species in- dicate some degree of segregation of spawning females at or near spawning time; spawners tended to be overrepresented in such samples. Segregation appears more extreme in nehu; the purse seine samples taken just before and after spawning time were almost all spawners. The greater percentage of nonspawners in some purse seine samples taken later after spawning and the nearly 1:1 ratio of spawners to nonspawners in most day samples in- dicate that spawning fish remix with others later during the night and that segregation of the next night's spawners does not occur until the mixed schools leave shallow day areas at or near sunset. The distribution of nonspawners early in the night is not known. The winter-summer differences in nehu fecundity were evident from both the comparison of relative fecundity and the regressions of fecundity on either length or weight. The G/S data for fish with ova >0.75 mm also showed a higher mean and broader range in summer. Other data (Clarke unpubl. data) indicate that spawned nehu eggs are about 20% heavier in winter, but this difference is insufficient to compensate for higher fecundity in summer fish. 134 CLARKE: FECUNDITY AND SPAWNING FREQUENCY OF NEHU The summer- winter ratio of mean relative fecundity was 1.54; rouglily corrected for the egg weight dif- ference, the ratio of mean effort per spawning would be 1.28 (1.54/1.20), about the same as the ratio of mean G/S, 1.31. There was no evidence that winter fish compensated for lower effort per spawning with higher frequency. The causes and adaptive value of the much greater range and, on the average, higher effort by sum- mer fish are not obvious. Similar differences have been reported between different populations of other species. For example, the northern population of E. mordax appears to be more fecund than the central population (Table 3). This difference is prob- ably genetic and appears to reflect the shorter spawning season (and lower number of batches) in the northern population (Laroche and Richardson 1980). Since nehu live <6 mo (Struhsaker and Uchi- yama 1976), it is difficult to postulate that the dif- ferences between summer and winter fish are genetic. It is, however, possible the winter fish may spawn for longer periods and thus to some degree compensate for lower effort per spawning. The winter-summer differences in nehu reproduc- tive effort per batch may simply be physiological consequences, perhaps with neutral or even nega- tive adaptive value, which result from seasonal differences in the environment. If output in nehu is closely linked to recent feeding success (see below), the output could be lower in winter fish if average daily ration were lower. There is, however, no evidence of major seasonal differences in stand- ing crop of the macrozooplankton upon which adult nehu feed (Hirota and Szyper 1976). Also, nehu feed almost exclusively at night (Clarke unpubl. obs.), and actually have a longer feeding period per diel cycle during the winter. Although the difference between summer maxima and winter minima of temperature in Kaneohe Bay is only about 5°C, it is possible that metabolic processes overall, and consequently both daily ration and reproductive output are slowed enough in winter to account for the observed differ- ence. Regardless of season, the relative fecundity data combined with minimal estimates of spawner abun- dance from purse seine catches predicts planktonic egg densities 2 or 3 orders of magnitude higher than those reported by egg surveys of Tester (1955) or Watson and Leis (1974). Assuming all fish in a ca. 300 m- area were captured, catches of several purse seine sets indicated 0.3-0.5 g dry weight of spawning females/m^ and predicted egg densities of 10^-10^/m". Studies in progress have shown that such egg densities do in fact occur routinely, but that most of the eggs are deeper than 5 m in the water column. Thus the earlier egg surveys, which used surface plankton tows, had missed over 90% of the spawned eggs. Comparable fecundity data are available for only a few other species of anchovies (Table 3), and most Table 3.— Fecundity-weight relationships for winter and summer nehu, Encrasicholina purpurea, and five other species of an- chovies. Means and standard deviations of relative fecundity and power cur\'es for fecundity vs. weight were calculated from available fecundity and weight data. Fish weight were ovary-free wet weights except for nehu, whose wet weights were estimated from dry ovary-free weight data, and Engraulis ringens, for which the data were given as total fish wet weight. Power curves are the antilog forms of equations based on Model II linear regressions of the natural logarithms; 95% confidence limits are for the exponents. Relative fecundities of the smallest and largest female from each group were calculated from the extremes of weight values and the appropriate power curve. Species N Fish weights (g) Relative fecundity (eggs/g) Mean ( + 2 SD), smallest- largest Fecundity vs. weight (95% C.L.) Reference Encrasicholina purpurea Summer 128 0.4-1.8 566 (±436), 284-1,043 F = 647 W'" (1.63-2.05) This study Winter 94 0.4-1.3 368 ( ± 266), 195-542 F = 431 W^8° (1.57-2.05) This study Engraulis mordax Central 67 9.3-31.9 421 (±295), 261-561 F = 65.6 W'^^ (1.36-1.93) Hunter and Macewicz 1980 North 21 14.4-31.3 826 (±449), 650-1.094 F = 108.9 W^^^ (1.19-2.34) Laroche and Richardson 1980 Engraulis ringens 83 11.8-41.5 651 (±404), 493-709 F = 241 W'^^ (1.09-1.53) Minano 1968 Cetengraulis mysticetus 86 24.5-69.5 863 ( ± 529), 613-1,233 F = 71.9 W'^^ (1.45-1 .93) Peterson 1961 Stolephorus heterolobus 9 1.6-6.3 469 (±173), 410-514 F = 379.4 W' '^^ (0.89-1.53) Muller 1976 Anchoa naso 12 0.8-5.6 885 ( ± 672), 1,257-618 F = 1,159 W°^ (0.43-0.94) Joseph 1963 135 FISHERY BULLETIN: VOL. 85, NO. 1 of these species are much larger than nehu. The reproductive size range of nehu overlaps slightly with only Stolephorus heterolobus and Anchoa naso. Unfortunately, previous studies of these two species involved very few specimens, and the summary statistics must be regarded as less reliable than those of the other species in Table 3. Mean relative fecundities for nehu appear to be lower than those of most species; however, the use- fulness of this parameter is questionable because the exponents of the power curves relating fecundity and weight are considerably (and significantly) greater than one in most of the species. Thus mean relative fecundity, a commonly used comparator, would be affected by the size range and size com- position of the sample of females upon which fecun- dity and weight are based. When two groups of similar size composition are compared, as in the case of summer and winter nehu, the difference in mean relative fecundity is similar to that indicated by com- parison of power curves, but otherwise, such as when comparing different-sized species, mean relative fecundities are likely to give erroneous or at best misleading results. Mean relative fecundity also ignores the differences between small and large individuals of the same species or population. The exponents of the power curves for nehu are considerably higher than those of any other species. Although the 95% confidence limits for these values do not exclude those for all the other populations, this indicates that the rate of increase in relative reproductive output with increasing size is greatest in nehu. The consequences are illustrated by the relative fecundities calculated for the smallest and largest fish of each population using the power curve for that species (Table 3). Relative fecundities of the largest females are 1.2-2.2 times those of the small- est in the other species but 2.8 and 3.7 times greater in winter and summer nehu, respectively. Both the smallest and largest winter nehu appear to be less fecund per unit weight than the smallest and largest females of all or most of the other species. Small summer nehu also have considerably lower relative fecundity than most of the others, but the value for large summer nehu is among the highest. Ignoring the rather questionable results for Anchoa naso (only 12 individuals), the value for the largest Cetengraulis mysticetus is the only one substantially greater than that of the largest summer nehu. Although these comparisons must be regarded as tentative because many between-species differences in power curve exponents are not significant, nehu seem to be distinguished from other anchovies not by differences in relative fecundity but rather by dif- ferences in the relation between relative fecundity and size. Speculation about the possible relation of this to differences in environment and other life history parameters, such as nehu's short life span and maturity soon after metamorphosis, is un- warranted without evidence that similar differences exist between large and small species in other taxa. Nevertheless, it seems possible that the pattern of allocation of resources between growth and repro- duction over the reproductive life span is yet another life history parameter which could be selected for by prevailing adult mortality rates, predictability of larval survival, etc. Comparison of fecundities alone does not ade- quately reflect differences in reproductive effort if there are differences in egg size. For example, nehu eggs average about two-thirds the egg weights calculated for E. mordax by Hunter and Leong (1981). Effort per batch would be best measured by relative cost in terms of dry weight, calories, etc., rather than numbers of eggs. Available data per- mit only crude comparisons of the two species. The intercept of the regression equation for G/S vs. fecundity of nehu with ova >0.75 mm is about 2.5% for fish from both seasons and nearly the same as the mean G/S (2.4%) of 21 other fish whose largest oocytes were 0.48-0.65 mm and had presum- ably just spawned. (G/S data were not available for fish used for POF analyses.) Using 2.5% as the mean G/S 2 days before spawning and subtracting this from mean G/S of nehu with ova >0.75 mm, i.e., those about to spawn, gives mean relative weights per batch of 3.8% of bodily dry weight in summer and 2.3% in winter. These estimated relative costs per batch are minimal since they do not include in- vestment in bringing oocytes to the size at 2 days before spawning. Hunter and Leong (1981) did not give relative cost per spawning of E. mordax in terms of dry weight, but data in their table 4 plus an assumption of dry bodily weight equal to 25% of wet weight yield an estimate of about 4.4% of bodily weight per spawn- ing for an average female. Hunter and Leong' s data in table 1 indicated that dry weight in E. mordax declined about 30% during the main spawning season due to loss of fat; this loss is shown to be equal to the calories required for about 13 spawn- ings. If this is also true for dry weight then the loss per batch would be about 2.3% of dry bodily weight. The above estimates of cost per batch in terms of dry weight are very crude and only indicate that nehu, particularly summer nehu, are probably similar to E. mordax. Additionally it is clear that nehu, like E. mordax, lose half or more of their ovary 136 CLARKE: FECUNDITY AND SPAWNING FREQUENCY OF NEHU weight with each spawning and must depend on bodily reserves and assimilation of food, rather than ovarian reserves, to continue spawning. As men- tioned above, Hunter and Leong (1981) showed that about 65% of the caloric cost of spawning is supplied by fat reserves. Even if the same were true for nehu, the additional requirements for continued spawn- ing would have to come from food assimilated and available for reproductive processes over a period of only 2 d rather than 7 d in £". mordax. Assuming cost per batch is 4% of dry bodily weight and that 65% of this comes from bodily reserves in both E. mordax and summer nehu, the average additional requirements per day would be 0.2% and 0.7%, respectively. The above suggests that all aspects of reproduc- tive output in nehu— batch fecundity, spawning fre- quency, and duration of spawning— would be very sensitive to any factors affecting availability of resources for reproduction. Parasite load, which has been shown to affect batch fecundity in cod (Hislop and Shanks 1981), apparently has only an insignifi- cant effect on nehu, but since a batch is formed only 2 or 3 days before spawning and the ova to be spawned on a given evening do not attain maximum size until just a few hours before spawning, even re- cent events could affect the number or the growth rate of oocytes in a batch. Some of the great varia- tion in fecundity and the indications that some fish spawn more or less often than normal could result from individual differences in recent feeding suc- cess, injury or stress from predators or the fishery, or perhaps the extent of inshore-offshore move- ments over the diel cycle. Unfortunately, none of these putative factors (except for serious injury) would leave any detectable trace on individual fish that might explain why fecundity or spawning fre- quency was higher or lower than average. ACKNOWLEDGMENTS I thank L. R. Johnson and K. C. Landgraf for assistance in different phases of this study. The re- search was supported by the University of Hawaii Sea Grant Program, PM/M-20 and PM/M-lBBl; by the University of Hawaii Research Council; and by the Hawaii Institute of Marine Biology. LITERATURE CITED Alheit, J., V. H. Alarcon, and B. J. Macewicz. 1984. Spawning frequency and sex ratio in the Peruvian an- chovy, Engraulis ringens. Calif. Coop. Oceanic Fish. In- vest. Rep. 25:43-52. HiROTA, J., AND J. p. SZYPER. 1976. Standing stocks of zooplankton size-classes and trophic levels in Kaneohe Bay, Oahu, Hawaiian Islands. Pac. Sci. 30:341-361. Hislop, J. R. G., and A. M. Shanks. 1981. Recent investigations on the reproductive biology of the haddock, Melanogrammus aeglefinis, of the northern North Sea and the effect on fecundity of infection with the copepod parasite Lemaeocera branchialis. J. Cons. Int. Ex- plor. Mer 39:244-251. Hunter, J. R., and S. R. Goldberg. 1980. Spawning incidence and batch fecundity in northern anchovy, Engraulis mordax. Fish. Bull.. U.S. 77:641- 652. Hunter, J. R., and R. Leong. 1981. The spawning energetics of female northern anchovy, Engraulis mordax. Fish. Bull., U.S. 79:215-230. Hunter, J. R., and B. J. Macewicz. 1980. Sexual maturity, batch fecundity, spawning frequen- cy, and temporal pattern of spawning for the northern an- chovy, Engraulis mordax, during the 1979 spawning season. Calif. Coop. Oceanic Fish. Invest. Rep. 21:139-149. Joseph, J. 1963. Contribution to the biology of the engraulid Anchoa naso (Gilbert and Pierson 1898) from Ecuadorian waters. Inter- Am. Trop. Tuna Comm., Bull. 8:1-30. Laroche, J. L., and S. L. Richardson. 1980. Reproduction of northern anchovy, Engraulis mordax, off Oregon and Washington. Fish. Bull., U.S. 78:603- 618. Leary, D. F., G. I. Murphy, and M. Miller. 1975. Fecundity and length at first spawning of the Hawaiian anchovy, or nehu {Stolephorus purpureus Fowler) in Kane- ohe Bay, Oahu. Pac. Sci. 29:171-180. Lecluse, F. 1979. Dry mass of yolked oocytes of the southwest African pilchard, Sardinops ocellata, in relation to maturity stages and spawning cycles, 1972-1974. Invest. Rep., Sea Fish. Branch, S. Afr. 119:1-29. MiNano, J. B. 1968. Estudio de la fecundidad y ciclo sexual de la anchoveta (Engraulis ringens J) en la zona de Chimbote. Bol. Inst. Mar, Peru 1:505-533. Muller, R. G. 1976. Population biology oi Stolephorus heterolohus (Pisces: Engraulidae) in Palau, Western Caroline Islands. Ph.D. Thesis, Univ. Hawaii, Honolulu, 174 p. Peterson, C. L. 1961. Fecundity of the anchoveta (Cetengraulis mysticetus) in the Gulf of Panama. Inter- Am. Trop. Tuna Comm., Bull. 6:55-68. RiCKER, W. E. 1973. Linear regressions in fishery research. J. Fish. Res. Board Can. 30:409-434. 1975. A note concerning Professor Jolicoeur's comments. J. Fish. Res. Board Can. 32:1494-1498. Struhsaker, p., and J. H. Uchiyama. 1976. Age and growth of the nehu, Stolephorus purpureus (Pisces: Engraulidae), from the Hawaiian Islands as indicated by daily growth increments of sagittae. Fish. Bull, U.S. 74:9-17. Tester, A. L. 1955. Variation in egg and larva production of the anchovy, Stolephorus purpureus Fowler, in Kaneohe Bay, Oahu, dur- ing 1950-1952. Pac. Sci. 9:31-41. 137 FISHERY BULLETIN: VOL. 85, NO. 1 Watson, W., and J. M. Leis. Yamashita, D. L. 1974. The ichthyoplankton of Kaneohe Bay, Hawaii. Univ. 1951. The embryological and larval development of the nehu, Hawaii Sea Grant Tech. Rep. UNIHI-SE AGRANT-TR-75- an engraulid baitfish of the Hawaiian Islands. M.S. Thesis, 01, Honolulu, 178 p. Univ. Hawaii, Honolulu, 65 p. 138 NOTES ON THE COMPATIBILITY OF A NEW EXPRESSION FOR GROSS CONVERSION EFFICIENCY WITH THE VON BERTALANFFY GROWTH EQUATION' Gross food conversion efficiency {K{) is defined by K^ = growth increment/food ingested (1) dW form dt II where / is the ingestion rate (Ivlev 1939; Ricker 1966); data from feeding experiments are usually fit to an allometric model of the form K^ = cW (2) where W is the body weight, and c and a are em- pirical constants which, however, have the disadvan- tage of always predicting values oi K^ > 0, al- though the fish and other aquatic animals to which the model is meant to apply usually experience size constraints and hence must reach a value of W where /fj = 0. It is therefore preferable to choose a functional form for K^ which falls to zero as W approaches W^. Furthermore, recent analysis of feeding studies of a number of fish species indicates that Ki can approach arbitrarily close to unity for the smallest fishes, which suggests the alternate equation K, = i- {w/wy (3) where W^ is the weight at which K-^ = 0, and ft is an empirical constant estimated from the slope of log (1 - K,) = (ilogW - p log W^ (4) (Pauly 1986). In this note we show that Equation (3) is com- patible with the von Bertalanffy growth function (VBGF), both in its standard (von Bertalanffy 1938) and generalized forms (Richards 1959; Pauly 1981), which is not true of Equation (2). We assume that the ingestion rate (/) can be ex- pressed as an allometric expression of weight of the iICLARM Contribution No. 316. FISHERY BULLETIN: VOL. 85, NO. 1, 1987. / = HW^, (5) where H and d are empirical constants. From Equa- tion (1) we then obtain for the growth rate dW/dt = K^ HW^ (6) which combined with Equation (3) gives dWIdt = (1 - (WIWJP) HW (7) and hence dW/dt = HW^ - kW"^ (8) where m = d + fi and k = HIW^. Equation (8) is the differential form of the VBGF, and can be in- tegrated for various values of the constants m and d. Setting d = 2/3 and m = 1 (i.e., ft = 1/3) yields the "normal" VBGF for weight, Wt = W^il - e-^(«-«o))3 (9) where K = kIS, while if m = 1 and 0 < d < 1 we get the generalized VBGF sensu Pauly (1981), Wt = W^{1 - e-^('-«o)) ZID (10) where Z) = 3(1 - d). This second form is probably more useful as it allows for the exponent of the allometric relationship linking ingestion and weight (Equation (5)) to take wider range of values, as needed to fit various data sets and/or to mimic various models in the literature (see, e.g., Paloheimo and Dickie 1966 or Ursin et al. 1985). The compatibility shown here between the recent- ly proposed Equation (3) expressing K^ as a func- tion of fish weight and the VBGF is encouraging, as it supports the method suggested by Pauly (1986) for combining these two equations when estimating the food consumption of fish populations and leads to a mathematically consistent approach for the analysis of feeding and growth data. LITERATURE CITED Bertalanffy, L. von. 1938. A quantitative theory of organic growth (Inquiries on 139 growth laws II). Hum. Biol. 10:181-213. IVLEV, V. S. 1939. Balance of energy in carps. [In Russ.] Zool. Zh. 18: 303-318. 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. 23:1209-1248. Pauly, D. 1981. The relationship between gill surface area and growth performance in fish: a generalization of von Bertalanffy's theory of growth. Meeresforschung 28:251-282. 1986. A simple method for estimating the food consumption of fish populations from growth data and food conversion experiments. Fish. Bull., U.S. 84:827-840. RiCKER, W. E. 1966. "The biological productivity of waters" by V. S. Ivlev. J. Fish. Res. Board Can. 23:1717-1759. [TransL] Richards, F. J. 1959. A flexible growth function for empirical use. J. Exp. Bot. 10:290-300. Ursin, E., M. Pennington, E. B. Cohen, and M. D. Grosslein. 1985. Stomach evacuation rates of Atlantic cod (Gadics morhua) estimated from stomach contents and growth rates. Dana 5:65-80. W. Silvert Department of Fisheries and Oceans Marine Ecology Laboratory Bedford Institute of Oceanography Dartmouth, Nova Scotia, Canada B2Y kA2 International Center for Living Aquatic Resources Management MC P.O. Box 1501 Manila, Philippines D. Pauly EFFECT OF A RIVER-DOMINATED ESTUARY ON THE PREVALENCE OF CARCINONEMERTES ERRANS, AN EGG PREDATOR OF THE DUNGENESS CRAB, CANCER MAGISTER Carcinonemertes errans is a host-specific nemertean that can destroy large numbers of Dungeness crab, Cancer magister, eggs (Wickham 1979, 1980). Al- though the ectosymbiotic nemertean is present on adult and juvenile crabs of both sexes, its only known detrimental effect is to the egg stage. Wick- ham (1979) estimated that the direct mortality to eggs of Dungeness crabs off central California was 55%. High egg mortalities in the San Francisco, CA, area were suggested as a possible cause of the drastic decline in Dungeness crab populations in that area (Fisher and Wickham 1976; Wickham 1979). From November 1983 through October 1985, the National Marine Fisheries Service (NMFS) con- ducted a comprehensive study of the distribution, abundance, and size-class structure of Dungeness crabs in the Columbia River estuary, a river-domi- nated estuary. Limited sampling was also done in adjacent coastal areas. As an incidental part of the study, we examined crabs for C. errans, and ob- served an effect of the river-dominated estuarine environment on the prevalence of C. errans on Dungeness crabs. Methods The study was done in the lower Columbia River estuary and adjacent coastal areas (Fig. 1). The estuary is a drowned river mouth that is dominated by river flows. Highest flows typically occur during the spring and lowest flows during late summer and fall. Estimated river flows (monthly averages) dur- ing the study period ranged from 3,121 m^/s (August 1985) to 14,091 m^/s (May 1985) (U.S. Geo- logical Survey, Portland, OR). Salinities fluctuate widely in the estuary depending on river flow, tidal stage, and distance from the river mouth (Neal 1972). Inversely related to river flows, the salinity intrusion is typically least during spring and great- est during late summer and fall. Sampling was done monthly at a maximum of 28 estuarine and ocean sites (Fig. 1). At 26 of the sites, an 8 m semiballoon shrimp trawl with stretched mesh size of 38.1 mm was used to collect samples; a 9.5 mm Hner was inserted in the cod end of the net to prevent escape of small Dungeness crabs. Sampling in the estuary was normally done during times of higher salinity (early flood to early ebb tide). Generally a subsample of at least 100 Dungeness crabs (^20 mm) from each trawl effort was mea- sured to the nearest mm (carapace width, anterior to the 10th anterolateral spines), weighed, sexed, and checked for eggs and C. errans. Specific body areas— the undersurface of the abdomen, the thoracic area covered by the abdomen, and the pleo- pods— were examined for C. errans. Dungeness crab catches at individual stations varied considerably, ranging from 0 to >100 crabs per trawl effort. Crabs <20 mm were measured and weighed, but were not routinely sexed or checked for C. errans. Dungeness crabs were separated into four size classes: I (<50 mm), II (50-99 mm). III (100-129 mm), and IV (>129 mm). We used the chi-square test to compare the prevalences of C. errans on crabs in the ocean and the estuary and to compare the level of infestation between males and females within the two areas. 140 FISHERY BULLETIN: VOL. 85, NO. 1, 1987, WASHINGTON "T Bottom trawl station * Intertidal station PACIFIC OCEAN 0 1 ? 3 Miles I 1 I I I — ! — 1 — I 1 1 0 12 3 4 5 Kilometers Figure 1.— Map of the Columbia River estuary and adjacent coastal areas, showing sampling sites for the 2-yr Dungeness crab study. Results and Discussion The prevalence of C. errans on Dungeness crabs collected in the estuary was significantly lower than the prevalence on crabs in the ocean (x^, df = 1, P < 0.001); average prevalences in the estuary and ocean were 6 and 79%, respectively (Tables 1, 2). Within the estuary, mean prevalence was highest at the mouth (stations 1,2, 23-26) where it averaged 25%. In the estuary, significantly more females (8%) were infested than were males (5%) (x", df = 1, P < 0.001), but in the ocean there was no significant difference (P > 0.05) in prevalence on males (80%) and females (76%). Only three egg-bearing females were collected during the study; they were collected December 1984 at the mouth of the estuary and in the ocean. One egg-bearing female had an obvious C. errans infestation. In both the estuary and ocean, size class I Dunge- ness crabs were least frequently infested. No chi- square comparison was done for this size class because of the small numbers of infested crabs. In addition, the total sample size of size class I crabs in the ocean was small (46 crabs). For the individual size classes II-IV, the prevalences of C. errans on crabs were significantly lower in the estuary than in the ocean (x^, df = 1, P < 0.001). In the estuary, the infestation by C. errans was highest in size class IV crabs (29%). The prevalence of C. errans found on Dungeness crabs in the ocean and the Columbia River estuary was lower than the prevalence reported by Wickham (1980) in the Bodega Bay, CA, area; he reported that all nonegg-bearing crabs >20 mm carapace width were infested with C. errans. In our study, some light infestations may have been missed by not ex- amining the entire exoskeletons of the Dungeness crabs. The major result of our examinations for C. errans was discovering the large difference in infestation levels between the ocean (79%) and the estuary (6%). Low salinities in the estuary, particularly upstream from the mouth, probably were the major cause of the lower infestation. During low river flows (about 4,400 m^/s), when salinity intrusion is greatest, minimum bottom salinities in most of the lower 22 km of the estuary generally range from 0.5 to 15 ppt, although maximum salinities are >S0 ppt. Dur- ing high river flows (about 8,800 m^/s), minimum bottom salinities in much of the lower 22 km of the estuary may be zero (Jay 1984). Wickham^ noted that "Pure fresh water will kill worms in 1-2 minutes depending on the worms' size." The lower prevalence in the estuary may have have little ef- fect on the overall prevalence in the ocean. Non- infested Dungeness crabs migrating from the estuary could be infested in the ocean by larval worms (Wickham 1980), or through copulation (Wickham et al. 1984). 'D. E. Wickham, Bodega Marine Laboratory, P.O. Box 247, Bodega Bay, CA 94923, pers. commun. November 1985. 141 Table 1 . — Prevalence of the egg predator Carcinonemertes errans on Dungeness crabs collected in the Columbia River estuary from November 1983 through October 1985. Number Number Percent examined infested infested Prevalence by month Nov. 1983 362 25 7 Dec. 1983 345 8 2 Jan. 1984 273 3 1 Feb. 1984 160 1 1 Mar. 1984 130 1 1 Apr. 1984 105 9 9 May 1984 84 8 10 June 1984 141 15 11 July 1984 146 6 4 Aug. 1984 248 18 7 Sept. 1984 306 9 3 Oct. 1984 169 11 7 Nov. 1984 218 5 2 Dec. 1984 158 4 3 Jan. 1985 264 1 0 Feb. 1985 59 0 0 Mar. 1985 311 3 1 Apr. 1985 135 6 4 May 1985 238 6 3 June 1985 287 5 2 July 1985 301 8 3 Aug. 1985 262 12 5 Sept. 1985 328 36 11 Oct. 1985 424 122 29 Total 5,454 322 mean, 6 Prevalence by size class Size class 1 1,273 4 0 Size class II 2,561 102 4 Size class III 1,225 101 8 Size class IV 395 115 29 Prevalence by sex Male 3,269 155 5 Female 2,185 167 8 Table 2.— Prevalence of the egg predator Carcinonemertes errans on Dungeness crabs collected in nearshore areas of the Pacific Ocean from December 1983 through Septem- ber 1985. Our data indicate that C. errans is a marine species that apparently cannot tolerate the lower salinities in the Columbia River estuary. It would be informative to examine Dungeness crabs from other Oregon and Washington estuaries with typi- cally higher salinities to determine if infestation levels are comparable to those in the Columbia River estuary. Acknowledgments We thank personnel at the Hammond, OR, Field Station of the Northwest and Alaska Fisheries Center (NMFS) for their assistance in field sam- pling. The U.S. Army Corps of Engineers (Portland District) provided partial financial support for this study. Literature Cited Fisher, W. S., and D. E. Wickham. 1976. Mortalities and epibiotic fouling of eggs from wild Number Number Percent examined infested infested Prevalence by month Dec. 1983 15 14 93 Jan. 1984 139 113 81 Mar. 1984 13 11 85 Apr. 1984 6 5 83 May 1984 4 3 75 June 1984 7 6 86 July 1984 10 6 60 Aug. 1984 20 17 85 Dec. 1984 5 2 40 Feb. 1985 3 0 0 Apr. 1985 16 . 15 94 May 1985 3 2 67 July 1985 37 29 78 Aug. 1985 11 7 64 Sept 1985 130 99 76 Total 419 329 mean, 79 Prevalence by size class Size class 1 46 1 2 Size class II 115 107 93 Size class III 151 124 82 Size class IV 107 97 91 Prevalence by sex Male 276 220 80 Female 143 109 76 populations of the Dungeness crab. Cancer magister. Fish. Bull., U.S. 74:201-207. Jay, D. 1984. Circulatory processes in the Columbia River estuary. Final Report on the Circulation Work Unit of the Columbia River Estuary Data Development Program. CREST, Astoria, OR, 233 p. Neal, V. T. 1972. Physical aspects of the Columbia River and its estuary. hi A. T. Pruter and D. L. Alverson (editors). The Columbia River estuary and adjacent ocean waters, bioenvironmen- tal studies, p. 19-40. Univ. Washington Press, Seattle. Wickham, D. E. 1979. Predation by the nemertean Carcinonemertes errans on eggs of the Dungeness crab Cancer magister. Mar. Biol. (Berl.) 55:45-53. 1980. Aspects of the life history of Carcinonemertes errans (Nemertea: Carcinonemertidae), an egg predator of the crab Cancer magister. Biol. Bull. (Woods Hole) 159:247-257. Wickham, D. E., P. Roe, and A. M. Kuris. 1984. Transfer of nemertean egg predators during host molt- ing and copulation. Biol. Bull. (Woods Hole) 167:331- 338. George T. McCabe, Jr. Robert L. Emmett Travis C. Coley Robert J. McConnell Northwest and Alaska Fisheries Center National Marine Fisheries Service, NOAA 2725 Montlake Boulevard East Seattle, WA 98112 142 SPRING AND SUMMER MOVEMENTS OF SUBADULT STRIPED BASS, MORONE SAXATILIS, IN THE CONNECTICUT RIVERA The Connecticut River has no known spawning population of striped bass, Morone saxatilis, but there is an annual run of subadults in the late spring and summer from Long Island Sound to Holyoke Dam, 140 km upstream (Moffitt et al. 1982). In 1980-82, 80-90% were age II (the remainder were age III); about 60% were males (Warner 1983). The biological reason for such a run is unknown, but feeding may be an important attractant to the river. The major foods of striped bass collected at Holyoke Dam are spottail shiners, Notropis hudsonius, and the scales and body parts of adult American shad, Alosa sapidissima, and blueback herring, A. aesti- valis, that result from injury or death at the hydro- power dam and fish lifts or from angling (Warner and Kynard 1986). Factors other than food are un- iContribution No. 94 of the Massachusetts Cooperative Fishery Research Unit, which is supported by the U.S. Fish and Wildlife Service, Massachusetts Division of Fisheries and Wildlife, Massa- chusetts Division of Marine Fisheries and the University of Massachusetts. doubtedly important influences on the riverine migration. The migration of subadult striped bass into natal or nonnatal rivers was documented by Raney et al. (1954) and Nicholas and Miller (1967), but the reasons for the movement are not clear. We hy- pothesized that detailed studies of subadult move- ments in the Connecticut River could help reveal some of the environmental factors that effect the movements. We used radio telemetry of subadults captured at Holyoke Dam to observe the use of river habitats, diel activity, and the rates of upstream and downward movements. We also investigated the passage of striped bass at the Holyoke fish lifts in relation to river temperature during 1979-86. Study Area Radio-tagged striped bass were observed after they were transferred above Holyoke Dam into the 53 km of the Connecticut River, between the Hol- yoke Dam and the Cabot Station hydroelectric facility which is below Turners Falls Dam (Fig. 1). The upstream 23 km reach is relatively straight, with few areas deeper than 4 m; the lower 30 km reach meanders, creating a deep channel and shoals (Fig. 1). Bottom type is rubble and gravel in the Cabot Stotion {8).,^ Monlogue Bridge (2) Eost Deerfield (3) Second Island (2)---..._ Sunderland Bridge (2)-.. S-Curve (4)-,,^ N. Hadley Stait ( l).'^---._ Hadley Turn (I), Northampton Turn (I) "j^ Shepherd Island (I ) Northampton Oxbow Outlet (5 Rt. 202 Bridge (I), Turners Falls Dam Deerfield River 1981 Release Site 1982 Release Site lOkm •Coolidge Bridge Holyoke Holyoke Tailrace "-^--Dam Figure 1.— The 53 km of the Connecticut River above the Holyoke Dam where the movements of radio-tagged striped bass were observed in 1981-82. The 13 holding areas where striped bass stopped are on the left side of the river (number of stops in parenthesis). FISHERY BULLETIN: VOL. 85. NO. 1. 1987. 143 upper stretch; and sand and areas of exposed rock ledge are in the lower section (Armour 1966). Methods The number of striped bass passed daily by the Holyoke fish lifts from 1979 to 1986 was counted by personnel of the Massachusetts Cooperative Fish- ery Research Unit (MCFRU). Maximum daily river temperature recorded at the dam was used to char- acterize the temperature regime for the striped bass lifted each day. The daily records of each year's run were used to make frequency distributions of the number of fish lifted and the daily maximum tem- perature. We used the statistics of mean, median, standard deviation, and range of temperatures to visually compare the temperatures when striped bass entered the lifts. All striped bass used for telemetry were captured during 1981 and 1982 in the fish trap at the lifts. To help reduce mortality caused by handling, we marked only the largest fish captured (280-365 mm fork length). Fish were held at the dam for a max- imum of 5 d in a 1,325 L circular tank supplied with river water. At the release sites (Fig. 1), we inserted into the fish a transmitter which went directly through the mouth and into the stomach, a proce- dure that did not interfere with subsequent feeding (Warner 1983). Radio transmitters were constructed using the design of Knight (1975) or with the modifications of Buckley (MCFRU). The transmitters measured 12 mm in diameter and 45 mm long, weighed 3.5-5.5 g in air, and transmitted for 7-21 d. Weight of the transmitters never exceeded 3.4% of the body weight of the fish. Individual fish were identified by 12 frequencies (30.05-30.25 MHz) and by variations in the pulse rate of each frequency. We tracked striped bass from a boat using an omnidirectional antenna (1/8-wave, base loaded) to locate fish to within about 100 m and a directional, tuned-loop antenna to locate fish to within about 10 m. Locations of fish were noted on contour maps of the river. Initially, we tracked striped bass from 4 to 30 h, but tracking each fish was not continuous and depended on the speed of dispersal. Later, we surveyed the study area daily. Some striped bass moved actively and others were sedentary; there- fore, we tracked the active fish continually for as long as 6 h, but only periodically noting the locations of others. In addition to the daily surveys, we ob- served some fish continually for 24 h to determine the diel movement; we conducted three diel surveys in 1981 and nine in 1982. The upstream and downstream rates of movement (ground speed) were determined by using the con- tinuous observations of striped bass that had been free longer than 1 h. Locations where striped bass remained longer than 90 min were designated as "holding areas". The physical characteristics of these areas were determined from visual observa- tions and contour maps. Results and Discussion Passage in the Lifts and Temperature Activity at the fish lifts appeared to be related to temperature (Fig. 2). Striped bass first entered the lifts when river temperatures were 17°-19°C (late May or early June), and in some years a few were still entering the lifts at 25°-28°C when lift opera- tion ceased. The mean temperature of activity when striped bass entered the lifts ranged from a low in 1980 of 10.0°C to a high in 1983 of 23.4°C (Fig. 2). For the 7-yr period, the mean temperature of peak movement was 21.3°C (SD, 1.7°C) with 72% of the fish passage from 20°C to 24°C (85% of passage between 19° and 24 °C). 15 o o -' 10 lU =) < tr UJ CL LlI I- 21 2-1 27 231 n--86 MO standard J -I- mean deviolion S V mcdion , range 493 346 355 110 \ ■o- 187 V - J_ _L J_ _1_ 1979 1980 1981 1982 I9B3 1984 1985 1985 YEAR OF RUN Figure 2.— Mean, median, standard deviation, and range of temperatures when subadult striped bass were passed in the Holyolie Dam fish lifts, 1979-86. A recent hypothesis proposed that, as striped bass advance in age, they prefer cooler water (Coutant 1985). Further, the thermal niche of subadults (43-68 cm total length) in Tennessee reservoirs was 20°- 24° C, when these temperatures were available (Cou- tant and Carroll 1980). This is essentially the same range as most upstream movement into lifts in this study. Although the movement of striped bass into 144 the lifts is more an indication of general activity than choice of preferred temperatures; nevertheless, the similarity in range of temperatures found in the two studies is remarkable. Striped bass began entering the fish lifts in late May when river temperatures were about 17°C. During mid- to late May 1979-83, the daily maximum temperatures in Long Island Sound near the mouth of the river were much cooler (12°-13°C, Millstone Laboratory, Northeast Utilities Service Co., Hartford, CT). While we do not know whether the striped bass overwinter in the lower river or enter fresh from Long Island Sound each spring, the movement of subadults from the cooler waters of the Sound into the warmer river is con- sistent with the thermal niche hypothesis of Cou- tant (1985). The only data from the Connecticut River that appears inconsistent with the hypothe- sis of Coutant (1985) is the capture of nine subadults in the lifts at Holyoke Dam in the fall of 1979 when river temperatures were 7°-10°C. Although prey abundance is high each fall at Holyoke Dam because of the outmigration and death of many juvenile American shad and blueback herring passed through the turbines (Taylor and Kynard 1985), the temper- atures when the striped bass entered the lifts were much colder than preferred. Did the food abundance cause some striped bass to remain in water tem- perature that would otherwise be avoided? Because no striped bass have been lifted in the fall since 1979, we concluded that the event must be rare, whatever the reasons. Radio Telemetry We tagged 63 striped bass with transmitters: 11 in 1981 and 52 in 1982. Three tags failed immedi- ately after release (all in 1982); therefore, 60 fish total were tracked. The study area was surveyed from late June to late July during 13 d in 1981 and 47 d in 1982. Individual striped bass were tracked for periods of 1-14 d: 35 were tracked for >1 d; and 25 were tracked for >2 d. Fish were tracked for an average of 4.3 d in 1981 (range: 1-14 d, N = 11) and 2.2 d in 1982 (range: 1-12 d, A^ = 46). Tracks offish ended because of tag regurgitation, tag failure, and move- ment out of the study area. Operating tags were regurgitated by 15 fish (4 in 1981, 11 in 1982) aver- aging 3.6 d before regurgitation. There were four known tag failures after an average of 3.5 d of obser- vations. Tracking the remaining 41 fish ended after they moved out of the study area or after undetected tag failure. No striped bass were observed moving upstream of the Cabon Station or into tributaries of the river. Surveys below Holyoke Dam located seven tagged fish, one 75 km downstream of Hol- yoke Dam near Hartford, CT. None of these fish returned to Holyoke Dam and they may have con- tinued moving downstream to the Long Island Sound. Twenty additional fish were last observed moving downstream toward Holyoke Dam, and we expected that they also continued past the dam and, possibly, to the Sound. Because there was no spillage over the dam when many tagged fish returned downstream to Holyoke Dam, they passed the dam by entering one or more hydroelectric turbines. Only striped bass tagged in 1982 moved upstream; the average upstream rate was 0.7 km/h (range: 0.30-1.2 km/h, A^ = 11). The mean rate of down- stream movement was 1.9 km/h in 1981 (range: 1.0-3.2 km/h, N = 9); and 2.3 km/h in 1982 (range: 1.0-3.8 km/h, N = 21). Mean rates during the 2 years did not differ significantly (Student's ^-test:P > 0.05). One striped bass, which was located nine times during July 1982, traveled at least 143 km in the study area during 14 d. Nine fish moving downstream in 1981 followed the channel of the river; 61 of 68 locations were in the channel at depths of 3-17 m. Although the actual proportion of deep-channel habitat compared with shoal habitat is unknown, there is much less chan- nel than shoal area. Therefore, the preference for the channel appears strong, as was also found in telemetry studies of adult striped bass in Watts Bar Reservoir, TN (Cheek et al. 1985). A total of 29 striped bass localized for periods of 90 min to 6 d in 13 different holding areas (Fig. 1). (Two localized at more than one area; therefore, a total of 33 such events were recorded.) Two fish were rarely found simultaneously at the same site. All holding areas, except the Route 202 Bridge site, were located near a bank of the river. The Cabot Station site accounted for 8 of the 33 localized periods and for most of the longest periods, i.e., two fish stayed at Cabot Station for 6 d each, one for 3 d, and three for 2 d. Fish activity in holding areas was highly variable. Fish at Cabot Station moved in a stop-and-go man- ner during the day and night, in shallow and in deep water, and in the fast water of the power station discharge and in the slackwater upstream of the sta- tion. They appeared to be feeding in the discharge of the hydroelectric station— we have observed striped bass feeding in the discharge water of the Hadley Falls Hydroelectric Station at Holyoke Dam. Five fish were tracked at the outlet of the North- ampton Oxbow (Fig. 1). All stayed in the main stem within 0.5 km of the outlet, moved in a stop-and-go 145 manner, and used three habitats: a low-flow turbid area, a 10 m deep channel, and sandy shoals up- stream from the outlet. Movements at the S-curve (Fig. 1) and at other holding areas were highly variable: some remained in a small discreet area, others were inactive for long periods, and the rest moved actively within a 0.5 km reach of the river— similar to the movements at Cabot Station and at the Northampton Oxbow. Striped bass followed several patterns of diel behavior. At Cabot Station, which has outside il- lumination at night, four moved actively during both the day and night. Koo and Wilson (1972) also found that adult striped bass were active at night in il- luminated areas. At sites with natural illumination, the movements of 10 striped bass were as follows: 9 moved actively during the day; 6 stopped and 3 were less active at night; and 1 moved only at night. Of the 14 striped bass that we observed for 24 h, 10 increased their activity at dawn, dusk, or both. Dudley and McGahee (1983) found that adults were most active in late afternoon or evening, but noted an increased activity at dawn. Because striped bass feed most actively at dawn and dusk (Raney 1952), the increase in activity during these periods was pro- bably related to feeding. Based on the results of fish passage at Holyoke Dam, behavioral observations using telemetry, and the general thermal niche of subadults reported by Coutant and Carroll (1980) and Coutant (1985), we hypothesize that the movement of subadult striped bass into the Connecticut River is due in part to ther- mal preferences. The upriver migration in May-July places subadults in temperatures closer to their preferred range than those found in Long Island Sound. Tracking of fish in the river indicates a diverse behavioral range of active swimming, rest- ing, and feeding that is consistent with a spring- summer period of high activity and growth. Local attraction to dam tailwaters provides access to abun- dant food (Warner and Kynard 1986), a feature that reinforces the advantages of following thermal cues into the riverine environment. The feeding advan- tages for striped bass will likely increase as the restoration program for American shad and blue- back herring results in an increased abundance of juveniles. Acknowledgments This research was supported by Federal Aid Pro- ject AFS-4-R-21 and Dingell-Johnson Project 5- 29328 to the Massachusetts Division of Fisheries and Wildlife and the Massachusetts Cooperative Fishery Research Unit. The radio telemetry study was a por- tion of a thesis submitted by John P. Warner in par- tial fulfillment of the requirements of the M.S. degree from the Graduate School, University of Massachusetts, Amherst. We thank Holyoke Water Power Co. for providing the space for our holding tanks. Literature Cited Armour, C. L. 1966. Some aspects of bottom types and benthos of a 32-mile segment of the Connecticut River. M.S. Thesis, Univ. Massachusetts, Amherst, 49 p. Cheek, T. E., M. J. Van Den Avyle, and C. C. Coutant. 1985. Influences of water quality on distribution of striped bass in a Tennessee River impoundment. Trans. Am. Fish. Soc. 114:67-76. Coutant, C. C. 1985. Striped bass, temperature, and dissolved oxygen: A speculative hypothesis for environmental risk. Trans. Am. Fish. Soc. 114:31-61. Coutant, C. C, and D. S. Carroll. 1980. Temperatures occupied by ten ultrasonic-tagged striped bass in freshwater lakes. Trans. Am. Fish. Soc. 109: 195-202. Dudley, R. G., and T. G. McGahee. 1983. Winter and altered spring movements of striped bass in the Savannah River, Georgia. Fish. Bull., U.S. 81:420- 425. Knight, A. E. 1975. A tuned-antenna radio telemetry tag for fish. Under- water Telem. Newslt. 5:13-16. Koo, S. S. Y., AND J. S. Wilson. 1972. Sonic tracking striped bass in the Chesapeake and Delaware Canal. Trans. Am. Fish. Soc. 101:453-462. Moffitt, C. M., B. Kynard, and S. G. Rideout. 1982. Fish passage facilities and anadromous fish restoration in the Connecticut River basin. Fisheries (Bethesda, MD) 7{6):2-ll. Nichols, P. R., and R. V. Miller. 1967. Seasonal movements of striped bass, Roccics saxatilis (Walbaum) tagged and released in the Potomac River, Mary- land, 1959-61. Chesapeake Sci. 8:102-124. Raney, E. C. 1952. The life history of the striped bass, Roccus saxatilis (Walbaum). Bull. Bingham Oceanogr. Collect., Yale Univ. 14:5-97. Raney, E. C, W. S. Woolcott, and A. G. Mehring. 1954. Migratory pattern and racial structure of Atlantic Coast striped bass. Trans. 19th North Am. Wild, and Nat. Resour. Conf., p. 376-396. Taylor, R. E., and B. Kynard. 1985. Mortality of juvenile American shad and blueback herring passed through a low-head Kaplan hydroelectric tur- bine. Trans. Am. Fish. Soc. 114:430-435. Warner, J. P. 1983. Demography, food habits, and movements of striped bass, Morone saxatilis Walbaum, in the Connecticut River, Massachusetts. M.S. Thesis, Univ. Massachusetts, Am- herst, 94 p. Warner, J. P., and B. Kynard. 1986. Scavenger feeding by subadult striped bass below a 146 low-head hydroelectric dam. Fish. Bull., U.S. 84:220-222. Boyd Kynard John P. Warner Massachusetts Cooperative Fishery Research Unit University of Massachusetts Holdsworth Hall Amherst, MA 01003 HABITAT PARTITIONING BY SIZE IN WITCH FLOUNDER, GLYPTOCEPHALUS CYNOGLOSSUS: A REEVALUATION WITH ADDITIONAL DATA AND ADJUSTMENTS FOR GEAR SELECTIVITY In 1970, Powles and Kohler hypothesized separa- tion of habitats of adult and juvenile witch flounder, Glyptocephalus cynoglossus, by depth based on sur- veys of Nova Scotia Banks and in the Gulf of St. Lawrence. Juveniles were sampled with a small mesh Icelandic shrimp trawl on the Nova Scotia Banks. These data were supplemented by data ob- tained from Squires' (1961) field records collected during shrimp surveys in the Cabot Strait and Gulf of St. Lawrence in the summers of 1957 and 1958 using a Norwegian deep-sea shrimp trawl. The authors concluded that during the summer months newly metamorphosed and small (<30 cm) witch flounder were found in the 180-288 m depth range. Adult witch flounder (>30 cm) were sampled with a No. 36 Yankee otter trawl on the Nova Scotia Banks from May to October and from November to April. Powles and Kohler (1970) concluded that adult witch flounder were most abundant at a depth range of 92-162 m. In winter months both adults and juven- iles were found together in deeper water while in the summer both groups were separated. Powles and Kohler (1970) suggested that this deepwater distribution of juvenile witch flounder could prevent direct competition with young of more abundant species such as Atlantic cod, Gadus morhua, and American plaice, Hippoglossoides platessoides , and provide a natural conservation against fishery exploitation. Their otter trawl catches, over a depth range of 36-450 m, yielded few juvenile witch flounder, although many small American plaice were captured. Escapement of juvenile witch flounder through the mesh in the wings of the trawl was ruled out because many small plaice were captured on the same grounds. The authors concluded that juvenile witch flounder were absent unless American plaice and witch flounder differed radically in behavior. Other studies of witch flounder depth distribution on the continental slope off Virginia (Markle 1975) and in the Gulf of St. Lawrence, NAFO (Northwest Atlantic Fisheries Organization) Divisions 4R and 4S (LaFleur and Lussiaa-Berdou 1982) supported the habitat separa- tion hypothesis. However, recent studies showed that a No. 36 Yankee shrimp trawl was more efficient in catch- ing juveniles whereas a No. 41.5 Yankee otter trawl was more efficient in catching adult witch flounder (Walsh 1984). In that study juvenile American plaice and witch flounder co-occurred in the shrimp trawl catches; differential catches of witch flounder in the otter trawl was due to the escapement of juveniles. Apparent depth separation proposed by Powles and Kohler (1970) may have been based on data biased by gear selection. Accurate descriptions of life history patterns of witch flounder are important for sound fisheries management, especially with regard to competition with other species and with regard to presumed mechanisms which protect from overfishing. Powles and Kohler (1970) derived their results from sum- mer and winter surveys, and the conclusions were tentative because of potential gear selectivity prob- lem. Therefore, I reevaluated the depth separation hypothesis with additional data taking gear selec- tivity into consideration. Materials and Methods Data used in the analysis were obtained from regular groundfish biomass surveys of the Gulf of St. Lawrence, NAFO Divisions 4R and 4S, by re- search vessels of the Northwest Atlantic Fisheries Centre, St. John's, Newfoundland, during the period 1978-80. In addition, two juvenile flatfish surveys were used: one in the northern Gulf of St. Lawrence, NAFO Division 4R, 1980; and one in the areas of Hermitage Bay and Fortune Bay, NAFO Division 3Ps, in 1981 (Fig. 1). Fishing Gears and Research Designs Groundfish surveys in September and October of 1978-80, NAFO Divisions 4R and 4S by the A. T. Cameron (side trawler) were conducted with a stan- dard No. 41.5 Yankee otter trawl with a stretched mesh size of 127 mm in the wings and reducing to 111 mm in the cod end and a 30 mm mesh cod end liner was used. A total of 188-30 min fishing sets FISHERY BULLETIN: VOL. 85, NO. 1, 1987. 147 CO -a c cs c o S. m > Q O 2 > > 0) I w OS D O 148 were made using a random stratification scheme (Fig. 1). Both juvenile flatfish surveys used a No. 36 Yankee shrimp trawl, with a 38.1 mm mesh through- out and a 12.7 mm cod end mesh liner. The Septem- ber 1980 survey was in the northern Gulf of St. Lawrence, NAFO Division 4R aboard the chartered stern trawler Zagreb. The trawl was equipped with a single tickler chain for approximately 53% of the fishing sets (see Walsh 1984). A total of 53 30-min fishing sets were made (Fig. 1). The October 1981 juvenile survey was in Hermitage Bay and Fortune Bay aboard RV R. V. Shamook. A total of 28 fish- ing sets in depths of 188-402 m were used in the analysis. Most of these sets were of <30-min dura- tion owing to otter doors being stuck in the heavy mud of these bays and the catches were adjusted upward based on the ratio of actual tow to the stan- dard 30 min (Fig. lA). Both juvenile surveys were based on line transects that ran perpendicular to depth contours so that all depth zones would be sampled. Stations in the Gulf of St. Lawrence were about 10 mi apart on each line while the surveys in the two bays were about 2 mi apart. The purpose of these two surveys was to test the use of a small mesh trawl and delineate depth distribution of all flatfishes in the area sampled. Method of Analysis Total lengths of witch flounder were grouped into 2 cm intervals for the analysis. Catches of witch flounder at each station were divided into two size categories: juveniles (<30 cm) and adults (>30 cm). Majority of witch flounder are sexually immature at 30 cm (Powles and Kohler 1970; Beacham 1983). Depths of catches were broken down at 20 m inter- vals and a Kolmogorov-Smirnov two-sample test was applied (Siegel 1956) to the cumulative distribu- tion of both size categories for each data set. The null hypothesis used states that there is no differ- ence in depth distribution of juvenile and adult witch flounder; i.e., the values of the population from which the juvenile sample and the adult sample were drawn have the same cumulative distribution. The alternative hypothesis used stated that there was a difference in depth distribution, i.e., the two- sample cumulative distributions were far apart and suggests the samples came from different popula- tions. The level of significance used was «: = 0.05. A catch frequency was calculated for each size group over 20 m depth intervals. The analysis was used on five data sets: 1) otter trawl catches for 1978-80 in both Division 4R and Division 4S were combined to increase sample size and coverage of the Gulf of St. Lawrence; 2) otter trawl catches in Division 4R, 1980 were used to com- pare with 3) shrimp trawl catches in Division 4R, 1980; 4) shrimp trawl catches in Division 3Ps; and 5) combination of the catches of both gears from sets north of lat. 50°N in Division 4R, 1980. The latter combination of data was used for two reasons: 1) There were no successful sets made by the shrimp trawl in depths <180 m owing to rough bottom while the otter trawl had sets in depths as shallow as 120 m, both vessels were in the same area at the same time, and 2) given a bias in gear selectivity, com- bination of catches of both gears should be repre- sentative of the population located in this small area of northern Esquiman Channel (Fig. 1). Results Trends in depth distribution of witch flounder using different fishing gears showed no significant difference in the cumulative distributions of juvenile and adult witch flounder in all data sets (P > 0.05) (Table 4). No. 41.5 Yankee Otter Trawl Divisions 4R and 45, 1978-80. Juveniles were found in a depth range of 102-464 m with a median located in the 241-260 m depth interval. Adults were distributed in a depth range of 91-484 m with the median located in the 181-200 m depth interval (Table 1, Fig. 2 A). Division 4R, 1980. Both juveniles and adults were distributed in a depth range of 122-464 m. Most of the juveniles were located in the 241-260 m depth interval while the median of the adult witch flounder was located in the 160-180 m depth interval (Table 1, Fig. 2B). No. 36 Yankee Shrimp Trawl Division 4R, 1980. Juvenile and adult witch flounder were widely distributed in a depth range of 187-502 m. The median of juveniles was located in the depth interval 241-260 m while for adults it was the 261-280 m interval (Table 2, Fig. 2C). Division 3Ps, 1981. Juvenile and adult witch flounder were widely distributed in a depth range of 188-402 m. The median of juvenile distribution was in the 281-300 m interval while that for adults was in the 261-280 m interval (Table 2, Fig. 2D). 149 Table 1 .—Cumulative frequency of juvenile (<30 cm) and adult (>30 cm) witch flounder catches over 20 m depth intervals using a No. 41.5 Yankee otter trawl. Depth 20 m NAFO Div. 4R and 4S, " 1978-80 NAFO Div . 4R, 1980 1 Nos. Cumulative Nos. Cumulative Nos. Cumulative Nos. Cumulative intervals <30 cm % % >30 cm % % <30 cm % % >30 cm % % 81-100 0 0 0 6 0.34 0.45 101-120 2 0.11 0.47 4 0.22 0.74 121-140 0 0 0.47 20 1.12 2.23 0 0 0 2 1.08 1.32 141-160 11 0.62 3.02 142 7.99 12.76 2 1.08 5.88 61 32.97 41,72 161-180 36 2.02 11.40 415 23.34 43.55 6 3.24 23.53 42 22.70 69.54 181-200 13 0,73 14.42 148 8.32 54.53 201-220 15 0.84 17.91 82 4.61 60.61 1 0.54 26.47 7 3.78 74.17 221-240 90 5.06 38.84 59 3.32 64.99 241-260 90 5.06 59.77 35 1.97 67.58 5 2.70 41,18 8 4.32 79.47 261-280 41 2.31 69.30 55 3.09 71.66 281-300 38 2.14 78.14 110 6.19 79.82 3 1.62 50,00 5 2.70 82.78 301-320 9 0.51 80.23 33 1.86 82.27 5 2,70 64,71 6 3.24 86.75 321-340 9 0.51 82.33 64 3.60 87.02 1 0,54 67,65 4 2.16 89.40 341-360 5 0.28 83.49 15 0.84 88.13 2 1,08 70.58 1 0.54 90.07 361-380 31 1.74 90.70 59 3.32 92.51 381-400 11 0.62 93.26 32 1.80 94.88 401-420 8 0.45 95.12 7 0.39 95,40 8 4.32 97.06 7 3.78 94.70 421-440 3 0.17 95.81 6 0.34 95,85 441-460 15 0.84 99,30 45 2.53 99,18 461-480 3 0.17 100.00 10 0.56 99,93 1 0.54 100.00 8 4.32 100.00 481-500 0 0 — 1 0.06 100,00 Total 430 1,348 34 151 Table 2.— Cumulative frequency of juvenile (<30 cm) and adult (>30 cm) witch flounder catches over 20 m depth intervals using a No. 36 Yankee shrimp trawl. Depth 20 m NAFO Div. 4R, 1980 NAFO Div. 3Ps, 1981 Nos, Cumulative Nos. Cumulative Nos, Cumulative Nos, Cumulative intervals <30 cm % % >30 cm % % <30 cm % % >30 cm % % 180-190 191-200 201-220 210 4.69 4.91 43 0.96 21.18 10 0.91 2.70 42 3.83 5.87 135 3.01 8,07 5 0.11 23,65 9 0.82 5.12 51 4.69 12.79 221-240 640 14.29 23,04 15 0.33 31.03 31 2,82 13.48 103 9.38 26.96 241-260 1,194 26.26 50.97 18 0.40 39.90 8 0,73 15.63 112 10.20 42.37 261-280 1,209 27.00 79.25 36 0.80 57.64 28 2.55 23.18 74 6,74 52.54 281-300 372 8.31 87.95 21 0.47 67.98 110 10.02 52.83 118 10,75 68.78 301-320 98 2.19 90.25 4 0.09 69.95 15 1.37 56.87 6 0.55 69.60 321-340 242 5.40 95.91 8 0.18 73.89 15 1.37 60.92 29 2.64 73.59 341-360 29 0,65 96.58 1 0.02 74.38 105 9,56 89.22 174 15,85 97.52 361-380 381-400 401-420 — — — — — — 40 3.64 100,00 18 1,69 100.00 411-440 441-460 461-480 76 1.70 98.36 39 0.87 93.60 481-500 19 0.42 98.81 2 0.04 94.58 501-520 51 1.14 100.00 11 0.25 100.00 Total 4,275 203 371 727 No. 36 Yankee Shrimp Trawl Combined with No. 41.5 Yankee Otter Trawl Division 4R, North of Lat. 50° N, 1980. The median of juvenile distribution was located in the 240-260 m interval while that for adults was located in the 180-200 m depth interval (Table 3, Fig. 3). Discussion The results of the analysis do not statistically sup- port the hypothesis that juvenile witch flounder prefer a deeper water habitat than adults, although the median usually shows adults shallower than juveniles. Only the shrimp trawl catches in NAFO Division 4R, 1980 show juveniles shallower than adult witch flounder (Fig. 2C). Combining the data 150 100 A -^^^^ 8 _. /^^^^^ 90 ' ^.-r/ ' 80 - ^^.-^ . ^-— / , // OTTER TRAWL ,-' / 70 - y / • /' ^y • 60 , Diy. 4RS , 1 f^ 1 1 Oiv. 4R . / / / 1 ( 1978-80 1 / 1 1 1980 50 " ■ 1 J 1 / - 40 1 f 1 / 1 / 1 / ■ 30 ■ / / ; / • 1 / 1 ^ - 20 • / / ■ / / - •- / ^/^ 1 1 ' 10 . / /^ . 1 1 f / , «j / / J oc / / &■ 0 ■ ■ « 111 • ■ ■ 1 1 1 1 1 1 ■ t 1 i 1 1 »»*■■■■ UJ 2 100 . -a* ^ J ,. t- ^ r' 'y^ i 90 . C ^/ ^ ' D 1 y^ 1 y^ M J / 3 .^ f f " 80 - . #/ ■ 70 - / / ^_-^ SHRIM ' TRAWL 1 1 )\ • 1 - 60 • / / / / Oi» 4R / J Oiv 3P - 50 ■ / / / / 1980 . 1 y^ / f 1 1 1981 40 // ■ 1 1 1 \ 1 \ 1 / • 30 ~ // / / / / ■ / / / / / / - 20 " ^ j • / / / / - 10 - • / / - 0 L. 1 1 I 1 1 1 . _J 1 1_ ' ' ' ' ■ . 1 , , — 1 1 1 1 — » 1 1 1 1 1 1 ■ ■ J 1 1 1 1 ^"^ ?P ^'? ^^^ S? ^*? ^^^ ..-P i^^ .\^ ^S- A- sV ^',- ^o>- ^^ DEPTH (METERS) Figure 2.— Cumulative frequency distribution of juvenile and adult witch flounder by both fishing gears. Dash line: adults (>30 cm); solid line: juveniles (<30 cm). A) No. 41.5 Yankee otter trawl, Division 4RS 1978-80 combined. B) No. 41.5 Yankee otter trawl, Division 4R 1980. C) No. 36 Yankee shrimp trawl, Division 4R, 1980. D) No. 36 Yankee shrimp trawl, Division 3Ps, 1981. sets from two fishing gears for analysis of the north- ern Esquiman Channel area of NAFO Division 4R in 1980 takes into account biases in gear selectivity (Fig. 3). In this study of the Gulf of St. Lawrence, juveniles are more vulnerable to shrimp trawls as Powles and Kohler (1970) showed in their study. Low catches of adult witch flounder in all data sets indicate that they are widely dispersed and not readily accessible in any large numbers regardless of fishing gears used. A discrete separation of adults and juveniles does not exist. Although a large percentage of adult witch flounder are found Figure 3.— Cumulative frequency distribution of juvenile and adult witch flounder in the Northern Esquiman Channel area (north of lat. 50°N). Catches of a No. 41.5 Yankee otter trawl and No. 36 Yankee shrimp trawl combined. Division 4R, 1980. 100 90 80 y- Z 70 ? 60 UJ > 50 i- «t = 40 z " 30 20 10 X^ , //' / // // / / / / / / / / ^y' / DIv 4R (Noflh of 50°) /' / SHRIMP TRAWL / / / OTTER TRAWL / / / / / / 1 1 1 1 1 j 1 ^X 1 j^ -^ ^ ^ ^ ^ ^' ^^ S" V -v t^ c^ '^ DEPTH (METERS) 151 Table 3. — Cumulative frequency of juvenile (<30 cm) and adult (>30 cm) witch flounder catches over 20 m depth intervals. Catches of a No. 41.5 Yankee otter trawl (8 sets) and a No. 36 Yankee Shrimp trawl (48 sets) combined for the northern Esquiman Channel area (lat. 50°N). Depth 20 m NAFO Div. 4R, 1980: Sets north of 50' 'N Nos. Cumulative Nos. Cumulative intervals <30 cm % % >30 cm % % 120-140 0 0 0 2 0.04 0.66 141-160 2 0.09 0.05 61 1.37 20,86 161-180 6 0.13 0.19 42 0.94 34,77 181-200 2.0 4.70 5.24 43 0.96 49,01 201-220 136 3.05 8.50 12 0.27 53.98 221-240 640 14.33 23.88 15 0.34 57.95 241-260 1,199 26.85 52.68 26 0.58 66.56 261-280 1,209 27.08 81.72 36 0.81 78.48 281-300 375 8.40 90.73 26 0.58 87.09 301-320 103 2.31 93.20 10 0.22 90.40 321-340 243 5.44 99.04 12 0.27 94.37 341-360 31 0.69 99.78 2 0.04 95.03 361-380 — — 99.78 — — 95.03 381-400 — — 99,78 — — 95.03 401-420 8 0.18 99,98 7 0.16 97.35 421-440 — — 99.98 — — 97.35 441-460 — — 99.98 — — 97.35 461-480 1 0.02 100.00 8 0.18 100.00 Total 4,163 302 Table 4.— Results of Kolmorgov-Smirnoff two sample test on each data set. Level of significance used was « = 0.05. Fishing gear NAFO Total Total Divi- no. no. D Table Significant Year sions juveniles adults statistic value at « = o.05 1,348 0.4270 0.0753 Not significant 185 0.4770 0.2582 Not significant 203 0.2220 0.0977 Not significant 727 0.2936 0,0868 Not significant 302 0.4448 0,0810 Not significant No. 41.5 Yankee otter trawl 1978-80 4RS 430 No, 41,5 Yankee otter trawl 1980 4R 34 No. 36 Yankee shrimp trawl 1980 4R 4,275 No. 36 Yankee shrimp trawl 1981 3Ps 371 No. 41,5 Yankee otter trawl 1980 4R 4,163 + No, 36 Sets north Yankee shrimp of lat. 50°N. trawl shallower than juveniles in the 100-200 m range, they are also found in sufficient numbers in all depths >200 m (Table 1). LaFleur and Lussiaa- Berdou (1982) research surveys in the Gulf of St. Lawrence using a Yankee 41.5 otter trawl found juveniles in the 200-300 m depth range and, while supporting Powles and Kohler's (1970) depth separa- tion hypothesis also noted that even at depths >300 m, a significant proportion of adult witch flounder were caught. W. R. Bowering (Department of Fish- eries and Oceans, St. John's, Newfoundland, pers. commun. 1986) has found that adult witch flounder in NAFO Division 2J3KL also exhibit two peak concentrations: one at 101-200 m and a second one in depths >300 m. That adult witch flounder are not concentrated during summer months is the reason why an economical commercial fishery only occurs during winter months (Powles and Kohler 1970; Bowering and Pitt 1974; Bowering and Brodie 1984). Catches of witch flounder in the 1981 shrimp trawl survey of the two deepwater bays in NAFO Divi- sion 3Ps shows that although adults are usually average shallower than juveniles, they are also dis- persed across deeper depth zones (Table 1). This sug- gests that in confined areas of deepwater bays, distribution patterns of adult witch flounder is more concentrated than in large open areas like the Gulf 152 of St. Lawrence. Similarly, adult witch flounder have been reported concentrated in the deepwater of St. Georges Bay, NAFO Division 4R during the summer months where a localized fishery occurs in depths of 300 m (Bowering and Brodie 1984). In conclusion, juvenile witch flounder are distri- buted differently than the adult population off con- tinental shelf areas of the Gulf of St. Lawrence. However, the two populations are not discretely separated as proposed by Powles and Kohler's (1970) niche separation hypothesis. Bowers (1960) concluded that witch flounder in the Irish Sea have no definitive separation. Heavy exploitation of juvenile witch flounder is prevented by the behavior of this size group making them less vulnerable to commercial otter trawls. The difference may be related to difference in preferred food items or dis- tribution of predators. Further research is required to establish the mechanisms for the difference in depth distribution documented by this study. the slope off Virginia. J. Fish Res. Board Can. 32:1447- 1450. Powles, P. M., and A. C. Kohler. 1970. Depth distribution of various stages of witch flounder (Glypocephalus cynoglossus) off Nova Scotia and in the Gulf of St. Lawrence. J. Fish Res. Board Can. 27:2053-2062. SlEGEL, S. 1956. Nonparametric statistics for the behavioral sciences. McGraw-Hill Book Co., N.Y., 312 p. Squires, H. J. 1961. Shrimp survey in the Newfoundland fishing area, 1957 and 1958. Bull. Fish Res. Board Can. 129:vii -i- 29 p. Walsh, S. J. 1984. Relative efficiency of two bottom trawls in catching juvenile and commercial sized flatfishes in the Gulf of St. Lawrence. J. Northw. Atl. Fish Sci. 5:181-188. S. J. Walsh Science Branch Department of Fisheries and Oceans P.O. Box 5667 St. John's, Newfoundland AlC 5X1, Canada Acknowledgments Special thanks for informative suggestions from J. A. Brown, MSRL. I am grateful to Jake Rice of the Northwest Atlantic Fisheries Center and two other referees for constructive comments and criticisms. Literature Cited Beacham, T. D. 1983. Variability in size and age at sexual maturity of witch flounder, Glyptocephalus cynoglossus, in the Canadian Mari- times region of the Northwest Atlantic Ocean. Can. Field- Nat. 97(4):409-422. Bowering, W. R., and W. B. Brodie. 1980. An evaluation of recent management strategy for witch in the Gulf of St. Lawrence (NAFO Div. 4RS). CAFSAC Res. Doc. 80/49, 20 p. 1984. Distribution of witch flounder in the northern Gulf of St. Lawrence and changes in its growth and sexual matur- ity patterns. North Am. J. Fish. Manage. 4A:399-413. Bowering, W. R., and T. K. Pitt. 1974. An assessment of witch (Glyptocephalus cynoglossus) for ICNAF Div. 2J-3KL. ICNAF Res. Doc. 74/48, Ser. No. 3255, 7 p. Bowers, A. B. 1960. Growth of the witch {Glyptocephalus cynoglossus L.) in the Irish Sea. J. Cons. Perm. Int. Explor. Mer 25:168- 176. Lafleur, p. E., and J. P. Lussiaa-Berdou. 1982. La plie grise (Glyptocephalus cynoglossus) dans le nord du Golfe du Saint-Laurent (Div. 4R et 4S de I'opano). Donnes sur I'ecologie et I'exploitation. Gov. Quebec Minist. Agric. Pecheries aliment., Cah. Inf. No. 97, 30 p. Markle, D. F. 1975. Young witch flounder, Glyptocephalus cynoglossus, on MOVEMENT OF TAGGED LINGCOD, OPHIODON ELONGATUS, IN THE PACIFIC NORTHWEST Lingcod, Ophiodon elongatus, is a commercially and recreationally important West Coast species. Most previous studies have indicated that lingcod is a relatively nonmigratory species (Hart 1943; Chat- win 1956; Phillips 1959). More than 90% of the adults remained within 5 mi (8.1 km) of the point of tagging for as long as several years. We tagged lingcod in the eastern Strait of Juan de Fuca and near San Juan Island, WA, from 1976 and 1981. We present results from tags returned by fishermen through 1985. The tag returns were analyzed primarily to show the extent of migration. We also analyzed recaptures by sex, size, direction of movement, and the effects of tag type and the location of tagging. Methods From 1976 to 1978, relatively small numbers of lingcod were tagged, incidental to a tagging study directed to rockfish {Sebastes sp.) (Mathews and Barker 1984), in which rod-and-reel with artificial lures was used to capture fish for tagging. From 1979 to 1981 tagging effort was for lingcod using a chartered commercial vessel trolling with a string of 6-10 jigs or other artificial lures from a hydraulic gurdy. FISHERY BULLETIN: VOL. 85, NO. 1, 1987 153 A total of 1,692 lingcod were tagged during 1976- 81. Most of the lingcod (over 90%) were tagged dur- ing March through May. When caught singly, they were immediately tagged and released. If several lingcod were brought aboard at the same time, they were held in a circulating seawater tank until tagged. All tagged fish were measured (fork length) to the nearest millimeter. From 1978 to 1981, sex was determined by the presence of the anal papillae in males. Only fish not injured by capture were tagged and released. Those that bled, or that were hooked in the gills or throat, or that otherwise ap- peared disabled were not tagged. Three types of spaghetti end tags were used: Anchor with #20 tubing (Floyi FD-67, Floy Co., Seattle, WA); small dart with #20 tubing (Floy FT-2); and large dart with #13 tubing (Floy FT-1). The tagging area and number tagged at each loca- tion are shown in Figure 1. The principal tagging locations were Middle Bank, a low relief, hard rub- ble bottom bank of about 6 km^ and 20-60 m deep; Hein Bank, of similar area to Middle Bank but shallower (6-30 m deep), having a softer bottom and extensive kelp beds; and San Juan Channel, a passage with high relief, rocky substrate 2-6 km wide, coursing among several of the San Juan Islands. Most of the tagging in San Juan Channel was done near Turn Island in about 30 m of water. A few lingcod were tagged at other locations near San Juan Island. Recapture information, including primarily the date and place of capture, was obtained from tags 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Figure L— Lingcod tagging area in relation to western Washington. Small numbers show depth contours in fathoms (1 fathom = 1.829 m) and large numbers show numbers tagged by location. 154 returned voluntarily by sport fishermen and com- mercial troll and trawl fishermen. A $2 reward was offered for the return of the tags. Several fishermen were personally contacted to clarify the information they provided and to seek specific information on where and how they fished; all the fishermen were cooperative. Assuming these fishermen were repre- sentative of all those who returned tags, we believe that the overall recovery information was accurate. The size and sex distributions of the tagged ling- cod are shown in Graphs I, II, and III of Figure 2. Eighty-six percent of all tagged lingcod were sexed, and of this sample 87% were males. The reported size ranges at maturity are 40-46 cm for males and 70-76 cm for females (Forrester 1969; Hart 1973). Operationally, we define migratory and nonmigra- tory lingcod as fish recaptured at distances greater than and <8.1 km (5 mi), respectively, from the tag- ging site. This reference distance has been used for similar purposes in previous tagging studies. Since the recovery locations were usually given by the name of a geographical location such as "Middle Bank" or "Turn Island", there was some impreci- sion in estimating the distance moved. However, the fishing area associated with such named locations is <8.1 km in diameter. Thus, for example, a fish tagged on Middle Bank and recaptured on Middle Bank was assumed to have travelled <8.1 km. Chi-square contingency table analysis was used for comparing recapture rates by tag type and sex, for comparing release-length frequency distributions of migratory and nonmigratory recoveries, and for comparing migrational tendencies by sex. A chi- square goodness of fit test was used to test the null hypothesis that the release-length distribution of all recaptured lingcod was the same as that of all tagged lingcod. For both of the length-frequency tests, lengths were grouped into 5 mm intervals, but at the tails of the distribution the intervals were wider than 5 mm to follow the rule for chi-square analysis that no expected cell frequency should be <1.0 and that no more than 20% of expected cell frequencies should be <5.0 (Zar 1974, p. 50). One- way analysis of variance was used to test the null hypothesis that the average time between tagging and recapture was the same for fish that had migrated different distances. Most of our tagged males and about half of our tagged females were large enough to be reproduc- tively mature when tagged. Results There were no significant differences among Length at maturity 70 Length cm 120 Figure 2.— Length-frequency distributions of tagged lingcod. I - known male lingcod tagged; II - known female lingcod tagged; III - all lingcod tagged; IV - release length distribution of all tagged lingcod recovered less than 8.1 km from release location; V - release length distribution of all tagged lingcod recovered more than 8.1 km from release location. recovery rates by tag type (x^ = 1.90 with 2 df; 0.26

8.1 km from the tagging location (Table 3). Of the 74 that migrated, 61 were recaptured 8.1-50 km from the tagging location and 13 were recaptured farther than 50 km from the tagging location. The extent of migration depended on the location of the tag- ging site. Only one of 15 recaptured lingcod tagged in San Juan Channel migrated, whereas 70 of 117 (60%) tagged at Middle Bank and Hein Bank migrated (Table 3). The predominant pattern of movement was west and south through the Strait of Juan de Fuca; 65 of the 74 migratory lingcod were recaptured south and west of the tagging site, but only 9 were re- covered north and east of the tagging site (Table 3). The null hypothesis that lingcod were as likely to go south/west as north/east was rejected (x^ = 42.4 with 1 df; P < 0.001). Five recaptures from the Pacific Ocean were reported; the one farthest from the tagging location was caught off Newport, OR, a migration of 564 km.^ The longest migration to the north/east was to Porlier Pass, BC, Canada, about 75 km from the tagging site. The greatest number of recaptures (34) was from Constance Bank, located in Canadian waters about 18 km west of Middle Bank. Most of the Constance Bank recap- tures were made by Canadian trawlers. About one third of the total reported recaptures were taken in Canadian waters and two thirds in U.S. waters, ^This unusual recovery location was verified by follow-up corre- spondence. The fish was recaptured 6.5 yr after tagging by a small coastal trawler that usually fished 3-8 mi off Newport's south jetty. Table 2.— Number of lingcod tagged, 1976-81, and recaptured through 1985 by year of recapture. Ypar No Number recaptured by year O/n tagged tagged 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 Unknown' Total recaptured 1976 41 1 2 0 1 0 0 1 0 0 0 0 5 12.2 1977 101 1 5 4 1 0 0 2 0 0 0 13 12.9 1978 87 1 4 0 2 0 0 0 0 0 7 8.0 1979 507 8 24 5 4 1 0 1 0 43 8.5 1980 535 19 14 3 4 0 1 0 41 7.7 1981 421 29 7 7 1 1 1 46 10.9 Unknown^ — 1 1 0 0 0 0 0 2 — Total 1,692 1 3 6 17 45 51 15 14 1 3 1 157 9.3 'Year of recapture not reported. 'Tag number unreadable when recaptured. Table 3.— Distribution of recoveries of tagged lingcod by location of tagging, distance of migration and direction of migration. No. recovered <8 1 km No. recovered farther than 8.1 from tagging location km Recovery location unknown relative to Tagging No. from tagging 8.1-50 km >50 km 8.1-50 km >50 km Total o/o location tagged location W and/or S W and/or S E and/or N E and/or N tag location recaptured recaptured Middle Bank 1,214 47 50 6 4 4 6 117 9.6 Hein Bank 253 10 3 2 1 0 0 16 6.3 San Juan Channel 153 14 1 0 0 0 0 15 9.8 Miscellaneous 72 4 2 1 0 0 0 7 9.7 Unknown' — — — — — — 2 2 — Total 1,692 75 56 9 5 4 8 157 9.3 'Tag number unreadable when recaptured. 156 which indicates that Ungcod in the study region are an international resource. The time between tagging and recapture averaged 18 mo (Table 4) and did not differ significantly by the distance traveled (^2,128 = 1-32; 0.25 < P < 0.50). The majority of the nonmigratory recaptures were caught in May-July, but the migratory fish were caught mostly in August-October. This difference could not be attributed to any seasonal pattern of migration and was probably a sampling artifact. Fishing effort by commercial troUers and sport fishermen in the tagging areas peaked during May- July, while the fishing effort of the trawl fleet on Constance Bank, from which many of the migratory recaptures came, peaked in late summer and fall (Smith 1981; Leaman 1982, 1983, 1984). We found that migratory tendency apparently did not depend on individual size. Figure 2 shows a com- parison of the release length-frequency distributions of lingcod recaptured <8.1 km (Graph IV) and more than 8.1 km (Graph V) from release location. The null hypothesis that migratory and nonmigratory lingcod have the same length distribution was ac- cepted ix- = 13.09 with 10 df; 0.10

50 km Total 75 55 11 131 15.1 18.6 18.0 18.0 0-71.2 3.2-76.0 2.2-76.6 0-76.6 Table 5.— Distributions by sex of the lingcod recaptured <8.1 km from the tagging location and of those recaptured >8.1 km from the tagging location. Sex No. tagged and sexed Number recaptured <8.1 km >8.1 km Total % recaptured Male Female 1,279 183 57 63 12 5 120 17 9.4 9.3 rates were virtually the same for the two sexes. The null hypothesis that male and female tagged re- coveries represent two populations with equal pro- portions of nonmigratory and migratory individuals was accepted (x" = 3.14 with 1 df; 0.05 Vondracek, G., S. R. Hanson, and P. B. Moyle. Sacramento squawfish, Ptychocheilus grandis, predation on juvenile chinook salmon, Oncorhynchus tshawytscha , below the Red Bluff Diver- sion Dam in the Sacramento River, California. Manuscr. in prep. Wildlife and Fisheries Biology, University of California, Davis, CA 95616. placing a numbered Floy^ anchor tag between the rays of the dorsal fin. After the prescribed digestion period Sacramento squawfish were netted from the small experimen- tal tanks and placed into the foam-lined trough. A catheter connected to a small water pump was in- serted into the anus. Digestion tract contents were flushed through the mouth and collected in a fine mesh net. The digestive tract contents were weighed (salmon were weighed individually if digested <30% and en masse if >30%) and placed in a drying oven in 60°C for 24 h. Dry weights did not change after 24 h. An initial dry weight of each ration was deter- mined by sacrificing 5 to 10 salmon prior to the digestive trials. Mean percent dry weight of the salmon was 20.8 ± 2.0% for the 10° and 15°C trials and 21.9 ± 2.1% for the 5° and 20°C trials. If the dry weight of the digested ration exceeded the esti- mated initial dry weight, the percent of the ration was set to 100%. The dry weight of the digested ration exceeded the estimated initial dry weight dur- ing 17 trials with a mean of about 104%. Digestive rates at each temperature were deter- mined by linear regression of the percent of the ra- tion remaining in the alimentary tract versus time after force feeding. The initial wet weights of the salmon fed to the squawfish were not used in the regression analysis. Time for alimentary tract evac- uation for each temperature was assumed to be the point where the extrapolated regression for diges- tion intersected the x axis (time after force feeding). Results The digestive rates of Sacramento squawfish were directly related to temperature, while the gastric evacuation times were inversely related to temper- ature (Fig. 1). The digestive rates were 1.8%/h at 5°C, 2.6%/h at 10°C, 6.3%/h at 15°C, and 8.2%/h at 20°C. Gastric evacuation times were 61 h at 5°C, 38 h at 10°C, 17 h at 15°C, and 14 h at 20°C. The digestive process appeared to involve at least two phases. During the initial phase, the wet weight of the ingested salmon increased. The duration of the initial phase was inversely related to the ex- perimental temperature. At 5°C the initial phase was at least 16 h, 4 h in duration at 10°C, 2 h at 15°C, and approximately 2 h at 20 °C. During the second phase the percent dry weight of salmon re- maining in the digestive tract decreased linearly with time. ^Reference to trade names does not imply endorsement by the National Marine Fisheries, Service, NOAA. 160 100 a z z < cc Q o o u. z LU o oc UJ a. lOOi 10C r=-.869 N:39 Y = 98.9-2.60X 20C r=-.971 Y:118.2-8.21X TIME(HOUR) Figure 1.— The percent dry weight of juvenile chinook salmon remaining in the digestive tract of force- fed Sacramento squawfish after specified digestion periods at 5°, 10°, 15°, and 20°C. Each squawfish was fed four salmon. The mean time for complete gastric evacuation over tlie temperature range examined can be cal- culated using a curvilinear equation in the form: logGE = 1.996 - 0.045T (1) where GE = gastric evacuation (h) T = temperature (°C) Although other equations may be equally applicable, Equation (1) fits the data well {r- = 0.978, Fig. 2). Discussion The digestion and gastric evacuation rates of Sacramento squawfish in relation to temperature differ in some respects than for the northern squaw- fish. The gastric evacuation times of the Sacramento squawfish in this study are different from the times calculated from equations presented in Falter (1969) 5.0 4.0 (V o 3.0 cc O X LLl I- 2.0 1.0 5 10 15 20 TEMPERATURE (ISX) Figure 2.— Estimated time for complete gastric evacuation of Sacramento squawfish force-fed four juvenile chinook salmon at 5°, 10°, 15°, and 20°C. 161 and Steigenberger and Larkin (1974) (Table 1); how- ever, the times noted in this study are bracketed by the other studies. Falter's equations predict that gastric evacuation would be complete in about 29 h at 6° and 10°C and about 10 h at 16.5° and 20°C. In contrast Steigenberger and Larkin's data predict that northern squawfish would complete gastric evacuation in 84, 51, 23, and 13 h at 6°, 10°, 15°, and 20°C, respectively. Table 1 .—Estimated time (hours) for total evacuation of stomach contents of Sacramento squawfish held at selected temperatures. Temperat ure (°C) Source 6 10 15 20 Falter 1969 30 29 9 (16.5°C) 11 Steigenberger and LarlP>0.05 (1.5) (5.5) (4.0) (1.0) (1.0) Worton Point 1984 0 11 0 0 1986 (0.5) 2 (1.4) (6.2) 12 (15.8) (3.4) 11 (8.6) (0.8) 3 (2.2) 10.51 P < 0.01 Combined 1984 3 31 6 0 0 1986 (1.8) 2 (18.7) 21 (14.7) 34 (1.5) 4 (3.9) 8 26.62 P < 0.01 (3.1) (32.3) (25.3) (2.5) (5.1) take coastal migrations (c.f. Setzler et al. 1980). In the Chesapeake Bay and Hudson River, tagging studies suggest that individuals less than age 2 do not migrate extensively from their natal tributaries (c.f. Setzler et al. 1980). After this sedentary period, females begin to leave the Chesapeake Bay for coastal waters and virtually all females older than age 4 return only to spawn (Kohlenstein 1980). Females do not mature sexually until age 3 at the earliest and most do not mature until age 4 or 5 (Jones et al. 1977). In contrast, few males leave the Chesapeake Bay until age 4 or 5 and virtually all age 2 are sexually mature. Tagging studies by Manseuti (1961) suggest that larger males (ages 3-4) moved greater distances within the Chesapeake than small males (ages 0-2). Massman and Pacheco (1961) supported this conclusion and also found that James and York River fish tended to migrate north- ward in the bay proper. These migration studies fit nicely with the data presented here, if indeed the changes in mtDNA frequencies were due to immi- gration from the James and York Rivers. Further study of striped bass population dynamics are needed to test the hypotheses outlined above. Of particular importance will be an assessment of populations from the James and York Rivers. Whatever the outcome, the data presented here will need to be considered in management plans for this economically important species. Acknowledgments I thank Harley Speir, Steve Early, John Foster and Louis Rugolo of Maryland DNR for their assis- tance in this study. Comments by two anonymous reviewers improved an earlier version of the manu- script. The research was supported by Maryland Sea Grants R/F-39 and R/F-57. Literature Cited Boone, J. G., and J. Uphoff. 1983. Estuarine fish recruitment survey. Maryland DNR Tidewater Admin. Annu. Fed. Aid. Rep. Proj. F-27-R-10. Chapman, R. W., and D. A. Powers. 1984. A method for the rapid isolation of mitochondrial DNA from fishes. Maryland Sea Grant Tech. Rep. UM-SG-TS- 84-05, 11 p. Grove, T. L., T. S. Berggren, and D. A. Powers. 1976. The use of innate genetic tags to segregate spawning stocks of striped bass (Morone saxatilis). Estuarine Pro- cess. 1:166-176. Jones, P. W., J. S. Wilson, R. T. Morgan II. H. R. Lunsford, AND J. Lawson. 1977. Potomac River fisheries study; striped bass spawning stock assessment. Interpretive report 1974-1976. Univ. Md. CEES Ref. No. 77-55-CBL. Kohlenstein, L. C. 1980. Aspects of the population dynamics of striped bass {Morone saxatilis) spawning in Maryland tributaries of the Chesapeake Bay. The Johns Hopkins University Applied Physics Laboratory, PPSE T-14. Mansueti, R. J. 1961. Age, growth and movements of the striped bass, Roc- cits saxatilis, taken in size selective fishing gear in Mary- land. Chesapeake Sci. 2:9-36. Massman, W. H., and A. L. Pacheco. 1961. Movements of striped bass tagged in Virginia waters of the Chesapeake Bay. Chesapeake Sci. 2:37-44. Morgan, R. P., II, T. S. Y. Koo, and G. E. Krantz. 1973. Electrophoretic determination of population of the striped bass, Morone saxatilis, in the upper Chesapeake 169 Bay. Trans. Am. Fish. Soc. 102:21-32. Setzler, E. M., W. R. Boynton, K. V. Wood, H. H. Zion, L. Lubbers, N. K. Mountford, P. Frere, L. Tucker, and J. A. MlHURSKY. 1980. Synopsis of biological data on striped bass, Morone saxatilis (Walbaum). FAO Fish. Synop. No. 121. 69 p. SiDELL, B. D., R. G. Otto, D. A. Powers, M. Karweit, and J. Smith. 1980. Apparent genetic homogeneity of spawning striped bass in the Upper Chesapeake Bay. Trans. Am. Fish. Soc. 109:99-107. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry. The principles and practice of statistics in biological research. Freeman, San Franc, 776 p. Robert W. Chapman Chesapeake Bay Institute The Johns Hopkins University Shady Side, MD 207 6J^ 170 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, hi.s/her affiliation, and mailing address, 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 American Fisheries Society Special Publication No. 12, A List of Com- mon and Scientific Names of Fishes fi^om the United States and Canada, Fourth Edition, 1980. Ifext footnotes should be typed separately from the text. 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 5% 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. Consistency in headings and format is desirable Vertical rules should be avoided, as they make the tables more expensive to print. Footnotes in tables should be numbered sequentially in arable numerals. lb avoid confusion with powers, they should be placed to the left of numerals. 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 sur- name under the heading "Literature Cited." Only the author's surname and initials are required in the literature cited. The accuracy of the literature cited is the respon- sibility of the author. Abbreviations of names of periodicals and serials should conform to Serial Sources for the BIOSIS Data Base. {Chemical Abstracts 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), %, 7oo, and so forth, should be used. Abbreviate units of measure only when used with numerals. Periods are only rarely used with abbrevia- tions. We prefer that measurements be given in metric units; other equivalent units may be given in parentheses. FORM OF THE MANUSCRIPT The original of the manuscript should be typed, double- spaced, on white bond paper. Please triple space above headings. We would rather receive good duplicated copies of manuscripts than carbon copies. The sequence 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 arabic 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— Continiied KYNARD, BOYD, and JOHN P. WARNER. Spring and summer movements of subadult striped bass, Morone saxatilis, in the Connecticut River 143 / WALSH, S. J. Habitat partitioning by size in witch flounder, Glyptocephalus cynoglosstcs: a reevaluation with additional data and adjustments for gear selectivity 147 MATHEWS, S. B., and M. LaRIVIERE. Movement of tagged lingcod, Ophiodon eJxmgatus, in the Pacific Northwest " 153 VONDRACEK, BRUCE. Digestion rates and gastric evacuation times in relation to temperature of the Sacramento squawfish, Ptychocheilus grandis 159 HANLON, ROGER T, PHILIP E. TURK, PHILLIP G. LEE, and WON TACK YANG. Laboratory rearing of the squid Loligo pealei to the juvenile stage: growth comparisons with fishery data 163 CHAPMAN, ROBERT W. Changes in the population structure of male striped bass, Morone saxatilis, spawning in the three areas of the Chesapeake Bay from 1984 to 1986 . . . 167 • GPO 791-008 .,< "«o». Fishery Bulletin Marine Bioloeicai Laboratory '^^ATES O^ ^ Marine Biological Laboratory LIBRARY JUL 2 8 1987 Woods Hole, Mass. Vol. 85, No. 2 April 1987 JONES, CYNTHIA, and EDWARD B. BROTHERS. Validation of the otolith increment aging technique for striped bass, Morone saxatilis, larvae reared under suboptimal feeding conditions 171 LAI, HAN-LIN. Optimum allocation for estimating age composition using age-length key 179 GREELEY, MARK S., JR., DANIEL R. CALDER, and ROBIN A. WALLACE. Oocyte growth and development in the striped mullet, Mugil cephalus, during seasonal ovarian recinidescence: relationship to fecundity and size at maturity 187 FEENEY, RICHARD F. Development of the eggs and larvae of the yellowchin sculpin, Icelinus quadriseriatvs (Pisces: Cottidae) 201 THEILACKER, GAIL H. Feeding ecology and growth energetics of larval northern anchovy, Engraulis mordax 213 WYLLIE ECHEVERRIA, TINA. Thirty-four species of California rockfishes: maturity and seasonality of reproduction 229 HALES, L. STANTON, JR. Distribution, abundance, reproduction, food habits, age, and growth of round scad, Decaptertcs punctatvs, in the South Atlantic Bight 251 BRANSTETTER, STEVEN, J. A. MUSICK, and J. A. COLVOCORESSES. A comparison of the age and growth of the tiger shark, Galeocerdo cuvieri, from off Virginia and from the northwestern Gulf of Mexico 269 PETERSON, CHARLES H., HENRY C. SUMMERSON, and STEPHEN R. FEGLEY Ecological consequences of mechanical harvesting of clams 281 IVERSEN, EDWIN S., EDWARD S. RUTHERFORD, SCOTT R BANNEROT, and DARRYL E. JORY. Biological data on Berry Islands (Bahamas) queen conchs, Strom- bus gigas, with mariculture and fisheries management implications 299 PEREZ FARFANTE, ISABEL. Revisions of the gamba prawn genus Psmdaristms, with description of two new species (Crustacea: Decapoda: Penaeoidea) 311 MOFFIT, ROBERT B., and JEFFREY J. POLOVINA. Distribution and yield of the deep- water shrimp Heterocarptis resource in the Marianas 339 MERRICK, RICHARD L., THOMAS R. LOUGHLIN, and DONALD G. CALKINS. Decline in abundance of the northern sea lion, Eumetopias juhatus, in Alaska, 1956-86 351 YORK, ANNE E., and PATRICK KOZLOFF On the estimation of the numbers of northern fur seal, Callorhinus ursinus, pups bom on St. Paul Island, 1980-86 367 {Continued on hack cover) Seattle, Washington U.S. DEPARTMENT OF COMMERCE Malcolm Baldrige, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Anthony J. Calio, Administrator NATIONAL MARINE FISHERIES SERVICE William E. Evans, Assistant Administrator 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. 1 103. 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. Bruce B. Collette National Marine Fisheries Service Dr. Robert C. FVancis 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 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 CJommerce 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 through 1 April 1986. Fishery Bulletin CONTENTS Vol. 85, No. 2 April 1987 JONES, CYNTHIA, and EDWARD B. BROTHERS. Validation of the otolith increment aging technique for striped bass, Morone saxatilis, larvae reared under suboptimal feeding conditions 171 LAI, HAN-LIN. Optimum allocation for estimating age composition using age-length key 179 GREELEY, MARK S., JR., DANIEL R. CALDER, and ROBIN A. WALLACE. Oocyte growth and development in the striped mullet, Mugil cephalus, during seasonal ovarian recrudescence: relationship to fecundity and size at maturity 187 FEENEY, RICHARD F. Development of the eggs and larvae of the yellowchin sculpin, Icelinus quadriseriatus (Pisces: Cottidae) 201 THEILACKER, GAIL H. Feeding ecology and growth energetics of larval northern anchovy, Engraulis mordax 213 WYLLIE ECHEVERRIA, TINA. Thirty-four species of California rockfishes: maturity and seasonality of reproduction 229 HALES, L. STANTON, JR. Distribution, abundance, reproduction, food habits, age, and growth of round scad, Decapterus punctatics, in the South Atlantic Bight 251 BRANSTETTER, STEVEN, J. A. MUSICK, and J. A. COLVOCORESSES. A comparison of the age and growth of the tiger shark, Galeocerdo cuvieri, from off Virginia and from the northwestern Gulf of Mexico 269 PETERSON, CHARLES H., HENRY C. SUMMERSON, and STEPHEN R. FEGLEY. Ecological consequences of mechanical harvesting of clams 281 IVERSEN, EDWIN S., EDWARD S. RUTHERFORD, SCOTT P BANNEROT, and DARRYL E. JORY. Biological data on Berry Islands (Bahamas) queen conchs, Strom- bus gigas, with mariculture and fisheries management implications 299 PEREZ FARFANTE, ISABEL. Revisions of the gamba prawn genus Psmdaristeus, vdth description of two new species (Crustacea: Decapoda: Penaeoidea) 311 MOFFIT, ROBERT B., and JEFFREY J. POLOVINA. Distribution and yield of the deep- water shrimp Heterocarpus resource in the Marianas 339 MERRICK, RICHARD L., THOMAS R. LOUGHLIN, and DONALD G. CALKINS. Decline in abundance of the northern sea lion, Eumetopias jubatus, in Alaska, 1956-86 351 YORK ANNE E., and PATRICK KOZLOFF On the estimation of the numbers of northern fur seal, Callorhinus ursinus, pups born on St. Paul Island, 1980-86 367 (Continued on next 'page) P^"*"'"^""^™""""^^^^^^^i^i^^^^ Marine Biological Laboratory | LIBRARY JUL 2 8 1987 Seattle, Washington 1987 Woods Hole, Mass For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington DC 20402— Subscription price per year: $21.00 domestic and $26.25 foreign. Cost per single issue: $6.50 domestic and $8.15 foreign. Contents— Cowiinwed Notes BOROWSKY, BETTY. Laboratory studies of the pattern of reproduction of the isopod crustacean Idotea baltica 377 O'LEARY, JOHN, and DOUGLAS G. SMITH. Occurrence of the first freshwater migration of the gizzard shad, Dorosoma cepedianum, in the Connecticut River, Massachusetts 380 WYLLIE ECHEVERRIA, TINA. Relationship of otolith length to total length in rock- fishes from northern and central California 383 LE BOEUF, BURNEY J., JOHN E. McCOSKER, and JOHN HEWITT Crater wounds on northern elephant seals: the cookiecutter shark strikes again 387 Notices-NOAA Technical Reports NMFS published from March to December 1986. . . 393 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. VALIDATION OF THE OTOLITH INCREMENT AGING TECHNIQUE FOR STRIPED BASS, MORONE SAXATILIS, LARVAE REARED UNDER SUBOPTIMAL FEEDING CONDITIONS Cynthia Jonesi and Edward B. Brothers^ ABSTRACT Striped bass, Morone saxatilis. larvae were reared in the laboratory for 97 days to validate the otolith increment aging technique for this species. Otolith-increment deposition rates were determined under optimal laboratory conditions for growth and under three conditions of restricted feeding and using both light and scanning electron microscopy (SEM). Under optimal laboratory conditions, increments were deposited daily from the fourth day after hatching through the first 2 months of life and were discernible with the light microscope. For larvae reared under restricted feeding regimes and readings done with the light microscope, counts did not reflect true age. Counts obtained from these same otoliths using SEM, however, more closely reflected true daily age. Results indicate that the use of light microscopy alone can result in inaccurate estimation of age for larvae that have experienced starvation episodes. When otolith increments in larval fish are deposited daily, with a known time of onset, precise age of each individual can be determined and the growth curves for the individuals may be generated. The ability to follow changes in growth of individuals and populations on as fine a scale as, say, a week may provide a means to improved understanding of the effects which environmental factors have on sur- vival. To apply this aging technique to larval striped bass, Morone saxatilis, daily deposition of incre- ments and the age at first increment deposition had to be confirmed in the laboratory with known-age larvae. Although daily depositional rates of otolith increments in known-age larval striped bass have not been previously reported, daily deposition has been noted for larvae and juveniles of 17 other species of fish reared in the laboratory (see Jones 1985 for review). Nonvalidated data exist to support the concept of daily increment deposition for field- captured striped bass (Brothers et al. 1976). How- ever, tests of depositional rate under suboptimal lab- oratory conditions, using light microscopy, have shown that depositional rates can be affected by the specific growth rate (Geffen 1982), by photoperiod (Radtke 1978), by food supply (Geffen 1982; Neilson and Geen 1982), and by temperature (Brothers 1978; Geffen 1983). Campana and Neilson (1985) stated that "few workers have critically assessed the 'Old Dominion University, Department of Oceanography, Nor- folk, VA 23508. ^EFS Consultants, 3 Sunset West, Ithaca, NY 14850. assumptions upon which the age and growth infer- ences are based or considered the potential for envi- ronmental modification of microstructural features." Of particular importance is the potential for count- ing fewer otolith increments when otolith growth rate is slowed to the extent that increments being deposited are too narrow to resolve with a light microscope. Inadequate resolution with the light microscope could lead to systematically low incre- ment counts and thus, result in overestimation of the growth and mortality rates, and underestima- tion of variance in growth, all of which have impor- tant biological implications. Hence, to demonstrate that striped bass larvae from the field could be aged accurately by the otolith increment technique, we found it necessary to determine the regularity and readability of otolith-increment deposition under simulated laboratory suboptimal field conditions. Lack of scanning electron microscopy (SEM) validation hinders the resolution of an important issue: Is daily formation of increments a robust biological rhythm common to most teleosts which requires serious and prolonged starvation to disrupt, or is it a more volatile physiological connection in which daily formation occurs only under optimal food concentrations as certain laboratory studies indicate? Factors which affect growth and survival of striped bass larvae have been studied extensively (see Westin and Rogers 1978 for review). Rogers (1978) raised larval striped bass under various tem- perature and feeding regimes to determine growth under laboratory conditions. Larvae grew well at Manuscript accepted December 1986. FISHERY BULLETIN: VOL. 85, NO. 2, 1987. 171 FISHERY BULLETIN: VOL. 85, NO. 2 temperatures between 16° and 22° C and with a minimum of 1,000-2,000 Artemia nauplii/L. The optimum saUnity range was between 3.5 and 14.0"/oo (Bayless 1972). Davies (1973) studied larval survival under combinations of temperature, pH, and dissolved solids. Optimum temperature was 17.0°C; optimum pH, 7.5. Eldridge et al. (1981) studied the growth of larvae under various feeding regimes and found growth rates which approx- imated field growth rates at concentrations of 5,000 ArtemialL. They found that the "point of no return" was ill defined and starved larvae could live for as long as 31 days. Dey (1981) has reported on growth and survival of wild larvae, using length and developmental stage to estimate growth. He found growth was temperature dependent and tempera- tures between 12° and 15°C resulted in massive mortalities. The purpose of this study was to determine the relationship between age, environmental condition, and otolith increment depositional rates in labora- tory-raised striped bass larvae. This was accom- plished by studying the increments of known-age lar- vae reared under both optimal laboratory conditions and restricted feeding regimes (laboratory-simulated suboptimal field conditions). Larvae were subjected to various periods of food deprivation to determine the potential dependence of increment depositional rates on nutritional condition. Specifically, incre- mental counts made with light microscopy and SEM were compared to evaluate the reality of apparent interruptions of daily deposition. METHODS Striped bass eggs were obtained from the Ver- plank Hatchery, Verplank, NY, within 24 hours of fertilization. Eggs were held in water obtained at the hatchery (O'Vi,,, salinity) at 18°C, under a 14L: lOD photoperiod. Light levels were 25-31 n Ein- steins/m- per second. This light level is approx- imately equal to light at a depth of 2-3 m in a coastal stream or 1 m in a coastal estuary depending on tur- bidity and season of the year. Eggs hatched within 24 hours of fertilization. Newly hatched larvae were transferred to 4 L jars and stocked at densities of 50 per liter. Over the first 8 days, salinities were gradually raised to 5°/oo by adding filtered seawater with 0"/()(i water. Seventy-six days after hatching of the larvae, salinities were gradually raised to lO'Voo over a span of 8 days. Water was changed at least every other day. Four feeding conditions were established. The food for all conditions was newly hatched brine shrimp, Artemia. Larvae were fed ad libitum (con- dition 1), other larvae were starved throughout the experiment (condition 2), and other larvae were starved for the first 15 days after hatching, then fed ad libitum (condition 3). Condition 4 consisted of lar- vae that were intermittently deprived of food. These larvae were not fed between 39-43, 51-55, and 62-66 days after hatching, for a total of 15 days out of the 68 days they were reared. For the remaining time they were fed ad libitum. Larvae were sampled according to the schedule listed in Table 1. Larvae that were sampled were anesthetized with Tricaine methanesulfonate (Cre- sent Research Chemical) and sacrificed. Total length was measured to the nearest 0.1 mm. Oto- liths were teased from the otic capsules with fine dissecting needles, cleared of tissue, washed in de- ionized water and transferred with a micropipette or with fine dissecting needles to a labeled micro- scope slide. Small otoliths were mounted permanently in Ewparol without grinding. Larger otoliths were mounted in Flowtex (Lerner Laboratories), ground with 600 grit sandpaper and read with the light microscope. A subsample of ground otoliths was re- moved from the Flowtex, mounted in Spurr's me- dium, and ground to the core on Beuhler lapidary wheels. Initial grinding on the wheels began at 180 ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Table 1 .—Sample size at age of striped bass larvae reared under four feeding regimes. N.T. = not taken. Continuously -eeding regime condition Age 1 Starved Intermittent (d) fed Starved then fed starved 3 3 4 8 5 5 5 3 6 5 5 7 5 5 8 5 5 9 5 5 10 5 5 14 5 5 18 5 5 2 22 5 5 2 2 26 5 6 2 2 33 5 2 2 38 4 N.T. N.T. 40 2 3 2 47 5 2 2 54 5 2 2 58 1 N.T. N.T. 64 N.T. 1 N.T. 68 3 3 2 97 5 2 172 JONES and BROTHERS: AGING TECHNIQUE FOR STRIPED BASS LARVAE grit and final polishing was done with 0.25 /um dia- mond paste. These otoliths were etched with 0.02N HCl, then mounted on SEM stubs and sputter coated with gold/palladium. Three light microscopes were used: a Zeiss, a Leitz, and an Olympus. The latter two were equipped with video viewing systems and polarized light sources. Readings were done with brightfield illumination at 400, 540, and 1,000 power. Video in- creased magnification to a maximum of 2700 x . The maximum resolution for the light microscopes was 0.5-1.0 nm. The SEM employed was a JOEL (JSM 200) equipped with both secondary electron image (SEI) and backscattered electron image (BE I) collectors. For light microscopy, slides were chosen at ran- dom and read double blind (age of the larvae and condition were unknown). Readings were done three times for each slide. Each slide was counted only once during each session so that replicate counts did not immediately follow each other. Thirteen of the twenty-four samples from conditions 3 and 4 were used for SEM analysis. For SEM examinations. counts were blind (ages of the larvae were un- known); condition, however, was selected by the in- vestigators to check the accuracy of the light micro- scope counts for conditions 3 and 4. RESULTS Light Microscopy The relationship between the number of otolith increments and age, in days, for the four experi- mental conditions is shown in Figure 1. Fully fed larvae {n = 63), condition 1, had a regression slope of 0.98 increments/day, and the smallest standard error (Table 2). Its confidence interval included 1 increment/day. Beyond 68 days of age sagittae became very difficult to read. Continuous counting paths or an appropriate series of transects were dif- ficult to find because the sagitta changes shape and develops new centers of deposition around the periphery of the otolith. This resulted in underesti- mates of true age (Table 2) for larvae older than 2 months of age. INCREMENT COUNTS VERSUS KNOWN AGE 90- • FED FOR 97 DAYS / ' ° STARVED / : 80 - D STARVED FOR 15 DAYS, THEN FED y/ AINTERMITTENT STARVATION X CO 1- 2: =) o o 70 - 60 - y / z LU LU cr o 50 - i*0 - 30 - 20 - 10 - 0 - • o / • o 0 0 0 o 0 B • o □ o r 1 1 1 1 1 1 1 1 1 1 0 10 20 30 40 50 60 70 80 90 100 AGE (days) Figure 1.— Relationship between otolith increment count in larval striped bass and true age for four feeding regimes, light microscope observations. 173 FISHERY BULLETIN: VOL. 85, NO. 2 Table 2. — Parameters for weighted regressions of increment counts on days from hatch of striped bass larvae reared under four feeding regimes. SE indicates standard error of the estimate, C.I. indicates confidence interval, N.S. indicates slope not significantly different than 1.0, * indicates P = 0.05. N Slope 95% C.l. Low High Intercept r^ Age Condition (counts/d) SE (slope) Counts SE (int) P > 0.05 68 d and 1 Always fed younger All ages 1 Always fed 2 Starved 3 Starved/fed 4 Intermittent 60 63 43 12 12 0.980 0.946 0.469 0.930 0.873 0.0243 0.0169 0.0402 0.1005 0.0586 0.931 1.029 0.912 0.980 0.388 0.550 0.711 1.149 0.745 1.000 -4.016 0.4482 -3.627 0.4068 -1.697 0.4325 -10.430 4.3906 2.579 2.4010 0.96 0.96 0.77 0.90 0.96 N.S. * N.S. N.S. The slope of the regression Une for starved lar- vae (condition 2, n = 43), 0.469 increments/day, differed significantly from 1.0 increment/day. Incre- ments appeared regularly spaced. Otoliths of starved larvae did not appear aberrant under the light microscope. The regression of increment counts versus true daily age for larvae, which v^ere starved then fed (condition 3, n = 12), had a slope of 0.930 incre- ments/day with confidence intervals which included 1.000 increments/day (Table 2). However, the re- gression intercept was - 10.430, an overestimate of age at first increment deposition. This leads to a 6-d underestimate of true age because depositional rates were underestimated during the first 2 weeks of life. The slope of the regression line for intermittent- ly starved larvae (condition 4, n = 12) was 0.873 in- crements/day. The slope of 1.0 increment/day fell at the very edge of the confidence interval. If a slightly smaller alpha level had been chosen, deposi- tion would not have been assumed daily. Initial increment formation began at 4 days after hatching with a 95% confidence interval that ranged Table 3. — Counting bias for larvae starved for the first 15 days after hatch (calculated as estimated age - true age^ Underesti- mate of age is indicated by - ; overestimate indicated by + . SEI = secondary elec- tron image; BEI = backscattered electron image. Sample no. Age (d) Microscopic technique Light SEI BEI 1 2 3 4 5 6 7 68 47 54 68 33 47 47 -17 0 +1 - 7 -5 -5 0 -5 -6 - 24 - 5 - 5 - 7 +4 +5 - 8 -2 -2 -7-1 0 Mean bias SE -10 -2 -2 7.9 3.4 4.0 from 3 to 5 days. Yolk-sac absorption occurs at 7 days after hatching at 18°C and first feeding begins at approximately the same time. However, initial increment deposition does not appear to be con- nected to these events. Two or three weakly defined increments were observed within the core in many SEM preparations. They were not counted in light or SEM readings. Scanning Electron Microscopy Results from the SEM study are qualitative rather than quantitative due to the small sample sizes, n = 13, used for the SEM. With SEM, otohth incre- ment counts for condition 3, larvae which were starved then fed (Table 3), and for condition 4, lar- vae which were intermittently starved (Table 4), yielded more accurate counts than those obtained on the same specimens with light microscopy. With light microscopy counts from larvae which were starved for 15 days resulted in an underestimate of true age by 10 days (Table 3). The variability was also high (SE = 7.9 days). SEM counts underesti- mated true age by 2 days. Variability was small; the standard error was 3.4 and 4.0 days for SEI and Table 4. — Counting bias for larvae intermit- tently starved (calculated as estimated age - true age'). Underestimate of age is in- dicated by - ; overestimate indicated by + . SEI = secondary electron image; BEI = backscattered image. Sample Age Microscopic technique no. (d) Light SEI BEI 8 68 + 2 -5 -6 9 47 - 5 -5 -3 10 54 -11 -1 -2 11 54 - 4 -4 0 12 47 + 2 -4 -4 13 47 - 5 +1 + 1 Mean bias - 4 -3 -2 SE 4.9 2.4 2.6 ^Estimated age = number of Increments + mean age at first increment formation. 'Estimated age = number of increments + mean age at first increment formation. 174 JONES and BROTHERS: AGING TECHNIQUE FOR STRIPED BASS LARVAE BEI, respectively. Light microscope-counted incre- ments of intermittently starved larvae underesti- mated age by 4 days (Table 4). The standard error of the mean was 4.9 days. For this sample, both SEM techniques also gave more accurate age estimates. SEI underestimated age by 3 days (SE = 2.4 days), and BEI underestimated age by 2 days (SE = 2.6 days). The comparison between SEI and BEI showed no significant difference in accuracy (Tables 3, 4). Figures 2A and 2B illustrate the increment struc- ture observed with these two methods of SEM. BEI Figure 2.— Comparison of otoliths of larval striped bass using two scanning electron micro- scope (SEM), techniques. A) Normal secondary emission (SEI) photomicrograph of an otolith of a larva starved for the first 15 days posthatch then fed ad libitum until sacrificed. Incre- ment width during starvation is approximately 5 /.im. B) Backscattered emission (BEI) photo- micrograph for a comparable otolith to A above. Legends in the micrographs indicate 1) length of scale bars in p^m, 2) accelerating voltage KV, 3) mm working distance, 4) coded photo number. 175 FISHERY BULLETIN: VOL. 85, NO. 2 enhances contrast, but does not allow the specimen to be tilted. With SEI, tilting can increase increment relief and visibility. Increments deposited during starvation were only 0.5 ^m in width, too closely spaced to be discerned with the light microscope (Fig. 3). Additionally, the material that is deposited appears to be more homogenous in density, probably containing a lower amount of matrix. When etched, less material was dissolved in the area corresponding to starvation periods. This resulted in a higher area of relief, form- ing a broad ridgelike structure, subdivided into finer increments. The etching properties were, therefore, different compared to the same area in the otolith of a fed larvae. This ridgelike structure consistent- ly indicated periods of starvation during the first 2 weeks of life. Ridges were not apparent for older larvae starved for shorter time intervals (Fig. 4). DISCUSSION Estimation of age obtained using the light micro- scope was not always accurate. When larvae grew well, the light microscope gave correct age esti- mates. However, otoliths of larvae reared under sub- optimal feeding conditions gave underestimates of true age. Age estimates were more accurate using SEM, and starvation episodes were easier to recog- nize in the otoliths. Light microscopy has been routinely used to esti- mate age in field-captured larval fish (Jones 1985). Only a few investigators have employed SEM. Al- though SEM improves the accuracy of age esti- mates, it is more costly, requires more precise prep- aration, and is more time consuming. However, for larvae as resistant to starvation as striped bass, SEM verification of age estimation obtained with the light microscope is necessary. In our view, inves- tigators using increments to estimate growth should check their results from the light microscope with SEM studies. SEM analysis could be performed on a randomly chosen subsample of otoliths. If prob- lems were uncovered, a more extensive analysis using SEM could be undertaken. Checks on a ran- dom sample using SEM are particularly important for field studies where application of the otolith increment technique to estimate field growth is relatively new. It could be argued that larvae which undergo periods of starvation are more vulnerable to preda- tion and may occur only infrequently in samples. Although this is quite likely, it is precisely during Figure 3.— SEM photomicrograph of an otolith of a larva starved for the first 15 days of life then fed ad libitum. A ridge, indicated by the circle, develops as the result of etching the otolith with 0.02 N HCl. This ridge corresponds to the period of starvation. Increment width during starvation is approximately 5 \m\. Legends in the micrographs indicate 1) length of scale bars in ^^m, 2) accelerating voltage KV, 3) mm working distance, 4) coded photo number. 176 JONES and BROTHERS: AGING TECHNIQUE FOR STRIPED BASS LARVAE Figure 4.— SEM photomicrograph of an otolith of an intermittently starved larva. Note the pattern of narrow bands, indicated by the bracket, typical of starvation episodes. Increment width during starvation is approximately 5 ^ II Q 0.15 n 0.14 0.13 0.12 H 0.1 1 0.10 0.09 0.08 0.07 0.06 0.05 0.04 0.03 - 0.02 - 0.01 0.00 " PACIFIC COD. FIXED SUBSAMPLE ♦ PACIFIC COD, RANDOM SUBSAMPLE SABLEFISH. FIXED SUBSAMPLE SABLEFISH. RANDOM SUBSAMPLE POLLOCK, FIXED SUBSAMPLE POLLOCK. RANDOM SUBSAMPLE 10000 30000 50000 TOTAL COST C 70000 (MINUTES) 90000 Figure 1.— The relationship of D (= yVartot) and total cost for Pacific cod, sablefish, and walleye pollock, using fixed or random age subsample. returns is reached beyond this total cost and the curves become flatter for C greater than 10,000. These results indicate that setting a precision at D = 0.02, 0.025, and 0.03 respectively for Pacific cod, walleye pollock, and sablefish using random age sub- samples would represent a reasonable compromise between cost and precision. Increasing total cost beyond this level will show no more gains from the ALK. DISCUSSION It is obvious that the random subsampling scheme is superior to the fixed subsampling scheme. How- ever, it is more important to realize that there is a cap on total cost for ALK. This cap represents the most effective budget for ALK. Vartot of estimated age composition will not decrease significantly for a greater budget. For the three species, total cost of 10,000 minutes (about 70 working days) is the up- per limit. This indicates that approximately 2,000 length observations and 800 random age subsamples for sablefish, 2,500 and 1,200 for walleye pollock, and 3,000 and 600 for Pacific cod represent the best compromise between cost and precision of estimates (VVartot = 0.03, 0.025, and 0.02 for the three species respectively). Although it can be argued that minimizing Vartot may not be sufficient for optimum sampling design for all age classes, it is necessary to consider that some age classes are rare in the commercial catch and are therefore difficult to sample precisely. How- ever, these age classes do not generally represent significant contributions to biomass, and it therefore seems reasonable to concentrate on the major age classes. If these rare age classes are important to population dynamics, the optimum allocation can be addressed as a multiple minima. The objective func- tion can be rewritten as M{N,n) = Z w, Var(p,), for i = 1, 2, . . ., A where w, is weighting factor. A larger w, must be given to those age classes which are of interest, whereas the mathematical expressions of optimum allocation are the same as Equations (7) to (14), and subject to the same cost function (Equation (6)), ex- 182 LAI: ESTIMATING AGE COMPOSITION cept that a's and b's are weighted by Wj. In fact, minimizing Vartot is a special case of minimizing M{N,n) = Vartot for all w, = 1. Another argument may relate to the possibility that the cost function may be more complicated so that traveling and overhead costs can be taken into account. In such cases, cost function may become a nonlinear form, and the explicit expressions of n* and A/'* cannot be obtained. However, the technique of nonlinear programming can be applied to find numerical solutions of w* and A'*. In general, it is to minimize M{N,n) = Z. w, Var(pj) subject to C = c{N,n) where C is total cost and c{N,n) is cost function. Many optimization programs can be employed from popular computer software packages for main frame computers. Bunday (1984) provided several BASIC programs for constrained optimization, which may be useful in personal computers. It should be noted, however, that the sufficient and necessary condi- tions of this constrained minimum must be proved. Theoretically, there is a unique minimum if objec- tive function is convex and constraint function is concave (Bunday 1984). ACKNOWLEDGMENTS This study was funded by the Northwest and Alaska Fisheries Center, NMFS, Seattle, WA. I am deeply indebted to Donald R. Gunderson, Vincent F. Gallucci, and Loh-Lee Low for their guidance and Pat Sullivan for his comments. Many thanks to the referees for their helpful comments, and particularly for rewording a more accurate description of Vartot. LITERATURE CITED Bunday, B. D. 1984. Basic optimisation methods. Edwards Arnold, Lend., 128 p. Cochran, W. G. 1977. Sampling techniques. 3d ed. John Wiley & Sons, Inc., N.Y., 428 p. Kendall, M., and A. Stuart. 1977. Theadvancedtheoryofstatistics. Vol. 3. 3ded. Grif- fin & Co., Lond., 585 p. Kimura, D. K. 1977. Statistical assessment of age-length key. J. Fish. Res. Board Can. 31:317-324. 1983. Determing appropriate sampling levels for estimating age distributions for trawl landings. Wash. Dep. Fish., Tech. Rep. 80, 18 p. Kutkuhn, J. H. 1963. Estimating absolute age composition of California salmon landings. Calif. Fish Game, Fish. Bull. 120, 29 p. Lal H. L. 1985. Evaluation and validation of age determination of sable- fish, pollock, Pacific cod and yellowfin sole; optimum sam- pling design using age-length key; and implications of aging variability in pollock. Ph.D. Thesis, Univ. Washington, Seattle, 426 p. SCHWEIGERT, J. F., AND J. R. SiBERT. 1983. Optimizing survey design for determining age struc- ture of fish stocks: an example from British Columbia Pacific herring {Clupea harengus ■pallasi). Can. J. Fish. Aquat. Sci. 40:588-597. Southward, G. M. 1963. Sampling landings of halibut for age composition. Int. Pac. Halibut Comm. Sci. Rep. 58, 31 p. Tanaka, S. 1953. Precision of age-composition of fish estimated by double sampling method using the length for stratification. Bull. Jpn. Soc. Sci. Fish. 19:657-670. 183 FISHERY BULLETIN: VOL. 85, NO. 2 APPENDIX A. Derivation of Vartot For a fixed age subsampling, substituting rij = n/L into Equa- tion (2), the Var(p,) is Var(p,) = ^ L I J q,j (1 - Qi^ Ij {q,j - p,y n N (A.l) Applying Equation (3), Vartot = J. Var(p,) 1=1 A L I I 1 = 1 J=l L Ij q,j (1 - q,j) Ij iq,j - Pj)^ n N A L A L 1 I [L if g, (1 - g,)] I I [i, (9y - py)2] i=l j=l i=l i=l + n N n N which is Equation (4). For a random age subsampling, substituting n-^ = n Ij into Equa- tion (2) and applying Equation (3), the Vartot is A L Vartot = X X i=i j=i h 0. (B.l) h h h i i>j Therefore, iLAl)(ZBl)>ii:A,B,f. h h h For a fixed age subsample, let Ai = \fajn; A2 = \/a^; B^ = \fc^; and B., = \Jc^N. Applying Equation (B.l), the product of D'- and C is D'-C = \— + ~\ {c^n + CjN) > (\/a^2 + \fWhitney Laboratory, The University of Florida, Route 1 , Box 121, St. Augustine, FL 32086. ^Whitney Laboratory, The University of Florida, Route 1, Box 121, St. Augustine, FL 32086, and Department of Anatomy and Cell Biolog}', College of Medicine, University of Florida, Gainesville, FL 32610. [Direct all reprint requests to Robin A. Wallace at the Whitney Laboratory.] has been the subject of induced breeding in the laboratory (Kuo et al. 1973, 1974a, b; Kuo 1982). General information on the biology of M. cephalus and related species of mullet can be found in Ander- son (1958), Stenger (1959), and Thomson (1966). Striped mullet have one breeding cycle per year lasting from 2 to 5 months depending on the loca- tion (Jacot 1920; Breder 1940; Bromhall 1954; Anderson 1958; Arnold and Thompson 1958; Stenger 1959; Tang 1964; Zhitenev et al. 1974; Pien and Liao 1975; Timoshek and Shilenkova 1975; Finucane et al. 1978; Apekin and Vilenskaya 1979; Azoury and Eckstein 1980; Chubb et al. 1981; Dindo and MacGregor 1981). In coastal waters of the southeast United States, spawning has been reported to occur from October through February as determined from the time of appearance and size of larvae and fry (Anderson 1958; Arnold and Thompson 1958), from the presence of migrating mullet with "developing" ovaries (Breder 1940; Arnold and Thompson 1958; Stenger 1959), and by monthly gonadosomatic index (GSI) changes (Dindo and MacGregor 1981). In view of the extensive interest in the mullet, it Manuscript accepted February 1987. fishery BULLETIN: VOL. 85, NO. 2, 1987. 187 FISHERY BULLETIN: VOL. 85, NO. 2 is somewhat surprising that so little information con- cerning its reproduction, other than the seasonal spawning time, is available. For instance, little attention has been paid to the dynamics of oocyte development and ovarian recrudescence in natural striped mullet populations. A staging system for the ovary itself that would simply and meaningfully represent the extent of ovarian recrudescence in this species is lacking. And, although there have been numerous references to size of age at sexual matur- ity (Jacot 1920; Stenger 1959; Timoshek 1973; Ape- kin and Vilenskaya 1979), these are generally not based on comprehensive sampling. Similarly, al- though there have been equally numerous reports of striped mullet fecundity (Nash et al. 1973; Pien and Liao 1975; review by Alvarez-Lajonchere 1982), most derive from counts of eggs (or oocytes) in only a very limited number of fish, while body size-related data are virtually nonexistent. We recently initiated several studies dealing with the reproduction of M. cephalus in coastal waters of northeast Florida and related topics. The specific purposes of this investigation include describing the patterns of oocyte growth during seasonal ovarian recrudescence, developing an ovarian staging sys- tem based on these patterns, and providing defin- itive determinations of both the size at maturity and the size-specific fecundity of female striped mullet in the area. MATERIALS AND METHODS Fish Female striped mullet approximately 18 cm stan- dard length (SL) or larger were captured by cast net from the Matanzas River Inlet south of St. Augus- tine, FL. Collections were made periodically from October 1984 through January 1985, and again from August 1985 through January 1986, for a total sam- ple of 340 fish. Standard length and fork length (FL) were determined to the nearest 0.1 cm, and body weight (BW) to the nearest 0.1 g for all fish. Ovaries of most (248) fish >20 cm SL were removed and transferred to a buffered salt solution (FO: Wallace and Selman 1978), and any adhering non-ovarian tissue was removed. Whole ovaries were then patted dry and weighed to the nearest 0.01 g. The GSI for each fish was calculated as GSI = (ovary weight/ body weight) x 100. Oocyte Size-Frequency Profiles Oocyte size-frequency profiles were constructed 188 for each fish in the following manner. A represen- tative (see below) piece was gently teased free from the middle of each newly collected ovary, patted dry, weighed to the nearest 0.1 mg, and placed in a petri dish containing FO solution. Pieces weighed from 1 to 9 mg and contained from 100 to 500 oocytes >0.10 mm in diameter. Individual follicle-enclosed oocytes were manually measured to the nearest 0.02 mm with an optical micrometer mounted on a dis- secting microscope. Oocyte size-frequency profiles were not determined for fish with largest oocyte diameters (LODs) <0.10 mm, although their LODs were noted. Profiles and LODs derived from a sample of ovary can be considered representative because oocyte development is known to occur uniformly through- out the mullet ovary (Ochiai and Umeda 1969; She- hadeh et al. 1973; Timoshek and Shilenkova 1975). Oocyte Stages The oocyte size at which vitellogenesis (the period of protein yolk accumulation) begins in the striped mullet was determined by the appearance of specific yolk protein bands in oocyte homogenates subjected to polyacrylamide gel electrophoresis (see Greeley et al. 1986b). Groups of small oocytes with mean diameters of 0.14, 0.16, 0.18, and 0.20 mm were isolated from surrounding ovarian tissue, homog- enized in a sodium dodecyl sulfate (SDS) contain- ing buffer solution, and heated at 100°C for 5 minutes. Samples were loaded onto a 0.75 mm thick polyacrylamide gradient gel (gradient: 3.5-20.4%) and were electrophoresed in SDS buffer with a con- stant applied current of 30 mAmps until the bromophenol blue marker migrated from the gel. Protein fixation, visualization with Coomassie blue, and molecular weight determinations were con- ducted as in Wallace and Selman (1985). Biochem- icals and reagents were highest available grades from Sigma Chemical Company^ and Bio-Rad Laboratories. The minimum oocyte size competent to resume meiotic maturation (leading to the development of a mature fertilizable egg) in response to steroid hor- mone stimulation was determined in vitro by the methods of Greeley et al. (1986a). Larger follicle- enclosed oocytes were isolated in FO and assigned to one of four pools with mean diameters of 0.52, 0.56, 0.60, and 0.64 mm. Oocytes from each pool were randomly subdivided into treatment (steroid 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. GREELEY ET AL.: STRIPED MULLET OVARY added) and control groups, and then transferred into separate 35 x 10 mm petri dishes containing 3 mL of a 75% L-15 culture medium (Sigma) at pH 7.5. A solution containing 3 i^g of 17a-hydroxy-20/?- dihydroprogesterone in 10 /uL of 95% ethanol was added to each treatment dish (for a final steroid con- centration of 1 ^ig/mL); 10 ^L of 95% ethanol was added to each control dish. Oocytes were incubated at 20°C for 48 hours and then examined for germ- inal vesicle breakdown (GVBD), which was easily noted by the absence of the germinal vesicle (GV) or nucleus in the highly transparent postmaturation oocyte, as an indicator of the resumption of meiotic maturation in response to the hormone. Fecundity Potential fecundity was estimated for each fish whose oocyte size-frequency profile demonstrated that recruitment of oocytes into vitellogenesis had ceased as potential fecundity = [(number of re- cruited oocytes in ovary piece) x (weight of whole ovary) ]/(piece weight). RESULTS Gonadosomatic Index (GSI) Ovarian recrudescence in the striped mullet occurs during autumn and early winter in coastal waters of northeast Florida, as indicated by seasonal changes in the GSIs of females collected from the Matanzas River Inlet (hereafter referred as Inlet) during the 1984-85 and 1985-86 breeding seasons (Fig. 1). In August of 1985, the earliest month of collections for this study, GSIs were <1 in all sam- pled fish. Ovaries represented by these low GSIs were quite small and were either translucent and colorless (smallest and least developed ovaries) or opaque with a red hue resulting from extensive vascularization (slightly larger and more devel- oped ovaries). Similar low GSIs continued to be found in all females collected through mid- Septem- ber. In late September the GSIs of a portion of the col- lected females rose sharply, to nearly 9. The much larger ovaries represented by these high GSIs were X UJ Q O I- < o o Q < O 20 15- 10- o 1984-1985 • 1985-1986 1. o o o o ^ t •: " 0 0 o o 20 AUG 10 20 SEP 10 20 OCT 10 20 NOV -T^-V 10 20 DEC fV-J 0 20 JAN Figure 1.— Variation in the gonadosomatic index (GSI) of female Mugil cepkalus >20 cm SL during prespawning ovarian recrudescence along the northeast Florida coast. 189 FISHERY BULLETIN: VOL. 85, NO. 2 now yellow to golden in color because of a prepon- derance of yellow yolky oocytes in the ovarian lamellae. This initial phase of GSI increases con- tinued through early November, at which time two distinct groupings of females within the population became apparent: those with GSIs remaining <2, and those with GSIs ranging between 10 and 21. Average GSIs in the latter group rose only slightly thereafter, ranging from 13 to 22 by early Decem- ber. The lower GSIs of the rest of the population remained unchanged throughout the remainder of this study. During the 1984-85 season, females with high GSIs continued to be caught until mid-January, at which time larger females (>20 cm SL) apparently left the Inlet because none were caught during the 3 additional weeks of intensive collecting. In con- trast, during the 1985-86 season, females with high GSIs were only collected through mid-December. Although a few mid-size females were collected through late January of this year after the expen- diture of considerable collecting time and effort, all proved to have small GSIs. Oocyte Stages A primary purpose of this project was to document the oocyte growth and development that accom- panied these GSI changes during prespawning ovarian recrudescence. Two oocyte development stages of primary interest to us were 1) the stage at which yolk accumulation or vitellogenesis begins, and 2) the stage during, or following, vitellogenic growth when the oocyte first becomes competent to resume meiotic maturation and thereby develop into a fertilizable egg. We previously showed two large proteins at M^ = 104,000 and 90,000 to be the major yolk proteins present in mullet oocytes at the end of vitellogenic growth (Greeley et al. 1986b). As evidenced by the de novo appearance of these marker proteins (Fig. 2), vitellogenesis in the striped mullet begins in oocytes that are between 0.16 and 0.18 mm in diam- eter. In this study, oocytes 0.16 mm in diameter and smaller are thus considered to be previtellogenic, a broad classification including both primary growth- and yolk vesicle-stage oocytes as described 200- CO 1 O X 116.25 H 92.5 X o 66.2 UJ ^ < 45 -J 3 O LU 31 _J o 2 21.5 14.4- 104 90 t 0.14 0.16 0.18 0.20 OOCYTE DIAMETER (mm) Figure 2.— Polyacrylamide gel electrophoresis of oocyte proteins indicating de novo appearance of yolk proteins in vitellogenic oocytes (0.18 mm in diameter) of the Mugil cephalus, with comparison to smaller previtellogenic and larger vitellogenic oocytes (see text). Molecular weight standards are indicated on the left for comparison; the molecular weights of the two major yolk proteins are shown on the right (arrows). 190 GREELEY ET AL.: STRIPED MULLET OVARY by Wallace and Selman (1981); oocytes 0.18 mm in diameter and larger are considered to be actively vitellogenic. Late vitellogenic oocytes become competent to resume meiotic maturation and develop into a fer- tilizable egg in response to an in vitro steroid chal- lenge at 0.60 mm in diameter, with larger oocytes being only marginally more responsive (Table 1). Therefore, oocytes 0.60 mm and larger that do not exhibit any signs of maturation such as yolk "clear- ing" or hydration are considered to be in a pre- maturation stage of development. Striped mullet eggs following meiotic maturation are even larger (0.90 to 1.00 mm in diameter) (Abra- ham et al. 1966; Nash et al. 1973; Pien and Liao 1975; Finucane et al. 1978; Greeley et al. 1986b), nearly transparent (yolk has cleared), hydrated, float in full-strength seawater, and thus are easily distin- guished from the smaller and more opaque pre- maturation oocytes. Females with matui'e eggs were not collected during this study. Table 1 .—Percentage of different-sized striped mullet oocytes undergoing germinal vesicle breakdown (GVBD) in vitro in response to treat- ment withi 17a-fiydroxy-20/3-difiydroproges- terone (1 ^ig/mL) or an ethanol control. Oocyte diameter (mm) Treatment N (%)GVBD 0.52 control steroid 55 50 0 0 0.56 control steroid 50 45 0 0 0.60 control steroid 326 301 0 65 0.64 control steroid 196 198 0 71 Largest Oocyte Diameter (LOD) The relationship of oocyte diameter to stages of oocyte development is clearly delineated in Figure 3. Three major developmental stages are shown: 1) previtellogenic growth, including both the pri- 0.8 ,_ F E OJ OC lij 0 6 \- UJ 5 < 0.5 Q liJ l- 0.4 > O o o 0.3 1- cn UJ 0.2 or < 1 0.1 ° 1984-1985 • 1985-1986 oo • 4 • o i • o o o o I I o 8 i Mature or Maturing Oocytes Prematuration Oocytes Vitellogenic Oocytes • 1 h ^ « • ° ^ °° o . .1 JLa cf{ to • 0 o . 8 o ' . . • Previtellogenic Oocytes f 20 AUG I I ITT 1 10 20 10 20 10 20 SEP OCT NOV I 1 T ( 1 1 10 20 10 20 DEC JAN Figure 3.— Variation in the largest oocyte diameter (LOD) of female Mugit cephalus >20 cm SL during prespawning ovarian recrudescence along the northeast Florida coast. Presented on the right are oocyte developmental stages corresponding to these oocyte sizes. Broken line between vitellogenic and prematuration designations signifies uncertainty concerning whether the latter is actually a substage of the former. The category "previtellogenic" encompasses both primary growth and yolk vesicle oocytes. 191 FISHERY BULLETIN: VOL. 85, NO. 2 mary growth and yolk vesicle phases as defined by Wallace and Selman (1981), 2) vitellogenic growth, and 3) maturation. An additional developmental category— prematuration— may or may not be con- sidered a substage of late vitellogenesis; although oocyte growth appears to level off at this time, we have no direct evidence that the oocyte completely ceases to take up yolk precursors at this time. Changes in the LOD (Fig. 3) during the period of seasonal ovarian recrudescence were similar to changes in the GSI. LODs were small in August and early September, representative of still previtello- genic oocytes. LODs then rose sharply in mid- September as vitellogenic oocytes began to appear in females captured in the Inlet. The range of LODs in the population at this time varied considerably, from <0.01 mm in some fish to nearly 0.56 mm in others. LODs continued to rise in one portion of the population through mid-November, before leveling off between 0.60 and 0.72 mm; in other females, LODs remained below 0.18 mm through the end of each study period. Females with high LODs were last collected in mid-December during the 1985-86 breeding season and in mid-January during the 1984-85 season. Body Size at Maturity An examination of the LODs as a function of stan- dard length on a month to month basis (Fig. 4) reveals several body size-related trends that bear on the results presented in Figure 3. During August, the first month of the collections, there was little difference between the LODs of different-sized females, all being low and representative of pre- vitellogenic oocytes. Only a few large striped mullet were collected during this month. This changed in September, with larger females over 32 cm SL becoming more prevalent in the Inlet and exhibit- ing a tendency to have higher LODs than smaller females. By October, smaller females (to a lower limit of 28 cm SL) also began to acquire vitellogenic oocytes, although their average LODs were still lower than those of larger females. During Novem- ber, LODs of larger females leveled off at 0.60 to 0.72 mm, but a few smaller females (now to a lower limit of 26 cm SL) continued to have intermediate LODs indicative of oocytes in the early stages of vitellogenesis. By December and January, recruit- ment of smaller females into sexual maturity ap- parently ceased because LODs were now uniform- either greater than the 0.60 mm prematurational oocyte size or less than the minimum 0.18 mm vitel- logenic size— in all females regardless of body size. E E a: UJ I- UJ < Q UJ I- > o o O (/) UJ o q: < 0.6- AUG ^ 0.4- 0.2- D- :r. • • • -==r 0.6- 0.4- SEP • • • • v.. 1 • 0.2- n - • •• . • X Q O m 2000i 500- ^ 1000 500 0 BW = 0.0I3(SL)^'26 r2=0.97 —^^ — I 1 1 1 — — I — 0 20 25 30 35 40 STANDARD LENGTH (cm) 45 50 E o X I- o z LU o 50 45- 40- 35- 30 25-1 20 0 B K FL=l.ll2(SL) + 0.950 r2 = 0.99 ^/^ — I — 0 20 25 30 35 40 45 STANDARD LENGTH (cm) — I 50 Figure 5.— Relationship of (A) body weight and (B) fork length to the standard length of female Mugil cephaliis (>20 cm SL) dur- ing prespawning ovarian recrudescence along the northeast Florida coast. Lines were fitted by least squares regression. 193 FISHERY BULLETIN: VOL. 85, NO. 2 nor does it provide information about the range of oocyte sizes to be found in individuals. To better understand the pattern of oocyte development lead- ing to the formation of a clutch of mature eggs, it is necessary to examine comprehensive oocyte size- frequency profiles. Representative profiles of Figure 6 illustrate both the pattern of oocyte development during ovarian recrudescence in the striped mullet and an ovarian staging system based upon these profiles. In ovarian stage I (previtellogenesis), yolk accu- mulation by developing oocytes has not yet begun, as all oocytes are less than the minimum previtello- genic size of 0.18 mm in diameter. An oocyte size- Entry into Vitellogenesis Minimunn Prennaturation Size 40 20 0 o o 20 H— o 5S 0 > 0 20 Z UJ -1 0 0 UJ q: 40 u_ Figure 6.— Representative oocyte size-frequency profiles corresponding to stages of ovarian recru- descence in Mugil rephalus. Dashed line on left represents the transition between previtellogenic and vitellogenic oocytes; dashed line on the right delineates the minimum prematuration oocyte size. See text for explanation of stages. I I ' I ' I I ' ' ' I I ' ' ' I I ' ' ' I I ' ' ' 0.2 0.3 0.4 0.5 0.6 0.7 OOCYTE DIAMETER (mm) 194 GREELEY ET AL.: STRIPED MULLET OVARY frequency profile from an ovary in this stage is shown in Figure 6. GSIs of females with stage I ovaries ranged up to 0.8. Stage II (early vitellogenesis) is characterized by recruitment of previtellogenic oocytes into vitello- genesis. A single clutch of vitellogenic oocytes starts to form, becoming distinct from the remaining mass of previtellogenic oocytes as the recruited oocytes increase in diameter due to yolk accumulation. This stage can be further divided into two substages on the basis of the appearance of the profiles. In stage Ila, recruitment into vitellogenesis has just begun; a clear separation between the developing (vitello- genic) and nondeveloping (previtellogenic) oocytes is not yet discernible. In stage lib, the developing clutch forms a distinct peak (or peaks) as recruit- ment and subsequent vitellogenic growth continues. LODs of females with stage II ranged from 0.18 to 0.56 mm; GSIs varied from 0.3 to 8.5. In stage III (mid- to late-vitellogenesis), recruit- ment into vitellogenesis ceases, although oocytes in the recruited clutch continue to increase in diam- eter due to further yolk accumulation. This stage can also be divided into two substages. In stage Ilia, recruitment has just ended: the recruited clutch is spread out in size, and multiple size-frequency peaks are apparent. In stage Illb, the recruited clutch tightens into a single peak; oocyte diameters con- tinue to increase. LODs of females with stage III ovaries varied from 0.40 to 0.59 mm; GSIs ranged from 2.1 to 12.1. In stage IV (prespawning), oocytes in the re- cruited clutch reach the minimum prematuration size of 0.60 mm in diameter and become capable of resuming maturation in response to a hormonal signal. Late in the stage, as shown in Figure 6, all oocytes in an ovary will be at, or above, the mini- mum prematuration size. LODs of females with stage IV ovaries were 0.60 and 0.72 mm; GSIs were more scattered, from a low of 11.4 to a high of 21.2. Stage V (spawning: not shown) is characterized by the presence in the ovary of maturing oocytes or mature eggs. No fish with ovaries in this stage were caught in the Inlet; however, this stage was produced in the laboratory by injection of human chorionic gonadotropin into prespawning females (Greeley et al. 1986b). Stage VI (postspawn) ovaries, upon gross exam- ination, are small, red, and flaccid in appearance with a thickened ovarian wall. Spawning is ap- parently complete— at least by the time the females return to the Inlet— as no partially spawned ovaries were collected, nor were large atretic oocytes or eggs ever observed. LODs of postspawn females were 0.12 to 0.14 mm; GSIs ranged from 0.1 to 1.6. Monthly changes in the relative frequency of these stages in females collected from Matanzas Inlet dur- ing ovarian recrudescence in 1985-86 are shown in Figure 7. The variation in these stages is similar to the variation observed in the LOD, although the ovarian stages provide a somewhat different infor- o o o >- o LU O LU q: 100 50H 0 100- 50- 0 lOOi ^ 50- 0 100 50- 0 lOO-i 50 0 00 50H 0 I m \ \ ■ i T r i Y7^ -\ r ^ P R^ F?^ i ^ i?^ r^ ^ ^ , ^ I r^ r- Jan n = 6 Aug n = l9 Sep n = 52 Oct n = 36 Nov n = 22 _Ep. Dec n = 22 I 1 1 1 r I n n EI 2 3Z[ OVARY STAGE (1985-1986) Figure 7.— Monthly variation in ovarian stages of adult Mugil cephalus (>20 cm SL) during prespawning ovarian recrudescence along the northeast Florida coast. Data pre- sented for the 1985-86 season only (see text). 195 FISHERY BULLETIN: VOL. 85, NO. 2 mation. In August, only fish with previtellogenic ovaries (stage I) were collected. During September and October, a variety of ovary stages were ob- served, from previtellogenic (stage I) to prespawn- ing (stage IV). Ovaries with active recruitment of oocytes into vitellogenesis (stages Ila and b) were not observed after October. The first postspawn fish (stage VI) was caught in late November, and by January only previtellogenic or postspawn fish were caught. Variation in ovary stages during the 1984-85 season (not shown) was similar, except that in this year females with prespawning (stage IV) ovaries were collected as late as mid-January. Fecundity Because oocyte size-frequency profiles indicate that only a single clutch of developing oocytes pro- ceeds through vitellogenesis in a season, and that this single clutch is eventually spawned in its entire- ty, it is possible to calculate the individual fecundity of a female striped mullet by counting the number of vitellogenic oocytes in stages Illa-IV ovaries (in which recruitment of oocytes into developing clutches has ceased). The annual potential fecundity, or number of eggs available to be spawned in a single breeding season, was thus found to be linearly related to body weight and geometrically related to standard length (Fig. 8). The lowest fecundity observed was 0.25 x 10^ eggs in a fish 264 g BW and 23.5 cm SL, and the highest fecundity was 2.2 X 10^ eggs in a fish 1,627 g BW and 44 cm SL. DISCUSSION The present results indirectly confirm that the spawning season of M. cephalus in coastal waters of northeast Florida extends from at least late November (when the first postspawn female was collected during the 1985-86 season) through mid- January (when the last prespawn female was col- lected during the 1984-85 season). However, there is probably a certain amount of year-to-year vari- ation within this range: the first postspawn female was not observed until December during one season (1984-85), while the last prespawn in another (1985-86) was collected in December rather than January. It is also probable that these dates are in reality only a conservative estimate of the actual range of the striped mullet spawning season in this area. Available evidence from other studies strongly sug- gest that striped mullet spawn offshore (Anderson 1958; Arnold and Thompson 1958; Finucane et al. 1978). If this is also true of striped mullet in north- east Florida, then postspawn females collected in November may have traveled extensively between spawning at offshore sites and their eventual cap- ture in the Inlet. Likewise, the prespawn females collected in January would have required some time to reach offshore spawning sites after leaving the Inlet. Therefore, adding a month to each end of the observed range to conservatively account for such migrations may be appropriate and would make our dates consistent with previously published reports of spawning times for striped mullet in the south- east United States ranging from October through February (Broadhead 1956; Anderson 1958; Dindo and MacGregor 1981). Do striped mullet of northeast Florida actually spawn offshore? The best evidence for offshore spawning migrations in this study was our failure to collect from the Inlet any females with spawn- ing ovaries or even partially spent ovaries, suggest- ing that spawning probably occurred some distance from the Inlet. Further evidence for an offshore spawning site were the abrupt disappearances from the Inlet of fish with prespawning ovaries during both years of the study (mid- January 1984-85 and mid-December 1985-86), as this behavior suggested that mass spawning migrations to offshore waters occurred at these times. If these disappearances did represent mass off- shore spawning migrations, then how can we explain the earlier appearances in the Inlet of a few post- spawn fish? Perhaps there are actually multiple spawning migrations, possibly by different popula- tions of striped mullet moving through the Inlet at intervals throughout the period. Or perhaps some inshore spawning also occurs: staff at the Whitney Laboratory* have occasionally observed what they considered to be striped mullet spawning activity in the Intracoastal Waterway near the Inlet, and there are a few anecdotal accounts of inshore spawn- ing in the hterature (Breder 1940; Gunter 1945; Timoshek and Shilenkova 1975). However, if inshore spawning does occur, it must be limited in scope; otherwise, we would have collected females with spawning ovaries in the Inlet during our own studies. It may be that striped mullet can spawn either inshore or offshore, with offshore spawning favored, depending on factors— such as salinity, temperature, winds, currents, tides, or some com- bination thereof— which vary from locale to locale and from year to year. ^W. Raulerson, Whitney Laboratory, University of Florida, Route 1, Box 121, St. Augustine, FL 32086, pers. commun. 196 GREELEY ET AL.: STRIPED MULLET OVARY IS) O Q O liJ UJ O 0. A 2.0- PF=I,025{BW) + r2 = 0.8l 62,309 ^ 1.5- ^^ • 1.0- 0.5- 0- T 1 1 r 1 1 1 1 2.0 .5- .0 0.5 0 500 1000 1500 BODY WEIGHT (g) B PF = 25.84 (SD^-^^ r2 = 0.83 25 30 35 40 STANDARD LENGTH (cm) Figure 8.— The relationship of the potential annual fecundity of Mugil cephalus from northeast Florida to (A) body weight and (B) standard length. Lines are drawn from regression equations. In fact, the apparent timing of the hypothetical final spawning migrations from the Inlet in each of the two seasons of the present study suggests there might be a tidal involvement in these events: one coincided with a set of new moon spring tides, and the other to a set of full moon spring tides. Such a tidal or lunar connection to spawning migrations of striped mullet has been proposed previously (Bromhall 1954), although supporting evidence is still inconclusive. Others have alternatively sug- gested wind and currents might be contributing fac- tors to the onset of spawning migrations (Apekin and Vilenskaya 1979); further work is needed to clarify these issues. Most workers agree that individual female M. cephalus spawn only once a year (Zhitenev et al. 1974; Timoshek and Shilenkova 1975; Chubb et al. 1981). Our results support this assumption, as we never observed more than a single clutch of devel- oping oocytes proceeding through vitellogenesis dur- 197 FISHERY BULLETIN: VOL. 85, NO. 2 ing the fall period of prespawning ovarian recrudes- cence. Our failure to collect any partially spawned females is also consistent with a single seasonal spawn. In this study, both the GSI and the LOD proved to be adequate, although not completely satisfac- tory, indicators of the reproductive condition of female striped mullet. However, of these two indices the LOD would appear to be preferable. Determina- tion of the LOD requires only the biopsy of a small piece of ovary (see Shehadeh et al. 1973) which can be easily accomplished without harm to the fish, while determination of the GSI requires the sacrifice of the fish. Furthermore, the validity of the GSI has been questioned (deVlaming et al. 1982) as to its accuracy in correcting for body size in a consistent manner over all reproductive stages. On the other hand, the speed and ease of obtain- ing the LOD are its only advantages over a more comprehensive indicator of reproductive condition —the oocyte size-frequency profile. Such a profile also requires only the biopsy of a small piece of ovary and is a much more accurate indicator of ovarian stage, especially during the active vitellogenic growth of the ovary when a developing clutch may be quite spread-out in size. An adequate understanding of the functional rela- tionship between oocyte size and stage is, of course, required for correct interpretation of either LODs or oocyte size-frequency profiles. Of particular in- terest to us during this study were the sizes of the oocyte at 1) the beginning of vitellogenesis and 2) the prematuration stage of development. Our data indicate that oocytes of the striped mullet are able to grow to a point immediately prior to vitellogenic growth, then temporarily arrest at that stage. In contrast, once vitellogenic growth begins, further development leading to a subsequent clutch of mature eggs is apparently ensured. Thus knowledge of this transition point is extremely important to in- vestigators attempting to predict the future repro- ductive status of these fish. Likewise, clearly iden- tifying the prematuration stage of oocyte is impor- tant because at this stage the oocyte is competent to resume meiotic maturation culminating in the for- mation of a fertilizable egg. We define the beginning of vitellogenesis in the striped mullet oocyte by the initial appearance of yolk proteins detectable by electrophoretic tech- niques and the prematuration stage of development by in vitro culture techniques. Our resulting pre- maturation stage is in essential agreement with the "functional maturity" stage of Kuo et al. (1974b), as is our 0.18 mm initial vitellogenic stage with the initial "yolk globule" stage (0.20 mm) of these authors. We did not examine smaller oocytes in detail and thus did not attempt to establish specific stages for previtellogenic oocytes. The oocyte size-frequency profiles of this study, plus the additional data relating oocyte sizes and stages, demonstrate that M. cephalus has a group- synchronous type of oocyte development, as ori- ginally defined by Marza (1938) and reiterated by Wallace and Selman (1981). In such an ovary, a single developing clutch of oocytes coexists with an apparently asynchronous pool of previtellogenic oocytes. However, it must be pointed out that in the mullet this pattern is not always straightforward. The movement of oocytes into vitellogenesis is quite prolonged in time, so that a truly discrete clutch, as characterized by the cessation of additional recruitment into the clutch, does not become ap- parent until the more developed oocytes within the clutch are already well into vitellogenic growth. Fur- thermore, even after recruitment into vitellogenesis ceases, the developing clutch can be quite hetero- genous in size (see oocyte size-frequency profile for ovarian stage Ilia, Figure 6) and thus may not ap- pear to be undergoing synchronous growth until ovarian stage Illb. Such a pattern of oocyte devel- opment could be very difficult to characterize with- out examination of ovaries from females collected throughout the period of ovarian recrudescence. To the best of our knowledge, the ovary staging system put forth in this paper is the only such com- prehensive system developed specifically for the striped mullet. Based on well-defined and physio- logically significant criteria, it is presented with the purpose of standardizing future studies dealing with reproduction in the striped mullet. The range (23 to 27 cm SL) we present for the size at maturity of striped mullet in northeast Florida is similar to the only other published report for the area (25 cm SL by Stenger 1959), but much lower than the values (30 to 46 cm SL) reported by Jacot (1920), Thomson (1951), Ochiai and Umeda (1969), Timoshek (1973), and Apekin and Vilenskaya (1979). Most investigators agree that female mullet reach sexual maturity at the end of their third year (Thomson 1951; Timoshek 1973). We made no at- tempt to age the fish because methods available were not always agreed upon (see Thomson 1966). However, according to the growth schedules of Thomson (1966), adapted from various primary sources, the age at maturity in this study apparently ranged from 2V4 years (23 cm SL) to 2V2 years (27 cm SL). It thus appears that female M. cephalus attain sex- 198 GREELEY ET AL.: STRIPED MULLET OVARY ual maturity in northeast Florida at a smaller size, and probably at an earlier age, than anywhere else in the world. However, caution must be taken in making such comparisons. Our results demonstrate that smaller fish typically lag behind larger fish in reaching a sexually developed state by as much as 2 months on a seasonal basis. In consequence, the determinations of size or age at maturity by other investigators, if based on only a limited number of samples or on samples obtained at only a limited number of dates within the prespawning period of ovarian recrudescence, might not be truly compar- able to our results. The annual fecundity of M. cephalus has been reported to be from 1.2 x 10^ to 2.8 x 10^ by most authors (Thomson 1966), although estimates ranged from as low as 0.5 x 10^ to as high as 14 x 10^ (from review by Alvarez-Lajonchere 1982). Unfor- tunately, the methods whereby these values were obtained are often not given, so many reports have to be considered suspect. In addition, data concern- ing size-related trends in the fecundity of striped mullet are nearly nonexistent. By contrast, our esti- mates of fecundity (0.25 x 10*^ to 2.5 x 10*^) are well-documented and demonstrate a clear and highly predictable relationship between individual fecun- dity and body size. In conclusion, this study describes the pattern of oocyte development during seasonal ovarian recru- descence in the striped mullet, proposes an ovarian staging system based on oocyte stages and size- frequency profiles, gives a range of values for the female size at maturity, and presents the only com- prehensive examination of size-related fecundity for M. cephalus to date. ACKNOWLEDGMENTS This research was supported in part by National Science Foundation Grant DCB-8511260 to Robin A. Wallace. Special thanks go to Lynn Milstead for her assistance in preparing the figures. LITERATURE CITED Abraham, M., N. Blanc, and A. Yashouv. 1966. Oogenesis in five species of grey mullets (Teleostei, Mugilidae) from natural and landlocked habitats. Isr. J. Zool. 15:155-172. Alvarez-Lajonchere, L. 1982. The fecundity of mullet (Pisces, Mugilidae) from Cuban waters. J. Fish Biol. 21:607-613. Anderson, W. W. 1958. Larval development, growth, and spawning of striped mullet {Mugil cephalus) along the south Atlantic coast of the United States. U.S. Fish. Wildl. Serv., Fish Bull. 58:501- 519. APEKIN, V. S., AND N. I. ViLENSKAYA. 1979. A description of the sexual cycle and the state of the gonads during the spawning migration of the striped mullet, Mugil cephalus. J. Ichthyol. 18:446-456. Arnold, E. L., and J. R. Thompson. 1958. Offshore spawning of the striped mullet, Mugil cephalus, in the Gulf of Mexico. Copeia 1958:130-132. AzouRY, R., and B. Eckstein. 1980. Steroid production in the ovary of the gray mullet Mugil cephalus during stages of egg ripening. Gen. Comp. Endocrinol. 42:244-250. Breder, C. M. 1940. The spawning of Mugil cephalus on the Florida west coast. Copeia 1940:138-139. Broadhead, G. C. 1956. Growth of the black mullet Mugil cephalus in west and northwest Florida. Fla. Board Conserv. Mar. Lab. Tech. Ser. 25:1-29. Bromhall, J. D. 1954. A note on the reproduction of the grey mullet, Mugil cephalus Linnaeus. Hong Kong Univ. Fish. J. 1:19-34. Chubb, C. F., I. C. Potter, C. J. Grant, R. C. J. Lenanton, AND J. Wallace. 1981. Age, structure, growth rates and movements of sea mullet, Mugil cephalus L., and yellow eye mullet, Aldrich- ettaforsteri (Valenciennes), in the Swan-Avon river system, western Australia. Aust. J. Mar. Freshw. Res. 32:605-628. Comp, G. S., and W. Seaman, Jr. 1985. Estuarine habitat and fishery resources of Florida. In S. Seaman, Jr. (editor), Florida aquatic habitat and fish- ery resources, p. 337-435. Florida Chapter of American Fisheries Society, Kissimmee, FL. DeVlaming, v., G. Grossman, and F. Chapman. 1982. On the use of the gonosomatic index. Comp. Biochem. Physiol. 73A:31-39. DiNDO, J. J., and R. MacGregor III. 1981. Annual cycle of serum gonadal steroids and serum lipids in striped mullet. Trans. Am. Fish. Soc. 110:403-409. FAO. 1984. 1983 yearbood of fisheries statistics. Catches and landings. Vol. 56. Rome, Italy. Finucane, J. H., L. A. Collins, and L. E. Barger. 1978. Spawning of the striped mullet, Mugil cephalus, in the northwestern Gulf of Mexico. Northeast Gulf Sci. 2:148- 150. Greeley, M. S., Jr., D. R. Calder, M. H. Taylor, H. Hols, and R. a. Wallace. 1986a. Oocyte maturation in the mummichog (Fundulus heteroclitus): Effects of steroids on germinal vesicle break- down of intact follicles in vitro. Gen. Comp. Endocrinol. 62:281-289. Greeley, M. S., Jr., D. R. Calder, and R. A. Wallace. 1986b. Changes in teleost yolk proteins during oocyte maturation: Correlation of yolk proteolysis with oocyte hydration. Comp. Biochem. Physiol. 84B:l-9. Gunter, G. 1945. Studies on marine fishes of Texas. Publ. Inst. Mar. Sci. Univ. Texas 1:1-194. Jacot, a. p. 1920. Age, growth, and scale characters of the mullets: Mugil cephalus and Mugil curema. Trans. Am. Microsc. Soc. 39: 199-299. Kuo, C.-M. 1982. Induced breeding of grey mullet, Mugil cephalus L. 199 FISHERY BULLETIN: VOL. 85, NO. 2 In C. J. J. Richter and H. J. Th. Goos (editors), Reproduc- tive physiology of fish, p. 181-184. Pudoc, Wageningen. Kuo, C.-M., Z. H. Shehadeh, and C. E. Nash. 1973. Induced spawning of captive grey mullet (Mugil cephalus L.) females by injection of human chorionic gonado- tropin (HCG). Aquaculture 1:429-432. Kuo, C.-M., C. E. Nash, and Z. H. Shehadeh. 1974a. A procedural guide to induce spawning in grey mullet {Mugil cephalv^ L.). Aquaculture 3:1-14. 1974b. The effects of temperature and photoperiod on ovarian development in captive grey mullet {Mugil cephalus L.). Aquaculture 3:25-43. Marza, V. D. 1938. Histophysiologie de la ovogenese. Hermann, Paris, 81 p. Nash, C. E., C.-M. Kuo, and S. C. McConnell. 1973. Operational procedures for rearing. In The grey mullet: Induced breeding and larval rearing 1972-1973, Vol. II, p. 23-34. Oceanic Inst. Tech. Rep. 01-73-128. OCHIAI, A., AND S. UMEDA. 1969. Spawning aspects of the grey mullet, Mugil cephalus L., living on the coastal region of Kochi Prefecture. [In Jpn, Engl. summ.]. Jpn. J. Ichthyol. 16:50-54. PlEN, P.-C, AND I.-C. LlAO. 1975. Preliminary report of histological studies on the grey mullet gonad related to hormone treatment. Aquaculture 5:31-39. Shehadeh, Z. H., C.-M. Kuo, and K. K. Milisen. 1973. Validation of an in vivo method for monitoring ovarian development in the grey mullet {Mugil cephalus L.). J. Fish Biol. 5:489-496. Stenger, a. H. 1959. A study of the structure and development of certain reproductive tissues of Mugil cephalus Linnaeus. Zoologica, N.Y. 44:53-70. Tang, Y. A. 1964. Induced spawning of striped mullet by hormone injec- tion. [In Jpn., Engl, summ.] Jpn. J. Ichthyol. 12:23-28. Thomson, J. M. 1951. Growth and habits of the sea mullet, Mugil dobula Gun- ther, in western Australia. Aust. J. Mar. Freshw. Res. 2: 193-225. 1966. The grey mullets. Oceanogr. Mar. Biol. Ann. Rev. 4: 301-335. TiMOSHEK, N. G. 1973. The distribution and migration of mullet in the Black Sea. [In Russ., Engl, summ.] Tr. Vses. Naucho-Issled. Inst. Morsk. Rybn. Khoz. 93:163-177. TiMOSHEK, N. G., AND A. K. Shilenkova. 1975. The nature of the oogenesis and spawning of Black Sea mullet. J. Ichthyol. 14:727-746. Wallace, R. A., and K. Selman. 1978. Oogenesis in Fundulus heteroclitus I. Preliminary observations on oocyte maturation in vivo and in vitro. Dev. Biol. 62:354-369. 1981 . Cellular and dynamic aspects of oocyte growth in tele- osts. Am. Zool. 21:325-343. 1985. Major protein changes during vitellogenesis and maturation of Fundulus oocytes. Dev. Biol. 110:492- 498. Zhitenev, a. N., D. S. Kalinin, and Y. I. Abeyev. 1974. The state of the gonads of the striped mullet {Mugil cephalus) and the sharpnose mullet {Mugil saliens) leaving estuaries to spawn, and their reaction to a pituitary injec- tion. J. Ichthyol. 14:232-239. 200 DEVELOPMENT OF THE EGGS AND LARVAE OF THE YELLOWCHIN SCULPIN, ICELINUS QUADRISERIATUS (PISCES: COTTIDAE) Richard F. Feeneyi ABSTRACT The development of the eggs and larvae of Icelinus quadriseriatus is described from laboratory-reared and field-collected specimens. The eggs have diameters from 1.08 to 1.17 mm, an adhesive chorion, and multiple oil globules. Before hatching the oil globules coalesce into one 0.14-0.19 mm in diameter. The embryo develops a patch of tubercles on the dorsal surface of the head that are lost immediately after hatching. The larvae hatch at 2.6-3.4 mm. Distinguishing characters are 1-6 rows of ventral gut melanophores, 25-63 postanal ventral melanophores, and lower jaw angle pigment. Larvae over 3.9 mm may develop chin and pectoral insertion melanophores. Nasal and parietal spines appear at 9.3 mm. Postflexion lar- vae develop three patches of pigment dorsolaterally on the body by 10.5 mm and transform to juveniles by 16.3 mm. Scorpaeniform fishes are represented in the North Pacific Ocean by a large group of endemic taxa whose early life histories are poorly known. The early life histories of many sculpins (Cottidae), the second largest family in the order, were recently described (Richardson and Washington 1980). Lar- vae of several species ofArtedius, Clinocottus, and Oligocottus have been described by Washington (1986). Synchirus gilli larvae were described by Marliave et al. (1985). Reared and field-collected lar- vae of Chitonotus pugetensis have been described (Misitano 1980; Richardson and Washington 1980); however, the larval stages of the closely related Icelinus, including nine described species and one undescribed species (Yabe et al. 1980; R. Rosen- blatt^), are unknown. The purpose of this paper is to describe the eggs and larvae of the yellowchin sculpin, Icelinus quadriseriatus, using both labora- tory-reared and field-collected material. Icelinus quadriseriatus occurs along the coast of California north to Sonoma County (lat. 38°23.5'N; long. 123°08'W) and south to Cabo San Lucas, Baja California (Miller and Lea 1972; Eschmeyer et al. 1983). Adults are usually collected at depths from 18 to 90 m (based on Natural History Museum of Los Angeles County (LACM) and California Acad- emy of Sciences (CAS) adult collection data); occa- 'Section of Ichthyology, Natural History Museum of Los Angeles County, 900 Exposition Boulevard, Los Angeles, CA 90007. 2R. Rosenblatt, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093, pers. commun. winter 1983. sionally they range beyond these limits, being found in the intertidal zone and as deep as 201 m (Love and Lee 1974). The period of peak occurrence of prespawning females for /. quadriseriatus ranges from January to April, but mature oocytes have been found in females in every month except Octo- ber (in one year of a 2-yr study) indicating an almost year-round spawning capability (Goldberg 1980). MATERIALS AND METHODS Adult /. quadriseriatus were collected by otter trawl off Santa Monica, CA, on 8 February 1981, 3 July 1981, and 11 March 1982 and off of Hunt- ington Beach, CA, on 19 March 1981. The females were separated from the males (easily recognized by their darkly pigmented anal fin and gill mem- branes) and their eggs stripped into petri dishes filled with seawater. Sperm stripped from the males was then added to selected clutches of eggs while other clutches were left unfertilized. One clutch from the spawn of July 1981 and two clutches from the spawn of March 1981 were split in half and only one- half was fertilized. The eggs were incubated in natural seawater in a refrigerated 227 L tank with undergravel filter and within a temperature range of 13°-16°C (the March 1981 spawn was kept at 14° -16°C). The egg clutches were kept separate in plastic containers each with its own airstone. For the last spawning, March 1982, the undergravel filter was not used and the gravel removed completely in an attempt to cut Manuscript accepted December 1986. FISHERY BULLETIN: VOL. 85, NO. 2, 1987. 201 FISHERY BULLETIN: VOL. 85, NO. 2 down on bacterial and nematode infestation. The seawater was initially UV sterilized and filtered before use in the tank. Ten percent of the water was replaced every week. The developing eggs of the first three spawnings were sampled every 2 hours for the first 12 hours and then only once a day. The eggs of the last spawning, March 1982, were sampled at 16, 38, 63, 89, and 110 hours and 6, 8, and 10 days. Living eggs were illustrated using a Wild M5^ dissect- ing microscope with camera lucida. Measure- ments of egg diameter, perivitelline space, and oil globule diameter were taken with an ocular microm- eter. After hatching the larvae were put into 10 L plastic buckets and later transferred to the main culture tank. Ten percent of the water was changed every day. The larvae were fed the rotifer, Brach- ionus plicatilis (Hunter 1976; Misitano 1978). For the March 1982 spawn the culture tank bloomed ini- tially with algae, including species of Tetraselmas and Isochrysus. Fresh rotifers and algae were added daily. After 20 days, Artemia salina nauplii were added in addition to the rotifers. The reared larvae were sampled and viewed at hatching and every day thereafter for 12 days, and then less frequently thereafter up to 35 days. A total of 40 larvae were preserved in 4% formalin after being tranquilized with dilute quinaldine. Twenty- five larvae were preserved for analysis of pigment characteristics and morphometric comparison. Length was recorded as notochordal (NL), flexion (FL), or standard (SL) depending on the stage of caudal fin development. Selected sizes of preserved specimens were drawn using the camera lucida. Field-collected larvae were obtained from the King Harbor Ichthyoplankton collection and the Bightwide Ichthyoplankton Program collection (LACM). Two juveniles (LACM 21639, 43579-1) were obtained from the LACM adult fish collection. Additional specimens were obtained from Marine Ecological Consultants (MEC) of Encinitas, CA. The larva from the King Harbor Ichthyoplankton collec- tion was collected in King Harbor using a single con- ical 1 m diameter plankton net with 335 ^m mesh towed just below the surface (McGowen 1978). A total of 420 larvae was collected by the Bightwide Ichthyoplankton Program along the California coast between Point Conception and the Mexican border using either an Auriga net (benthic sampler) or a 70 cm diameter bongo net for oblique and middepth tows (R. J. Lavenberg pers. commun.* and G. E. McGowen pers. commun.^). A 16.3 mm juvenile was collected by the Bightwide Ichthyoplankton Pro- gram with an Auriga net set with 2 mm diameter mesh. Five specimens from MEC were collected off San Onofre, CA, and the Santa Margarita River, CA, using an Auriga net (W. Watson pers. com- mun.^). Transforming larvae >10.5 mm and <16.3 mm were absent from the above collections. Sixteen specimens were double-stained for bone and cartilage using alizarin red and alcian blue stains (Dingerkus and Uhler 1977) including one juvenile (LACM 43579-1). Eleven specimens including the juvenile were used for meristic counts. Field-collected larvae were identified by compar- ing the pigmentation and myomere counts of smaller size larvae (2.7-5.8 mm) with reared larvae and by comparing larger size larvae (<9.3 mm) with cleared and stained specimens identified using vertebral, dorsal fin, anal fin, and pelvic-fin ray counts. A 10.5 mm larva was identified by meristics, including radiograph vertebral counts, and consistent spine and pigment development. A total of 425 field- collected larvae was examined; 55 larvae and 2 juveniles were observed for detailed pigment char- acteristics, morphometries, and meristic counts. Definition of terms Preanal length = distance from the snout to a ver- tical line through the anus. Body depth = depth of body at the pectoral fin base. Pectoral fin length = horizontal distance from up- per fin base to posterior edge of fin or end of longest ray. Head length = distance from snout to posterior edge of opercle. Flexion length (FL) = distance from the snout to the posterior tip of notochord during the stage when the posterior notochord starts to bend up- ward until the stage when the hypural plates are formed and in their permanent orientation, their posterior edges almost vertical. Eye diameter = horizontal diameter of eye. Pectoral insertion = ventral attachment of pectoral fin. 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. ^R. J. Lavenberg, Natural History Museum of Los Angeles Coun- ty Section of Fishes, 900 Exposition Boulevard, Los Angeles, CA 90007, pers. commun. winter 1982. ^G. E. McGowen, Natural History Museum of Los Angeles Coun- ty, Section of Fishes, 900 Exposition Boulevard, Los Angeles, CA 90007, pers. commun. winter 1982. ^William Watson, Marine Ecological Consultants, 531 Encinitas Boulevard, Suite 110, Encinitas, CA 90024, pers. commun. sum- mer 1981. 202 FEENEY: EGGS AND LARVAE OF YELLOWCHIN SCULPIN Trunk = muscular section of body not including the head, abdominal cavity or caudal fin. RESULTS Fertilization Eight whole egg clutches were externally fer- tilized with freshly stripped sperm, and five con- tained eggs that developed successfully into the embryo stage. Three whole egg clutches were left unfertilized, and none contained eggs that developed successfully into the embryo stage. Three whole egg clutches were split in half; one half was fertilized with sperm and the other was left unfertilized. Eggs in all three of the fertilized halves developed suc- cessfully into the embryo stage; eggs in the unfer- tilized halves did not. Therefore, there was no evidence of internal fertilization. Egg Description Icelinus quadriseriatus eggs are adhesive and negatively buoyant after being stripped from the female and formed a single clutch of 200-250 eggs (excluding eggs that may have been left in the ab- domen; Goldberg [1980] reported an average clutch size of 284 eggs in 28 gravid fish). The eggs are 1.08-1.17 mm in diameter and are initially trans- parent, but the chorion becomes more opaque and textured during development. A pale-green yolk fills most of the egg except for a small perivitelline space (0.024-0.096 mm). At spawning there are about 15 yellow oil glob- ules, the largest of which is 0.14 mm in diameter, which coalesce to one oil globule of 0.14-0.19 mm diameter by the eighth day of development. An opaque, flocculant mass is suspended in the yolk next to the oil globules. Embryonic Development The eggs develop for 12-13 days in 13° -16°C sea- water before hatching. Sixteen hours after artificial fertilization the blastodisc is well formed (Fig. lA). By 38 h the eggs are in the crescent stage of gastrulation (Fig. IB). The germ ring can be seen making its way around the yolk. At 63 hours somites begin to form along the embryo and the eye capsule is present (Fig. IC). At 89 hours the heart (Fig. ID), brain, and otic capsules are visible, and the body is lined with 30 or more myomeres. The heart begins beating after 110 hours, and the vitelline veins can be seen coursing across the surface of the yolk (Fig. IE). At 6-7 days the anus forms, the eyes become darkly pigmented (Fig. IF), and the tail begins flex- ing back and forth. A patch of tubercles forms on the interorbital section of the head and persists until immediately after hatching. At 7-8 days the embryos develop pectoral buds (Fig. IG) and melanophores begin forming on the anus and surrounding yolk sac. The eyes turn a metallic green and blood begins circulating through the ventral veins and arteries of the body. Numerous melanophores cover the yolk sac, anus, and postanal ventral midline (Fig. IH) in late-stage embryos. The head flattens against the chorion and the tubercles spread from the snout to just dorsal to the otic capsule. After 10-12 days eggs reared at 13° -16°C begin to hatch. Description of Larvae The larvae hatch with the oil globules positioned at the anterior of the yolk sac (Fig. 2A). The yolk is absorbed after 6-7 days. The oil globules disap- pear along with the yolk in reared larvae; whereas, the oil globule(s) in field-collected larvae move about the abdominal cavity, are fragmented, possibly in- crease in diameter, or are absent completely (Fig. 2B-C). Evidence of the globule's increase in diameter in the field-collected larvae is found in some specimens (6.1-8.0 mm SL) with oil globules from 0.26 to 0.36 mm diameter, about twice the diameters of oil globules in the eggs and yolk-sac larvae. The percentage of field larvae with oil globules (based on a subsample of 129 larvae) drops from 93% in larvae with yolk sacs (A^ = 15) to 70% in larvae <3.5 mm {N = 79) to 45% in larvae 3.6-6.0 mm (TV = 20) to 27% in larvae 6.1-8.0 mm (N = 15). Lar- vae over 8.0 mm possess minute or no oil globules. Reared larvae hatch at 2.7-3.4 mm NL (after pres- ervation); field-collected larvae are found as small as 2.6 mm. The larvae shorten to a varying degree, depending probably on the reaction to quinaldine and Formalin. Reared larvae shrink 5-17% after anesthesia and fixation. Myomere counts range from 31-37. Double-stained specimens have vertebral counts of 33-35. Morphometries are given in Table 1. From 6.0 to 9.4 mm the larvae become more deep-bodied (27- 30% SL) and the head length increases to 34% of the standard length. A pigmented preanal finfold does not preserve in the field larvae as it does in the laboratory-reared larvae. There is a skin connection between the anus and the rest of the gut, usually with one or more 203 FISHERY BULLETIN: VOL. 85, NO. 2 .0 mm Figure 1-Eggs of Icelinus quadriseriatus: A) 16 hours (LACM 44159-1); B) 38 hours (LACM 44159-2); C) 63 hours (LACM 44159-3); D) 89 hours (LACM 44159-4), arrow indicates heart; E) 110 hours (LACM 44159-5); F) 6 days (LACM 44159-6), arrow in- dicates anus; G) 8 days (LACM 44159-7), arrow indicates pectoral bud; H) 10 days (LACM 44159-8). melanophores, but it is not formed into a finfold as in the reared larvae. Pigmentation Pigment in the larvae at hatching is restricted to the postanal ventral midline, the ventral abdominal surface, the dorsoposterior peritoneum, the anus, and the lower jav^ angle (Fig. 2 A). The postanal ven- tral midline melanophores are positioned in a single row, number from 25 to 37 in yolk-sac larvae and increase to a maximum of 63 in larvae 3.5-4.0 mm in length. These melanophores decrease to a mini- mum of 25 in larvae 5.5 mm or larger. The ventral Figure 2.— Field-collected and reared Icelinus quadriseriatus larvae: A) 2.8 mm (LACM 025-RB-36-AU-01); B) 3 days old (reared), 3.9 mm (LACM 44160-1); C) 4.6 mm (LACM 012-88-36- BB-01); D)6.6mm(LACM012-88-36-BB-01); E) Ventral view, 6.6 mm. 204 FEENEY: EGGS AND LARVAE OF YELLOWCHIN SCULPIN B C D 205 FISHERY BULLETIN: VOL. 85, NO. 2 Table 1 .—Morphometries of Icellnus quadriseriatus larvae, represented as a mean percentage (Xp) of body length, with a range (r) of percentages and a standard deviation (SD). Specimens between dashed lines are undergoing flexion of the notochord. * = reared larvae. Size range (mm) Preanal length Body depth Pectoral fin length Head length Eye diameter N SD SD SD SD SD 2.5-2.9 •2.5-2.9 3.0-3.4 •3.0-3.4 3.5-3.9 •3.5-3.9 4.0-4.4 •4.0-4.4 4.5-4.9 •4.5-4.9 42.0 47.5 41.2 45.9 43.0 39.5 43.0 39.7 46.0 45.0 38.5-46.1) 41.1-51.7) 39.1-45.0) 41.2-51.3) 41.0-44.0) 34.4-50.5) 40.0-47.5) 36.8-43.2) 41.9-47.8) 40.9-49.1) 4.1 4.7 2.7 3.8 1.4 4.9 3.3 3.2 3.2 5.8 29.6 25.3 23.8 21.7 22.9 19.0 24.9 20.4 23.8 24.6 25.2-31.9 23.5-27.5 21.8-25.5 19.1-24.2 21.1-26.6 15.8-28.5 21.0-28.2 15.2-25.2 22.1-27.3 23.1-26.0 3.0 1.7 2.0 2.1 2.5 4.2 3.0 5.0 2.4 2.1 5.9 8.4 9.1 9.4 7.4 10.7 9.3 10.5 7.4 12.5 (5.2-6.8) (4.3-13.1) (7.9-11.0) (7.1-12.1) (6.6-8.7) (7.4-16.7) (7.3-10.3) (6.6-14.8) (6.8-8.0) (10.7-14.2) 0.9 3.9 1.5 1.7 0.9 2.7 1.4 4.1 0.6 2.5 24.0 25.3 23.0 23.6 22.8 21.3 24.5 23.1 22.4 27.5 20.4-26.2 22.1-28.5 21.8-24.7 18.2-26.7 21.1-25.1 17.6-31.5 23.0-26.1 19.1-26.6 19.4-25.9 27.3-27.6 2.5 2.6 1.3 2.8 1.7 4.6 1.3 3.8 2.8 0.2 11.3 12.2 10.9 11.3 10.2 9.7 9.9 9.9 9.3 11.7 (9.6- (11.1- (10.3- (10.0- (9.7- (8.2- (9.0- (8.2- (8.9- (11.6- 12.2) 14.1) 11.3) 12.4) 10.9) 13.6) 10.8) 11.1) 10.0) 11.8) 1.2 1.3 0.4 0.9 0.5 1.8 0.7 1.5 0.9 0.1 5.0-5.4 5.5-5.9 •5.5-5.9 6.0-6.4 6.5-6.9 7.0-7.4 4 4 1 4 4 4 45.6 44.9 48.6 49.2 46.0 46.7 44.0-48.7) 43.6-48.0) 46.1-54.7) 44.3-49.9) 44.1-52.2) 2.1 2.1 3.7 2.7 3.2 24.9 26.5 26.6 27.3 27.2 27.8 21.1-27.7; 25.3-28.0; 25.9-28.3; 24.4-31.2 25.4-29.1 2.8 1.2 1.4 3.0 1.5 7.4 8.7 14.8 9.1 11.5 12.1 (5.5-9.2) (6.5-10.5) (7.5-11.4) (8.1-15.7) (9.3-14.9) 1.8 4.6 2.9 23.4 25.4 29.7 27.5 29.4 27.9 20.4-26.2 22.5-27.5 24.4-30.2 25.3-33.9 24.7-31.5 2.6 2.3 2.4 3.9 2.5 9.1 9.3 11.0 8.7 8.9 8.8 (8.5-10.0) (8.7-10.0) (8.3-9.2) (8.4-9.4) (8.5-9.2) 0.7 0.6 0.4 0.4 0.3 7.5-7.9 3 48.7 8.0-8.4 2 50.7 8.5-8.9 4 51.5 9.0-9.4 4 52.2 9.5-10.4 — — 10.5 1 51.4 16.3 1 51.5 18.7 1 46.5 47.0-49.7) 1.5 27.5 47.7-53.6) 4.2 29.0 49.4-53.2) 1.6 30.8 50.9-54.0) 1.3 29.7 25.6 22.7 26.7 27.2-27.9 27.3-30.7; 27.4-33.9 29.3-30.0 0.4 17.2 (16.3-18.3) 1.0 30.2 2.4 19.7 (15.9-23.5) 5.4 29.8 2.7 22.3 (21.1-24.2) 1.3 34.4 0.3 23.9 (22.5-25.1) 1.2 33.8 29.5-31.2 29.6-30.O 32.4-37.3 31.1-36.7 0.9 9.3 (8.7-9.6) 0.5 0.3 9.6 (8.9-10.2) 0.9 2.4 9.5 (8.7-9.9) 0.5 2.3 9.3 (8.7-10.2) 0.7 32.2 39.1 11.1 28.8 36.8 11.0 24.6 37.9 9.1 abdominal pigment consists of 1-6 rows of melano- phores aligned anteroposteriorly (Fig. 2E). The dor- sal peritoneal pigment consists of 2-5 melanophores in a double row over the posterior half of the gut, increasing to a maximum of 17 in a 5.8 mm reared larva. There are usually 3-5 melanophores surround- ing the anus in newly hatched larvae and up to 12 in the larger larvae. There is always a distinct melanophore on the lower jaw angle. As the larvae grow they develop melanophores on the isthmus (throat), chin, pectoral insertion, ante- rior gut (usually 2 melanophores slightly internal from the ventral abdomen), head, and dorsal body (Table 2). Reared larvae >4.0 mm possess consider- ably more pigment (based on the number of melano- phores) on the dorsal body and the anterior gut areas than field larvae of the same size. The larvae undergo flexion between 5.2 mm and 7.6 mm (reared larvae 4.5-5.8-1- mm). Melanophores (1-4) are usually present on the caudal fin anlage and later at the base of the caudal fin. In flexion and postflexion larvae the caudal fin base melanophores are present in over 95% of the specimens. Postflexion larvae (8.0-9.3 mm SL) develop numerous small punctate melanophores over the midbrain portion of the cranium (Fig. 3A). Melano- phores form between the otic capsule and the hind- brain. Melanophores occur on the preopercle be- tween the eye and the fourth preopercular spine. There are 1-3 melanophores at the pectoral inser- tion, 3-6 melanophores along the pectoral base, and a circle of 7-9 small melanophores dorsal to the pec- toral origin. There are 4-11 small melanophores ven- tral to the eye and dorsal to the maxillary bone and 1 melanophore at the posteroventral edge of the maxillary. Several melanophores occur along the edge of the mandible from the articular to the den- tary bone. Four to five minute melanophores are situated between the eye and premaxillary bone. The ventral abdomen becomes sprinkled with 40-45 melanophores. Transition from larval to juvenile pigmentation starts at about 9.2-9.5 mm SL. Transforming lar- vae develop four patches of melanophores on the dorsal trunk and three patches on the lateral trunk (Fig. 3B). Melanophores appear on the dorsal and caudal fins. In larger specimens (Fig. 3C) the head and dorsolateral pigment takes on the juvenile pat- tern. The dorsal and lateral trunk melanophores merge forming three dorsolateral bars on the body. Distinct patches of melanophores cover the hypural Figure 3.— Field-collected Icelinus quadriseriatus postlarvae and juvenile (D): A) 9.3 mm (LACM 012-88-36-BB-Ol); B) 9.2 mm (MEC 148 Stn. E Rep. #2); C) 10.5 mm (MEC 1-89, E-LS EPI); D) 18.7 mm (LACM 21639). 206 FEENEY: EGGS AND LARVAE OF YELLOWCHIN SCULPIN B D 207 FISHERY BULLETIN: VOL. 85, NO. 2 Table 2.— Presence of melanophores at described locations in Icelinus quadriseriatus larvae, represented as a percentage of larvae showing the melanophoes. * = reared larvae. Size Lower Pectoral range jaw Isthmus inser- Anterior Head Dorsal (mm) N angle (throat) Chin tion gut (dorsal) trunk 2.5-2.9 4 100.0 0.0 0.0 0.0 25.0 0.0 0.0 •2.5-2.9 4 100.0 25.0 0.0 0.0 0.0 0.0 0.0 3.0-3.4 4 100.0 0.0 0.0 0.0 0.0 0.0 0.0 •3.0-3.4 7 100.0 14.3 42.9 14.3 0.0 14.3 14.3 3.5-3.9 4 100.0 75.0 50.0 25.0 0.0 0.0 0.0 •3.5-3.9 8 100.0 75.0 62.5 50.0 12.5 37.5 25.0 4.0-4.4 4 75.0 25.0 50.0 0.0 0.0 0.0 0.0 •4.0-4.4 3 100.0 100.0 100.0 100.0 66.7 33.3 0.0 4.5-4.9 4 100.0 25.0 50.0 0.0 0.0 0.0 0.0 •4.5-4.9 2 100.0 100.0 100.0 100.0 50.0 50.0 50.0 5.0-5.4 4 100.0 50.0 25.0 25.0 0.0 0.0 0.0 5.5-5.9 4 100.0 100.0 0.0 25.0 25.0 0.0 0.0 •5.5-5.9 1 100.0 100.0 100.0 100.0 100.0 100.0 100.0 6.0-6.4 4 100.0 100.0 25.0 50.0 25.0 0.0 0.0 6.5-6.9 4 100.0 100.0 25.0 50.0 0.0 0.0 0.0 7.0-7.4 5 100.0 80.0 40.0 0.0 20.0 0.0 20.0 7.5-7.9 3 100.0 100.0 0.0 66.7 0.0 0.0 0.0 8.0-8.4 2 100.0 100.0 0.0 100.0 0.0 0.0 0.0 8.5-8.9 3 100.0 100.0 66.7 100.0 33.3 33.3 33.3 9.0-9.4 5 100.0 100.0 80.0 100.0 60.0 80.0 20.0 was nine. It is possible that there is some slight variability in this meristic character or the last (in- conspicuous) dorsal spine may have been overlooked in the former's x-ray counts. plates. The ventral abdomen and postanal ventral melanophores begin to fade and become less num- erous. Transformed juveniles (including a 16.3 mm speci- men, LACM 056-OB-75-JA01) display most adult characters (Bolin 1944) including a double row of scales just ventral to the dorsal fin (Fig. 3D). A gap Table 3.-Meristics of icelinus quadriseriatus larvae. ND = no in the scale row is located below the insertion of the ^^^^ (specimens between the dashed lines are undergoing flexion , , 1 ^- Tv,r 1 1 1 , T of the notochord). * = reared larvae; -n = cleared and stained second dorsal tin. Melanophores almost disappear larvae- ° = x-rayed from the postanal and ventral midline and concen- trate on the head, dorsal body, and fins. There is ^ =- <" 1 . c c m^^'^Z 1 f "2 <„ ?"q. E s g ^ '1 r_.cS-.cS-o « ? S S-s8 E r n .£ 2^ <5r^-^ •- -^ "- 0) 3 -s > 5 mm were grouped by 0.3 mm size class. Numbers of larvae per group ranged between 2 and 12; the larger the larval weight, the fewer larvae I grouped together. After drying to a constant weight at 60°C (Lovegrove 1966), larvae were re- moved from the slides by using a single-edged razor blade and weighed on a Cahn electrobalance to ± 2 Feeding To estimate feeding rates and daily consumption, each prey item removed from the larva's gut was counted and its width measured. Width is the dimen- sion that limits a fish larva's selection of prey (Beyer 1980; Hunter 1981). Prey defecated onto the slide were included in the total number eaten. Gut con- tents in dry weight were determined by summing the width- specific weights of the prey eaten each day. I used the width-specific fresh dry weights and caloric values for Brachionus and Tigriopus given by Theilacker and Kimball (1984) and reproduced here in Table 1. Because northern anchovy larvae eat continuous- ly, asymptotic curves (c = C^ax x (1 - e"^')) were used to describe food intake, i.e., the relation be- tween observed gut contents c and time t, where C^ax is the asymptotic gut contents (contents in gut at steady state after filling) and k is the instan- taneous rate of gut filling. Using Marquardt's algorithm for fitting nonlinear models, parameters Cn,ax and k were estimated for 0.5 mm length classes (Table 2). Fish with empty stomachs were included in this analysis because all fish were used in the growth estimates. Daily mean gut contents c were calculated by integrating the area beneath Table 1 .—Width-specific dry weight and caloric value of rotifers, Brachionus plicatilis, and copepods, Tigriopus californicus\ Per individual Width class Dry weight Prey i^m mQ (SE) Volume X 1 0^ Mm X Caloric value 10"^ cal Brachionus plicatilis (4.4 cal/mg) Rotifers 74.3-109.7 0.10(0.01) 109.8-146.7 0.22 (0.04) 146.8-183.8 0.41 (0.06) 183.9-195.0 0.47(0.08) Tigriopus californicus (4.9 cal/mg) Nauplii 74.3-109.7 0.04 (0.01) 109.8-146.7 0.13(0.01) 146.8-183.8 0.25 (0.01) 183.9-195.0 0.38(0.00) Copepodites 146.8-183.8 0.63 (0.15) 183.9-221.0 1.20(0.26) 0.65 1.73 2.96 3.99 0.20 0.55 1.17 1.77 3.38 6.21 0.44 0.97 1.80 2.07 0.20 0.64 1.23 1.86 3.09 5.88 iProm Theilacker and Kimball 1984, table 2. 214 THEILACKER: FEEDING AND GROWTH OF NORTHERN ANCHOVY Table 2.— Estimates ot the asymptotic gut contents (C^^) and in- stantaneous rate of gut filling (k) for northern anchovy (n) fed several diets'. Prey Prey level per mL Prey wei ght SL (mm) number mq Diet n ^max k ^max k Rotifers^ 35 4.00 89 6.16 0.33 5.00 236 9.90 0.32 6.00 131 11.91 1.13 7.00 219 15.92 3.73 8.00 159 28.33 3.78 9.00 171 46.36 2.56 10.00 89 52.22 1.72 Rotifers 25 4.25 58 ^15.55 0.12 1.80 0.29 4.75 69 7.79 0.67 2.12 0.67 5.25 54 12.93 0.62 2.36 0.83 5.75 35 14.01 0.64 2.92 0.58 6.50 28 16.97 1.96 3.31 3.76 8.00 32 27.80 1.84 6.68 1.61 Rotifers 2 4.25 33 7.44 0.35 0.50 0.18 4.75 19 4.97 1.61 1.35 0.98 5.25 18 8.86 1.29 2.42 0.73 8.00 16 9.94 1.30 4.00 0.88 Copepods" 2 4.25 21 1.54 1.27 0.08 0.49 4.75 63 4.89 0.42 0.52 0.37 5.25 44 4.45 0.45 0.94 0.57 5.75 32 3.86 0.41 1.15 1.49 ^6.25 31 4.94 1.01 1.99 0.88 8.00 14 7.22 1.30 3.66 0.66 'Includes zero gut contents; observed gut contents c at time t = C^^ x (1 - e-"). ^Prey width not measured. 387% rotifers <150 ^m width. *Flsh eating Gymnodinium exclusively were removed to estimate C^^ and k. sCopepod concentration <0.1/mL. the curves and dividing by the 12-h feeding period. Equations were derived for each diet from the relation of C^ax and k with fish size and duration of feeding; these equations allow the calculation of gut contents c at time t for all fish lengths between 4 and 8 mm. The parameters (Table 3) were found using the derivative-free nonlinear regression pro- gram (BMDPAR) by Biomedical Computer Pro- grams (BMDP, 1981). Daily consumption F^, was estimated using a modification of an equation for consumption devel- oped by Stauffer (1973) and discussed by Elliott and Persson (1978) and Jobling (1981), F^ = rt + Cjg, where r is the /ug evacuated per hour calculated as mean gut contents c divided by the empirically determined rate of gastric evacuation (see Evacua- tion Rates), t is the duration of feeding, and C12 is the dry weight of the food remaining in the stomach at the end of the 12-h feeding period. Growth Length data for all feeding treatments were fit to exponential growth curves, SL = Iqb''^ where ^0 is length at hatching, k is the instantaneous growth rate and t is the age. Length-weight data were fit to a power equation where weight W = a(SL)''; parameters for the growth equations are given in Table 4. Table 3. — Parameters for equation' relating observed gut contents c in ^^ g dry weight at time t to standard length (SL) of northern anchovy fed three diets. Diet Age n (SD) (SD) (SD) "=0.69 0.33 0.326 , PjSLd-e -(p, +P2-SL)(/) (SD) Rotifers 25/mL 5-14 280 -1.90 (0.70) 0.51 (0.18) 0.57 (0.16) 0.29 (0.03) 2/mL 5-14 103 — — Copepods 2/mL 5-9 171 2.84 (0.88) -0.39 (0.12) ^1.26 ^(0.75) 1.18 (0.10) 'c = p^e' 2(P, + Pj) fixed at 0.69, the mean k from Table 2; asymptotic standard deviations could not be computed because k was held constant. 3x 10'^ Table 4.— Parameters for northern anchovy growth equations where SL is standard length in mm, W is weight in /jg, and t is age in days. Diet Age (d) Age length SL = /o©*' Length-weight W = a(SL)'' In (SD) (SD) (SD) (SD) Rotifers 25/mL 5-14 280 3.06 0.05 0.06 0.001 253 0.197 0.040 3.16 0.11 2/mL 5-14 103 3.14 0.07 0.05 0.002 84 0.379 0.030 2.80 0.04 Copepods 2/mL 5-9 138 3.06 0.15 0.07 0.010 109 0.297 0.097 2.88 0.20 215 FISHERY BULLETIN: VOL. 85, NO. 2 Metabolism I determined metabolic rates for the larvae, which ranged in age from first-feeding (3 days after hatch- ing) to 25 days using the Winkler technique. I chose the Winkler technique where large vessel volumes could be used and there was no need to shake the vessels during the experiment, as required for mano- metric techniques. Pearcy et al. (1969) found no dif- ferences between Winkler and Warburg estimates of oxygen consumption. Oxygen consumption was estimated at 16°C during 18-23 h experimental periods with a 12 h light-dark cycle. The respiration vessels were attached to a large, slowly rotating wheel. Young larvae, 0.02-0.14 mg dry weight were tested in 40 mL vessels in groups of 10-50, while larvae older than 16 days (larger than 0.14 mg) were tested individually in 60-150 mL vessels. All fish tested had empty guts. Data were not used when mortalities occurred during the experiment. To express metabolism (Q) as a function of dry weight, I used a nonlinear regression to fit a power equation to the data (see parameters for Model 1 in Table 5). The data points were weighted by their sample size {n = 10-50). The Model 1 fit was un- satisfactory for the whole size range, presumably because each data point for the young larvae (n = 72) was a group mean, and the model was signifi- cantly weighted toward the young larvae, causing it to overestimate oxygen consumption for the few large larvae {n = 17). Because the experimental technique differed (i.e., respiration was measured for groups of young larvae or individual older lar- vae), I also fitted two separate curves. These curves (Model 2) gave a good fit to the data (Table 5); the Model 2 equation for younger larvae was used in the present study. An alternate approach for estimating metabolic requirements is to starve larvae of known size (weight), determine the size-specific weight loss, and convert the weight loss to calories. This approach eliminates the need to restrict larval swimming ac- tivity in a respiration vessel. Presumably the weight Table 5.— Parameters for equation Q = aw'' where Q is metabolic rate in ^L Oj/h for northiern anchovy and w is their fresh dry weight. N Size group (mg dry wt) w Parameters Model a(SE) to(SE) 1 2 2 89 72 17 0.02-2.70 0.02-0.14 0.14-2.70 ^3.844 -1- 0.100 2.897 + 0.344 4.269 ± 0.325 0.858 ± 0.029 0.834 + 0.057 0.697 + 0.107 'The sum of the residuals In Model 1 does not equal zero, thus the pro- gram calculation of SE's Is biased. loss in caloric units would equal the loss due to metabolic costs, excluding the metabolic cost of attacking prey. Using this approach, I fed control northern anchovy larvae ad libitum on Gymno- dinium and Brachionus and starved the test larvae; both groups were maintained in 100 L rearing tanks at 15.5°C. Live standard length and dry weight of groups of the same length were measured daily, as described earlier in this Methods section, for 10-50 larvae sampled daily from each treatment. I calculated the caloric equivalent of northern an- chovy tissue using the caloric values given by Hunter and Leong (1981) for fat-free anchovy tissue, 4.129 cal/mg, and for anchovy lipid, 9.227 cal/mg. For ex- ample, northern anchovy larvae weighing 25 ^g con- tained 6 fig of lipid (unpubl. data: John Hakanson, UCSD, Scripps Institution of Oceanography); using the above caloric equivalents for 19 /jg of fat-free tissue and 6 ng of lipid yields 5.36 cal/mg as the energy equivalent of anchovy tissue. In a 20-d laboratory experiment, Hakanson found that lipid weight appeared to increase proportionally with an- chovy weight, thus the caloric content of anchovy tissue would be approximately constant for the age range studied. Lipid content seems to be lower in older northern anchovy larvae. The only other in- formation I found was for 40-60 d-old northern an- chovy where the caloric content averaged 4.9 cal/mg (unpubl. data: John Hunter, Southwest Fisheries Center). I used 5.4 cal/mg as the caloric value of an- chovy tissue for larvae between 5 and 14 days of age. Evacuation Gut clearance times were determined for active- ly feeding fish of various ages fed the rotifer and copepods diets. Larvae were transferred from the 100 L rearing tank to a 10 L test tank. Because northern anchovy larvae are sensitive to handling, handling was restricted to one transfer. Transferred larvae were kept in the test tank for 18 hours prior to an evacuation experiment because injured larvae usually die within 8-10 hours after transfer. First, larvae were fed a low concentration of prey that had been dyed with National Fast Blue (Laurence 1971). After larvae had filled their guts, eating most of the dyed prey, a known density of undyed prey was added. Larvae were sampled at 5-min intervals, and the time required for them to void their guts of the dyed prey was determined. The number of prey in the full guts was counted and converted to dry weight. Evacuation rates are given as fjg prey cleared through the gut per hour. Rates were re- lated to fish size and to prey type. 216 THEILACKER: FEEDING AND GROWTH OF NORTHERN ANCHOVY The timing of the second prey addition was not critical for determining gut clearance rates at high prey densities. But when the timing was not cor- rect for the tests that used low prey concentrations, deciphering the meaning of the gut contents was problematical. Results from most of these low- density tests could not be used. A series of evacuation experiments also was con- ducted with nonfeeding northern anchovy that were removed from their food source, rotifers, to filtered seawater. To reduce the incidence of injury during transfer, I constructed a cylindrical, clear plastic container (15 mm high and 7 mm diameter) with handle and a removable bottom grooved to fit the circumference of the cylinder. Larvae to be transferred were sur- rounded by the cylinder, and then the bottom was fitted onto the cylinder. The container with fish was transferred and lowered into the test tank. Remov- ing the bottom and slowly raising the cylinder re- leased the fish. Prey that were transferred into the experimental tank with the larvae were removed by an air-lift pump that slowly recirculated water and was screened to prevent the removal of larvae (O'Connell and Paloma 1981). Growth Efficiency To determine growth efficiencies, I used the in- formation on growth (Table 4), daily food consump- tion estimated from the general equations (Table 3) and evacuation rates of 1.15 hours for the high- density rotifer diet and 1.5 hours for the low-density diet. Gross growth efficiency was estimated based on dry weight and on caloric estimates. It is the ratio of growth to ingestion. To estimate assimilation ef- ficiency, I used the information on weight-specific metabolic rates and simply combined the energy of metabolism and growth and divided it by the energy consumed. Assimilated energy lost as feces and urine was not accounted for. RESULTS Feeding Sizes of prey fed to larval northern anchovy in these experiments ranged from 50 /um for Gymno- dinium, to 74-195 ^m for rotifers and copepod nauplii, and 147-221 f^m for copepodites. In contrast to larvae fed the rotifer diets, fish offered the cope- pod diet did not switch from eating Gymnodinium at first feeding on day 3 to eating larger prey on day 5. All fish fed copepods and sampled on day 5 contained Gymnodinium in their gut, and 96% of these guts were full of Gymnodinium (Table 6). By day 7, the number of copepod-fed northern anchovy eating Gymnodinium had decreased to 80%, with 10% full. These data reveal that young northern an- chovy were unsuccessful at capturing Tigriopus nauplii at 1/mL until 7-8 days of age, and because of this behavior and the failure of the copepod culture on day 9, 1 was unable to quantify daily con- sumption for northern anchovy fed copepods. Northern anchovy reared on the low-density (2/mL) and high-density (25/mL) rotifer diets ate only a few Gymnodinium cells after day 4. Between rotifer treatments, incidence of feeding on Gymno- dinium after day 4 was higher for fish fed the low- density diet (Table 6). On day 5, northern anchovy concentrated on eating rotifers in the 75-150 /^m size range; between 84 and 97% of the rotifers eaten by larvae sampled from both rotifer treatments were <150 Mm width (Table 7). Only 7% of the rotifers available to these larvae were smaller than 150 /^m width. Hence on day 5 larvae were selecting small rotifers in higher proportions than were available; in the two treatments, the apparent density of roti- Table 6.— Presence of Gymnodinium in guts of larval northern anchovy related to age of larvae and to diet. Diets Rotifers 25/mL^ Rotifers 2/mL^ Copepods 2/mL^ Aqe Incidence Gut full Incidence Gut full Incidence Gut full (d) n (%)' (0/0)5 n (%r (0/0)5 n (%r (o/of 5 21 48 0 29 83 0 27 100 96 6 82 59 0 11 9 0 57 77 16 7 10 10 0 5 0 0 49 80 10 8 75 9 0 0 — — 5 0 0 9 27 7 0 38 6 0 33 0 0 'Apparent density of rotifers <150 ^im = 2/mL. ^Apparent density of rotifers <150 jjm = 0.1/mL. ^Apparent density of nauplii < 150 ^jm = 1/mL. ••Percent of larvae having Gymnodinium cells In guts with rotifers or copepods. ^Percent of larvae having guts full vtrith Gymnodinium cells. 217 FISHERY BULLETIN: VOL. 85, NO. 2 Table 7. — Width frequency of prey eaten by northern anchovy (n) related to their diets and to their age and size. Rotifers 25/mL^ Rotifers 2/mL^ Copepods 2/mL=' Aqe Prey composition (%) Prey composition (%) Prey composition (%) (d) n <150 nm >150 urn n <150 pm >150 ^m n <150 ^m >150Mm 5 21 83.5 16.5 29 97.3 2.7 27 88.0 12.0 6 82 11.0 89.0 11 75.0 25.0 57 60.0 40.0 7 10 5.2 94.8 5 38.7 61.3 49 59.0 41.0 8 75 33.1 66.9 0 — — 5 33,4 56.6 9 27 40.9 59.1 38 12.2 87.8 33 31.8 68.2 Length (mm) 4.1-4.5 60 49.7 50.3 37 91.4 8.5 21 80.0 20.0 4.6-5.0 77 25.7 74.3 20 23.4 76.6 69 68.3 13.7 5.1-5.5 55 43.9 56.1 23 15.1 84.9 45 57.3 42.7 5.6-6.0 35 36.6 63.4 9 11.8 88.2 30 40.0 60.0 6.1-6.5 23 38.2 61.8 0 — — 17 36.4 63.6 'Apparent density of prey <150 ^jm = 2/mL. ^Apparent density of prey <150 nm = 0.1/mL. ^Apparent density of prey <150 ^im = 1/mL. fers <150 ^m was equivalent to about 2 and 0.1/mL. In sum, the larvae switched to eating prey >150 yim on day 6 in the high-density rotifer treatment, day 7 in the low-density treatment, and day 8 in the copepod treatment (Table 7). If the analysis is based on larval size instead of larval age, selection for prey >150 /um occurs at 4.6-5.0 mm for northern anchovy fed the rotifer diets and at 5.6-6.0 mm for those fed the copepod diet (Table 7; Fig. 1). Feeding intensity was highly variable on all diets. The observed stomach contents can differ by as much as a factor of three (Fig. 2) from the pre- dicted stomach contents (C^ax) used in the feeding models. At high rotifer densities, northern anchovy filled their guts at a faster rate and consumed more. On average, fish eating the high-density rotifer diet were full within 2 hours, while those eating a low- density diet were full in about 3 hours. Comparing the average observed stomach contents, and ex- cluding empty stomachs, for fish of equal age shows that all larvae eating at the high prey density ate more than their counterparts eating at lower prey densities (Table 8). Growth Hunter (1976) showed that the length-weight rela- 100 u o. a. S ID (/) Z o o >• m a. a 4.0 5.0 6.0 7.0 8.0 9.0 ANCHOVY STANDARD LENGTH (mm) 10.0 Figure L— Size of copepod prey eaten by northern anchovy related to fish size. See Table 1 for copepod width classes. 218 THEILACKER: FEEDING AND GROWTH OF NORTHERN ANCHOVY (/) 5.25 z UJ 4. SO h- z u o sz 3.75 X u> o < 01 5 3.00 o >. ■o 2.25 1- 3. >- * — 1.50 > o X .750 o z < 3.000 - • - • • - t • •• • • • •• -- - • • • • • • • • C max • • -^•l • • 1 - •• 1 1 r •" 1.50 2" 3.00 3" ,50 5.25^^0 6.75 ^ ^^ TIME (hours) Figure 2.— Observed stomach contents of 4.0-4.5 mm northern anchovy fed 25 rotifers/mL, predicted rate of gut filling k and predicted maximum gut content C^^. Each point is one larva. Table 8. — Mean dry weight of food (^g) observed in stomachs' of northern anchovy (n) related to their age. Rotifers Rotifers Copepods Copepods Age (d) 25/mL 2/mL 2/mL 0.2/mL n Mg n MQ n MQ 5 9 1.75 17 0.60 8 0.18 6 57 2.23 4 0.27 39 0.55 7 8 3.56 5 1.41 37 0.67 8 51 2.10 — — 3 0.87 1 0.08 9 26 2.99 27 2.60 31 1.88 6 0.78 10 — — — — 9 1.51 11 6 3.60 — — 9 0.42 12 22 5.14 4 ^6.39 3 0.90 ■> Average stomach contents (excludes empty stomachs) calculated for f = >3 hours; 3 hours is a reasonable time for northern anchovy to fill guts eating at lowest prey densities tested (see text). 2Day 13. tion for northern anchovy was curviUnear on a log- log plot, and he used a Laird Gompertz model to describe both growth in length and in weight over 75 days. Because I am describing only the first 2 weeks of growth, I used a simpler exponential model (Table 4) which probably should not be used for lar- vae beyond 2 weeks of age. Larvae grew at 0.35 mm/day on the copepod diet {k = 0.07) and at 0.33 and 0.25 mm/day on the high- {k = 0.06) and low-density (A; = 0.05) rotifer diets (Table 4). On the average, the dry weight of larval northern anchovy was proportional to length to the third power. Depending on diet, the length exponent ranged from 2.80 to 3.16 (Table 4). Daily percent increases in dry weight for northern anchovy fed the rotifer diets were 15% for larvae fed 2/mL and 21% for larvae fed 25/mL. However, northern anchovy grew the most, 23% per day, on the 2/mL copepod diet. For the three diets tested, daily growth in dry weight as a percent was con- stant over the size range. An analysis of covariance was used to test for diet- induced differences in the relation between natural logarithms of length and weight for larval northern anchovy between days 5 and 9 when prey concen- trations were controlled. Larvae which fed on cope- pods were significantly longer and heavier at age than larvae eating rotifers. There were no length or weight differences at age 7 days between the two rotifer treatments (Table 9), but there were differ- ences in growth thereafter. Larvae raised on the two rotifer diets were the same weight at 4.9 mm (30.75 and 33.25 yig; P = 0.3), but larvae raised on cope- pods weighed less at 4.9 mm (28.54 ^g) than larvae fed on the rotifer diets (P = 0.11 and <0.01). I compared the distribution of my northern an- chovy dry weights at length for the rotifer and Table 9. — Effect of diet on northern anchovy standard length (SL) and dry weight {W). Age = 7 days Probabilities = SL and = W Diet Density per mL No. cases' X SL X W (mm) (^xg) 1 vs. 2 SL/tV 1 vs. 3 SUW 2 vs. 3 SL/kV 1 -Rotifers 2-Rotifers 3-Copepods 25 2 2 34 14 28 4.78 28.36 4.61 27.33 5.28 35.11 0.20/0.71 0.00/0.00 0.00/0.00 SL = 4.9 mm Probabilities xW (ixg) 1 vs. 2 1 vs. 3 2 vs. 3 1 -Rotifers 2-Rotifers 3-Copepods 25 2 2 34 14 28 30.75 33.25 28.54 0.30 0.11 0.00 'Case numbers contain 2-12 larvae depending on number in group that were weighed. 219 FISHERY BULLETIN: VOL. 85, NO. 2 copepod diets with those in a study by Hunter (1976) where northern anchovy were fed Brachionus, 50-100/mL, and copepods, Tisbe 0.01/mL, at 17°C and Gymnodinium was fed as the first food. The curves show that among experiments there appear to be diet-induced differences in weight at length (Fig. 3). Metabolism The caloric equivalent of metabolism for northern anchovy larvae ranging in age from first feeding to 25 days was determined using the relation between the metabolic rate and fresh dry weight (Model 2, Table 5) and by converting the oxygen uptake to calories using an oxycalorific equivalent of 0.00463 cal/f.- cc "^ 100 >• > o O 80 z < 60 - 40 • Rotifers 50-100/ml & copepods 0. 1/ml 180 A Rotifers 25/ml 0 Rotifers 2/ml 160 — • Copepods 2/ml "S Is Copepods 0.1./ml a. 140 - 1- I g LLI 5 120 4 5 6 7 8 9 STANDARD LENGTH (mm) Figure 3.— Relation between dry weight and standard length for northern anchovy fed several diets where Gymnodinium was the first food. I also determined the caloric equivalent of metab- olism for first-feeding northern anchovy by starving them, determining the weight loss, and converting the loss to calories. Starving larvae lost an average of 10% of their body weight per day for 3 days, after which larvae must have continued to lose weight, but no decrease could be measured (Fig. 4). The time to maximum weight loss was 3-4 days, the time of irreversible starvation for northern anchovy from the onset of feeding (Lasker et al. 1970; Theilacker and Dorsey 1980). Larvae that weighed an average of 0.0211 mg at 3 days of age weighed an average of 0.0148 mg on day 6. The weight loss (0.0063 mg) X 5.4 cal/mg, which is the assumed caloric equiv- alent for northern anchovy tissue, equals 0.0339 calories, or a metabolic demand of 0.011 cal/day. This value, determined at a slightly lower temper- ature, corresponds well with the value obtained for first-feeding northern anchovy weighing 0.0211 mg using respiration measurements (0.013 cal/day). Evacuation Rates Gut clearance times for northern anchovy larvae appeared to be independent of larval age; however, the weight of food evacuated per hour increased with age because the stomach contents increased. Because the larvae fed at a constant rate after the gut was filled and defecated continuously, the gut clearance rate for actively feeding northern anchovy larvae was constant. The average gut clearance time for anchovy feeding on 25 rotifers/mL was 1.15 hours (SE = 0.13; range 0.7-1.5 hours; n = 6 tests). Reducing the prey density to 2/mL increased the average gut clearance time to 1.5 hours (range 1.2- 1.8 hours; n = 2 tests) for the rotifer diet and 2.73 hours (SE = 0.26; range 2.0-3.3 hours; n = 4 tests) for the copepod diet. Nonfeeding northern anchovy cleared their guts in 2.8-5.8 hours, depending on their size and stomach capacity (2.8 h/4.8 mm; 4 h/6.3 mm; 5 h/7.9 mm; 5.8 h/8.5 mm). Daily Consumption of Rotifers and Growth Efficiency Daily consumption was less on the low-density diet and, as a percent of body weight eaten per day, con- sumption ranged between 31 and 86%, depending on prey concentration and fish size (Tables 10, 11). For both rotifer diets, weight-specific consump- tion decreased with increasing body weight (Fig. 5). Gross growth efficiencies were higher for north- 220 THEILACKER: FEEDING AND GROWTH OF NORTHERN ANCHOVY em anchovy fed the low-density diet (Tables 10, 11). Mean gross-growth efficiency (days 5-14) based on dry weight was 0.30 for the high-density diet and 0.37 for the low-density diet. Based on calories, mean gross-growth efficiencies were 0.37 and 0.46 respectively. 35 ^ 30 o> 3. 'q) 5 20 Q 15 FED 1 4.5r- E ^ 4.0 o X) o 3.0 STARVED • = 5-20 larvae Day 3 = first feeding 5 6 7 3 4 5 AGE IN DAYS Figure 4.— Changes in standard length and weight of fed and starved north- ern anchow larvae over time. o < CO UJ cr O _i < o < > < a X g UJ X LU O z < X u .09 .08 .07 .06 .05 .04 .03 .02 .01 Rotifers 25/ml Rotifers 2/ml .05 .10 .15 .20 CONSUMPTION/DAY AS CALORIES (F^.) Figure 5.— Gross growth efficiencies (W^/F^ ; Tables 10, 11) of north- ern anchovy fed rotifers at 25/mL and at 2/mL. 221 FISHERY BULLETIN: VOL. 85. NO. 2 in CM n ui (O •o ® >> > o j: o c r o c w > o c a> o "55 o E Q. (« CO rj <■ — to c 5 6-^ i < s < 3. (0 O c g Q. E 3 (O c o O O) o _^ 3 CO '(^ o E cn o) > Q 0)3 ^ 73 O) E t^ C e CO d) S w crno < b c\j'a-mr^coT-'>tCDCX3 0JCM(MC\lC\JCOCr)COCO OCDOCDCDOOCDCD 0''5fCOOCM'-;(r)h-. OJC\jC0CO'^Lf>CC>t~^CK)C3i cocx>cni-c\icomh-c3)i- OOO'--'— •^'-^'-;C\I OCDOcnJcDOOOOO T-cDcooocor^cvjco'*'- oooooor-.h-cDCDir)ini/j Ti-'-'i-o>r^cocNj'a-'Tr~- T--^t-cq •^co-r^cocbocooocvicb i-7-CyJC\JC\JCOC0CO'*'^ coocncji-^ooh-cocoin r~.0'-cocD(r)'-ir)C3>'^ ■i-^c\ic\Jcvic\ic\JcococO'a- (35'^CvJh-COOJ'^'-'t' o r^i-^iDCDr-^iri'^uicTicb i-CMCVJCOCO'tinCDh-CJ) cocjjcDLniDcxjCMCJiooa) •■-cocoo5C\imcDC\icpo L0CDI~--00CJ)O'-CNJC0' .i2. ra CO E 2 CO o5 "to 0) i_ O c o CO o c g in CD H. 0) J2 O O o E to E i s ■9 « TO ■ — ■ II d C o CJ CD o c o 1- ~ 0) S 3. CO ss ^, CO O) ^^ 5 E 1 s fvj ■C3 CO o O O 15 CO _J c ^ O (M 5s 3. (0 2 in X < X II CO C3) 1 II ^ + II o § o II u:* o^ o _J 6 ^ ^" i ^ U. 3. o < < < < en o ^ '^ g> 5 CO E a. E 3 (/) C x: ° CD C .S T '^ ■'- a I CO -■• ~ II ^ JZ fc: cn„ O o -~ CJ> O) CD ^ c p " ., " 11 " ■K II II ^ CO " ^ " ^ •_ I -.- S .. lo O U. + ° II 222 THEILACKER: FEEDING AND GROWTH OF NORTHERN ANCHOVY , a £^ E S ^ CDCDTD-OCOCDIOCOCM <^l cotocor^t^ooooooo5 .iT^ U." ioa)oocor^oo-.-cor>- u> >. J" cnmm'^-^rtinmin "—^ 0> " s.i oooddddod < _^^ Oj^ u.* a)T-c\iLna)05c\j'cot^OLnot^ _i E ^ c:h « 1 ■r-7-T-C\JC\JCOC0M--<3- ,^ 0) - O) qoooooooo "x 0)$ -Si "^ odddddddc} B 0) M, CM c S 9 (TJ f re OB C

cocMcqc\iqir)LO'*{0 (0 E (0 ^z^ 0;;= ? < ojcooM-LfiLricbr^cd "re (U 1 -^ o c 0 o ^ C\ICOlOCDC3>i-COCOOCO CO ^i to g c 0 w ,^~^ 0) :^ 6 •i-^^-,--,-c\jc\jcgcoco in ^ Q. ■a '^_^ CO « O pppooqoooo E CO c 0 re E 0) o dddddddddd P- 1- II > o o CO 5 CM ocoeocooi-CDooeooo ra 1- to 0 o o 2 o LOoo-r-tnoino(Dcoi- ^^ _l CC 0 ra c (0 ^"^ _l c\Jc\Jcoco'srTririiritDt^ a> >> c 3 CO o " — ' re ■D O) a> O) ^ a E co(DO'5tcocoo>inooc\i E ff d ■0 0 0 o c -^ >; ::^ . " ■"^■^LnLniocDcot^ooa) "co <» . c O) O CO U. qoooooqooo o si. i re ■^ 1 c o to o dddddddddd 1- 'a- X CM O) in X < Q. m O) < w oo E 13 COOlOCOOOOCD'a-CDi- X II II CO •si- :5 S CD S + 0 ^6 o o 0) c o o in'^'*--cotocDLOC\ja)c\jco 01 0^9 < < 0 <7i ^ < CL a) ^;^ O O , i TOinco-^-r-coqqoooo •" •" " ^ £31^ 03dT-c\ico'ij-ir5t---o6d ■.-■.-7--.-,-T-T--,-C\J ^ <4— CO O O) ^ (0 0) (1) CD-'-i-CM'rtO'a-CO'a-Oi JZ c o c\jcottinqr-.p^coif) s O) T^^^T-'^T^^CMCVJOJ "o >s o "□ c o ^^ 3 ooir)'coirooh- Q. 1— •c r~~qqinoqT-Tro-% 0) — 1 "*" re ~" s 10 o w T3 X O O a 8 < S o8 6 H . 5 > o I o z < • • • • • • •• •• • •• • • • • • • •• J L ? I* I 0.0 10 2.0 3° 4.0 50 6.0 ^O 8.0 ^-^ 10.0 "012.0 TIME(hours) Figure 6.— Number of copepod nauplii observed in stom- achs of 4.0-5.0 mm northern anchovy fed 2 copepods/mL. Each point is one larva. Northern anchovy raised on copepods ate mainly Gymnodinium, augmenting their diet with copepod nauplii (Table 12). The stomachs of 96% of the 5-d- old larvae were fuR with Gymnodinium cells (Table 6). Fish were obtaining 60-90% of their daily caloric intake from Gymnodinium. This is evident by com- paring the daily caloric intake (FJ for day 5-6 lar- vae of equal size or weight that were fed copepods (Table 12) with those fed rotifers (Tables 10, 11). Consumption of 2-4 cells/minute can account for this energy input (4.2 x 10"^ call Gymnodinium cell [Vlymen 1977]), and successful feeding acts of this magnitude have been directly observed by Hunter (1981). Lasker and Zweifel (1978) developed a model (a modified version of Vlymen's [1977] model) to describe survival at sea in areas of various concen- trations and proportions of large and small prey and concluded that large prey made very little contribu- tion to survival of first-feeding larvae when suffi- cient (40 mL) small prey were available. As observed here, the ingestion of 10-20 copepod nauplii/day in addition to small Gymnodinium cells resulted in a growth rate of 0.35 mm/day, which is comparable to a rate of 0.37 mm/day reported for wild north- ern anchovy of similar age (Methot and Kramer 1979), and lends additional credence to Lasker and Zweifel' s (1978) hypothesis that survival of northern anchovy depends on patchy (layered) distributions of small, abundant prey like Gymnodinium. For all diets, an equivalent number of small prey. 225 FISHERY BULLETIN: VOL. 85, NO. 2 Table 12. — Estimate of caloric input from copepods eaten by northern anchovy. C) e) e) C) e) e) C) e) e) Gut Gut contents Consumption Standard Dry clearance Body Age length weight rate Dally mean Copepod s weight Copepods Diet (d) (mm) 0^9) (>^g/h) {^'9) (mQ) (%/d) (cal/d) A / W r c est c obs F. ^wlW Fc Copepods 5 4.34 20.35 0.07 0.19 0.16 1.07 05 0.0052 2/mL 6 4.66 24.99 0.10 0.28 0.30 1.55 06 0.0076 and 7 4.99 30.43 0.15 0.41 0.43 2.26 07 0.0111 Gymnodinium 8 5.36 37.39 0.23 0.62 0.68 3.44 09 0.0169 9 5.75 45.77 0.35 0.96 1.08 5.32 12 0.0261 'Hatching = Day 0. % = 3.06 e°°''. ^W = 0.297 /^°^ '^r = c est/2.73 h (see text). ^Estimated using equation in Table 3. ^Average stomach contents calculated for f = >3 h; includes empty stomachs. 'F^ = 1 2 r + c est. 8% body weight eaten/d as copepods. ^Calories/d, copepods; F x 4.9 cal/mg (Table 1). Gymnodinium cells, was available and the concen- tration of large prey was varied. Availability of large prey of a suitable size in the copepod diet was 10 times the number available in the low-density rotifer diet. Because 4 mm (day 5) larvae fed copepods ate mainly Gymnodinium, and those fed the low-density rotifer diet ate mainly rotifers (Tables 2, 6), cope- pods nauplii must be more difficult to catch than rotifers, and consequently larvae consumed the more abundant Gymnodinium cells. Larvae which consume prey as they are encountered, rather than choosing a diet that maximizes the energy gained per unit foraging time, have been labelled "number maximizers" as opposed to "energy maximizer" in the parlance of Griffiths (1975) and Hughes (1979). Additional evidence that points to northern anchovy feeding as "number maximizers" is that, when prey of the proper size were available, their feeding rates paralleled prey abundance (Tables 10, 11). The energy budget I calculated for northern an- chovy fed rotifers at two concentrations gives in- formation on their growth requirements that can be translated to growth requirements in the field. My data support Boehlert and Yoklavich's (1984) and Checkley's (1984) conclusions that larval fish may exhibit a high growth rate or a high growth effi- ciency, but not both at the same time. Boehlert and Yoklavich studied Pacific herring, which are 2-4 times the weight of northern anchovy, but like an- chovy feed continuously, and found that as consump- tion increased, the total amount of food assimilated continued to increase despite a decrease in the effi- ciency of the assimilation. Checkley studied Atlan- tic herring and found that the gross growth efficien- cies of Atlantic herring increased with increasing consumption, but he showed that the relation was peaked, and by incorporating results from the literature for other species, he also described a decrease in growth efficiency at high consumption for larval fishes. ACKNOWLEDGMENTS I wish to thank Patricia Hadley Hansen for her capable assistance during the laboratory phase of this work, Bob Lucas for conducting the evacuation experiments on nonfeeding larvae. Amy Kimball for assisting with the data analysis, and Nancy C. H. Lo for her consultation and valuable suggestions. John Hunter, Pattie Schmitt, Ken Frank, Timothy C. Lambert, and two anonymous referees critically reviewed the manuscript and offered suggestions. I appreciated the help of the manuscript support staff, in particular Jean Michalski and Lorraine Prescott. LITERATURE CITED Arthur, D. K. 1976. Food and feeding of larvae of three fishes occurring in the California Current, Sardinops sagax, Engraulis mor- dax and Trachurus symmetricus. Fish. Bull., U.S. 74:517- 530. Beyer, J. E. 1980. Feeding success of clupeoid fish larvae and stochastic thinking. Dana 1:65-91. Blaxter, J. H. S., AND J. R. Hunter. 1982. The biology of clupeoid fishes. Adv. Mar. Biol. 20:2- 223. Boehlert, G. W., and M. M. Yoklavich. 1984. Carbon assimilation as a function of ingestion rate in larval Pacific herring, Clupea harengus pallasi. J. Exp. Mar. Biol. Ecol. 79:251-262. Brett, J. R., and T. D. D. Groves. 1979. Physiological energetics. In W. S. Hoar and D. J. Ran- dall (editors), Fish physiology, Vol. 8, p. 279-352. Acad. Press, N.Y. and Lond. 226 THEILACKER: FEEDING AND GROWTH OF NORTHERN ANCHOVY Buckley, L. J., and D. W. Dillmann. 1982. Nitrogen utilization of larval summer flounder, Para- lichthys dentatus (Linnaeus). J. Exp. Mar. Ecol. 59:243- 256. Checkley, D. M., Jr. 1984. Relation of growth to ingestion for larvae of Atlantic herring, Clupea harengus and other fish. Mar. Ecol. Prog. Ser. 18:215-224. Chitty, N. 1981. Behavioural observation of feeding larvae of bay anchovy v4 wc/ioa mitchilli and big eye anchovy ylwc/ioa lam- protaenia. In R. Lasker and K. Sherman (editors), The early life history offish, p. 320-321. Rapp. P. -v. Reun. Cons, int. Explor. Mer 178:320-321. Eldridge, M. B., J. A. Whipple, and M. J. Bowers. 1982. Bioenergetics and growth of striped bass, Morone sax- atilis, embryos and larvae. Fish. Bull, U.S. 80:461-474. Elliott, J. M., and L. Persson. 1978. The estimation of daily rates of food consumption for fish. J. Anim. Ecol. 47:977-991. Griffiths, D. 1975. Prey availability and food of predators. Ecology 56: 1209-1214. Haegler, C. W., and B. N. Outram. 1978. The effects of diet and ration on the growth and sur- vival of Pacific herring, Clupea harengus pallasi, larvae. Fish. Mar. Ser\'. Tech. Rep. 767, 41 p. Houde, E. D., and R. C. Schekter. 1981. Growth rates, rations and cohort consumption of marine fish larvae in relation to prey concentrations. Rapp. P.-v. Reun. Cons. int. Explor. Mer 178:441-453. 1983. Oxygen uptake and comparative energetics among eggs and larvae of three subtropical marine fishes. Mar. Biol. 72:283-293. Hughes, R. N. 1979. Optimal diets under the energj' maximization premise: the effects of recognition time and learning. Am. Nat. 1 13: 209-221. Hunter, J. R. 1972. Swimming and feeding behavior of larval anchovy, 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. 1977. Behavior and survival of northern anchovy, Engraulis mordax, larvae. Calif. Coop. Oceanic Fish. Invest. Rep. 19: 138-146. 1981. Feeding ecology and predation of marine fish larvae. In R. Lasker (editor), Marine fish larvae, p. 34-77. Univ. Wash. Press, Seattle and London. Hunter, J. R., and R. Leong. 1981. The spawning energetics of female northern anchovy Engraulis mordax. Fish. Bull, U.S. 79:215-230. Jobling, M. 1981. Mathematical models of gastric emptying and the estimation of daily rates of food consumption for fish. J. Fish. Biol. 19:245-257. Kramer, D., and J. R. Zweifel. 1970. Growth of anchovy larvae (Engraulis mordax Girard) in the laboratory as influenced by temperature. Calif. Coop. Oceanic Fish. Invest. Rep. 14:84-87. Lasker, R. 1964. An experimental study of the effect of temperature on the incubation time, development and growth of Pacific sar- dine embryos and larvae. Copeia 1964:399-405. Lasker, R., H. M. Feder, G. H. Theilacker, and R. C. May. 1970. Feeding growth and survival of Engraulis mordax lar- vae reared in the laboratory. Mar. Biol. 5:345-353. Lasker, R., and J. Zweifel. 1978. Growth and survival of first-feeding northern anchovy larvae {Engraulis mordax) in patches containing different proportions of large and small prey. In John H. Steele (editor). Spatial pattern in plankton communities, p. 329-354. Plenum Publ. Corp. Laurence, G. C. 1971. Digestion rate of larval largemouth bass. New York Fish Game J. 18(l):52-56. 1975. Laboratory growth and metabolism of the winter flounder Pseudopleuronectes americanus from hatching through metamorphosis at three temperatures. Mar. Biol. 32:223-229. 1977. A bioenergetic model for the analysis of feeding and survival potential of winter flounder, Pseudopleuronectes americanus, larvae during the period from hatching to meta- morphosis. Fish. Bull., U.S. 75:529-545. 1979. Larval length-weight relations for seven species of northwest Atlantic fishes reared in the laboratory. Fish. Bull., U.S. 76:890-895. Leong, R. 1971 . Induced spawning of the northern anchovy, Engraulis mordax Girard. Fish. Bull., U.S. 69:357-360. LOVEGROVE, T. 1966. The determination of the dry weight of plankton and the various factors on the values obtained. In H. Barnes (editor). Some contemporary studies in marine science, p. 429-467. Hafner Publ. Co., N.Y. Methot, R. D., and D. Kramer. 1979. Growth of northern anchovy, Engraulis mordax, lar- vae in the sea. Fish. Bull, U.S. 77:413-423. Moffatt, N. M. 1981. Survival and growth of northern anchovy larvae on low zooplankton densities as affected by the presence of a Chlorella bloom. Rapp. P.-v. Reun. Cons. int. Explor. Mer 178:475-480. Moksness, E. 1982. Food uptake, growth and survival of capelin larvae (Mallorus villosus muller) in an outdoor constructed basin. Fiskeridir. Skr. Ser. Havunders. 17:267-285. O'Connell, C. p., and Pedro A. Paloma. 1981 . Histochemical indications of liver glycogen in samples of emaciated and robust larvae of the northern anchovy, Engraulis mordax. Fish. Bull, U.S. 79:806-812. O'Connell, C. P., and L. P. Raymond. 1970. The effect of food density on survival and growth of early post yolk-sac larvae of the northern anchovy (En- graulis mordax Girard) in the laboratory. J. Exp. Mar. Biol. Ecol. 5:187-197. Pearcy, W. G., G. H. Theilacker, and R. Lasker. 1969. Oxygen consumption of Euphausia pacifica: The lack of a diel rhythm or light-dark effect, with comparison of ex- perimental techniques. Limnol. Oceanogr. 14:219-223. Stauffer, G. D. 1973. A growth model for salmonids reared in hatchery envi- ronments. Ph.D. Thesis, Univ. Washington, Seattle. Theilacker, G. H., and K. Dorsey. 1980. Larval fish diversity. In Workshop on the effects of environmental variation on the survival of larval pelagic fishes, p. 105-142. Intergov. Oceanogr. Comm. Rep. No. 28. UNESCO, Paris. Theilacker, G. H., and A. S. Kimball. 1984. Comparative quality of rotifers and copepods as foods for larval fishes. Calif. Coop. Oceanic Fish. Invest. Rep. 15:80-86. 227 FISHERY BULLETIN: VOL. 85, NO. 2 Theilacker, G. H., and M. F. McMaster. tality of larval fish. J. Fish. Res. Board Can. 32:2503- 1971. Mass culture of the rotifer, Brachionus plicatilis, and 2512. its evaluation as a food for larval anchovies. Mar. Biol. Werner, R. G., and J. H. S. Blaxter. 10: 183-188. 1980. Growth and survival of larval herring (Clupea harengiis) Vlymen, W. in relation to prey density. Can. J. Fish. Aquat. Sci. 37: 1977. A mathematical model of the relationship between lar- 1063-1069. val anchovy {E. mordax), growth, prey microdistribution and Zweifel, J. R., and R. Lasker. larval behavior. Environ. Biol. Fishes 2:211-233. 1976. Prehatch and posthatch growth of fishes— a general Ware, D. M. model. Fish. Bull., U.S. 74:609-621. 1975. Relation between egg size, growth and natural mor- 228 THIRTY-FOUR SPECIES OF CALIFORNIA ROCKFISHES: MATURITY AND SEASONALITY OF REPRODUCTION Tina Wyllie Echeverriai ABSTRACT The viviparous rockfishes (Sebastes spp.) differ among species in age and size at maturity, and in the timing of peal< spermatogenesis, fertihzation, and larval extrusion. Age at 50% maturity ranges from 2 years in S. jordani to 9 years in S. diploproa. Within species, males usually mature either at the same age and size as females or at a younger age and smaller size. Rockfishes have two major seasons of lar- val extrusion, winter (November-March) or spring (April-July). The reproductive season for a particular species will fall within one of the major seasons throughout its geographic range. Within the major season, annual variations in the peak month of larval extrusion was observed for individual species. A long reproductive season and variations in the annual timing of that season are evidence of plasticity in the reproductive biology of rockfishes. Reproductive development at the cellular level was compared with the coincident changes in the gross morphology of the gonads. The resulting description of the developmental sequences of the testes and ovaries enables the determination of maturity stage in the field. Reproductive parameters such as age and size at maturity have been shown to be adaptive character- istics and are responsive to external pressures. For example, reduced population size due to fishing pres- sure may be associated with increased growth rate, reduced age at maturity, decreased fecundity, or a change in the gonadal index (Adams 1980; Gunder- son 1980). For haddock, Melanogrammus aeglefinis, age at maturity was reduced and growth rates in- creased as the fishery increased (Templeman and Bishop 1979; Beacham 1983). Clupeoids shifted spawning location or time, and reduced age at maturity (Murphy 1977; Blaxter and Hunter 1982). A study of depleted populations of Pacific mackerel, Scomber japonicus, suggested a direct relationship between population size and age at maturity (Par- rish and MacCall 1978). Pacific halibut, Hippoglossns stenolepis, stocks also showed reduced age at matur- ity and increased growth rates with reduced popu- lations (Schmitt and Skud 1978). Age at maturity may thus be a useful indicator of heavy fishing mor- tality. Rockfishes exhibit a variety of life history pat- terns, but only a few species have been studied in detail (Chen 1971; Miller and Geibel 1973; Patten 1973; Moulton 1977; Larson 1980; Love and West- phal 1981; McClure 1982). Previously, the most com- prehensive work on the maturity of rockfishes was ^Southwest Fisheries Center Tiburon Laboratory, National Marine Fisheries Service, NOAA, 3150 Paradise Drive, Tiburon, CA 94920. by Phillips (1964) and Westrheim (1975). Phillips (1964) sampled market landings from northern and central California over several years. For each of the 10 species he investigated, maturity was re- ported for the sexes combined, and ages at matur- ity were derived from back-calculated von Berta- lanffy growth curves. Westrheim (1975) summarized 10 years of data gathered on trawl-caught fish off British Columbia and the Gulf of Alaska, for which he reported on size at maturity and reproductive seasonality. Most of the fish in his study are com- mercially important species for British Columbia and, except for three species, do not occur off California. There are some difficulties in assessing maturity stages and reproductive seasonality in rockfishes. One problem is the use of external morphology of gonads to determine maturity stages. Some re- searchers have questioned the accuracy with which immature fish can be distinguished from resting, mature fish during the nonreproductive months (Gunderson et al. 1980; Rosenthal et al. 1982). Also, the potential reproductive season may be protracted and the peak time of reproduction may shift within this season, so that short-term studies may be mis- leading. Thus, reported variation in length and age at maturity between studies could be the result of uncertainty in maturity-stage determinations. In this study, I clarify the determination of sexual maturity stages, determine age and size at sexual maturity, and survey the reproductive seasonality for 34 species of rockfishes from the waters off Manuscript accepted February 1987. FISHERY BULLETIN: VOL. 85. NO. 2, 1987. 229 FISHERY BULLETIN: VOL. 85. NO. 2 northern and central California, from Port San Luis to Crescent City. I outline the reproductive patterns among species and annual variation within species. MATERIALS AND METHODS Data for this study were collected between July 1977 and July 1982 from three sources: 1) a coast- wide survey for adult rockfish made in July and August 1977 (Gunderson and Lenarz 1980); 2) an ongoing cooperative program, initiated in 1977 by the California Department of Fish and Game and the National Marine Fisheries Service, to sample the commercial (Sen 1984) and sport rockfish landings in northern and central California; and 3) a 1980 ex- pansion of the cooperative program, to include data on gonad condition, prey items, seasonal fluctuation of interstitial fat, and gonad volumes. Additional col- lections were taken during cooperative survey trips with the National Marine Fisheries Service, South- west Fisheries Center, and the California Depart- ment of Fish and Game. Fish were collected to sup- plement rarely sampled species and subadults that were not well represented in the three surveys. Information on maturity was gathered primarily for seven species of commercial and sportfishing im- portance. Data for 27 additional species are pre- sented, but these were inadequately sampled for statistical treatment of the data (Table 1). For each fish sampled from 1977 to 1982 the species, sex, and total length (mm) was determined in the field and otoliths were collected for age determination. The viscera from each of the 16,444 fish sampled from 1980 to 1982 were removed, preserved in 10% buf- fered formalin, and sent to the laboratory for anal- ysis. In the laboratory the sex was verified and the maturity stage of the gonads was determined based on the criteria defined in this paper. Total length was measured from the most anterior part of the jaw to the dorsal tip of the caudal fin. When neces- sary, total lengths were converted to fork length to compare this study with others (Echeverria and Lenarz 1984). Ages were determined for all fish investigated in this study. Most estimates of age were made from the surface of whole otoliths immersed in 70% ethyl alcohol, under a dissecting microscope at 12 x . Stan- dard techniques for counting annuli on whole otoliths were followed (Kimura et al. 1979; Shaw and Archibald 1981). Certain species such as 5. diplo- proa required thin sectioning techniques for count- ing annuli (Beamish 1979). In some species it is dif- ficult to determine what constitutes an annulus (Table 1), and consequently those ages have not been validated. The "break-and-burn" method of age determination (Chilton and Beamish 1982) is useful and more accurate than surface ages when aging fish older than 16 years (Tagart 1984). Ages were not redetermined in this study because the age at maturity occurs before 16 years. Maximum ages, however, may be underestimated. The gonad conditions described for males are im- mature, maturing, mature (peak spermatogenesis), spent, and resting. For females, they are immature, maturing, mature (fertilized), ripe (eyed larvae), recently spent, and resting (Lyubimova 1965; West- rheim 1975; Gunderson et al. 1980). Histological sec- tions were examined to define seasonal maturation and the sections were analyzed based on the criteria established by Moser (1967b) for S. paucispinis and by Lisovenko (1970) on 5. alutus for mature fish. The development of germ cells into spermatozoa or mature oocytes was examined to determine the developmental sequence leading to maturity. Cel- lular development was compared with the external morphology (Westrheim 1975; Gunderson et al. 1980) to understand clearly the developmental se- quence and to aid in the interpretation of maturity stages in the field. Gonads were subsampled from fish in all maturity stages. Whole gonads were sec- tioned and stained with hematoxylin. Histological sections were examined from 519 testes and 708 ovaries from 30 species (Table 1). Egg diameters were routinely measured from histological sections and checked against whole eggs using an ocular micrometer. To determine age and size at maturity and repro- ductive seasonality, the approximate age and size when 50% and 100% of the males and females were mature were estimated for each species. For those commercial species for which large samples were available, I checked the accuracy of the rough ma- turity estimates by applying the method of Gunder- son et al. (1980). These authors used the logistic model 1 + e"*-^* where P^ = proportion mature at size x, and a and b are constants, to estimate the length at 50% maturity for a sample. I applied a transformation of this equation In 1 I = ax + b to obtain estimates of a and b for size and age by 230 WYLLIE ECHEVERRIA: REPRODUCTION OF CALIFORNIA ROCKFISHES Table 1.— Sample characteristics used to determine age and size at maturity and seasonality for Sebastes from central California. Sampled Total number Hi stology Species of Age Size Immature Mature Males Females Sebastes (common name) (yr) (cm TL) (A/) (W) (A/) {N) alutus (Pacific ocean perch) 5-16 26-54 2 83 2 2 aurlculatus (brown rockfish) 1-19 10-52 140 281 16 20 aurora (aurora rocl:■:■:■:■: J 1 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec maturing | 1 spent 100 n B UJ O 50 CC lU Q. I Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec illlilliil fertilized maturing with eyed larvae spent Figure 4.— A. Reproductive seasonality of mature males otSebastes melanops during 1980-82. Each bar shows the percent of mature males sampled that were maturing (Gonad stage 3 + 4) and spent (Gonad stage 6 + 7). See Table 2 for further definition of stages. B. Reproductive seasonality of mature females of S. melanops during 1980-82. Each bar shows the percentage of mature females sampled that were maturing (Gonad stage 3), fertilized (Gonad stage 4), with eyed larvae (Gonad stage 5), and spent (Gonad stage 6 + 7). See Table 1 for further definition of stages. 237 FISHERY BULLETIN: VOL. 85, NO. 2 100n LU O 50 q: LU Q. 0 1 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec maturing [ [ spent 100 n B z LU O 50 cc LU a aa Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec fertilized maturing with eyed larvae spent Figure 5.— A. Reproductive seasonality of mature males ofSebastes mystinus during 1980-82. Each bar shows the percent of mature males sampled that were maturing (Gonad stage 3 + 4) and spent (Gonad stage 6 + 7). See Table 2 for further definition of stages. B. Reproductive seasonality of mature females of S. mystinus during 1980-82. Each bar shows the percentage of mature females sampled that were maturing (Gonad stage 3), fertilized (Gonad stage 4), with eyed larvae (Gonad stage 5), and spent (Gonad stage 6 -t- 7). See Table 1 for further definition of stages. 238 WYLLIE ECHEVERRIA: REPRODUCTION OF CALIFORNIA ROCKFISHES 100n O 50 QC UJ 0. *AA**ikHHMM*Mi* Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec ilij maturing I I spent 100 1 B Z liJ O 50 GC m Q. I'll M Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec maturing fertilized with eyed larvae spent Figure 6.— A. Reproductive seasonality of mature males of Sebastes paucispinis during 1980-82. Each bar shows the percent of mature males sampled that were maturing (Gonad stage 3 + 4) and spent (Gonad stage 6 + 7). See Table 2 for further definition of stages. B. Reproductive seasonality of mature females of S. paticipinis during 1980-82. Each bar shows the percentage of mature females sampled that were maturing (Gonad stage 3), fertilized (Gonad stage 4), with eyed larvae (Gonad stage 5), and spent (Gonad stage 6 -i- 7). See Table 1 for further definition of stages. 239 FISHERY BULLETIN: VOL. 85, NO. 2 100-1 LU O 50H cc LU Q. 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec maturing I I spent 100-1 B \- z LU O 50 a: LU Q. Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec maturing fertilized with eyed larvae spent Figure 7.— A. Reproductive seasonality of mature males ofSebastes pinniger during 1980-82. Each bar shows the percent of mature males sampled that were maturing (Gonad stage 3 + 4) and spent (Gonad stage 6 + 7). See Table 2 for further definition of stages. B. Reproductive seasonality of mature females of S. pinniger during 1980-82. Each bar shows the percentage of mature females sampled that were maturing (Gonad stage 3), fertilized (Gonad stage 4), with eyed larvae (Gonad stage 5), and spent (Gonad stage 6 + 7). See Table 1 for further definition of stages. 240 WYLLIE ECHEVERRIA: REPRODUCTION OF CALIFORNIA ROCKFISHES 2) has spermatogonia! cysts developing throughout the testis. There is no sign of spermatozoa in the sperm duct. Externally, the testes appear slightly swollen and white; in cross section they are translu- cent in the center because of the absence of sperm in the sperm duct. In a maturing male (Gonad Stage 3) spermatozoa appear throughout the testis in spermatozoan cysts. The sperm duct often contains remnants of residual sperm from the previous reproductive season. In a cross section the testis is swollen and whitish at the periphery and brownish off-white in the center because of the presence of residual sperm in the sperm duct. As the reproductive season approaches (Gonad Stage 4) the spermatozoan cysts burst open, re- leasing spermatozoa into the efferent ducts and the sperm duct. Externally the testis is large, soft, and very white, sperm flows freely when the testis is pressed or cut. During the reproductive season sper- matogenesis has ceased and the spermatozoa have moved from the periphery towards the sperm duct so that the periphery becomes hard and discolored. The central area may be swollen with sperm. At the end of the reproductive season (Gonad Stage 7), the testis undergoes resorption and re- organization, wherein smooth muscle cells, connec- tive tissue, and scattered residual spermatozoa, con- stituting cellular detritus, are evident in the histo- logical sections. At the periphery, a new generation of germ cells reorganizes along the spermatogen- ic tubules. Externally the testis is a compact, ir- regular triangular shape that appears gray or brown. An understanding of the developmental sequence on the cellular level aids in the interpretation of gonad stage in the field. The four, difficult to inter- pret, transitional stages can be clarified: 1) The first reproductive year is indicated by a white periphery (sperm) and an absence of sperm in the center of a cross section of testis, a thin ovarian wall and no residual pigmented eyes in ovaries. 2) The center of the prespawned testis is firm and a dark cream color (residual sperm); the periphery is white and swollen with sperm. The postspawned testis usual- ly shows signs of white fluid (sperm) in the sperm duct and is firm and a dark cream color at the peri- phery. 3) Unfertilized eggs are opaque yellow or white held tightly in grapelike clusters while fer- tilized eggs are a translucent yellow or white and the outer eggs can be separated from each other. 4) Vitellogenesis is indicated by a deep yellow color and swelling of the eggs so that the ovarian wall fits tightly around the eggs; spermatogenesis is in- dicated by a softening and swelling of the testis and a whitening at the periphery. Reproductive Maturity and Seasonality Maturity was observed over a broad age and size range within species throughout the years sampled. The age and size at first maturity, 50% maturity and 100% maturity were estimated for males and females of each species (Table 4). Males reached 50% maturity either at the same or younger age than females. Size at 50% maturity is generally similar or somewhat smaller for males than for females of the same species. The standard linear regressions were run on the transformed logistic for the seven principal species occurring in this study, resulting in similar estimates of age at 50% maturity for S. entomelas, S. flavidus, S. goodei, S. melanops, S. mystinus, S. paucispinis, and 5. pinniger (Table 5), and similar, if not exact, sizes at 50% maturity as estimates derived from the raw data. Maturity for species without sufficient data for statistical treat- ment are estimated from the raw data. The reproductive season in Sebastes can be long with larval extrusion (parturition) seen in females of some species for up to 9 months (Table 6). From all the data collected between 1977 and 1984 a sum- mary of principal month of spermatogenesis, fer- tilization, and parturition was determined for 32 species (Table 7). A span of 1 to 5 months between peak spermatogenesis and fertilization is seen. The time, when males ripen and mate, is not dependent upon the eggs being fully mature. The time between fertilization and parturition is usually about 1 month (Moser 1967a). Reproductive seasonality for the principal species sampled is displayed graphically in Figures 1-7. Seasonality histograms are available, upon request, for most species investigated. The general trend in the seasonality oi Sebastes is a prolonged reproduc- tive period for each maturity stage. This trend is seen in the seven most abundant species sampled (Figs. 1-7). Spermatogenesis (Gonad Stage 3) occurs over 3 to 5 months before the testes are fully ripe (Gonad Stage 4). The timing of mating is estimated from the appearance of testes swollen with sperm. At least part of the male population is ready for mating for a period of 2 to 4 months. In females, generally, fertilized eggs (Gonad Stage 4) are found 1 to 3 months after mating. Eyed larvae (Gonad Stage 5) appear from 1 to 4 months after fertilized eggs were observed and were present in the sam- pled population for 3 to 6 months, usually with a 241 FISHERY BULLETIN: VOL. 85. NO. 2 Table 4.— Estimated age and size at 1st, 50% and 100% maturity for given in cm Male Female Species of Sebastes 1st 50% 100% 1st 50% 100% yr TL yr TL yr TL yr TL yr TL yr TL alutus 5 28 7 32 7 26 7 26 7 32 auriculatus 3 26 5 31 10 38 3 26 5 31 10 38 aurora 5 28 — — 5 28 5 28 — — 5 28 babcocki 3 27 4 31 5 36 3 32 4 34 6 41 carnatus 4 17 4 17 5 21 4 17 4 17 5 21 caurinus 3 30 4 32 7 40 5 31 6 34 8 41 chlorostictus 4 25 6 27 12 38 5 26 6 28 9 34 chrysomelas 3 14 3 16 5 20 3 14 3 15 5 19 constellatus 6 28 7 30 12 36 5 23 6 27 9 34 crameri 3 25 4 27 7 36 3 24 4 27 6 34 diploproa 7 20 9 22 10 29 6 18 7 19 9 23 elongatus 7 23 7 23 10 27 5 18 7 23 10 27 entomelas^ 3 31 5 36 8 41 3 29 5 37 8 40 flavldus^ 4 30 6 35 11 43 4 27 7 36 11 42 goodei'^ 2 26 3 31 7 38 2 29 3 34 6 39 helvomaculatus 7 22 7 22 10 27 5 20 8 23 10 27 hopkinsi 4 15 5 16 5 17 5 17 5 18 7 21 'Denotes samples from 1980-82 study only. peak month of larval extrusion. The external ap- pearance of individual ovaries indicates that larvae are in the eyed stage and ready for release through- out the ovary. The total number of mature females observed dur- ing the reproductive months, and the percentage containing eyed larvae is presented for three com- mercially important species (Figs. 8-10). For the years 1981-85, variations in the months of parturi- tion and/or the peak month is seen. SebaMes ento- melas (Fig. 8) showed an annual variation in the months of parturition but not in the peak months (January-February). The peak month for S. goodei (Fig. 9) varied from December to February, and in iS. paucispinis (Fig. 10) from December to March. A variation in which months larval extrusion occurs is seen in all three species. Chi-square tests showed the percentages to be dependent upon year and month. Therefore, there is a relationship between the percent-number of mature females with eyed lar- vae seen in a particular month, and reproductive year. DISCUSSION The reproductive biology of Sebastes follows the sequence of spermatogenesis, vitellogenesis, mating, ovulation, fertilization, and larval extrusion (Moser 1967b). As in other \dviparous fishes (Turner 1947), sperm can apparently survive within the ovary for many months. In S. mentella, ovulation and the activation of sperm coincide with a change in pH (Sorokin 1967). The histological evaluation of the gonads in Sebastes of northern California confirms that spermatogenesis is generally completed and mating occurs before the completion of vitello- genesis. The length of time males are fully ripe can be up to 2 months in various species, and the delay between the time males are fully ripe and fertiliza- tion (sometimes up to 4 months) indicates that mating does not coincide with fully mature ova (Figs. 1-7). Testes are observed in decreasing degrees of ripeness after spermatogenesis has ceased. This indicates that one mating does not void the testes; males may mate more than once per season. Ovulation in teleosts is regulated by steroids and prostaglandins, which in turn are influenced to some degree by temperature, pheromones, or spatial/temporal cues (Stacey 1984). The presence of sperm in the ovaries of Sebastes does not trigger final oocyte maturation and ovulation; therefore, ovulation is probably influenced by environmental conditions. Thus knowing which conditions influence ovulation is necessary to determine some of the fac- tors contributing to a successful reproductive year, as measured by the percent mature females vdth eyed larvae. The costs of reproduction for females include the development of a highly vascular network through- out the ovaries, the nourishing of developing em- bryos (Boehlert and Yoklavich 1984), and the metab- olism of waste products from the embryos (Moser 1967a). The ovaries must accommodate the added gonadal weight and volume until larval extrusion. 242 WYLLIE ECHEVERRIA: REPRODUCTION OF CALIFORNIA ROCKFISHES Sebastes from northern and central California collected 1977-82. Lengths total length (TL). Male Female Species of 1st 50% 100% 1st 50% 100% Sebastes yr TL yr TL y TL yr TL yr TL yr TL jordani 1 12 2 14 5 20 2 12 3 14 4 19 levis 4 32 4 32 4 32 4 32 4 32 5 37 maliger 4 22 4 22 7 31 6 26 6 26 7 28 melanops^ 3 25 6 36 10 43 5 30 7 41 11 48 melanostomus 6 29 7 33 10 39 7 30 8 35 9 36 minlatus 5 35 5 38 8 43 5 37 5 37 9 46 mystinus^ 4 22 5 27 9 32 5 22 6 29 11 35 nebulosus 3 26 4 27 6 30 3 26 4 27 6 30 ovalis 4 28 4 28 5 30 4 28 4 28 5 29 paucispinis'' 3 32 3 42 7 55 3 36 4 48 8 60 pinniger^ 4 28 7 40 9 45 4 27 9 44 13 54 rosaceus 4 16 6 20 8 25 4 15 6 20 8 25 ruberhmus 6 36 7 40 8 46 6 36 7 40 8 46 rubhvinctus 5 29 5 30 6 32 5 29 8 34 10 38 rufus — 27 3 31 4 36 2 32 3 34 4 41 saxicola 3 15 3 16 4 17 2 17 2 17 3 18 serranoides 4 32 5 33 8 38 4 32 5 35 8 39 Table 5.— Standard linear regression used to estimate age (yr) and size (cm) at 50% maturity for males (M) and females (F) in seven species of Sebastes. Slope and y-intercept were estimated from the equation In ^ = ax + b. r = correlation coefficient, N = number of ages or sizes used in the regressions; each N represents at least 10 observations per age or size. Species of Sebastes Sex Variable y-intercept Slope 50% maturity r N entomelas M age size 4.5598 10.1042 -1.0006 -0.3162 5 33 -0.9588 -0.7649 6 11 F age size 5.4800 28.4810 -0.7982 -0.7810 5 37 -0.9647 -0.9667 5 4 flavidus M age size 2.9748 13.7834 -0.4994 - 0.3964 6 35 -0.9577 -0.9718 10 14 F age size 5.3731 15.7482 -0.7757 -0.4331 7 36 -0.9541 -0.9723 9 13 goodei M age size 1.1620 7.7618 - 0.4093 -0.3044 3 31 -0.9263 -0.8730 9 10 F age size 3.7624 12.2400 -1.1129 - 0.3740 3 34 - 0.9775 - 0.9650 6 15 melanops M age size 4.3841 13.7147 -0.8011 -0.3814 6 36 -0.9769 -0.9411 6 6 F age size 5.0627 1 1 .0749 -0.7149 - 0.2720 7 41 -0.9760 -0.9520 8 16 mystinus M age size 1 .3482 4.3885 -0.3669 -0.2049 5 27 -0.8530 -0.6778 10 11 F age size 5.0097 9.5090 - 0.8022 - 0.3368 6 29 -0.9680 -0.9218 9 17 paucispinis M age size 1.7138 10.7040 - 0.6677 -0.2564 3 42 -0.9431 -0.9196 8 16 F age size 3.5175 13.7028 -0.8232 -0.2876 4 48 -0.9232 - 0.9565 7 20 pinniger M age size 3.7804 1 1 .2047 - 0.5845 -0.2837 7 40 -0.9710 - 0.9500 9 16 F age size 5.1221 1 1 .5621 - 0.5897 -0.5897 9 44 -0.9800 -0.9460 11 15 243 FISHERY BULLETIN: VOL. 85, NO. 2 Table 6.— Months of parturition for species of Sebastes thiat occur in the northeastern Pacific Ocean. All available data are listed by area (results of this study in parentheses). Species of Gulf of British North-Central Southern Sebastes Alaska Colunfibia Washington Oregon California California aleutianus Apr.' May^ alutus May^, Apr.-May^ Mar.3 Mar." Jan. -Apr. ^ (Jan. -Mar.) auriculatus June^ May^ (Dec. -Jan.; May-July), May'' aurora >June' (Mar. -May) babcocki Apr.' Apr.2 (May) borealis Apr.' brevisplnis >May' carnatus (Mar. -May), May'' caurinus Mar.«, Apr.sss (Feb.) chlorostictus Apr.2 (Apr.-Sept.) Apr.-July'°, July" chrysomelas (Feb. -Mar.), Jan. -May'' constellatus (Apr. -May) Mar. -1- May'°" crameri June' Apr.-May'2 (May-July) ensifer Feb. + May'° entomelas Apr.' Jan.2'" (Dec. -Apr.), Nov.-Mar.'^ eos Apr. -June" fl avid us Mar.' Jan.-Apr."* Jan.-Mar.^ (Jan.-July), Nov.-Mar.'^ goodei (Nov.-June), Nov.-Mar.'^ Oct.-Mar." helvomaculatus >May' >May' (May-June) hopkinsi (Feb.-Mar.) jordani Feb.2 (Feb.-Apr.), Nov.-Mar.'^ lentiginosus Mar.'° levis (Dec. -Feb.) Dec. -Jan." maliger May-July^5 Apr.''6 , May^ (Apr. -July) melanops Jan.'^ Feb.' Dec. -Apr. ^ (Dec. -Mar.), Nov.-Mar." proriger Apr-July^ (July-Aug.) reedi May' Feb.-May"^ rosaceus (Apr. -July) Mar. -h May'°'" ruberrimus June-Aug.'^ May' July^ Mar. -Apr. ^, Apr.-May'^ (Apr. -June) rubrivinctus May' Apr.-May'2 (July) Mar.-June'° rufus (Dec. -May) saxicola Feb.' Feb.2 (Jan.-Mar.), Nov.-Mar.'^ serranoides (Jan. -Mar.) Jan.'« simulator Feb.-Mar.'° umbrosus Apr.'° wilsoni June' zacentrus >May^ July' Mar.-July^, Apr.-May'2 (May- June) iWestrheim 1975. 2W. H. Barss, Oregon Department of Fish and Wildlife, Marine Science Dnve, Newport, OR 97365, pers. commun. 1985. ^Lyubimova 1965. ^Gunderson et al. 1980. ^Washington et al. 1978. •^Delacy et al. 1964. 'Larson 1980. apatten 1973. ^IVloulton 1977. 'oChen 1971. "Moser 1967a. '2Hltz 1962. '^Phillips 1964, '"Barss and Echeverria 1987. 'SRosenthal et al. 1981. '^Love and Westphal 1981. "Dunn and HItz 1969. lewales 1952. '^Miller and Gelbel 1973. ^oBurge and Schultz 1973. 244 WYLLIE ECHEVERRIA: REPRODUCTION OF CALIFORNIA ROCKFISHES Table 7.— Reproductive seasonality for Sebastes from central California collected 1977-84. Listed by taxonomic order (Barsukov 1981). Principal month(s) of Species of Sebastes spermato- genesis fertil- ization parturi- tion melanostomus Nov. Feb. aurora Apr. ? ruberrimus Dec. Apr. chrysomelas ? Jan. carnatus Dec. ? nebulosus 7 Jan. auriculatus l\/lay May caurinus Dec. Jan. maliger Dec. -Jan. Apr. elongatus ? Apr. babcocki Jan. 7 saxicola ? Dec. diploproa June June crameri ? Dec. alutus Sept.-Oct. 7 pinniger Oct. Dec. miniatus July 7 levis Sept.-Oct. Oct. rosaceus May May constellatus Dec. Feb. chlorostictus Feb. Apr. goodei Oct. Dec. jordani Nov. Jan. paucispinis Oct. Dec. oval is Nov. 7 rufus Nov. Nov. hopkinsi Dec. Feb. entomelas Oct. Jan. mystinus Aug. Jan. melanops Oct. Jan. fl avid us Sept. Jan. serranoides Nov. Jan. Feb. Jan. Feb. Feb. Feb. Females are exposed to changing environmental fac- tors from year to year, so flexibility in the timing of their greatest reproductive involvement may be advantageous. The apparent flexibility of the period between mating and larval extrusion may be a mech- anism to optimize reproductive success. A long and variable period of larval extrusion is exhibited by the rockfish group (Figs. 8-10). The evidence of two spawnings per season have been reported for 5. paucispinis (Moser 1967a), S. ovalis, and S. con- stellatus (MacGregor 1970). Multiple broods are in- dicated by the presence of eyed larvae undergoing resorption in the ovary concurrently with vitelloginic eggs at least 0.4 mm in diameter. During the years and throughout the area of this study, evidence of multiple broods was rare: only one 5. paucispinis gonad showed evidence of a multiple brood. Moser' s (1967b) detailed description of the histology of multi- ple broods indicates the development and extrusion of a second brood follows within 2 months of the 40 30 20 10 O Feb. lU 60 Apr. < June > cc SO Feb. < 40 Mar. -J 1 30 Jan. a June LU 20 Feb. >■ LLI 10 Apr. X 0 May H May Jan. ^ 50 July m 40 Jan. -1 Mar. < 20 Dec. 10 Sept. u. Dec. LU OC June 3 Apr. h- May < 40 Jan. zo Feb. z 20 Feb. UJ May O CC 10 Feb. LU o Mar. 0. 40 30 20 10 o 1981 44 147 128 83 WM^ " J2^_ DEC JAN FES MAR NOV DEC NOV DEC JAN FEB MAR APR 1982 0 183 278 i 332 1 VI 1 1 1 1 1 1 1983 g ■ 24 = i? ^ 334 1 • 1984 CD 20 137 1 130 1 300 280 MAR APR 1985 14 255 15 1 I 1 198 54 NOV DEC JAN FEB MAR APR Sebastes entomelas Figure 8.— Percent of mature female Sebastes entomelas contain- ing eyed larvae for the reproductive months during 1981-85. Total number of mature females observed are indicated over bars. first. In this study, two distinct seasons of larval ex- trusion, December and June, were noticed in 5. auriculatus sampled from a limited area (Pt. Reyes- Half Moon Bay) (Table 6) in north-central Califor- nia. The entire reproductive sequence in Sebastes seems to reflect a plasticity for reproduction that may enable the individual to respond adaptively to environmental factors. The consequences of the reproductive biology in Sebastes include the absence of strong spawning pulses, the possibility of multiple mating of males, and flexibility in the timing of fertilization. Ap- parently mating is not restricted to a short period 245 FISHERY BULLETIN: VOL. 85, NO. 2 30 ♦O 30 20 10 lU 0 < > QC < 30 _J 20 a LU 10 >- 0 LU I h- SO ^ 40 LU 30 -1 < 20 2: LU 10 U. 0 LU CC 3 H 30 < 20 1- 10 Z LU O oc LU 50 a 40 30 20 10 0 1981 37 79 145 1 168 I 92 NOV DEC 1982 20 29 OCT NOV DEC JAN FEB MAR APR 1983 10 0 116 1 86 1 13 i 1 84 DEC JAN 1984 320 178 ■ 207 206 116 1 i 126 105 OCT NOV DEC JAN FEB MAR APR 226 1985 619 189 525 ^ i 189 339 MAR APR Sebastes goodei Figure 9.— Percent of mature female Sebastes goodei containing eyed larvae for the reproductive months during 1981-85. Total number of mature females observed are indicated over bars. lU < > QC < -I I Q LU > Ui m u. ui oc I- < tu o QC LU a 40 T 30 20 ■■ 10 ■■ o 40 30 20 10 O 40 30 20 10 O 40 30 20 10 O 50 40 30 20 10 O 1981 118 1982 16 123 85 31 96 14 13 1983 • 0 70 71 109 48 ^ 1984 . 53 127 147 m 78 77 1 19 85 77 1 55 P i 39 21 AK Sebastes paucispinis Figure 10.— Percent of mature female Sebastes patuyispinis con- taining eyed larvae for the reproductive months during 1981-85. Total number of mature females observed are indicated over bars. of time and elaborate displays to induce ovulation are not necessary. It is not surprising, therefore, to find so little information regarding mating be- havior in the most speciose and populous group of fishes in north-central California (Helvey 1982). Ages and sizes at maturity from this study gen- erally agree with findings in the literature, but some- times exhibit large discrepancies (Table 8). The variability has possible sources: differences in age determination techniques, length measurements, identification of the immature gonad stage, or sam- pling at times of the year when it is difficult to distinguish immature from resting fish. Previous studies on maturity often focused on adult popula- tions or were based on samples taken only during summer, when most species are not in reproductive condition (Westrheim 1975; Gunderson et al. 1980; Rosenthal et al. 1981, 1982). The differentiation of an immature gonad stage versus mature "resting" gonads is most difficult and subject to error during the nonreproductive months. Age and size at 50% maturity by area and year were calculated for S. Jlavidus, S. goodei, and S. paucispinis (Table 9). Age and size at maturity did not differ geographically but did vary between years. Age varies up to 1 year and size varies up to 3 cm. Apparently, there is some variability in age and size at maturity within a population between years. The normal range of variation of size and age 246 WYLLIE ECHEVERRIA: REPRODUCTION OF CALIFORNIA ROCKFISHES Table 8.— Age and size at 50% maturity for male (M) and female (F) Sebastes by area. All lengths are converted to fork length for comparison (Echeverria and Lenarz 1 984). Species of Sebastes Alaska British Columbia Washington Oregon California this study auriculatus M F entomelas M F flavidus M goodei M F melanops M mystinus M F pauclspinis M F pinniger M F { 11-15 yr^ 41-45 cm FL 10-13 yr, 40-42 cm FL^ 9-12 yr, 39-42 cm FL 37 cm FL^ 38 cm FL 40 cm FL, 40.5 cm FL^'^ 42 cm FL, 45 cm FL <57 cm FL^ <62 cm FL 41 cm FL^'^ 48 cm FL 4 yr, 23 cm FL^ 4 yr, 25 cm FL 3 yr, 21 cm FL'' 3-4 yr, 33 cm FL 5 yr, 45 cmFL^ 6 yr, 48 cm FL 4 yr, 33 cm FL^ 7 yr, 38 cm FL 26 cm FL^ 37 cm FL 5 yr^ 6 yr 3yr^ 4 yr 12 yr, 39 cm FL^ 10 yr, 43 cm FL 4 yr, 29 cm FL" 5 yr, 31 cm FL" 4 yr, 28 cm FL" 4 yr, 30 cm FL 6-7 yr^ 6 yr [4yr" 1 37 cm FL 15-6 yr" (33 cm FL 5 yr, 31 cm FL 5 yr, 31 cm FL 5 yr, 32 cm FL 5 yr, 36 cm FL 6 yr, 32 cm FL 7 yr, 33 cm FL 3 yr, 29 cm rL 3 yr, 32 cm FL 6 yr, 35 cm FL 7 yr, 40 cm FL 5 yr, 26 cm FL 6 yr, 28 cm FL 3 yr, 39 cm FL 4 yr, 45 cm FL 7 yr, 39 cm FL 9 yr, 43 cm FL 'Washington et al. 1978; ^Westrheim 1975; ^Barss and Echeverria 1987; ^Phillips 1964; ^Rosenthal et al. 1982; sQunderson m al. 1980; 'Barker 1979; sMcClure 1982; sMiller et al. 1967. at maturity should be determined before subtle shifts in these parameters are studied. Later spawning in high latitude populations oc- curs in some teleosts as a response to temperature and photoperiod (Wootton 1984). The existence of geographical trends for spawning months can be determined for a species if data exist for the same months and years. Parturition occurred somewhat earlier in the southern end of the range of S. Table 9. — Age and size of three Sebastes species at 50% matur- ity derived from linear regressions for males (f^) and females (F) by area and by year for the area betv/een Crescent City and Morro Bay. S. flavidus S. goodei S. paucispinis Data base Sex yr cm TL yr cm TL yr cm TL North of M 4 3 43 Point Arena F — — 4 — 4 48 South of M — 4 31 3 43 Point Arena F — — 4 33 4 47 April 1980- 1^ 6 35 4 31 3 43 March 1981 F 7 37 4 34 4 48 April 1981- M 7 38 5 33 3 42 March 1982 F 6 36 4 32 4 47 April 1980- M 6 35 4 31 3 43 Mach 1982 F 7 36 4 34 4 48 maliger, S. ruberrimus, and S. entomelas. Sebastes maliger extruded larvae between May and July off Alaska (Rosenthal et al. 1981, 1982) and between April and July off California during 1982. Sebastes ruberrimus extruded larvae between June and August off Alaska (Rosenthal et al. 1981, 1982) and between April and July off California during 1982. Sebastes entomelas extruded larvae in February off Oregon and in January off California in 1982 (Barss and Echeverria 1987) (Table 6). For species where comparable data exist, parturition is earlier in the southern end of the species range. Reproductive seasonality can be classified into one of two broad seasons, early (winter) or late (spring- summer) (Phillips 1964), which seems to hold true throughout a species range (Table 6). The duration of larval extrusion varies from 1 to 9 months and is species-specific. Closely related species have similar seasons of parturition, but the peak month may differ (Table 7). Data on parturition may be useful when investi- gating recruitment by estimating annual reproduc- tive success. To predict year-class strength for species of Sebastes, it must be possible to identify species in the juvenile stages. Juvenile identifica- tions have been described for 18 species of rockfish 247 FISHERY BULLETIN: VOL. 85, NO. 2 that occur off California (Moser and Ahlstrom 1978; Richardson and Laroche 1979; Laroche and Richard- son 1980, 1981; Anderson 1983); the difficulties of differentiating between similar species are evident in these studies. Knowledge of the principal month of parturition may be a useful tool when identify- ing species in the age 0 population. Principal month of parturition would have to be documented for the year that recruitment is investigated. Life history parameters that are interrelated, such as growth, maturity, and fecundity, can be influ- enced by external conditions, such as temperature, prey abundance, and predation (Stearns and Cran- dall 1984). The manner in which they can be affected is species-specific and may change in a predictable manner. Observed changes that coincide with re- duced population sizes include increase in growth rates (Templeman and Bishop 1979), decrease in age at maturity (Murphy 1977; Parrish and MacCall 1978; Schmitt and Skud 1978; Templeman and Bishop 1979), and decrease in size at maturity (Aim 1959; Pitt 1975). Stearns and Crandall (1984) pro- posed that changes in age and/or size at maturity are determined by genetic as well as environmental factors, so that populations will respond in a predict- able manner. A shift in either the age or size at maturity in Sebastes may be an indication of a change in population densities. In order to detect any population changes, age and size at maturity for species should be determined yearly and within a well-defined geographic area. Fish of the esti- mated age and size at first maturity should be included— sampling from market fish tends to yield only mature fish. Ages should be determined from the same fish that are sampled for maturity. Fecundity in poikilotherms is generally related to size; changes in growth rates and size at maturity will affect fecundity. Fecundity often relates more to body size in short-lived species and to available energy in long-lived species (Ware 1980). Fecundity increases with size in at least some species of Sebas- tes (Phillips 1964), but annual reproductive success may be linked to available energy. Gonad volumes of female S. flavidus were reported for 1981 (Guille- mot et al. 1985) and compared with volumes mea- sured during the El Nifio winter of 1983-84 (Lenarz and Wyllie Echeverria 1986); this comparison showed reduced gonad volumes in 5. flavidus for the 1983 reproductive season. Whether the decreased gonad volume was due to egg size or number was not determined. Shifts in age and/or size at maturity may occur in species that have a multigenerational, late-matur- ing population. In his studies of flatfish populations. Roff (1982) predicted that size at maturity would be primarily influenced by size-dependent mortality and that changes in size at maturity would occur in species where growth to a minimal size is more adap- tive than early reproduction. Changes in age at maturity will more likely occur in species that mature early. Changes in size, rather than age, at maturity would most likely occur in Sebastes sub- jected to overfishing or long-term environmental stress. General changes in life history parameters may be predictable according to a species' position on the r-/iC selection continuum. Increased fishing mortality resulting in decreased populations may affect life history parameters by increasing growth rates, re- ducing age at first maturity, increasing fecundity at age (Adams 1980; Gunderson 1980), and reducing variability in the gene pool by reducing the number of spawning groups in the more /C-selected species (Leaman and Beamish 1984). The reproductive strategy of Sebastes reflects more A'-type character- istics, which include later maturity, slower growth rates, lower individual fecundity, or some degree of parental care (Garrod and Horwood 1984). The K- type reproductive strategy enables a species to minimize the effects of a poor reproductive year (Roff 1984). A disadvantage for a heavily fished K- type species is the late age at maturity, as exists in Sebastes, so that the advantage of many reproduc- tive seasons must be balanced against adult popu- lation size to obtain an allowable harvest. The reproductive strategy of Sebastes, with multi- ple generations reproducing simultaneously and the plasticity of annual timing, results in a buffered system. The populations of exploited stocks of Sebastes should be able to recover from a single year of high mortality due either to poor recruitment or to adult mortality. However, overfished populations or long periods of poor recruitment could result in a reduced size at maturity and a corresponding re- duced fecundity. ACKNOWLEDGMENTS Among the many people who contributed to this research, I wish to thank the biologists and seasonal samplers of the California Department of Fish and Game, the personnel who assisted me in the collec- tion and processing of these samples, including but not limited to Lisa Andrade, Todd Anderson, Patrick Guillemot, Steven Pace, Dana Tryde, Mickey Singer, Stuart Running, Marianne McKean, Karen Novak, Sonia Linnik, Michael King, Sally Krenn, Lisa Natanson, Kathleen Mathews, Laura 248 WYLLIE ECHEVERRIA: REPRODUCTION OF CALIFORNIA ROCKFISHES Young, and Jerry Kwiecien. I wish to thank William H. Lenarz for support and guidance, S. J. West- rheim and Ralph Larson for their insightful reviews, and Sandy Wyllie Echeverria and Rahel Fischer for editorial assistance. LITERATURE CITED Adams, P. B. 1980. Life history patterns in marine fishes and their conse- quences for fisheries management. Fish. Bull., U.S. 78:1-12. Alm, G. 1959. Connection between maturity, size and age in fishes. Inst. Freshwater Res. Drottningholm Rep. 40, 145 p. Anderson, T. W. 1983. Identification and development of nearshore juvenile rockfishes (genus Sebastes) in central California kelp forests. M.A. Thesis, San Francisco State Univ., San Francisco, CA, 216 p. Barker, M. W. 1979. Population and fishery dynamics of recreationaily ex- ploited marine bottomfish of northern Puget Sound. Diss. Abstr. Int. B Sci. Eng. 40(6):2497. Barss, W. H., and T. Echeverria. 1987. Marturity of widow rockfish {Sebastes entomelas) from the northeast Pacific 1977-1981. In W. H. Lenarz and D. R. Gunderson (editors), Proceedings of the Widow Rockfish Workshop, 11-12 December 1980, Tiburon, CA. U.S. Dep. Commer., NOAA-TM-NMFS-SWFC 48:13-18. Barsukov, V. V. 1981. A brief review of the subfamily Sebastinae. J. Ichthyol. 21:1-26. Beacham, T. D. 1983. Variability in size and age at sexual maturity of had- dock {Melanogramrnus aeglefinus) on the Scotian Shelf in the northwest Atlantic. Can. Tech. Rep. Fish. Aquat. Sci. 1168, 33 p. Beamish, R. J. 1979. New information on the longevity of Pacific ocean perch (Sebastes alutus). J. Fish. Res. Board Can. 36:1395- 1400. Blaxter, J. H. S., and J. R. Hunter. 1982. The biology of the clupeoid fishes. Adv. Mar. Biol. 20:1-223. BOEHLERT, G. W., AND M. YOKLAVICH. 1984. Reproduction, embryonic energetics, and the maternal fetal relationship in the viviparous genus Sebastes (Pisces: Scorpaenidae). Biol. Bull. (Woods Hole) 167:354-370. BURGE, R. T., AND S. A. SCHULTZ. 1973. The marine environment in the vicinity of Diablo Cove with special reference to abalones and bony fishes. Calif. Dep. Fish Game, Mar. Resour. Tech. Rep. 19, 433 p. CHEN, L. C. 1971. Systematics, variation, distribution, and biology of the subgenus Sebastomus (Pisces, Scorpaenidae, Sebastes). Bull. Scripps Inst. Oceanogr. 18, 115 p. Chilton, D. E., and R. J. Beamish. 1982. Age determination methods for fishes studied by the groundfish program at the Pacific Biological Station. Can. Spec. Publ. Fish. Aquat. Sci. 60, 102 p. DeLacy, a. C, C. R. Hitz, and R. L. Dryfoos. 1964. Maturation, gestation, and birth of rockfish (Sebastodes) from Washington and adjacent waters. Wash. Dep. Fish., Fish. Res. Pap. 2(3):51-67. Dunn, J. R., and C. R. Hitz. 1969. Oceanic occurrence of black rockfish (Sebastes mela- nops) in the central north Pacific. J. Fish. Res. Board Can. 26:3094-30^7. Echeverria, T., and W. H. Lenarz. 1984. Conversions between total, fork, and standard lengths in 35 species oi Sebastes from California. Fish. Bull., U.S. 82:249-251. GaRROD, D. J., AND J. W. Horwood. 1984. Reproductive strategies and the response to exploita- tion. In G. W. Potts and R. J. Wootton (editors). Fish reproduction: strategies and tactics, p. 367-384. Acad. Press, Lond. Guillemot, P. J., R. J. Larson, and W. H. Lenarz. 1985. Seasonal cycles of fat and gonad volume in five species of northern California rockfish (Scorpaenidae; Sebastes). Fish. Bull., U.S. 83:299-311. Gunderson, D. R. 1980. Using r-K selection theory to predict natural mortality. Can. J. Fish. Aquat. Sci. 37:2266-2271. Gunderson, D. R., P. Callahan, and B. Goiney. 1980. Maturation and fecundity of four species of Sebastes. Mar. Fish. Rev. 42(3-4):74-79. Gunderson, D. R., and W. H. Lenarz. 1980. Cooperative survey of rockfish and whiting resources off California, Washington, and Oregon, 1977: Introduction. Mar. Fish. Rev. 42(3-4): 1. Helvey, M. 1982. First observations of courtship behavior in rockfish, genus Sebastes. Copeia 1982:763-770. Hitz, C. R. 1962. Seasons of birth of rockfish (Sebastodes spp.) in Oregon coastal waters. Trans. Am. Fish. Soc. 91:231-233. Kimura, D. K., R. R. Mandapat, and S. L. Oxford. 1979. Method, validity, and variability in the age determina- tion of yellowtail rockfish, Sebastes flavidus, using otoliths. J. Fish. Res. Board Can. 36:377-383. Laroche, W. a., and S. L. Richardson. 1980. Development and occurrence of larvae and juveniles of the rockfishes Sebastes flavidus and Sebastes melanops (Scorpaenidae) off Oregon. Fish. Bull., U.S. 77:901-924. 1981. Development of larvae and juveniles of the rockfishes Sebastes entomelas and S. zacentrus (family Scorpaenidae) and occurrence off Oregon, with notes on head spines, of S. mystinus, S. flavidus, and S. melanops. Fish. Bull., U.S. 79:231-257. Larson, R. J. 1980. Competition, habitat selection, and the bathymetric segregation of two rockfish (Sebastes) species. Ecol. Monogr. 50:221-239. Leaman, B. M., and R. J. Beamish. 1984. Ecological and management implications of longevity in some northeastern Pacific groundfishes. Int. N. Pac. Fish. Comm. Bull. 42:85-97. Lenarz, W. H., and T. Wyllie Echeverria. 1986. Comparison of visceral fat and gonadal fat volumes of yellowtail rockfish, Sebastes flavidus, during a normal year and a year of El Nino conditions. Fish. Bull., U.S. 84:743- 745. Lisovenko, L. a. 1970. A study of spermatogenesis in Pacific rockfish Sebas- todes alutus (Gilbert) from the Gulf of Alaska. In Moiseev, P. A. (editor), Soviet fisheries investigations in the north- eastern Pacific, Part V, p. 248-266. Israel Program Sci. Transl. Jerusalem. 249 FISHERY BULLETIN: VOL. 85, NO. 2 Love, M. S., and W. V. Westphal. 1981. Growth, reproduction, and food habits of olive rockfish, Sebastes serranoides, off central California. Fish. Bull, U.S. 79:533-545. L\T?BIMOVA, T. G. 1965. Main stages in the life cycle of the rockfish Sebastodes alutus (Gilbert) in the Gulf of Alaska. Tr. VNIRO 49:85- 111. MacGregor, J. S. 1970. Fecundity, multiple spawning, and description of the gonads in Sebastodes. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 596, 12 p. McClure, R. E. 1982. Neretic reef fishes of central Oregon: aspects of life histories and the recreational fishery. M.S. Thesis, Oregon State Univ., Corvallis, OR, 94 p. Miller, D. J., M. W. Odemar, and D. W. Gotshall. 1967. Life history and catch analysis of the blue rockfish (Sebastodes mystiniis) off central California 1961-1965. Calif. Dep. Fish Game, MRO Ref. No. 67-14, 130 p. Miller, D. J., and J. J. Geibel. 1973. Summary of blue rockfish and lingcod life histories; a reef ecology study; and giant kelp, Macrocystis pyrifera, ex- periments in Monterey Bay, California. Calif. Dep. Fish Game, Fish Bull. 158, 137 p. Moser, H. G. 1 967a. Reproduction and development of Sebastodes paucis- pinis and comparison with other rockfishes off southern California. Copeia 1967:773-797. 1967b. Seasonal histological changes in the gonads of Sebas- todes paucispinis (Ayres), an ovoviviparous teleost (family Scorpaenidae). J. Morph. 123:329-351. Moser, H. G., and E. H. Ahlstrom. 1978. Larvae and pelagic juveniles of blackgill rockfish, Sebas- tes melanostomus, taken in midwater trawls off southern California and Baja California. J. Fish. Res. Board Can. 35:981-996. Moulton, L. L. 1977. An ecological analysis of fishes inhabiting the rocky near shore regions of northern Puget Sound, Washington. Ph.D. Thesis, Univ. Washington, Seattle, 181 p. Murphy, G. L 1977. Clupeoids. In J. A. Gulland (editor), Fish population dynamics, Ch. 12, p. 283-308. John Wiley & Sons, Lond. Parrish, R. H., and a. D. MacCall. 1978. Climatic variation and exploitation in the Pacific mack- erel fishery. Calif Dep. Fish Game, Fish Bull. 167, 110 p. Patten, C. G. 1973. Biological information on copper rockfish in Puget Sound, Washington. Trans. Am. Fish. Soc. 102:412-416. Phillips, J. B. 1964. Life history studies on ten species of rockfish (genus Sebastodes). Calif Dep. Fish Game, Fish Bull. 126, 70 p. Pitt, T. K. 1975. Changes in abundance and certain biological character- istics of Grand Bank American plaice, Hippoglossoides pLatessoidea. J. Fish. Res. Board Can. 32:1383-1398. Richardson, S. L., and W. A. Laroche. 1979. Development and occurrence of larvae and juveniles of the rockfishes Sebastes crameri, Sebastes pinniger, and Sebastes helvomaculatus (Family: Scorpaenidae) off Oregon. Fish. Bull., U.S. 77:1-46. ROFF, D. A. 1982. Reproductive strategies in flatfish; a first synthesis. Can. J. Fish. Aquat. Sci. 39:1686-1698. 1984. The evolution of life history parameters in teleosts. Can. J. Fish. Aquat. Sci. 41:989-1000. Rosenthal, R. J., L. J. Field, and D. Myer. 1981. Survey of nearshore bottomfish in the outside waters of southeastern Alaska. Prepared for Alaska Dep. Fish Game, Comm. Fish. Div., Juneau, AK, 91 p. Rosenthal, R. J., L. H. Haldorson, L. J. Field, V. Moran- O'CoNNELL, M. G. Lariviera, J. Underwood, and M. C. Murphy. 1982. Inshore and shallow offshore bottomfish resources in the southeastern Gulf of Alaska (1981-1982). Prepared for Alaska Dep. Fish Game, Comm. Fish. Div., Juneau, AK, 166 P- Schmitt, C. C, and B. E. Skud. 1978. Relation of fecundity to long-term changes in growth, abundance and recruitment. Int. Pac. Halibut Comm. Sci. Rep. 66, 31 p. Sen, a. R. 1984. Sampling commercial rockfish landings in California. U.S. Dep. Commer., NOAA-TM-NMFS-SWFC-45, 95 p. Shaw, W., and C. P. Archibald. 1981. Length and age data of rockfishes collected from B.C. coastal waters during 1977, 1978, and 1979. Can. Data Rep. Fish. Aquat. Sci. 289, 119 p. SOROKIN, V. P. 1967. Some data on gametogenesis and sexual cycles of the Pacific Ocean Scorpaenidae. Tr. PINRO 20:304-315. Stacey, N. E. 1984. Control of the timing of ovulation by exogenous and endogenous factors. In G. W. Potts and R. J. Wootton (editors), Fish reproduction: strategies and tactics, p. 207- 222. Acad. Press, Lond. Stearns, S. C, and R. E. Crandall. 1984. Plasticity for age and size at sexual maturity: a life- history response to unavoidable stress. In G. W. Potts and R. J. Wootton (editors). Fish reproduction: strategies and tactics, p. 13-33. Acad. Press, Lond. Tagart, J. V. 1984. Comparison of final ages assigned to a common set of Pacific ocean perch otoliths. Wash. Dep. Fish. Tech. Rep. 81, 36 p. Templeman, W., and C. A. Bishop. 1979. Sexual maturity and spawning in haddock, Melano- grammits aeglefinus, of St. Pierre bank. ICNAF Res. Bull. 14:77-83. Turner, C. L. 1947. Viviparity in teleost fishes. Sci. Mon. 65:508-518. Wales, J. H. 1952. Life history of the blue rockfish Sebastodes mystimis. Calif Fish Game 38:485-498. Ware, D. H. 1980. Bioenergetics of stock and recruitment. Can. J. Fish. Aquat. Sci. 37:1012-1024. Washington, P. M., R. Gowan, and D. H. Ito. 1978. A biological report on eight species of rockfish Sebas- tes spp. from Puget Sound, Washington. Northwest Alaska Fisheries Center, 50 p. [Processed rep.] Westrheim, S. J. 1975. Reproduction, maturation, and identification of larvae of some Seba.stes (Scorpaenidae) species in the northeast Pacific Ocean. J. Fish. Res. Board Can. 32:2399-2411. Wootton, R. J. 1984. Introduction: strategies and tactics in fish reproduction. In G. W. Potts and R. J. Wootton (editors). Fish reproduc- tion: strategies and tactics, p. 1-12. Acad. Press, Lond. 250 DISTRIBUTION, ABUNDANCE, REPRODUCTION, FOOD HABITS, AGE, AND GROWTH OF ROUND SCAD, DECAPTERUS PUNCTATUS, IN THE SOUTH ATLANTIC BIGHT' L. Stanton Hales, Jr.^ ABSTRACT Five years of bottom trawling indicated that round scad were abundant and widely distributed throughout the South Atlantic Bight in summer and fall, but less abundant and restricted to deeper (28-110 m), warmer (>15°C) waters in winter and spring. Adults and juveniles were spatially segregated, with adults dominating catches in inner and outer shelf regions and juveniles dominating midshelf regions year round. Catches over sponge-coral habitat were significantly greater than catches over sand bottom in winter, whereas catches over the two bottom types were similar in other seasons. This seasonal change in distribu- tion may relate to higher productivity and temperature stability of live bottom habitats. Stomach con- tents indicated that round scad are diurnally feeding zooplanktivores; diets changed seasonally and increased in prey diversity with growth. Round scad spavra repeatedly from March through September. Daily growth analysis revealed that both sexes mature in 4-5 months at approximately 1 1 cm fork length. The life span of round scad could not be determined because the growth record of otoliths of most adults was irregular. Fishes of the genus Decaptertcs occur in most neritic and some oceanic waters of tropical, subtropical, and temperate latitudes. Little is known of the biology of most species, except for those species which sup- port fisheries in the Hawaiian Islands, the Philip- pines, Japan, and the west coast of Africa (Yama- guchi 1953; Tiews et al. 1970; Akaoka 1971; Boely et al. 1973). Although the taxonomy of Indo-Pacific species is unclear (Berry 1968), three species are recognized in the western North Atlantic: the red- tail scad, Decapte7tis tabl Berry; the mackerel scad, D. macarellus (Cuvier); and the round scad, D. punc- tatus (Agassiz). The round scad occurs in the western Atlantic from Nova Scotia to Rio de Janeiro, Brazil, and throughout the West Indies and Bermuda (Berry 1968); however, little information is available con- cerning its basic biology. The distribution of the species has been determined from purse seine catches in the Gulf of Mexico (Klima 1971), where it supports a bait fishery, and from bottom trawl catches over sand bottom habitat in the South Atlan- tic Bight (Wenner et al. 1979a, b, c, d, 1980). The location and duration of the spawning season has 'Contribution No. 71 from the Marine Resources Research In- stitute and No. 222 from the Grice Marine Biological Laboratory. ^Grice Marine Biological Laboratory, The College of Charleston, Charleston, SC 29412 and Marine Resources Research Institute, P.O. Box 12559, Charleston, SC 29401; present address: Depart- ment of Zoology and Institute of Ecology, University of Georgia, Athens, GA 30602. been ascertained from ichthyoplankton surveys in the eastern Gulf of Mexico (Aprieto 1974; Leak 1981). In addition. Leak (1981) determined larval mortality and production, and estimated biomass and potential yield of round scad in the eastern Gulf of Mexico. The objectives of this study were to pro- vide information on seasonal distributions, rela- tive abundance, reproduction, feeding habits, age, and growth of round scad in the South Atlantic Bight. METHODS Seasonal Distribution and Relative Abundance A stratified random sampling design (Grosslein 1969) was used to assign trawling stations within six depth zones (9-18 m, 19-27 m, 28-55 m, 56-110 m, 111-183 m, 184-366 m) on nine seasonal cruises (Table 1). A total of 739 stations were completed on the continental shelf and upper continental slope between Cape Fear, NC and Cape Canaveral, FL. Fishes were captured in a 3/4 scale version of a "Yankee No. 36" otter trawl (Wilk and Silverman 1976) with a 11.9 m headrope, a 16.5 m footrope, and a 1.3 cm stretch mesh cod end liner. Although catches of pelagic fishes by bottom trawls seldom provide accurate estimates of absolute abundance of these species, they probably do reflect relative Manuscript accepted December 1986. FISHERY BULLETIN: VOL. 85. NO. 2, 1987. 251 FISHERY BULLETIN: VOL. 85, NO. 2 Table 1 .—Catch statistics for Decapterus punctatus in the South Atlantic Bight, n = number of trawls; x.,, = stratified mean catch tow"\ X|n = stratified mean catch tow"' for ln(x + 1) transformed data; Xgi.ss = Bliss (1967) estimate of the stratified mean catch tow"'; CV = coefficient of variation for untrans- formed data; CV^ = coefficient of variation for ln(x +1) trans- formed data. Cruise n 'Bliss CV cv„ Fall 1973 (23 Oct. -16 Nov.) Spring 1974 (1 Apr. -9 May) Summer 1974 (13 Aug.-19 Sept.; Winter 1975 (16 Jan.-2 Feb.) Spring 1975 (31 Mar.-IO Apr.) Summer 1975 (30 Aug.-19 Sept.; Winter 1976 (12 Jan.-7 Feb.) Summer 1976 (28 Aug.-21 Sept.; Winter 1977 (17 Jan.-9 Mar.) 67 2.83 0.59 1.86 31.6 1.68 89 1.21 0.25 0.58 35.8 1.70 69 3.13 0.59 1.97 25.5 1.69 52 0.24 0.11 0.79 0.8 2.02 40 0.08 0.06 0.07 0.4 1.31 68 1.17 0.37 0.89 7.9 1.47 69 1.73 0.17 0.52 62.4 2.81 69 1.51 0.34 0.87 25.7 1.70 72 3.30 0.14 1.18 197.0 9.01 abundance and distribution (Wenner et al. 1979a). All trawling was conducted from the RV Dolphin, a 32.6 m converted tug, for approximately 30 minutes at 6.5 km/hour. Weight and fork length (FL) were measured for each individual, except for large catches which were subsampled. Surface and bottom temperatures were taken after each tow. Length-frequency distributions were compared by season and by depth zone. An index of relative abun- dance (IRA = lln X In (x -i- 1); n = # trawls in each depth zone, x = weight of fish for each tow) was calculated from the catches for each depth zone (Musick and McEachran 1972). The stratified mean catch per tow and the estimated variance of the stratified mean catch per tow (Cochran 1977) were calculated from untransformed and In (x + 1) trans- formed data to reduce the effects of contagion (Elliott 1971). The coefficient of variation was used to compare variation in catches (Clark and Brown 1977). The Bliss (1967) approximation retrans- formed the data from logarithmic to arithmetic units. The Wilcoxon rank sum statistic (Hollander and Wolfe 1973) was used to compare catches, depths, and temperatures of sponge-coral (Wenner 1983) and sandy open-shelf (Struhsaker 1969) habi- tats. Habitat designations of Wenner et al. (1979a) were used. Catches collected north and south of lat. 31°30'N were compared for winter and sum- mer cruises to determine if seasonal migration occurred. Reproductive Biology Specimens used for reproductive analyses were collected in 1980 by several research and commer- cial vessels, frozen, and examined in the laboratory. Specimens were measured (nearest mm) and weighed (nearest 0.1 g). Ripe ovaries were fixed in Gilson's solution (Bagenal and Braum 1978) for fecundity determination. All other gonads were fixed in formol-alcohol, stained with a modified Harris hematoxylin and counterstained with eosin (Humason 1972). Maturity stages for testes follow Hyder (1969); maturity stages for ovaries were based on Wallace and Selman (1981) and the fre- quency distributions of oocyte diameters. Frequen- cy distributions of oocyte sizes were determined from randomly selected ovaries in each maturity stage. Analysis of variance revealed no differences in means and variances of oocyte diameters taken from different regions of ovaries in any stage. Therefore, oocyte distributions for each ovary were determined from two or three randomly selected sections. Proportions of fish in each maturity stage were determined bimonthly, and gonadosomatic indices [GSI = gonad wt/(body wt - gonad wt)] were deter- mined for each maturity stage (except stage 1 when gonadal tissues weighed <0.1 g). Ova numbers were estimated by modifying the methods of Macer (1974). Both ripe ova and devel- oping oocytes with diameters >0.115 mm were counted. Developing oocytes were included in fecun- dity estimation because they exhibited characteris- tics of secondary growth phase (Wallace and Selman 1981) and were atretic in spent ovaries. Feeding Habits Stomach contents of 457 fish collected in 1980 were fixed in 20% formalin, and stored in 50% iso- propyl alcohol. Frequency of occurrence (%F0) and percentage composition by number (%A'^) were com- puted for major prey categories. Volumetric displacement (%VOL) of prey categories from a representative subsample of 30 stomachs were determined by using a 0.1 cm- grid (Windell 1971). Seasonal and ontogenetic change in diets were com- pared with an index of relative importance [IRI = (%A^ + %VOL)(%FO)], computed from the sums of each prey category (Pinkas et al. 1971). Feeding periodicity was determined by plotting the percentage of empty stomachs collected per time period, using 377 stomachs with known collection times. Distributions of fish lengths collected in dif- 252 HALES: ROUND SCAD IN THE SOUTH ATLANTIC BIGHT ferent time intervals were compared to evaluate size bias that may occur with this method (Jenkins and Green 1977). mined from 156 individuals (13-185 mm FL) random- ly selected within 10 mm size classes from all specimens (total = 1047) collected in 1980. Age and Growth Utricular otoliths (lapilli) of specimens collected in 1980 were used for age determination. Otoliths were stored in 95% ethyl alcohol and prepared for viewing using a modification of the methods of Haake et al. (1981), which resulted in a thin sagit- tal section containing the core of the otolith em- bedded in "Spurr" (Spurr 1969). Otolith length was measured to the nearest 0.1 mm at 100 x with an ocular micrometer. Otolith images were projected on a high resolution television screen with a high resolution camera, which produced a total viewing magnification of 1088 x or 2176 x. Otoliths ex- amined by scanning electron microscopy were pre- pared by the methods of Haake et al. (1981). Two counts of growth increments were made by the author, and an additional count was made by other experienced readers. Mean counts were used in all analyses, and specimens were discarded if individual counts for a specimen differed by more than 10%. Different readers usually showed agreement between counts: percentage difference between readers averaged 8%. Counts of otolith increments were obtained from 71 juvenile and adult round scad, 13-143 mm FL. Sixty specimens (121-180 mm FL) could not be assigned ages because of the numerous growth interruptions in outer regions of the otolith. Incre- ment formation was validated by examination of the margins of otoliths of juveniles (13-55 mm FL) collected at different times of day. Consistent measurements of the marginal increment could not be made because of the irregular shape of the lapilli; thus, only the occurrence of an incremental or dis- continuous zone (terminology of Mugiya et al. 1981) could be noted. The SAS NLIN regression procedure with DUD and Marquardt options (Helwig and Council 1979) was used to determine parameters for the von Ber- talanffy (1957) and Gompertz (Zweifel and Lasker 1976) growth equations. Because similar patterns of variation were observed in plots of the residuals of both models, r^ values were used to evaluate model performance (Grossman et al. 1985). Instan- taneous growth rates (%FL d'^ and %WT d^^) were calculated according to Ricker (1979). Weights were converted from lengths by using the least squares regression (Sokal and Rohlf 1981), In wt (in g) = 2.96 In FL (in mm) - 11.2 (r^ = 0.99), deter- RESULTS Seasonal Distribution and Relative Abundance A total of 57,460 round scad were captured at 230 of the 739 stations in depths from 11 to 267 m; over 99% of the catch came from <92 m. Fish ranged from 2 to 26 cm FL (x = 11.4), with 99% of the fish 6-17 cm FL. Round scad were more widely distributed and abundant in summer and fall than in winter and spring. Indices of relative abundance (Fig. 1) were consistently high in summer and fall at shallow depths (<55 m) where D. punctatus were captured at 121 of 220 trawl stations. Indices of relative abun- dance during summer and fall in 56-110 m depths were quite variable, and catches in waters >110 m were rare (3 of 78 trawls), small (52 individuals cap- tured), and occurred only in summer. The highest indices in winter and spring occurred in 19-110 m depths (usually 19-27 m), but were lower than values in summer and fall. Round scad were rarely collected in 9-18 m depths and never collected in waters >110 m in winter. Differences among un transformed (Xgf), trans- formed (xin), and Bliss (ieiiss) estimates of the stratified mean catch per tow (Table 1) revealed additional seasonal changes in the distribution of round scad. Transformed and Bliss estimates of the stratified mean catch per tow were higher in sum- mer and fall than in winter and spring. However, untransformed values (Xg/) indicated that total catches in winter often exceeded total catches in summer. Such differences in catch statistics resulted from the relatively high frequency and low vari- ability of catches in summer and fall, and the rela- tively low frequency and high variability of catches in winter and spring. This result was generally con- sistent with the coefficients of variation (CV and CVin), which indicated increased clumping in the winter catches. The seasonal distribution of round scad appeared affected by temperature (Table 2). Over 97% of winter catches occurred in waters warmer than 15°C, over 99% of spring catches were made in waters warmer than 17° C, and over 99% of sum- mer and fall catches occiirred in waters warmer than 20°C. Habitat affected the distribution of round scad in 253 WINTER 1975 0 6 0 4 0 2 4/16 6/12 4, ,4 LIJ 0 m 1 B 0/3 0/3 0/4 o z < Q z WINTER 1976 3 m < 0.6 ai > 0 4 sne 5/15 < _l 0.2 0 1/17 ■ 1 a.i|o/» 0/8 oc U. O WINTER 1977 X Ol Q z 06 0 4 7/25 ^ 3/13 0 2 ■ H A 0/17 mMM''- 0/1 1 SPRING 1974 0.6 f 0.4 0.2 12/23 11/19 ^ ^1'?9 2/18 0/13 0/14 SPRING 1975 9- 19- 28- 56- 111- 184- 18 27 55 HO 183 366 9- 19- 28- 56- 111- 184- 18 27 55 110 183 366 FISHERY BULLETIN: VOL. 85, NO. 2 e/,g SUMMER 1975 0 6 SUMMER 1976 10/23 1/14 0/10 ^''^ FALL 1973 13/18 13/19 ^_ 5/1 2 9- 19- 28- 56- 111- 184- 18 27 55 110 183 366 DEPTH ZONES Figure 1.— Indices of relative abundance (Musick and McEacliran 1972) oiDecapterus pwnctatvs by depth zone from MARMAP trawl survey for all cruises. Fractions are the number of trawls with round scad/total number of trawls. Table 2.— Surface and bottom temperatures where Decapterus punctatus was collected in the South Atlantic Bight. Upper figure of each pair is surface value, lower figure is bottom value; x = mean, s = variance. Winter Spring Summer Fall 1975 1976 1977 1974 1975 1974 1975 1976 1973 mini- mum maxi- 17,4 17.2 2.5 2.0 13.3 13.3 21.7 mum 19.9 19.1 17.0 3.5 2.8 12.1 12.3 23.3 22.8 17.9 16.4 2.5 2.1 11.5 11.6 18.9 18.9 19.6 19.1 1.5 1.3 17.4 17.0 23.2 22.0 18.7 17.4 2.7 1.3 14.9 14.9 27.9 25.5 0.5 1.9 26.9 19.3 28.1 25.4 26.0 25.4 23.8 23.7 23.8 29.1 19.2 28.2 0.6 0.6 1.9 4.9 3.5 1.9 26.9 26.0 17.2 9.8 8.5 17.2 29.3 28.7 26.5 29.2 27.4 26.3 winter. Catches over sponge-coral habitat were sig- nificantly larger than catches over sandy habitat in winter during 1976-77, whereas catches over the two bottom types were similar in other seasons (Table 3). Habitat types did not differ in tempera- ture during either winter (Wilcoxon rank sum tests, P = 0.25 and 0.20, df = 19 and 12, respectively). Thus, temperature alone did not appear to account for the observed difference in winter catches over the two bottom types. Table 3.— Comparison of catches of Decapterus punctatus over sponge-coral and sand habitats by cruise. If the Wilcoxon rank sum (Z) is significant (*), the direction of the difference is indicated. ZC; = sum of catches over sponge-coral habitats (n, = # trawls), XCj = sum of catches over sand habitats (n^ = trawls). Cruise IC, ICc Winter 1975 1976 1977 Spring 1974 1975 Summer 1974 1975 1976 Fall 1973 3 7,201 10,936 239 2 445 614 663 215 (3) (11) (11) (11) (6) (14) (18) (8) (10) 672 597 39 4,717 46 8,023 4,805 5,094 13,139 (49) (75) (82) (91) (34) (74) (69) (81) (77) 0.23 3.57* 5.40* 0.05 0.36 0.61 0.92 0.86 0.74 (L>S) (L>S) Round scad did not appear to undertake seasonal longshore migrations in the South Atlantic Bight. Catches north and south of lat. 31°30'N (which roughly bisects the South Atlantic Bight) showed occasional differences (Table 4), but no consistent 254 HALES: ROUND SCAD IN THE SOUTH ATLANTIC BIGHT seasonal pattern. Catch differences indicative of a southward migration in winter and northward in summer (either N < S in winter and spring or N > S in summer and fall) occurred only in winter 1975 and summer 1976. Distributions opposite to the above patterns occurred in summer 1975. Similar- ity in the numbers (Chi-square test, P = 0.50, 1 df) of gill rakers, a variable character (Berry 1969), from specimens collected off South Carolina in winter (i = 37.2, s^ = 1.2, n = 33) and summer (x = 36.9, s^ = 1.4, n = 38) also suggested that discrete stocks were not migrating through the South Atlantic Bight. Although adults and juveniles were caught at all depths throughout the year, length-frequency dis- tributions by depth (Fig. 2) showed a similar pat- tern for nearly every cruise: fish size decreased from 9-18 m to 19-27 m depths, then increased with in- creasing depth to 110 m. Catches in 9-18 m consisted primarily of adults in fall and winter, whereas both juveniles and adults were captured in spring and summer. Juveniles predominated in 19-27 m, where- as adults composed most of the catch in deeper waters. Table 4.— Comparison of catches of Decapterus punctatus north and south of lat. 31°30'N by cruise. If the Wilcoxon rank sum (Z) is significantly different (*), the direction of the difference is in- dicated. SCn = sum of the catches north (n^ = # trawls north), ZCj = sum of the catches south (n^ = # trawls south). Cruise 2:c„ "n ^c. "s Z Winter 1975 8 (17) 667 (35) 561.80* (S>N) 1976 6,149 (42) 1,649 (44) 0.07 1977 10,728 (48) 247 (45) 0.31 Spring 1974 993 (62) 3,963 (40) 0.33 1975 44 (36) 4 (4) 0.25 Summer 1974 1,892 (47) 6,576 (41) 0.15 1975 1,915 (49) 3,504 (38) 1.35* (S>N) 1976 4,395 (46) 1,362 (43) 2.52* (N>S) Fall 1973 6,184 (47) 7,170 (40) 0.85 Reproductive Biology All stages of ovarian maturity had different fre- quency distributions of oocyte diameters (Fig. 3). Resting ovaries contained primary or first growth phase oocytes approximately 25-115 /^m in diameter. Few larger (>115 fxra) developing or atretic oocytes occurred. Oocytes in developing ovaries ranged from 30 to 375 nm in diameter, and exhibited character- istics of first or second growth phase. Either one or two modes were present in the frequency distri- butions of the sizes of second growth phase oocytes, 100-375 ^m in diameter. Germ cells in ripe ovaries ranged from 30 to 495 /im in diameter. Ripe ovaries contained oocytes in both growth phases and matur- ing ova. Two or three modes were present in the frequency distibution of germ cells of ripe ovaries. Spent ovaries contained small oocytes in primary growth phase and occasionally larger oocytes under- going atresia. Germ cells in these flaccid ovaries were usually 30-255 /im in diameter, although larger cells were observed. Gonadosomatic indices (Table 5) changed as expected: indices increased from rest- ing through ripe stages, then decreased for spent fish. Maturity stages of testes were more difficult to distinguish because testes often contained all stages of spermatogenesis; therefore, stage determination was based upon subjective interpretation of the relative quantities of spermatocytes, spermatids, and sperm. Although more variable, gonadosomatic indices of males (Table 5) were similar to those of females. Table 5. — Gonadosomatic indices (gonad wt/(total body wt - gonad wt)) of Decapterus punctatus by stage maturity, x = mean, s = variance, n = sample size. Testes Ovaries X s n X s n Resting Developing Ripe Spent 0.008 0.018 0.028 0.015 0.004 0.012 0.014 0.009 64 103 74 18 0.007 0.027 0.045 0.016 0.003 0.011 0.019 0.008 60 49 64 13 Seasonal occurrence of maturity stages showed good agreement between males and females (Fig. 4), and indicated a protracted spawning period. Developing gonads were found from February through August, and ripe individuals of both sexes were collected from March through August. Examination of gonads indicated that both species mature at approximately 110 mm FL. Frequency distributions of length by sex (Fig. 5) indicate that both sexes mature over a narrow size range. Speci- mens <100 mm FL were immature, whereas more than 90% offish 110-119 mm FL were mature. Both ripe males and females were collected in the size range at which gonadal development begins (100- 109 mm FL). Fecundity estimates (# ova female"^ yr"^) for 32 ripe females (119-174 mm FL) ranged from 6,200 to 51,000 per female and were highly variable for specimens of similar sizes. The distribution of the sizes of the oocytes in these ovaries was not deter- 255 FISHERY BULLETIN: VOL. 85, NO. 2 WINTER 1975 SPRING 1974 10 15 20 SUMMER 1975 111-184 0 5 to 15 20 0 5 10 15 20 WINTER 1976 SPRING 1975 SUMMER 1976 60- 56-367 10 15 20 0 5 10 15 20 5 10 15 20 SUMMER 1974 FALL 1973 WINTER 1977 0 5 10 15 20 FORK LENGTH (cm) '0 S 10 IS 20 FORK LENGTH (cm) FORK LENGTH (cm) Figure 2.— Frequency distributions of fork lengths of Decapterus punctatus by depth zone, for all cruises. Numbers above each distribution indicate range of the depth zone in m; N is sample size; x is mean length of individuals. 256 HALES: ROUND SCAD IN THE SOUTH ATLANTIC BIGHT 60n RESTING 40- 20- UJ 30H a. < 20H UJ > o o o u. o 10- 20- 10- 60- 40- 20- DEVELOPING RIPE SPENT I ' I ' I ' I — I — i — I — I — I — r 155 235 315 395 475 OOCYTE DIAMETER(ijm) Figure 3.— Frequency distributions of oocyte diameters from four randomly selected ovaries of Decapterus punctatus in each matur- ity stage. The total number of oocytes measured in each ovary was 306 (oocytes from 3 sections within the anterior, central, and posterior regions of each ovary). mined, and it was unknown if these individuals had spawned. The regression equations of fecundity on length and weight for 33 specimens were as follows: logio fee = -10.9 + 5.63 logjo FL {r^ = 0.46); and logio fee = -1.14 + 1.56 logio wt (r- = 0.55). Feeding Habits Approximately 91% of round scad (39-189 mm FL) contained identifiable prey. The highest indices of relative importance for all specimens were for cope- pods (0.37), mollusk larvae (0.19), amphipods (0.06), and ostracods (0.04). The most numerous prey groups were mollusk larvae (29%, predominantly gastropod and pelecypod veligers), copepods (25%), barnacle cyprids (14%), and ostracods (10%). Chae- tognaths (35%), copepods (28%), mollusk larvae MALES FEMALES JAN - FEB N=43 MAR -APR N=67 MAY-JUN N=50 JUL -AUG N=93 NOV -DEC N=28 JAN - FEB N=51 MAR -APR N=43 MAY-JUN N=60 JUL-AUG N = 44 NOV- DEC N=22 GONAD STAGES DEVELOPING RIPE SPENT RESTING Figure 4.— Seasonal occurrence of maturity stages of gonads of Decapterus punctatus. N is sample size for each bimonthly period. No gonads were examined from specimens collected in Septem- ber and October 1980-81. CO 70- 50- 30- 10- § 70- Q -z. L. 30- o (7) 10- QC LiJ GQ 3 180- -z. 120- 60- FEMALE JZl MALE IMMATURE 35 55 75 95 115 135 155 175 FORK LENGTH (mm) Figure 5.— Frequency distributions of fork length of Decapterus punctatus by sex. Both sexes mature at approximately 110 mm FL. 257 FISHERY BULLETIN: VOL. 85, NO. 2 (12%), and amphipods (10%, predominantly hyperi- ids and caprellids) contributed the greatest volumes of prey. Copepods (70%), mollusk larvae (47%), am- phipods (30%), and decapod larvae (25%) occurred most frequently. Some differences were found in stomach contents of fishes of different size (Table 6). Small fish preyed almost exclusively on copepods. Medium-sized fish preyed predominantly on copepods, but less fre- quently and to a lesser extent than small fishes. Large fish fed on a variety of prey and consumed large prey items (such as chaetognaths). Mollusk larvae and copepods dominated the diets in all seasons except spring (Table 7). In spring, round scad fed on copepods, ostracods, chaeto- gnaths, and barnacle cyprids. The preponderance of copepods in diets of round scad in summer is due in part to the large number of juveniles included in the analysis. The mean size of fish in the summer sample v^ras smaller than the mean size of fish in all other seasons (Student-Nev^man-Keuls tests; ^453 4 = 14.1 for winter vs. summer, ^453 3 = 13.2 for summer vs. spring, and ^453 9 = 11.1 for summer vs. fall). In all other seasons, dietary analyses (Table 5) were based on fish samples with similar size distributions (Student-Newman-Keuls tests; 94533 = 0.7 for winter vs. fall, ^453 2 = 0.7 for winter vs. spring, and ^453 2 = 0.1 for fall vs. spring). The mean number of prey items showed considerable seasonal variation from 4.6 in fall to 104 in spring. The percentage of empty stomachs varied as a function of time of day (Fig. 6). Few empty stomachs (2-7%) were collected from midmorning to early evening, whereas 13-29% of stomachs were empty from early evening to midmorning. Size effects are unlikely to have caused the observed differences in the percentages of empty stomachs: samples with lower (2, 5, and 7) and higher (13, 20, and 29) per- centages were comprised of fish of similar size (ANOVA, F,i,4) = 0.01, P > 0.75). Age and Growth Validation of the daily growth marks on otoliths of round scad was provided in two ways. Examina- tion of marginal increments of lapilli from small specimens (13-55 mm FL) collected at different times of day suggested daily periodicity of increment formation. The margin consisted of the transparent incremental zone from midafternoon until early morning and the dark discontinuous zone in mid- morning (Table 8). The allometric relationship be- tween otolith and fish length also validates the use of otoliths for age determination. Otolith length (OL) Table 6.— Index of relative importance (IRI), frequency of occurrence (%F0), volumetric displacement (%VOL), and relative abundance of prey (%A/) by sizes. N = sample size (# empty stomachs), x = mean fork length in mm, s = vari- ance of fork lengths, and n = total number of prey. N VOL FO Length/prey % % % IRI 40-89 mm FL: N = 82(12) , X = 50 mm. s = 13 mm, n = 1,272 Copepoda 74 80 75 1.15 Molluscs 13 5 24 0.04 Decapoda 3 7 27 0.03 Amphipoda 2 2 15 0.01 Eggs 3 2 24 0.01 Ostracoda 1 1 10 <0.01 Other 4 3 32 0.02 90-139 mm FL: N -- = 192(15), X = 114 mm, s = 13 mm, n = 8,003 Copepoda 34 57 74 0.67 Mollusca 20 13 47 0.16 Cirripedia 31 13 20 0.09 Ostracoda 9 8 22 0.04 Amphipoda 2 3 28 0.01 Decapoda 1 3 25 0.01 Other 3 3 32 0.02 140-189 mm FL: N = 183(14), > r = 156 mm, s = = 10 mm, n = 12,850 Mollusca 36 12 59 0.28 Copepoda 14 12 64 0.17 Amphipoda 14 12 31 0.08 Chaetognatha 15 53 10 0.07 Ostracoda 11 5 30 0.05 Cirripedia 4 2 25 0.02 Decapoda 1 1 32 0.01 Other 5 4 48 0.04 258 HALES: ROUND SCAD IN THE SOUTH ATLANTIC BIGHT Table 7.— Frequency of occurrence, volumetric displacement, and relative abun- dance of prey of Decapterus punctatus by season. Abbreviations as in Table 6; r = range of fork lengths. N VOL FO Season/prey % % % IRI Winter: N = 116(10), x = 140 mm, s = 25 mm, r = 85-185 mm. n = 4,357 Mollusca 47 32 48 0.38 Copepoda 17 31 52 0.25 Ostracoda 15 14 31 0.09 Cirripedia 12 5 33 0.06 Decapoda 2 7 35 0.03 Amphipoda 1 2 17 0.01 Other 6 9 47 0.07 Spring: N = 112(5), x = 135 mm, s = 18 mm, r = 97-175 mm, n = 10,999 Copepoda 21 18 92 0.36 Amphipoda 17 15 52 0.16 Ostracoda 13 6 45 0.09 Chaetognatha 17 51 13 0.09 Cirripedia 22 5 25 0.07 Mollusca 6 2 67 0.05 Decapoda 1 2 26 0.01 Other 3 2 45 0.02 Summer: N = 161(16) X = 87 mm, s = 40 mm r = 35-163 mm, n = 6,505 Copepoda 55 73 79 1.01 Mollusca 37 18 43 0.24 Decapoda 1 3 22 0.01 Amphipoda 1 1 21 <0.01 Ostracoda 1 1 10 <0.01 Cirripedia 1 1 9 <0.01 Other 2 2 28 0.01 Fall: N = 68(10), x = 138 mm. s = 33 mm, r = 48-180 mm, n = 264 Mollusca 65 44 21 0.23 Copepoda 19 34 42 0.22 Other 16 22 36 0.14 CO X o < CO I- Q- :e LlI 30- 2 5- 20- 0- 5- (40) (52) 0300 0700 1100 1500 1900 2300 TIME OF COLLECTION(E.S.T) Figure 6.— Percent of empty stomachs of Decapterus jmnctatiis collected at different times of day. Samples were pooled over 4-h intervals, of which midpoints are given, (n) is sample size. Table 8.— Appearance of marginal increments of otoliths of Decapterus punctatus collected at different times of day. Data in- dicate increments are formed daily. N = number of specimens, times are Eastern Standard Time. FLof specimens (A/) Time of capture Marginal increment 42-55 mm (4) 17-22 mm (11) 13-24 mm (4) 35-53 mm (8) 0239-0257 0812-0817 1506-1516 2045-2053 wide, transparent thin, dark wide, transparent wide, transparent was proportional to fork length throughout the size range (14-143 mm FL) for which age determination was possible (logio OL (in mm) = 0.82 logjo FL (in mm) - 1.61; r^ = 0.99, n = 71). The thickness and structure of growth increments changed in consistent ways in lapilli (Fig. 7). Ten to twelve faint daily increments surrounded a cen- tral core, which contained the primordium (Fig. 8). The next 10-15 increments increased in thickness, and the following 20-25 increments gradually de- creased in thickness. A distinct change occurred at this point, and increments became thinner and more regular. Increments appeared uniform in thickness 259 FISHERY BULLETIN: VOL. 85, NO. 2 Figure 7.— Sagittal section through a lapillus of Decapteriis punctatus. The pattern of fine, regular growth increments is interspersed increasingly with heavy, irregular growth interruptions (crosshatches) in outer regions of the otolith. Bar indicates 0.10 mm, P is the primordium. 260 HALES: ROUND SCAD IN THE SOUTH ATLANTIC BIGHT Figure 8.— Scanning electron micrograph of a sagittal section through the primordium of a lapillus oiDecapterus punctatus. P is the primordium and crosshatches denote daily increments. The pattern of otolith growth of juveniles was consistent in most otoliths. to approximately increment 100, then gradually became thinner and more difficult to count. Growth interruptions appeared in outer portions of otoliths of large fish, and growth records were more irreg- ular. Round scad grew rapidly for 120-150 days until reaching sexual maturity at approximately 110 mm FL (Fig. 9). The von Bertalanffy [FL = 161 (1 - exp (-0.012 (age - 29.5)))] and Gompertz [FL = 1.17 exp [4.76 (1 - g- 0.026 (age)) jj g^wth equations provided good and nearly identical fits (r^ = 0.96 and 0.97, respectively) to the observed data. Specific growth rates (Table 9) for juveniles were initially high, decreased sharply until sexual maturity, then decreased more gradually throughout the time period for which age determination was possible. The largest specimen for which reliable counts could be determined was 143 mm FL, although most specimens at that size could not be assigned an age. The age of round scad for the entire size range (FL up to 21 cm) that was collected in the South E E. X t- o z LU _l CO o 160 120 80 40 - FL = 161 (1 - exp 1 -0.012 (age - 24.3)]} — FL = 1.17 exp (4.76 [1 - exp (-0,026 * age)]) 50 100 150 200 250 AGE(d) Figure 9.— Von Bertalanffy (thin line) and Gompertz (heavy line) growth equations. Sexual maturity of Decapterus punctatus is reached at approximately 110 mm FL in 120-150 days. Atlantic Bight could not be determined from daily increments on the lapillus. The age of all specimens <120 mm FL could be determined, but only half of 261 FISHERY BULLETIN: VOL. 85, NO. 2 Table 9.— Instantaneous growth rates (Ricker 1 979) of Decapterus punctatus predicted from the von Bertalanffy equation (fig. 11). Weights were con- verted from lengths by using the regres- sion: In Wt (g) = 2.96 In FL (mm) - 11.2. Age FLd-^ Wtd"' (d) (0/0) (%) 35-60 6.3 13.2 61-85 1.9 5.6 86-105 0.8 2.6 106-130 0.7 2.2 131-155 0.4 1.6 156-180 0.3 1.2 the specimens 121-143 mm FL and no individuals >143 mm FL could be assigned an age. Replicate counts of otolith increments from adults often dif- fered by >10%, and were considered unreliable. In addition, frequent growth interruptions occurred in the outer portions of the otoliths of large specimens, and the timing of formation of such marks was not known. DISCUSSION Seasonal Distribution and Relative Abundance Round scad apparently migrate shoreward across the continental shelf as sea temperatures increase in spring, then migrate into warm (>15°C) midshelf (28-110 m) depths as inshore temperatures decline in winter. Intermediate shelf depths in the South Atlantic Bight are fairly warm year round, unlike inshore waters which are seasonally cooled by cold fronts (Atkinson et al. 1983) and outer shelf waters which are intruded upon by cold-water upwellings (Blanton et al. 1981). Magnuson et al. (1981) re- ported that catches of round scad were proportional to temperature, and laboratory studies (Wyllie et al. 1976) have suggested a preferred temperature of 27°C. Seasonal onshore and offshore movements are made by round scad in the Gulf of Mexico (Klima 1971) and by other pelagic fishes in the South Atlan- tic Bight (Wenner et al. 1979a, b, c, d, 1980) and elsewhere (Allen and DeMartini 1983). Longshore migration does not appear to be a con- sistent feature of the movements of round scad along the southeastern Atlantic states. Differences of catches (Table 4) made north and south of lat. 31°30'N are inconsistent, and limited meristic data (see Results) provides no evidence of movements by discrete stocks through the South Atlantic Bight. Although fish communities of live bottom habitats may change (Chester et al. 1984), latitudinal differ- ences in the distribution of demersal fishes in sandy, open shelf habitats of the South Atlantic Bight are not apparent (Wenner et al. 1979a). More substan- tive information (tag-recapture studies, etc.) than the limited data given above is needed to determine accurately the movement patterns of round scad. At the present time, available information suggests that seasonal migration in round scad involves main- ly onshore or offshore movement. The abundance of round scad in the South Atlan- tic Bight is undoubtedly underestimated by catch statistics (Table 2). First, benthic otter trawling is generally inadequate for determining the abundance of small, mobile pelagic species (Wenner et al. 1979a). Second, the attraction of round scad to sponge-coral habitats in winter may have exag- gerated the apparently large fluctuations in the seasonal abundance of round scad in the South Atlantic Bight. The sampling protocol assumes equal probability of capture over the different habitat types, but the probability of capturing round scad over live-bottom habitats varies seasonally. An in- creased proportion of live-bottom areas should be sampled in winter to obtain more reliable estimates of the abundance of round scad. The distribution, extent, and adequacy of sampling of live-bottom habitats are not well known (Wenner 1983). Thus, sampling inadequacies and the seasonal attraction of round scad to live-bottom habitat result in under- estimation of abundance. Several factors may influence the attraction of round scad to live-bottom habitat in winter. First, round scad utilize live-bottom habitat in winter, when invertebrate biomass has peaked and poten- tial competitors and predators have decreased (George and Staiger 1978^; Wenner et al. 1980, 1983, 1984; Sedberry and Van Dolah 1984). Second, winter temperatures at live-bottom stations, though not significantly different from temperatures at sand-bottom stations, tend to be warmer, and scad prefer warm waters. The greatest densities of round scad occur in the midshelf where seasonal temper- atures are generally warmest and the most highly productive live-bottom areas are located (Miller and Richards 1980; Sedberry and Van Dolah 1984). Finally, the relief of live-bottom habitats (albeit low) may serve to attract round scad; many coastal pelagic fishes, including round scad, have an affin- ^George, R. V., and J. C. Staiger. 1978. Epifaunal benthic invertebrate and demersal fish populations in the Georgia Bight continental shelf environment. South Atlantic Benchmark pro- gram, Volume 3, Texas Instruments Inc. Draft Report, p. 211-254. 262 HALES: ROUND SCAD IN THE SOUTH ATLANTIC BIGHT ity for structure (Klima and Wickham 1971; Feder et al. 1974; Hastings et al. 1976). Although fish sizes within each depth zone over- lapped, the observed pattern was fairly consistent and distinct for all cruises; adults composed the catch in 9-18 m depths, and fish in deeper waters (mostly juveniles) showed a positive size-depth corre- lation. Adults migrate inshore in spring to feed and spawn, and offshore in winter to avoid cold waters. However, the apparent movement of juveniles to deeper waters is not understood. Correlations be- tween fish size and depth are numerous in aquatic habitats (Helfman 1978), but explanations based on changing physiological tolerances (Bullis and Struh- saker 1970), foraging strategies (Polloni et al. 1979), and predation responses (Hobson 1972) have been difficult to demonstrate in most fishes. Reproductive Biology Although results (Fig. 3) indicate that D. punc- tatiis spawn primarily from March through August, spawning probably occurs through September and to a lesser extent throughout the year. Collections were not made during September and October of 1980, but water temperatures in August and Sep- tember in the South Atlantic Bight are generally similar (Atkinson et al. 1983). Round scad larvae have been collected in winter in the South Atlantic Bight (Fahay 1975; Powles and Stender 1976) and the eastern Gulf of Mexico (Leak 1981). Larval oc- currence has been correlated with water temper- ature in the eastern Gulf of Mexico (Leak 1981), and sufficient water temperatures (>20°C) occur throughout the year in parts of the South Atlantic Bight. The pattern of oocyte development is generally quite variable and complex in serial spawners: oocytes develop asynchronously or synchronously in groups, and ova are released in batches. Three observations suggested that round scad are serial spawners: 1) the occurrence of three distinct modes in the frequency distributions of oocyte diameters from ripe ovaries collected in spring; 2) two modes in those distributions from ripe fish collected in late summer; and 3) evidence of spawning in ovaries hav- ing a frequency distribution of oocyte diameters similar to developing ovaries. Spawning was in- dicated by disorganized ovarian septa with con- spicuous spaces, debris in the ovarian lumen, residual atretic oocytes, and brown bodies. Although estimates of fecundity in round scad are comparable to those of D. pinnulatiis (Yamaguchi 1953), D. macrosoma, and D. rtisselli of similar size (Tiews et al. 1970), the conventional method applied here probably underestimated fecundity. Because gonads used in fecundity estimation were not ex- amined histologically, it was not possible to deter- mine if spawning had occurred recently in specimens used for fecundity measures. In addition, sufficient numbers of specimens were not examined to deter- mine spawning frequency from running-ripe (DeMartini and Fountain 1981) or postovulatory (Hunter and Goldberg 1979) females. Previous studies on other serial spawning fishes (Hunter and Goldberg 1979; Hunter and Leong 1981; DeMartini and Fountain 1981; Conover 1985) have shown that estimates of annual fecundity (total number of ova spawned in 1 year) can differ from conventional fecundity estimates (which ignore multiple spawn- ing) by an order of magnitude. Serial spawning fishes generally have low relative ovary weights (Martinez and Houde 1975; Smith and Lasker 1978; DeMartini and Fountain 1981), but can expend over 100% of their body weight per year in eggs (Hubbs 1976; DeMartini and Fountain 1981). If observed fecundity in round scad (6,200-51,000) is extra- polated from the 4.3% relative ovary weight (Table 7) to total body weight, then fecundity estimates of 142,000-1,173,000 would result. If observed fecun- dity is divided by the proportion of oocytes in the most advanced developmental mode (32%, from Figure 3), then batch fecundities of approximately 2,000-16,000 per female (130-230 eggs/g body weight) and annual fecundity estimates (based on 10 d spawning cycle for 6 months) of 36,000-288,000 would result. Both estimates are entirely specula- tive, but support the contention that the conven- tional method underestimated fecundity, and emphasize the need for additional studies on the fecundity of round scad. Round scad mature at a smaller size than reported for other species of Decapteriis, which reach matur- ity at 18-20 cm (Yamaguchi 1953; Tiews et al. 1970). The small size at which round scad become sexual- ly mature suggests that they are under strong selec- tion pressure to mature rapidly. The natural mor- tality of round scad in the eastern Gulf of Mexico is high (Houde et al. 1983). Compared with tem- perate and boreal species, many tropical clupeoids also mature at small sizes, seldom attain large size and have high adult mortality rates (Blaxter and Hunter 1982; Houde et al. 1983). Feeding Habits Zooplanktivores feed during the day or at night, but seldom during both periods (de Silva 1973; Hob- 263 FISHERY BULLETIN: VOL. 85, NO. 2 son and Chess 1976; Helfman 1986), probably due to visual limitations (Durbin 1979). The absence of nocturnal prey and the strongly diurnal periodicity of stomach fullness (Fig. 6) indicate that D. punc- tatus feeds during the day. Round scad rarely con- sumed mysids, tanaids, and cumaceans, which are abundant in the water column at night but dwell in the benthos during the day (Kaestner 1970; Hobson and Chess 1976). The diel periodicity in empty stomachs (Fig. 6) is not biased by sizes of fish differ- ing among collection times (see Results). Individuals of different sizes have similar gut evacuation rates (Perrson 1981) but different gut capacities; there- fore, small individuals empty their guts more quickly than large individuals. The lack of a size difference in this analysis substantiates the daily feeding period by round scad. Scales were a common item in round scad stomachs, but were deleted from analyses because several observations indicated that round scad were feeding on debris (including scales) generated by trawling: 1) oral chambers of specimens often con- tained scales; 2) scale size and type varied; and 3) the extent of presumed lepidophagy was not corre- lated with fish size. In addition, scales were seldom found in latter portions of the gut. Most other lepi- dophages generally prey on a small number of species (Sazima 1983), or for only a portion of their life history (Carr and Adams 1972). Thus, round scad probably consume scales on occasion, but not to the extent that the data would indicate. It seems likely that round scad were feeding on scales abraded from fishes during their avoidance of or capture by the trawl. Yamaguchi (1953) also attributed the occur- rence of scales in the stomachs of D. pinnulatus to gear bias. Ontogenetic changes in the diet occurred with growth. Larger individuals had more diverse diets which included larger prey. Small fish fed primar- ily on small, abundant copepods and copepodites. Stomach contents of several (10) small juveniles (13-26 mm FL) collected during an ichthyoplankton survey in 1973 also contained mostly copepods (S. Hales pers. obs.). Data from these specimens were not included in previous analyses due to the possi- bility of differential prey digestion. Seasonal changes in the diets of round scad prob- ably reflect fish size and the relative abundance of zooplankton components. Copepods and mollusks were the most important prey in all seasons except during spring when barnacle cyprids were the most numerous and chaetognaths contributed the great- est volume of prey. Zooplankton volumes and diver- sity reach their peak in the spring (Deevey 1960; Reeve 1964), and cyprids are more abundant in the spring than in any other season (Lang and Acken- hausen-Johns 1981). The preponderance of copepods in the diet of round scad in summer partially re- flects the abundance of juvenile round scad, which appear to feed primarily on small abundant cope- pods. Molluscan veligers may be an exception to the general pattern of prey selection. They have not been reported to be abundant in the zooplankton of the South Atlantic Bight (Paffenhofer 1980, 1981) or elsewhere (Deevey 1960; Reeve 1964), yet oc- curred frequently in the diets of round scad. Shells of both gastropods and bivalves and the opercula of gastropods were all that remained in the stomachs of round scad on occasion; such parts are apparent- ly digested slowly and retained in the stomach. How- ever, such an explanation alone does not account for the frequency and abundance of mollusk veligers. Brewer and Kleppel (1986) have reported the para- dox of low bivalve density yet frequent occurrence in the guts of some larval fishes, and suggested that bivalve veligers may occur in microscale patches just above the bottom. Planktonic stages of gastropods are important prey for horse mackerel {Trachurus trachurus), which school near the bottom (Macer 1977). Thus, the frequency and abundance of mol- lusk veligers in the stomachs of round scad may be attributed to both the abundance of mollusk veli- gers being greater than generally recognized and their low digestibility and long retention in stom- achs. Age and Growth Round scad grow rapidly to sexual maturity at 11 cm FL in 4-5 months, and apparently achieve a major proportion of their total size in their first year. Because of the problems encountered in age deter- mination in this study, little can be said about the age of most adults. The asymptote predicted by the von Bertalanffy model (161 mm FL) is much shorter than the maximum size observed in the South Atlan- tic Bight. Thus, the growth rates observed in this study should not be extrapolated to older and larger fish. Houde et al. (1983) reported the mean size of round scad in the eastern Gulf of Mexico to be 136 mm FL at age 1, 160 mm FL at age 2, and 177 mm FL at age 3. In addition, they reported considerable variation (10-20 mm) in the mean lengths at age. Dif- ferences in the growth rates observed in the two studies are believed to be due mainly to method- ology, but may also be due to slight differences in the growth of round scad between the two areas. 264 HALES: ROUND SCAD IN THE SOUTH ATLANTIC BIGHT Growth rates reported in this study for juvenile round scad are similar to those of juvenile Selar crumenopthalmus (Kawamoto 1973) and Trachurus trachurus (Macer 1977), and would enable spring- spawned round scad encountering favorable condi- tions to spawn in the fall (Leak 1981). Other species of Decapterus are reported to grow more slowly (Yamaguchi 1953; Tiews et al. 1970; Ingles and Pauly 1984), but such studies used length-frequency information only and may have underestimated growth. These species {D. pinnulatus, D. russelli, D. macrosoma) reach sexual maturity at 17-20 cm FL in their first or second year, and attain 25-35 cm FL in 3-5 years (Yamaguchi 1953, Tiews et al. 1970). The rapid growth of round scad to sexual maturity suggests that this species is under strong selection to mature early. Mortality estimates for round scad (Leak 1981; Houde et al. 1983) are high even in comparison with other coastal pelagic fishes, which do not achieve as large a size as round scad during their first year. Daily growth increments were easily distinguished in juveniles and small adults, but could not be used to determine ages of most adults. Frequent spawn- ing involving the high energetic expenditures reported for other pelagic fishes (Hubbs 1976; DeMartini and Fountain 1981) would result in slow growth of adults. Specific growth rates based on Houde et al. (1983) (0.04% FL d'^ for 1-2 yr olds, and 0.03% FL d"^ for 2-3 yr olds) are much slower than the rate of young adults (0.28% FL d'^) observed in this study. A reduction in increment thickness with growth (Brothers 1979; Campana 1985) might result in the exceedingly fine incre- ments observed in the outer portions of the otoliths of adults. Another possibility is that growth of adults is insufficient to maintain the pattern of daily incre- ment formation. McGurk (1984) has reported that daily increment formation in larval Clupea harengus may be altered when absolute growth rates are <0.36 mm d'^. Occurrence of growth interruptions in the outer portions of otoliths of most adults also hindered age determination. The period of formation of such marks was not known. If such marks are not formed in 1 day, then inclusion of such marks as daily in- crements would result in underestimation of the ages of adults. Growth interruptions may result from spawning (Panella 1974), lunar or tidal rhythms (Rosenberg 1982), or stress (Ralston and Miyamato 1983); however, correlations between such events and the growth record of round scad could not be determined. ACKNOWLEDGMENTS I thank W. D. Anderson, Jr., C. A. Barans, C. K. Biernbaum, N. A. Chamberlain, and C. A. Wenner for advice and encouragement, and D. E. Facey, G. S. Helfman, G. R. Sedberry, W. A. Roumillat, and the anonymous reviewers for their comments on various drafts. Personnel of the BLM (contract AA551-CT9-27) and MARMAP (contract NA84- WCC-06101) programs at the Marine Resources Research Institute, Charleston, SC provided much needed specimens and assistance. K. Swanson, J. Lay, E. Peters, B. Sanders, and J. Barrett prepared the figures, and J. Evans typed the manuscript. K. Piatt provided invaluable assistance throughout the course of this study. Financial support from the Department of Biology at The College of Charles- ton, the Charleston Higher Education Consortium, the Slocum-Lunz Foundation and Sea Grant #NA80AA-0091 to G. Helfman is gratefully ac- knowledged. Portions of this manuscript were taken from a thesis submitted in partial fulfillment of the requirements for the M.S. degree. This is contribu- tion No. 222 from the Marine Resources Research Institute and No. 71 from the Grice Marine Bio- logical Laboratory. LITERATURE CITED Akaoka, T. 1971. The purse seine fishery in Japan, /n H. Kristjonsson (editor), Modern fishing gear of the world. Vol. Ill, p. 161- 163. Fishing News Ltd., Lend. Allen, L. G., and E. E. DeMartinl 1983. Temporal and spatial patterns of nearshore distribu- tion and abundance of the pelagic fishes off San Onofre- Oceanside, California. Fish. Bull., U.S. 81:569-586. Aprieto, V. L. 1974. Early development of five carangid fishes in the Gulf of Mexico and the south Atlantic coast of the United States. Fish. Bull., U.S. 72:415-444. Atkinson, L. P., T. N. Lee, J. 0. Blanton, and W. S. Chandler. 1983. Climatology of the southeastern United States con- tinental shelf waters. J. Geophys. Res. 88:4705-4718. Bagenal, T. B., and E. Braum. 1978. Eggs and early life history. In T. B. Bagenal (editor), Methods for assessment of fish production in fresh waters, p. 165-201. Blackwell Sci. Publ., Lend. Berry, F. H. 1968. A new species of carangid fish (Decapterus tabl ) from the western Atlantic. Contrib. Mar. Sci. 13:145-167. Blanton, J. 0., L. P. Atkinson, L. J. Pietrafesa, and T. N. Lee. 1981. The intrusion of Gulf Stream water across the continen- tal shelf due to topographically induced upwelling. Deep- Sea Res. 28:393-405. Blaxter, J. H. S., and J. R. Hunter. 1982. The biology of clupeoid fishes. Adv. Mar. Biol, Acad. Press, N.Y., 20:1-224. 265 FISHERY BULLETIN: VOL. 85, NO. 2 Bliss, C. I. 1967. Statistics in biology, Vol. I. McGraw Hill, N.Y., 558 p. BoELY, T., A. Wysokinski, and J. Elwertowskl 1973. Les chinchards des cotes Senegaloises et Mauritan- iennes. [In Fr.] Trav. Doc. ORSTOM Dakar-Thiaroze 43, 74 p. Brewer, G. D., and G. S. Kleppel. 1986. Die! vertical distribution of fish larvae and their prey in nearshore waters of southern California. Mar. Ecol. Prog. Ser. 27:217-226. Brothers, E. B. 1979. Age and growth studies on tropical fishes. In S. B. Saila and P. M. Roedel (editors). Stock assessment for tropical small-scale fisheries, p. 119-135. Univ. Press, Kingston, RI. BULLIS, H., AND p. StRUHSAKER. 1970. Fish fauna of the western Caribbean upper slope. Q. J. Fla. Acad. Sci. 33:43-76. Campana, S. E. 1985. Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Carr, W. E. S., and C. a. Adams. 1972. Food habits of juvenile marine fishes: evidence of the cleaning habit in the leatherjacket, Oligoplites saurus, and the spottail pinfish, Diplodus holbrooki. Fish. Bull., U.S. 70:1111-1120. Chester, A. J., Huntsman, G. R., Tester, P. A., and C. S. Manooch, ni. 1984. South Atlantic Bight reef fish communities as repre- sented in hook-and-line catches. Bull. Mar. Sci. 34:267- 279. Clark, S. H., and B. E. BROv^fN. 1977. Changes in biomass of finfishes and squids from the Gulf of Maine to Cape Hatteras, 1963-74, as determined from research vessel survey data. Fish. Bull., U.S. 75:1-21. Cochran, W. C. 1977. Sampling Techniques. John Wiley and Sons, N.Y., 229 P- CONOVER, D. 0. 1985. Field and laboratory assessm.ent of patterns in fecun- dity of a multiple spawning fish: the Atlantic silverside Menidia menidia. Fish. Bull, U.S. 83:331-341. Deevey, G. B. 1960. Plankton studies. Bull. Bingham Oceanogr. Collect., Yale Univ. 17:1-86. DeMartini, E. E., and R. K. Fountain. 1981. Ovarian cycling frequency and batch fecundity in the queenfish, Seriphus politus: attributes representative of serial spawning fishes. Fish. Bull., U.S. 79:547-560. de Silva, S. S. 1973. Food and feeding habits of the herring Clupea haren- gus and sprat C. sprattus in inshore waters of the westcoast of Scotland. Mar. Biol. (Berl.) 5:689-705. DURBIN, A. G. 1979. Food selection by plankton feeding fishes. In H. Clep- per (editor). Predator-prey systems in fishery management, p. 203-218. Sport Fishing Inst., Wash., D.C. Elliott, J. M. 1971. Statistical analysis of samples of benthic invertebrates. Freshwater Biological Association of the United Kingdom, Publication #25, 144 p. Fahay, M. p. 1085. An annotated list of larval and juvenile fishes captured with surface-towed meter nets 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. Feder, H. M., C. H. Turner, and C. Limbaugh. 1974. Observations on fishes associated with kelp beds in Southern California. Calif. Dep. Fish Game, Fish Bull. 160, 144 p. Grosslein, M. D. 1969. Groundfish survey of the Bureau of Commercial Fish- eries, Woods Hole. Commer. Fish. Rev. 31(7):22-35. Grossman, G. D., M. J. Harris, and J. E. Hightower. 1985. The relationship between tilefish, L 310 cm): Immature - TL Mature - TL 17.7 CR + 20.18 (n = 44, r = 0.972) 7.6 CR + 190.21 {n = 20, r = 0.796). Figure 1.— Relationship between fork length and total length for Galeocerdo cuvieri taken in the Gulf of Mexico and off the Virginia coast. E o ^-^ 300 -J LL X h- o 200 z 111 _J ^ cc o 100 LL Galeocerdo cuvieri FL = .860(TL) - 11.5 rr.996 n-99 _] L _l L J I L 100 200 300 400 TOTAL LENGTH (TL) (cm) 271 FISHERY BULLETIN: VOL. 85, NO. 2 400 350 ^-^ E S 300 5 250 Z LU _j 200 -J < H 150 O 100 50 Galeocerdo cuvieri •^> 8 10 12 14 16 CENTRUM RADIUS 18 20 22 24 Figure 2.— Relationship of centrum dorsal radii to total length for Galeocerdo cuvieri taken in the Gulf of Mexico and off the Virginia coast. Centrum radii measurements are in ocular micrometer units (omu). 1 omu =1.2 mm. See text for discussion of the different regressions. Neonatal tiger sharks had only one annulus. Back calculations of length at the formation of this an- nulus indicated that it was formed at birth. Prebirth marks, which formed at placentation (Radtke and Cailliet 1984; Casey et al. 1985; Branstetter 1987c; Branstetter and Stiles in press), were not found in this aplacentally developing species; a condition also noted for the aplacental Alopias vulpinus (Cailliet et al. 1986). Marginal increment analysis on all but neonatal tiger sharks (Fig. 3) indicated that the annuli formed in late fall or early winter (October-December) be- came visible off the centrum edge by January and were farthest from the centrum edge in summer. This "winter" annulus was consistent throughout the size range of the sample (Beamish and McFar- lane 1983). Therefore, the first band bordered by the birth mark and the first winter annulus repre- sented approximately 6 months growth; remaining bands formed annually. Annuli along the periphery of centra in large (old) tiger sharks were closely spaced, making counts for these individuals more difficult. Annulus counts between the two readings were identical except for some of the larger individuals. In these cases, results of a third count matched one of the two previous counts, and this was the value accepted. c (0 f tof- w CO c o E 0) o "cO c 'O) k- cO .8 .6- .4 .2- •• ••• • •• •• J ' F M" AM' J J A' S O N D Figure 3.— Marginal increment widths as a ratio of the width of the last fully formed band in vertebral centra of Galeocerdo cuvieri compared by month. Specimens from the two regional samples are combined. The two regional samples of tiger sharks exhibited similar growth rates. By combining observed length at age data from both samples a single von Berta- lanffy curve could be fitted by using Fabens (1965) 272 BRANSTETTER ET AL.: AGE AND GROWTH OF TIGER SHARK Table 1 . — Back calculations by age class for the Virginia and Gulf of Mexico samples of tiger sfiarks, Galeocerdo cuvieri. Ages are based on age at the formation of the winter mark. Lengths to nearest cm TL. Significantly different mean lengths at age between samples in- dicated by asterisks ('* P < 0.001; * P < 0.01). Winter n Age at the formation of the ! winter mark mark B 0 + 1 + 2 + 3 + 4 + 5 + 6 + 7 + 8 + 9 + 10 + 11+ 12+ 13 + 14 + 15 + 0 1 0 2 76 126 2 2 74 130 156 3 6 73 125 159 185 4 9 73 122 161 189 209 5 1 84 126 160 192 219 237 ATLANTIC 6 2 78 135 178 205 223 244 262 7 2 75 120 151 186 218 243 259 274 8 2 82 118 163 198 224 242 261 278 294 9 1 73 122 148 165 196 224 249 271 290 309 10 5 72 116 154 186 211 229 250 270 290 305 314 11 6 72 121 159 184 211 239 262 280 300 316 326 336 12 1 78 118 144 176 211 230 245 270 284 296 311 322 330 13 1 73 128 183 219 238 256 284 302 317 326 335 348 356 366 14 1 62 107 153 187 207 253 270 281 299 314 324 337 345 351 362 15 2 77 123 158 189 214 236 261 287 301 313 325 334 346 355 364 372 16 1 72 111 150 180 198 214 244 260 274 285 304 318 331 346 359 364 378 X 74 122 159* * 188* * 213* * 237* * 258* * 277* * 295* 310* 321 334 342 355 362 369 378 cm/yr '(48) 37 29 25 24 21 19 18 15 11 13 8 13 7 7 9 cm/yr '(49) 51 35 29 24 18 18 17 16 13 9 X 73 122 173* * 208* * 237* * 261* * 279* * 297* * 314* 330* 343 352 0 1 77 1 4 70 114 2 3 74 129 169 3 5 74 125 180 212 4 2 80 124 170 214 241 I GULF 5 2 77 125 169 208 235 267 6 1 68 115 166 198 227 256 281 7 2 74 130 172 197 233 249 274 288 8 1 61 105 179 202 229 250 275 300 307 9 2 74 126 178 209 237 263 280 301 317 329 10 0 — — — — — — — — — — — 11 1 68 113 166 215 254 283 290 306 316 333 343 352 'Six months growtti. procedure (Fig. 4) to adequately describe the growth rate. However, young age classes in the Gulf of Mex- ico (hereafter referred as Gulf) sample were slight- ly larger at age than their Atlantic counterpart. Independent von Bertalanffy curves for each data set had different parameter estimates. Regressions of the curves, linearized by log transformation, were analyzed for covariance (SAS Institute 1985) and were significantly different (P < 0.0001). Simul- taneous nonlinear regression analysis of the two von Bertalanffy curves derived from back-calculated mean lengths at age produced parameter estimates with nonoverlapping simultaneous confidence inter- vals (Bernard 1981). Back calculations by age class for each sample (Table 1) also showed the more rapid growth of the Gulf juveniles. Mean lengths at the formation of the winter annuli were significantly different (P < 0.01 or P < 0.001) for early age classes of the two samples. Neonates in both samples increased near- ly 50 cm in length the first 6 months, and the Age I Gulf tiger sharks continued to grow at 50 cm/year, but Atlantic Age I individuals grew <40 cm/year. Gulf tiger sharks continued to grow approximately 4 cm/year faster than the Atlantic population until the fourth year. Growth rates then became similar; Gulf tiger sharks were simply larger at age. Observed and back-calculated lengths at age (Table 2) corresponded within each sample. Com- parisons of observed and back-calculated lengths did not indicate the occurrence of Lee's phenomenon. Differences in observed and back-calculated lengths at age were attributable to the fact that most speci- mens in both samples were taken in summer; there- fore, observed lengths at age were larger than back- calculated lengths at age based on the winter formed annuli. The growth rate estimated from centra was val- idated with one tag-recapture. A female, tagged 3 November 1978, was estimated to be 230 cm, and 273 FISHERY BULLETIN: VOL. 85, NO. 2 £ o ^^ X I- o z LU < I- o I- 450 400 350 300 250 200 150 100 Galeocerdo cuvieri ? c? 9 O ^ ir Gulf of Mexico O • Atlantic coast •• U« o , • o •* o o • ■sir • • * T von Bertalanffy parameters 1 g Gulf K =.184- L„=388 t» = -1.13 Atlantic .107 440 -2.35 Combined !l58 388 -1.73 Winter marks 5 J , i_ 9 11 13 15 J I I I I r I L 7 9 11 AGE (yr) 13 I 17 Figure 4.— Length at age for Galeocerdo cuvieri from the Gulf of Mexico and the Atlantic coast of Virginia. In- dividuals are plotted by their estimated actual ages (time elapsed since formation of the last winter mark). Birthdays set at 1 June. at recapture, 7 April 1984, was estimated to be 320 cm from the weight/length relationship. The tiger shark grew 90 cm in 5.4 years. By using the age/ length relationship estimated by the growth curve (Fig. 4), the shark would have been 3.4 years of age when tagged and 8.8 years of age at recapture. Even with the relatively rapid growth rate ex- hibited by this species, a length-frequency analysis for both samples (Fig. 5) did not distinguish age classes. The size distribution did indicate that young juvenile tiger sharks occur only rarely in the Virginia region. Males matured at approximately 310 cm, females at 315-320 cm, and the differences in growth rates between the two samples meant that they reached maturity at different ages. For the Gulf of Mexico, the smallest mature males (310, 311 cm) were 8.0 and 7.8 years old. The largest male aged (340 cm) was only 8.8 years old. Back calculations indicated that this individual grew relatively rapidly compared with smaller individuals in the sample, and the only larger male collected (363 cm) was not aged. The smallest mature female (325 cm) was 8.8 years old, the largest (355 cm) was 11.2 years old. For the Atlantic sample, two immature males (310, 311 cm) were not aged, but a 312 cm mature male was 10.1 years old. The largest male (381 cm) was 15.1 years old. The largest immature female (307 cm) was 8.1 years of age, the smallest mature females (318, 319 cm) were 9.0 and 11.1 years of age, and the largest female (381 cm) was 16.1 years old. The rapid linear growth early in life did not cor- respond to a great increase in the weight of the in- dividuals (Fig. 6). Growth from the third through the seventh winter decreased from 30 to 20 cm/year, and weights increased during this period. As the 274 BRANSTETTER ET AL.; AGE AND GROWTH OF TIGER SHARK Table 2.— Comparison of length at age for observed and back-calculated data for cuvieri. Lengths are to the nearest cm TL. Values indicate Atlantic and Gulf populations of the tiger shark, Galeocerdo low-mean-high (n) for each age class. Winter mark Age 0 0 1 0 + II 1 + III 2 + IV 3 + V 4-1- Gulf observed back calculation Atlantic observed back calculation 91-99-106 50-73-85 NA(0) 60-74-84 (2) (25) (44) 100-121-140 (4) 96-122-137 (23) 125-140-155 (2) 101-122-149 (44) 150-179-199 (3) 149-173-184 (19) 156-165-173 (2) 138-159-188 (42) 205-220-240 (5) 192-208-228 (16) 180-192-225 (6) 161-188-220 (40) 248-249-250 (2) 225-237-254 (11) 205-216-229 (9) 183-213-238 (34) 278-279-279 (2) 239-261-283 (9) 237 (1) 202-237-256 (25) Winter mark Age- VI 5-H Vil 6 + VIII 7-1- IX 6 + X 9■^ XI 10-H Gulf observed back calculation Atlantic observed back calculation 288 (1) 272-279-290 (7) 250-276-302 (2) 221-258-284 (24) 285-298-310 (2) 283-297-306 (6) 278-282-286 (2) 245-277-302 (22) 311 (1) 307-314-318(4) 292-300-307 (2) 270-295-317 (20) 325-333-340 (2) 325-330-333 (3) 318 (1) 292-310-327(18) NA (0) 343 (1) 307-322-335 (5) 304-321-341 (17) 355 (1) 352 (1) 319-341-354 (6) 315-334-349 (12) Winter mark: Age: XII 11 + XIII ^2 + XIV 13-1- XV 14-h XVI 15-H Atlantic observed back calculation 338 (1) 340-343-356 (6) 368 (1) 346-355-366 (5) 368 (1) 359-362-369 (4) 370-376-381 (2) 364-369-378 (3) 381 (1) 378 (1) (0 z LU o LU CL (/) LL O CC LU m TOTAL LENGTH (cm) Figure 5.— Length frequency of Galeocerdo cuvieri collected off Virginia and in the Gulf of Mexico. Specimens are grouped into 10 cm size classes. 275 FISHERY BULLETIN: VOL. 85, NO. 2 400 350 300 ? 250 X 200 LLI ^ 150 100 50 Galeocerdo cuvieri log wt. = 3.24(log TL) - 5.85 wt. = 1.41 X lO'^TL^'^") n = 120 r = .944 50 100 150 200 250 300 350 400 TOTAL LENGTH (cm) Figure 6.— Weight/length relationship for Atlantic and Gulf of Mexico Galeocerdo cuvieri, sexes combined. animals matured at 310-320 cm, linear growth slowed from 15 cm/year to <10 cm/year, and weights increased dramatically. DISCUSSION An isometric relationship between centrum growth and length has been noted for many shark genera (Cailliet et al. 1983; Gruber and Stout 1983; Bran- stetter and McEachran 1986). The slight curvilinear relationship between centrum growth and length noted for Galeocerdo cuvieri suggested there were two distinct growth stanzas. A similar relationship was also noted for hums oxyrinchus (Pratt and Casey 1983). The point of inflection in the curve is generally at the length corresponding to the onset of maturity, a decreased linear growth rate and an increased weight gain rate. Apparently, centrum growth is correlated to the structural support neces- sary for length increases, but an increasing rate of weight gain does not require additional strengthen- ing of the vertebral column. Marginal increment analysis of annulus periodicity demonstrated that one growth band, consisting of one calcified opaque zone and one less calcified translucent zone, formed annually. A similar period- icity for growth bands or annuli has been verified for several shark genera (Gruber and Stout 1983; Cailliet et al. 1986; Branstetter and McEachran 1986) and validated using tetracycline injected Negaprion brevirostris (Gruber and Stout 1983), Triakis semifasciatus (Smith 1984), Rhizoprionodon terraenovae, and Carcharhinus plumbeus (Bran- stetter 1987a). In contrast, Parker and Stott (1965) and Pratt and Casey (1983) provided evidence that lamnoids produce two band pairs per year, and Natanson (1984) could find no regular periodicity in centrum bands of Squatina californica. Our estimates indicated the tiger shark doubles in length the first year of life. This is supported by growth of a full-term embryo (69 cm) placed in an aquarium by Clark and von Schmidt (1965) on 21 May, where it survived 12 weeks growing to 89 cm. Rapid linear growth for juvenile tiger sharks may be necessary for adequate cohort survival. With a 13-16 mo gestation period (Clark and von Schmidt 1965) and a mating season which occurs before full- term females have pupped, the female reproductive 276 BRANSTETTER ET AL.: AGE AND GROWTH OF TIGER SHARK cycle is at least 2 years. Considering the litter size (40-70 pups) (Kauffman 1950; Bass et al. 1975; Bran- stetter 1981), natural mortality must be high for young age classes. Pups are born in coastal waters at a relatively large size (>70 cm) which reduces some predation, but the elongate, flexible body pro- duces an inefficient anguilliform swimming motion. Additionally, early in life, the caudal fin is extremely flexible and has a low thrust angle (Thompson and Simanek 1977). The combination of these charac- teristics precludes rapid swimming speeds, thus making the pups vulnerable to predation by the abundant coastal sharks including their own species. Not only does rapid linear growth make them larger than most potential predators, it may help decrease predation by increasing swimming efficiency and speed through increased body rigidity (producing a more carangiform motion) and increased caudal fin thrust angle. Linear growth continues at >20 cm/year until the tiger sharks are near maturity. Such rapid growth is similar to that noted for several lamnoids (Parker and Stott 1965; Gruber and Compagno 1981; Pratt and Casey 1983; Cailliet et al. 1985), but contrasts sharply to the slow growth rates estimated for several carcharhinids and sphyrnids (Thorson and Lacy 1982; Gruber and Stout 1983; Schwartz 1983a). Even the more rapidly growing carchar- hinids do not have such large relative increases in length (Parsons 1985; Branstetter and McEachran 1986). The mean lengths at age between the Gulf of Mex- ico and Atlantic tiger sharks were significantly different, and probably represent ecophenotypic dif- ferences between the two regions. However, the two regional groups are not isolated. Our one tag- recapture was tagged off Mobile Bay, AL and recap- tured in the Florida Straits off Havana, Cuba, and there are similar tag returns of tiger sharks that moved between the Gulf of Mexico and the Atlan- tic (J. Casey pers. commun."*). However, long migra- tions between the two regions may be restricted to larger individuals with juveniles remaining in their respective regions. If juvenile tiger sharks do remain in their respec- tive regions early in life, growth rate differences between the two regions may be caused by dif- ferences in early life histories. In the Gulf of Mex- ico, the pups apparently only migrate short distances inshore-offshore seasonally. In the Atlantic, the pups ^J. Casey, Northeast Fisheries Center Narragansett Laboratory, National Marine Fisheries Service, NOAA, South Ferry Road, Narragansett, RI 02882, pers. commun. June 1986. are born south of Cape Hatteras, probably in the Florida region (Dodrill 1977). These neonates may not migrate north during their first year, as small individuals, <150 cm, are rare in the Virginia region (Fig. 5). During this time, the growth rates for both groups are similar. The extensive northern migration for l-i- year old Atlantic juveniles, 150-200 cm, may be energetically costly, hinder- ing growth. Therefore, the Gulf young that do not migrate great distances are able to attain greater lengths during this time period. The increased swimming efficiency attained with lengths >250 cm could possibly explain why growth rates become similar. For juveniles of both regions, the energy require- ments for the inefficient swimming motion and rapid linear growth apparently restrict any great increase in weight (Fig. 6). Only after the tiger sharks reach lengths >200 cm (3 -i- years of age) does weight in- crease substantially, and correspondingly linear growth begins declining. After reaching maturity (310-320 cm) linear growth is <10 cm/year while weight growth is substantial, corresponding to the change in centrum radius/length relationship (Fig. 3). The von Bertalanffy parameter estimates for the two collections closely bracket known life history characteristics. With sexes combined, the L^ for the Gulf of Mexico collection and for both samples combined (388 cm) is smaller than many reported large individuals, but is a reasonable compromise between the maximum reported lengths for males and females: 419 cm individual (McCormick et al. 1964); 370 cm male, 410 cm female (Bass et al. 1975); 410 cm female (Branstetter 1981); and a 381 cm male and female from this study. However, the tiger shark is thought to attain lengths in excess of 450 cm (Bigelow and Schroeder 1948; Castro 1983), more in agreement with the L^ for the Atlantic sample (440 cm). The t^ value for the Gulf sample (-1.13 years) is accurate, but the 13-16 mo gesta- tion period is overestimated for the Atlantic sam- ple (-2.35 years). The ^o value for many shark species overestimates the gestation period (Casey et al. 1985; Branstetter 1986). The iiT values for each analysis reflect the rapid growth rate of this species and are similar to some of the more rapidly grow- ing Carcharhinus species such as C. limbatus, C. brevipinna (Branstetter 1987c), C. falciformis (Branstetter 1987b), and C. acronotus (Schwartz 1983b). At the estimated growth rate for the largest in- dividuals (5-10 cm/year), exceptionally large speci- mens, 400-450 cm, would be 20-25 years of age. The 277 von Bertalanffy curve using observed lengths at age produced an estimated age at L^ for the Gulf sample of 28 years, and 37 years for the Atlantic sample. This would mean that the species matures at 30-50% of its maximum age, and with a reproduc- tive cycle of greater than 2 years, a female would reproduce less than 10 times. On the other hand, von Bertalanffy curves derived using back- calculated lengths at age for both samples produced estimated ages at L^ of 45-50 years. Exceptional- ly high ages at L^ may be due to the exponential function of the model, or it is also possible that as tiger sharks attain sizes near their maximum weight or length, centrum growth and band formation do not accurately represent age. Because no excep- tionally large individuals were aged, we are unable to determine which is the case. Even so, the data indicate that the tiger shark is long-lived with a relatively low fecundity, and natural mortality for the young may be high. As with many other elasmo- branchs, this combination of /^'-selected character- istics may result in an overexploitation of this species under increased recreational and commer- cial fishing pressure (Musick and Colvocoresses 1986). ACKNOWLEDGMENTS Appreciation is extended to commercial longline vessel owners and captains (Hollis Forrester, Bill Templeton, Lacy Smith, Arnold Davila, Olin Wel- born. Buddy and Tinker Lindley, Joe Gayman, Henry Jones, and Floyd Condit), Mark's Sharks, Inc., the Biggs Pensacola Shark Rodeo, the Dauphin Island Deep Sea Rodeo, the Galveston Monster Fishermen Club Shark Tournament, and the Vir- ginia Beach Sharkers for allowing access to and ex- amination of numerous tiger sharks. We thank Marine Advisory Service personnel Gary Graham and Tony Reisinger for their help during the course of this study. Debbie Branstetter provided con- tinuous laboratory and field assistance in the Gulf of Mexico study, and John Gourley and Marta Nam- mack provided invaluable aid in the Virginia pro- gram. The senior author wishes to acknowledge the support and encouragement provided by John D. McEachran. Different portions of this study were supported by the Gulf and South Atlantic Fishery Development Foundation, Inc., the Texas A&M Sea Grant College Program, the Virginia Sea Grant Program, the Graduate College and Association of Former Students of Texas A&M University, and the State of Texas. FISHERY BULLETIN: VOL. 85, NO. 2 LITERATURE CITED Bass, A. J., J. D. D' Aubrey, and N. Kistnasamy. 1975. Sharks of the east coast of southern Africa. III. The families Carcharhinidae (excluding Cmrcharhinus and Muste- lus) and Sphyrnidae. Oceanogr. Res. Int. (Durban) Invest. Rep. 38, 100 p. Beamish, R. J., and G. A. McFarlane. 1983. The forgotton requirements for age validation in fish- eries biology. Trans. Am. Fish. Soc. 112:735-743. Bernard, D. R. 1981. Multivariate analysis as a means of comparing growth in fish. Can. J. Fish. Aquat. Sci. 38:233-236. Bigelow, H. B., and W. C. Schroeder. 1948. Sharks. In Fishes of the Western North Atlantic. Part I, No. 1, p. 59-546. Mem. Sears Found. Mar. Res., Yale University. Branstetter, S. 1981 . Biological notes on the sharks of the north central Gulf of Mexico. Contrib. Mar. Sci. 24:213-34. 1986. Biological parameters of the sharks of the northwestern Gulf of Mexico in relation to their potential as a commercial fishery resource. Unpubl. Ph.D. Thesis, Texas A&M University, 138 p. 1987a. Age and growth validation of new born sharks held in laboratory aquaria, with comments on the life history of the Atlantic sharpnose shark, Rhizoprionodon terraenovae. Copeia 1987:293-300. 1987b. Age, growth and reproductive biology of the silky shark, Carcharhinus fakiformis, and the scalloped hammer- head, Sphyrna lewini, from the northwestern Gulf of Mex- ico. Environ. Biol. Fishes 19:161-174. 1987c. Age and growth estimates for blacktip, Carcharhinus limbatus, and spinner, C. brevipinna, sharks from the north- western Gulf of Mexico. Copeia 1987. Branstetter, S., and J. D. McEachran. 1986. Age and growth of four carcharhinid sharks common to the Gulf of Mexico: a summary paper. In T. Uyeno, R. Aral, T. Taniuchi, and K. Matsuura (editors), Indo-Pacific Fish Biology: Proceedings of the Second International Con- ference on Indo-Pacific Fishes, p. 361-371. Ichthol. Soc. Jpn., Tokyo. Branstetter, S., and R. Stiles. In press. Age and growth estimates for the bull shark, Car- charhinus leucas, in the northern Gulf of Mexico. Environ. Biol. Fishes. Cailliet, G. M., L. K. Martin, J. T. Harvey, D. Kusher, and B. A. Welden. 1983. Preliminary studies on the age and growth of the blue, Prionace glauca, common thresher, Alopias vulpinus, and shortfin mako, Isums oxyrinchus, sharks from California waters. In E. D. Prince and L. M. Pulos (editors), Pro- ceedings of the International Workshop on Age Determina- tion of Oceanic Pelagic Fishes: Tunas, Billfishes, and Sharks, p. 179-188. U.S. Dep. Commer. NOAA Tech. Rep. NMFS 8. Cailliet, G. M., L. J. Natanson, B. A. Welden, and D. Ebert. 1985. Preliminary studies on the age and growth of the white shark, Carcharodon carcharias, using vertebral bands. South. Calif. Acad. Sci. Mems. 9:49-60. Cailliet, G. M., R. L. Radtke, and B. A. Welden. 1986. Elasmobranch age determination and verification: a review. In T. Uyeno, R. Aral, T. Taniuchi, and K. Matsuura (editors), Indo-Pacific Fish Biology: Proceedings of the Sec- ond International Conference on Indo-Pacific Fishes, p. 345-360. Ichthol. Soc. Jpn., Tokyo. 278 BRANSTETTER ET AL.: AGE AND GROWTH OF TIGER SHARK Carlander, K. D. 1969. Handbook of freshwater fishery biology. Volume 1. Iowa St. Univ. Press, Ames, 752 p. Casey, J. G. 1964. Angler's guide to the sharks of the northeastern United States, Maine to Chesapeake Bay. U.S. Fish Wildl. Serv. Circ. 179, 32 p. Casey, J. G., H. L. Pratt, Jr., and C. E. Stillwell. 1985. Age and growth of the sandbar shark (Carcharhinus plumbeus) from the western North Atlantic. Can. J. Fish. Aquat. Sci. 42:963-975. Castro, J. I. 1983. The sharks of North American waters. Texas A&M Univ. Press, College Station, 180 p. Clark, E., and K. von Schmidt. 1965. Sharks of the central Gulf coast of Florida. Bull . Mar . Sci. 15:13-83. DODRILL, J. W. 1977. A hook and line survey of the sharks found within five hundred meters of shore along Melbourne Beach, Brevard County, Florida. Unpubl. MS Thesis, Fla. Inst. Tech., 304 P- DODRILL, J. W., AND R. G. GiLMORE. 1978. Land birds in the stomachs of tiger sharks, Galeocer- do cuvieri (Peron and Lesueur). Auk 95:585-586. Fabens, a. J. 1965. Properties and fitting of the von Bertalanffy growth curve. Growth 29:265-289. GrUBER, S. H., and L. J. V. COMPAGNO. 1981. Taxonomic status and biology of the bigeye thresher, Alo-pias superciliosus. Fish. Bull., U.S. 79:617-640. Gruber, S. H.,, AND R. G. Stout. 1983. Biological materials for the study of age and growth in a tropical marine elasmobranch, the lemon shark, Nega- prion brevirostris (Poey). InE.D. Prince and L. M. Pulos (editors). Proceedings of the International Workshop on Age Determination of Oceanic Pelagic Fishes: Tunas, Billfishes, and Sharks, p. 193-205. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 8. Gudger, E. W. 1949. Natural history notes on tiger sharks, Galeocerdo tigrinis, caught at Key West, Florida, with emphasis on food and feeding habits. Copeia 1949:39-47. Kauffman, D. E. 1950. Notes on the biology of the tiger shark {Galeocerdo ar- ticus) from Philippine waters. U.S. Fish Wildl. Serv. Res. Rep. 16, 10 p. McCoRMiCK, H. W., T. Allen, and W. Young. 1963. Sharks on a line. In Shadows in the sea - the sharks, skates, and rays. Chap. 4, p. 89-114. Weathervane Books, N.Y. MUSICK, J. A., AND J. A. Colvocoresses. In press. Seasonal recruitment of subtropical sharks in Chesa- peake Bight USA. In A. Yanez-Anancibia and D. Pauly (editors), Proceedings of a workshop in tropical coastal demersal communities. Carman, Campeche, Mexico, 1986. MusicK, J. A., J. A. Colvocoresses, E. F. Lawler, and W. G. Raschi. 1985. Distribution of sharks in Chesapeake Bight. [Abstr.] Proceedings of the 65th Annual Meeting of the American Society of Ichthyologists and Herpetologists. Natanson, L. J. 1984. Aspects of the age, growth, and reproduction of the Pacific angel shark, Squatina califomica, off Santa Barbara, California. Unpubl. MA Thesis, San Jose St. Univ., San Jose, CA 71 p. Parker, H. W., and F. C. Stott. 1965. Age, size and vertebral calcification in the basking shark, Cetorhinus maximus (Gunnerus). Zool. Meded. 40: 305-319. Parsons, G. R. 1985. Growth and age estimation of the Atlantic sharpnose shark, Rhizoprionodon terraenovae: a comparison of tech- niques. Copeia 1985:80-85. Pratt, H. L., Jr., and J. G. Casey. 1983. Age and growth of the shortfin mako, Isurus oocyrin- chvs, using four methods. Can. J. Fish. Aquat. Sci. 40:1944- 1957. Radtke, R. L., and G. M. Cailliet. 1984. Age estimation and growth of the gray reef shark, Car- charhinus amblyrhynchos, from the northwestern Hawaiian Islands. In Proceedings Second Symposium on Resource Investigations of the Northwestern Hawaiian Islands, p. 121- 127. UNI-SEAGRANT-MR-84-01. SAS Institute, Inc. 1985. SAS User's Guide: Basics, SAS User's Guide: Statis- tics, SAS/Graph Software Version 5. SAS Institute Inc., Gary, NC. Schwartz, F. J. 1983a. Shark ageing methods and age estimation of scalloped hammerhead, Sphyma lewini, and dusky, Carcharhinus obscurus, sharks based on vertebral ring counts, /n E. D. Prince and L. M. Pulos (editors. Proceedings of the Inter- national Workshop on Age Determination of Oceanic Pelagic Fishes: Tunas, Billfishes, and Sharks, p. 167-174. U.S. Dep. Commer. NOAA Tech. Rep. NMFS 8. 1983b. Occurrence, abundance, and biology of the blacknose shark, Carcharhinus acronotus, in North Carolina. North- east Gulf Sci. 7:29-47. Smith, S. E. 1984. Timing of vertebral-band deposition in tetracycline in- jected leopard sharks. Trans. Am. Fish. Soc. 113:308-313. Snedacor, G. W., and W. G. Cochran. 1980. Statistical methods. Iowa State Univ. Press, Ames, 507 p. Springer, S. 1960. Natural history of the sandbar shark, Eulamia mil- berti. U.S. Fish Wildl. Serv. Fish. Bull. 61:1-38. 1963. Field observations on large sharks of the Florida- Caribbean region. In P. W. Gilbert (editor). Sharks and siu-- vival. Sect. II. Ch. 3, p. 95-113. Heath and Co., Boston. Thompson, K. S., and D. E. Simanek. 1977. Body form and locomotion in sharks. Am. Zool. 17: 343-354. Thorson, T. B., and E. J. Lacy, Jr. 1982. Age, growth rate and longevity of Carcharhinus leitcas estimated from tagging and vertebral rings. Copeia 1982: 110-116. 279 ECOLOGICAL CONSEQUENCES OF MECHANICAL HARVESTING OF CLAMS Charles H. Peterson, Henry C. Summerson, and Stephen R. Fegleyi ABSTRACT A field experiment was performed in 1,225 m^ plots in each of two shallow estuarine habitats, a seagrass bed and a sand flat, in Back Sound, North Carolina (USA), to test the impact of clam raking and two different intensities of mechanical harvesting of clams ("clam kicking") for up to 4 years on 1) hard clam, Mercenaria mercenaria, recruitment, 2) seagrass biomass, 3) the density of benthic macroinverte- brates, and 4) the density of bay scallops, Argopecten irradians. The removal of adult hard clams with the contingent sediment disturbance had ambiguous effects on the recruitment of hard clams: in the sand flat recruitment tended to be lower (but not significantly) in intense-clam-kicking matrices than in controls, whereas in seagrass recruitment of hard clams did not not show a clear response to treat- ment. In the raking and light-clam-kicking matrices, seagrass biomass fell immediately by =25% below controls but full recovery occurred within a year. In the intense-clam-kicking matrices, seagrass biomass fell by =65% below levels expected from controls; recovery did not begin until more than 2 years passed, and seagrass biomass was still =35% lower than predicted from controls 4 years later. Clam harvest did not affect either the density or species composition of small benthic macroinvertebrates from sedi- ment cores, probably because of their rapid capacity for recolonization and generally short life spans. In all treatments, densities of benthic macroinvertebrates (mostly polychaetes) were substantially higher in the seagrass than in the sand flat during October samplings but equal during March samplings. Bay scallop density declined with declining seagrass biomass across harvest treatments, but the intense-clam- kicking matrices contained even fewer bay scallops than their seagrass biomass would predict, perhaps because of enhanced patchiness of the remaining seagrass. The relative inertia of the change in seagrass biomass following extensive destruction in the intense- ly kicked matrices suggests that seagrass replanting may be an extremely important means of returning disturbed, unvegetated areas to seagrass systems. Emergence during summer of a between-habitat gradient in infaunal densities (higher in seagrass than in sand) supports the hypothesis that seagrass provides a partial prey refuge for infaunal invertebrates. The failure of the benthic macroinvertebrate density to respond to clam harvest treatments in both sand flats and seagrass beds implies that the polychaetes which dominate recover rapidly from disturbance and are probably not adversely affected by clam harvest. The negative and long-lasting impact of intense hard clam harvest on seagrass biomass with its effects on other fisheries, including bay scallops, implies that hard clam fisheries should be managed to minimize the intensity of harvest within seagrass beds. Technological innovation is frequently accompanied by an increased risk of harm to various aspects of the natural environment (e.g., Dickie 1974). While such innovation can be considered economically desirable and even inevitable, environmental managers still require ecological inputs to enable them to reach properly informed compromises between uncontrolled application of new technology and unnecessarily cautious protection of natural ecosystems. Because of its inherent lack of general principles and paradigms, ecology is rarely able to provide immediate answers to practical questions of the probable impact of new technology. Conse- quently, careful studies of the ecological impact of the application of each specific new technology are 'Institute of Marine Sciences, University of North Carolina at Chapel Hill, Morehead City, NC 28557. Manuscript accepted January 1987. FISHERY BULLETIN: VOL. 85, NO. 2, 1987. often necessary. Such studies can not only provide necessary applied information but also contribute to a better basic understanding of the specific system that is being explored. Although fisheries biologists are renowned for managing harvests in a way that will sustain a max- imum yield or maximize yield per recruit (Ricker 1975), studies are only occasionally undertaken to compare the environmental damage caused by alter- native fishing gears and technologies (e.g., Caddy 1973; Peterson et al. 1983a). Such studies are most common in estuarine and other shallow-water fish- eries, where high coastal productivity of diverse stocks induces intensive exploitation of a common area by multiple, potentially interfering fisheries. As technological advances in fishing gear have been made, this potential for interfishery competition has grown, as has the need for understanding the envi- 281 FISHERY BULLETIN: VOL. 85, NO. 2 ronmental consequences of the utilization of new, alternative technologies. Fisheries for the hard clam, Mercenaria mercen- aria (L.), and other sedentary benthic invertebrates require the use of either hand implements (rakes, hoes, etc.) or boat-drawn gear (dredges, trawls, etc.). Managers of benthic invertebrate fisheries may turn to the subdiscipline of benthic ecology to seek predic- tions of the relative environmental and ecological consequences of utilizing various alternative fishing gears or of permitting technologically new substitu- tions for traditional fishing methodologies. Unfor- tunately, benthic ecologists are frequently unable to provide confident answers to many questions, often either because the fisheries applications in- volve a far larger scale than can be or has been prac- tically accommodated in basic experimental research designs or because the questions fall into an area of current debate and ongoing study in the basic science of the field. One might take, as an example of the poor predic- tive capacity of benthic ecology, the question of whether widespread adoption of mechanical har- vesters by commercial M. mercenaria fishermen will affect the future recruitment success of M. mercen- aria in the local area of harvest. Most fisheries biologists agree that the mechanical harvesters are more efficient in gathering hard clams from a given area and cause more physical disruption of the bot- tom than the alternative hand methods of raking and tonging. Even given these assumed differences, ben- thic ecology provides mixed and conflicting predic- tions of the impact of switching to mechanical harvesters. Basic studies of adult-larval interactions, including some among suspension-feeding bivalves (Woodin 1976; Williams 1980; Peterson 1982b), might suggest that removal of large, adult suspen- sion feeders would enhance the survivorship of settling larvae and thereby increase the recruitment success of M. mercenaria in the efficiently harvested areas. Yet, the experimental results on which such a prediction is based were achieved on a much smaller spatial scale and probably depend upon ab- solute density (or feeding rate) of all suspension feeders in an unspecified way. It is conceivable that the virtual removal of M. mercenaria over a substan- tial area might remove an important settlement cue (produced by adults) needed for larval habitat selec- tion (e.g.. Meadows and Campbell 1972; Gray 1974). If this were true, recruitment success of M. mercen- aria would decline with the intensity of harvest. Similarly, benthic ecology provides conflicting pre- dictions about the effects of the increased physical disburbance of mechanical harvesting on recruit- ment success of M. mercenaria. On the one hand, M. mercenaria recruits might be expected to suffer increased mortality from burial during massive sedi- ment disturbance (Rhoads 1974; Myers 1977; Thistle 1981; Wilson 1981). Yet, larvae of many species settle more densely into disturbed bottoms (Gray 1974; McCall 1977; Hulberg and Oliver 1980). Again, these signals are conflicting but, even more impor- tantly, experimental benthic ecology is unable to predict adequately whether the scale and intensity of disturbance during commercial clam harvesting are appropriate to invoke either of these processes. Because of the restricted scale of past field ex- periments and the consequent limitations of benthic ecology in the applied arena, we designed controlled field experiments to test the impact of mechanical clam harvesting on a large scale, sufficient to pro- vide environmental data to resource managers and to extend simultaneously the scope of basic experi- mental, benthic ecology. Specifically, we tested on a 1,225 m^ scale whether the harvest of M. mercen- aria, with its attendant physical disruption of the bottom, affected the 1) recruitment success of M. mercenaria, 2) biomass of seagrasses, 3) density of bay scallops, and 4) density of all other benthic macroinvertebrates. We tested these harvest effects in each of two common estuarine habitats, a sand flat and a seagrass bed, and followed not only the immediate response to harvesting but also the changes in most variables over a subsequent 3.5-yr period. Thus, the need for ecological data to use in fisheries management provided an opportunity to expand the temporal and spatial scale of experi- ments in marine benthic ecology and thereby eval- uate our ability to extrapolate from previous theory based on smaller scales. METHODS To test whether the type and/or intensity of hard clam, Mercenaria mercenaria (L.), harvest has any detectable effect on 1) its own recruitment, 2) sea- grass biomass, 3) bay scallop, Argopecten irradians, density, or 4) density of small benthic macroinverte- brates, we performed a large-scale field experiment at sites along the southern (barrier island) margin of Back Sound near Beaufort, NC (Fig. 1). This ex- periment was conducted in a seagrass meadow and in an unvegetated sand flat approximately 500 m to the west to permit a test of whether effects of harvest vary with habitat. This general area and its physical characteristics are described in several previous publications (Sutherland and Karlson 1977; Nelson 1979; Peterson et al. 1983b, 1984). Back 282 PETERSON ET AL.: IMPACT OF MECHANICAL CLAM HARVESTING 35' 34''45 40-- 75° 40 30' 25' Figure 1.— The locations of the study sites in eastern North CaroHna, near Cape Lookout. BSS indicates the sand-fiat and BSG the seagrass-bed locations. Tick marks on the margins of the figure denote minutes of N. latitude and W. longitude. Sound is a shallow marine lagoon with a lunar tide of about 0.6 m range, little salinity variation (28- 34''/oo), and a wide seasonal temperature range from a winter monthly minimum of 2°-4°C to a summer monthly maximum of 29°-30°C. In Jan- uary 1980, we selected in each habitat 6 square plots (matrices) of 1,225 m^ area, each of which had a virtually constant water depth of about 0.1-0.3 m at low tide and homogeneous surface appearance. Specifically, all seagrass matrices held a spatially uniform cover of a seasonally varying mixture of two seagrasses, eelgrass Zostera marina and shoalgrass Halodule wrightii, whereas no sand-flat matrix con- tained seagrasses. These seagrass matrices had been continuously vegetated from at least 1974 until 1980 and the seagrass cover had not extended over the sand flat during that same period (Peterson et al. 1984). Before harvest treatment, we subsampled all 6 matrices in each habitat to test whether there were any initial differences among matrices in response variables. This sampling occurred between 22 Feb- ruary and 31 March 1980 in the sand flat and from 1 April to 6 May 1980 in the seagrass bed. A fixed number (9 or 36) of uniformly distributed 0.25 m^ subsamples was taken from each matrix to estimate abundance of hard clams, bay scallops, and seagrass (Table 1). A uniform sampling array was chosen to reduce the field effort and to avoid risk of sampling at or even near (<1 m) the same locations during subsequent sampling. A grid of marked ropes at- tached to equally spaced stakes was placed around the circumference of each matrix and moved to a new, randomly chosen set of positions for each new sampling date, thus producing a "frame shift" of the sampling template. 283 FISHERY BULLETIN: VOL. 85, NO. 2 Table 1 .—Temporal design of data collections and of experimental treatments for both habitats, 1980-84. Entries are numbers of samples^ taken per matrix. Harvest Harvest Spring 1980 treatment Fall 1980 treatment Spring 1981 Fall 1981 Fall 1982^ Fall 1983^ Fall 1984^ Parameter 22 Feb.- 12-30 20 Oct- 19 Dec- 2-13 4 Oct.- 20-29 28-31 22-29 estimated 6 May May 10 Nov. 22 Feb. Mar. 3 Nov. Oct. Oct. Oct. Total hard clam density 36 Density of hard clam recruits 36 Seagrass dry mass 36 Bay scallop density 0 Density of benthic macro- invertebrates 6 Sediment size distribution parameters 3 36 36 36 36 36 9 36 9 9 0 9 36 9 9 9 9 36 9 9 0 6 6 0 0 0 0 0 0 0 0 'In all cases where 36 or 9 samples were taken per matrix, these were 'A m^ samples distributed uniformly across the matrix such that no sample fell within 1 m of any previous sample location. Where 6 or 3 samples were taken, these were chosen at random from a group of 9 uniformly distributed samples positioned in a similar way to avoid any overlaps. All sediment samples were cores of 5 cm diameter x 20 cm deep. Macroinvertebrate samples were cores of 10 cm diameter x 25 cm deep. ^Data taken from only the seagrass habitat on these dates. To collect a repeatable sample, we first inserted a 0.25 m^ circular metal sampling frame pene- trating to a depth of 15 cm and used an hydraulic suction dredge to excavate the complete contents to that same depth. The material was collected in a 3 mm nylon mesh bag (for description and sam- pling efficiency, see Peterson et al. 1983b). All living M. mercenaria and A. irradians were removed from the mesh bag and placed in separate, labeled plastic bags for return to the laboratory. For all M. mercen- aria we measured length in the longest antero- posterior dimension, and for all A. irradians we measured the distance from the flat top of the hinge to the ventral margin using vernier calipers. Sea- grass material from the mesh bag was packaged in marked plastic bags in the field and returned to the laboratory, where it was gently rinsed in freshwater to remove attached salt and sediments, and dried to constant weight (2-4 days) at 105°C. To estimate densities of small benthic macroinver- tebrates, we took 9 uniformly distributed samples from each matrix in each habitat on 4 sampling dates (Table 1). We processed and analyzed a randomly chosen subset of 6 of these 9 samples for each matrix. The strategy of taking more samples than one expects to analyze is optimal when marginal costs of additional sampling are low, because extra replicates are then available for later analysis if among-sample variation proves so unexpectedly high as to reduce statistical power to an unaccept- able level. Benthic invertebrates were collected using 10 cm diameter cores taken to a depth of 25 cm. Complete contents of each core were placed in separate plastic bags and gently sieved, in the lab- oratory, through 1 mm mesh. Sieve contents were held in bottles containing rose bengal in 10% buf- fered formalin until animal tissues were adequate- ly stained and hardened. We later picked and iden- tified to class (and to species in a subset of the samples) all animals in each sample. In spring 1980, we also took 8 randomly located sediment cores (5 cm in diameter to a depth of 20 cm) from each matrix to characterize initial sedi- ment conditions. Cores were transferred into in- dividual plastic bags and frozen at - 10°C until analysis of sediment size distribution by weight. We split each sample by coning and quartering (Ingram 1971) and then used standard Rotap dry sieving and pipetting procedures (Folk 1974) to estimate dry weights of sediments in each of several size classes. In addition, percent organic content was measured by weight loss on ignition at 550 °C for 4 h (Gross 1981). Because our (customary) use of small-diam- eter cores to sample sediments failed to include large shell fragments and because such biogenic calcium carbonate appeared to be extremely common in 1 seagrass matrix, we designed a sampling procedure to estimate the relative degree of coarse shell. In October 1985, we used the suction dredge to ex- cavate 3 haphazardly located 0.25 m^ quadrats to a depth of 12 cm in each of the 6 matrices in each habitat. All shell fragments collected on a 3 mm 284 PETERSON ET AL.: IMPACT OF MECHANICAL CLAM HARVESTING mesh were then cleaned with freshwater, dried at 60 °C, and weighed to provide a quantitative indica- tion of the relative degree of coarse shelliness in each matrix. After our initial sampling in spring 1980, we ap- plied harvest treatments on 2 occasions, 12-30 May 1980 and 19 December 1980-22 February 1981 with a single sampling of response variables in between (Table 1). We then sampled on 5 subsequent occa- sions to test for the existence and persistence of any treatment effects without applying any additional harvest treatments (Table 1). Of the 6 matrices in each habitat, 2 were left untouched as controls, 2 were given intense applications of "clam kicking", and the remaining 2 were subjected to lower but equal harvest intensities (judged by estimated per- centage of spring 1980 M. mercenaria removed) of different types ("clam kicking" in one and hand raking in the other). Clam kicking is a mechanical form of clam harvest (described in detail in Guthrie and Lewis 1982) practiced in North Carolina which involves the modification of boat engines in such a way as to direct the propeller wash downwards in- stead of backwards. The propeller wash is sufficient- ly powerful in shallow water to suspend bottom sediments and clams into a plume in the water col- umn, which allows M. mercenaria to be collected in a trawl net towed behind the boat (see Figure 2). To reproduce this process, we employed a commer- cial clam kicker and his boat. We measured in a crude way the relative intensity of the harvest treat- ment by counting all legally marketable (>2.54 cm in thickness in North Carolina) M. mercenaria removed and then estimating the percent removed of those available using the initial spring 1980 sam- pling (Table 2). We also recorded the number of Figure 2.— Aerial photograph of a clam kicking boat in operation, showing the sediment plume in the wake and the tracks of previous kicking passes in the surrounding bottom. 285 FISHERY BULLETIN: VOL. 85, NO. 2 Table 2.— The intensity of clam harvest treatments. All numbers and percents refer to legally harvested Mercenaha mercenaria >2.54 cm in thickness. Treatment date and parameter estimated May 1980 Winter 1980-81 Both applications pooled Est. % of Est. % of Est. % Of No. of spring 1980 Effort No. of spring 1980 Effort No. of spring 1980 Effort Habitat and clams clams required clams clams required clams clams required harvest treatment removed removed in harvest removed removed in harvest removed removed in harvest Sand flat Control 1 0 0 0 0 0 0 0 0 0 Control II 0 0 0 0 0 0 0 0 0 Raking 191 16 170 min 140 11 210 min 331 27 380 min Light-Kicking 140 17 2 passes 9 min 177 22 4 passes 30 min 317 39 6 passes 39 min Intense-Kicking 1 176 65 4 passes 20 min 165 61 3 passes 30 min 341 125 7 passes 50 min Intense-Kicking li 384 47 8 passes 43 min 394 48 9 passes 87 min 778 95 17 passes 130 min Seagrass bed Control 1 0 0 0 0 0 0 0 0 0 Control II 0 0 0 0 0 0 0 0 0 Raking 134 1.9 275 min 925 13 2,125 min 1,059 15 2,400 min Light-Kicking 91 1.4 2 passes 9 min 963 15 18 passes 121 min 1,054 16 20 passes 130 min Intense-Kicking 1 136 2.6 4 passes 22 min 2,608 49 32 passes 179 min 2,744 52 36 passes 201 min Intense-Kicking II 1,033 12 12 passes 73 min 3,168 36 23 passes 156 min 4,201 48 35 passes 230 min passes of the kicking boat and the minutes of clam kicking applied (Table 2). All M. mercenaria col- lected were returned to the laboratory for size- frequency estimates. The cumulative removals from the 2 clam harvesting applications produced relative treatment intensities acceptably close to our initial intentions (Table 2). For the hand raking treatment, we used short-handled rakes with 6-10 prongs of =14 cm in length separated by 3.5 cm gaps (see descrip- tion and photograph of "pea digger" in Peterson et al. 1983a). We attempted to equalize the inten- sities of the raking and light-kicking treatments by removing equal percentages of the legally harvest- able M. mercenaria from each of these two treat- ment matrices (Table 2). We also recorded the length of time actually spent raking as another indication of treatment intensity (Table 2). RESULTS Initial Sampling and Estimation of Shelliness Within each habitat (sand flat and seagrass bed), one-way ANOVA was used on log {x + l)-trans- formed data (which eliminated heteroscedacity in Cochran's tests) to assess whether any response variables differed significantly among the 6 matrices in spring 1980 prior to application of harvest treat- ments. There was no significant (a = 0.05) initial variation among sand-flat matrices in any param- eter: average total density of hard clams, average density of hard clam recruits (length <2.5 cm), aver- age dry mass of seagrass, average density of all ben- thic macroinvertebrates, and sediment size (0) (Table 3). Furthermore, the average percent organic con- tent of sediments did not vary significantly among sand-flat matrices (P > 0.05 in ANOVA on angular- transformed proportions). Bay scallops were so rare in this initial sampling that we do not even record their densities in Table 3: bay scallops showed no significant difference among matrices in either habitat. The seagrass matrices exhibited significant initial variation in all parameters except average total density of hard clams and bay scallop density (Table 3). Variation in the other 4 parameters was not consistent across all seagrass matrices. A poste- riori Duncan's tests, used to identify how specific seagrass matrices differed, show that the control II and raking matrices had significantly higher den- sities of hard clam recruits than all other seagrass matrices in spring 1980. Average seagrass biomass was significantly greater in intense-kicking I and significantly lower in control I than in all other sea- grass matrices in the initial sampling. Control I also initially possessed a significantly higher average density of benthic macroinvertebrates, about 3 times the levels in the other seagrass matrices (Table 3). Duncan's test on mean 0s revealed that in seagrass the raking and light-kicking matrices possessed 286 PETERSON ET AL.: IMPACT OF MECHANICAL CLAM HARVESTING significantly higher initial 0 values (finer sediments), although the differences among matrices were small. Percent organic content did not differ sig- nificantly {P > 0.05) among seagrass matrices in a one-way ANOVA on angular-transformed propor- tions. The results of this initial sampling in spring 1980 prior to any application of clam harvest treatments imply that the sand-flat matrices were initially quite homogeneous. Consequently, any treatment effects can be expected to appear as significant differences that emerge among matrices in some or all sam- plings after application of the treatments. However, the initial differences among seagrass matrices im- ply that treatment effects may not be so readily identified. For those variables that exhibited initial differences among matrices, we performed two dif- ferent tests of the effects of treatment. We per- formed simple ANOVA' s to test for differences following treatment and we also, by subtraction of matrix means for spring 1980, adjusted the data from each matrix for initial differences and tested by ANOVA for significant changes in the differences among matrices. The first approach is appropriate if one believes that initial differences among matrices do not reflect intrinsic between-matrix dif- ferences that require adjustment, whereas the sec- ond approach assumes that initial differences among matrices would be expected to persist or recur in the absence of any treatment. An examination of how replicate matrices vary over time helps resolve which test procedure is more appropriate, but we performed both tests to provide a more robust set of conclusions. Although all matrices in each habitat were chosen to be homogeneous in surface appearance, our Octo- ber 1985 estimates of coarse shelliness of the sur- face (0-12 cm) sediments demonstrated that sea- grass control I had almost 10 times the amount of coarse shell than any of the other seagrass matrices. The average ( ± SE) mass of shell fragments >3 mm in the top 12 cm of the 0.25 m^ area in seagrass control I was 5,257 g ( ± 701) compared vdth a range of 375 (±70) to 777 (±135) g across the other 5 seagrass matrices. This substantially larger amount of shell (P < 0.001 in a one-way ANOVA) seemed to be present during the entire experiment. Because surface shell fragments could greatly influence sea- grass growth and especially M. mercenaria recruit- ment and survival (see Castagna and Kraeuter 1977), this physical anomaly of seagrass control I renders it a questionable control for the various treatment matrices. Similar data on surface shelli- ness taken from the sand matrices in October 1985 revealed no significant differences (P > 0.05) among matrices in a one-way ANOVA, with mean (±SE) Table 3. — Contrasts among replicate matrices within each habitat before application of harvest treatments. Data are sample means ( ± SE) from spring 1980 (22 Feb. -6 May). Sample sizes appear in Table 1 . Superscripts A and B indicate significant differences among matrices in Duncan's test at a = 0.05, with those means sharing capital letter superscripts not differing significantly. Where ANOVA was non- significant, no means differ significantly. Habitat and sample average for each 1 paramentet Sand flat Seagrass bee i Future matrix designation Total hard clam density per '/< m^ Density of hard clam recruits' per Va m^ Seagrass dry mass (g per '/4 m^) Density of benthic inverte- brates per 0.008 m^ Graphic mean sediment size m Total hard clam density per '/< m^ Density of hard clam recruits' per Va w? Seagrass dry mass (g per Va m^) Density of benthic inverte- brates per 0.008 m^ Graphic mean sediment size (U) Control 1 0.50 (0.14) 0.22 (0.10) 0.00 6.00 (1.34) 2.14 (0.00) 2.42 (0.44) 0.17^ (0.06) 10.36^ (2.38) 16.50^^ (4.32) 2.75^ (0.21) Control II 0.33 (0.10) 0.17 (0.07) 0.00 4.67 (0.76) 2.16 (0.03) 2.28 (1.72) 0.81^ (0.16) 14.37^ (2.23) 4.33^ (0.88) 2.94^ (0.12) Raking 0.47 (0.13) 0.17 (0.06) 0.00 4,67 (0.84) 2.17 (0.04) 2.19 (0.38) 0.53^s (0.14) 16.01^ (3.49) 4.83^ (1.11) 3.38* (0.07) Light-Kicking 0.25 (0.09) 0.06 (0.04) 0.00 6.00 (1.26) 2.16 (0.02) 2.28 (0.33) 0.39^ (0.11) 19.56^ (2.62) 5.67^ (2.08) 3.46* (0.07) Intense-Kicking 1 0.36 (0.14) 0.17 (0.07) 0.00 3.17 (0.83) 2.10 (0.02) 1.83 (0.28) 0.39^ (0.10) 41.22* (4.03) 6.50^ (1.52) 2.84^ (0.15) Intense-Kicking II 0.47 (0.14) 0.17 (0.07) 0.00 5.67 (1.38) 2.18 (0.01) 2.56 (0.36) 0.27^ (0.09) 28.44^ (4.17) 5.67^ (2.33) 2.69^ (0.05) Statistical significance^ NS NS NS NS NS NS * * • * • * * * • 'Recruits defined as <2.5 cm in length (see Peterson et a! 1983b for size data on 0 year class as support). 2* - P < 0.05, " - P < 0.01, *** - P < 0.001. NS - P > 0.05 in one-way ANOVA comparing matrices before experimental initiation. 287 FISHERY BULLETIN: VOL. 85, NO. 2 mass of shell fragments >3 mm ranging from 28 (±7) to 157 (+121) across the 6 sand-flat matrices. Our field plots were closed to all commercial and recreational shellfishing during the 4 years of the experiment by proclamation of the North Carolina Division of Marine Fisheries to avoid disruption of the experiments. However, on 7 occasions out of 50 days of observation, we observed clammers within the boundaries of our plots: 5 times in seagrass con- trol matrix I and once in both the seagrass raking matrix and the intense-kicking II matrix. This represents significantly more illegal clamming in control I than would be expected by chance alone (P < 0.01 in a binomial test). Thus, the seagrass con- trol I matrix may not represent a true control for our experiment. Posttreatment Sampling Mercenaria mercenaria Recruitment In the sand-flat habitat there were only two Octo- bers during which M. mercenaria recruits were sampled: October 1980 after the initial application of the clam harvest treatments and October 1981 after both treatment applications. In neither sam- pling did a one-way ANOVA on log {x + l)-trans- formed counts (which removed heteroscedacity in Cochran's tests) reveal significant (a = 0.05) vari- ation in average density of recruits among sand-flat matrices (Table 4). Furthermore, a two-way ANOVA on log (x + l)-transformed counts from both time periods, done to increase the power of the test of matrix differences, also failed to reveal any signifi- cant variation in average recruitment among sand- flat matrices. Despite the failure to demonstrate statistical significance in M. mercenaria recruitment among sand-flat matrices, the average density of recruits in the control matrices during these two Octobers was more than double (on untransformed scale) the average density in the 2 high-intensity clam kicking matrices (Fig. 3). Some of this differ- ence may have been present even before treatments were applied (Fig. 3), but it is also possible that the high local variability in recruitment lowers the power of this test of harvest treatment to a degree that even a twofold difference is undetectable. During 4 Octobers, M. mercenaria recruitment was estimated in the seagrass habitat (Table 4). One of these, October 1980, fell after the first harvest treatment (which Table 2 shows to have been very light in the seagrass plots) but before the second, more intense treatment. The other 3 samplings came in successive years, increasingly far from the actual time of application of the harvest treatments. Because of the preexisting significant differences Table 4. — The impact of clam harvesting on recruitment of Mercenaria mercenaria. Entries are mean numbers (± SE) of recruits per % m^. Recruits are defined as all individuals <2.5 cm in length in October of each year. For 1980 and 1981 , n = 36 samples from each treatment matrix in each habitat, whereas for 1982 and 1983, n = 9 for seagrass and 0 for sand flat. Habitat and date Sand flat Seagrass bed Treatment matrix 1980 1981 Unweighted average 1980 1981 1982 1983 Unweighted average Control 1 0.33 (0.11) 0.17 (0.06) 0.25 0.94 (0.21) 0.61* (0.13) 0.67*'^ (0.24) 0.67 (0.24) 0.72 Control II 0.36 (0.11) 0.06 (0.04) 0.21 0.72 (0.15) 0.28^ (0.09) 1.33*'^ (0.47) 1.56 (0.77) 0.97 Raking 0.44 (0.13) 0.14 (0.08) 0.29 0.81 (0.14) 0.22^ (0.07) 0.78*^ (0.32) 1.67 (0.55) 0.87 Light-Kicking 0.19 (0.08) 0.08 (0.05) 0.14 0.61 (0.13) 0.11^ (0.05) 2.11^^ (0.68) 0.33 (0.17) 0.79 Intense-Kicking 1 0.11 (0.05) 0.03 (0.03) 0.07 0.42 (0.11) 0.39*'^ (0.10) 0.22^ (0.15) 0.67 (0.17) 0.43 Intense-Kicking II 0.22 (0.07) 0.08 (0.05) 0.15 0.56 (0.14) 0.33'^-^ (0.10) 0.56^ (0.24) 0.33 (0.24) 0.45 Statistical significance^ NS NS NS NS * * * NS * ' * - P < 0.05, " - P < 0.01 in one-way ANOVA's on each date and two-way ANOVA's over all dates, reported in the unweighted average column. These analyses were performed on log-transformed data, which eliminated or reduced heteroscedacity in Cochran's tests Superscripts A and B indicate significant differences in Duncan's test at o = 0.05. No Duncan's test results are given for the unweighted averages in the seagrass bed because the two-way ANOVA ex- hibited highly significant (P < 0.001) interaction between date and treatment. 288 PETERSON ET AL.: IMPACT OF MECHANICAL CLAM HARVESTING SEAGRASS 3 O o "o o to o E o V— > 0.05) less in the 2 intense-kicking matrices than expected from the 2 controls in all sampling periods after application of both harvest treatments. This test provides the statistical justification for our presentation of differences in Figure 4. Table 5. — The impact of clam harvesting on the average seagrass dry mass (± SE) per % m^ within the seagrass habitat. Data presented for each date and matrix are the mean ( + SE) dry mass of seagrass per sample minus the mean dry mass in spring 1980 for that particular matrix (from Table 3). Sample sizes appear in Table 1 . Clam harvesting treatments occurred between spring 1980 and fall 1980 and again between fall 1980 and spring 1981. Superscripts A-D indicate significant differences among matrices in Duncan's test at o = 0.05, with those means sharing capital letter superscripts not differing significantly. Treatment matrices Fall 1980 Spring 1981 Fall 1981 Fall 1982 Fall 1983 Fall 1984 Control 1 2.2(2.9)^° 19.5(11.2)* 7.4(3.3)^ 11.9(6.1)* 1.2(4.4)^ 8.0(7.2)^ Control II 19.7(2.7)* 25.8(4.0)* 20.0(2.0)* 22.8(5.2)* 40.1(5.9)* 40.8(3.8)* Raking 7.7(1.8)^'^ 10.5(3.2)* 15.2(2.0)*'^ 13.6(5.1)* 38.2(7.0)* 41.1(5.4)* Light-Kicking 13.8(2.9)*'^ 14.4(6.0)* 15.3(3.0)*'^ 13.6(3.5)* 31.9(4.7)* 35.7(5.1)* Intense-Kicking 1 -1.5(2.7)° -21.2(6.9)^ -18.8(4.3)'^ -29.0(6.1)'^ -18.8(6.0)^ 5.6(9.1)^ Intense-Kicking II 7.1(3.7)^'^ -9.1(5.4)^ -12.2(3.1)'^ - 1 1 .3(7.2)^ 9.2(10.8)^ 1 .5(4.4)^ Statistical significance^ * * * * * ♦ • • * * • « * ** • * • ' * * ' - P < 0.001 In one-way ANOVA's on untransformed dry masses, comparing the matrix means on each separate date. ANOVA's were performed on the differences from spring 1980 matrix means because of pre-existing significant differences among matrices in spring 1980 before application of harvest treatments. 290 PETERSON ET AL.: IMPACT OF MECHANICAL CLAM HARVESTING I40n 120 e 100 g 80 O CD o Q. 60 40 20- — Light Kicking — Raking ■- Intense Kicking I — Intense Kicking n / / Sp'l980t FI980i Spl98l FI98I FI982 FI983 FI984 minor mojor treatment Figure 4.— Percent difference between observed average biomass of seagrass in each treat- ment matrix and expected biomass based on the assumption that initial differences between the two control matrices and each treatment matrix would be expected to remain constant across time. The expected biomass is then plotted as 100% (the no effect line). Times of the two clam harvest treatments are indicated with arrows on the x-axis. Benthic Macroinvertebrates In the sand-flat habitat, the average density of benthic macroinvertebrates never varied significant- ly among matrices (Table 6) in any of the 3 post- treatment sampling dates [one-way ANOVA's were run on log (x -i- l)-transformed counts, using a separate analysis for each date]. The sums over all 3 posttreatment dates of the average macroinverte- brate densities per core are nearly identical for each sand-flat matrix and a two-way ANOVA on log (x + l)-transformed densities from all 3 time peri- ods revealed no significant difference among ma- trices. In the seagrass habitat, analogous one-way ANOVA's done separately for each date, demon- strated that the average density of benthic macro- invertebrates did not differ significantly among seagrass matrices in fall 1980 or spring 1981 (Table 6). A significant difference among matrices did ap- pear in fall 1981, and in a two-way ANOVA on all 3 posttreatment dates together. Despite the statis- tical significance of 2 of 4 ANOVA's, actual differ- ences in mean densities among seagrass matrices were proportionately small. Furthermore, Duncan's tests revealed a pattern of differences among ma- trices (Table 6) that was identical to the initial pat- tern of significant differences in the spring 1980 sampling before treatment (see Table 3). Although the sums of the sample means from each of the 3 posttreatment sampling dates (Table 6) im- ply that benthic macroinvertebrate densities in the seagrass habitat were about double those in the sand flat, this pattern was not consistent across seasons. Nested ANOVA's, done on log {x + 1)- transformed counts and performed separately for each sampling date, showed that there was no significant differ- ence between habitats during either spring sampling period (spring 1980 or 1981), whereas average den- sities of benthic macroinvertebrates were signifi- cantly greater (P < 0.001 in fall 1980 and P < 0.005 in fall 1981) in the seagrass habitat in both of the Octobers. Although the clam harvesting treatments did not affect total density of benthic macroinvertebrates in either habitat, species composition might still have been altered. We identified all individuals in 16 cores in each habitat from the spring 1980 pretreatment sampling (4 cores randomly chosen from each con- trol matrix and from each intense-kicking matrix) and in 16 cores in each habitat from the spring 1981 posttreatment sampling (drawn equally from each of the same matrices). This comparison holds season constant and permits us to test for any gross shifts 291 FISHERY BULLETIN: VOL. 85, NO. 2 Table 6.— The impact of clam harvesting on average density ( + SE) of benthic macroinvertebrates per 0.008 m^. n = 6 samples for each treatment matrix at each sampling date. Samples were taken to 25 cm and passed through 1 mm mesh. Superscripts A-C indicate significant differences among matrices in Duncan's test at a = 0.05, with those means sharing capital letter superscripts not differing significantly. Habitat and date Sand flat Seagrass bed Treatment matrix Fall 1980 Spring 1981 Fall 1981 Sum Fall 1980 Spring 1981 Fall 1981 Sum Control 1 8.0 (2.7) 5.7 (1.2) 8.0 (1.4) 21.7 34.3 (7.8) 9.3 (1.8) 16.2* (2.9) 59.8* Control II 11.7 (1.5) 7.7 (1.4) 4.2 (0.8) 23.6 19.0 (2.0) 10.5 (1.6) II.O'^'^ (1.6) 40.5^ Raking 6.5 (0.5) 8.2 (1.0) 4.8 (0.8) 19.5 39.8 (5.1) 6.8 (1.1) 12.0*'^ (2.4) 58.6^ Light-Kicking 12.3 (2.6) 11.5 (3.0) 4.5 (0.9) 28.3 29.5 (8.6) 5.8 (1.3) jqB.C (1.3) 44.1^ Intense-Kicking 1 7.8 (0.7) 8.7 (1.7) 6.3 (1.4) 22.8 23.5 (4.8) 8.7 (2.9) 6.5^ (1.2) 38.7^ Intense-Kicking II 9.7 (2.2) 6.0 (1.3) 5.0 (0.8) 20.7 34.5 (10.7) 6.3 (1.1) 6.0^ (0.9) 46.8^ Statistical significance^ NS NS NS NS NS NS ♦ * * * * ' *• - P < 0.01, •*• - P < 0.001, NS - P > 0.05 in one-way ANOVA's (for each separate date) and two-way ANOVA's (for sums) on average macroinvertebrate counts per core (transformed by log (x + 1)). in species composition as a function of the intense- kicking treatment. Table 7 presents the results of these species identifications and shows that no major shift in species composition of the most abundant species occurred in either the sand-flat or seagrass habitat following the application of the intense-kicking treatment. Polychaetes domi- nated the fauna of both habitats and the same species of polychaetes tended to be represented at similar densities both before and after intense clam kicking. Bay Scallop Densities Bay scallops were never encountered in sampling the sand-flat matrices, so we have no test of whether clam harvest treatment affects bay scallops in areas lacking seagrass. One-way ANOVA's on log {x + 1)- transformed counts (which removed heteroscedacity in Cochran's tests) demonstrated significant (a = 0.05) differences among seagrass matrices in aver- age bay scallop density on only 2 sampling dates, fall 1980 and fall 1983 (Table 8). Duncan's test on the fall 1980 data showed that bay scallop density in control I was significantly (P < 0.05) lower than in every other matrix except intense-kicking II, and that there were no other significant differences between pairs of matrices. Because the fall 1980 sampling occurred before the major application of clam harvest treatments (see Table 2), this sampling period may be considered a pretreatment sampling. Extremely low seagrass biomass in control I in fall 1980 (Table 5) may explain the significantly lower bay scallop densities in that matrix on that date. The fall 1983 sampling occurred after a period of more successful bay scallop recruitment than oc- curred before any other sampling date (Table 8) and, thus, provided more "substrate" on which effects of clam harvest treatments may have operated. Dun- can's test on mean bay scallop densities for fall 1983 demonstrated that the matrices split into two separate groups: a low-density group, made up of control I and the 2 intense-kicking matrices, and a high-density group, comprised of control II, the raking, and light-kicking matrices (Table 8). Within each group, no matrices differed significantly (a = 0.05) from any other, but all differences between groups were statistically significant. Because fall 1983 bay scallop densities were so much greater than at any other sampling date, the sums over all five sampling periods also exhibited significant dif- ferences among matrices in an analogous two-way ANOVA, and Duncan's tests separated the matrices into groupings virtually identical to those detected for the fall 1983 data set alone (Table 8). A contrast of the bay scallop results of fall 1980 and fall 1983 demonstrates that after application of the second intense-kicking treatment in the seagrass habitat in winter of 1980-81, bay scallop densities declined to join the already low value of control I, 292 PETERSON ET AL.: IMPACT OF MECHANICAL CLAM HARVESTING which together formed a group of low-density bay scallop matrices. About 84% of the variance in bay Table 7— For each habitat, total numbers of individuals found in four randomly chosen cores from each of the two controls and the two intense-kicking matrices on two dates, one before and one after clam-harvest treatment. All species with total counts greater than two are listed separately. Spring 1980 Spring 1981 Before treatment After treatment Intense- Intense- Species Controls kicking Controls kicking Sand-flat habitat Aricidia frag His 7 1 6 5 Notomastus hemlpodus 3 4 5 5 Platynereis dumerilii 1 0 5 7 Axiothella sp. 3 2 1 7 Dhlonereis magna 5 5 0 1 Spiochaetopterus oculata 0 2 3 3 Arabella iricolor 3 1 1 2 Glycera sp. 5 1 0 0 Others' 5 2 2 1 Seagrass habitat Axiothella sp. 23 8 7 4 Platynereis dumerilii 12 8 14 1 Notomastus hemipodus 20 7 4 0 Tharynx marioni 3 4 3 2 Nereis falsa 0 1 5 5 Glycera sp. 3 4 2 1 f\/lelinna maculata 1 0 6 3 Onuptiis jenneri 0 1 3 3 Lumbrinereis sp. 0 0 4 3 Spiocfiaetopterus oculata 1 0 4 0 Spionidae 4 0 0 1 Sttienelais limicola 0 1 1 1 Arabella iricolor 0 2 1 0 Poecilochaetus sp. 3 0 0 0 Onuphidae 0 0 3 0 Others' 1 2 1 1 scallop densities in fall 1983 is explained by seagrass biomass in a simple linear regression. Figure 5 pre- sents the relationship between average seagrass biomass and bay scallop densities on a 1,225 m^ scale, which suggests that the 2 intense-kicking matrices contained even fewer bay scallops than predicted from their reduced seagrass biomass. This is similarly illustrated from calculations of the mean numbers of bay scallops per 100 g of seagrass in each matrix in fall 1983: control I (5.7), control II (5.1), raking (4.7), light kicking (4.3), intense-kicking I (2.0), and intense-kicking II (2.3). 'These include molluscs, an amphipod, and additional polychaetes. 10 20 30 40 50 60 Av. seagrass dry moss (g/0.25m^) Figure 5.— Relationship between the average density of bay scallops, Argopecten irradians, and the average biomass of seagrass in fall 1983 samplings of each control and treatment matrix of the clam harvest experiment in the seagrass matrix. Clam harvest treatments had been applied in spring 1980 and again in winter 1980-81. Table 8.— The effect of clam harvesting in the seagrass habitat on average bay scallop, Argopecten irradians, den- sity per V4 m^ (±SE). Sample sizes per treatment matrix were 36 in fall 1980 and fall 1981 and 9 in spring 1981, fall 1982, and fall 1983. No data are presented for the sand flat because of the rarity of bay scallops in that habitat. Superscripts A-C indicate significant differences among matrices in Duncan's test at o = 0.05, with those means sharing capital letter superscripts not differing significantly. TrpQtmpnt Sampling date matrix Fall 1980 Spring 1981 Fall 1981 Fall 1982 Fall 1983 Sum Control 1 0.1 1(± 0.05)^ 0.44(±0.24) 0.05( + 0.04) 0.11(±0.16) 0.66(±0.33)^ 1.37^ Control II 0.63( + 0.11)* 1.00(±0.37) 0.14(±0.07) 0.22(±0.15) 2.89(±0.51)^ 4.88^ Raking 0.53(±0.14)* 0.78(±0.28) 0.16(±0.06) 0.44( + 0.24) 2.56(±0.67)'^ 4.47^^ Light-Kicking 0.75(±0.18)* 0.33(±0.33) 0.14(±0.07) 0.89(±0.42) 2.22(±0.46)* 4.33^ Intense-Kicking 1 0.50(±0.12)'^ 0.22(±0.15) 0.03(±0.03) 0.00(±0.00) 0.44( + 0.29)^ 1.19^'^ Intense-Kicking II 0.39(±0.11)*^ 0.56(±0.18) 0.14( + 0.07) 0.55( + 0.24) 0.88(±0.35)^ 2.52^ Statistical significance' * * NS NS NS * * * * • * 1 '• - P < 0.01, *•* - P < 0.001, NS - P > 0.05 in one-way ANOVA on log (x + 1 )-transformed sample counts, comparing matrix means on each date and in a two-way ANOVA over all dates (in the sums column). 293 FISHERY BULLETIN: VOL. 85, NO. 2 DISCUSSION The one-way ANOVA's which we performed to test the significance of differences in parameter means among matrices at any given samphng date can demonstrate heterogeneity among matrices. If there is no significant heterogeneity, we probably can conclude safely that there was no effect of treat- ment on that parameter at that sampling date, assuming that equivalent levels of the parameter prevailed before application of the treatment (which was not always true). If, on the other hand, the one- way ANOVA demonstrates significant differences among matrices, this result does not necessarily im- ply that the treatment was the cause. Replication in these ANOVA's is generated from subsamples within each individual matrix. These subsamples taken from within a given matrix are not indepen- dent because of their spatial proximity. Consequent- ly, matrices can diverge in various ways from one another over the course of an experiment, caused by extraneous events that act on the scale of the plot (matrix) to destroy independence among subsam- ples. This experimental design would be termed pseudoreplication (Hurlbert 1984), and permits a test of whether plots differ significantly and does not allow an unambiguous assignment of observed differences to the treatment applied (but see Stewart-Oaten et al. 1986). For that reason, we replicated both our control matrices and our intense- kicking matrices in each habitat. These permit us to use a priori contrasts, with replication of 2 sep- arate, independent plots, to test unambiguously whether the most important treatment (intense clam kicking) was responsible for observed changes. Ap- preciation of the differences between these two sorts of analyses is necessary to interpret properly the results of this study. Although we designate our heavier clam-kicking treatment "intense", it probably falls well short of the effort that commercial clammers would apply to a productive seagrass bottom; we took only an estimated 50% of the clams legally available for harvest (Table 2). Consequently, the intensity of harvest that we applied in the seagrass is not un- reasonably high. In the sand-flat system, we took approximately 100% of the estimated numbers of legally available clams in our intense treatments. Although higher than the percent taken in the sea- grass, this probably better approximates the fish- ing intensity that is applied to productive unvege- tated areas by commercial clammers. Efficiency of returns remained high even in the high-intensity kicking matrices, as compared with hand raking. In the sand flat, light kicking produced an average of 8.1 clams per minute and intense kicking 6.2 clams per minute, compared with a return of only 0.9 clams per minute from hand raking (Table 2). In the seagrass bed, light kicking yielded an average of 8.1 clams per minute and intense kicking 16.1 clams per minute, in contrast to a return of only 0.4 clams per minute from hand raking (Table 2). Thus, efficiency of harvest, defined as clams caught per unit of time, was clearly greater by over an order of magnitude with the mechanical technique than with the tradi- tional hand method. The improved efficiency dur- ing clam kicking in the seagrass as harvest inten- sity increased from taking about 15% to about 50% of available clams is probably caused by the gradual removal of seagrasses which, when present, reduce the efficiency of clamming. To test whether hard clam harvest affects its own recruitment in the area of harvest, we counted new recruits (<2.5 cm in length, Peterson et al. 1983b). Recruitment, when estimated in this fashion, con- founds both larval (and postlarval) settlement with subsequent early mortality from time of settlement until October. Consequently, we do not directly test the hypothesis that natural densities of adult hard clams inhibit larval settlement in their vicinity. Fur- thermore, our clam harvest treatment not only removes many larger hard clams, but it also disturbs the bottom sediments. Consequently, there are several plausible mechanisms by which our clam harvest treatments may affect October recruitment of hard clams: 1) reduction of adult hard clam den- sity may affect hard clam settlement (positively, if negative adult-larval interactions predominate, as suggested by most past studies: Woodin 1976; Williams 1981; Peterson 1982b) or survivorship from settlement until October (no a priori prediction from the literature on what direction this effect may take), or 2) disturbance of the bottom may alter hard clam settlement (positively, if hard clam larvae select disturbed sediments, which seems unlikely, or negatively if hard clam larvae avoid disturbed sediments) or early survivorship (negatively, if the clam harvest buries small clams too deeply to reemerge or if disturbance has removed protective seagrass or shell materials and thereby made juvenile hard clams more vulnerable to predators (Peterson 1982a; Summerson and Peterson 1984)). Our data on hard clam recruitment are sufficiently ambiguous to preclude any definitive answers to the question of how clam harvest affects subsequent recruitment. In the sand flat, there was no signifi- cant effect of harvest treatment, but the 2 intense- ly kicked matrices yielded only 50% of the recruits 294 PETERSON ET AL.: IMPACT OF MECHANICAL CLAM HARVESTING produced by the 2 controls (Fig. 3). In the seagrass, M. mercenaria recruitment may also have been reduced by harvest treatments (Table 4), but the conclusion depends upon the assumption that the shelly control I was an adequate control for recruit- ment data. Given the enhanced survivorship of M. mercenaria recruits in shell (Castagna and Kraeuter 1977) and the significant illegal clamming in sea- grass control I, this assumption is questionable. It is possible that removal of adult hard clams enhances larval settlement over a larger spatial scale than the 1,225 m- experimental plots because depletion of larvae by feeding from the water col- umn should extend over a larger spatial scale (Peter- son 1982b). Although it is possible that our sampling was on too fine a scale to detect such an effect, our sampling occurred on a far larger spatial scale by 3 orders of magnitude than any previous experi- mental test of adult-larval interactions and, thus, should have provided for greater opportunity to detect any positive effect of adult hard clam removal. The failure to demonstrate a response in the sand flat may be a different consequence of scale. Newly recruited hard clams may settle more heavily where adult densities have been reduced but the effect may be diffused away by the physical dispersal of new recruits by tidal currents and waves. As a consequence of such multiple interpre- tations, we can best conclude that on the scale of our experiments no dramatic increase in hard clam recruitment occurs vdth intense mechanical harvest of adult hard clams in seagrass and harvest may even reduce recruitment in both unvegetated and vegetated areas. The effect of various clam harvest treatments in the seagrass bed on seagrass biomass (Fig. 4) is the most obvious result of this study. Clam harvest of all types had an immediate impact in reducing the seagrass biomass. Reduction of seagrass increased with harvest intensity, as was demonstrated both by the enhanced effect of the second treatment ap- plication, which was much more intense than the first, and also by the larger effects of intense kick- ing as compared with the other treatments (Fig. 4). Although the seagrass biomass in the raking and light-kicking matrices recovered to levels predicted from the controls within a year's time, the seagrass biomass in the intense-kicking matrices did not even begin to recover for 2 years and had not fully re- turned to predicted, control levels after 4 years. These results imply that if sufficient seagrass is destroyed, recovery is slow. Because our intense- kicking treatment removed only an estimated 50% of available hard clams and because the efficiency of hard clam capture per unit time of harvest was greater in the intense treatment than in the light treatment in the seagrass habitat, we suspect that commercial clam kickers would apply even more harvest intensity than we did in the this intense- kicking treatment. Consequently, the effects of com- mercial clam kicking in seagrass beds are probably underestimated by our data (Fig. 4). Furthermore, by using both control matrices (including the shelly one) in estimating the effects of harvest on seagrass biomass, we intentionally provide an additional con- servative bias. Clam kicking at a low level (=15% of available hard clams harvested) does not appear to be any more destructive of seagrass than hand raking that same number of clams, but the lack of replication of these two types of treatment matrices renders this a tentative conclusion. The extremely slow recovery of seagrass in the intensely kicked seagrass matrices raises the possi- bility that seagrass beds and unvegetated sand flats may exist as alternative stable states (Sutherland 1974; Connell and Sousa 1983; Peterson 1984) on many of the same shallow bottoms of sounds and coastal lagoons. That is, a given shallow bottom may exist as either a seagrass bed or an unvegetated sand flat, but whichever state it occupies it is likely to retain for a relatively long period of time. Trans- formation from one state to another may require some input of external energy. Because great changes in current regime and surface sediment character are associated with the presence and growth of seagrasses (Ginsburg and Lowenstam 1958; Orth 1977; Fonseca et al. 1983; Peterson et al. 1984; Eckman in press), it is reasonable to hypothesize that destruction of seagrass may result in sufficiently higher energy at that site that natural reestablishment could be difficult. Certainly, the slow return of seagrass following intense clam kick- ing in our experiments implies that seagrass re- covery even in previously vegetated areas is ten- uous. If seagrass beds and unvegetated bottoms do tend to represent alternative stable states for large areas of the estuarine and sound bottom, then denuding of vegetation would have long-lasting ef- fects, even beyond what we have demonstrated. Furthermore, transplantation of relatively dense seagrass may be necessary to produce rapid rever- sion back into a vegetated system (for reviews of disturbance, recovery, and transplantation of sea- grasses see Zieman 1982; Thayer et al. 1985). Because of the important roles that seagrasses play in promoting estuarine productivity and coastal fisheries (Thayer et al. 1975), intense clam kicking in vegetated areas could have long-lasting and 295 FISHERY BULLETIN: VOL. 85, NO. 2 serious impacts on many commercially important fisheries. Our own data imply a potentially negative impact on hard clam recruitment (Table 4) and a clear reduction in bay scallop abundance (Table 8) in part because of reduction in seagrass biomass. Clam harvesting had no detectable effect on the abundance of small benthic invertebrates. The den- sity data did not even suggest an effect (Table 6) and the composition of the most abundant species did not change, even with intense clam kicking (Table 7). This lack of response is probably a conse- quence of the dominance of small polychaetes in these invertebrate data. Small polychaetes make up most of the total infaunal density and all of the most abundant species. Small polychaetes tend to exhibit rapid turnover, quick colonization and short life spans, relative to molluscs, echinoderms, and many other invertebrates; consequently, they may be ex- pected to recover more rapidly after disturbance. The large seasonal variability in total macroinver- tebrate density at our seagrass sites is a reflection of the short-term response times of this fauna, which is known to exhibit large seasonal fluctuations in density in North Carolina (Commito 1974). Like several previous studies of the densities of benthic infauna (Kikuchi 1966; Warme 1971; Orth 1977; Reise 1977, 1978; Stoner 1980; Summerson and Peterson 1984), our data demonstrate higher densities inside the seagrass bed than on unvege- tated bottoms in October. However, the difference in infaunal density between habitats appears to vary seasonally, as shown previously (Reise 1978; Stoner 1980). In spring, the two habitats had approximately equal densities of infauna. Because estuarine den- sities of epibenthic predators, both fishes (Adams 1976; Orth and Heck 1980) and crustaceans (Heck and Orth 1980), also vary seasonally such that our fall samplings occur after months of high density and our spring samplings after a low-density season for epibenthic consumers, these new observations provide further support for the hypothesis (see review of concepts in Kikuchi 1980; experimental evidence in Reise 1977; Orth 1977; Summerson and Peterson 1984) that seagrass provides a natural refuge from predation for infaunal invertebrates. Intense clam kicking caused a substantial decline in the average density of bay scallops in the seagrass habitat (Table 8). Most of the variation among matrices in the total densities of bay scallops and in the fall 1983 densities, when numbers were high, could be readily explained by the variation among matrices in average seagrass biomass. Bay scallops recruit to seagrass blades where they remain at- tached by byssal threads for the first few months of life. In addition, adult bay scallops, which are mobile, tend to be found in seagrass beds, as our failure to encounter them in the sand-flat samples illustrates. Their feeding may be more efficient in the slower currents of the seagrass environment (Kirby-Smith 1972). Consequently, it is not surpris- ing that reductions in bay scallop density accom- panied the declines in average seagrass biomass in our experiments. However, the apparent effect (Fig. 5) of intense clam kicking that persists even after the seagrass biomass effect is removed was a sur- prise. Because the application of clam kicking is necessarily patchy (it forms a trail behind the path of the boat) and, thus, produces an increase in the patchiness of the vegetation (see standard errors in Table 5), we suspect that this residual effect of in- tense clam kicking is a reflection of that enhanced seagrass patchiness. We hypothesize that the aver- age biomass of seagrass present in our plots is more attractive (in a broad sense) to bay scallops when it is more uniformly distributed over a given area than when it is clumped into more discrete patches at least on the 0.25 m- scale of our samples. The implications of this study for the management of the hard clam fishery depend upon the specific values attributed to various factors. Our data show clearly the enhanced efficiency that the mechanical clam harvesting process known as clam kicking brings to the fisherman who adopts it instead of hand raking. Yet the enhanced efficiency may itself be a danger if the resource is thereby overfished beyond its capacity to sustain harvest. Our data on the negative impacts of clam harvest do not permit one method to be selected in preference to another except to the degree that hand raking might never reach the same harvest intensity and, therefore, might not cause the same magnitude of effects on seagrass beds and their fauna. Outside seagrass beds, clam kicking does not appear to have any serious negative impacts on other parameters of ecological value with the possible exception of hard clam recruitment. This effect is probably a necessary price to pay for the harvest of the adult, marketable clams. Inside seagrass beds, effects of clam kicking on seagrass biomass and bay scallop abundance are quite serious and long-lasting. Because seagrass con- tributes so substantially to the production of many coastal fisheries (Thayer et al. 1985), any regulation that might limit the intensity of clam fishing in that habitat would probably be beneficial. Restriction of the much more efficient mechanical clam harvesters to unvegetated bottoms may be a suitable mech- anism for limiting the total harvest pressure in seagrass beds and, thereby, preserving other fish- 296 PETERSON ET AL.: IMPACT OF MECHANICAL CLAM HARVESTING eries in the face of emerging new technology, which has the potential to enhance greatly the user con- flicts for limited and interdependent coastal and estuarine resources. ACKNOWLEDGMENTS We thank D. Guthrie for his advice and help in applying the clam kicking treatments using his boat and gear. Field and laboratory assistance in this pro- ject was provided by W. G. Ambrose, Jr., B. F. Beal, K. L. Bowers, P. Boyd, M. E. Colby, P. B. Duncan, S. H. Larson, J. Purifoy, G. W. Safrit, Jr., P. S. Smith, and N. T. Sterman. Loans of equipment by H. J. Porter and D. Spitsbergen are gratefully acknowledged. W. Garner, Jr., M. Marshall, F. Munden, and D. Spitsbergen provided useful advice concerning the data needed for hard clam manage- ment. H. E. Page and V. Page provided graphics services. We thank D. Colby and S. Muga for im- proving our presentation and sharpening our infer- ences. R. B. Deriso and D. Colby provided statis- tical advice. The project was supported by the Institute of Marine Sciences of the University of North Carolina at Chapel Hill and by the Office of Sea Grant, NOAA, U.S. Department of Commerce under grant #NA81AA-D-00026 and the North Carolina Department of Administration. LITERATURE CITED Adams, S. M. 1976. The ecology of eelgrass, Zostera marina (L.), fish com- munities. I. Structural analysis. J. Exp. Mar. Biol. Ecol. 22:269-291. Caddy, J. F. 1973. Underwater observations on tracks of dredges and trawls and some effects of dredging on a scallop ground. J. Fish. Res. Board Can. 30:173-180. CaSTAGNA, M., AND J. N. Kraeuter. 1977. Mercenaria culture using stone aggregate for predator protection. Proc. Nat. Shellfish. Assoc. 67:1-6. COMMITO, J. A. 1976. Predation, competition, life-history strategies, and the regulation of estuarine soft-bottom community structure. Ph.D. Thesis, Duke Univ., Durham. CONNELL, J. H.. AND W. P. SOUSA. 1983. On the evidence needed to judge ecological stability or persistence. Am. Nat. 121:789-824. Dickie, L. M. 1974. Interaction between fishery management and environ- mental protection. J. Fish. Res. Board Can. 30:2496-2506. ECKMAN, J. E. In press. The role of hydrodynamics in recruitment, growth and survival of Argopecten irradians (L.) and Anomia simplex (D'Orbigny) within eelgrass meadows. J. Exp. Mar. Biol. Ecol. Folk, R. L. 1984. Petrology of sedimentary rocks. Hemphill Publ. Co., Austin. FoNSECA, M. S., J. S. Fisher, J. C. Zieman, and G. W. Thayer. 1983. Influence of the seagrass Zostera marina L. on cur- rent flow. Estuarine Coastal Shelf Sci. 15:351-364. GiNSBURG, R. N., and H. a. Lowenstam. 1958. The influence of marine bottom communities on the depositional environment of sediments. J. Geol. 66:310- 318. Gray, J. S. 1974. Animal-sediment relations. Oceanogr. Mar. Biol. Annu. Rev. 12:223-262. Gross, M. G. 1971. Carbon determination. In R. E. Carver (editor), Pro- cedures in sedimentary petrology, p. 573-596. Wiley-Inter- science, N.Y. Guthrie, J. F., and C. W. Lewis. 1982. The clam-kicking fishery of North Carolina. Mar. Fish. Rev. 44(1):16-21. Heck, K. L., Jr., and R. J. Orth. 1980. Structural components of eelgrass (Zostera marina) meadows in the lower Chesapeake Bay - Decapod Crustacea. Estuaries 3:289-295. HULBERG, L. W., and J. S. OLIVER. 1980. Caging manipulations in marine soft-bottom com- munities: importance of animal interactions or sedimentary habitat modifications. Can. J. Fish. Aquat. Sci. 37:1130- 11.39. HURLBERT, S. H. 1984. Pseudoreplication and the design of ecological field ex- periments. Ecol. Monogr. 54:187-211. Ingram, R. L. 1971. Sieve analysis. In R. E. Carver (editor). Procedures in sedimentary petrology, p. 49-94. Wiley-Interscience, N.Y. KiKUCHI, T. 1966. An ecological study on animal communities of the Zostera marina belt in Tomioka Bay, Amakusa, Kyushu. Publ. Amakusa Mar. Biol. Lab. 1:1-106. 1980. Faunal relationships in temperate seagrass beds. In R. C. Phillips and C. P. McRoy (editors), Handbook of seagrass biology: an ecosystem perspective, p. 153-172. Garland STPM, N.Y. Kirby-Smith, W. W. 1972. Growth of the bay scallop: the influence of experimen- tal water currents. J. Exp. Mar. Biol. Ecol. 8:7-18. McCall, p. L. 1977. Community patterns and adaptive strategies of the in- faunal benthos of Long Island Sound. J. Mar. Res. 35:221- 266. Meadows, P. S., and J. I. Campbell. 1972. Habitat selection by aquatic invertebrates. Adv. Mar. Biol. 10:271-382. Myers, A. C. 1977. Sediment processing in a marine sub tidal sand bottom community: II. Biological consequences. J. Mar. Res. 35: 633-647. Nelson, W. G. 1979. An analysis of structural pattern in an eelgrass (Zostera marina) amphipod community. J. Exp. Mar. Biol. Ecol. 39:231-264. Orth, R. J. 1977. The importance of sediment stability in seagrass com- munities. In B. C. Coull (editor), Ecology of marine ben- thos, p. 281-300. Univ. South Carolina Press, Columbia. Orth, R. J., and K. L. Heck, Jr. 1980. Structural components of eelgrass (Zostera marina) 297 FISHERY BULLETIN: VOL. 85, NO. 2 meadows in the lower Chesapeake Bay - Fishes. Estuaries 3:278-288. Peterson, C. H. 1982a. Clam predation by whelks {Busycon spp.): experimen- tal tests of the importance of prey size, prey density, and seagrass cover. Mar. Biol. 66:159-170. 1982b. The importance of predation and intra- and inter- specific competition in the population biology of two infaunal suspension-feeding bivalves, Protothaca staminea and Ckione undatella. Ecol. Monogr. 52:437-475. 1984. Does a rigorous criterion for environmental identity preclude the existence of multiple stable points? Am. Nat. 124:127-133. 1985. Enhancement oi Mercenaria mercenaria densities in seagrass beds: Is pattern fixed during settlement season or altered by subsequent differential survival? Limnol. Oceanogr. 31:200-205. Peterson, C. H., H. C. Summerson, and S. R. Fegley. 1983a. Relative efficiency of two clam rakes and their con- trasting impacts on seagrass biomass. Fish. Bull., U.S. 81:429-434. Peterson, C. H., P. B. Duncan, H. C. Summerson, and G. W. Safrit, Jr. 1983b. A mark-recapture test of annual periodicity of inter- nal growth band deposition in shells of hard clams, Mer- cenaria mercenaria, from a population along the south- eastern United States. Fish. Bull., U.S. 81:765-779. Peterson, C. H., H. C. Summerson, and P. B. Duncan. 1984. The influence of seagrass cover on population struc- ture and individual growth rate of a suspension-feeding bivalve, Mercenaria mercenaria. J. Mar. Res. 42:123-138. Reise, K. 1977. Predation pressure and community structure of an intertidal soft bottom fauna. In B. F. Keegan, P. 0. Ceidigh, and P. J. S. Boaden (editors). Biology of benthic organisms, p. 513-519. Pergamon Press, N.Y. 1978. Experiments on epibenthic predation in the Wadden Sea. Helgol. Wiss. Meeresunters. 31:55-101. Rhoads, D. C. 1974. Organism-sediment relations on the muddy sea floor. Oceanogr. Mar. Biol. Annu. Rev. 12:263-300. Ricker, W. E. 1975. Computation and interpretation of biological statistics offish populations. Bull. Fish. Res. Board Can. 191:1-382. Stewart-Oaten, A., W. W. Murdoch, and K. R. Parker. 1986. Environmental impact assessment: "pseudoreplica- tion" in time? Ecology 67:929-940. Stoner, a. W. 1980. The role of seagrass biomass in the organization of ben- thic macrofaunal assemblages. Bull. Mar. Sci. 30:537-551. Summerson, H. C, and C. H. Peterson. 1984. Role of predation in organizing benthic communities of a temperate-zone seagrass bed. Mar. Ecol. Prog. Ser. 15:63-77. Sutherland, J. P. 1974. Multiple stable points in natural communities. Am. Nat. 108:859-873. Sutherland, J. P., and R. H. Karlson. 1977. Development and stability of the fouling community at Beaufort, North Carolina. Ecol. Monogr. 47:425-446. Thayer, G. W., D. A. Wolfe, and R. B. Williams. 1975. The impact of man on seagrass systems. Am. Sci. 63:288-296. Thayer, G. W., W. J. Kenworthy, and M. S. Fonseca. 1985. The ecology of eelgrass meadows of the Atlantic coast: a community profile. U.S. Fish Wildl. Serv., Wash., D.C. FWS/OBS-84/02, 147 p. Thistle, D. 1981. Natural physical disturbances and communities of marine soft bottoms. Mar. Ecol. Prog. Ser. 6:223-228. Warme, J. E. 1971. Paleoecological aspects of a modern coastal lagoon. Univ. Calif. Publ. Geol. Sci. 87:1-131. Williams, J. G. 1980. The influence of adults on the spat of the clam. Tapes japonica. J. Mar. Res. 38:729-741. Wilson, W. H., Jr. 1981 . Sediment-mediated interactions in a densely populated infaunal assemblage: The effects of the polychaete Abaren- icola pacifica. J. Mar. Res. 39:735-748. WOODIN, S. A. 1976. Adult-larval interactions in dense infaunal assemblages: Patterns of abundance. J. Mar. Res. 34:25-41. ZlEMAN, J. C. 1982. The ecology of the seagrasses of south Florida: a com- munity profile. U.S. Fish Wildl. Serv., Off. Biol. Serv., Wash., D.C. FWS/OBS-82/85, 158 p. 298 BIOLOGICAL DATA ON BERRY ISLANDS (BAHAMAS) QUEEN CONCHS, STROMBUS GIGAS, WITH MARICULTURE AND FISHERIES MANAGEMENT IMPLICATIONS Edwin S. Iversen,^ Edward S. Rutherford,^ Scott P. Bannerot.^ AND DARRYL E. JORYI ABSTRACT Biological data designed to assess the mariculture potential of queen conchs, Strombus gigas, and to aid in management of stocks in the Berry Islands, Bahamas, were collected from March 1980 to February 1983. Juveniles congregated in shallow areas adjacent to cays with strong currents. Growth of queen conchs differed among cays and seemed related to conch density. Average growth rates from several cays in the Berry Islands showed that growth was slower than that reported for queen conchs in other areas in the Caribbean. Estimated survival of juvenile queen conchs (about 10 cm) was 57-80% per month, or 2-9% annually. Yield per recruit from this population can be maximized by harvesting the animals at about 15 cm, which is the size at onset of lip formation but may be below the size at maturity. Presently, potential for increasing queen conch production through intensive and/or extensive mariculture seems low because of high hatchery costs, lack of dependable mass-rearing techniques, high predation on young released in nature, and slow growth of penned conchs. The queen conch, Strombus gigas, sl giant marine snail which is a major food resource in the Carib- bean, Bahamas, and some Central American na- tions, has been exploited by subsistence and com- mercial fishermen for centuries. During the last several decades, recreational conch fisheries have developed and expanded considerably, placing high fishing pressure on these stocks. Until recently there has been little scientific research directed at improv- ing production from existing stocks. The present study was designed to obtain biological data to ful- fill this need in the Berry Islands, Bahamas. Based on its high fecundity, feeding habits, limited migration habits, and high market demand, queen conch appears to be a desirable candidate for both intensive mariculture (enclosed) and extensive mari- culture (released into nature to augment natural stocks) (Berg 1976; Brownell 1977; Brownell et al. 1976; Brownell and Stevely 1981). Success of either type of mariculture is dependent upon technical ability to mass-rear queen conch inexpensively from eggs on a dependable basis, and on knowledge of optimal natural habitats for raising juveniles to a 'University of Miami, Rosenstiel School of Marine and Atmo- spheric Sciences, 4600 Rickenbacker Causeway, Miami, FL 33149-1098. ^University of Miami, Rosensteil School of Marine and Atmo- spheric Sciences, 4600 Rickenbacker Causeway, Miami, FL; present address: Everglades National Park, Research Center, Homestead, FL 33034. sufficiently large size for either release in nature or for grow-out for market. Our research on hatchery methods and potential of queen conch mariculture is described in Siddall (1983) and Iversen (1983), and the role of predators in limiting the size of conch populations is described in Jory (1982), Jory and Iversen (1983), and Iversen et al. 1986. Much of the information needed to assess feasibility of increasing queen conch production through mariculture is directly relevant to manage- ment of wild stocks. Specific objectives of the Berry Islands field work were to obtain data on age and growth, survival, and optimal habitat for rapid growth and high survival of early life stages. Based on this information, we make recommendations for management of wild stocks. Our study area, the Berry Islands, lies on the northeastern edge of the Great Bahamas Bank (lat. 25°35'N, long 77°45'W) about 190 km east of Miami, FL (Fig. 1). This area is characterized by small cays, shallow sand flats (2-4 m deep) with abundant turtle grass, Thalassia testudinium, beds. The 30 plus islands are located on the west side of the N.E. Pro- vidence Channel and north of the Tongue of the Ocean. The islands are generally low-lying and covered with dense undergrowth, Australian pines, and palm trees. Tidal currents, frequently quite strong, set in and out of the openings between cays. Most cays are privately owned and sparsely popu- lated. Manuscript accepted December 1986. FISHERY BULLETIN: VOL. 85, NO. 2, 1987. 299 FISHERY BULLETIN: VOL. 85, NO. 2 The habitat of the juvenile queen conch in our study area consisted of a large shallow plain sur- rounded by deep offshore waters. With a few ex- ceptions, large adults were found in these deep areas and channels. METHODS AND MATERIALS Collections and observations were made at sta- tions on adjacent cays: Little Whale Cay, Whale Cay, Vigilant Cay, Little Cockroach Cay, Bird Cay, Cat Cay, and Frazer's Hog Cay (Chub Cay) (Fig. 1). Twenty-three field trips were made to the Berry Islands from February 1980 through February 1983, each lasting 4-5 days. The two methods used to obtain growth estimates were tagging-recapture and size-frequency analysis (Cassie 1954). After trying several different tags for conchs, we found that a thin plastic tag measuring 9.5 X 22.3 mm, obtained from the Floy^ Tag Co. 'Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. (Seattle, WA) was satisfactory. It was easily seen and suitable for even the small (ca 2-3 cm) conchs we tagged. A spot on the spire of conchs to be tagged was cleaned and dried, and the tag was affixed with underwater epoxy. We found these tags remained on wild conchs for about 2 years with in- dications of only a few being shed, of the 2,775 conchs we tagged and released at the sites men- tioned above. For growth estimates we measured queen conchs to the nearest mm along the anterior-posterior axis using a measuring board. Significant differences in growth rates of tagged conchs among cays and size classes were tested by analysis of variance. A Student-Neuman-Keuls test was used to detect sig- nificant differences. Differences were considered significant for all statistical tests at the P = 0.05 level. Mean values include 95% confidence inter- vals. We derived whole animal weight-shell length rela- tionships for queen conchs by measuring whole animal weight after removing conchs from their shells. We then removed everything but the foot to 2QO, 1- N 27' 26" 25'»- 24" 23' \J>^ Freeport , 'Miami iBimini „ '•>,. Berry ' \ Islands p^ BERRY ISLANDS Little Cockroacti • Coy Cockroach Cay V Bond s Cay Channel Little Whaie J, Coy "^Whaie Coy FRAZERS HOGfy . '^'^'^"n^v'l / Frazers ^aj ^ • ^°^ rot 'Channel ^ Coy Channel WHALE CAY Scale- km ANDROS Nassau eO'W 79" 78° 77° 76° 75° 74° Figure 1.— Location of study site. 300 IVERSEN ET AL.: BIOLOGICAL DATA ON QUEEN CONCHS measure market weight. Significant differences in weight-length and whole animal weight-market weight relationships of conchs among cays were determined by analysis of covariance. Survival rates were obtained from the decrease in numbers of tagged queen conch at each of the tagging sites using Jackson's formula for monthly estimates and Heincke's formula for annual estimates as described in Everhart et al. (1975). We used pens and cages of varying sizes to evaluate the feasibility of intensive mariculture. Six pens, each 25 m-, were constructed with walls of monofilament webbing 30 cm high, held up by buoys. Pen walls were held in close contact with the bottom by heavy chains and stakes driven into the bottom, and were stocked with conchs 10-15 cm at densities of 1 or 2 conchs/m^. Tagged conchs in this size range released in the vicinity of the pens served as controls. Two additional large pens, 90 and 100 m^ in area, were planted with 1 conch/m^, the conch in the size range of 10-15 cm. Various studies on growth and survival in pens ran from 1 to 15 months. Three wooden floating cages were used to mea- sure growth and survival of small conchs (2-5 cm) over a 1-yr period. They were covered with fine NITEX 4 mm screening and measured 1 x 1 x 0.6 m, stocked with 50 conchs; 0.61 x 0.61 x 0.61 m, stocked with 10 conchs; and 1.6 x 1.2 x 0.6 m, stocked with 100 conchs. Searches for small, young-of-the-year queen conch (<3 cm) were made by towing a dredge, by siev- ing sand samples with 4 mm mesh, by towing divers, by walking and digging on tide flats, by tow- ing a shrimp try net (3 m opening and 1.3 cm stretched mesh), and by a suction dredge (Iversen et al. 1986). To assess density of wild queen conch stocks in shallow water, counts were made along 100 m transects perpendicular and parallel to the shore. All queen conchs lying within 1 m of either side of the transect were counted. Significant differences in density of conchs among cays were tested by analysis of variance. Most searches for queen conch were made dur- ing the day. To determine if this animal's bury- ing activity varied between day and night, we con- ducted day-night counts at several cays and in our pens, and found no differential burying activity. Previous studies in the Virgin Islands (Randall 1964) and Puerto Rico (Appeldoorn and Ballantine 1983) reported no day-night differences in burying activity. RESULTS AND DISCUSSION Queen Conch Distribution and Movement Queen conchs sampled ranged from 2 to 26 cm in length (Fig. 2). The smallest conchs (<10 cm) were found on tidal flats, in shallow waters (<1 m), mostly on sandy bottoms with depressions. The largest juveniles were found in high concentrations near shores of cays, many exposed on low tides. Concen- trations of adults with flared lips almost without ex- ception were found in deep water (>3 m). Juveniles were found associated with cays having tidal flats, available food (microalgae and detritus), beaches with a gradual slope, and good water circulation. None was found in the large, open shallow- water areas between cays. On all 23 field trips, young-of-the-year queen conchs were sought in the course of our regular field activities. The largest concentrations of young were found on the tidal flats between Bird Cay and Cat Cay (Iversen et al. 1986). Lack of shell epibionts on conchs and extensive searching suggested that small queen conchs live in the substrate and in rubble depressions until they are about 0.5 yr old, or about 3-5 cm long, at which time they are found on the tidal flats and nearshore areas in the Berry Islands. Size-frequency distributions (Table 1) showed that smallest individuals spawned the previous year (estimate based on laboratory-reared queen conchs by Siddall [1983], Brownell [1977], and others) ap- peared in winter, spring, and early summer. Large juvenile queen conchs (10-18 cm) were easily located on the substrate surface all year long, generally in shallow water. Relatively few lipped queen conchs (A'' = 109; mean size = 19.3 ± 0.5 cm) were found during the study, most in channels 6 m deep although a few individuals were seen in shallow waters characteristic of most of our study sites. At least one lipped conch was recorded for all areas ex- cept Frazer's Hog Cay and Bird Cay-Cat Cay tidal flats. The smallest lipped conchs were found at Cat Cay (Z = 14.8 ± 2.6 cm; AT = 6). Lipped conchs were found every month except February, July, September, and December, with most found in April {N = 46) and October (A = 39). The distribution and seasonal occurrence of lipped conchs may reflect fishing pressure as much as potential reproductive activity. Studies by Randall (1964), D'Asaro (1965), Brown- ell (1977), and Weil and Laughlin (1984) indicated that queen conchs have a protracted spawning season as long as March to October. Average length 301 FISHERY BULLETIN: VOL. 85, NO. 2 10 8 S O z: LU Z) o LLI DC HAY 1980-FEB 1963 CAT CAY (N - 8,440) UTTLE COCKROACH CAY (N - 3^11) 40 60 80 100 120 140 160 180 200 220 240 JU>fE 1980 - FEB 1983 VIGILANT CAY - OFFSHORE WEST (N = 1,031) VIGILANT CAY - ONSHORE WEST (N = 1,193) ♦0 60 80 100 120 140 160 180 200 220 SHELL LENGTH CLASS MIDPOINTS CmmD of lipped conchs reported for the Virgin Islands was 20.4 cm (Randall 1964). Randall noted that lipped conchs sampled in the Berry Islands were smaller in length than conchs taken elsewhere in the Baha- mas, or in the Virgin Islands. Without exception, tagged queen conchs stayed at the cays where they were released, including transplanted queen conchs from nearby cays. It is possible that we did not observe migration because the majority of conchs we sampled were juveniles. Hesse (1979) reported that adult queen conch in Turks and Caicos ranged farthest (about 2 km) from the tagging site and made seasonal migrations off- shore in fall and inshore in spring, while juveniles moved <1 km. Weil and Laughlin (1984) reported similar movements for adult and juvenile queen conchs in Venezuela. Queen Conch Density by Areas Since the density of queen conch in local areas can affect growth (Alcolado 1976; Weil and Laughlin 1984; Appeldoorn and Sanders 1984), we made den- sity estimates at each of our sampling sites. The mean density of queen conchs at all locations studied, based on 100_m transects taken perpen- dicular to shore, was X = 7.9 ± 1.2 conchs/10 m^. Highest mean densities were found at Bird Cay 302 IVERSEN ET AL.: BIOLOGICAL DATA ON QUEEN CONCHS o LU o LU CC LJL 14 12 10 LnTLi; WHALE CAY (N = 184) MAY. JUNE, AUG &. SEFI 1980 BOND CAY CHANNEL (N = 175) APRIL 19B0. APRIL & OCT 1981 V 40 60 80 100 120 140 160 180 200 220 240 260 SHELL LENGTH CLASS MIDPOINTS Cmm) Figure 2.— Shell length distribution of queen conchs in the Berry Islands. Table 1 .—Size-frequency distribution of queen conch at Little Cockroach Cay (May 1980-April 1981). Length 1980 1981 (cm) May June July' Aug. Sept. Oct. Nov. Dec. Jan. Feb.' Mar. Apr. 8.0-9.0 1 2 9.0-9.9 1 1 10.0-10.9 7 3 1 6 3 2 11.0-11.9 7 7 2 10 18 16 16 12.0-12.9 31 16 4 1 18 45 54 36 13.0-13.9 32 27 14 5 5 1 8 25 58 34 14.0-14.9 35 25 25 12 11 5 7 12 19 20 15.0-15.9 54 34 32 17 27 17 11 13 16 6 16.0-16.9 66 42 48 24 21 15 7 9 11 5 17.0-17.9 37 34 45 25 36 27 5 4 2 4 18,0-18.9 8 8 35 22 44 36 2 3 4 3 19.0-19.9 3 2 9 19 22 18 2 1 1 2 20.0-20.9 2 3 8 15 3 3 2 2 21.0-21.9 1 3 3 1 1 Totals 282 198 217 128 177 137 74 140 189 131 Lipped conchs 0 0 1 0 1 3 0 1 0 0 'No data. Channel, X = 19.6 ± 2.6 conchs/10 m^, followed by Vigilant Cay west, X =_13.5 + 3.1 conchs/10 m^, and Vigilant Cay east, X = 12.2 + 2.9 conchs/ 10 m-. Lowest mean densities were found at Little Cockroach Cay, X = 1.5 ± 0.7 conchs/10 m^ and Little Whale Cay north, Z = 3.3 ± 0.7 conchs/10 m2 (Table 2). The mean queen conch density for all locations combined varied between June, September, and November. Density was highest in June {X = 9.9 ± 0.3_conchs/10 m^, N = 978) followed by Novem- ber (X = 6.7 ± 0.4 conchs/10 m^, A^ = 673) and September {X = 6.0 ± 0.3 conchs/10 m^; N = 722). Queen conchs were randomly distributed over the 100 m transect at all locations except Cat Cay and Vigilant Cay, according to the results of a serial ran- 303 FISHERY BULLETIN: VOL. 85, NO. 2 domness test (Zar 1974). Conchs were especially clumped at Cat Cay, all appearing within 10 m of shore, effectively making their densities much higher than were reported for 100 m transects (Table 2). Queen conch densities reported for other areas in the Caribbean were generally lower, ranging from 0.8-5.2 conchs/10 m^ in Cuba (Alcolado 1976), 0.01 conchs/10 m- in U.S. Virgin Islands (Wood and Olsen 1983), 0.9 conchs/10 m^ in the Turks and Caicos (Hesse 1979) to 0.1-21 conchs/10 m^ {X = 4.2 conchs/10 m-) in Los Roques, Venezuela (Weil and Laughlin 1984). Growth of Queen Conch by Season, Location, and Size Seasonal Growth Based on all data collected between February 1980 and June 1982, there was a significant (P < 0.001 ANOVA) seasonal difference in mean length of un- tagged individuals. Queen conchs measured during winter were smaller than those measured during other seasons (Table 3). Nearly all growth of juvenile queen conchs in our study took place during the warm summer months, May-September. At Cat Cay, for example, mean growth of tagged conchs ranged from 0.44 to 1.63 cm per month during the summer, and from 0.18 to 0.30 cm per month during the remainder of the year. This is consistent with studies by Randall (1964) on queen conch in St. Croix, U.S. Virgin Islands; by Alcolado (1976) in Cuba; and by Appel- doorn (1985) on small juveniles in Puerto Rico. Our small caged conchs (2.4-3.6 cm at tagging), held for 1 year, increased 3.56 cm on the average; 92% of this increase (3.27 cm) took place between April and October. Growth by Location and Size To examine the effect of location and size on growth, mean monthly growth of penned and un- penned tagged queen conch was compared within 3 size groups (<9.6 cm, 9.7-15.3 cm, >15.4 cm) by location. Densities of penned conchs (10-20 conchs/ 10 m-) were higher than densities of unpenned conchs (2-20 conchs/10 m^, X = 8) measured in the field. In every size class, unpenned conchs grew significantly faster (P < 0.001, ANOVA). Among unpenned conchs, there was a significant interaction effect of location and size on mean monthly growth. Large conchs (>15.4 cm) at Little Cockroach Cay, Table 2. — Average density of queen conchs by sampling sites. Average density (conchs/10 m^ + Sampling site 95% C.I.) N Little Cockroach Cay 1.5 + 0.7 88 Little Whale Cay (North) 3.3 + 0.7 66 Vigilant Cay (West) 13.5 + 3.1 808 Vigilant Cay (East) 12.2 + 2.9 733 Bird Cay Channel 19.6 -1- 2.6 391 Cat Cay (North) 4.2 + 1.2 249 Cat Cay (East) 6.2 + 3.3 372 Whale Cay 6.2 + 1.7 123 Table 3.— Mean length of untagged queen conchs collected in each season between February 1980 and June 1982. Mean length Season Months (cm) N Winter December January 11.9 + U.& 1,432 February Spring March April 12.9 + 1.1 3,475 May Summer June July 12.6 + 1.2 3,317 August Fall September October 13.0 + 1.3 2,943 November ' + 95% confidence interval. where density was lowest, grew faster than all other sizes. Small conchs (<9.6 cm) grew the next fastest, followed by intermediate-sized conchs (Table 4). Queen conchs at Cat Cay and Vigilant Cay offshore west, where densities were higher, grew slower as Table 4. — Comparison of mean monthly growth rates (cm) for un- penned queen conchs. Underlined locations indicate significant difference in monthly growth between locations as determined by Student-Newman-Keuls test. Size class Tagging locations 9.6 cm Vigilant Cay Little Whale Cat Cay Offshore West Cay (0.40) + '0.3 (0.48) ± 0.03 (0.50) ± 0.04 A/ = 198 A/ = 13 N = 114 9.6-15.3 cm Cat Cay (0.25) ± 0.02 N = 385 Little Whale Little Cockroach Cay Cay (0.40) ± 0.03 (0.48) ± 0.02 A/ = 186 N = 248 15.3 cm Little Whale Little Cockroach Cay Cay (0.31) ± 0.07 (0.50) ± 0.03 A/ = 23 A/ = 146 '95% confidence interval. 304 IVERSEN ET AL.: BIOLOGICAL DATA ON QUEEN CONCHS a group than queen conchs at Little Whale Cay and Little Cockroach Cay (Table 4). Alcolado (1976) also reported that queen conchs in Cuba grew slower in areas of high density (5.2 conchs/10 m^) than in areas of low density (0.08 conchs/10 m^). Appel- doorn and Sanders (1984) reported similar results in a laboratory experiment on small juvenile conchs. There was no significant interaction effect be- tween size and location on growth of penned queen conchs. Smallest conchs in cages grew fastest, followed by intermediate-sized conchs (Table 5). There were insufficient data for large penned conchs to estimate their mean monthly growth. Among the intermediate-sized conchs where density was known, mean monthly growth was highest in pens with the lowest density (0.1/10 m- compared with 0.2/10 m-). Randall's (1964) penned queen conchs (mean length 6.2 and 7.5 cm, range 5.2-8.0 cm; N = 25) grew slowly (0.26 cm/month), but these measure- ments were made during winter months. Pen size was not specified. In another experiment^ Randall placed 16 tagged conchs (19.0-20.0 cm, X = 19.4 cm) in a "60 ft by 140 ft elliptical fenced area" dur- ing winter and reported average growth of 0.1 cm/ month through April when the experiment was discontinued. Growth rates for our larger penned conchs (mean length 10.3 cm) approximated Randall's rates (0.1 and 0.2 cm/month), even though our data were re- corded throughout the year. Growth rates were higher for our smaller conchs (mean length = 4.6 cm; mean growth = 0.4 and 0.2 cm/month) than for larger conchs. modes of 634 queen conchs measured in October 1980 were present at 7.6, 12.5, and 17.0 cm, sug- gesting length at ages 1, 2, and 3, respectively. Parameters in the von Bertalanffy (1938) equa- tion were estimated by fitting a Walford (1946) line to tagging data from Cat Cay, Vigilant Cay, Little Whale Cay, and Little Cockroach Cay {N = 117). Fitting a Walford growth line requires that the growth rate decreases with age. Since the largest queen conchs at Little Cockroach Cay grew faster than the middle-sized juveniles, we excluded these data from our calculations and obtained the fol- lowing estimates of average length by ages (Table 6). Age Lt (cm) I 8.3 II 12.2 III 15.4 IV 18.1 With L„ = 30.0 K = 0.20 ta = -0.65 Our estimates of length at age from both length- frequency analyses and von Bertalanffy estimates of tagging data (excluding Little Cockroach Cay data) indicate that queen conch in the Berry Islands grow more slowly than those in the Virgin Islands and some of the areas in Cuba where density was low (0.8 conchs/10 m^). We suggest that the higher densities of queen conch and cooler water temper- atures in the Berry Islands may slow their growth relative to other areas. Length at Age Estimates of length at ages 1-3 were obtained for Berry Islands queen conchs by length-frequency analysis (Cassie 1954) and by fitting the von Bertal- anffy equation to tagging data. Distinct length Length-Weight Relationship Whole animal weight(minus the shell)-shell length relationships were derived for queen conch sampled at Chub Cay (N = 39), Frazer's Hog Cay (A^ = 32), and Bird-Cat Cay Channel (A'' = 34). Log^o Table 5. — Comparison of mean monthly growth rates (cm) for penned queen conch. Underlined locations indicate no significant difference in monthly growth between locations as determined by Student-Newman-Keuls test. Size class Tagging location 9.6 cm Pen 7 (0.04) ± '0.07 W = 25 Pen 9 (0.21) ± 0.06 A/ = 25 Small Wood Cage (0.24) ± 0.06 A/ = 38 Large Wood Cage (0.35) ± 0.06 A/ = 66 9.7-15.3 cm Pen 5 (-0.05) ± 0.04 A/ = 38 Pen 6 (-0.01) ± 0.01 A/ = 38 Pen 7 (0.04) ± 0.02 A/ = 64 Pen 2 (0.08) ± 0.02 A/ = 48 Pen 9 (0.11) ± 0.04 A/ = 56 Pen 3 (0.15) ± 0.03 A/ = 45 Pen 1 (0.17) ± 0.05 A/ = 34 15.4 cm Insufficient data '95% confidence interval. 305 FISHERY BULLETIN: VOL. 85, NO. 2 Table 6. — Estimates of queen conch length (cm) at age from the Caribbean. Berry Islands,^ Ypar Bahamas Piiprtn U.S. Virgin Islands class a b Rico^ St. John^ St. Thomas'* Cuba^ Venezuela® 1 7.3 8.3 8.8 10.8 9.0 7.9-11.2 7.6 II 12.5 12.2 12.6 17.0 12.6 12.50-18.8 12.8 III 17.0 15.4 18.0 20.5 15.7 15.5-24.3 18.0 IV 18.1 17.4-28.3 L (cm) 30.3 26.0 20.8-38.3 K 0.20 0.52 0.287-0.571 to -0.65 0 -0.12-0.13 N 634 103 193 104 301 63-284 161 'This study • size frequence (a) and von Bertalanffy fit to tagging data (b). 2Berg 1976 - size frequency. ^Berg 1976 and Brownell et al. 1976 - von Bertalanffy fit to Randall's (1964) tagging data. ■•Wood and Olsen 1983 - size frequency. ^Alcolado 1976 - von Bertalanffy fit to tagging data from 7 locations. ^Brow/nell 1977 - size frequency. (weight)-Logio (length) relationships best fit the data. Analysis of covariance showed that queen conch at Frazer's Hog Cay and Chub Cay had similar whole animal weight-shell length relationships but that both differed from conchs at Bird-Cat Cay Channel. Therefore, two relationships were devel- oped. Frazer's Hog-Chub Cay Logio (whole animal weight) = -2.40 + 3.57 X Logio (shell length) r = 0.95 A^ = 71 Bird-Cat Cay Channel Logio (whole animal weight) = -1.36 -i- 2.84 X LogiQ (shell length) r = 0.93 N = 34. Mean lengths of queen conch at Frazer's Hog Cay {X = 15.6 ± 0.7 cm) and Chub Cay (X = 18.6 ± 0.8 cm) were significantly (P < 0.001) larger than those at Bird-Cat Cay Channel {X = 13.6 ±0.1 cm). Shell length-whole animal weight relationships changed with size. Smaller conchs (X = 13.6 ±0.1 cm) increased in weight per unit length faster than did larger conchs {X = 17.0 ± 0.7 cm). We found a close linear relationship between whole animal weight and meat weight of Berry Islands queen conch which did not vary among areas: market weight = 0.65 (whole animal weight) + 6.00 A^ = 105; r = 0.97. The relationship between shell length and animal weight, although significant (P < 0.001) was not as close: market weight = 11.47 (shell length) - 50.69 N = 105; r = 0.84. Table 7 gives the numbers of different aged queen conchs in the Berry Islands needed for 1 pound of market meat. Using the whole animal weight-shell length and whole animal weight-market weight rela- tionships developed above and assuming size at lip formation (14.8-19.3 cm) is the size at harvest, 4-10 queen conchs are needed to produce 1 pound of meat (Table 7). In the Berry Islands 6-8 conchs are needed to make 1 pound of market meat, as opposed to 2-3 and 3-4 conchs/pound from other areas in the Bahamas (Berg 1981). The high numbers of queen conchs per pound of market meat from the Berry Islands may be partially explained by their stunted growth. Table 7. — Number of Berry Island queen conch required to make 1 pound of market meat.^ von Bertalanffy Whole animal No. of conchs estimated length weight/conch Market to make 1 lb Age at age (g) wt(lb) of meat 1 8.3 ^17.8 0.04 26 II 12.2 ^53.1 0.09 11 III 15.4 =^73.0 0.12 9 IV 18.1 ^130 0.20 5 IV^ 19.3 164 0.25 4 'Market meat = (0.65) (wfiole animal weight - shell weight) + 6. 2Shell length (cm) converted to whole animal weight (gm) with Bird-Cat Cay Channel regression log,^ (weight) = -1.36 + 2 84 log,„ (shell length). ^Shell length (cm) converted to whole animal weight (g) with Frazer's Cay- Chub Cay regression log,^ (weight) = -2.40 + 3.57 log,(, (shell length). "Mean size of lipped queen conchs sampled in Berry Islands. 306 IVERSEN ET AL.: BIOLOGICAL DATA ON QUEEN CONCHS Survival-Mortality Estimates of monthly and annual survival of un- penned and penned queen conchs in the Berry Islands were derived from tagging studies. The estimates assume that tags are not overlooked, that tags do not fall off or affect survival, that the tagged population is similar to the untagged population in all other respects and that no emigration occurs dur- ing the experiments. Monthly survival rate of unpenned queen conchs ranged from 57 to 80%, depending on location (Table 8). The estimates for Little Cockroach, Vigilant, and Cat Cays are the most reliable, because more conchs were tagged over a longer period of time at these three locations than at others. Annual survival was low for these areas, ranging from 2 to 9%. These proportions result in estimates of total instantane- ous mortality rate, Z, from 2.41 to 3.91, considerably higher than those reported by Alcolado (1976) for queen conch in Cuba (annual survival 15-35%, Z from 1.06 to 1.90) or by Wood and Olsen (1983) for recruited queen conch in St. Thomas, U.S. Virgin Islands (Z from 0.22 to 1.80). Appeldoorn (1985) found mortality of small juven- ile queen conchs (<6.4 cm) in Puerto Rico higher at Z = 8.62, or an annual survival of 0.02%. In a re- cent study of large juveniles and adults, Appeldoorn (in press) estimated annual Z = 2.67, with M plus emigration = 1.53 and F = 1.14. Our survival estimates are probably low because of problems in- herent to tagging studies mentioned above. Survival of penned queen conchs in 4 of our pens was much higher than survival of unpenned animals (Table 8). Monthly survival ranged from 90 to 97% for penned animals, and annual survival ranged from 28 to 73%. However, all conchs died in the 2 deeper water pens. We attribute much of the in- creased survival rate of queen conch in our best pens to reduced predation, although an undetermined portion of the increase is due to eliminating emigra- tion and the increased probability of finding tagged animals in an enclosure. However, the positive in- fluence of increased survival must be balanced with the slow growth rates of all but the smaller (4.5-8 cm) sizes in pens. Seasonality of burying, as demonstrated in our penned conch data and by Hesse (1979), using un- penned queen conch, can affect the estimated sur- vival rate because some of the animals cannot be found. Based on our results in pens, the possible error in survival estimate due to burying would probably be relatively small because over about a 1-yr period, a total of 25 out of 200 queen conchs (about 12%) were buried; however, on any single monitoring trip only 1 or 2 individuals were buried. Causes of Mortality It is well documented that predation plays a sig- nificant role in the survival of queen conchs (Jory 1982; Jory and Iversen 1983; Iversen et al. 1986) and will not be detailed here. After settlement, at all sizes, even flared lipped, thick-shelled adults are subject to predation by large turtles and fishes; however, the rate of predation is significantly higher on the small, thin-shelled juveniles (2-5 cm) and decreases as the animals grow. Based on our observations, those of our colleagues and reports in the literature, predation, rather than abiotic factors of the environment, or parasites and disease, seem to be the most important causes of queen conch mortality. Hence, stock size after settle- ment appears to be predator-controlled. This is not an unusual finding when the wide range of species of queen conch predators feeding on all sizes of conch is considered (Randall 1964; Jory and Iversen 1983; Iversen et al. 1986), together with the impor- tant role that predation plays in the mortality of many other mollusks (Jory et al. 1984). Table 8. —Survival estimates foi r Berry Islands queen conchs. Unpenned conchs Little Whale Cay penned conchs Little Little Small Large Cockroach Vigilant Cat Whale wooden wooden Cay Cay Cay Cay cage cage Pen 1 Pen 2 Pen 3^ Pen 4 Monthly survival^ 0.80 0.72 0.80 0.57 0.96 — 0.93 0.97 0.92 0.90 Annual survival 0.02 0.09 0.09 0.02 0.69 0.13 0.36 0.73 0.36 0.28 Number tagged _ conch released 282 169 418 59 26 15 25 30 50 25 X size (cnn) 15.0 8.4 11.4 13.0 7.2 7.6 10.5 10.2 10.2 10.5 Range (cm) 8.2-22.3 6.5-11.7 8.1-19.1 8.6-20.6 5.0-8.0 4.5-9.0 8.5-12.0 6.0-12.5 6.0-12.0 8.5-12.5 'Only 11 months involved. ^Monthly survival estimates made using Jackson's formula. ^Annual survival estimates made using Heincke's formula. 307 FISHERY BULLETIN; VOL. 85, NO. 2 Management Implications: Yield per Recruit Yield-per-recruit analysis, based on our estimate of the von Bertalanffy growth equation and total mortality from tagging, was conducted for the shallow-water queen conch populations in the south- ern Berry Islands. Yield per recruit was computed from the model given by Beverton and Holt (1957) which assumes that growth rate, susceptibility to capture, and natural mortality remain constant after age of recruitment. We believe that the combined data from unpenned queen conch in all areas (excluding Little Cockroach Cay) gave us accurate estimates of growth and mor- tality. We estimated maximum meat weight for Berry Island conchs to be 463 g, based on the shell length to whole animal weight regression from the Frazer's Hog Cay-Chub Cay area, which was the largest sample. Age at recruitment was assumed to be 3 years (corresponding to a length of 15.4 cm). Maximum age of queen conch, based on L^ = 30.3 cm and our von Bertalanffy equation, was 11 years. Using an overall mean annual survival of 7%, Z is 2.66. Estimates of yield per recruit were obtained for a range of values of M between 0.50 and 2.6. F varied between 0.16 and 2.66, by increments of 0.50, and ^0 varied between 1 and 5 years by increments of 1.10 for each level of M. This analysis probably encompassed any value of M and F that actually ex- isted during the study. AtM = 0.50, age liable to capture that maximizes yield in weight per recruit is 3 years, over a range of F from 0.16 to 2.66. Thus, if fishermen take queen conch of approximately 15 cm and larger, they would maximize the yield available from the popula- tion. At all values of M above 0.50, results indicate a stage of underfishing because yield in weight per recruit increases over the range of F and decreases with increasing age liable to capture. The results of yield-per-recruit analyses are limited for several reasons. First, larger, faster growing queen conchs from Little Cockroach Cay were excluded from the analyses; the von Bertal- anffy equation did not describe well the growth of conchs from the full data set. Second, the range of sizes in the tagging studies did not accurately reflect the range of sizes in the total queen conch popula- tion in the southern Berry Islands. Most data were collected on immature conchs (before lip formation) that were living on shallow flats near islands. While the purpose of analyzing prospects for mariculture were adequately fulfilled by these data, they should not be used for fisheries management because they, for the most part, do not include the larger adults found in deeper channels between cays. While these data are preliminary, they indicate an important management principle that was also determined for queen conch in the U.S. Virgin Islands by Wood and Olsen (1983); namely, that maximum yield per recruit is obtained at age of first harvest, which is just at onset of lip formation. In the Virgin Islands they found the maximum yield could be obtained by harvest between 3 and 5 years, at an average length of 15.78-19.1 cm. Maximum yield per recruit may occur below onset of matur- ity, however, since there is some evidence that queen conch may not be reproductively active until some time after lip formation (Wilkins et al., in press). Mariculture Potential Queen conch mariculture potential, one objective of this study, was investigated as a possible means to increase conch production in the Bahamas. A hatchery was established at the University of Miami. Techniques were developed for mass-rearing queen conch from egg masses collected in the wild through the larval stages (Siddall 1983). However, because of the high hatchery costs and high mortality associated with planting young mollusks in the wild (Iversen et al 1986; Jory et al. 1986), supplementing natural conch stocks by extensive mariculture does not appear to be economically feasible at this time. We placed juvenile conchs in pens at densities slight- ly higher than those in nature and found very slow growth, meanwhile experiencing considerable dif- ficulty in physically maintaining the pens. Further, complete mortalities occurred in some pens, which we cannot explain. Our results suggest that, for the numbers of queen conch required for supplementing natural stocks, the techniques available could probably only be success- ful in certain well-protected areas. In Bonaire, a degree of success has been reported by Hensen (1983). For intensive mariculture, unless a special area is found with good water exchange, natural food availability, where most predators can be ex- cluded, and large juvenile conchs released, this tech- nique of attempting to enhance production does not appear to have much potential at this time. With additional research, particularly on developing dependable hatchery techniques and cost-efficient means of predator protection, intensive mariculture may some day play a useful role in increasing production. 308 IVERSEN ET AL.: BIOLOGICAL DATA ON QUEEN CONCHS ACKNOWLEDGMENTS This research was sponsored by the Wallace Groves Aquaculture Foundation whose contribution in research funds and logistic support is gratefully acknowledged. Special thanks go to A. Albury, K. Lightbourne, and other employees on Little Whale Cay for their enthusiastic and unstinting support during this study. Able assistance in the field was provided by C. Berg, P. Bizzell, I. Brook, W. Brown- ell, L. Cresswell, L. FitzGerald, C. Grail, R. Hensen, E. J. Iversen, E. S. Iversen, Jr., J. Iversen, K. Orr, and G. Woon. S. Siddall set up the required com- puter programs to synthesize data. R. Hensen made conch density and length-weight measurements and was the first to locate the habitat of the young-of- the-year in Bird Cay-Cat Cay Channel. LITERATURE CITED Alcolado, p. M. 1976. Crecimiento, variaciones morfologicas de la concha y algunos datos biologicos del cabo, Str-ombus gigas, L. (Mollusca: Mesogastropoda). Acad. Cienc. Cuba, Ser. Oceanol. 34, 36 p. Appeldoorn, R. S. 1984. The effect of size on mortality of small juvenile conchs (Strombus gigas Linne and S. costatus Gmelin). J. Shellfish Res. 4(l):37-43. 1985. Growth, mortality and dispersion of juvenile laboratory- reared conchs, Sfrombiis gigas and S. costatus, released at an offshore site. Bull. Mar. Sci. 37:785-793. In press. Ontogenetic changes in natural mortality rate of queen conch, Strombus gigas (Mollusca: Mesogastropoda). Bull. Mar. Sci. Appeldoorn, R. S., and D. L. Ballantine. 1983. Field release of cultured queen conch in Puerto Rico: Implications for stock restoration. Proc. Gulf Carib. Fish. Inst. 35:89-98. Appeldoorn, R. S., and I. M. Sanders. 1984. Quantification of the density-growth relation in hatch- ery-reared juvenile conch {Strombus gigas Linne and S. costatus Gmelin). J. Shellfish Res. 4(l):63-66. Berg, C. J., Jr. 1976. Growth of the queen conch, Strombus gigas, vdth a discussion of the practicality of its mariculture. Mar. Biol. (Berl.) 34:191-199. 1981. Proceedings Queen Conch Fisheries and Mariculture Meeting. Wallace Groves Aquaculture Foundation, Bahamas, 46 p. Bertalanffy, L. von. 1938. A quantitative theory of organic growth (inquiries on growth laws II). Human Biology 10:181-213. Beverton, R. J. H., and S. J. Holt. 1957. On the dynamics of exploited fish populations. Invest. Minist. Agric. Fish Food (G.B.) Ser. II, 19, 533 p. Brownell, W. N. 1977. Reproduction, laboratory culture, and growth of Strom- bus gigas, S. costatus, and S. pugilis in Los Roques, Vene- zuela. Bull. Mar. Sci. 27:668-680. Brownell, W. N., C. J. Berg, Jr., and K. C. Haines. 1976. Fisheries and aquaculture of the conch, Strombus gigas, in the Caribbean. Paper presented to CICAR-II Sym- posium, Progress in Marine Research in the Caribbean and Adjacent Regions, Caracas, Venezuela, July 12-16, 1976, p. 59-69. Brownell, W. N., and J. M. Stevely. 1981. The biology, fisheries and management of the queen conchs, Strombus gigas. Mar. Fish. Rev. 43(7): 1-12. Cassie, R. M. 1954. Some uses of probability paper in the analysis of size frequency distributions. Aust. J. Mar. Freshwater Res. 5:513-522. D'ASARO, C. N. 1965. Organogenesis, development, and metamorphosis in the queen conch, Strombus gigas, with notes on breeding habits. Bull. Mar. Sci. 15:359-416. Everhart, W. H., a. W. Eipper, and W. D. Youngs. 1975. Principals of fishery science. Comstock Publ. Assoc, Cornell Univ. Press, Ithaca, NY, 228 p. Hensen, R. R. 1983. Queen conch management and culture in the Neth- erlands Antilles. Proc. Gulf Carib. Fish. Inst. 35:53- 56. Hesse, K. 0. 1979. Movement and migration of the queen conch Strom- bus gigas, in the Turks and Caicos Islands. Bull. Mar. Sci. 29:303-311. Iversen, E. S. 1983. Feasibility of increasing Bahamian conch produc- tion by mariculture. Proc. Gulf Carib. Fish. Inst. 35:83- 88. Iversen, E. S., D. E. Jory, and S. P. Bannerot. 1986. Predation on queen conchs, Strombus gigas, in the Bahamas. Bull. Mar. Sci. 39:61-75. Jory, D. E. 1982. Predation by tulip snails, Fasciolaria tulipa. on queen conchs, Strombus gigas. M.S. Thesis, Univ. Miami, Coral Gables, FL, 73 p. Jory, D. E., and E. S. Iversen. 1983. Queen conch predators: not a roadblock to mariculture. Proc. Gulf Carib. Fish. Inst. 35:108-111. 1985. Molluscan mariculture in the greater Caribbean: An overview. Mar. Fish. Rev. 47(4): 1-10. Jory, D. E., M. R. Carriker, and E. S. Iversen. 1984. Preventing predation in molluscan aquaculture. J. World Mariculture Soc. 15:421-432. Randall, J. E. 1964. Contribution to the biology of the queen conch, Strom- bus gigas. Bull. Mar. Sci. Gulf Carib. 14:246-295. Siddall, S. E. 1983. Biological and economic outlook for hatchery produc- tion of juvenile queen conch. Proc. Gulf Carib. Fish. Inst. 35:46-52. Walford, L. a. 1946. A graphic method of describing the growth of animals. Biol. Bull. (Woods Hole) 90:141-147. Weil, E., and R. G. Laughlin. 1984. Biology, population dynamics, and reproduction of the queen conch Strombus gigas Linne in the Archipelago de los Roques National Park. J. Shellfish Res. 4(1):37- 43. WiLKiNS, R. M., M. H. Goodwin, and D. M. Reid. In press. Research applied to conch resource management in St. Kitts/Nevis. Proc. Gulf Carib. Fish. Inst. 309 FISHERY BULLETIN: VOL. 85, NO. 2 Wood, R. S., and D. A. Olsen. Zar, J. H. 1983. Application of biological knowledge to the management 1974. Biostatistical Analysis. Prentice-Hall, Inc. Englewood of the Virgin Islands conch fishery. Proc. Gulf Carib. Fish. Cliffs, NJ, 620 p. Inst. 35:112-121. 310 REVISION OF THE GAMBA PRAWN GENUS PSEUDARISTEUS, WITH DESCRIPTION OF TWO NEW SPECIES (CRUSTACEA: DECAPODA: PENAEOIDEA) Isabel Perez Farfante' ABSTRACT The genus Pseudaristeus (family Aristeidae) is widespread in the Indo-West Pacific where five species— P. crassipes, P. gracilis, P. kathleenae n. sp., P. protensus n. sp., and P. sibogae—have been found at depths between 719 and 1,785 m; another species, P. speciosus, occurs in the southwestern Atlantic, where it was taken at 4,847 m. Females of Pseudaristeus possess a styliform, long rostrum, 0.70-1.40 as long as the carapace; most males have shorter ones 0.20-0.57, but some have been found in which the rostrum is as long as that of the females. This suggests that in this genus, as in AristeiLs, at certain stages of the life cycle, males develop a long rostrum. Following a revised definition of the genus, a key to the species, a synonymy, the location of the types, type-locality, and a list of specimens examined are given for each species. Detailed morphological accounts, including intraspecific variation, accompany statements of maximum sizes, and geographic and bathymetric ranges. The description of P. gracilis is the first to take into consideration adult material. In discussing relationships, P. crassipes, P. kathleenae, and P. protensTis, sympatric off the coast of India, are shown to constitute a rather homogenous group somewhat distantly related to the other very distinctive members of the genus. The deep-sea "gamba prawns" of the genus Pseuda- risteus are widely distributed at depths of 719-1,785 m in the Indo-West Pacific— from the Gulf of Aden and off Natal, South Africa to the Philippines— where five (P. crassipes, P. gracilis, P. kathleenae n. sp., P. protensus n. sp., P. sibogae) of the six species recognized herein are found. The sixth (P. speciosus) occurs in the southwestern Atlantic, off northeastern Argentina. It is unlikely that members of the genus occur in the northwestern Atlantic, in- cluding the Gulf of Mexico and the Caribbean, where, although intensive collecting have been con- ducted, no Pseudaristeus have been taken. The large gamba prawns (reaching as much as 47.5 mm carapace length, about 150 mm total length) that occur at shallower depths might, in the not too distant future, make a minor, but highly esteemed, contribution to the commercial catches of penaeoids in certain areas of the Indo-West Pacific, as do members of other deep-sea genera {Aristaeomorpha, Aristeus, and Plesiopenaeus) of the family Aristeidae. The genus Pseudaristeus has been poorly under- stood largely because the original descriptions of the first three recognized species, P. speciosus, P. gracilis, and P. crassipes, are inadequate for sep- ^Systematics Laboratory, National Marine Fisheries Service, NOAA, U.S. National Museum of Natural History, Washington, DC 20560. arating them, and the illustrations of the latter two, although well rendered, are of little help. Incomplete descriptions and inadequate illustrations were pri- marily responsible for subsequent assignment of specimens of two new species described here to P. crassipes, which, like them, occurs in the waters off India. Availability of the rich collections of Pseudaristeus made by the U.S. Bureau of Fisheries steamer Albatross during the Philippine Expedition, 1907-10, and the loan of critical material from several museums have enabled me to make detailed studies of all six species of the genus. Included in the ac- counts of each are numerous diagnostic characters which have not been previously recognized in the four described species. One of these characters is the sinuous ventral antennular flagellum found in male P. gracilis, which not only allows a ready iden- tification of these animals but constitutes another significant element for the interpretation or a better understanding of the relations of the members of Pseudaristeus to those of the closely allied Aristeus. PRESENTATION OF DATA In the account of the species, most of the termi- nology used follows that proposed and illustrated by Perez Farfante (1969, 1977). The anterolateral carina, a unique feature of one member of Pseuda- risteus, which has not been cited by me previously, Manuscript accepted January 1987. FISHERY BULLETIN: VOL. 85, NO. 2, 1987. 311 FISHERY BULLETIN: VOL. 85. NO. 2 extends between the gastro-orbital and branchi- ostegal-hepatic carinae. The names of various parts of the eye, adopted from Young (1956, 1959), were recently employed and illustrated by Perez Farfante (1985). The measurement of rostrum length (RL) is the linear distance from apex to orbital margin, that of carapace length (CL) is the distance between or- bital margin and the midposterior margin of the carapace, and, finally, that of total length (TL) is the distance from the apex of the rostrum to poste- rior end of the telson. All measurements are made to the nearest 0.5 mm. The petasmata have been described and all but one depicted unfolded; the il- lustrations were made from stained specimens. Be- cause more than one species have been found in lots reported by various authors under a single name and new species are described herewith from waters from which records have been previously cited, the map is based on only the specimens examined by me. Scales accompanying the illustrations are in milli- meters. Abbreviations of the repositories of the specimens examined during this study are as follows: BMNH - British Museum (Natural History), London. MP - Museum National d'Histoire Naturelle, Paris. USNM - National Museum of Natural History, Smithsonian Institution, Washington, D.C. ZMA - Zoologisch Museum, Amsterdam. ZMB - Zoologisches Museum der Humboldt - Uni- versitat, Berlin. ZSI - Zoological Survey of India, Calcutta. GENUS PSEUDARISTEUS CROSNIER, 1978 Hemipenaeus Bate 1881:186 [part]. Aristaeus Wood-Mason and Alcock 1891:278 [part]. Aristeus Anderson 1896:91. Pseudaristeus Crosnier 1978:81 [type species, by original designation, Aristaeus crassipes Wood- Mason 1891:281. Gender masculine]. Diagnosis.— Body slender, covered with densely set minute setae. Rostrum long in females and short or long in males; armed with 2 dorsal teeth; epigastric tooth distinctly posterior to first rostral, situated about 0.1 CL from orbital margin. Antennal and branchiostegal spines present; orbital, pterygo- stomian, and hepatic spines lacking. Cervical sulcus crossing postrostral carina (rarely only reaching it); postcervical sulcus extending to postrostral carina; gastro-orbital, antennal, branchiostegal-hepatic, and branchiocardiac carinae strong; hepatic sulcus long. usually fusing with branchiocardiac sulcus and descending obliquely almost to margin of branchi- ostegite. Abdomen with dorsomedian carina extend- ing from fourth through sixth somites; elongate sixth somite bearing pair of long cicatrices. Telson produced posteriorly in sharp, median spine and with posterior 0.4 of length armed with 4 pairs of small, movable, lateral spines. Eye with well- developed cornea and dorsoventrally depressed peduncle bearing mesial tubercle; basal article not produced in scale. Antennular peduncle length about 0.55 CL; prosartema rudimentary, consisting of short stump bearing brush of long setae; dorsal flagellum short, about 0.4 length of antennular pe- duncle, and flattened; ventral flagellum long, no less than 2.75 CL, and filiform. Mandibular palp (Fig. L4 ) reaching to about base of ischiocerite (third ar- ticle of antennal peduncle), distal article suboval and much smaller than basal. Palp of first maxilla un- segmented (Fig. IB). Second maxilla and first and second maxillipeds as illustrated (Fig. IC-E). Exo- pods on all maxillipeds but lacking on pereopods. Petasma with dorsomedian lobule short, only about 0.4 length of petasma; ventromedian lobule with rib narrow proximally, broadening to level of distal end of dorsomedian lobule where reaching mesial mar- gin, and continuing almost to end of lobule; ventral costa distally free, not attached to dorsolateral lobule; endopod of second pleopod bearing appen- dices masculina and interna. Thelycum of "open type", with large, lanceolate median plate on ster- nite XIII. Well-developed podobranchia on second and third maxillipeds and first and second pereo- pods, those on pereopods subequal in size. One ar- throbranchia on somite VII and two on VIII through XIII, all well developed except very small anterior one on somite VIII. Pleurobranchia on somites IX through XIV, that on XIV well developed, remain- ing ones much smaller. Nonbifurcate, large epipod on somites VII through XII, that on XII subequal in size to that on XI. (Modified from Crosnier 1978.) It seems worth emphasizing that although the rostrum is long in females, 0.70-1.40 CL, and rela- tively short in most of the males examined, 0.20- 0.57 CL, I have found one male in each of two species with long rostra, 0.70 and 1.40 CL. Five species known from the Indo-West Pacific are P. crassipes, P. gracilis, P. kathleenae n. sp., P. protensus n. sp., and P. sibogae; and the one from the southwestern Atlantic is P. speciosus. The species from the Indo-West Pacific have been taken at depths between 750 and 1,785 m and that from the southwestern Atlantic at 4,847 m. These depths and all others cited are those noted for the stations 312 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS -A-C D E -f i ' Figure \.—Pseudaristeus kathleenae n. sp., 9 35 mm CL, Lagonoy Gulf, east of southern Luzon, Philippines. A. Mandible. B, First maxilla. C, Second maxilla. D, First maxilliped. E, Second maxilliped. /, Rudimentary arthrobranchia. /', Enlargement of/. (All from left side). Scales: A-E = 3 mm;/' = 1 mm. at which the specimens were obtained, but because open nets were used in collecting, it is not possible to ascertain the actual level at which the shrimp entered the net. Pseudaristeus is quite close to Aristeus, having the same branchial formula, but differing from it in exhibiting a well-marked cervical sulcus which reaches the dorsomedian carina of the carapace and in possessing a postcervical sulcus. Among the mem- bers of Pseudaristeus only males of P. gracilis ex- hibit a sinuous ventral antennular flagellum, a feature characteristic of the males of all species of the genus Aristeus. This similarity, together with a branchial formula common to the two genera, in- dicates a close affinity between members of the genera Pseudaristeus and Aristeus and is a con- vincing basis for the postulate that they have had a more recent common ancestor than either share with Hemipenaeus. The species now assigned to Pseudaristeus were placed in the genus Hemipen- aeus by almost all authors in this century until Crosnier (1978) proposed that they be removed to his new genus. Key to species of Pseudaristeus 1. Ventral extremity of cervical carina blunt, not forming sharp-edged arc. Anterolateral carina present. Posterior part of hepatic sulcus not fusing with branchiocardiac sulcus but extending longitudinally subparallel to latter P. speciosus Ventral extremity of cervical carina forming 313 FISHERY BULLETIN: VOL. 85, NO. 2 sharp-edged arc. Anterolateral carina ab- sent. Posterior part of hepatic sulcus fusing with branchiocardiac sulcus and continuing posteroventrally oblique to latter 2 2. Optic calathus long, mesial margin at least 1.5 width of distal extremity. Pereopods covered with minute setae P. sibogae Optic calathus short or relatively so, mesial margin 1.3 or less width of distal extremity. Pereopods not covered with setae 3 3. Thelycum with median plate of sternite XIII very long (length more than 4 times basal vddth), almost reaching spine on sternite XII, and narrow, maximum width 0.40 length. [Males unknown] P. protensus Thelycum with median plate of sternite XIII relatively short (length less than 3.5 times basal width) and broad or relatively broad, maximum width more than 0.40 length. ... 4 4. Third article of antennular peduncle expand- ed laterally, forming strong, subtriangular projection in males and variably developed rounded prominence in females. Petasma with ventral costa neither strongly inclined distomesially nor contracted preapically. Thelycum with median plate of sternite XIII moderately broad (maximum width 0.65-0.75 length) and lacking posterolateral promi- nences P. kathleenae Third article of antennular peduncle not ex- panded laterally, forming neither strong prominence in males nor rounded promi- nence in females. Petasma with ventral costa strongly inclined distomesially or contracted preapically. Thelycum with median plate of sternite XIII either relatively narrow (max- imum width 0.45-0.55 length) or, if broad (0.80-0.93), bearing posterolateral promi- nances 5 5. Males with ventral antennular flagellum sinuous proximally and bearing narrow band of densely set small setae distal to dorsal flagellum; petasma with ventral costa markedly contracted preapically. Females with median plate of sternite XIII expanded in pair of conspicuous posterolateral promi- nences P. gracilis Males with ventral antennular flagellum straight proximally and lacking band of small setae; petasma with ventral costa not markedly contracted preapically. Females with median plate of sternite XIII lacking posterolateral prominences P. crassipes Pseudaristeus kathleenae, new species Figures 1-3, 4C, 5-9 Aristaeus crassipes. Alcock 1901a:50 [part]. Hemipenaeus crassipes. De Man 1911:24. Kemp and Seymour Sewell 1912:17 [part], pi. 1, fig. 8. De Man 1913, pi. 2, fig. 4a-c and in legend to fig. 5 [under Remarks for H. sibogae]. Balss 1925:229 [part]. Materials. Holotype: o-, USNM 216710, 23.5 mm CL, 6.5 mm RL, about 88 mm TL; type-locality: Teluk Bone, Sulawesi (Celebes), Indonesia; 3°19'40"S, 120°36'30"E; 900 m; gray mud; 19 December 1909; Albatross stn 5657. Paratypes: India-1 9, USNM 216711, collected with holotype. Ict, ZSI 7806/10, W of Cape Com- orin, Tamil Nadu, India; 7°46'N, 76°37'E; 1,225 m; 26 April 1911; Investigator stn 388. 2 o-, ZSI, S of Cape Comorin, Tamil Nadu, India; 7°36'N, 78°05'E; 556-595 m; 10 April 1900; Investigator stn 268. Indonesia-1 9, ZMB, off W Sumatra; 0°39'S, 98°52'E; 750 m; 31 January 1899; Valdivia stn 191. 1 CT, ZMA, eastern Flores Sea; 7°24'00"S, 118°15'12"E; 794 m; fine gray mud, with some radiolariae and diatomes; 6 April 1899; Siboga stn 45. 19, USNM, Teluk Bone, Sulawesi (Celebes); 3°17'40"S, 120°36'45"E; 885 m; gray mud; 19 December 1909; Albatross stn 5656. 3 9, USNM, Teluk Bone, Sulawesi (Celebes); 3°32'40"S, 120°31'30"E; 933 m; gray mud; 19 December 1909; Albatross stn 5658. 1 9, USNM, W of Halmahera; 0°35'00"N, 127°14'40"N, 127°14'40"E; 795 m; fine gray sand, mud; 27 November 1909; Albatross stn 5619. Philippines— 2 9, USNM, Cagayan Is, Sulu Sea; 9°38'30"N, 121°ir00"E; 929 m; gray mud, coral sand; 31 March 1909; Albatross stn 5423. 1 ct 1 9, USNM, Lagonoy Gulf, E of southern Luzon; 13°32'30"N, 123°58'06"E; 1,033 m; gray mud; 10 June 1909; Albatross stn 5460. 1 9, USNM, Lagonoy Gulf, E of southern Luzon; 13°40'57"N, 123°57'45"E; (549 m); sand; 16 June 1909; Albatross 314 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS stn 5463. 1 9, USNM, Lagonoy Gulf, E of southern Luzon; 13°39'42"N, 123°40'39"E; 914 m; gray mud; 17 June 1909; Albatross stn 5465. 1 9, USNM, Verde Island Passage, N of Mindoro; 13°36'11"N, 120°45'26"E; 622 m; fine sand; 20 January 1908; Albatross stn 5114. 4 9, USNM, Lagonoy Gulf, E of southern Luzon; 13°35'27"N, 123°37'18"E; 878 m; gray mud; 18 June 1909; Albatross stn 5467. 2 o* 5 9, USNM, Lagonoy Gulf, E of southern Luzon; 13°36'48"N, 123°38'24"E; 914 m; green mud; 18 June 1909; Albatross stn 5469. Diagnosis.— Optic calathus relatively short, mesial margin 1.0-1.3 times distal width. Anterolateral carina lacking. Ventral extremity of cervical carina forming sharp-edged arc. Posterior extremity of hepatic sulcus turned ventrally. Third article of antennular peduncle expanded laterally, forming large subtriangular projection in males, but weak to rounded prominence in females; males with ven- tral antennular flagellum never sinuous, and ulti- mate article of third maxilliped strongly curved, spatulate and bearing patch of strong, rigid setae. Pereopods not covered with minute setae. Petasma with ventral costa only slightly inclined distomesial- ly, and ventral surface of dorsolateral lobule lack- ing setae. Thelycum with plate of sternite XIV rather long and produced at either side in short anterolateral hood; median plate of sternite XIII moderately long (not nearly reaching spine on ster- nite XII), rather broad (maximum width 0.65-0.75 length) but not expanded posterolaterally in con- spicuous prominences. Description.— Body slender (Fig. 2), densely studded with minute setae. Rostrum in males usually short, its length 0.25-0.45 CL (but in one male, 24 mm CL, 1.4 CL), and roughly lanceolate; in females long (Fig. 3), 1-1.15 CL (but in one female 23 mm CL, 0.70 CL), relatively deep and convex basally, styliform and slightly unturned anteriorly. Rostral plus epigastric teeth 3; rostral teeth situated variably in males, basally in females. Adrostral carina strong, in males almost reaching apex, in females (and in male with long rostrum) extending just anterior to second tooth. Antennal spine sharp; branchiostegal spine as long as or longer than antennal and acutely pointed. Cervical sulcus crossing postrostral carina (rarely only reaching it) at about 0.45 CL from or- bital margin, with ventral part turning anteriorly; accompanying carina blunt except for sharp strongly arched ventral extremity; weak postcervical sulcus reaching, but not crossing, postrostral carina at about 0.70 CL from orbital margin. Postrostral carina, extending 0.75-0.80 CL from orbital margin, well marked and sharp to cervical sulcus, low and blunt posteriorly, and followed by small tubercle situated near posterior margin of carapace. Antero- lateral carina lacking; gastro-orbital carina well defined; antennal carina rather short; branchios- tegal-hepatic carina long, raised and sharp. Orbito- antennal sulcus shallow; hepatic sulcus fusing with branchiocardiac sulcus, then turning obliquely almost ventrad, forming small branch nearly reach- ing margin of branchiostegite; branchiocardiac sulcus, accompanied by carina, long, extending pos- teriorly to near margin of carapace; blunt, dorsally concave ridge (disposed dorsal to posterior part of hepatic sulcus and anterior part of branchiocardiac sulcus) delimited dorsally by weak groove, latter approaching cervical sulcus anteriorly and ending about level of postcervical sulcus posteriorly. Eye (Fig. AC) with optic calathus relatively short, length of mesial margin 1.0-1.3 times distal width; mesial tubercle strong and situated between distal 0.25 and 0.30 length of margin. Gnathal appendages, except third maxilliped, illustrated in Figure 1. Antennular peduncle with stylocerite produced in sharp spine falling conspicuously short of mesial base of distolateral spine; latter small and sharp; third article in adult males uniquely produced in large subtriangular or ax-head shaped projection (Fig. 2) directed ventrolaterally, in females (Fig. 5B) expanded laterally in broadly rounded prominence. Dorsal flagellum reaching between distal 0.20 and 0.15 of scaphocerite; ventral flagellum straight and long, although incomplete in all specimens ex- amined, in one male with 19 mm CL its length at least 2.75 times CL. Scaphocerite extremely long, surpassing anten- nular peduncle by as much as 0.4 its own length; strong lateral rib ending in sharp spine falling con- siderably short of distal end of lamella. Antennal flagellum at least 1.25 times TL of shrimp. Third maxilliped sexually dimorphic. In males (Fig. 5A ) with penultimate article often slightly to strongly inflated proximally, compressed and pro- duced in strong, subelliptical or acuminate process (overhanging proximal part of ultimate article) distally; article also bearing brush of long thickly set setae on both mesial and lateral surfaces, and dense row of setae along distal margin of process. Ulti- mate article subspatulate, strongly arched, not ex- panded basally, bearing distally dense patch of strong, rigid setae on lateral surface, proximalmost setae short and more distal ones considerably longer; terminal margin with tuft of very long flex- 315 FISHERY BULLETIN: VOL. 85, NO. 2 C/2 d J a> s o -a c "53 3 oT c o H O 00 b "o I w Qi O 316 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS Figure Z.—Pseudaristeus kathleenae, 9 37 mm CL, Verde Island Passage, north of Mindoro, Philippines. Anterior region, lateral view. Scale = 10 mm. c Figure 4.— Eyes. A, Pseudaristeus crassipes, cr 28.5 mm CL, south of Cape Comorin, Tamil Nadu, India. B, P. gracilis, cr 20 mm CL, eastern Mindanao Sea, Philippines. C, P. kathleenae, holotype o* 23.5 mm CL, Teluk Bone, Sulawesi, Indonesia. D, P. protenstcs, holotype 9 40 mm CL, west of Everal Gujarat, India. E, P. sibogae, 9 47.5 mm CL, south of Pulau Muna, Sula- wesi, Indonesia. Scales: A, B, C = 1 mm; D, E = 2 mm. 317 FISHERY BULLETIN: VOL. 85, NO. 2 Figure h.—Pseudaristeus kathleenae-A, cr 26 mm CL, Lagonoy Gulf, east of southern Luzon, Philippines, distal articles of left third maxilliped, lateral view. B, Q 39 mm CL, Teluk Bone, Sulawesi, Indonesia, distal articles of right antennular peduncle and flagella, dorsal view. C, same 9, distal articles of left third maxilliped, dorsal view. Scales = 2 mm. ible setae, and entire ventral margin supporting numerous ones. In females (Fig. 5C), ultimate arti- cle slender, flattened, broadening slightly from nar- row base, then tapering gently to blunt apex. Pereopods not covered with setae; first and sec- ond with broad, compressed merus bearing small, slender, distomesial spine. Abdomen with sharp dorsomedian carina extend- ing from posterior 0.75 of fourth somite posterior- ly through sixth somite and produced in spine on caudal margin of last 3 somites; sixth also bearing pair of minute posteroventral spines and 2 elongate cicatrices. Telson with median sulcus weak, usual- ly distinct along anterior 0.75 length of telson, and flanked by paired longitudinal dorsolateral ridges; bearing 4 pairs of movable spines: 3 at about 0.60, 0.75, 0.85 length from basal margin of telson, fourth flanking short terminal part. Mesial ramus of uropod surpassing apex of telson by as much as 0.40 its own length; lateral ramus overreaching mesial ramus by as much as 0.33 its own length. Petasma (Figs. 6,1A,B) with dorsomedian lobule cincinnulate along entire mesial margin. Ventro- median lobule extending distally as far as dorso- lateral lobule and bearing elongate, lapel-like flap Figure 6.—Pseudaristeus kathleenae, cr 24.5 mm CL, Lagonoy Gulf, east of southern Luzon, Philippines. Petasma and proximal part of first pleopods, dorsal view. Scale = 2 mm. 318 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS Figure 7 .—Pseudaristeus kathleenae, holotype o", 23.5 mm CL, Teluk Bone, Sulawesi, Indonesia. A, Petasma, dorsolateral view of left half. B, Ventral view. C, Right appendices masculina and interna, and basal sclerite of endopod, dorsal view. D, Ventral view. Scales = 1 mm. distoventrally along mesial margin. Dorsolateral lobule sclerotized, expanding distolaterally before tapering to subangulate mesial apex, its distolateral margin strongly curved; ventral surface, exhibiting arched slender rib, lacking setae. Ventral costa gent- ly sinuous, only slightly inclined distomesially, its terminal part forming truncate blade lying free but against ventral surface of dorsolateral lobule. Appendix masculina (Fig. 7C, D ) roughly obovate, with proximal part curving ventrally embracing ap- pendix interna; its distal margin bearing long setae and mesial margin bearing short, more numerous ones. Appendix interna roughly triangular and sub- equal in length to appendix masculina. In males, sternite XIV with setose anteromedian tubercle; plate of sternite XIII elongate (length 2-3 times basal width), broadly rounded anteriorly and produced in minute apical spine bearing tuft of long setae. Thelycum (Fig. 8) with setose, moderately long plate of sternite XIV either traversed by shallow groove posteriorly or broadly depressed, and with anteromedian margin varying from distinctly con- cave to slightly convex, plate produced at either side in short anterolateral hood; fossa immediately ante- rior to plate very conspicuous and bearing pair of small oblique ridges. Median plate of sternite XIII, ^>jrtl^--'W-^i^iiri.<44^^^t\^ 'T^'-'^^-'-^^'^ Figure 8.—Pseudaristeus kathleenae, 9 47 mm CL, Lagonoy Gulf, east of southern Luzon, Philippines. Thelycum. Scale = 2 mm. 319 FISHERY BULLETIN: VOL. 85, NO. 2 also covered with setae, moderately long (length 3-3.5 times basal width), falling considerably short of spine on sternite XII, broadly lanceolate (max- imum width 0.65-0.75 length), and flat or slightly excavate; posterolateral margins of plate, some- times turned ventrally, abutting slender ridges ex- tending posteromesially before curving laterally on margin of sternite XIII. Sternite XII minutely setose, bearing median keel ending anteriorly in anteroventrally directed sharp spine. Maximum lengths.— Msdes, 24 mm CL; females, 46.5 mm CL. Geographic and bathymetric ranges.— From west of Cape Comorin, India, through Indonesia, and north- ward to east of Luzon, Philippines (Fig. 9). It has been obtained at depths between 549 and 1,225 m. Discussion.— This species differs from P. crassipes, with which it has been confused previously, in the following unique characteristics. In males, the third article of the antennular peduncle is strikingly pro- duced in a subtriangular or roughly ax-head shaped lateral projection, and in females it is expanded laterally in a broadly rounded prominence. This is a feature by which the females of P. kathleenae can be infallibly distinguished from those of P. crassipes which, otherwise, are quite similar. In P. crassipes, the third article of the antennular peduncle of both sexes is uniform in width proximally and gradually tapers distomesially. In males of P. kathleenae, the penultimate article of the third maxilliped is com- pressed distally and produced in a strong subellip- tical or acuminate process which overhangs the ultimate article; the latter is subspatulate, conspicu- ously curved throughout, almost uniform in width, and bears long rigid setae on the lateral surface. In the males of P. crassipes the penultimate article of the third maxilliped is subtriangular in cross section throughout and does not project distally in a con- spicuous process but instead is rounded distally; the ultimate article is twisted, expanded basally, and weakly to conspicuously so distally, and bears lateral rows of spinules on its ventral surface. Pseudaristeus kathleenae differs further from P. crassipes in features of the petasma. In the former, the dorsolateral lobule is expanded distolaterally before tapering to a subangulate mesial apex; also the ventral costa is sinuous and does not turn strong- ly distomesially from its basal part. In contrast, the dorsolateral lobule is not expanded distolaterally in P. crassipes, tapers to broadly elliptical apex, and the ventral costa is almost straight basally before turning rather abruptly distomesially. Finally, in females of P. kathleenae the median plate of ster- nite XIII is broader than that in P. crassipes, its maximum width ranging from 0.67 to 0.75 rather than from 0.45 to 0.55 as it does in the latter. In females of Pseudaristeus the rostrum is long, almost as long or considerably longer than the cara- pace, whereas, in males it is usually short, less than 0.33 the length of the carapace. Among the 8 males of P. kathleenae examined in this study, one pos- sesses a rostrum that is 1.4 times the length of the carapace (longer than that of any female examined), and among the three available males of P. crassipes, the rostrum of one, although proportionally not so long, is 0.7 as long as the carapace. Perhaps these males with long rostra are not as rare as one might anticipate. This suggestion is based on a study by Burukovsky and Romensky (1972) in which they noted considerable variation in the length of the rostrum in another aristeid, Aristeus varidens, occurring in the eastern Atlantic from Ghana to Angola. It was well known that in all genera of Aris- teidae the rostrum of females is much longer than that of males; however, they found that although the majority of males of this gamba prawn have a relatively short rostrum, in almost 30% of them it is long. Also, they noted that in both sexes the rostrum decreases proportionately with increasing size (a fact well established for penaeoid species) and that this age-dependent variation is different for the sexes: in small females the rostrum may exceed 1.5 times the length of the carapace, whereas in large ones it is about as long as the carapace; in small males the length of the rostrum may be almost 1.5 times that of the carapace, but as they grow it decreases to 0.1-0.5 times. Nevertheless, in individ- uals of the same size they found that the range of variation is smaller in females than in males. Their study demonstrates that sexual dimorphism in the length of the rostrum, thought to be typical of aris- teids, disappears at least in small males of A. varidens. This too might obtain in members of Pseudaristeus, but because of the lack or paucity of males of all 6 species I am unable to conduct a mean- ingful investigation of variations in RL/CL. Because no correlation was observed between variations in the length of the rostrum and changes in the petas- ma or in the structure of the gonads in males with short and long rostrum which would indicate sex- ual change, Burukovsky and Romensky suggested that the reduction of the rostrum, more marked in males than in females, might be associated with the transition from a benthic to a bathypelagic existence undergone by members of the family Aristeidae. 320 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS 321 FISHERY BULLETIN: VOL. 85, NO. 2 Only a few of the available specimens of this species still have entire pereopods, and even fewer have retained all 5 pereopods; in almost all, the fourth and fifth are missing, or only two or three podomeres are represented. The data obtained by me, however, seem to confirm those presented by De Man (1911) and Crosnier (1978), which indicate that the pereopods of P. kathleenae are more slender than those of P. sibogae. Their data were based on a female 31 mm CL collected by the Siboga Expedi- tion and identified by them as "crassipes". This specimen was examined during the present study and is assigned herein to P. kathleenae. Etymology.— I take pleasure in naming this shrimp for my daughter Kathleen P. Canet. Pseudaristeus crassipes (Wood-Mason 1891) Figures AA, 9-13 Aristaeus crassipes Wood-Mason 1891:281, fig. 7 [syntypes: 1 9, ZSI 6713/9, Andaman Sea; 11°25'05"N, 92°47'06"E; 405 fm (741 m); green mud; 9 December 1890; Investigator stn 116. 1 9, ZSI 3171/9 (could not be located in ZSI, 24 November 1984, Maya Deb of ZSI, pers. com- mun.2); off SW Sri Lanka; 6°29'00"N, 79°34'00"E; 597 fm (1,092 m)]. Alcock and Anderson 1894: 147. Faxon 1895:198. Anderson 1896:91. Alcock 1901a:50 [part]; 1902:268, figs. 63, 64a, b. [l]Aristaeus crassipes. Alcock 1898:74. Alcock and Anderson 1899:3. Doflein 1906:259. Aristaetis {Hemipeneus) crassipes. Alcock 1901b:33. Aristeus {Hemipeneus) crassipes. Alcock and McAr- dle 1901, pi. 49, figs. 1, 2. Aristeus crassipes. Lloyd 1907:2. Hemipeneus crassipes. Kemp and Sewell 1912:17 [part], pi. 1, fig. 9. Hemipenaeus crassipes. Balss 1925:224 [part]. Ramadan 1938:49. Anderson and Lindner 1945: 301. Ramadan 1952:15, fig. 18. Sewell 1955:203. Burukovsky 1974:48. Silas and Muthu 1979:78. Not Hemipenaeus crassipes Monod 1974:118, figs. 7-11 [= Aristeus virilis (Bate 1881) and Aristeus mabahissae Ramadan 1938; fide Crosnier 1978:85]. Pseudaristeus crassipes. Crosnier 1978:83, fig. 30d; 1986:862. Pseudaristeus sp. Crosnier 1984:22. ^Mayo Deb, Zoologist, Zoological Survey of India, Calcutta, India, pers. commun. 24 November 1984. Material. Gulf of Aden-3 9, BMNH + 19, USNM, off Djibouti; 13°06'12"N-13°03'00"N, 46°24'30"E- 46°21'42"E; 1,061 m; 7 May 1934; green mud; John Murray Exped. stn 193. 1 9, ZMB, off Yemen (Aden); 13°02'N, 46°41'W; 1,469 m; 4 April 1899; Valdivia stn 271. 1 9, BMNH, off Yemen (Aden); 13°41'N-13°40'N, 48°17'E-48°19'E; 1,295 m; 15 October 1933; John Murray Exped. stn 33. 1 9, BMNH, off Yemen (Aden); 14°36'06"N-14°38'42"N, 51°00'18"E-50°57'42"E; 1,269 m; 4 May 1934, John Murray Exped. stn 184. India— 1 9, ZMB, Arabian Sea, Investigator [stn not given]. 2 o", ZSI, S of Cape Comorin, Tamil Nadu; 7°36'N, 78°15'E, 1,017-1,088 m; green mud, sand; 10 April 1900; Investigator stn 268. 1 9, MP, GulfofMannar;8°ll'N, 79°03'E; 1,035 m; 28 July 1981; Safari II stn 4, CP 06. 1 c, ZSI, off Chidam- baran, Tamil Nidu; 11°29'45"N, 80°02'30"E; 816 m; 19 March 1901; Investigator stn 280. 1 o* 1 9, BMNH, S of Andaman Is, 10°06'N,92°29'E; 1,239 m; Investigator stn 315. Sri Lanka-2 9, ZSI, NW of Colombo; 6°54'30"N, 79°34'30"E; 878 m; green mud; 20 October 1898, Investigator stn 250. Indonesia— 1 9, MP, Strait of Makassar; 2° 04' 24"S, 118°46'54"E; 1,710-1,730 m; 9 November 1980; Corindon II stn 286. Diagnosis.— Optic calathus relatively short, mesial margin 0.9-1.1 times distal width. Anterolateral carina lacking. Ventral extremity of cervical carina forming sharp-edged arc. Posterior extremity of hepatic sulcus turned ventrally. Third article of antennular peduncle not expanded laterally; males with ventral antennular flagellum not sinuous and ultimate article of third maxilliped twisted, marked- ly dilated proximally, less so distally, and bearing ventrolateral rows of minute spines. Pereopods not covered with minute setae. Petasma with distal two- thirds of ventral costa turned rather abruptly disto- mesially and ventral surface of dorsolateral lobule lacking setae. Thelycum with plate of sternite XIV long and produced in short anterolateral hoods; me- dian plate of sternite XIII moderately long (not near- ly reaching spine on sternite XII), rather narrow (maximum width 0.45-0.55 length), and not ex- panded posterolaterally in conspicuous prominences. Description.— Body slender, densely studded with minute setae. Rostrum in males (complete only in two) relatively short, its length 0.2 and 0.7 CL, and tapering gradually to sharp apex; in females longer. 322 PEREZ FARFANTE; REVISION OF GENUS PSEUDARISTEUS 1.15-1.25 CL, moderately deep and convex basally, styliform and slightly upturned anteriorly, but occa- sionally with apical extremity curving downward. Rostral plus epigastric teeth 3; rostral teeth, situ- ated variably in males, basally in females. Adrostral carina strong, in males with short rostrum almost reaching apex, in females and in male with long rostrum extending just anterior to second tooth. Antennal spine sharp; branchiostegal spine longer than antennal, acutely pointed. Cervical sulcus crossing postrostal carina (rarely only reaching it) at about 0.45 CL from orbital margin, with ventral part turning anteriorly; accompanying carina blunt, except for sharp, strongly arched ventral extrem- ity; postcervical sulcus reaching, but not crossing, postrostral carina at about 0.7 CL from orbital margin. Postrostral carina, extending 0.8-0.9 CL from orbital margin, well marked and sharp to cer- vical sulcus, low and blunt posteriorly, followed by small tubercle situated near posterior margin of carapace. Anterolateral carina lacking; gastro- orbital carina strong; antennal carina relatively short; branchiostegal-hepatic carina long, raised and sharp. Orbito-antennal sulcus shallow; deep hepatic sulcus fusing with branchiocardiac sulcus before turning obliquely almost ventrad, forming small branch nearly reaching margin of branchiostegite; branchiocardiac sulcus, accompanied by strong carina, deep and long, extending posteriorly to near margin of carapace; blunt, dorsally concave ridge (disposed dorsal to posterior part of hepatic sulcus and anterior part of branchiocardiac sulcus) delimited dorsally by shallow groove, latter extend- ing anterodorsally almost to cervical sulcus and con- tinuous posteriorly with postcervical sulcus. Eye (Fig. 4A ) with optic calathus relatively short, length of mesial margin 0.9-1.1 times distal width; mesial tubercle strong and situated between 0.30 and 0.45 length of mesial margin from cornea. Antennular peduncle with stylocerite produced in sharply pointed spine falling distinctly short of, to almost reaching, mesial base of distolateral spine; latter slender and sharp. Third article never pro- duced laterally (Fig. IOC); dorsal flagellum reach- c /< Figure 10.— PseudaristeiLS crassipes (Wood-Mason): A, cr 26.5 mm CL, south of Cape Comorin, India, distal articles of right third maxilliped, ventral view. B, Same cr distal articles of right third maxilliped, dorsal view. C, Lectotype 9, Andaman Sea, India, distal articles of right antennular peduncle and flagella, dorsal view (prepared from camera lucida drawings by H. C. Ghosh). D, 9 25 mm CL, south of Andaman Islands, India, distal articles of left third maxilliped, ventral view. Scales: A, B, D = 1 mm, C = A mm. 323 FISHERY BULLETIN: VOL. 85, NO. 2 ing between base of distal 0.25 and end of scapho- cerite; ventral flagellum straight and long, although broken in all specimens studied, in male with 22 mm CL its length 2.1 times CL. Scaphocerite extremely long, surpassing anten- nular peduncle by as much as 0.4 its own length; strong lateral rib ending in sharp spine falling considerably short of distal end of lamella. An- tennal flagellum incomplete in all specimens ex- amined. Third maxilliped sexually dimorphic: Males (Fig. lOA, B) with penultimate article subtriangular in cross section, its distomesial margin not produced in conspicuous process but bluntly rounded; ultimate article twisted, expanding from short, narrow base, then narrowing and becoming concave laterally before again expanding (sometimes almost imper- ceptibly) distally; ventral surface with proximal, transverse comb of long setae continuous with lateral rows of minute spines extending to and around distal extremity. Females with ultimate ar- ticle (Fig. lOD), slender, but broadening slightly mesially from short, narrow base, then tapering to blunt apex. Pereopods not covered with setae; first and sec- ond pereopods with broad depressed merus bearing small, slender, distomesial spine. Abdomen with sharp dorsomedian carina extend- ing along posterior 0.75 of fourth somite through sixth, and produced in spine on posterior margin of last 3 somites; sixth somite also bearing pair of minute posteroventral spines and 2 elongate cica- trices. Telson with median sulcus shallow, usually distinct only along anterior 0.25 length of telson, and flanked by pair of longitudinal dorsolateral ridges; bearing 4 pairs of movable spines: 3 situated at about 0.60, 0.75, 0.85 length from basal margin of telson, fourth flanking short terminal part. Mesial ramus of uropod surpassing apex of telson by as much as 0.40 its own length; lateral ramus over- reaching mesial ramus by as much as 0.33 its length. Petasma (Fig. 1L4, jB) with dorsomedian lobule cincinnulate along entire mesial margin. Ventro- median lobule extending distally almost as far as dorsolateral lobule and bearing elongate, lapel-like flap distoventrally along mesial margin. Dorsolateral lobule sclerotized, not expanded distolaterally, with lateral margin only slightly curved and distalmost part broadly rounded forming subelliptical mesial apex; ventral surface lacking setae, exhibiting con- spicuous, slender, arched rib. Ventral costa almost straight basally along 0.4 of its length, then turn- ing somewhat abruptly distomesially, its rather broad terminal part truncate (sometimes vdth disto- FlGURE II.— Psevdaristeiis crassipes, cr 26.5 mm CL, south of Cape Comorin, India. A, Petasma, dorsal view of left half. B, Ventral view (specimen slightly distended). Scale = 1 mm. 324 PEREZ FARFANTE; REVISION OF GENUS PSEUDARISTEUS mesial angle slightly produced) lying free but against ventral surface of distolateral lobule. Appendices masculina and interna like those of P. kathleenae. In males, plate of sternite XIII flat, ovate but produced in minute apical spine, its length 1.8-1.9 basal width; sternite XIV bearing antero- median tubercle or low, anteriorly produced prom- inence. Thelycum (Fig. 12A, 5) with setose plate of ster- nite XIV long, transversely depressed (occasional- ly with median elevation), anteriomedian margin straight or slightly convex, plate produced at either side in short anterolateral hood; fossa immediately anterior to plate short and bearing pair of small, oblique, ridges. Median plate of sternite XIII, also covered with setae, moderately long (length 2.9-3.5 basal width) but falling distinctly short of spine on sternite XII, rather narrow (maximum width 0.45- 0.55 length), lanceolate, tapering anteriorly from near base, occasionally from near midlength, and strongly produced in sharp apical spine; postero- lateral margins of plate, usually turned ventrally, flanked by, or interlocking with, slender ridges curving laterally on margin of sternite XIII. Ster- nite XII minutely setose, strongly raised and crested by median carina ending anteriorly in minute spine. Color in life crimson (Wood-Mason 1891). Maximum lengths. - mm CL. -Males, 29 mm CL; females 37 Geographic and bathymetric ranges.— From the Gulf of Aden to the Strait of Makassar, Indonesia (Fig. 9). It has been found at depths between 741 and 1,710-1,730 m. Variation.— This species exhibits marked variation in shape of the last 2 articles of the third maxilliped in armature of sternite XIV of males, and also in the shape of the thelycal plate of sternite XIII in females. Those articles range from moderate to very narrow widths and both the proximal and distal parts of the ultimate article may be imperceptibly to conspicuously dilated. Sternite XIV may bear either an anteromedian tubercle or a low, anterior- ly produced prominence. In the few males examined, the presence of a very slender ultimate article of the third maxilliped seems to be correlated with the presence of a prominence, instead of a tubercle, on sternite XIV. Furthermore, in females the thelycal plate of sternite XIII, although always lanceolate, may be broadest near the base, (in most specimens), or as far anteriorly as midlength. The female from the Strait of Makassar, Indo- nesia, cited in "Material", was made available to me -: -■'■■- ■ r-^ ' ■'- - f \ ■/■■y:- y^^^l^vv^^^^ B Figure 12.— Psevdaristeus crassipes: A, 9 35 mm CL. off Djibouti, Gulf of Aden. B. 9 3.5 mm CL, Gulf of Mannar, India. Thelyca. Scale = 2 mm. 325 FISHERY BULLETIN: VOL. 85, NO. 2 through the kindness of A. Crosnier, who discussed this shrimp in his work of 1984. He pointed out features that he beheved would distinguish it from P. crassipes and P. sibogae: the absence of setae from the integument; a narrower median plate on sternite XIII (Fig. 13); the absence of setae on the pereopods [ typically present in P. sibogae but lack- ing in P. crassipes]; and a robust optic calathus which resembles that of P. crassipes. The Indonesian specimen definitely does not belong to P. sibogae, but its relation to P. crassipes is not entirely clear. A few specimens of the latter species are glabrous, a condition, as noted by Crosnier, unlikely to have been attained acciden- tally, but absence of setae is not typical of any species of the genus. Variations in the length/width ratio of the optic calathus of P. crassipes embrace that of the Indonesian specimen, the mesial margin length of which is equal to the distal width. The max- imum width of the median thelycal plate of sternite XIII in most females of P. crassipes ranges from 0.50 to 0.55 its length (in one setose specimen, Figure 12B, however, it is only 0.47), whereas the maximum width, 0.45, falls below this range in the Indonesian female. The latter exhibits on the plate of sternite XIV a median ridge that ends in a minute anterior spine, the plate is not produced at either side in an anterolateral hood, and the contour of the median plate on sternite XIII is almost uniformly broad from the base to about midlength. These features differ from those of typical P. crassipes females in which the plate of sternite XIV is un- ornamented, a well-developed anterolateral hood is produced at either side, and the contour of the me- dian plate on sternite XIII broadens from a narrow base posterior to its midlength, then tapers to its apex (cf. Figs. 12, 13). Additional material from the Indonesian locality, including males, might provide evidence for assigning this form to a new taxon. Discussion.— The males of P. crassipes differ strikingly from those of its congeners in that the ultimate article of the third maxilliped is twisted (forming a strong concavity laterally), conspicuously expanded proximally, weakly to markedly dilated distally, and studded with minute spines ventro- laterally. The petasma differs from that of P. kathleenae and P. gracilis, the other 2 species for which adult males are known; the dorsolateral lob- ule is not expanded distolaterally, as it is in the former, and it tapers gently to a broadly obtuse apex instead of narrowing rapidly as it does in P. gracilis. Furthermore, the ventral costa turns somewhat abruptly distomesiad rather than forming a gentle, Figure \2,.—Pseudaristeus crassipes, 9 35 mm CL, Strait of Makassar, Indonesia. Thelycum. Scale = 2 mm. sinuous curve or simple arc, and, in contrast to the costa of P. gracilis, its terminal part, which is also truncate as it is sometimes in the latter species, is not set off by a conspicuous constriction. The females of P. crassipes can be distinguished readily from those of P. kathleenae by the shape of the third article of the antennular peduncle, which in the former is uniform in width proximally and tapers distomesially, but in the latter is expanded laterally in a broadly rounded prominence. Also, although the thelyca of these 2 species exhibit a marked resemblance, the median plate of sternite XIII is narrower in P. crassipes than that in P. kathleenae, its maximum width ranging from 0.45 to 0.55 instead of 0.67 to 0.75. Remarks.— In the original description of Aristeus crassipes, Wood-Mason (1891) cited two females, one from Investigator station 116 and another taken at lat. 6°29'N, long. 79°34'E [off southeastern Sri Lanka], perhaps collected by the Investigator. He did not designate either specimen as the holotype and consequently they must be considered syntypes. Later, Alcock (1901b) recorded the registration numbers of the various lots of specimens of this 326 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS species in the Zoological Survey of India, and next to the number 6713/9, corresponding to the first female cited by Wood-Mason, he added "Types of the species". Immediately following it is the regis- tration number, 3175/9, of the second female (Maya Deb fn. 2). Alcock's referral to "Types" indicate that he did not intend to select the first female as a lecto- type; moreover, since he did not specifically desig- nate it as such, according to the Article 74a of the International Code of Zoological Nomenclature of 1961 it cannot be so considered. Maya Deb has also informed me that the female from Investigator sta- tion 116 is in the Zoological Survey of India accom- panied by the registration number cited above, but that the other female recorded by Wood-Mason could not be located. Because two other species, P. kathleenae and P. protensus, have been confused with A. crassipes and conceivably because the miss- ing syntype might prove to be conspecific with one of them, I hereby designate the female in the Zoo- logical Survey of India assigned registration num- ber 6713/9 as the lectotype oi Aristaeus crassipes Wood-Mason, 1891. Through the kindness of Maya Deb and H. C. Ghosh, both of the ZSI, who examined the lectotype and provided me with information on specific mor- phological features, clear drawings, and photo- graphs, I have been able to ascertain the identity of P. crassipes. As stated above, the characters used by Wood-Mason in the description of P. crassipes were inadequate for distinguishing it from 2 other species and this deficiency was no doubt responsible for the assignments of closely allied forms to P. crassipes (De Man 1911, 1913; Kemp and Sewell 1912; Balss 1925). Kemp and Sewell, however, noted that among males of P. crassipes in the ZSI there were two types of third maxillipeds which they described and illustrated. They stated further that it was possible that the males exhibiting one of the maxilliped types should be recognized as belonging to a new variety. The "new variety" is described herein as P. kathleenae. Pseudaristeus protensus, new species Figures 4D, 9, 14 Material. Holotype: 9, USNM 42681, 40 mm CL, length of median plate of sternite XIII 7.1 mm, basal width 1.5 mm; type-locality W of Everal Gujarat, India (Arabian Sea); 19°51'30"N, 69°07'30"E; 1,569 m; sand and mud; 14 April 1906; Investigator stn 370. Paratype: 9, MP, off Godavar, India; 869 m; Investigator. Diagnosis.— Optic calathus relatively short, mesial margin length equal distal width. Anterolateral carina lacking. Ventral extremity of cervical carina forming sharp-edged arc. Posterior extremity of hepatic sulcus turned ventrally. Third article of antennular peduncle in females not expanded later- ally. Pereopods not covered with minute setae. Petasma unknown. Thelycum with plate of sternite XIV very short and produced in long anterolateral hoods; median plate of sternite XIII very long (almost reaching spine on sternite XII), narrow (maximum width 0.40 length), and not expanded posterolaterally in conspicuous prominences. Description.— Body of holotype and paratype (only two specimens available) slender, studded with minute setiferous punctations and extremely minute setae. Rostrum broken. Anntenal spine broken; branchiostegal spine long, slender, and acutely pointed. Cervical sulcus crossing or just reaching postrostral carina at about 0.45 CL from orbital margin, with ventral part turning anteriorly; accom- panying carina blunt except for sharp, strongly arched ventral extremity; postcervical sulcus deep, almost reaching, but not crossing, postrostral carina at about 0.70 CL from orbital margin, and con- siderably extending anteriorly. Postrostral carina, extending to 0.85 CL from orbital margin, well marked and sharp to cervical sulcus, low and blunt posteriorly, and followed by small tubercle situated near posterior margin of carapace. Anterolateral carina lacking; gastro-orbital carina strong; anten- nal carina relatively short; branchiostegal-hepatic carina long, raised and sharp. Orbito-antennal sulcus shallow; deep hepatic sulcus fusing with branchio- cardiac sulcus, where turning obliquely almost ventrally forming small branch nearly reaching branchiostegite; branchiocardiac sulcus, accom- panied by sharp carina, deep and long, extending posteriorly to near margin of carapace; strong, arched ridge (disposal dorsal to posterior part of hepatic sulcus and anterior part of branchiocardiac sulcus) delimited dorsally by deep groove, latter con- tinuous posteriorly with postcervical sulcus but not extending anteriorly to cervical sulcus. Eye (Fig. AD) with optic calathus relatively short, length of mesial margin equal width of distal ex- tremity; mesial tubercle strong and situated at distal 0.33 length of mesial margin. Antennular peduncle with stylocerite produced in sharply pointed spine almost reaching or falling 327 FISHERY BULLETIN: VOL. 85. NO. 2 short of mesial base of well-developed, sharp, disto- lateral spine; third article not produced laterally; dorsal flagellum extending to distal 0.2 of scapho- cerite; ventral flagellum although incomplete, long, and straight, not mesially curved (concave) just distal to apex of dorsal flagellum. Scaphocerite extremely long, surpassing anten- nular peduncle by as much as 0.4 its own length, strong lateral rib ending in acutely pointed small spine falling considerably short of distal end of lamella. Antennal flagellum broken. Third maxilliped with ultimate article slender but slightly broadening mesially from narrow base, then tapering gently to blunt apex. Pereopods not covered with setae; first and sec- ond pereopods with broad, depressed merus armed with small, slender, distomesial spine. Abdomen with sharp dorsomedian carina extend- ing full length of fourth somite through sixth, and produced in spine on posterior margin of last 3 somites; sixth somite also bearing pair of minute posteroventral spines and 2 elongate cicatrices. Telson with median sulcus shallow anteriorly, in- distinct posteriorly, and flanked by paired longitu- dinal dorsolateral ridges (posterior part of telson lacking in types). Lateral ramus of uropod surpass- ing mesial ramus by about 0.3 its own length. Thelycum (Fig. 14) with setose plate of sternite XIV very short, deeply excavate transversely, bear- ing small anteromedian notch, and produced at either side in elongate anterolateral hood; fossa im- mediately anterior to plate very short and armed with pair of small, oblique lateral ridges. Median plate of sternite XIII very long (length 4.5-4.9 times basal width), narrowly lanceolate (maximum width 0.4 length), strongly produced in sharp apical spine, almost reaching anteromedian spine on sternite XII and covered by very thickly set setae; posterolateral margins of plate raised in slender ridges merging with similar ones extending posteromesially before curving laterally following margin of sternite XIII. Sternite XII minutely setose, strongly raised and bearing low median carina ending anteriorly in minute spine. Geographic and bathymetric ranges.— Known only from the type-locality, located in the Arabian Sea, and from off Godavari, in the Bay of Bengal, at depths of 1,569 and 869 m respectively (Fig. 9). Discission.— Like P. crassipes, P. kathleenae, and P. gracillis but unlike P. sibogae, the anterodorsal extremity of the groove dorsal to the posterior part of the hepatic sulcus does not join the cervical sulcus in P. protensus; the optic calathus is relatively short, and the pereopods are not covered by setae. Pseuda- risteus protensus differs strikingly from all its cogeners in several distinctive thelycal features: the plate of sternite XIV is short, bears a small median notch on the anterior margin, and is produced in long anterolateral hoods; the median plate of ster- nite XIII is very long (4.5-4.9 times the basal width, rather than 1.8-3.5), almost reaching the antero- median spine on sternite XII, and narrower (max- imum width 0.40 instead of 0.45-0.75) than its length. Moreover, it is very densely setose (more so than in its cogeners) and bears a pair of very conspicuous, vertically directed posterolateral ridges. Although I have examined only two specimens of P. protensus, distinct thelycal differences between this gamba prawn and those of other members of the genus leave no doubt that it represents a new species. Etymology.— hsitm protensus, stretched forth, refer- ring to the unusual length of the thelycal plate of sternite XIII. Figure 14. —Pseudaristeus protensus, n. sp., holotype 9 40 mm CL, west of Everal Gujarat, India. Thelycum. Scale = 2 mm. 328 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS Pseudaristeus gracilis (Bate 1 888) Figures 4fi, 9, 15-17 Hemipenaeus gracilis Bate 1888:302, pi. 44, fig. 2 [syntypes 3 0-3 9, BMNH; type-locality: off Tablas I, Philippines; 12°21'N, 122°15'E; 1,240 m; blue mud; 16 January 1975; Challenger stn 207]. De Man 1911:26. Estampador 1937:493. Anderson and Lindner 1945:301. Burukovsky 1974:48. Hemipeneus gracilis. Faxon 1895:198. Pseudaristeus gracilis. Crosnier 1978:76, fig. 27 bis. 30e. Material. Philippines— 8 9, USNM, Iligan Bay, northern Mindanao; 8°15'20"N, 123°57'E; 750 m; gray mud, sand; 7 August 1909; Albatross stn 5511. 1 9, USNM, Iligan Bay, northern Mindanao; 8°34'48"N, 124°01'24"E; about 1,280 m; 8 August 1909; Alba- tross stn 5515. 1 o", USNM, Macajalar Bay, north- ern Mindanao; 8°41'30"N, 124°35'40"E; 1,013 m; green mud, fine sand; 4 August 1909; Albatross stn 5499. 1 o- 2 9, USNM, eastern Mindanao Sea; 9°06'30"N, 125°00'20"E; 1,785 m; gray mud; 2 August 1909; Albatross stn 5495. 1 9, USNM, N of Siquijor I; 9°12'45"N, 123°45'30"E; 1,472 m; green mud, globigerina; 11 August 1909; Albatross stn 5526. 19, USNM, eastern Mindanao Sea; 9°12'45"N, 125°20'E; 1,344 m; gray mud; 1 August 1909; Albatross stn 5492. 1 9, USNM, between Bohol and Siquijor Is; 9°22'30"N, 123°42'40"E; 719 m; globigerina ooze; 11 August 1909; Albatross stn 5527. 2 0-19, USNM, eastern Mindanao Sea; 9°24'N, 125°12'E; 1,346 m; green mud, coral; 1 August 1909; Albatross stn 5491. 19, USNM, off Panaon Is, S of Leyte; 9°58'00"N, 125°07'40"E; 1,417 m; green mud; 10 April 1908; Albatross stn 5203. 2 9, USNM, Sogod Bay, southern Leyte; 10°N, 125°06'45"E; 1,412 m; green mud; 31 July 1909; Albatross stn 5488. 1 c, USNM, Sogod Bay, southern Leyte; 10°02'45"N, 125°05'33"E; 1,339 m; green mud; 31 July 1909; Albatross stn 5487. 1 o- 1 9, ZSI, Sogod Bay, southern Leyte; 10°10'00"N, 125°04'15"E; 1,013 m; gray sand, mud; 10 April 1908; Albatross stn 5201. 2 9, MP, SW of Tablas I; 12°09'N, 122°14'E; 1,404 m; 6 June 1985; MUSORSTOM III, stn CP 136. 3 o- 3 9 syntypes. 1 9, MP, SE of Bondoc Point, Luzon; 13°02'08"N, 122°37.1'E; 1,030-1,190 m; 25 November 1980; MUSORSTOM II stn 39. 3 o- 3 9, MP, NE of Bon- doc Point, Luzon; 13°23.2'N, 122°20.7'E; 820-760 m; 26 November 1980; MUSORSTOM II stn 44. Diagnosis.— Optic calathus relatively short, mesial margin 1.0-1.3 times distal width. Anterolateral carina lacking. Ventral extremity of cervical carina forming sharp-edged arc. Posterior extremity of hepatic sulcus turned ventrally. Third article of antennular peduncle not expanded laterally; males with ventral antennular flagellum sinuous and ulti- mate article of third maxilliped straight and slight- ly broadening proximomesially before tapering to apex. Pereopods not covered with minute setae. Petasma with distalmost part of dorsolateral lobule narrowing to subangular apex, and ventral surface studded with minute setae; ventral costa slightly in- clined distomesially and contracted just proximal to spatulate or paddlelike terminal process. Thelycum with plate of sternite XIV short and produced in moderately long anterolateral hoods; median plate of sternite XIII relatively short (not nearly reach- ing spine on sternite XII), broad (maximum width 0.80-0.93 length), thickened and expanded postero- laterally in conspicuous prominences. Description.— Body slender, densely studded with minute setae. Rostrum in males short, its length 0.25-0.30 CL and tapering gradually to sharp apex; in females long, 0.90-1.50 CL, relatively deep and convex basally, styliform and slightly upturned ante- riorly. Rostral plus epigastric teeth 3; rostral teeth situated variably in males, basally in females. Adros- tral carina strong, in males almost reaching apex of rostrum, in females extending just anterior to second tooth. Antennal spine sharp; branchiostegal spine longer than antennal, acutely pointed. Cervical sulcus crossing postrostral carina (rarely only reach- ing it) at about 0.45 CL from orbital margin, ven- tral part turning anteriorly; accompanying carina blunt, except for sharp, strongly arched ventral ex- tremity; postcervical sulcus reaching, but not cross- ing, postrostral carina at about 0.7 CL from orbital margin. Postrostral carina, extending to 0.8-0.9 CL from orbital margin, well marked and sharp to cer- vical sulcus, low and blunt posteriorly, and followed by small tubercle situated near posterior margin of carapace. Anterolateral carina lacking; gastro- orbital carina strong; antennal carina relatively short; branchiostegal-hepatic carina long, raised and sharp. Orbito-antennal sulcus shallow; deep hepatic sulcus fusing with branchiocardiac sulcus before turning obliquely almost ventrad forming small branch nearly reaching margin of branchiostegite; branchiocardiac sulcus, accompanied by carina, deep and long, extending posteriorly to near margin of carapace; blunt arched ridge, disposed dorsal to posterior part of hepatic sulcus and anterior part 329 FISHERY BULLETIN: VOL. 85, NO. 2 of branchiocardiac sulcus, delimited dorsally by very shallow, sometimes indistinct, groove. Eye (Fig. 4B) with optic calathus relatively short, length of mesial margin 1.0-1.3 times distal width; mesial tubercle strong and variably situated between 0.15 and 0.35 length of mesial margin from base of cornea. Antennular peduncle with stylocerite produced in sharply pointed slender spine falling distinctly short of, or almost reaching, mesial base of distolateral spine; latter acutely pointed. Third article never pro- duced laterally (Fig. 15A); dorsal flagellum about 0.4 length of antennular peduncle, reaching between distal 0.2 and terminal margin of scaphocerite; ven- tral flagellum long (although incomplete in all specimens examined, in one male 20 mm CL it length 3 times CL), uniquely sinuous, slightly broadened just distal to apex of dorsal flagellum and bearing narrow band of densely set small setae on mesial margin of broadened part (Fig. 15^). Scaphocerite extremely long, surpassing anten- nular peduncle by as much as 0.4 its own length; strong lateral rib ending in sharp spine falling con- siderably short of distal end of lamella. Antennal flagellum incomplete in all specimens examined. Third maxilliped in males (Fig. 155) with penulti- mate article subtriangular in cross section, and not produced in distal process; ultimate article slender, slightly broadening mesially from short, narrow base before tapering to apex (sometimes slightly dilated proximal to tip); in females (Fig. 15C) pen- ultimate article more slender than in males, ultimate article broadening slightly from narrow base then tapering to blunt apex. Pereopods not covered with setae; first and sec- ond pereopods with broad, compressed merus bear- ing small, slender, distomesial spine. Abdomen with dorsomedian carina extending from fourth through sixth somite, carina low on fourth, sharp and somewhat higher posteriorly, and produced in small spine on caudal margin of each somite; sixth also bearing pair of minute postero- A B C ' M m Figure 15.— Pseudaristeus gracilis (Bate): A,cr20 mm CL, eastern Mindanao Sea, Philippines, last article of right antennular peduncle and flagella, dorsal view. B, same o", distal articles of left third maxilliped, dorsolateral view. C, 9 35.5 mm CL, between Bohoi and Siquijor Islands, Philippines, distal articles of left third maxilliped, dorsal view. Scale = 2 mm. 830 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS lateral spines and 2 elongate cicatrices. Telson with median sulcus well defined only on anterior 0.33 length of telson and flanked by paired longitudinal, slender ridges reaching base of third of 4 pairs of movable, marginal spines: 3 situated at about 0.65, 0.80, and 0.90 length from basal margin of telson, fourth flanking short terminal part. Mesial ramus of uropod surpassing apex of telson by about 0.40 its own length; lateral ramus overreaching mesial ramus by about 0.33 its own length. Petasma in adults (Fig. 16/1, 5) vdth dorsomedian lobule cincinnulate along entire mesial margin. Ven- tromedian lobule, extending distally as far as dor- solateral lobule, and bearing elongate, lapel-like flap distoventrally along mesial margin. Dorsolateral lobule sclerotized, broad proximally to about base of distal 0.25, then tapering to subacute mesial apex, ventrally bearing conspicuous longitudinal, arched rib and studded with setae. Ventral costa broadly curved along almost entire length, inclined disto- FlGURE l^.—Pseudaristeus gracilis, Petasmata: A, cr20 mm CL, eastern Mindanao Sea, Philippines, dorsal view of left half. B, Ventral view. C, Syntype o" 13 mm CL, Tablas Island, Philippines, dorsal view of left half. D, Ventral view. Scales: A, B = 1 mm; C, D, = 0.5 mm. 331 FISHERY BULLETIN: VOL. 85, NO. 2 mesially and with distal part, lying free but against ventral surface of dorsolateral lobule, markedly con- tracted just proximal to spatulate or paddlelike terminal process. Petasma in juveniles lacking cincinnuli, specific characters seemingly absent (Fig. 16C, D): dorso- median and ventromedian lobules not completely differentiated, but narrow, lapel-like flap present along ventromesial margin; dorsolateral lobule broad proximally, tapering gradually to rounded mesial tip, and with distolateral margin gently curved. Ventral costa arched throughout its length, narrower and contracted distally. Appendixes masculina and interna as in P. kathleenae. In males, plate of sternite XIV often bearing in- conspicuous anteromedian tubercle; plate of sternite XIII flat, roughly lanceolate, produced in sharp spine, its length 1.5-2.3 basal width. Thelycum (Fig. 17) with densely setose plate of sternite XIV short, deeply grooved transversely, its sharp anteromedian margin turned ventrally, plate produced at either side in moderately long, antero- lateral hood; fossa preceding plate long, deep, and bearing pair of small, obliquely disposed ridges. Me- dian plate of sternite XIII, also covered with thick- ly set setae, concave, and produced apically in acute spine; relatively short (length 2.0-2.7 basal width), falling considerably short of spine on sternite XII, broad (maximum width 0.80-0.93 length), and uniquely expanded in strong posterolateral promi- nences continuous with slender ridges extending into fossa of sternite XIV. Sternite XII minutely setose, strongly keeled and crested by median carina ending anteriorly in slender, anteroventrally directed spine. The morphological account above is the first to in- clude adult features. This gamba prawn was named by Bate (1888) on the basis of 6 small juveniles, and the characters pointed out by him have proven in- adequate to recognize the species. Subsequent cita- tions to P. gracilis have been based only on Bate's information. The material available to me have allowed the detailed descriptions of the petasma, thelycum, and the ventral flagellum of the male which is unique among the members of Pseudaris- teus. Maximum lengths.— Males, 21 mm CL; females, 35 mm CL. Geographic and bathymetric ranges.— Known only from waters of the Philippines (Fig. 9). It has been taken at depths between 719 and 1,785 m. Figure n .—Pseudaristeus gracilis, 9 35.5 mm CL, between Bohol and Siquijor Islands, Philippines. Thelycum. Scale = 2 mm. Discussion.— Ferhaps the most conspicuous differ- ence between the males of P. gracilis and those of the other species of the genus of which males are known is in the ventral antennular flagellum, which is sinuous and bears a dense band of closely set setae on the mesial margin of the slightly broadened part. Males also may be distinguished from those of their cogeners by the shape of the dorsolateral lobule of the petasma, which tapers rapidly to the subangular apex and exhibits a setose ventral surface, and also by the ventral costa, which is markedly contracted distally. Characteristic of the females are the short plate of sternite XIV, in which the anteromedian margin is turned ventrally, and the unique strong posterolateral prominences of the median plate of sternite XIII. The prominences are clearly defined in juveniles as small as the syntype with a 10.3 mm CL. Pseudaristeus sibogae De Man 1911 Figures 4£, 9, 18 Hemipenaeus sibogae De Man 1911:25 [9 holotype, ZMA De. 102.462, E Savu Sea, Indonesia; 9°03'24"S, 119°56'42"E; 1,000 m; globigerina; 20 332 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS April 1899; Siboga Exped. stn 52]. De Man 1913, pi. 2, fig. 5, 5a-c. Ramadan 1938:48. Anderson and Lindner 1945:301. Burukovsky 1974:48. Pseudaristeus sibogae. Crosnier 1978:83, figs. 27a, 30a-c; 1984:22. De Freitas 1985:12, fig. II-5, A-H. Material. Madagascar— 1 9, MP, NW of Ankazomanga; 21°26'30"S, 43°11'00"E; 810-1,020 m; 26 November 1973; Vauban stn 92. 1 o- 1 9, MP, SW of Bale des Assassins; 22°16'48"S, 42°56'00"E; 1,200 m; 30 November 1973; Vauban stn 109. Indonesia— 9 holotype. 1 9, USNM, S of Pulau Muna, Sulawesi (Celebes); 5°31'30"S, 122°22'40"E; 834 m; green mud; 16 December 1909; Albatross stn 5646. 10-2 9, USNM, Selat Butung, Sulawesi (Celebes); 5°34'00"S, 122°18'15"E; 950 m; green mud; 16 December 1909; Albatross stn 5647. 1 9, USNM, off southern Buru; 3°47'15"S, 126°23'40"E; 946 m; fine gray sand; 10 December 1909; Albatross stn 5638. Diagnosis.— OTptic calathus long, mesial margin 1.5-1.7 times distal width. Anterolateral carina lack- ing. Posterior extremity of hepatic sulcus turned ventrally. Third article of antennular peduncle not expanded laterally; males with ventral antennular flagellum never sinuous; ultimate article of third maxilliped straight and slightly broadening proximo- mesially before tapering to apex. Pereopods covered with minute setae. Adult petasma unknown. Thely- cum with plate of sternite XIV moderately long and produced in short anterolateral hoods; median plate of sternite XIII relatively short (not nearly reach- ing spine on sternite XII) and broad (maximum width 0.60-0.70 length) but not expanded postero- laterally in conspicuous prominences. Description.— Body slender, densely studded with minute setae. Rostrum in males straight, moderately long, 0.48 and 0.57 CL in 2 specimens, 21 and 22 mm CL, respectively, and roughly lanceolate; in females longer, in one 37 mm CL its length 1.07 CL, rather deep and usually convex, occasionally almost straight basally, styliform and moderately upturned anteriorly. Rostral plus epigastric teeth 3; 2 rostral teeth in males situated at 0.1-0.2 and 0.4 RL respec- tively, basally in females. Adrostral carina strong, in both males and females extending just anterior to second tooth. Antennal spine sharp; branchios- tegal spine longer than antennal, acutely pointed. Cervical sulcus crossing postrostral carina at about 0.45 CL from orbital margin, with ventral part turn- ing anteriorly; accompanying carina blunt, except for sharp, arched ventral extremity; rather weak postcervical sulcus reaching, but not crossing, post- rostral carina at 0.7-0.8 CL from orbital margin. Postrostral carina, extending 0.8-0.9 CL from or- bital margin, well marked and sharp to cervical sulcus, low and blunt posteriorly, and followed by small tubercle situated near posterior margin of carapace. Anterolateral carina lacking; gastro- orbital carina strong; antennal carina relatively short; branchiostegal-hepatic carina long, raised and sharp. Orbito-antennal sulcus shallow; deep hepatic sulcus fusing with branchiocardiac sulcus before turning obliquely almost ventrad, forming small branch nearly reaching margin of branchiostegite; branchiocardiac sulcus, accompanied by strong carina, deep and long, extending posteriorly to near margin of carapace; blunt, dorsally concave ridge (disposed dorsal to posterior part of hepatic sulcus and anterior part of branchiocardiac sulcus) de- limited dorsally by groove, latter deep and abutting cervical sulcus anterodorsally but becoming shallow posteriorly and indistinct close to postcervical sulcus. Eye (Fig. AE) with optic calathus long, length of mesial margin 1.50-1.75 times distal width; mesial tubercle small and situated between distal 0.40 and 0.55 length of margin. Antennular peduncle with stylocerite produced in sharp, slender spine falling conspicuously short to almost reaching mesial base of distolateral spine; lat- ter acutely pointed; third article never produced laterally. Dorsal flagellum about 0.4 length of anten- nular peduncle, reaching between distal 0.25 and 0.20 of scaphocerite; ventral flagellum long and straight along entire length. Scaphocerite extremely long, exceeding anten- nular peduncle by about 0.30-0.35 its own length; strong lateral rib ending in sharp spine falling con- siderably short of distal margin of lamella. Anten- nal flagellum broken in specimens examined. Third maxilliped in both sexes with penultimate article convex dorsally, slightly flattened ventral- ly, and not produced in distal process; ultimate article also convex dorsally, slightly excavate ven- trally, and slender but broadening slightly from relatively elongate, narrow base before tapering to rather blunt apex. All pereopods covered with minute setae. First and second pereopods with compressed merus bear- ing distomesial spine. Abdomen with dorsomedian carina extending from fourth through sixth somites, carina quite low 333 FISHERY BULLETIN: VOL. 85, NO. 2 but clearly distinct on anterior part of fourth, sharp and rather high more posteriorly, and produced in short but strong spine on caudal margin of each somite; sixth also bearing pair of minute postero- ventral spines and 2 elongate cicatrices. Telson with median sulcus weak, usually limited to anterior half, flanked by paired longitudinal ridges reaching base of second of 4 pairs of movable, marginal spines situated at about 0.55, 0.75, 0.85, and 0.90 length from basal margin. Mesial ramus of uropod surpass- ing apex of telson by as much as 0.40 its own length; lateral ramus overreaching mesial ramus by as much as 0.33 its own length. Petasma of young individual lacking cincinnuli, similar to juvenile petasma of P. gracilis. Petasma of adults unknown. Curiously, only male available 21.5 mm CL with petasma still quite undeveloped. Appendices masculina and interna as in P. kath- leenae. In small juvenile males, sternite XIV bearing large, minutely setose prominence, semicircular in outline; median plate of sternite XIII setose, elon- gate (length 2-3 times basal width), and produced in conspicuous apical spine. Thelycum (Fig. 18) with setose, moderately long plate of sternite XIV broadly depressed and pro- duced at either side in short anterolateral hood, anteromedian margin varying from weakly convex to concave or biconcave; fossa preceding plate rela- tively short and bearing pair of small, oblique ridges. Median plate of sternite XIII, also covered with setae, relatively short (length 1.8-2.1 times basal width), falling considerably short of anterior margin of sternite XII; broadly lanceolate (maximum width 0.60-0.70 length), produced apically in acute spine, and flat or slightly excavate; posterolateral margins of plate, lacking bosses, abutting slender ridges ex- tending posteromesially before curving laterally on margin of sternite XIII. Sternite XII minutely setose, strongly convex and crested by slender me- dian carina ending anteriorly in sharp spine. Color, orange (Crosnier 1978). Maximum lengths.— Males, 22 mm CL (only juve- niles known; the 33.5 mm CL cited by Crosnier (1978) is a misprint, the specimen is a female); females, 47.5 mm CL. Geographic and bathymetric ranges.— This species has been found off Natal, South Africa, western Madagascar, and in the waters of Indonesia to southern Buru (Fig. 9). It occurs at depths between 834 and 1,200 m and was also taken in a trawl be- tween 810 and 1,020 m. i?rn ^^ Wf It^t^HH^fM-^--^ -^^f '" >^;s Figure 18.— Pseudaristetis sibogae (De Man), holotype 9 34 mm CL, east Savu Sea, Indonesia. Thelycum. Scale = 2 mm. Discvssion.-The minutely setose pereopods and the disposition of the deep groove lying dorsal to the posterior part of the hepatic sulcus, which abuts the cervical sulcus, distinguish P. sibogae from all the species previously described. It also is distinctive in having a longer optic calathus, the length of the mesial margin being at least 1.45 times its distal width instead of not more that 1.30. The tubercle of the calathus in P. sibogae is almost always situ- ated near its midlength, between the distal 0.4 and 0.6 length of the mesial margin rather than only as far as 0.4 or more often less, except in the eye of P. protenstis in which it is placed about at midlength. In females of P. sibogae the median plate of ster- nite XIII is shorter, its length 1.8-2.1 times the basal width, than in females of its congeners, in which the ratio is usually more than 2.1; in occasional speci- mens of P. gracilis it is 2, overlapping the highest ratios observed in P. sibogae. As stated above, the petasma of the adult of this species is not known; however, the very large prom- inence of sternite XIV, present in the 2 males ex- amined, appears to be a diagnostic feature. These specimens are 21 and 21.5 mm CL and, curiously, 334 PEREZ FARFANTE: REVISION OF GENUS PSEUDARISTEUS their petasmata are still little developed, lacking cin- cinnuli and apparently exhibiting no specific char- acter. In other species, males of this size may be identified by petasmal features. It seems worth men- tioning that in the 2 males of this species examined by me, the rostrum is slightly longer, 0.48 and 0.47 CL, than it is in most of the males of its congeners in which it ranges between 0.25 and 0.45 CL. De Man (1911) indicated that the rostral teeth were less prominent in the female holotype of P. sibogae than in the female of P. crassipes ( = P. kathleenae) available to him, and that they were situated in a horizontal line, whereas in the latter "a line uniting the tips of the teeth appears distinctly arcuate". Actually, the arrangement of the teeth in females with the same carapace length varies slight- ly between individuals of the same species, they are usually disposed in an arc, including P. sibogae, but sometimes they are arranged in an almost straight line. De Man also noted that in the holotype of P. sibogae the rostrum is much shorter [RL/CL = 0.75] and less slender than in the female of "P. crassipes". Crosnier (1978), on the basis of the comparison of a female 37 mm CL (RL/CL = 1.07) with the holo- type, believed that the difference in the length of the rostrum seemed invalid, that in the holotype the rostrum was in the process of being generated after having been broken. Pseudaristeus speciosus (Bate 1881) Figure 19 Hemipenaeus speciosus Bate 1881:186 [syntypes 1 o- 1 9 (BMNH); type-locality: E of Rio de la Plata, Argentina; 36°44'S, 46°16'W; 2,650 fm (4,847 m); 2 March 1876; Challenger stn 325]. Bate 1888: 303, pi. 37, Fig. 3, pi. 44, fig. 3. Murray 1896: 388. De Man 1911:26. Estampador 1937:493. Anderson and Lindner 1945:301. Burukovsky 1974:48. Hemipeneus speciosus. Faxon 1895:198. Ma^eriaL— Argentina Basin— o* syntype (BMNH). Diagnosis.— Optic calathus relatively long, mesial margin 1.4 times distal width. Anterolateral carina present. Ventral extremity of cervical carina broad and blunt rather than forming sharp-edged arc. Third article of antennular peduncle in females not expanded laterally. Posterior extremity of hepatic sulcus extending posteriorly subparallel to branchio- cardiac sulcus, instead of turning ventrally. Petas- ma and thelycum unknown. Description.— Based on few notes by Bate (1881), my observations of his illustration and examination of the incomplete cephalothorax of the male syntype. Body slender, lacking setae. Rostrum (Fig. 19) in male relatively short, its estimated length 0.40 CL and roughly lanceolate. Rostral plus epigastric teeth 3; rostral teeth situated at about 0.35 and 0.75 from orbital margin. Adrostral carina strong almost reaching apex. Antennal spine sharp; branchiostegal spine longer than antennal, acutely pointed. Cervical sulcus reaching but not crossing postrostral carina at about estimated 0.50 CL from orbital margin and well-marked dor sally; accompanying carina weak, its ventral extremity blunt instead of forming sharp- edged arc; postcervical sulcus reaching, but not crossing, postrostral carina at about estimated 0.70 CL from orbital margin. Postrostral carina well marked and sharp to cervical sulcus, low and blunt posteriorly. Anterolateral carina (ventral to gastro- FlGURE 19 .—Pseudaristetts speciosvs (Bate), syntype o- "total length = 63 mm" (Bate 1881), off east coast of Buenos Aires. Anterior part of anterior region, lateral view. Scale = 1 mm. 335 FISHERY BULLETIN: VOL. 85, NO. 2 orbital) dor sally concave, rather strong; gastro- orbital carina blunt but well defined; antennal carina relatively short, and branchiostegal-hepatic carina strong and sharp only anteriorly. Orbito-antennal sulcus quite shallow; hepatic sulcus not fusing with branchiocardiac sulcus and extending posteriorly, almost longitudinally rather than turning ventral- ly, subparallel to anterior part of branchiocardiac sulcus; branchiocardiac sulcus and accompanying carina long, extending posteriorly to near margin of carapace. Eye with optic calathus relatively long, length of mesial margin 1.4 times distal width; mesial tuber- cle situated almost at midlength. Antennular peduncle with stylocerite produced in sharp spine reaching mesial base of distolateral spine; latter small and sharp; third article in females not expanded laterally; dorsal and ventral flagella incomplete. Scaphocerite long, conspicuously surpassing antennular peduncle; strong lateral rib ending in sharp spine falling considerably short of distal end of lamella. Antennal flagellum incomplete. Geographic and bathymetric ranges.— Pseudaristeus speciosus is known only from the type-locality. Discission.— This species, tentatively assigned to the genus Pseudaristeus, can be readily distin- guished from the other members of the genus in possessing an anterolateral carina; the branchios- tegal-hepatic carina is strong and sharp only ante- riorly; the ventral extremity of the cervical sulcus is almost straight, instead of turning anteroventral- ly, and is accompanied by a very weak, rather than sharp, and strongly arched carina; also the posterior part of the hepatic sulcus extends subparallel to the branchiocardiac sulcus instead of fusing with it before turning ventrally. Pseudaristeus speciosus was described from 2 specimens, one of which is no longer extant and the other has disintegrated except for the anterior part of the carapace to which are attached the eyes, antennules and antennae, and the dismembered distal part of the third maxillipeds. Despite the poor condition of the available syntype, the distinctive features of the carapace, which are clearly repre- sented in Bate's (1888) illustration of the entire animal, are sufficient to conclude that P. speciosus is a valid species. Because the branchiae of the syn- type are lacking, it is not possible, as noted by Crosnier (1978), to determine with certainty the genus to which it should be assigned, but because of the supraspecific characters exhibited by the carapace, I am almost convinced that it is congeneric with the five Indo-West Pacific species studied herein. It should be noted that the syntypes of P. speci- osus were found at 4,847 m, a depth considerably beyond the greatest depth, 1,785 m, at which any of the assumed relatives are known to occur. ACKNOWLEDGMENTS Without the generous cooperation of various colleagues this study would not have been possible. I am much indebted to Maya Deb of the Zoological Survey of India for providing descriptions, drawings and photographs of certain morphological features of the lectotype of P. crassipes, which permitted a confirmation of the true identity of the species, and for the loan of critical collections from the waters off India; to Alain Crosnier of the Office de la Recherche Scientifique et Technique Outre Mer and the Museum National d'Histoire Naturelle for his hospitality during a working visit to the latter in- stitution, for the loan of specimens, and for review- ing the manuscript; and to Anthony A. Fincham of the British Museum (Natural History), H.-E. Gruner of the Zoologisches Museum de Humboldt Univer- sitat, and S. Pinkster of the Zoologisch Museum, Amsterdam, for providing materials, including types, from their respective institutions; to H. C. Ghosh of the Zoological Survey of India for a draw- ing of the antennular peduncle of the lectotype of P. crassipes. Horton H. Hobbs, Jr. of the Smith- sonian Institution once again offered invaluable ad- vice and innumerable suggestions during the course of my studies and preparation of the paper; Fenner A. Chace, Jr. aided me in solving technical problems and commented on the first draft; Bruce B. Collette and Austin B. Williams of the National Marine Fish- eries Service Systematics Laboratory critically read the manuscript. Keiko Hiratsuka Moore, with her artistic talent and devotion to accuracy, made all the illustrations except those of the eyes of various species and the gnathal appendages and of the carapace of P. kathleenae which were rendered by Maria M. Dieguez. Ruth E. Gibbons prepared the map. Virginia R. Thomas and Arleen S. McClain pa- tiently typed several drafts of the manuscript. To all of them goes my deep gratitude. LITERATURE CITED Alcock, a. W. 1898. A summary of the deep-sea zoological work of the royal Indian Marine Survey Ship "Investigator" from 1884 to 336 PEREZ FARFANTE: REVISION OF OENUS PSEUDARISTEUS 1897. Sci. Mom. Med. Off. Army India 11:45-93. 1901a. Zoological gleanings from the royal Indian Marine Survey Ship Investigator. Sci. Mem. Med. Off. Army India 12:35-76. 1901b. A descriptive catalogue of the Indian deep-sea Crustacea Decapoda Macrura and Anomala, in the Indian Museum. Being a revised account of the deep-sea species col- lected by the royal Indian Marine Survey Ship "Investi- gator". Indian Mus., Calcutta, 286 p. 1902. A naturalist in Indian seas or, four years with the royal Indial Marine Survey Ship "Investigator". John Murray, Lond., 328 p. Alcock, a. W., and a. R. S. Anderson. 1894. Natural history notes from H. M. Indian Marine Survey Steamer "Investigator", Commander, C. F. Oldham, R.N., commanding. Series II, No. 14. An account of a recent col- lection of deep sea Crustacea from the Bay of Bengal and Laccadive Sea. J. Asiat. Soc. Bengal 63(Part II, 3):141-185. 1899. Natural history notes from H. M. Indian Marine Survey Steamer "Investigator," Commander T. H. Heming, R.N., commanding. Series III, No. 2 An account of the deep-sea Crustacea dredged during the surveying-season of 1897-98. Ann. Mag. Nat. Hist., Ser. 7, 3:1-27, 278-292. Alcock, A. W., and A. F. McArdle. 1901. Illustrations of the zoology of the royal Indian Marine Survey Ship "Investigator," under the command of Com- mander T. H. Heming, R.N., Crustacea. Part IX, plates 49-55. Off. Supt. Gov. Print. India, Calcutta. Anderson, A. R. S. 1896. Natural history notes from the R.I.M. Survey Steamer "Investigator," Commander C. F. Oldham, R.N., command- ing. Series II, No. 21. An account of the deep sea Crustacea collected during the season 1894-95. J. Asiat. Soc. Bengal 65(Part II, 1):88-106. Anderson, W. W., and M. J. Lindner. 1945. A provisional key to the shrimps of the family Penae- idae with especial reference to American forms. Trans. Am. Fish. Soc. 73:284-319. Balss, H. 1925. Macrura der Deutschen Tiefsee-Expedition, 2. Natan- tia. Tei] A. Wiss. Ergeb. Dtsch. Tiefsee Exped. "Valdivia" 20:217-315. Bate, C. S. 1881. On the Penaeidea. Ann. Mag. Nat. Hist., Ser. 5, 8: 169-196. 1888. Report on the Crustacea Macrura collected by H.M.S. "Challenger" during the years 1873-76. Rep. Sci. Res. Voyage H.M.S. "Challenger" during the years 1873-76 (Zool.) 24, 942 p. BURUKOVSKY, R. N. 1974. Opredelitel Krevetok, langoustov i omarov [Keys to shrimps and lobsters]. Moskva Pishch. Promyst'., 126 p. A. A. Balkema (ed.), Rotterdam, 1983, Russ. Trans. Ser. 5, 173 p. BURUKOVSKY, R. N., AND L. L. ROMENSKY. 1972. The rostrum variability in the shrimp Aristeus varidens (Decapoda, Penaeidae) [In Russ.]. Rybokhoz. Issled. Atl. Okeane, Tr. AtlantNIRO 42, p. 156-160. Translated by U.S.A. Natl. Mar. Fish. Serv. Int. Act., Wash., D.C., TT 72-50101. Crosnier, a. 1978. Crustaces decapodes peneides Aristeidae (Benthe- sicyminae, Aristeinae, Solenocerinae). Faune de Madagas- car 46:1-197. 1984. Penaeoid shrimps (Benthesicymidae, Aristeidae, Solenoceridae, Sicyoniidae) collected in Indonesia during the Corindon II and IV Expeditions. Mar. Res. Indonesia No. 24, p. 19-47. 1986. Crevettes peneides d'eau profond recoltees dans I'ocean Indien lors des campagnes Benthedi, Safari I et //, MD 32lReunion. Bull. Mus. Nat. Hist. Nat., Paris, 4e Ser. 7, Sect. A [1985] (4):838-877. DE FREITAS, a. J. 1985. The Penaeoidea of southeast Africa. II. The families Aristeidae and Solenoceridae. S. Afr. Assoc. Mar. Biol. Res., Oceanogr. Res. Inst., Invest. Rep. No. 57, 69 p. DE Man, J. G. 191 1. Family Penaeidae. The Decapoda of the Siboga Expedi- tion. Part 1. Siboga Exped. Monogr. 39a, 131 p. 1913. Explanation of plates of Penaeidae. The Decapoda of the Siboga Expedition. Supplement to Part 1. Family Penae- idae. Siboga Exped. Monogr. 39a, (Suppl.), 10 pi. DOFLEIN, F. 1906. Ostasienfahrt. Erlebnisse und Beobachtungen eines Naturforschers in China, Japan und Ceylon, 511 p. ESTAMPADOR, E. P. 1937. A checklist of Philippine crustacean decapods. Philipp. J. Sci. 62:465-559. Faxon, W. 1895. The stalk-eyed Crustacea.Reports on an exploration off the west coasts of Mexico, Central and South America, and off the Galapagos Islands, in charge of Alexander Agassiz, by the U.S. Fish Commission steamer "Albatross," during 1891, Lieut. Commander Z. L. Tanner, U.S.N., commanding, XV. Mem. Mus. Comp. Zool., Harv. Coll. 18, p. 1-292. Freitas, A. J. DE. See de Freitas, A. J. Kemp, S., and R. B. Seymour Sewell. 1912. Notes on Decapoda in the Indian Museum. III. The species obtained by R.I.M.S.S. "Investigator" during the survey season 1910-11. Rec. Indian Mus. 7:15-32. Lloyd, R. E. 1907. Contributions to the fauna of the Arabian Sea, with descriptions of new fishes and Crustacea. Rec. Indian Mus. 1:1-12. Man, J. G. de. See de Man, J. G. MONOD, Th. 1974. Sur quelques crustaces neo-caledoniens de profondeur. Cah. ORSTOM, Ser. Oceanogr. 11:117-131. Murray, J. 1896. On the deep and shallow-water marine fauna of the Kerguelen region of the Great Southern Ocean. Trans. R. Soc. Edinb. 38:343-500. Pe'rez Farfante, I. 1969. Western Atlantic shrimps of the genus Penaeus. U.S. Fish Wildl. Serv., Fish. Bull. 67:461-591. 1977. American solenocerid shrimps of the genera Hymeno- penaeus, Haliporoides, Pleoticics, Hadropenaeiis new genus, and Mesopenaeus new genus. Fish. Bull., U.S. 75:261-346. 1985. The rock shrimp genus Sicyonia (Crustacea: Decapoda: Penaeoidea) in the eastern Pacific. Fish. Bull., U.S. 83:1-79. Ramadan, M. M. 1938. Crustacea: Penaeidae. Sci. Rep. John Murray Exped. 1933-34, 5(3):35-76. 1952. Contribution to our knowledge of the structure of the compound eyes of Decapoda Crustacea. Lunds Univ. Arsskr. New Ser. Sect. 2, 48(3):l-20. Sewell, R. B. S. 1955. A study of the sea coast of southern Arabia. Proc. Linn. Soc. Lond. 165:188-210. Silas, E. G., and M. S. Muthu. 1979. Notes on a collection of penaeid prawns from the Andamans. J. Mar. Biol. Assoc. India [1976] 18:78-90. 337 FISHERY BULLETIN: VOL. 85, NO. 2 Wood-Mason, J. 1891. Phylum Appendiculata. Branch Arthropoda. Class Crustacea. In J. Wood-Mason and A. Alcock (editors), Natural history notes from H. M. Indian Marine Survey Steamer "Investigator," Commander R. F. Hoskyn, R.N., commanding. Series II, No. 1. On the results of deep-sea dredging during the season 1890-91, p. 269-286. Ann. Mag. Nat. Hist., Ser. 6, 8. Wood-Mason, J., and A. Alcock. 1891. Natural history notes from H.M. Indian Marine Survey Steamer "Investigator," Commander R. F. Hoskyn, R.N., commanding. No. 21. Note on the results of the last season's deep-sea dredging. Ann. Mag. Nat. Hist., Ser. 6, 7:1-19, 186-202, 258-272. Young, J. H. 1956. Anatomy of the eyestalk of the white shrimp, PenaetLS setifencs (Linn. 1758). Tulane Stud. Zool. 3:171-190. 1959. Morphology of the white shrimp Penaeus setiferus (Linnaeus 1758). U.S. Fish Wildl. Serv., Fish. Bull. 59: 1-168. 338 DISTRIBUTION AND YIELD OF THE DEEPWATER SHRIMP HETEROCARPUS RESOURCE IN THE MARIANAS Robert B. Moffitt and Jeffrey J. Polovina^ ABSTRACT A shrimp trapping survey was conducted at 22 islands and banks in the Mariana Archipelago during a 2-year field period. Three species of deepwater shrimp were found in abundance at various depths: Heterocarpiis ensifer at 366-550 m, H. laevigatus at 550-915 m, H. longirostris >915 m. Heterocarpus laevigatus was the largest and most abundant of the three and has the greatest economic potential. Estimates of the unexploited biomass of this species by bank were calculated from estimates of catchabOity, relative abundance, and habitat area. An archipelago average of the unexploited trappable biomass was estimated to be 0.8 t/nmi". Evaluation of length-frequency distributions produced estimates of asymp- totic length (Loo) of 55 mm carapace length, instantaneous growth constant (K) of 0.3 yr" ', and instan- taneous total mortality (Z) of 0.75 yr"^ A recommended yield of 162.0 t/year (0.2 t/nmi" per year) for the entire archipelago was calculated using the Beverton and Holt yield-per-recruit equation based on minimum spawning stock considerations. Of this yield, 85% would come from the southern islands (e.g., Guam and Saipan), 13% from the northern islands (e.g.. Pagan and Anatahan), and 2% from the western seamounts (e.g., Arakane Reef and Pathfinder Reef). The Mariana Archipelago in the western Pacific Ocean stretches from Guam in the south at lat. 13°N to Farallon de Pajaros (also called Uracas) in the north at lat. 20°N (Fig. 1). Within the approximately 270,000 nmi- area of the 200 mi zone around the archipelago are two political entities— the Territory of Guam and the Commonwealth of the Northern Mariana Islands (CNMI)— and three geological formations— the southern island chain, the northern island chain, and the western seamount chain (Karig 1971). The purpose of this study was to assess the standing stock and sustainable yield of the deep- water pandalid shrimp resources in the Marianas. Pandalid shrimp catches account for about 9% of the world shrimp landings or about 155,000 1 in 1982 (FAO 1984). About 98% of this catch is of a few species of the genus Pandalus trawled at depths of 70-240 m in the cold-water areas of the North Atlan- tic, North Pacific, and Bering Sea. The next largest pandalid fishery is the trawl fishery for Hetero- carpus reedi conducted at depths of 155-424 m in the waters off Chili and Peru (Holthuis 1980). The landings from this fishery were 3,450 1 in 1982 (FAO 1984). In recent years, pandalid shrimp resources with commercial potential have been identified from deepwater trapping surveys conducted at depths of 200-1,200 m in the central and western Pacific (Clarke 1972; Wilder 1977; King 1980, 1981a, 1981b; Moffitt 1983). The primary component of these catches has been species of the genus Heterocarptcs, including H. laevigatus, H. ensifer, H. sibogae, H. longirostris, and H. gibbosus. Trawling for these species has produced poor results (Struhsaker and Yoshida 1975) which may be due to the depths in- volved, the rough bottom surrounding the Pacific islands, or behavioral characteristics of the shrimp. In Hawaii a rapidly expanding commercial trap fishery has been established with 1983 annual land- ings of about 135 t. Catches of 1,350 t have been projected for the near future by the Western Pacific Regional Fishery Management Council (WPRFMC 1984). This projected yield has not materialized and does not appear to be forthcoming since the larger shrimp trapping vessels have left the fishery for economic reasons. Commercial ventures in Guam and the CNMI have been sporadic and short lived; landings of 0.3 1 were reported in 1982, the last year that the resource was fished. ^ SAMPLING GEAR AND METHODS Shrimp trapping operations in the Mariana Archi- 'Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 96822-2396. ^Western Pacific Fishery Information Network data on file at the Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 96822-2396. Manuscript accepted January 1987. FISHERY BULLETIN: VOL. 85, NO. 2, 1987. 339 FISHERY BULLETIN: VOL. 85, NO. 2 Stingray Shoal Bank C Bank D Pathfinder Arakane Reef Bank A o Farallon de Pajaros : Supply Reef oMaug I. 'Asuncion 0 Agrihan I. ^P Pagan o Alannagan ° Guguan I. o Sarigan I. *^ Anatahan I. 38 fm Bank — Esmeralda X •Farallon de Medinilla /ySaipan I. -\}-Tinian I. — Aguijan I. ^ Rota I. GUAM "Cocos I. iGalvez Banks . 143° 144' '- Santa Rosa Reef 145" 146' I 147° E -20" -19" ■17" -I6°- -I5°- I4°- -13" Figure 1.— Chart of the Mariana Archipelago. 340 MOFFITT ET AL.: DISTRIBUTION AND YIELD OF DEEPWATER SHRIMP pelago were conducted on seven cruises of the NOAA ship Townsend Cromwell (TC) between April 1982 and Augnst 1984 and one charter cruise of the University of Guam vessel Pesquedot (PQ) in August 1984. The standard gear used consisted of strings of five canvas-covered, half-round shrimp traps set about 40 m apart. The traps were constructed of a reinforcing bar frame (about 90 cm long, 65 cm wide, and 45 cm high) wrapped with 2.5 x 1.3 cm mesh, 18-gauge welded wire. An entry cone with an open- ing of approximately 10 cm was located at each end of the trap. A figure of this general trap design is shown in Gooding (1984). The ground lines attach- ing the traps together and the main lines attaching the ground line to the surface buoys were of 13 mm polypropylene line. Strings were usually set in the afternoon and retrieved the following morning. Normal soaking times ranged from 15 to 24 hours. Traps were baited with Pacific mackerel, Scomber japonicus. Ordinarily, subsamples of 100 specimens of each of the three major species of Heterocarpus {H. laevi- gatus, H. ensifer, and H. longirostris) were saved each day from each depth sampled. Two sampling sites, Esmeralda Bank and Pagan Island, were visited on each of six Cromwell cruises and large subsamples of 400 H. laevigatus were saved on each visit. All specimens were returned to the laboratory where the carapace length, sex, and reproductive condition were recorded. The areas of suitable habitat for H. laevigatus at each island and bank location was estimated from charts using a computer aided planimeter. In this study, the yield assessment approach described by Polovina and Ralston (1986) was em- ployed. A systematic survey of 22 islands and banks gave information on the depth range of the primary species and their relative abundance by area. An in- tensive fishing experiment produced an estimate of catchability using the Leslie method (Ralston 1986). This information combined with an estimate of the area of suitable habitat for each island or bank was used to estimate available biomass by location. Esti- mates of growth were obtained by application of Elefan I (Pauly 1982) to a site specific time series of length-frequency data. The ratio of mortality to growth and asymptotic length was estimated from a large length-frequency sample (Wetherall et al. in press). Equilibrium yield as a function of fishing mortality was determined from the Beverton and Holt (1956) yield-per-recruit equation as the product of yield per unexploited trappable biomass and the trappable recruited biomass estimate obtained from the systematic sampling and intensive fishing (Polo- vina and Ralston 1986). Estimates of recommended yield from the equilibrium yield equation were ob- tained based on marginal yield and minimum spawn- ing stock biomass considerations. RESULTS^ Throughout the course of this survey a total effort of 2,508 trap-nights was expended at 527 shrimp trapping stations. The total catch of pandalid shrimp was 5,188 kg for an overall catch rate of 2.07 kg/trap-night. Over 99% of this catch was composed of//, ensifer, H. laevigatus, and//, longirostris. A complete list of shrimp species taken during this study is given in Table 1. Table 1. — List of shrimp species taken (C = common, F = frequent, R = rare). Depth of Species Frequency abundance (m) Pandalidae Plesionika serratifrons C 90-270 Plesionika longirostris C 180-360 Plesionika ensis F 450-630 Plesionika martia R 360-450 Heterocarpus ensifer C 360-540 Heterocarpus gibbosus R 360-450 Heterocarpus sibogae R 630 Heterocarpus lepidus F 540-720 Heterocarpus laevigatus C 540-810 Heterocarpus dorsalis F 630-900 Heterocarpus longirostris C >900 Heterocarpus tricarinatus R 900 Opiopfioridae Acanthephyra eximia F 810-990 Aristaeidae Plesiopenaeus edwardsianus R 630-810 DEPTH AND SIZE DISTRIBUTION A plot of catch per unit effort (CPUE) versus depth for the three major species (Fig. 2) shows that they inhabit different depth strata. Heterocarpus en- sifer is the shallowest dwelling of these species. The maximum catch rate for this species was 0.17 kg/ trap-night at a depth of 366 m (200 fathoms). Un- fortunately, 366 m was the shallowest depth tar- geted throughout most of the survey. A small amount of effort was expended in the 137-274 m (75-150 fathoms) depth range on cruises TC-84-02 and PQ-84-01. Heterocarpus ensifer catches in this depth range were negligible (only five shrimp in 37 trap-nights). Heterocarpus laevigatus, the most abundant shrimp taken in our survey, was caught ^Portions of this section are also presented in Polovina et a!. 1985. 341 FISHERY BULLETIN: VOL, 85, NO. 2 3.0 -, 300 500 600 700 DEPTH (m) Figure 2.— Catch rate of three species of Heterocarpus by depth. in large numbers at depths between 549 and 777 m (300 and 425 fathoms). A maximum catch rate of 2.33 kg/trap-night was obtained at 777 m. Hetero- carpus longirostris was the deepest dwelhng of the Heterocarpus species taken, ranging from 823 m (450 fathoms) down past 1,097 m (600 fathoms), which was the greatest depth targeted in this study. Maximum catch rates of H. longirostris were ob- tained at 1,052 m (575 fathoms), but the deeper end of this species' range was not sampled and, as with H. ensifer, it is uncertain that the depth of maximum catch rate obtained here is indicative of depth of maximum abundance. Heterocarpus ensifer was the smallest of the three major species taken. Carapace lengths of 3,401 in- dividuals ranged from 11 to 38 mm with a mean of 26.4 mm. Heterocarpus laevigatus was the largest species taken. The mean carapace length was 38.2 mm and the range was 13 to 61 mm {N = 16,405). The mean carapace length of H. longirostris was similar to that oiH. laevigatus (i = 37.5 mm), but the size range of 1,443 individuals was more re- stricted (20 to 50 mm). The mean size of H. longi- rostris taken in this study is probably higher than the mean of catchable shrimp of this species for all depths. We set traps only in the shallower end of this species' depth range (<1,143 m) where the population is dominated by females that grow to a larger size than the males. It has been suggested that the size of shrimp varies with depth, i.e., larger shrimp occur within the range of maximum abundance while smaller in- dividuals were found in shallower or deeper water (Clarke 1972; Wilder 1977). This type of distribu- tion was not observed for any of the three major species of Heterocarpus in the Marianas. Linear regressions of mean carapace length with depth by sex were computed for each of the three species. A significant decrease in carapace length with increas- ing depth was obtained for H. longirostris females, although the full depth range of this species was not sampled. In all other cases, no significant change in size with depth was observed (Table 2). Wilder (1977) reported high male to female ratios of 3-4 to 1 for H. ensifer and H. laevigatus taken around Guam. He also stated that small individuals were nearly all males and large individuals almost all female. This was not true for the shrimp ex- amined in this study. The overall sex ratios (ex- Table 2. — Results of mean carapace length by depth regressions for three species of Heterocarpus by sex. Species Sex Regression coefficient fl2 Prob- ability H. ensifer Male Female 0.008 0.002 0.41 0.03 0.06 0.64 H. laevigatus Male Female -0.007 -0.017 0.26 0.16 0.06 0.16 H. longirostris Male Female -0.004 -0.04 0.02 0.67 0.83 0.02 342 MOFFITT ET AL.: DISTRIBUTION AND YIELD OF DEEPWATER SHRIMP pressed in percent males) for the three major species were 52.8% for H. ensifer (N = 3,302), 55.2% for H. laevigatus {N = 12,555), and 24.2% for//, longi- rostris {N = 1,408). These ratios tended to hold true for all size classes except the very largest of each species which were indeed nearly all females. Sex ratios did differ by depth showing that the two sexes tend to occupy different areas. In all three major species, females were more abundant at the shal- lower end of the depth range and males at the deeper end (Fig. 3). This relationship is much more obvious for H. laevigatus and H. longirostris than for H. ensifer. For H. longirostris, this may explain the small percentage of males taken in this study since the deeper end of the depth range of this species was not sampled. 100 80 [2 60 -I < s v9 40 20 100 80-1 (A uj 60 < S S? 40 20 100 80 ffi 60 -I < S a? 40 - 20- H. ensifer ■',' II I H. laevigatus I 1 ."I H. longirostris 400 500 600 700 800 900 1000 1100 1200 DEPTH (m) Figure 3.— Sex ratio by depth for three species of Heterocarpus with 95% confidence limits. REPRODUCTION Pandalid shrimp are typically considered to be pro- tandrous hermaphrodites. This is indeed the case for the Pandalus species taken in the subarctic areas (Butler 1964). In these species the shrimp spend the first few years of life as functional males, trans- forming into functional females for the last year or two of life. Clarke (1972) and Wilder (1977) sug- gested that the tropical Heterocarpus shrimp also are protandrous hermaphrodites. The sex ratios ob- tained in our studies, particularly the near even ratio for the smaller individuals led us to believe that this was not the case. King and Moffitt (1984) examined males of several species of Heterocarpus and Plesionika for relative growth of the appendix masculina on the second pleopod (a secondary sex characteristic). If these species were indeed pro- tandrous hermaphrodites, the relative size of the male appendage should decrease with increasing carapace length (as the shrimp transforms from male to female). This was not so for any of the tropical pandalids examined. Instead, the relative size of the male appendage increased with increased carapace length indicating maturation as a male. The Marianas data for the ratio (R) of the appen- dix masculina to the appendix interna versus carapace length (CL) was fit to the logistic model with three parameters, a, b, and c (Gunderson et al. 1980). Table 3 lists the parameters obtained when fitted to the model R X 100 = a I + g-6(CL-c) The fit to the nonlinear regression for H. longi- rostris was not particularly good due to the lack of small males in our collection. When the data for this species are fit to a linear regression, however, the slope of the regression is positive indicating rela- tive growth of the secondary sexual characteristics Table 3. — Parameter values for the nonlinear regression of the relative length of the appendix masculina versus carapace length. Species a' b c^ Plesionika longirostris Heterocarpus ensifer Heterocarpus laevigatus Heterocarpus longirostris ^Asymptotic value for the ratio of ttie lengths of the appendix masculina and the appendix interna. ^Carapace length at the inflection point at 50% of the asymptotic ratio. 106.83 0.75 11.67 80.35 0.43 15.05 104.97 0.29 28.09 158.32 0.05 19.36 343 FISHERY BULLETIN: VOL. 85, NO. 2 with increasing size (regression coefficient = 1.64, r- = 0.52). Based on the assumption that the relative growth of the appendix mascuhna correlates directly with maturity, we chose the point where this appendage is 90% of its asymptotic value to define the length at maturity for males. The 90% level was used in- stead of 50% of the asymptotic value, as used for the females below, because the 50% point would be where the males are 50% mature, not where 50% of the males are mature. Using this definition, the carapace length at maturity (L^j) for males of H. ensifer is 20.2 mm and that ofH. laevigatus is 35.7 mm. For H. longirostris, the length at maturity is estimated at 31 mm. The length at maturity of females is perhaps more important in assessment work since the females are directly responsible for the production of recruits to the population. For this study we used the pres- ence of eggs (berried) as the indicator of maturity. The length at maturity is defined as the size where 50% of the females are mature (Gunderson et al. 1980). When using the presence of eggs as the measure of maturity, the sample must be restricted to females collected during the time of year that egg bearing can be expected. For H. laevigatus, the breeding season is relatively discrete (November to February), whereas for H. ensifer and H. longiros- tris there are peaks in December and May (Fig. 4). Data for each species were fitted to the same non- linear regression model used for the males. Asymp- totic values for percent berried by carapace length are 66% for H. ensifer, 92% for H. laevigatus, and 55% fori/, longirostris. The carapace lengths asso- ciated with values equal to one-half of the asymp- totic values are the L^ for the various species. These are 23.9 mm for H. ensifer, 42.7 mm for H. laevigatus, and 37.4 mm for H. longirostris. YIELD ASSESSMENT The assumptions and methods of yield assessment used in this study are presented in Polovina and Ralston (1986) and Wetherall et al. (in press). Be- cause H. laevigatus yielded the highest catch rates and because it is generally regarded as having superior market acceptability, most of our fishing effort targeted this species. Hence estimates of total biomass and sustainable yield for the pandalid re- source are restricted to this species. GROWTH AND MORTALITY Estimates of asymptotic size (L^) and the ratio 344 uu 80 H. ensi er 60 ■ 40 1 1 20 - f — 1 uu ■ 80 ■ H. laevigatus 60 ■ ; 40 - \ 20 \ \ \ f f 1 r— — » . 100 80 60 ■ 40 • 20 H. longirostris I I Nov Dec Jan Feb Mar Apr May Jun Jul Aug MONTH Figure 4.— Percentage of females bearing eggs by month for three species of Heterocmyus with 95% confidence limits. of instantaneous total mortality to the instantaneous growth constant {ZIK) were obtained by examining the descending limb of the length-frequency distri- bution using the regression method based on the Beverton and Holt (1956) model (Wetherall et al. in press). Table 4 lists the values of L^ and ZIK for males, females, and pooled sexes for each of the three major species. As anticipated, L^ values for females are larger than those for males of the same species. Estimates of length at recruitment to the exploitable population (L^j) were obtained by apply- ing the method of Gulland (1969) to the ascending limb of the same length-frequency distributions. Heterocarpus ensifer is recruited into the fishery at a carapace length of 23 mm, H. laemgatus at 29 mm, and H. longirostris at 34 mm. By fixing the L^ value at the estimates obtained from the large length-frequency sample and using MOFFITT ET AL.: DISTRIBUTION AND YIELD OF DEEPWATER SHRIMP Table 4. — L^ and ZIK estimates for the three major species of Heterocarpus by sex from length-frequency data. Table 5.— Catch rates, habitat areas, and unexploited biomass estimates for Heterocarpus laevigatas by location. Species U ZIK Heterocarpus ensifer Males 34.5 2.6 Females 37.5 3.0 Sexes combined 36.6 2.9 Heterocarpus laevigatas Males 51.3 2.1 Females 55.4 1.9 Sexes combined 55.2 2.5 Heterocarpus longirostris Males 41.0 1.7 Females 48.1 1.9 Sexes combined 48.6 2.2 Elefan I (Pauly 1982) to fit the von Bertalanffy growth curve to the time series data for H. laeviga- tus collected at Esmeralda Bank and Pagan Island, the growth constant i^ can be estimated. When ap- plied to the sexes separately, multiple estimates were obtained for each category ranging from 0.19 to 0.31 yr"^ The inconsistency of these estimates within area and sex groupings was most likely due to the small sample size. When sexes were pooled, however, K was estimated at 0.30 yr"^ for both areas. Estimates of the age at recruitment and maturity were obtained by solving the von Berta- lanffy equation for the particular carapace lengths estimated above. For female H. laevigatus, with a L^ = 29 mm, the T^ = 2.0 years, and with Ljij = 43 mm, the 7^^ = 4.5 years. With a K estimate of 0.3 yr-i and ZIK of 2.5, an estimate for Z of 0.75 yr"^ is obtained. Because there is no fishery for H. laevigatus in the Marianas, Z is equivalent to natural mortality (M). UNEXPLOITED BIOMASS Because standard trapping techniques were used throughout our study, CPUE values from various locations could be used as a measure of relative abundance. The unexploited biomass of the H. laevi- gatus resource for each area is then calculated as the product of the area of suitable habitat and rela- tive abundance divided by the coefficient of catch- ability (g = 0.001945 trap-night~^) estimated from the Alamagan Island intensive trapping operation (Ralston 1986). Although the catch rate from the western seamounts is about twice that of the south- ern island chain, the fiftyfold greater area of suit- able habitat around the southern islands more than compensates for the low catch rates in producing a higher biomass estimate for the southern islands (Table 5). Catch rate Area Biomass (kg/trap-night) (nmi^) (t) Northern Banks Maug 1.88 3.83 3.7 Asuncion 2.11 5.93 6.4 Agrihan 1.96 12.39 12.5 Pagan 2.17 16.19 18.0 Alamagan 2.18 11.43 12.8 Guguan 2.52 5.60 7.2 Sarigan 1.45 4.55 3.4 Anatahan 2.36 10.89 13.2 38 Fathom Bank 2.12 6.37 6.9 Esmeralda Bank 1.35 2.03 1.4 Mean = 2.01 Total = 79.21 85.5 Southern Banks Farallon de Medinilla 0.97 88.55 44.2 Saipan 2.06 213.99 226.5 Tinian 1.81 73.80 68.4 Aguijan 1.61 39.36 32.6 Rota 1.02 197.31 103.6 Guam 0.48 44.24 16.2 Galvez and Santa Rosa 1.78 50.77 84.5 Mean = 1.39 Total = 708.02 576.0 Seamounts Bank C 2.07 2.71 2.9 Bank D 2.72 2.71 3.8 Pathfinder 2.79 2.71 3.9 Arakane 2.83 2.10 2.1 Bank A 1.43 3.33 2.4 Mean = 2.37 Total = 13.56 15.1 EQUILIBRIUM YIELD With the values for K, L^, M, and Tf., the Bever- ton and Holt yield-per-recruit equation can be used to compute the ratio of equilibrium yield to unex- ploited recruited biomass (Y/B) as a function of fishing mortality (F) (Polovina and Ralston 1986). Because the shrimp resource in the Marianas is not fished, the estimates of the biomass for each bank represent the unexploited trappable biomass (B), and hence, the product of Y/B and B gives the equi- librium yield as a function of F (Table 6). As F in- creases, the equilibrium yield increases rapidly for low levels ofF. The relationship between F and the equilibrium yield estimated from the Beverton and Holt yield-per-recruit equation assumes that recruit- ment is unchanged as F increases and does not take into account any economic considerations. Ideally a spawner-recruit relationship is needed to account for changes in yield because of the changes in re- cruitment which might occur as F increases. How- ever, in the absence of a knowledge of the spawner- recruit curve, two approaches can be used to estimate recommended yield. One approach esti- 345 FISHERY BULLETIN: VOL. 85, NO. 2 Table 6.— Equilibrium yield of Heterocar- pus laevigatas and relative spawning stock biomass as a function of fishing mortality (F). Total yield Relative spawning F (t) stock biomass 0.1 56.2 0.70 0.2 95.6 0.51 0.3 124.2 0.37 0.4 145.4 0.27 0.5 161.6 0.20 0.6 174.1 0.15 0.7 184.1 0.11 0.8 192.0 0.09 0.9 198.5 0.07 1.0 203.8 0.05 mates the recommended yield from the yield-per- recruit derived yield equation as the yield which corresponds to that level of effort where an increase in one unit of effort will increase the catch by 0.1 of the amount caught by the very first unit of ef- fort (Gulland 1983, 1984). This effort is denoted as Fq 1 and the corresponding yield as Yf, i- The value of Fq 1 for H. laevigatus in the Marianas is 0.8 and Fo.i is 192 t annually (Table 6) from areas within the depth range of 500-825 m. A second approach to estimate recommended yield uses a computation of the spawning stock biomass. With an age estimate of the onset of sexual matur- ity, the spawning stock biomass for a level of F relative to the spawning stock biomass, in the ab- sence of fishing, can be computed from the Bever- ton and Holt yield-per-recruit equation (Polovina and Ralston 1986). This relative spawning stock biomass can be used to determine the maximum value of F before a substantial decline in recruitment occurs. The recommended yield can then be estimated as the yield from the constant recruitment yield curve which corresponds to that maximum value of F. This is a conservative approach because it does not in- corporate any density dependent compensation, i.e., size at onset of sexual maturity does not decrease as density decreases. The relationship between the relative spawning stock biomass and recruitment is not known for H. laevigatus, but it has been sug- gested that as a lower bound the relative spawning stock biomass should not be reduced below 20% of the unexploited level if a substantial reduction in recruitment is to be avoided (Beddington and Cooke 1983). When F is 0.5, the relative spawning stock biomass is estimated to be 20% of the unexploited level, and the equilibrium yield at this level of fishing is estimated at 162 1 annually (Table 6) for the depth range of 500-825 m. To be conservative, the lower yield estimate of 162 t annually from the Mariana Archipelago will be used. Given the habitat area from 500 to 825 m, this yield is equivalent to 0.20 t/nmi^. An approximate variance for this yield, and hence an approximate confidence interval, can be computed from a Taylor series expansion of the yield estimator if it is assumed that the variance of the yield estimate is due primarily to variances in bank CPUE and catchability. The yield at each bank is computed as Yield = (CPUE/g)(Area)(y/B). Thus the variance of the yield (y(Yield)) can be ex- pressed as y (Yield) = (Area)2(y/£)2y(CPUE/g), = (AreajHYmy r y(CPUE) (CPUE)^ Viq) X O 1 A Estimates of F(CPUE) were obtained from the repeat sampling at each bank and ranged from 0.02 at Guam to 0.64 at Sarigan, while V{q), estimated at 5.5 X 10"", was obtained from the intensive trapping work. The variance of the total yield was estimated as the sum of the individual bank vari- ance. The 95% confidence interval, derived from the estimate ± 1.96 times the standard deviation of the estimate of total yield, resulted in a targeted yield range of 102 to 218 t (0.12 to 0.27 t/nmi^) per year. About 85% of this yield would come from the south- ern islands and banks, 13% from the northern islands, and about 2% from the western seamounts (Table 7). DISCUSSION Although trap design, depth fished, and species present undoubtedly affect catch rate, catch rates reported in other studies using differing trap designs in various areas fall within a fairly tight range of 1.2 to 6.6 kg/trap-night (Table 8). This indicates that the productivity of deepwater pandalids is relative- ly uniform throughout the tropical central and western Pacific, and the first estimate of recom- mended yield of 0.2 t/nmi^ obtained from the Mari- ana Archipelago can be applied to other Pacific islands, though the relative importance of the vari- ous species may differ greatly from area to area. In our study H. ensifer, H. laevigatus, and H. 346 MOFFITT ET AL.: DISTRIBUTION AND YIELD OF DEEPWATER SHRIMP Table 7.— Equilibrium yield for Heterocar- pus shrimps In the 500-825 t depth range for a fishing mortality of 0.5. Bank Yield (t/yr) Northern Banks Ivlaug 0.9 Asuncion 1.5 Agrihan 3.0 Pagan 4.3 Alamagan 3.0 Guguan 1.7 Sarigan 0.8 Anatahan 3.1 38 Fathom 1.7 Esmeralda 0.3 Total 20.3 Southern Banks Farallon de Medinilla 10.6 Saipan 54.1 Tinian 16.3 Aguijan 7.8 Rota 24.7 Guam 3.9 Galvez and Santa Rosa 20.2 Total 137.6 Seamounts Bank C 0.7 Bank D 0.9 Pathfinder 0.9 Arakane 0.5 Bank A 0.6 Total 3.6 Archipelago total 161.5 Table 8.— Catch rates of pandalid shrimp in the tropical Pacific. Catch rate Area (kg/trap-night) Source Tonga 1.2 King 1981b Samoa 1.4 King 1980 Fiji 1.5 King 1983 Hawaii (main islands) 1.5 Clarke 1972 Guam 2.1 Wilder 1977 Mariana Archipelago 2.1 Present study Hawaii (IMWHI) 2.5-3.5 Oishi 1983 Vanuatu 2.8 King 1981a Hawaii (main islands) 3.5-5.0 Hawaiian Divers, Inc. 1983 (text fn. 4) Hawaii (main islands) 6.6 Struhsaker and Aasted 1974 longirostris proved to be the major components of the catch within their respective depth zones. Heterocarpus ensifer was the shallowest dwelling and the least abundant of the three major species in the Marianas. Its range of abundance was 366 to 503 m (200 to 275 fathoms) and the peak catch rate was 0.17 kg/trap-night at 366 m. In Hawaii, H. en- sifer appears to be the most abundant species. Average catch rates are between 1.5 and 6.6 kg/trap-night in a somewhat wider reported depth of abundance of 274 to 600 m (Clarke 1972; Struhsaker and Aasted 1974; Gooding 1984). In the Southern Hemisphere; H. ensifer is not found in great abundance and is replaced in the 300 to 500 m depth range by a very closely related species, H. sibogae (King 1983). Heterocarpus laevigatus is an important part of the catch throughout the central and western Pacific. In the Marianas, this was the most common species. Its abundance peaked between 549 and 777 m (300 and 425 fathoms) and the maximum catch rate of 2.33 kg/trap-night was obtained at 777 m. In the Northwestern Hawaiian Islands (NWHI), the same standard half-round traps caught just under 1.0 kg/trap-night in an optimum depth range of 500 to 800 m (Gooding 1984). Commercial vessels using larger traps obtained catches of 2.5 to 5.0 kg/trap- night in the Hawaii area and found that the opti- mum depth range is shallower in the main Hawaii- an Islands (530 to 622 m) than in the NWHI (640 to 732 m) (Hawaiian Divers 1983*; Oishi 1983; Good- ing 1984). In the South Pacific, H. laevigatus is reported to be abundant at depths of 549 to 640 m and catch rates range from 0.4 to 1.1 kg/trap-night depending on the area studied (King 1983). Before this study, H. longirostris had been known to science from only four specimens taken in the Indian Ocean (Moffitt 1983). In the Marianas, it occurs in sufficient quantity to suggest a commer- cial potential. Heterocarpus longirostris is probably present in many other areas in the Pacific but has not been found because its optimum depth range is below those sampled. In the Marianas, sex ratio varied with depth for all three of the major species of Heterocarpus. For each species, a larger percentage of the catch is com- posed of females at the shallower end of the species' depth range. A similar distribution in Hawaii has been reported for H. ensifer (Clarke 1972) and H. laevigatus (Dailey and Ralston 1986). Changes in the mean size with depth have been reported for H. ensifer and H. laevigatus. Wilder (1977) and Gooding (1984) reported increases in size with increasing depth for H. ensifer from Guam and the NWHI, respectively. Clarke (1972), on the other hand, found that a higher proportion of larger in- dividuals in Hawaii occupied the depth of greatest abundance, while smaller individuals were found in shallower or deeper water. Gooding (1984) noted a ^Hawaiian Divers, Inc. 1983. Deepwater shrimp utilization study for Hawaii. Report prepared under NOAA Cooperative Agreement No. 80-ABH-00065 for the Southwest Region, West- ern Pacific Program Office, National Marine Fisheries Service, NOAA, 47 p. 347 FISHERY BULLETIN: VOL. 85, NO. 2 decline in size (kilograms/individual) with increas- ing depth for H. laevigatus in the NWHI. In all of these studies, changes in sex ratio with depth were not taken into account. As we have shown, sex ratio does change with depth and the sexes do grow at different rates. The observed changes in size with depth may be due to changes in sex ratio rather than size-specific stratification. Dailey and Ralston (1986) examined the sexes separately and found that for H. laevigatus in Hawaii the carapace length of males and egg-bearing females displayed no apparent change with depth, whereas that of nonegg-bearing females showed a strong inverse relationship. In the Marianas, significant changes in mean carapace length with depth were not observed for either sex of the three species of Heterocarpus, except for female H. longirostris (Table 2). For this group, an inverse relationship was observed much like that found for nonegg-bearing female H. laevigatus in Hawaii (Dailey and Ralston 1986). The estimates of growth parameters for H. laevi- gatus obtained in this study correspond well with those of other authors (Table 9). Using the regres- sion method (Wetherall et al. in press), we estimated L^ to be 51.3 mm CL for males, 55.4 mm CL for females, and 55.2 mm CL for the pooled population in the Marianas. Using the same method, Dailey and Ralston (1986) obtained estimates of 57.9 mm CL for males, 62.5 mm CL for females, and 61.7 mm CL for the combined sexes in Hawaii. Apparently, H. laevigatus grows about 7 mm larger in Hawaii than in the Marianas. King (1983), using the Bever- ton and Holt method, estimated L^ = 57 mm CL for H. laevigatus in Fiji. Estimates of Z/K for H. laevigatus in the Mari- anas were 2.1 for males, 1.9 for females, and 2.5 Table 9. — Asymptotic length (Z-oo), instantaneous growth constant (K), length at maturity {L,^), and age at maturity (T^) of Hetero- carpus laevigatus. Location Loo (mm) K(yr-^) Lm Tm Fiji (King 1983) Sexes combined 57 0.27 Females 40.5 4.6 Males 24 Hawaii (Dailey and Ralston 1986) Sexes combined 61.7 Females 62.5 0.25 40 4 Males 57.9 0.35 Mariana Archipelago (present study) Sexes combined 55.2 0.30 Females 55.4 43 4.5 Males 51.3 36 3.5 for the sexes combined. In Hawaii, ZIK estimates of 4.3, 2.9, and 2.6 were obtained for the same categories, respectively (Dailey and Ralston 1986). The ZIK estimates for the combined sexes are nearly identical in the two studies. In our study, Z/K esti- mates for the two sexes were similar to each other and lower than that of the sexes combined, whereas in the Hawaii study they were very different from each other and both larger than that of the combined sexes. In our study, if we assume that instantaneous growth of the two sexes is similar, then mortality will also be similar and close to the 0.75 yr~^ value estimated for the pooled sexes. In the Hawaii study, ZIK estimates for males and females differed wide- ly. Mortality estimates differed considerably as well, Z = 1.51 yr"^ for males and 0.73 yr"^ for females. The K parameter for H. laevigatus in the Mari- anas was estimated as K = 0.30 yr"^ for the com- bined sexes (Table 9). Estimates of K for the in- dividual sexes were ambiguous and inconsistent. King (1983) estimated K as 0.27 yr-i for H. laevi- gatus in Fiji and Dailey and Ralston (1986) estimated K as 0.35 yr " ^ for male and 0.25 yr" ^ for female H. laevigatus in Hawaii. Because growth estimates for H. laevigatus are similar for the various areas studied, it is not surprising that estimates of age at maturity are also similar (Table 9). King (1983) reported the age at maturity for female H. laevi- gatus in the South Pacific as 4.6 years (40.5 mm CL). He further suggests that males mature at about 24 mm CL (age not calculated). Dailey and Ralston (1986) found that females in Hawaii mature at 40 mm or about 4 years. In the Marianas female matur- ity is estimated at 43 mm CL or about 4.5 years. Males mature earlier at 35.7 mm CL or about 3.0 to 3.5 years. Male maturity estimates by King (1983) and this study are based on the relative growth of the appendix masculina on the second pleopod. King appeared to have chosen 50% of the asymptotic value as the point of maturity much as the point where 50% of the female shrimp are bearing eggs is used to define maturity for females. We feel that 90% of the asymptotic value is a better estimate of male maturity and have used that point in our esti- mate for the Marianas. ACKNOWLEDGMENTS This manuscript is a result of the Resource Assess- ment Investigation of the Mariana Archipelago (RAIOMA) program conducted by the National Marine Fisheries Service, Southwest Fisheries Center Honolulu Laboratory. We thank the facul- ty, staff, and students of the University of Guam 348 MOFFITT ET AL.: DISTRIBUTION AND YIELD OF DEEPWATER SHRIMP Marine Laboratory for their assistance and par- ticipation during the field period. We also thank the staffs of the Division of Aquatic and Wildlife Re- sources of the Territory of Guam and the Division of Fish and Wildlife of the Commonwealth of the Northern Mariana Islands for their cooperation throughout the study. LITERATURE CITED Beddington, J. R., AND J. G. Cooke. 1983. The potential yield of fish stocks. FAO Fish. Tech. Pap. 242, 47 p. Beverton, R. J. H., AND S. J. Holt. 1956. A review of methods for estimating mortality rates in fish populations, with special reference to sources of bias in catch samplings. Rapp. P.-v. Reun. CIEM, 140:67-83. Butler, T. H. 1964. Growth, reproduction and distribution of pandalid shrimps in British Columbia. J. Fish. Res. Board Can. 21: 1403-1452. Clarke, T. A. 1972. Exploration for deep benthic fish and crustacean re- sources in Hawaii. Univ. Hawaii, Hawaii Inst. Mar. Biol., Tech. Rep. 29, 18 p. Dailey, M. D., and S. Ralston. 1986. Aspects of the reproductive biology, spatial distribu- tion, growth, and mortality of the deepwater caridean shrimp Heterocarpus laevigatus in Hawaii. Fish. Bull., U.S. 84:915-925. FAO. 1984. 1982 Yearbook of fishery statistics; catches and land- ings. FAO, Rome 54, 393 p. Gooding, R. M. 1984. Trapping surveys for the deepwater caridean shrimps, Heterocarpus laevigatus and H. ensifer, in the Northwest- ern Hawaiian Islands. Mar. Fish. Rev. 46(2): 18-26. Gulland, J. A. 1969. Manual of methods for fish stock assessment. Part 1. Fish population analysis. FAO Fish. Ser. 3, 154 p. 1983. Fish stock assessment. A manual of basic methods. John Wiley & Sons, N.Y., 223 p. 1984. Advice on target fishery rates. Fishbyte [ICLARM] 2(1):8-11. GuNDERSON, D. R., p. Callahan, and B. Goiney. 1980. Maturation and fecundity of four species of Sebastes. Mar. Fish. Rev. 42(3-4):74-79. HOLTHUIS, L. B. 1980. Shrimps and prawns of the world. An annotated cat- alogue of species of interest to fisheries. FAO Fish. Synop. (125) vol. 1, 261 p. Karig, D. E. 1971. Structural history of the Mariana Islands arc system. Bull. Geol. Soc. Am. 82:323-344. King, M. G. 1980. A trapping survey for deepwater shrimp (Decapoda: Natantia) in Western Samoa. Rep. Inst. Mar. Res., Univ. South Pacific, Fiji, 26 p. 1981a. Deepwater shrimp resources in Vanuatu: a prelim- inary survey off Port Vila. Mar. Fish. Rev. 43(12): 10-17. 1981b. The deepwater shrimps of Tonga: a preliminary survey near Nuku'alofa. Rep. Inst. Mar. Res. Univ. South Pacific, Fiji, 29 p. 1983. The ecology of deepwater caridean shrimps (Crustacea: Decapoda: Caridea) near tropical Pacific islands with par- ticular emphasis on the relationship of life history patterns to depth. Ph.D. Thesis, Univ. South Pacific, Suva, Fiji, 258 P- 1984. The species and depth distribution of deepwater cari- dean shrimps (Decapoda, Caridea) near some southwest Pacific islands. Crustaceana 47(2):174-191. King, M. G., and R. B. Moffitt. 1984. The sexuality of tropical deepwater shrimps (Decapoda: Pandalidae). J. Crust. Biol. 4(4):567-571. Moffitt, R. B. 1983. Heterocarpus longirostris MacGilchrist from the North- ern Mariana Islands. Fish. Bull., U.S. 81:434-436. OiSHi, F. 1983. Pacific Tima Development Foundation - State of Hawaii shrimp industry development project. Division of Aquatic Resources, Department of Land and Natural Resources, State of Hawaii, 22 p. Pauly, D. 1982. Studying single-species dynamics in a tropical multi- species context. In D. Pauly and G. I. Murphy (editors), Theory and management of tropical fisheries, p. 33-70. ICLARM and CSIRO, Manila. PoLoviNA, J. J., R. B. Moffitt, S. Ralston, P. M. Shiota, and H. A. Williams. 1985. Fisheries resource assessment of the Mariana Archi- pelago. 1982-85. Mar. Fish. Rev. 47(4):19-25. Polovina, J. J., and S. Ralston. 1986. An approach to yield assessment for unexploited re- sources vrith applications to the deep slope fishes of the Marianas. Fish. Bull., U.S. 84:759-770. Ralston, S. 1986. An intensive fishing experiment for the caridean shrimp, Heterocarpus laevigatus, at Alamagan Island in the Mariana Archipelago. Fish. Bull., U.S. 84:927-934. Struhsaker, p., and D. C. Aasted. 1974. Deepwater shrimp trapping in the Hawaiian Islands. Mar. Fish. Rev. 36(10):24-30. Struhsaker, P., and H. 0. Yoshida. 1975. Exploratory shrimp trawling in the Hawaiian Islands. Mar. Fish. Rev. 37(12):13-21. Western Pacific Regional Fishery Management Council. 1984. Status of fisheries assessment of development and management needs for selected crustacean species in the Western Pacific region. Western Pacific Regional Fishery Management Council, Honolulu, HI, 60 p. Wetherall, J. A., J. J. Polovina, and S. Ralston. In press. Estimating growth and mortality in steady state fish stocks from length-frequency data. In D. Pauly and G. R. Morgan (editors), Length-based methods in fishery research. ICLARM Conf. Proc. Manila, vol. 1. Wilder, M. J. 1977. Biological aspects and fisheries potential of two deep- water shrimps Heterocarpus ensifer and Heterocarpus laevigatus in waters surrounding Guam. M.S. Thesis, Univ. Guam, Agana, 79 p. 349 DECLINE IN ABUNDANCE OF THE NORTHERN SEA LION, EUMETOPIAS JUBATUS, IN ALASKA, 1956-86 Richard L. Merrick, i Thomas R. Loughlin.i and Donald G. Calkins^ ABSTRACT Aerial, ship, and onshore surveys were conducted to assess the abundance of northern sea lions, Eumetopias jubatus, in southwestern Alaska, from the central Gulf of Alaska through the central Aleu- tian Islands, during June-July of 1984-86. Counts of northern sea lions from these surveys were com- pared with counts made in 1956-62 and 1975-79. These data indicated that the number of adults and juveniles onshore declined 52% from 140,000 animals in 1956-60 to 68,000 in 1985— an annual rate of decline of at least 2.7%. Numbers have declined throughout the region, with the greatest declines in the eastern Aleutian Islands (79%) and the least in the central Aleutian Islands (8%). This was not due to emigration because significant increases have not been noted elsewhere. Between the 1960s and mid-1970s, there were large decreases in the eastern Aleutian Islands and western Gulf of Alaska, and a major increase in the central Aleutian Islands. Beginning in the late 1970s declines occurred in all areas. The causes of the declines are unknown, but they may be associated with disease, prey availability or quality, or a combined effect of these and other factors. Factors which may contribute to the declines include the pre-1973 commercial harvests, entanglement of juveniles in marine debris, incidental takes in fisheries, and killing by fishermen. The northern or Steller sea Hon, Eumetopias juba- tus, breeds from the Kuril Islands and Okhotsk Sea through the Aleutian Islands and Gulf of Alaska, and south to California. Loughlin et al. (1984) estimated the maximal population in 1974-80 at 290,000 (in- cluding some pups), of which more than 196,000 were in Alaska. The number of northern sea lions counted in Alaska was unchanged since the surveys of Kenyon and Rice (1961) and Mathisen and Lopp (1963) in 1956-60, even though significant declines had occurred in the eastern Aleutian and Pribilof Islands (Kenyon 1962; Braham et al. 1980). These declines were offset by increases in northern sea lion numbers in the central and western Aleutian Islands (Fiscus et al. 1981). Concern over the decline in northern sea lion numbers in the eastern Aleutian Islands prompted the National Marine Mammal Laboratory (NMML) and the Alaska Department of Fish and Game to conduct surveys in 1984, 1985, and 1986 at sites throughout southwestern Alaska. These included aerial, ship, and onshore surveys of rookeries and major haul-out sites from Kiska Island in the cen- tral Aleutian Islands to the Barren Islands in the central Gulf of Alaska, as well as observations dur- ing two breeding seasons at Ugamak Island, a major rookery in the eastern Aleutian Islands. Together with earlier data for the Aleutian Islands (Kenyon and Rice 1961; Kenyon 1962^; Kenyon and King 1965*; Braham et al. 1980; Fiscus et al. 1981) and for the Gulf of Alaska (Mathisen and Lopp 1963; Calkins and Pitcher 1982^), these data present a 30-yr record of counting northern sea lions, albeit sporadically, in Alaska waters. The objectives of this paper are 1) to report the results of surveys con- ducted between 1984 and 1986, thus describing the current distribution and numbers of northern sea lions in much of Alaska, 2) to compare those counts with the historical data, and 3) to discuss the state of knowledge on causes of the decline in sea lion numbers. 'National Marine Mammal Laboratory, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. ^Alaska Department of Fish and Game, 333 Raspberry Road, Anchorage, AK 99502. 'Kenyon, K. W. 1962. Sea otter studies, population and distri- bution (with notes on Steller sea lion and emperor goose). Un- publ. Rep., U.S. Fish Wildl. Serv., Branch Wildl. Res., Seattle, 47 p. Available from Northwest and Alaska Fish. Cent. Natl. Mar. Mammal Lab., NMFS, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. ^Kenyon, K. W., and J. G. King, Jr. 1965. Aerial survey of sea otters and other marine mammals, Alaska Peninsula and Aleu- tian Islands, 19 April to 9 May 1965. Processed Rep., U.S. Fish Wildl. Serv., Bur. Sport Fish. Wildl., Seattle, 52 p. Available from Northwest and Alaska Fish. Cent. Natl. Mar. Mammal Lab., NMFS, NOAA, 7600 Sand Point Way, N.E., Seattle, WA 98115. ^Calkins, D. G., and K. W. Pitcher. 1982. Population assess- ment, ecology and trophic relationships of Steller sea lions in the Gulf of Alaska. Alaska Dep. Fish and Game, Final Rep. RU243, 128 p. Alaska Department of Fish and Game, 333 Raspberry Road, Anchorage, AK 99502. Manuscript accepted December 1986. FISHERY BULLETIN: VOL. 85, NO. 2, 1987. 351 FISHERY BULLETIN: VOL. 85. NO. 2 STUDY REGION AND METHODS Study Region The study included northern sea lion hauling sites in southwestern Alaska from Kiska Island in the Aleutian Island chain eastward to the Barren Islands in the central Gulf of Alaska (Fig. 1). This region was subdivided for analysis into four areas: 1) central Gulf of Alaska, 2) western Gulf of Alaska, 3) eastern Aleutian Islands, and 4) central Aleutian Islands. Two general types of northern sea lion sites on land were recognized— rookeries and haul-outs (Loughlin et al. 1984). Rookeries were areas where adult males actively defended territories and most females gave birth and mated. Haul-outs were sites where few pups were present and where little breeding took place. Some islands included more than one distinct rookery and haul-out. A total of 114 sites, of which 28 were rookeries (on 27 islands), were surveyed during 1984-86. Ugamak Island was a site for NMML field studies during the northern sea lion breeding seasons in 1969, 1977, 1978, 1985, and 1986. The island is located in the eastern Aleutian Islands (long. 164°50'W, lat. 54°14'N), about 110 km east of Dutch Harbor, AK. The island contained the largest aggre- gation of breeding sea lions in the Aleutian Islands as late as 1969. Survey Methods Aerial photographic surveys of northern sea lion rookeries and haul-outs in the eastern Aleutian Islands area (Fig. 1) were conducted 7-12 July 1984, using a Bell 205*^ helicopter flown off of the NOAA ship Surveyor. A survey of the entire study region 1 Central ■ Gulf of Alaska 4 Central Aleutian Islands 3 Eastern Aleutian Islands O Western Gulf of Alaska Figure 1.— The four Alaskan study areas (left) and 28 northern sea lion rookery sites (right) counted during 1984-86. Rookery island name and number as in Table 2. 352 MERRICK ET AL.: DECLINE OF NORTHERN SEA LION was conducted during 9-13 June 1985, using a Grum- man Widgeon for the eastern Aleutian Islands and Gulf of Alaska and a Piper Navajo for the central Aleutian Islands. All surveys were conducted be- tween the hours of 1000 and 1800 Alaska Daylight Saving Time (ADT) and from mid-June to mid-July, when the most adult and juvenile sea lions were expected to be onshore (Withrow 1982). Survey methods were those of Braham et al. (1980). A shipboard survey also was conducted of rook- eries and major haul-outs from Ugamak Island to Kiska Island between 25 June and 15 July 1985 (Loughlin et al.'^). Weather permitting, observers landed at each site and counted the number of north- ern sea lions present and the number of animals en- tangled in debris. Pups were counted by first walk- ing through rookeries to drive off the adult and juvenile animals, and then returning to count the pups. This survey was timed to occur after most pups had been born, and before they had begun to enter the water (Withrow 1982). Pups were counted at northern sea lion rookeries in the central and western Gulf of Alaska between 3 July and 9 July 1984 (Calkins^) and between 29 June and 10 July 1986. Access was provided by a skiff launched from a larger vessel or by helicopter. ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 'Loughlin, T. R.. P. J. Gearin. R. L. DeLong, and R. L. Merrick. 1986. Assessment of net entanglement on northern sea lions in the Aleutian Islands, 25 June-15 July 1985. Processed Rep. 86-02, 50 p. Northwest and Alaska Fisli. Cent. Natl. Mar. Mammal Lab., NMFS, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. *Calkins, D. G. 1985. Draft final report. Steller sea lion pup counts in and adjacent to Shelikof Strait. Submit, to North Pac. Fish. Manage. Coun., March 8, 1985, 13 p. Alaska Department of Fish and Game, 333 Raspberry Road, Anchorage, AK 99502. 2 Alaska /"/ 20. Peninsula ^^ p^V Shumagin Is. 17 Unalaska I. Unimak \. 19 14 15 19 Ugamak 20 Sea Lion Rock 21 Clubbing Rocks 22 Pinnacle Rock 23 Chernabura 24 Atkins 25 Chowiet 26 Chirikof 27 Marmot 28 Sugarloaf 353 FISHERY BULLETIN: VOL. 85, NO. 2 Pups were counted by driving off the adult animals just ahead of the counting team. Counts of northern sea lions were made daily by observers on Ugamak Island between 1 June and 3 July 1985 and between 16 June and 26 July 1986. Hourly counts were made between the hours of 0700 and 2400 ADT for 6 days in 1986. Animals were counted from the cliffs above the sites, using 7 x 35 binoculars, a 15-60 power spotting scope, and unassisted vision. Counts were made of animals ac- cording to five types: adult territorial male, other adult male, adult female, juvenile, and pup (Merrick 1984). Animals in the water were excluded from counts. Freshly dead pups were recorded when seen, with pup mortality estimated as the total number of dead pups divided by the maximal number of pups counted (living and dead). Data Analysis The number of northern sea lions counted in the 1984-86 surveys were compared with counts from surveys in 1956-79 conducted by Kenyon and Rice (1961), Kenyon (fn. 3), Mathisen and Lopp (1963), Braham et al. (1980), Fiscus et al. (1981), and Calkins and Pitcher (fn. 5). Differences in survey areas complicated comparisons for the entire region, so comparisons were generally performed by area. Comparisons also were complicated by the differ- ences in the counting methods used. Some were counts from land (Fiscus and Johnson 1968^; Fiscus et al. 1981; Withrow 1986io), while others were estimates from ships (Fiscus and Johnson fn. 9; Calkins and Pitcher fn. 5), and counts from aerial photographs (Mathisen and Lopp 1963; Braham et al. 1980; Calkins and Pitcher fn. 5). The most ac- curate were visual counts from land and from aerial photos (Withrow 1982); these were the methods used in the 1984-85 aerial surveys. Several assumptions were made in the analysis of these data. The first was that all sites with more than a few animals were surveyed in 1984-85. Sec- ond, the dates and times of peak seasonal and daily abundance were considered to have remained con- stant throughout the 30-yr period. The 1984-85 surveys and those conducted by Braham et al. (1980) were scheduled to coincide with these peaks; where- as, those of Kenyon and Rice (1961) and Kenyon (fn. 3) were conducted in the spring and without regard to time of day. A smaller proportion of animals were probably onshore in the spring than in the summer (Mathisen and Lopp 1963; Braham et al. 1980). Third, the proportion of the population onshore was assumed to have remained unchanged in the 30 years of counting. Finally, double counting was con- sidered to be negligible in all the surveys because large areas were surveyed in a single day. Counts presented here are indices of population size because they exclude animals at sea and because it is difficult to count at the exact time peak num- bers are ashore. There are few data on the propor- tion of animals that are at sea at the time the peak number is onshore. Consequently, it was necessary to assume that the proportion had not changed over time. Even during the period when maximal num- bers of animals were expected onshore there was variation due to weather and tidal affects (Withrow 1982; Merrick 1984), so that it was unlikely that a survey would occur on the day and time of peak numbers. However, because the sites on Ugamak Island were counted daily, the maximal number counted there was a closer approximation than the aerial survey counts of the actual peak number of animals onshore during the breeding season. Thus the Ugamak Island data were used to determine if seasonal and daily variation in northern sea lion hauling patterns had changed and to assess the potential amount of error (due to counting at the wrong time) in the aerial photo counts. Rates of decline between two points in time were calculated using the formula where A^o t d = N, = Nod< count in base year count in future year t number of years between the base year and year t rate of change, with the percent an- nual change calculated as {d - 1) x 100. 'Fiscus, C. H., and A. M. Johnson. 1968. Site for research on the Steller sea Hon, June-July 1968. Processed Rep., U.S. Fish Wild). Serv., Bur. Commer. Fish., Mar. Mammal Biol. Lab., Seattle, 33 p. Northwest and Alaska Fish. Cent. Natl. Mar. Mam- mal Lab., NMFS, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. '"D. Withrow, Northwest and Alaska Fisheries Center National Marine Mammal Laboratory, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115, pars, comm. January 1986. Area counts were regressed as a linear function of time to determine if trends in population sizes ex- isted. Student's ^test was used to assess the signif- icance of the regressions. Wilcoxon's signed rank test was used for between year comparisons of paired site counts within an area (Hollander and Wolfe 1973). 354 MERRICK ET AL.: DECLINE OF NORTHERN SEA LION RESULTS 1984-86 Survey Findings The 1984 survey of the eastern Aleutian Islands resulted in a count of 9,833 adult and juvenile north- ern sea lions on 16 sites (Table 1). The six rookery islands surveyed included 7,934 animals (Table 2), 91% of the total count. One rookery, Adugak Island, and several haul-out sites were not surveyed owing to inclement weather. A total of 10,802 adult and juvenile northern sea lions were seen in the 1985 survey of the eastern Aleutian Islands (Table 1), which was not significant- ly different from the 9,833 animals counted there in 1984 (P > 0.05). A total of 67,617 animals were counted at 105 sites in the entire study region. Most (60%) of the animals were associated with the 27 rookeries surveyed; the largest rookeries were in the central Gulf of Alaska, notably at Marmot Island, and in the Central Aleutian Islands. The rookery on Semisopochnoi Island was not surveyed; eight previously identified haul-out sites were un- occupied. Observers in the 1985 shipboard survey counted 4,950 pups at six rookeries in the eastern Aleutian Islands and 9,170 pups at nine rookeries in the cen- tral Aleutian Islands (Table 3). In the 1984 and 1986 pup surveys of six rookeries in the Gulf of Alaska Table 1 .—Counts and percent declines of adult and juvenile north- ern sea lions at all sites in spring and summer 1956-85 in the Aleutian Islands and Gulf of Alaska. Reg ion Central Western Eastern Central Gulf of Gulf of Aleutian Aleutian Refer- Year Alaska Alaska Islands Islands ence^ 1956 ^24,320 1 1957 35,150 1 1959 28,115 2 1960 ^52,530 2 1962 31,040 3 1975 21,221 4 1976 30,677 9,480 22,142 4,5 1977 23,922 4 1978 14,917 5 1979 ^41,677 6 1984 9,833 7 1985 24,389 6,667 10,802 25,759 7 Decline'' Overall -31% - 73% - 79% -8% Annual -1.3% - 4.4% -6.1% - 0.3% 'Reference: 1— Mathisen and Lopp (1963); 2— Kenyon and Rice (1961); 3— Kenyon (text fn. 3); 4— Braham et al. (1980); 5— Calkins and Pitcher (text fn. 5); 6— Fiscus et al. (1981); 7— this study. ^Significant difference (P < 0 05) from 1985 using W/ilcoxon signed rank test. ^Declines calculated from earliest survey date (Table 3) 16,278 (excluding Clubbing Rocks and Pinnacle) and 12,025 pups were counted, respec- tively. Trends in Regional Numbers Comparison of the 1984-85 aerial surveys with historical data (Table 1) shows that significant (P < 0.05) declines have occurred in northern sea lion numbers in the western Gulf of Alaska ( - 73%), and eastern Aleutian Islands ( - 79%). The central Gulf of Alaska (-31%) and central Aleutian Island popu- lations may have also declined since 1957-59, though the decreases ( - 31% and - 8%, respectively) were not statistically significant (P > 0.05). Note that the central Aleutian Islands numbers increased 34% between 1959 and 1979. This suggests that either the population increased markedly, was supple- mented by immigration from other areas (e.g., the eastern Aleutian Islands), or was an artifact of the 1979 survey methodology (i.e., a shipboard survey). Linear regression models (Fig. 2) fitted to these counts indicate that the trends of all areas, other than the central Aleutian Islands, exhibit significant negative slopes (P < 0.05). The number of adult breeding animals has de- clined in all areas since 1957 except the central Aleu- tian Islands (Table 2). Numbers at the central Aleu- tian Islands rookeries increased by 88% between 1959 and 1985, a significant increase (P < 0.01). As the overall population in the central Aleutian Islands decreased between 1959 and 1985, this may indicate that a larger proportion of the population is now breeding than in the past. Rookery populations in the eastern Aleutian Islands have declined 79% since 1957, a significant decline (P < 0.05). A loss of 15,000 animals occurred at the Ugamak and Akutan Island rookeries alone between 1968 and 1975. Numbers at rookeries in the western and central Gulf of Alaska decreased 66% and 47%, respective- ly, between 1956-57 and 1985. Trends in Regional Pup Production Pup counts are available for only a few sites prior to 1984. These data (Table 3) show that the number of northern sea lion pups counted in the central Gulf of Alaska decreased between 1979 and 1986 by 44%, from 18,998 to 10,600. The number of pups has also declined at sites in the western Gulf of Alaska and eastern Aleutian Islands. Pupping decreased 52% at Bogoslof Island between 1968 and 1985, 89% at Walrus Island (in the Pribilof Islands) between 1960 355 FISHERY BULLETIN: VOL. 85, NO. 2 Table 2. — Counts of adult and juvenile northern sea lions on individual rookeries for selected surveys from spring and summer 1956 to 1985 in the Aleutian Islands and Gulf of Alaska. Year Island' 1956-59^ 1960-62^ 1968" 1976-79^ 1984^ 1985^ Central Aleutians: Kiska (1, 2) 1,000 600 7,155 3,066 Ayugadak (3) 600 1,005 1,463 702 Amchitka (4) 600 1,515 1,943 728 Semisopochnoi (5) 2,500 3,700 1,223 nc Ulak (6) 1,500 550 3,068 2,729 Tag (7) 400 200 1,740 944 Gramp (8) 700 0 2,235 1,290 Adak (9) 0 0 972 964 Kasatochi (10) 200 2,000 2,166 1,170 Agiigadak (11) 250 3,000 993 514 Seguam (12) 25 1,275 6,493 2,942 Yunaska (13) 800 110 2,249 1,071 Total ^8,575 13,955 ^31,700 16,120 Eastern Aleutians: Adugak (14) 1,275 1,000 nc« 1,177 nc 955 Ogchul (15) 2,966 2,000 nc 1,109 712 547 Bogoslof (16) 2,136 1,100 3,310 3,308 1,379 1,287 Akutan (17) 9,275 15,720 10,316 4,019 2,533 1,710 Akun (18) nc 2,000 1,900 1,050 760 435 Ugamak (19) 14,536 13,400 13,553 4,760 1,252 1,429 Sea Lion Rock (20) 2,871 2,000 nc 2,076 1,298 538 Total 33.059 ''37,220 29,079 '17,499 7,934 6,901 Western Gulf: Clubbing Rocks (21) 1,556 2,663 1,251 Pinnacle Rock (22) 3,142 3,692 1,588 Chernabura (23) 4,806 2,758 487 Atkins (24) 4,995 3,943 1,562 Total 14,499 13,056 4,888 Central Gulf: Chowiet (25) 6,014 4,441 2,059 Chirikof (26) 1,695 5,199 2,346 Marmot (27) 3,866 6,381 4,983 Sugarloaf (28) 1 1 ,963 4,374 2,991 Total 23,538 20,395 12,379 'Rookery island name and code (within parentheses) as in Figure 1. ^Mathisen and Lopp (1963). 'Kenyon (text fn. 3), for central Aleutian Islands in 1962; Kenyon and Rice (1961), for eastern Aleutian Islands in 1960 ■•Fiscus and Johnson (text fn. 9). sFiscus et al. (1981), for central Aleutian Islands in 1979: Braham et al. (1980), for eastern Aleu- tian Islands in 1976; Calkins and Pitcher (text fn. 5), for western Gulf of Alasl 60 c ■t-' c 50 o (J 40 30 20 - Central Gulf of Alaska Count = 36,016 - 347.7 (year - 1955) r^ = 0.924 p <0.05 Central Aleutian Islands Count = 27,340 + 170.6 (year - 1955) p>0.05 10 0 Western Gulf of Alaska Count = 24,385 - 653.8 (year - 1955) r^ =0.971 p< 0.05 J I '2 =0.068 Eastern Aleutian Islands Count = 57,716 - 1701.2 (year - 1955) r^ = 0.973 p<0.01 A J I I I I 1956 61 66 71 76 81 85 1956 61 66 71 76 81 85 Year Figure 2.— Trends in total adult and juvenile northern sea lion abundance by area in Alaska, for spring and summer surveys conducted from 1956 to 1985. empty. Although some rookeries have been reduced in size and decreases have been observed in all age and sex groups, the ratio of adult females to terri- torial males has remained relatively constant be- tween 1977-78 (14.4-17.3) and 1986 (16.4). Between 1985 and 1986, however, the number of adult males remained constant, while the number of females decreased by 26%. Island totals of juveniles were similar in 1985 (87 animals) and in 1986 (110 animals). However, an analysis of comparable areas of the island (i.e., N1-N3 and A2-A4; Fiscusi^) indicated that the juvenile portion of the population had decreased significantly (x^ = 4.09, P < 0.05) from 9.0% in 1977 to 1.4-1.6% in 1985-86. A comparably low rate (2%) has only been observed at Ano Nuevo Island (Gentry 1970) and that rate was calculated for a single primary rookery rather than a whole island. Rates such as the 22% observed at Marmot Island in 1983 or the 25% observed for British Columbia in 1956 seem more typical for northern sea lion rook- eries (Pike and Maxwell 1958; Merrick 1984). Counts of pups were available for six rookeries '^Fiscus, C. H. 1970. Steller sea lions at Ugamak Island, Aleu- tian Islands, Alaska, June 1969. Unpubl. Rep., U.S. Fish Wildl. Serv., Bur. Commer. Fish., Mar. Mammal Biol. Lab., Seattle, 78 p. Northwest and Alaska Fish. Cent. Natl. Mar. Mammal Lab., NMFS, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. 358 MERRICK ET AL.: DECLINE OF NORTHERN SEA LION on the island for 1969, 1977, 1985, and 1986 (Fiscus fn. 12; Withrow fn. 10). The number of pups declined by 72% between 1969 and 1977, with two rookeries abandoned. A decline of 35% occurred between 1977 and 1986, with three more rookeries abandoned. Pupping has probably declined on other parts of the island, because the number of breeding animals decreased at other rookeries (Table 4) where pups were not counted in 1968 and 1977. Between 1985 and 1986, the only years with complete island surveys, there was a decline of 18% in pup numbers. Despite the decline in the number of pups born, the ratio of pups to adult females has increased from 0.75 in 1968 and 0.73-0.81 in 1977-78 to 0.95-1.06 in 1985-86. The Ugamak Island pup mortality dur- ing the first 1-2 mo postpartum in 1985-86 was 3.4-4.5%. Median pupping dates were similar in 1977-78 (13 June) and 1985 (12 June). Finally, the Ugamak Island data allowed an evalu- ation of the times of peak abundance and variability in the counting methods. Counts peaked in 1985-86 between the third week of June and the first week of July, with 90% of the maximal breeding season population onshore at midday during this period. Hourly counts between 1000 and 2000 ADT during this period were always within 10% of the day's maximum. These occupancy patterns were the same as those observed by Withrow (1982) at Ugamak Island in 1977-78. The Ugamak Island data also were used to pro- vide an estimate of the accuracy of the 10 June 1985 aerial survey of the island. There was only a 3% difference between the aerial photo count and the simultaneous ground count. Thus, as noted by Withrow (1982), aerial photo counts accurately reflect ground counts. Despite this accuracy, the aerial photo count of 10 June 1985 was 24% lower than the maximal ground count of 25 June 1985. This bias was likely caused by the aerial survey oc- curring slightly before the period of peak abundance. DISCUSSION The number of northern sea lions found at sites within the study region, which included at least 140,000 animals circa 1958, totaled less than 68,000 in 1985-a decline of 52% (-2.7% per year). All in- dicators (regional numbers, breeding animals on rookeries, pup production, and the Ugamak Island data) confirm this decline. The rates of decline are probably underestimated because declines probably did not begin until after 1958. For example, the Ugamak Island population showed no decline be- tween 1957 and 1969, and then declined at least 10.1% per year between 1969 and 1986. This in- dicates that the eastern Aleutian Island northern sea lion population may have begun to decline in the early 1970s rather than in 1958, the base year for rate calculations. Data are insufficient to calculate regional or area rates from 1969 to later dates. Declines may have occurred in two phases. The first phase may have begun in the 1970s and been confined to the eastern Aleutian Islands and west- ern Gulf of Alaska. Numbers for the entire study region fell by 25% (-1.6% per year) between 1958 and 1977. Numbers in the eastern Aleutian Islands appeared to stabilize in the mid-1970s, while those in the central Aleutian Islands and western Gulf of Alaska may have increased. A second phase of the decline may have begun during the late 1970s, with all areas being affected and overall numbers falling 36% (-5.2% per year) between 1977 and 1985. Results of the 1986 pup survey and 1986 Ugamak Island study indicate that the decline is still con- tinuing. Alternative Explanations for the Declines in Northern Sea Lion Numbers Consideration has been given as to whether the declines could be explained by counting errors or biases, changes in northern sea lion behavior, or emigration. However, errors in counting or, for that matter, changes in technique do not explain the decline in numbers. Because counts taken in the 1950s and 1960s were conducted in the spring (before abundance had peaked on land) and because sites were missed, we believe that they underesti- mated sea lion numbers. Braham et al. (1980) esti- mated that only 42% of the eastern Aleutian Island sites were surveyed in 1957 (Mathisen and Lopp 1963). Animals present were probably accurately counted in the pre-1970 surveys because observers were all experienced. Kenyon and Rice (1961) com- pared their visual counts to concurrent aerial photos taken and found that their error was between 6 and 10%. Even if they had overestimated numbers the error may have been counterbalanced by any under- estimate from counting too early in the year. The counts of the 1970s and 1980s were probably more comparable because they were made during the period of peak numbers onshore. Also there was little variation during this period in methods and personnel were experienced in these survey tech- niques. Furthermore, the methods used were be- lieved to have been the most accurate. Even if all of the aerial counts were lower than the maximal 359 FISHERY BULLETIN: VOL. 85, NO. 2 count for the season (as observed at Ugamak Island in 1985), the amount of error would be insufficient to explain the low counts. In any event, the declines in northern sea lion numbers observed in aerial surveys since the 1970s have been confirmed by the counts made from land at Ugamak Island and by the pup counts taken in the eastern Aleutian Islands and the Gulf of Alaska. Pup counts can provide a reliable index of popu- lation change since almost all pups born can be counted in surveys scheduled prior to the pups going to sea. However, the index may be biased (Berkson and DeMaster 1985). Briefly stated, if precount pup survival (e.g., live birth rates) is density-dependent the counts would overestimate the rate of decline. If postcount pup survival (e.g., juvenile survival) is density-dependent then the counts would underesti- mate the rate of decline. Finally, if adult survival is density-dependent there would be little bias. Few data exist on density-dependent population regula- tion in northern sea lions, so we cannot be sure which, if any, of these mechanisms are operative. However, available data suggest that while precount survival in recent years is either unchanged or has improved, postcount survival has decreased. In both cases, the effect would be for pup counts to under- state the rate of decline. It is unclear whether northern sea lion hauling behavior has changed sufficiently to affect the counts. The 1977-78 and 1985-86 Ugamak Island data show that seasonal and daily hauling patterns and the timing of critical events (e.g., the median pupping date) have not changed. Similar data are not available for the earlier surveys. Animals may have dispersed to other sites, but they still would have been counted in the regional surveys. There is some evidence from Ugamak Island that females may spend less time ashore now than before (e.g., the high ratio of pups to adult females), which would decrease the number of adult animals counted. The low number of juveniles counted at Ugamak Island in 1985-86 may simply reflect increased juvenile dispersal away from the site due to changing prey resources and earlier weaning. The decline in northern sea lion numbers onshore in the region has not been due to the emigration of animals to other regions because significant in- creases have not been noted elsewhere. Numerical decreases have been noted at the western extent of the breeding range in the Kuril and Commander Islands (Perlov 1982; Kuzin et al. 1984; Chelnokov 1984). Numbers began to fall in this region circa 1972, and by 1981-82 had fallen by 50% or more. Abundance at sites in the western Aleutian Islands declined 34-61% from 1979 to 1985 (Kletti=^). Adult northern sea lions at the Pribilof Islands have de- clined in number from approximately 7,000 animals in 1960 to 1,100 in 1981 (Kenyon 1962; Loughlin et al. 1984). Most of these animals were located at Walrus Island which was occupied by 4,000-5,000 northern sea lions in 1960, around 1,500 in 1975, and only 600 in 1982 (Kenyon 1962; NMML fn. 11). In 1960, 2,866 pups were counted at Walrus Island, but this number had fallen to 334 by 1982 (Kenyon 1962; NMML fn. 11). Northern sea lions are regular- ly seen farther north at St. Matthew and St. Law- rence Islands during the ice-free season, but the total has rarely exceeded 300 animals (Loughlin et al. 1984). Frost et al.^* estimated 2,000 animals were in northern Bristol Bay during the 1970s, with 1,500 observed in the summers of 1980-82. The only rookery in Bristol Bay, Sea Lion Rock (Table 1), has decreased in size by 81% since the counts of 1957 reported by Mathisen and Lopp (1963). The 1982 count of 7,962 animals in southeast Alaska was not substantially different from the 1973 estimate of 8,430 (Calkins and Pitcher^^). The number of north- ern sea lions in British Columbia and in the lower United States (i.e., Washington, Oregon, and Calif- ornia) do not appear to be increasing from 5,700 and 5,410 animals, respectively (Loughlin et al. 1984; Bigg 1985). Proximate Causes of the Decline The declines of northern sea lion could either be due to falling reproductive rates or reduced survival of pups, juveniles, and adults (especially females). There does not, however, appear to have been sig- nificant declines in reproductive rates or in pup sur- vival 1-2 mo postpartum. The pregnancy rate of females taken in the Gulf of Alaska during April- May 1985 was 62% (n = 62), which was not signif- icantly different from the 67% (n = 102) found there in 1975-78 (x^ = 0.002, P > 0.50; Pitcher and Calkins 1981; Goodwin and Calkins^^). In addition. "E. Klett, U.S. Fish and Wildlife Service, Aleutian Island Unit, Alaska Maritime National Wildlife Refuge, Box 5251 Naval Air Station, Adak, AK 98791, pers. comm. March 1986. '"Frost, K. J., L. F. Lowry, and J. J. Burns. 1982. Distribution of marine mammals in the coastal zone of the Bering Sea during summer and autumn. Alaska Dep. Fish Game, Final Rep. RU613, 188 p. Alaska Department of Fish and Game, 1400 College Road, Fairbanks, AK 99701. '^Calkins, D. G., and K. Pitcher. 1983. 1982 pinniped inves- tigations in southern Alaska. Unpubl. rep., 15 p. Alaska Depart- ment of Fish and Game, 333 Raspberry Road, Anchorage, AK 99502. '"Goodwin, E. A., and D. G. Calkins. 1985. Preliminary results of ongoing investigations of San Miguel Sea Lion Virus, Lepto- spirosis, and Chlamydiosis in Alaska Steller sea lions and their rela- 360 MERRICK ET AL.: DECLINE OF NORTHERN SEA LION the ratio of pups to adult females at Ugamak Island has increased since 1968 and 1977-78. The 3.4-4.5% rate of pup mortality at Ugamak Island was low when compared with the 11% found at Marmot Island, 10% at Ano Nuevo Island, CA, and the 13-14% at Wooded Island, AK (Gentry 1970; Sande- gren 1970; Merrick 1984). The declines of northern sea lion may be due to reduced survival of pups (after they go to sea), juveniles, and adults. Changes in survival rates are difficult to assess. The precipitous declines in pup- ping in the Gulf of Alaska (Table 3) may indicate that there have been large declines in the female popula- tion there. Numbers of adult females at Ugamak Island also declined between 1985 and 1986. The Ugamak Island data also seem to indicate that juvenile abundance was much lower in 1985-86 than at other sites and in other years, which may indicate unusually high mortality is occurring after pups leave the rookery. Investigation of the declines in juveniles and adult females may hold the greatest promise for further study. Ultimate Causes of the Decline The causes of the reduced fecundity or survival of northern sea lions are presently unknown, but there are several possibilities— disease, changes in prey resources, and the combined effects of these and other factors. Disease and prey limitations are particularly plausible causes of the decline because of their potential for widespread impacts (hence declines in other regions) and because they could be implicated in the apparent declines of other Bering Sea and North Pacific Ocean pinniped populations. The number of northern fur seals, Callorhinus ursinus, breeding at the Pribilof Islands and on Robben Island in the Sea of Okhotsk have de- creased since the mid- to late 1970s (Fowler 1985). Since the 1970s, harbor seal, Phoca vitulina richardsi, numbers may have decreased in Bristol Bay (Pitcher 1^) and have declined substantially in the central Gulf of Alaska at Tugidak Island (Calkins and Pitcher^^). Diseases resulting in reproductive failures and neonate, juvenile, and adult mortality could be a significant source of mortality. Antibodies to two types of bacteria (Leptospira and Chlamydia) and one marine calicivirus virus (San Miguel Sea Lion Virus) which could produce such mortality were present in blood taken from northern sea lions in Alaska (Fay et al. 1978^9; Goodwin and Calkins fn. 16; Barlough et al. in press). Leptospires are spiro- chete bacteria and are suspected agents of abortion and adult mortality in California sea lions, Zalopkus calif ornianus, (Smith et al. 1974a) and in northern fur seals (Smith et al. 1974b). San Miguel Sea Lion Virus may also be associated with reproductive failures or neonatal deaths in California sea lions and northern fur seals, although the evidence is limited (Smith et al. 1973). Chlamydia has not been studied previously in sea lions. These and other agents are being examined for their possible adverse effects on northern sea lion populations. The decline in northern sea lion numbers may be related to changes in the quantity and size of their prey. The few studies of the food habits of northern sea lions indicated that their primary prey are wall- eye pollock, Theragra chalcogramma, in the Bering Sea, Gulf of Alaska, and North Pacific Ocean (Klumov 1957; Pitcher 1981; Calkins et al.^o). This fish is also a major prey item of harbor seals and northern fur seals (Pitcher 1980; Kajimura 1984). Walleye pollock biomass in the eastern Bering Sea rose from less than 5 million metric tons (t) in the 1960s to a peak of over 13 million t in the early 1970s and has since declined to about 8 million t in 1985 (Bakkala et al. in press). While the population bio- mass remains high, sporadically low abundance of age-1 walleye pollock between 1979 and 1984 could mean that in some years (e.g., 1981, 1982, and 1984) there would be fewer fish in the 10-35 cm range (Bakkala et al. in press). This size range includes the mean sizes consumed by northern sea lions and har- bor seals (Pitcher 1981; Frost and Lowry 1986). Declines in abundance and increases in fish length have also been noted since 1981 in the Shelikof Strait region of the Gulf of Alaska (Nelson and Nun- nallee 1985). However, there are few data on north- ern sea lion foraging patterns in the Bering Sea and tionship to declining pup counts. Presented at the Sixth Biennual Conference on Biology of Marine Mammals, Vancouver, B.C., Nov. 1985. Alaska Department of Fish and Game, 333 Raspberry Road, Anchorage, AK 99502. "K. Pitcher, Alaska Department of Fish and Game, 333 Rasp- berry Road, Anchorage, AK 99502, pers. comm. February 1986. '^Calkins, D. G., and K. Pitcher. 1985. Pinniped investigations in southern Alaska: 1983-84. Unpubl. rep., 19 p. Alaska Depart- ment of Fish and Game, 333 Raspberry Road, Anchorage, AK 99502. I'Fay, F. H., R. A. Dieterich, L. M. Shults, and B. P. Kelley. 1978. Morbidity and mortality in marine mammals. U.S. Dep. Commer. and U.S. Dep. Inter. (OCSEAP), Environ. Assess. Alaska Cont. Shelf, Ann. Rep., Mar. 1978, 1:39-79. ^oCalkins, D. G., G. A. Antonelis, Jr., and G. W. Oliver. 1981. Preliminary report of the Steller sea lion/ice seal research cruise of the ZRS Zvyagino. Unpubl. rep., 22 p. Alaska Department of Fish and Game, 333 Raspberry Road, Anchorage, AK 99502. 361 FISHERY BULLETIN: VOL. 85, NO. 2 Gulf of Alaska, so a relationship between the de- clines in northern sea lion numbers and changes in the abundance of their prey cannot be rejected or confirmed. The declines of northern sea lion may not have a single cause, but may be due to the effects of a combination of these and other factors. Sources of mortality which alone seem insufficient to account for the declines but which could be important in a combined effect include the pre- 1973 commercial sea lion harvests, entanglement in marine debris, inci- dental taking in fisheries, and the killing of sea lions for bait and predator control. Northern sea lions were commercially harvested in the eastern Aleutian Islands and Gulf of Alaska from 1959 to 1972. Six hundred and sixteen adult males were taken in an experimental harvest in 1959 (Thorsteinson et al. 1961). A total of 45,178 north- ern sea lion pups of both sexes were harvested in the eastern Aleutian Islands and Gulf of Alaska be- tween 1963 and 1972 (FEIS^i). The largest harvests were conducted between 1963 and 1972 at Sugar- loaf and Marmot Islands where 16,763 and 14,180 pups, respectively, were killed, and between 1970 and 1972 at Ugamak and Akutan Islands where 3,773 and 6,036 pups, respectively, were killed. The pup harvests, which sometimes reached 50% of the total pup production from a rookery (e.g., at Sugar- loaf Island in 1965 and 1968), could have depressed recruitment in the short term. This may partially explain the declines experienced at some sites through the mid-1970s. However, it is unclear why numbers declined in areas where no harvest oc- curred (e.g., the north side of Ugamak Island), while no declines were observed at some harvest sites (e.g.. Marmot Island). In any event, those harvests should not currently be affecting the decline, be- cause populations should have stabilized 3-5 years after the cessation of harvesting as unharvested year classes reached breeding age. Furthermore, these harvests probably cannot explain the declines in numbers counted in the western and central Aleu- tian Island populations. Little information exists on the effect of entangle- ment in marine debris on northern sea lions. Despite debris commonly being found in areas northern sea lions frequent (Calkins 1985; Merrell 1985), data from NMML surveys suggest that this is not a prob- lem, at least for adult sea lions. Observed entangle- ment rates were 0.07% in the 1985 ship survey (Loughhn et al. fn. 7), 0.09-0.17% in the 1985-86 Ugamak Island surveys, and 0.12% at Marmot Island in 1983 (Merrick 1984). Numerous northern sea lion pups were seen in the November 1985 east- ern Aleutian Island entanglement survey (Lough- lin et al. fn. 7), but none were entangled. Never- theless, it is possible that entangled northern sea lion pups drown and are not observed. Numerous northern sea lions have been taken in- cidental to fisheries in the Bering Sea and North- east Pacific Ocean since the late 1960s and early 1970s (FEIS fn. 21). In 1978-81 the estimated aver- age annual mortality for all foreign vessels was 724 animals (Loughlin et al. 1983). This does not, how- ever, include animals taken by U.S. fishermen fish- ing in joint ventures or independently. Loughlin and Nelson (1986) documented the take in the Shelikof Strait joint venture walleye pollock fishery where an estimated 958 to 1,436 northern sea lions were caught by U.S. trawlers in 1982. This take declined to less than 400 animals per season in 1983 and 1984, probably due to changes in fishing technique and the area and times fished. The cumulative im- pact of foreign independent and joint venture fish- eries in the Bering Sea and North Pacific Ocean probably now accounts for less than 500 deaths per year (NMML fn. 11). Domestic fishermen now work- ing independently probably take less since they generally are involved in fisheries that catch few sea lions. However, as foreign fishing is phased out of U.S. waters, the domestic take will probably in- crease. The foreign and domestic incidental take contributes to but cannot totally account for the decline. We are uncertain how the killing of northern sea lions by fishermen has affected the population. Fish- ermen have been observed to kill adult animals at rookeries, haul-outs, and in the water near boats, but the magnitude of this take is generally unknown. Trawl fisheries attract many northern sea lions dur- ing haulback operations and shooting at these animals is a common occurrence. One of the few estimates of shooting mortality comes from Matkin and Fay22 who calculated that 305 northern sea lions were killed directly (shot) while interferring with fishing operations in the spring 1978 Copper River Delta salmon gill net fishery. Northern sea lions at 2>Final environmental impact statement (FEIS). 1977. Con- sideration of a waiver of the moratorium and return of manage- ment of certain marine mammals to the State of Alaska. Vol. II. U.S. Dep. Commer. and U.S. Dep. Inter., Interagency Task Group, Wash., D.C., 251 p. 22Matkin, C. 0., and F. H. Fay. 1980. Marine mammal-fishery interactions on the Copper River and in Prince William Sound, Alaska, 1978. Final rep. for contract MMC-78/07 to Mar. Mam- mal Comm., 71 p. Available from National Technical Information Service, U.S. Department of Commerce, Springfield, VA 22161, as PB80-159536. 362 MERRICK ET AL.: DECLINE OF NORTHERN SEA LION sites in the eastern Aleutian Islands also would have been prime sources of bait for crab fishermen. Thus it may be more than a coincidence that the onset of the northern sea lion decline in the eastern Aleu- tian Islands began at the time of peak landings in the Bering Sea king crab (Paralithodes spp., Lithodes aequispina) and Tanner (snow) crab {Chio- noecetes spp.) fisheries. Killing "nuisance" northern sea lions continues to this date (R. L. Merrick pers. obs.). This killing may have a significant effect on local populations (e.g., the eastern Aleutian Islands and central Gulf of Alaska) and might have caused animals to disperse away from traditional rookeries and haul-outs. It should have little effect, however, in areas that have not been heavily fished (e.g., the western and central Aleutian Islands). Sources of mortality that we think are of minor or unknown importance in the decline include changes in oceanographic or climatic conditions, in- creased predation, harassment, subsistence har- vests, and chemical pollutants. Prospects for the Future Many pinniped species have experienced popula- tion declines within recent history, and in most cases the population has been able to rebuild. Overex- ploitation has been a cause of long-term but tem- porary declines in many species, including Southern Hemisphere fur seals (Arctocephalus spp.), elephant seals (Mirounga spp.), and northern fur seals (Bon- ner 1982). Other human activities have caused declines, such as that of ringed seals, Phoca hispida, in the Baltic Sea, where organochlorines may have caused a high rate of reproductive failures (Helle et al. 1976). Natural mortality and temporary local declines have resulted from influenza outbreaks in northwest Atlantic Ocean harbor seals, Phoca vitulina concolor, (Geraci et al. 1983), and Lepto- spirosis epizootics in California sea lions (Vedros et al. 1971). Decreased prey abundance may have re- duced the ringed seal and bearded seal, Erignathus barbatus, populations in the eastern Beaufort Sea in 1974-75 (Stirling et al. 1982). Thus the northern sea lion decline in southwest- ern Alaska is not unique among pinnipeds, but the causative factor remains difficult to identify. Based on these other examples we can estimate what the ultimate effect of the most plausible hypotheses will be on the population. If one of the causes of the decline is disease, then the population will stabilize and begin to increase once the epizootic has run its course. If a change in prey quantity or quality has reduced the carrying capacity of the Bering Sea, Gulf of Alaska, and North Pacific Ocean for north- ern sea lions, then the population should stabilize if the critical resource stabilizes. If the decline is caused by a combination of factors, then the out- come cannot be determined. Though serious, the current reduced status of the stock in southwestern Alaska does not yet imperil the population, because a large reservoir of adult breeding animals remains to rebuild the population should the decline abate. ACKNOWLEDGMENTS G. Antonelis, R. Bakkala, J. Balsinger, H. Bra- ham, M. Dahlheim, R. DeLong, F. Fay, C. Fiscus, C. Fowler, P. Gearin, D. Kimura, L. L. Low, L. Lowry, R. Miller, M. Perez, K. Pitcher, D. Rugh, A. Smith, and D. Withrow provided critical review of this manuscript. 0. Siebert and D. Blackburn flew the survey aircraft in 1985. J. Sinks and R. Crowe flew the helicopters transporting personnel and gear to and from Ugamak Island. Officers and crew of the NOAA ship Surveyor supported the 1984 survey team. R. V. Miller, D. MacAlister, R. Everitt, and M. Perez provided valuable assistance in the con- duct of the aerial surveys or in the analysis of the aerial photos. P. Gearin, D. Withrow, and S. Osmek made many of the Ugamak Island ground counts. C. Fiscus and R. DeLong were instrumental in designing the aerial, ship, and Ugamak Island surveys. J. Reeves assisted in the analysis of data on the Bering Sea and Gulf of Alaska crab fisheries. The Aleutian Islands Unit, Alaska Maritime Na- tional Wildlife Refuge (U.S. Department of Interior, Fish and Wildlife Service), and the Akutan, Aleut and Chaluka Corporations granted permission for surveys to be made on islands under their control. The North Pacific Fisheries Management Council and the U.S. Marine Mammal Commission pro- vided funding for the 1984 Gulf of Alaska pup counts. LITERATURE CITED Bakkala, R., V. Wespestad, and L. L. Low. In press. Historical trends in abundance and current condi- tion of walleye pollock in the eastern Bering Sea. Fish. Res. Barlough, J. E., E. S. Berry, E. A. Goodwin, R. F. Brown, R. L. DeLong, and A. W. Smith. In press. Antibodies to marine caliciviruses in the Steller sea lion (Eumetopias jubatus Schreber). J. Wildl. Dis. Berkson, J. M., AND D. P. DeMaster. 1985. Use of pup counts in indexing population changes in pinnipeds. Can. J. Fish. Aquat. Sci. 42:873-879. Bigg, M. A. 1985. Status of the Steller sea lion (Eumetopias jvhatus) and California sea lion {Zakrphus califomianus) in British Colum- bia. Can. Spec. Publ. Fish. Aquat. Sci. 77, 20 p. 363 FISHERY BULLETIN: VOL. 85, NO. 2 Bonner, W. N. 1982. Seals and man: A study of interactions. Univ. Wash. Press, Seattle, 170 p. Braham, H. W., R. D. Everitt, and D. J. Rugh. 1980. Northern sea lion population decline in the eastern Aleutian Islands. J. Wildl. Manage. 44:25-33. Calkins, D. G. 1985. Steller sea lion entanglement in marine debris. InR. S. Shomura and H. 0. Yoshida (editors). Proceedings of the Workshop on the Fate and Impact of Marine Debris 27-29 November 1984, Honolulu, Hawaii, p. 308-314. U.S. Dep. Commer., NOAA Tech. Memo. NMFS SWFC-54. Chelnokov, F. G. 1984. Abundance of sea lions and their relationship with fur seals at the southeastern hauling ground of Mednyi Island (Commander Islands). Sov. J. Mar. Biol. 9:191-195. [Engl, transl. Biol. Morya.] Fiscus, C. H., D. J. Rugh, and T. R. Loughlin. 1981. Census of northern sea lion (Eumetopias jubatus) in central Aleutian Islands, Alaska, 17 June-15 June 1979, with notes on other marine mammals and birds. U.S. Dep. Com- merce, NOAA Tech. Memo. NMFS F/NWC-17. 109 p. Fowler, C. W. 1985. An evaluation of the role of entanglement in the population dynamics of northern fur seals on the Pribilof Islands. In R. S. Shomura and H. 0. Yoshida (editors), Pro- ceedings of the Workshop on the Fate and Impact of Marine Debris 27-29 November 1984, Honolulu, Hawaii, p. 291-307. U.S. Dep. Commer., NOAA Tech. Memo. NMFS SWFC-54. Frost, K. J., and L. F. Lowry. 1986. Sizes of walleye pollock, Theragra chalcogramma, con- sumed by marine mammals in the Bering Sea. Fish. Bull., U.S. 84:192-197. Gentry, R. L. 1970. Social behavior of the Steller sea lion. Ph.D. Thesis, Univ. California, Santa Cruz, 113 p. Geraci, J. R., D. J. St. Aubin, I. K. Barker, R. G. Webster, V. S. HiNSHAW, W. J. Bean, H. L. Runhke, J. H. Prescott, G. Early, A. S. Baker, S. Madoff, and R. T. Schooley. 1982. Mass mortality of harbor seals: pneumonia associated with influenza A virus. Science (Wash., D.C.) 215:1129- 1131. Helle, E., M. Olsson, and S. Jensen. 1976. DDT and PCB levels and reproduction in ringed seal from the Bothnian Bay. Ambio 1976, p. 188-189. Hollander, M., and D. A. Wolfe. 1973. Nonparametric statistical methods. John Wiley & Sons, N.Y., 503 p. Kajimura, H. 1984. Opportunistic feeding of the northern fur seal, Cal- lorhinus ursinus, in the eastern North Pacific Ocean and eastern Bering Sea. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-779, 49 p. Kenyon, K. W. 1962. History of the Steller sea lion at the Pribilof Islands, Alaska. J. Mammal. 43:68-75. Kenyon, K. W., and D. W. Rice. 1961 . Abundance and distribution of the Steller sea lion. J. Mammal. 42:223-234. Klumov, S. K. 1957. Registration of the shore rookeries of sea lions (Eumetopias jubatus) in the Kuril Islands and tentative determination of their numerical magnitude. [In Russ.] Dokl. Akad. Nauk. SSSR 117(2):345-348. (Transl. L. V. Sagen, avail. Natl. Mar. Mammal Lab. Library.) KuziN, A. E., M. K. Maminov, and A. S. Perlov. 1984. The number of pinnipeds and sea otters on the Kuril Islands, /w V. E. Rodin, A. S. Periov, A. A. Berzin, G. M. Gavrilov, A. I. Shevchenko, N. S. Fadeev, and E. B. Kucheriavenko (editors), Marine mammals of the Far East, p. 54-72. [In Russ.] Tikhookean. Nauchno-Issled. Inst. Rybn. Khoz. Okeanogr. (TINRO), Vladivostok. (Transl. F. H. Fay., avail. Natl. Mar. Mammal Lab. Library.) Loughlin, T. R., L. Consiglieri, R. L. DeLong, and A. T. Actor. 1983. Incidental catch of marine mammals by foreign fishing vessels, 1978-81. Mar. Fish. Rev. 45(7-8-9):44-49. Loughlin, T. R., D. J. Rugh, and C. H. Fiscus. 1984. Northern sea lion distribution and abundance: 1956-80. J. Wildl. Manage. 48:729-740. Loughlin, T. R., and R. Nelson, Jr. 1986. Incidental mortality of northern sea lions in Shelikof Strait, Alaska. Mar. Mammal Sci. 2:14-33. Mathisen, 0. A., AND R. J. Lopp. 1963. Photographic census of the Steller sea lion herds in Alaska, 1956-58. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 424, 20 p. Merrell, T. R., Jr. 1985. Fish nets and other plastic litter on Alaska beaches. In R. S. Shomura and H. 0. Yoshida (editors). Proceedings of the Workshop on the Fate and Impact of Marine Debris 27-29 November 1984, Honolulu, Hawaii, p. 160-182. U.S. Dep. Commer., NOAA Tech. Memo., NMFS SWFC-54. Merrick, R. L. 1984. Observations on the status of the northern sea lion population in Alaska. M.S. Thesis, Oregon State Univ., Cor- vallis, 85 p. Nelson, M. 0., and E. P. Nunnallee. 1985. Acoustic-midwater trawl survey of spawning walleye pollock in the Shelikof Strait Region, 1980-81 and 1983-84. In R. L. Major (editor). Condition of ground fish resources of the Gulf of Alaska region as assessed in 1984, p. 179-206. U.S. Dep. Commer., NOAA Tech. Memo., NMFS F/NWC- 80. Perlov, A. S. 1982. Reduction in the number of sea lions. [In Russ.] In G. B. Agarkov, V. A. Arsen'ev, V. M. Bel'kovich, V. A. Zem- skiyi, L. A. Popov, I. V. Smelova, A. S. Sokolov, A. G. Tom- lin, and A. K. Yablokov (editors). The study, preservation, and rational exploitation of marine mammals, p. 284-285. Vses. Nauchno-issled. Inst. Morsk. Rybn. Khoz. Okeanogr., Akad. Nauk. SSSR, Astrakhan. (Transl. by S. Pearson, Natl. Mar. Mammal Lab., NMFS, NWAFC, Seattle.) Pike, G. C, and B. E. Maxwell. 1958. The abundance and distribution of the northern sea lion [Eumetopias jubata) on the coast of British Columbia. J. Fish. Res. Board Can. 15:5-17. Pitcher, K. W. 1980. Food of the harbor seal, Phoca vitulina richardsi, in the Gulf of Alaska. Fish. Bull., U.S. 78:544-549. 1981. Prey of the Steller sea lion, Eumetopias jubatiis, in the Gulf of Alaska. Fish. Bull., U.S. 79:467-472. Pitcher, K. W., and D. G. Calkins. 1981 . Reproductive biology of Steller sea lions in the Gulf of Alaska. J. Mammal. 62:599-605. Sandegren, F. E. 1970. Breeding and maternal behavior of the Steller sea lion (Eumetopias jubata) in Alaska. M.S. Thesis, LIniv. Alaska, College, AK, 138 p. Smith, A. W., S. H. Madin, and T. G. Akers. 1973. Pinnipeds as a possible natural reservoir for a virus 364 MERRICK ET AL.: DECLINE OF NORTHERN SEA LION disease of domestic swine. Proceedings of the 10th Annual Conference of the Biological Sonar and Diving Mammals, Stanford Research Inst., Menlo Park, CA. Smith. A. W., R. J. Brown. D. E. Skilling, and R. L. DeLong. 1974a. Leptospini pomona and reproductive failure in California sea lions. J. Am. Vet. Med. Assoc. 165:996-998. Smith. A. W., C. M. Prato, W. G. Gilmartin, R. J. Brown, and M. C. Keyes. 1974b. A preliminary report on potentially pathogenic micro- biological agents recently isolated from pinnipeds. J. Wild. Dis. 10:54-59. Stirling, I., M. Kingsley, and W. Calvert. 1982. The distribution and abundance of seals in the eastern Beaufort Sea, 1974-79. Can. Wildl. Serv., Occas. Pap. 47, 25 p. Thorsteinson, F. v., R. W. Nelson, and D. F. Lall. 1961 . Experimental harvest of the Steller sea lion in Alaska waters. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 371, 15 p. Vedros, N. a., a. W. Smith, J. Schonewald, G. Migaki, and R. C. Hubbard. 1971. Leptospirosis epizootic among California sea lions. Science (Wash., D.C.) 172:1250-1251. WiTHROW, D. E. 1982. Using aerial surveys, ground truth methodology, and haul out behavior to census Steller sea lions, Eumetopias jubatus. M.S. Thesis, Univ. Washington, Seattle, 102 p. 365 ON THE ESTIMATION OF NUMBERS OF NORTHERN FUR SEAL, CALLORHINUS URSINUS, PUPS BORN ON ST. PAUL ISLAND, 1980-86 Anne E. Yorri and Patrick Kozloff^ ABSTRACT Since 1962, the numbers of northern fur seal, Callorhimis ursinus, pups born on St. Paul Island have been determined using a mark-recaptiu-e procedure. We investigate the feasibility of determining estimates of the total pup population on the 14 rookeries of St. Paul Island from subsamples of rookeries. Estimates are derived from simple random sampling and stratified (by rookery size) random sampling using stan- dard ("blow up") estimation procedure, and ratio and regression estimates (based on the same sampling procedure but taking advantage of a strong relationship between numbers of breeding males and live pups on the various rookeries). Evaluation of the sampling schemes and estimation methods is based on the performance of the estimators for 3 years (1965, 1970, 1975) of data for which the mark-recapture estimates from all 14 rookeries were available. Ratio estimates are preferred to estimates obtained from the standard procedure for both simple random sampling and stratified random sampling. Furthermore, estimates from sampling plans based on three strata proved more satisfactory than those based on either unstratified or two-strata sampling. The ratio methods are applied to data collected during 1980-86. The number of northern fur seal pups born on St. Paul Island decreased at approximately 7.5% per year during 1975-81. There was no statistically detectable trend in numbers born during 1981-86. The number of northern fur seals, Callorhinus ursinus, born on St. Paul Island (approximately 80% of the total Pribilof Islands herd production) has been determined in a variety of ways since the United States assumed direct management of the fur seal herd in 1910 (Parker 1946). The history of northern fur seal population estimation during 1912-47 and analyses of the reliability of methods then proposed for estimating numbers of pups are presented in Kenyon et al. (1954). The evolution of the "shearing-sampling" method, a variant of the mark-recapture technique, is discussed in Chapman (1964) and Chapman and Johnson (1968). Since 1962, the estimate of the size of the pup population has been obtained using the "shearing- sampling" method. The safety of the crew, the accuracy of the estimate, and the minimization of disturbance to rookeries are major concerns; hence, the work is done as the breeding structure breaks up, but before pups spend most of their time in the water. During early August, a large number of pups (approximately 10% of the population) are marked by shearing a small patch of hair from the top of 'National Marine Mammal Laboratory, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. ^National Marine Mammal Laboratory, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115; present address: Tanadgusik Cooperation, St. Paul, AK 99660. their heads; this exposes the pale underfur and pro- duces an easily identifiable mark. The marking ef- fort is allocated throughout the rookery so that each pup has an approximately equal chance of being marked. A few days later, each rookery is sampled twice during different periods to estimate the pro- portion of marked animals on the rookery. Thus, estimates of the population size and its variance can be calculated for each rookery. The estimate of the population present at the time of shearing is the number of sheared animals divided by the propor- tion of sheared pups among all those resighted— the normal Petersen estimate. The variance of this estimate is one-fourth the squared difference of the two estimates. The purpose of this paper is to demonstrate the feasibility of obtaining accurate estimates of the total pup population on St. Paul Island from "shearing-sampling" estimates on a few sample rookeries. The advantages of obtaining estimates of the population from a subsample of rookeries in- clude 1) less disturbance on the total northern fur seal population (each season that pup production is estimated on a particular rookery, crews must traverse the rookery four times— once to do the marking, twice to estimate the proportion of marked pups among the population, and once to count the number of dead pups); and 2) considerable savings in time, energy, and funds. Manuscript accepted February 1987. FISHERY BULLETIN; VOL. 85, NO. 2, 1987. 367 FISHERY BULLETIN: VOL. 85, NO. 2 METHODS AND AVAILABLE DATA The data (Table 1) used for evaluating procedures for estimating the size of the pup population on St. Paul Island were the counts of breeding males made in mid-July and the estimate of the size of the pup population made in early August. We assumed that for any year, the sum of the estimated numbers of live pups from each of the 14 rookeries, T, was the known or "true" size of the population. Estimates of the variances of each rookery were available; we assumed that the counts from each rookery were independent and estimated the variance of the total population, o^, as the sum of the estimated variances on the 14 rookeries. An approximate 95% confidence interval for the total population was T ± ^(0.975,14) o, where ^(0.975,14) is the 97.5 percentile of Student's t distribution with 14 degrees of freedom. Two sampling schemes and three estimation pro- cedures were investigated. In particular, estimates based on the standard procedure or the "blow-up" estimate, were compared with ratio and regression estimates, which take advantage of a strong corre- lation between the numbers of breeding bulls and numbers of pups on the rookeries (Figs. 1, 2); these estimation procedures were compared under both simple random sampling and stratified random sam- pling schemes. For each sampling scheme, all possi- ble subsamples of the 14 rookeries were generated, and the distributions of estimates for each of the 3 years of data were constructed. To determine how well the estimates predict the "true" population, we computed the fraction of (nominal) 95% confidence intervals about the estimate which contained the "true" value (the actual confidence level of the 95% confidence region). In addition, we computed the variance, bias, and the average half-width of the nominal 95% confidence interval of each estimator under the given sampling design. "Blow-up" estimates of the total numbers of pups on the rookeries under simple random sampling, T^ij, were calculated in the following way (Cochran 1977): Let (Pj, Pg- • • ■ -P«) be estimates of pup numbers on n sample rookeries. The total on all rookeries was approximated by multiplying the aver- age number of pups on the sample rookeries by the total number of rookeries: '^BU - 14 1 P. = 14 P. n i=i (1) The estimate of the variance of this estimate is Y2.t{Tbu) 14 n 14 n n{n - 1) 1 = 1 I (P, - P)2 14 '^ + — 2. Var(P,). n 1=1 (2) When sampling was stratified, the above pro- cedure was applied to each stratum. The total num- ber of pups on all rookeries was estimated as the sum of the estimates on all strata; the variance was approximated by applying Equation (2) to each stratum and then summing over the strata. Other methods of estimating the total number of pups on all rookeries, when a total count of breed- ing males was available, were suggested through an examination of the regression equations of numbers Table 1 .—Numbers of northern fur seal pups counted and their standard deviations, numbers of breeding bulls, and ratio of pups counted to breeding males for the rookeries of St. Paul Island, AK for 1965, 1970, and 1975. 1965 1970 1975 Rookery Pups SD Bulls Ratio Pups SD Bulls Ratio Pups SD Bulls Ratio Vostochni 34,208.0 2,091.6 1,434 23.9 33,808.5 4,797.7 791 42.7 41,356.0 2,300.9 799 51.8 Tolstoi 25,122.0 294.2 876 28.7 22,194.0 1,759.3 570 38.9 31,107.5 1,375.3 621 50.1 Zapadni 25,066.0 4,228.5 978 25.6 33,665.5 1,112.3 664 50.7 36,815.5 4,413.1 610 60.4 Reef 29,032.5 488.6 1,179 24.6 24,907.0 4,464.7 716 34.8 27,561.0 1,050.8 622 44.3 Morjori 15,434.5 204.4 739 20.9 14,894.0 3,624.6 352 42.3 21,284.5 3,926.6 376 56.6 Polovina CI. 18,547.5 491.4 650 28.5 17,092.5 1,880.2 390 43.8 24,869.5 4,017.1 461 53.9 L. Zapadni 14,306.0 1,937.5 551 26.0 15,240.0 739.6 325 46.9 21,168.0 2,115.7 363 58.3 Kitovi 11,361.0 244.7 486 23.4 12,713.0 1,678.7 241 52.8 12,965.0 2,511.0 267 48.6 Gorbatch 16,929.0 1,347.7 674 25.1 15,027.5 1,248.0 385 39.0 17,038.5 761.6 387 44.0 Ardiguen 2,680.5 997.7 105 25.5 3,106.5 77.1 108 28.8 2,774.0 297.0 85 32.6 Lukanin 5,290.0 895.2 204 25.9 5,508.5 1,608.7 107 51.5 5,704.0 868.3 112 50.9 Zapadni Reef 5,259.0 58.0 221 23.8 4,191.5 560.7 106 39.5 7,223.0 657.6 139 52.0 Polovina 5,291.0 2,426.8 220 24.1 3,707.5 222.7 87 42.6 4,354.5 1,130.7 88 49.5 L. Polovina 6,117.5 236.9 236 25.9 3,848.0 257.4 103 37.4 3,415.0 43.8 88 38.8 214,644.5 6,019.9 8,553 25.1 209,904.0 8,479.6 4,945 42.4 257,636.0 8,558.0 5,018 51.3 368 YORK and KOZLOFF: NORTHERN FUR SEAL PUPS of pups as a function of numbers of breeding males (Figs. 1, 2). The analyses of variance of these regres- sions indicated that the quality of the fits was ex- cellent and that the relationship might be used for predictive purposes. No intercepts, except that for 1916 data, were significantly (P > 0.95) different from 0. We were interested in subsampling the rookeries (possibly conducting the estimation on as few as four rookeries) and therefore, if a regression estimator were to be used, it was desirable to reduce the number of parameters as much as possible. In- asmuch as the intercepts were not different from 0, the simpler model with no intercept was con- sidered appropriate. Since the variance of the pup estimates was not constant for each rookery, weighting appeared necessary. The variance of the estimates of pup numbers was roughly proportional to the number of bulls, and in such cases (Draper and Smith 1966), the best estimate of the slope of regression line is the average number of pups divided by the average number of bulls (equivalent to the ratio of the total number of pups to the total number of bulls). In this case, the total number of pups on all rookeries was estimated in the follow- ing manner: Let P^, . . . , P„ and Bi, . . . , B„ be as above, and B a count of the total number of bulls on all rookeries. Then the total number of pups on all rookeries may be estimated as B 1 P, T^ = ^^^ = rB. (3) One estimate of the variance of this ratio estimator is Var (Tj,) = ^ f\ I Var(P,) " \ "^ 1=1 ,,T*i^M (Cochran 1977). (4) When stratified random sampling was used in- stead of simple random sampling, we calculated the estimator in the same way since the ratio of pups to breeding males did not vary significantly between strata. The difference was due to the evaluation pro- cedures; the number of logical sampling combina- tions differed and the analysis was restricted to those combinations of sample rookeries that were consistent with the sampling design (e.g., one small rookery, one medium-sized rookery, and two large rookeries). Another way to estimate the ratio and its variance is with jackknife methods (Hosteller and Tukey 1977). Let r_, be the ratio of pups to breeding males on all but the i^^ rookery, and r the ratio of pups to breeding males on all the sample rookeries (as in Equation (5)): I {nP - P,) r-i = 1=1 I (nB - B,) i=l (5) Then, the i'^ pseudovalue is r* = nr - (n - 1) r_i. The jackknife estimate of the ratio, r*, is the mean of the r,*'s and the variance of r* is (Mosteller and Tukey 1977): Z (r* - r*y Var(r*) 1=1 n (n - 1) Thus, the jackknife estimate of the total numbers of pups on all rookeries, Tj, is Tj = r* B, and Var(T^) = B^ Var(r*). The advantage of the jackknife estimate over the ordinary ratio estimate is the reduction of bias and a simple method of calculating the variance. The ordinary regression estimate (assuming that the intercept is 0) of the ratio of pups to bulls is I P, B, i = l S = n i=l Thus, the regression estimator of total numbers of pups is Tn, = sB and Var(T;j„) = B^ Var(s). The estimate of the variance of s is calculated from the mean square residual of the regression equation. RESULTS Regressions of numbers of northern fur seal pups 369 FISHERY BULLETIN: VOL. 85, NO. 2 20r 1912 15 10 0 20r 1913 i_ • 195 + 69x J I I I 50 100 150 200 250 20 r 15- 10 5 - 0 1914 / -648 + 69x :/ 100 200 25 r 300 300 150 200 250 300 200 400 600 800 Count of harem bulls Figure 1.— Relationship of counts of northern fur seal pups born to counts of harem bulls for the various rookeries of St. Paul Island, AK, during 1912-22 (data from Lander 1980). 370 YORK and KOZLOFF: NORTHERN FUR SEAL PUPS 3 O o in O. a H— o / 20 • / 10 n •/ 493 H • f 1 -41x 1 50 r 1970 40 500 1000 30 20 10 0 » • / 105 + / 46x X X 50 r 1975 40 30 20 10 0 500 1000 1500 Count of harem bulls -736 + 58x X J 500 1000 1500 Figure 2.— Relationship of estimates of northern fur seal pups born and counts of harem bulls for the various rookeries of St. Paul Island, AK, during 1963-75 (data from Lander 1980). 371 FISHERY BULLETIN: VOL. 85, NO. 2 versus numbers of breeding males for those years in which data were collected on all rookies indicated a strong relationship that could be used for predic- tion of total pup production if only subsamples of rookeries were censused. The relationship held for those years when censuses of pups were conducted by counting (Fig. 1), and for later years when the shearing-sampling method was used (Fig. 2). Al- though the slopes varied substantially from year to year (they ranged from 71 in 1913 to 29 in 1963), the variance about the regression line within any particular year was very small. We compared the various estimators and sampling plans by analyzing the bias and variance of the esti- mates and the half-width and coverage properties of nominal 95% confidence intervals for 3 years (1965, 1970, 1975) of data when all rookies were sampled. Detailed statistics on the performance of the estimators under all sampling plans appear in a manuscript report available from the authors^. Under simple random sampling, the "blow-up" estimate is unbiased. The various ratio estimates are all slightly biased (in most cases less than 1%) with the regression estimate exhibiting the largest degree of bias. In Figure 3 the percentage of bias of the three ratio estimates for the 1975 data is shown as a function of sample size (under simple ran- dom sampling). Estimates based on 1975 data were the most biased among the 3 years analyzed, and these biases are exhibited as a worst case. The regression estimate was the most biased, and for these data the bias increased as the sample size in- creased; however, the bias was only about 1% and is not serious. Confidence intervals were constructed for each subsample and a count was made of the number of nominal 95% confidence intervals containing the "true" population. The observed coverage was near 95% for most procedures. Confidence intervals for the regression estimate tended to be conservative, i.e., a higher than 95% coverage rate, while the coverage rate for the ordinary ratio estimate tended to be less than 95%. Coverage rates for the jack- knife and blow-up estimates were near 95% or a bit higher. This indicates that the estimate of the vari- ance of the regression estimate tended to be too large, that of the ordinary ratio estimate was too small, and that the estimates of the variance of the ^York, A. E., and P. Kozloff. 1985. Estimation of numbers of fur seal pups bom on St. Paul Island, 1980-84. Unpubl. manuscr. Available National Marine Mammal Laboratory; 7600 Sand Point Way N.E., Seattle, WA 98115. (Background paper submitted to the 28th Annual Meeting of the Standing Scientific Subcommit- tee of the North Pacific Fur Seal Commission, March-April 1985, Tokyo, Japan.) blow-up and jackknife estimates tended to be un- biased. The half-widths of confidence intervals for the ratio estimates were nearly equal. All were less than one-half the length of the half -width of the con- fidence interval of the blow-up estimate. The rookeries were stratified by population size. Two methods for stratifying the rookeries were in- vestigated: one using two strata (small and large rookeries) and the other using three strata (small, medium, and large rookeries). As in the case of sim- ple random sampling, the ratio estimators were superior to the blow-up estimates. The estimators under the three-strata sampling plans were less variable than under the two-strata sampling plans. In addition, the computed levels of the nominal 95% confidence intervals were higher and the size of the confidence intervals smaller. Under the three-strata sampling plans, the standard deviations of the esti- mates were about 10% smaller than under simple random sampling with the same size sample. This resulted in a similar reduction in the size of the con- fidence intervals. These results indicated that reasonable estimates of the size of the pup popula- tion can be made using any of the ratio estimators under various sampling plans. The superior plans use three strata: two small, one medium, and one large rookery; one small, two medium, and one large l.5r , Jacknife 0.9 c Q. m ■0.3 -0.9 -1.5 12 15 Sample size Figure 3.— Percent bias of the jackknife estimates ( ), ordinary ratio estimates ( • • • ), and regression estimates ( ) based on sim- ple random sampling of 1975 northern fur seal data. 372 YORK and KOZLOFF: NORTHERN FUR SEAL PUPS rookery; and, one small, one medium, and two large rookeries. A subsampling estimation procedure was devel- oped for 1980-84: rookeries were grouped into three strata— large, medium, and small rookeries; one small, one medium, and two large rookeries were sampled each year. Furthermore, in order that some rookeries were not disturbed inordinately more than others, each rookery was sampled at least once, but no more than twice during the 5-yr period. We had intended to census all rookeries in 1985, but logistic difficulties permitted a sampling of only seven rookeries. A summary of data collected during 1980-86 with the ordinary ratio, jackknife ratio, and regression estimates of the ratio of pups to breeding males ap- pears in Table 2. The estimates based on the three methods are approximately equal within each year; in most cases, the jackknife estimate lies between the ordinary ratio and regression estimates. Esti- mates of the total number of pups born were ob- tained by adding counts of dead pups to number of pups alive at the time of census (based on jackknife ratios); approximate 95% confidence intervals were calculated (Table 2). In Figure 4, estimated 95% confidence intervals Table 2— Summary of the total number of breeding northern fur seal males, ratios of the number of pups alive at the time of sampling to the number of breeding males counted, estimated number of pups alive at the time of sampling, counted number of dead pups, and estimated number of pups born, and approximate 95% confidence interval based on the jackknife standard errors, St. Paul Island, 1980-84. Rat OS of pups to Total no. of breeding males Numbe r of pups breeding Jackknife Ordinary Regres- Year males ratio ratio sion Live Dead Born 1980 5,490 35.695 35.896 35.580 195,966 7,859 203,825 ± 36,838 1981 5,120 33.720 33.821 33.563 172,646 6,798 179,444 -1- 20,054 1982 5,767 34.035 33.896 34.147 196,280 7,301 203,581 + 9,665 1983 4,827 33.135 32.766 33.448 159,944 5,997 165,941 + 19,216 1984 4,803 34.803 33.861 34.167 167,159 6,115 173,274 ± 22,531 1985 4,372 40.482 40.292 41.071 176,992 5,226 182,258 + 18,887 1986 4,603 34.735 34.936 34.498 167,656 7,771 167,656 + 16,272 ■a 3 o Q. 3 a. 350 r 305 260 215 170 125 1 I' 1967 1971 1975 1979 1983 1987 Year FiGt^iRE 4.— Approximate 95% confidence intervals and estimates of numbers of northern fur seal pups born on St. Paul Island, AK, 1970-86. (We include only those years for which data were available to compute estimates accord- ing to the methods developed in this paper.) 373 FISHERY BULLETIN: VOL. 85, NO. 2 of numbers of northern fur seal pups born on St. Paul Island since 1970 are presented; estimates for 1970-79 are based on data from Lander (1980). We computed estimates for those years in which cen- suses were made on all rookeries or for which data were available to compute estimates according to the methods developed in this paper. Regressions of logarithms of numbers of pups born versus time indicated a statistically significant decrease during 1975-81— a decrease of 7.5% per year with a stan- dard error of 2%. During 1981-86, there is no statis- tically significant decreasing or increasing trend; the estimate of the slope is - 1.8% with a standard er- ror of 1.8%. This slope is statistically different from the - 7.5% slope calculated for 1975-81 (P > 0.90). DISCUSSION Our study indicates that we can obtain reasonable estimates of the total number of northern fur seal pups born from subsampling as few as four rook- eries of St. Paul Island if estimates of numbers of pups and breeding males are available for the sam- ple rookeries and if a total bull count is available for the island. Subsampling is successful because within a given year, pup production is predictably propor- tional to numbers of breeding males. Some refine- ments in the reduction of bias and variance can be made by restricting the subsamples to stratified designs over large, medium, and small rookeries. The advantages of subsampling rookeries for cen- susing northern fur seal numbers are considerable. Most important is the reduction of total disturbance on the northern fur seal population on the island. Our sampling schedule over several years attempts to apportion disturbance approximately equally so that rookeries are neither under- or oversampled through time. This is an important aspect of the sampling design, since it is not known how great the long-term impact of disturbance is. In addition, subsampling requires a smaller crew for the shear- ing and less time for resampling, resulting in con- siderable savings of resources. Ratios of numbers of males to pups, and conse- quently breeding females, vary considerably over time, even in successive years. It is difficult to inter- pret the meaning of these changes. During the period covered by Figure 1 (1912-22), numbers of pups born on St. Paul Island were increasing rather rapidly. Since males begin to breed at an older age than females, part of the increase in the ratio of breeding males to pups may be explained by the number of breeding males lagging a few years behind the number of breeding females. Signifi- cantly different ratios from one year to the next could also be due to differences in counting methods or abilities among individual counters, or to differ- ent survival rates among separate cohorts (e.g., harvest rates). Figures 1 and 2 also imply a certain consistency and a rather uniform rate of usage of rookeries by breeding males and females, in that, if a rookery accounts for 10% of breeding males within a year, it will account for approximately 10% of the total pup production within the same year. A rookery's relative contribution to both these populations may change but the correlation between them does not appear to change. The recent history of the population of numbers of pups on St. Paul Island in Figure 4 shows a de- crease of about 7.5% per year during 1975-81. No significant trend is detectable after 1981, although the number born in 1982 was significantly higher than in 1981 or 1983-86. The causes of the decline are unknown. There is no evidence that pregnancy rates have changed significantly since the 1950's (Goebel and Gentry 1984"*). Thus, considerable at- tention has centered on potential causes of increased mortality of northern fur seals: entanglement in debris (e.g.. Fowler 1985), effects of weather (Trites 1984; York 1985^), and direct effects on food avail- ability from competition with fisheries in the North Pacific Ocean (York and Hartley 1981; Swartzman and Harr 1983; Kajimura 1984; Loughlin and Livingston 1985^). One may also speculate that the pattern of decline and possible stabilization in num- bers of pups born resulted from a new disease which abated or was controlled by an immune response of the population (c.f., Geraci et al. 1982). Of the aforementioned explanations for the de- cline, only entanglement has been cited as a major contributing factor with an attributed mortality of «Goebel, M. E., and R. L. Gentry. 1984. The use of longitu- dinal records of tagged females to estimate fur seals survival and pregnancy rates. Unpubl. manuscr. National Marine Mammal Laboratory, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. (Background paper submitted to the 27th Annual Meeting of the Standing Scientific Committee of the North Pacific Fur Sea! Commission, March-April 1984. Moscow, U.S.S.R.) ^York, A. E. 1985. Forecast of the 1985 harvest on St. Paul Island. Unpubl. manuscr. National Marine Mammal Laboratory, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. (Background paper submitted to the 28th Annual Meeting of the Standing Scientific Subcommittee of the North Pacific Fur Seal Commission, March-April 1985, Tokyo, Japan.) *Loughlin, T. R., and P. A. Livingston (editors). 1986. Sum- mary of joint research on the diets of northern fur seals and fish in the Bering Sea during 1985. NWAFC Processed Report 86-19, 92 p. Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. 374 YORK and KOZLOFF: NORTHERN FUR SEAL PUPS about 5.5% per year (Fowler 1985). It is possible that entanglement in debris was indeed responsible for the decline during 1975-81, however, the data in Figure 4 do not seem to support this hypothesis. If entanglement were the principal cause of this decline, we would have expected the population to have continued to decrease at the pre- 1981 rate since the observed entanglement rates have re- mained stable since 1976. We may never know the cause of the 1975-81 decline in northern fur seal production. In general, estimates of population size are highly variable so several censuses are required to detect a statistically significant decrease in a population; thus, the fact of a decline, unless it is sudden and dramatic, is not usually known for several years following its initia- tion. Post facto studies are invariably subject to criticism for flaws in experimental design; thus, careful continual monitoring of the many aspects of the biology of a population is the best hope for ascribing a particular cause to a population change. Comparisons of the population dynamics, food habits, incidence of diseases, and entanglement rates of northern fur seals with other pinniped species which share their habitat in the North Pacific Ocean might shed additional light on the various hypotheses. ACKNOWLEDGMENTS We thank R. H. Lander for suggesting this study and for indicating the value of the relationship between numbers of breeding males and numbers of pups. Review of this manuscript and helpful sug- gestions for its improvement were provided by H. Braham, R. L. DeLong, C. W. Fowler, G. Y. Harry, Jr., D. Kimura, R. H. Lander, and R. Pearson. LITERATURE CITED Chapman, D. G. 1 964 . A critical study of the PribOof fur seal estimates. Fish. Bull., U.S. 63:657-669. 1973. Spawner recruit models and estimation of the level of maximum sustainable catch. Rapp. P. -v. Reun. Cons. int. Explor. Mer 164:325-332. Chapman, D. G., and A. M. Johnson. 1968. Estimation of fur seal pup populations by randomized sampling. Trans. Am. Fish. Soc. 97:264-270. Cochran, W. G. 1977. Sampling Techniques. 3d ed. John Wiley and Sons, N.Y., 428 p. Draper, N. R., and H. Smith. 1966. Applied regression analysis. John Wiley and Sons, N.Y., 407 p. Fowler, C. W. 1985. An evaluation of the role of entanglement in the popula- tion dynamics of northern fur seals on the Pribilof Islands. In R. S. Shomura and H. 0. Yoshida (editors). Proceedings of the workshop on the fate and impact of marine debris, 27-29 November 1984, Honolulu, Hawaii, p. 291-307. U.S. Dep. Commer., NOAA Tech. Memo. NMFS SWFC-54. Geraci, J. P., D. J. St. Auben, I. K. Barker, R. G. Webster, V. S. HiNSHAW, W. J. Bean, H. L. Ruhnke, J. H. Prescott, G. Early, A. S. Baker, S. Madoff, and R. T. Schooley. 1982. Mass mortality of harbor seals: Pneumonia associated with influenza A virus. Science (Wash., D.C.) 215:1129- 1131. Kajimura, H. 1984. Opportunistic feeding of the northern fur seal, Cal- lorhinus ursinus; in the Eastern North Pacific Ocean and Eastern Bering Sea. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF-779, 49 p. Kenyon, K. W., V. B. Scheffer, and D. G. Chapman. 1954. A population study of the Alaskan fur seal herd. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Wildl. 12, 77 p. Lander, R. H. (editor). 1980. Summary of northern fur seal data and collection pro- cedures. Volume I: land data of the United States and Soviet Union (excluding tag and recovery records). U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-31, 315 p. MOSTELLER, F., AND J. W. TUKEY. 1977. Data analysis and regression. Addison- Wesley Publ. Co., Reading, MA, 588 p. Parker, G. H. 1946. The world expands; recollections of a biologist. Har- vard Univ. Press, Cambridge, MA, p. 112-116. SWARTZMAN, G. L., AND R. T. HAAR. 1983. Interactions between fur seal populations and fisheries in the Bering Sea. Fish. Bull., U.S. 81:121-132. Trites, a. W. 1984. Population dynamics of the Pribilof Islands North Pacific fur seal (Callorhinus ursimis). M.S. Thesis, Univ. British Columbia, Vancouver, 123 p. York, A. E., and J. R. Hartley. 1981. Pup population following harvest of female northern fur seals. Can. J. Fish. Aquat. Sci. 38:84-90. 375 NOTES LABORATORY STUDIES OF THE PATTERN OF REPRODUCTION OF THE ISOPOD CRUSTACEAN IDOTEA BALTICA The isopod Idotea baltica is a cosmopolitan species that can be an important component of fishes' diets in the field (summarized in Sywula 1964 and Strong and Daborn 1979). Tinturier-Hamelin (1963) reported that /. baltica extends along all European coasts, from Finland to Gibraltar, including Great Britain; it is present in the Baltic, the Mediterranean, and the Black and Caspian Seas. In North America it is present from Nova Scotia to North Carolina. In addition, Sywula (1964) reported that it is also found in South America, Bermuda and Barbados, the Red Sea, Australia, New Zealand, and Java. Investigators who have observed I. baltica in the field have reported the species' association with the dominant plants of the community. Interestingly, the type of associated plant varies with geographic locality. In Nova Scotia it is found on Ascophyllum nodosum (Strong 1978), in the Baltic on Fucus vesi- culosus (Salemaa 1979), in Poland on Zostera (Mobius 1873, as reported in Sywula 1964), and on Ulva lactuca in Jamaica Bay, NY (present study). Generally, the animals' principal food is the plant on which they are found, and this species is often the principal primary consumer of its community (Strong and Daborn 1979), occupying a critical link in local fish food chains. The present study was undertaken to provide in- formation about /. baltica' s reproductive behavior and physiology under laboratory conditions in order to determine the feasibility of developing it as a fish food for mariculture systems. Materials and Methods All animals were collected in July and August 1985, by removing attached Ulva lactuca thalli from the fouling community attached to submerged piers at the Barrens Island Marina, Jamaica Bay. The ani- mals were sorted from these collections in the labor- atory, and placed in individual 22.5 cm diameter glass culture dishes of ambient seawater (29 ppt) either in heterosexual pairs (30 pairs), or in isosexual pairs (20 male and 20 female isosexual pairs). The pairs were maintained at room temperature (x - 24.3 ± 2°C SD), with a light cycle of 15:9 L:D. They were fed Viva lactuca thalli ad libitum. These pairs were observed 2 times a day, 12 hours apart, in the light, and maintained until one of the members of the pair died. Observations consisted of noting the occurrence of molts and ovulations, as well as any reproductive behaviors exhibited. Intermolt periods were calculated by counting the number of days between the first and second molts only of animals maintained in heterosexual pairs. This was done to minimize any artifacts of culture conditions. In addition to the pairs, 60 females were isolated in individual 10 cm diameter culture dishes. These females were used to determine the variability in timing of molts, ovulations, and expression of repro- ductive behavior. Females were observed at 12-h intervals, and the dates and times of their molts noted. Males were introduced either on the day the females molted, on day 1 or day 2 after the molts, or no males were introduced at all (12, 19, 13, and 16 different females observed, respectively). The oc- currence and timing of copulations, ovulations, and subsequent brood developments were noted for all four groups of females. Finally, to determine the timing of copulation with respect to the sequence of the shedding of the two parts (see following section) of the female's exo- skeleton, males were introduced to five females between the first and second partial molts and the males' responses noted. Results Molts Intermolt />eWo 0.05). Nature of the molts.— The exoskeleta of both sexes were cast off the same way. First the posterior half of the exoskeleton (from the fifth segment back) was shed, then the remaining anterior portion was cast off. The anterior part included the first four pairs FISHERY BULLETIN: VOL. 85. NO. 2, 1987. 377 of oostegites of the female's brood pouch; the poste- rior part included the fifth and last pair. There was an interval between the shedding of the two parts (females: i = 6.5 ± 6.1 SD hours, n = 35, range = 0-12 hours; males: x = 8.3 ± 6.9 SD hours, n = 36, range = 0-24 hours). There was no significant difference between the intervals of the two sexes (Student's f-test: t = 1.41, df = 69, P < 0.05). Final- ly, the observations suggest that there is no terminal molt in either sex, because all individuals molted until they died. Amplexus Males picked up and held onto females until the females molted. During amplexus, females were held ventral to the male, and only the males' move- ments resulted in locomotion. There was no spe- cialized point of attachment on the female. The male inserted the angle formed by the dactyl held slight- ly extended from the palm of its second gnathopod into the space between the curved, lateral edges of the female's first and second pereonal segments. Sometimes males held onto females with posterior pereiopods as well, but this was only for brief periods of time. The occurrence of amplexus was correlated prin- cipally with female intermolt stage. Amplexus began about ^ = 1.9 ± 1.4 SD days before the female molt(n = 15, range = 0.0-4.5 days) and all couples separated within 12 hours after the female discarded its anterior cast. Thus, females were in amplexus for approximately the latter 14.2% of their inter- molt periods. Most females were in amplexus 24 hours before their posterior molts and separated from the males 24 hours after their anterior molts (Table 1; Fisher's exact probability test, P < 0.001). In contrast, most males were not in amplexus either before or after their molts (Table 1; 83.3% and 64.3% not in am- plexus before and after the molt, respectively). However, there was a significantly greater propor- tion of males in amplexus after than before their molts (xi = 3.94, P < 0.05). Thus, amplexus was Table 1. — Frequencies of occurrences of amplexus 12 hours before the posterior molt and 12 hours after the anterior molt in Idotea baltica maintained in heterosexual pairs. Number of molts Before After Sex Number in Number amplexus apart Number in Number amplexus apart Females Males 33 42 31 2 7 35 1 32 15 27 principally correlated with female intermolt stage, but male intermolt stage had an affect as well. Finally, amplexus was never observed in either male or female isosexual pairs. Copulations During copulation the male held the female's body in a perpendicular, ventral position, and inserted its pleopods into the posterior part of the female's brood pouch about five times in rapid succession. This was followed by a rest period of about 5 sec- onds during which the male retained its hold on the female. Then the sequence was repeated two or three more times. Copulations occurred within minutes after males were introduced to females that had shed both parts of their exoskeleta, regardless of how many days ago the molt had occurred. In contrast, no copula- tions occurred in the five females that had just molted the posterior portions. Amplexus was ini- tiated with these females instead, and copulation occurred only after the anterior portion of the exo- skeleton was shed. Ovulations Isolated females ovulated about 2.9 + 0.5 SD days after their molts {n = 14, range = 2-3.5 days). In contrast, females maintained with males in hetero- sexual pairs ovulated x = 0.12 ± 0.2 SD days after their anterior molts (n = 26, range = 0-0.5 days). The difference was significant (Student's ^-test: t = 25.92, df = 38, P < 0.001). Another difference between isolated females and females maintained in heterosexual pairs was that none of the broods of the former group developed, while most of the latter broods did (Table 2: 0 of 14 vs. 24 of 26 broods, respectively; Fisher's exact probability test: P < 0.001). Eggs ovulated by isolated females were no longer observed in the brood pouches about 5.0 + 1.8 SD days after the females' molt (n = 12, range = 3-6 days) or about 2.0 ± 1.5 SD days after ovula- Table 2.— Frequencies of ovulations and viable broods of isolated females and females paired with males. Number of Number of ovulated broods that No. of ovula- no ovula- do not females tions tions develop develop Isolated females 16 14 2 0 14 Paired females 26 26 0 24 2 378 tion (n = 12, range = 0.5-5.0 days). These observa- tions show 1) that females ovulate in the absence of a male, but 2) delay their ovulations under those conditions; and 3) such eggs disappear from the brood pouches a few days after ovulation. The introduction of a male to a female any time during the approximately 3-d period that females could delay their ovulations stimulated ovulation. Copulations generally occurred within 5 minutes after males were introduced to females 0, 1, or 2 days past their molts, and all ovulations occurred within 3 hours of copulation. There was no signif- icant difference in the frequencies of copulations and ovulations among females isolated for different lengths of time (Table 3; xl = 0.008, P > 0.05 and xl = 0.712, P > 0.05 for copulations and ovulations, respectively). In contrast, the broods of females who copulated and ovulated 0 and 1 day after their molts devel- oped significantly more often than did the broods of 2-d postmolt females (Table 3;xi = 3.61, P< 0.05). This suggests that unfertilized eggs aged past 2 days have reduced viability. Table 3.— Frequencies of copulations, ovulations, and viable broods of isolated females introduced to males 0, 1, and 2 days past the females' molts. Number of Days after No. of Copulations Ovulations Broods that develop molt females Yes No Yes No Yes No 0 12 10 2 10 2 10 2 1 19 16 3 14 5 12 7 2 13 11 2 11 2 5 8 Discussion The present study has shown that males and females remain apart until towards the end of the female intermolt period, when amplexus is begun. This continues until the female molts, when copula- tion, followed by ovulation occtirs. Females repeated this cycle in the laboratory until they died. Some females produced three broods in succession, with- out a rest period. In contrast, females in the field in Nova Scotia produce only one brood (Strong 1978), and in the northern Baltic most have one brood (although some may have two (Salemaa 1979)). It is impossible to say whether the difference in the number of broods produced in the field and in the laboratory is caused by environmental and/or genetic differences. In support of the first explana- tion, a recent symposium has demonstrated the im- portance of photoperiod in regulating the timing of reproductive activities of marine animals (Marcus 1986), and, specifically, in the number of broods of some peracarids (Steele and Steele 1986). The photo- period of Nova Scotia may limit the number of broods there. In support of the second explanation, Healy and O'Neill (1984) noted that two populations of /. granulosum produced different numbers of broods in Ireland and in Britain, although there was no significant difference in temperature or latitude at the two locations. Further, while Tinturier- Hamelin (1963) found no reproductive barriers among /. baltica from different parts of Europe that were hybridized in the laboratory, she did conclude that the subspecies were genetically distinct. Only laboratory culture, under identical conditions, of representatives from different geographic localities will reveal whether the difference in the number of broods is genetic. The present observations also reveal that /. baltica females spend relatively little of their intermolt periods in heterosexual pairs (14%) compared with other peracarids (the mean of seven other peracarids was 32% (Borowsky 1986)). Strong (1978) reported that a short amplexus period was correlated with a high risk of fish predation in the amphipod Hyallela azteca and suggested that couples are more visible than single individuals. This may be the case here as well. Fish and predaceous shrimp are abun- dant in the community from which individuals for the present study were collected (D. Franz pers. commun.^). Both sexes cast off the posterior parts of their exo- skeleta before the anterior parts. Since the female's anterior portion includes most of the brood pouch, the new marsupium is not completely exposed un- til both parts are shed. Males did not copulate with females that had not molted the anterior cast. The sequence of molts and copulation makes sense, because the present study shows that copulations stimulate ovulations. If copulation occurred before the anterior portion of the pouch were shed, the new eggs might be secreted into the old brood pouch, and then discarded along with the anterior molt. The present study reveals some flexibility in the timing of key reproductive events. Females could copulate until about 36 hours after they cast off their anterior molts without incurring a significant reduc- tion in fecundity. However, if copulation did not occur by about 3 days after their molts, females 'D. Franz, Department of Biology, Brooklyn College, City University of Nevi^ York, Bedford Avenue and Avenue H, Brook- lyn, NY 11210, pers. commun. August 1986. 379 ovulated spontaneously. These eggs did not develop and disappeared from the brood pouches after a few days. Thus, there appears to be no sperm storage in /. baltica, and females must be accompanied by a male at the time of their molts to ensure the development of their broods. One interesting observation was that males en- gage in amplexus significantly more often after than before their molts. This may be explained by the observation that neurons become detached from the exoskeleton a few days before the molt (Guse 1983). Thus, if contact and/or water-borne pheromones are secreted by receptive female /. baltica as they are in some other peracarid females (Borowsky 1984, 1985, 1986), it is possible that the males cannot sense the stimuli produced by females shortly before their own molts, and therefore are less likely to engage in amplexus at that time. Conclusion The results of the present study show that /. baltica adults can be maintained in the laboratory, and will reproduce freely with minimal effort and at minimal cost. Females fed exclusively on Ulva lac- tuca produced many broods in succession in non- aerated, uncycled water. While further study is necessary to determine whether juveniles will develop under these conditions, and, if so, what the yield will be, the observations reported here suggest the feasibility of culturing this species for fish mariculture systems. Literature Cited Borowsky, B. 1984. The effects of receptive females' secretions on some reproductive behaviors in the amphipod crustacean Micro- deuto-pus gryllotalpa (Costa). Mar. Biol. 84:183-187. 1985. The responses of the amphipod crustacean Gammarus palustris to conspecifics' and congenerics' secretions. J. Chem. Ecol. 11:1545-1552. 1986. Laboratory observations of the pattern of reproduction oi Elasmopus Levis (Crustacea: Amphipoda). Mar. Behav. Physiol. 12:245-256. GusE, G. W. 1983. Ultrastructure development and molting of the aesthe- tascs oi Neomysis integer and Idotea baltica (Crustea: Mala- costraca). Zoomorphology 103:121-133. Healy, B., and M. O'Neill. 1984. The life cycle and population dynamics of Idotea pelagica and Idotea. granuloma (Isopoda: Valvifera) in south- east Ireland. J. Mar. Biol. Assoc. U.K. 64:21-33. Marcus, N. H. 1986. Introduction to the symposium: photoperiodism in the marine environment. Am. Zool. 26:387-388. MoBius, K. 1873. Die wirbellosen tiere der ostsee. Jahresber. Comm. Wiss. Unters. Dtsch. Meere 2 Kiel. Salemaa, H. 1979. Ecology of Idotea-spp- isopoda in the northern Baltic. Ophelia 18:133-150. Steele, V. J., and D. H. Steele. 1986. The influence of photoperiod on the timing of reproduc- tive cycles in Gamrnarus species (Crustacea, Amphipoda). Am. Zool. 26:459-467. Strong, K. W. 1978. Breeding and bionomics of Idotea baltica (Pallas) (Crustacea: Isopoda). Proc. N.S. Inst. Sci. 28:217-230. Strong, K. W., and G. R. Daborn. 1979. Growth and energy utilization of the intertidal isopod Idotea baltica (Crustacea: Isopoda). J. Exp. Mar. Biol. Ecol. 41:101-124. Sywula, T. 1964. A study on the taxonomy, ecology and geographical distribution of species of the genus Idotea fabricius (Isopoda, Crustacea) in the Polish Baltic. 1. Taxonomical part. 2. Eco- logical and zoogeographical part. Bull. Soc. Amis Sci. Lett. Poznan ser D 4:141-200. Tinturier-Hamelin, E. 1963. Polychromatisme et determination genetique du sexe chez I'espece polytypique Idotea baltica (Pallas) (Isopode Valvifere). Cah. Biol. Mar. 4:473-591. Betty Borowsky Osbom Laboratories of Marine Sciences New York Aquarium Boardwalk at West 8th Street Brooklyn, NY 11224 OCCURRENCE OF THE FIRST FRESHWATER MIGRATION OF THE GIZZARD SHAD, DOROSOMA CEPEDIANUM, IN THE CONNECTICUT RIVER, MASSACHUSETTS' Occurrence of a freshwater migration of the gizzard shad, Dorosoma cepedianum (Lesuer) (Clupeidae), is documented for the first time in a New England river system. Adult gizzard shad were observed and collected at the Connecticut River fishlift facility in Holyoke and upstream in Massachusetts during 1985 and 1986. It is believed that the Connecticut River migrants are derived from a population re- cently observed in Long Island Sound and already occurring in the Hudson and Connecticut River estuaries and Nan tic Bay. The gizzard shad is a widely distributed species occurring in marine and tidal freshwaters along the 'Contribution No. 104 of the Massachusetts Cooperative Fishery Research Unit, which is supported by the U.S. Fish and Wildlife Service, Massachusetts Division of Fisheries and Wildlife, Massa- chusetts Division of Marine Fisheries, and the University of Massachusetts. 380 FISHERY BULLETIN: VOL. 85, NO. 2, 1987. middle, southern, and gulf coasts of eastern North America (Megrey 1979). Landlocked freshwater populations are known from the Mississippi River drainage (Miller 1956; Megrey 1979) and the Great Lakes (Miller 1956, 1960). On the North American Atlantic coast, the gizzard shad has been reliably reported north to northern New Jersey and New York Harbor (Breder 1938; Miller 1956) (Fig. 1). Recent evidence indicates that the gizzard shad has ventured into the estuaries of certain major rivers draining into Long Island Sound. Dew (1974) reported that the species was first observed in the lower Hudson River estuary at Indian Point (river km 64.5) between 1969 and 1971 (Fig. 1). Subse- quent surveys suggested that the lower Hudson River population is increasing and that reproduc- tion was possibly occurring in the estuary (Dew 1974). However, George (1983) believed that the giz- zard shad in the lower Hudson River are derived from fish which migrated through the Erie Canal to the Mohawk River and down the Hudson River. If George's (1983) theory is correct then the lower Hudson River population would have been founded by landlocked freshwater animals, and not by mi- grating "anadromous" adventives from New York Harbor. Results and Discussion In the Connecticut River, adult gizzard shad were first observed near the mouth (river km 2.4, Fig. 1) in 1976 by commercial fishermen using gill nets set for American shad, Alosa sapidissima, (Whit- worth et al. 1980). In 1984 and 1985, gizzard shad Figure 1.— Recent reports of the gizzard shad, Dorosoma cepedianum, in New England: 1. Whitworth et al. (1980), Connecticut River, river km 2.4. 2. Gephard (text fn. 2), Connecticut River, river km 26. 3. O'Leary and Smith (this paper), Connecticut River, river km 139.4. 4. Gauthier (text fn. 3), Millstone nuclear power plant. 5. S. Henry (Assistant Aquatic Biologist, Massachusetts Division of Fisheries and Wildlife, Field Headquarters, Route 135, Westboro, MA 01581), Lawrence fishway, Merrimack River. 6. O'Leary and Smith (this paper), Connecticut River, Northampton Oxbow, river km 150. 381 were subsequently collected by fishermen using gill nets farther up the Connecticut River estuary (river km 26; S. Gephard pers. commun.^; Fig. 1) and en- trained at the Millstone nuclear power plant in Nan- tic Bay, CT (C. Gauthier pers. commun.^; Fig. 1). In October 1985, a single specimen was captured at the Lawrence fishway on the Merrimack River, Lawrence, MA. This specimen has been deposited into the Museum of Zoology, University of Massa- chusetts. During late May and June of 1985 and 1986, over 70 subadult gizzard shad were observed at the Holyoke Dam Fishlift on the Connecticut River in Holyoke, MA (river km 139.4) approximately 69 km above the head of the tide (Fig. 1). Four live and one dead gizzard shad— two females, two males, and one unknown— were collected at the fishlift; all have been deposited into the Museum of Zoology, Univer- sity of Massachusetts. Mean total length of the live fish was 418 mm (range 395-460 mm) and all were sexually mature. The mean total length of the live fish is near the maximum size reported for this species from freshwater (Miller 1960; Bodola 1965) and larger than the Mohawk River specimens dis- cussed by George (1983). Later in July 1986, a single juvenile gizzard shad (50 mm TL) was captured in the Northampton Oxbow of the Connecticut River (river km 150, T. Savoy pers. commun.*; Fig. 1). The specimen is in the collections of the Connecticut Department of Environmental Protection. A follow- up survey in September by O'Leary at the same locality produced no juveniles, but two small adults (300 and 348 mm TL) were captured and these two specimens have been divided among the Museum of Zoology, University of Massachusetts and the Museum of Comparative Zoology, Harvard Univer- sity. The collected juvenile specimen provides evidence that the species is breeding in the fresh- water portion of the Connecticut River, and the co-occurrence of adults suggests that the Northamp- ton Oxbow is an area where reproduction is occur- ring. Cooper (1983) suggested that the gizzard shad has been extending its range northward along the east coast of North America in response to warming climate. Whether the species has moved into the ^S. Gephard, Fishery Biologist, Marine Fisheries Office, State of Connecticut, Department of Environmental Protection, P.O. Box 248. Waterford, CT 06835, pers. commun. August 1985. 'C. Gauthier, Scientist, Northeast Utilities Environmental Laboratory, P.O. Box 128, Waterford, CT 06385, pers. commun. August 1985. ■•T. Savoy, Fishery Biologist, Marine Fisheries Office, State of Connecticut, Department of Environmental Protection, P.O. Box 248, Waterford, CT 06385, pers. commun. August 1986. Hudson River estuary (Dew 1974) while migrating northward or has entered the river from Lake Erie through the Mohawk River (Erie Canal) (George 1983) is unresolved. The species could have entered the Connecticut River only from the estuary as no inland connection between the Connecticut River and the Great Lakes or the Hudson River exists. The same argument would apply for the origin of other species encountered along the New England coast. The lack of any sightings of gizzard shad prior to 1985 at the Holyoke fishlift leads us to believe that the 1985 and 1986 migrations represent the first in- disputable movement into freshwaters of gizzard shad from a marine stock occurring off the southern New England coast. These findings support Cooper's (1983) contention that the gizzard shad is extending its range northward along the eastern North American coastline. Acknowledgments We thank Steven Gephard and Thomas Savoy, Connecticut Department Environmental Protection; Christine Gauthier, Northeast Utilities Service Com- pany; and Stephen Henry, Massachusetts Division of Fish and Game for providing information on giz- zard shad collections. We also thank Alan Richmond for assistance in the field. Literature Cited BODOLA, A. 1965. Life history of the gizzard shad, Dorosomxi cepedianum (Le Seur), in western Lake Erie. U.S. Fish. Wildl. Serv., Fish Bull. 65:391-425. Breder, C. M., Jr. 1938. The species of fish in New York Harbor. New York Zool. Soc. Bull. 41:23-29. Cooper, E. L. 1983. Fishes of Pennsylvania and the Northeastern United States. Pennsylvania State Univ. Press, University Park, 243 p. Dew, C. B. 1974. Comments on the recent incidence of the gizzard shad Dorosoma cepedianum, in the lower Hudson River. Third Symposium on Hudson River Ecology, Bear Mountain, NY, Pap. 20, 10 p. George, C. J. 1983. Occurrence of the gizzard shad in the lower Mohawk Valley. New York Fish Game J. 30:113-114. Megrey, B. a. 1979. Dorosoma cepedianum (Lesueur) gizzard shad. In D. S. Lee, C. R. Gilbert, C. H. Hocutt, R. E. Jenkins, D. E. McAlister, and J. R. Stouffer, Jr. (editors), Atlas of North American fishes, p. 69. North Carolina State Museum Natural History, Raleigh. Miller, R. R. 1956. Origin and dispersal of the alewdfe, Alosa pseudoharen- gtis, and the gizzard shad, Dorosoma cepedianum, in the 382 Great Lakes. Trans. Am. Fish. Soc. 86:97-111. 1960. Systematics and biology of the gizzard shad {Dorosoma cepedianum) and related fishes. U.S. Fish Wiidl. Serv., Fish Bull. 60:371-392. WHITWORTH, W. R., p. MlNTA, AND R. Orciari. 1980. Further additions to, and notes on, the freshwater ich- thyofauna of Connecticut. In P. M. Jacobson (editor), Studies of the ichthyofauna of Connecticut, p. 27-29. Storrs Agricultural Experiment Station, University of Connecticut, Storrs, Bull. 457. John O'Leary Department of Forestry and Wildlife Management University of Massachusetts Amherst, MA OlOOS Douglas G. Smith Museum of Zoology University of Massachusetts Amherst, MA 01003-0027 RELATIONSHIP OF OTOLITH LENGTH TO TOTAL LENGTH IN ROCKFISHES FROM NORTHERN AND CENTRAL CALIFORNIA Knowing the relationship between otolith length and total length of a fish is useful for two reasons: 1) Fish size can be estimated from otolith lengths measured from otoliths encountered in predator stomachs, in core samples, archaeological sites, etc., and 2) the length of a fish can be verified when the age determined from the otolith lies outside ex- pected values. The otolith/total length relationship is useful in predator-prey and archeological studies if fish size can be extrapolated from otolith length. Otoliths are often the only part of a prey fish remaining in a predator's gut (Ainley et al. 1981; Treacy and Crawford 1981) or at cooking sites of archeological middens (Fitch 1972). Fish lengths could be esti- mated from otoliths found as remains of prey or in coastal archaeological excavations (Fitch and Brownell 1968). Existing keys (e.g., Morrow 1979) allow identification of fish species from otoliths. With these keys, personal reference collections, and the length relationships described in this paper, in- vestigators will be able to verify species and size data collected in field sampling, and obtain more complete knowledge of prey species of marine mam- mals, birds, and fishes. Large-scale surveys, such as the California coop- erative survey (Sen 1984) that samples commercial rockfish landings in northern California, are prone to errors at several levels. Problems that may be encountered in collecting otoliths and measuring fish lengths include errors in recording lengths and the mixing up of otoliths. Some errors can be corrected by measuring the otolith and estimating the size of the fish it came from. Every effort should be made to eliminate erroneous data from the database before curves are constructed or cohort analysis is performed. In this paper, I report the results of my investi- gation of the relationship between otolith length and total length for 30 rockfish species of the genus Sebastes. Linear regression statistics are presented for all fish of the species encountered. Methods Specimens were collected during a life history study on the rockfishes of northern and central California conducted at the Southwest Fisheries Center Tiburon Laboratory. Fish were sampled from the commercial trawl fishery, the commercial sport fishery, skiffs, and research cruises from 1977 to 1980. Specimens were identified to species, and then total lengths of frozen— then thawed— car- casses were measured on a meter board in milli- meters (mm). Otoliths were measured to the near- est 0.1 mm with an ocular micrometer. The greatest length of the otolith was measured from the ante- rior tip to the most posterior projection (Kimura et al. 1979) (Fig. 1) as if the otolith were flat, without compensating for the curvature. Linear regressions were run on total length (y) versus otolith length (x) for 30 rockfish species. Outliers (±3.0 standard deviations) from the line were assumed to result from measurement or recording errors and were discarded (2% of the observations). Figure 1.— The length of an otolith is measured from the anterior tip to the posterior projection. FISHERY BULLETIN: VOL. 85, NO. 2. 1987. 383 Table 1 gives the sample size (N) and the mini- mum and maximum total lengths used in the anal- ysis for each species and each sex. Table 2 shows estimates of ^/-intercept (a), slope (6), standard er- ror of estimate (Sy^), correlation coefficient (r), and F for each species and sex. Analysis of covari- ance was used to determine if separate lines for males and females significantly reduced the variance from a common line (Kleinbaum and Kupper 1978). Analysis of covariance was also used to test for significant differences in the relationship of otolith length to total length between the sexes at the P = 0.05 level and the P = 0.01 level (Table 2). The highest values of r and examination of scattergrams (Fig. 2) indicate that the length relationships are linear over the observed range of values. Limiting the application of these regressions to the ranges of observed values is advised. Results and Discussion Linear regressions were run on each sex in order to investigate possible sexual differences. In 17 of the 30 species investigated, the relationship between otolith length and fish length is significantly differ- ent between males and females (Table 2). Sexual size dimorphism has been observed in 11 of the 17 species in Table 2. These species (plus S..alutus) in- clude most commercially and sport-caught rock- fishes in the northeastern Pacific Ocean. The six species for which growth curves have yet to be con- Table 1 —Sample sizes and size ranges used in the linear regres- sions of total lengtli versus otolith length for Sebastes. Measure- ments are in millimeters. Males Females Total length Total length Species of Mini- Maxi- Mini- Maxi- Sebastes N mum mum N mum mum auriculatus 34 257 477 44 179 523 aurora 27 203 378 44 230 398 camatus 100 112 289 103 109 279 caurinus 67 281 507 65 135 542 chlorostictus 73 155 450 101 162 458 chrysomelas 72 162 256 94 141 268 constellatus 54 186 422 45 177 430 crameri 42 206 445 47 134 505 diploproa 34 125 343 44 131 381 elongatus 25 188 326 73 135 378 entomelas 38 245 464 68 284 524 flavidus 163 254 504 221 232 539 goodei 26 227 385 52 227 556 hopkinsi 13 119 195 46 134 294 jordani 118 147 281 65 160 321 levis 14 267 773 15 237 900 maliger 13 317 481 21 226 478 melanops 120 334 534 89 197 607 melanostomus 34 250 442 46 297 538 miniatus 64 328 644 35 360 691 mystinus 141 248 480 63 213 375 nebulosus 25 270 391 23 257 500 ovalis 18 228 355 66 241 456 paucispinis 46 287 733 40 296 786 pinniger 92 249 585 81 251 622 rosaceus 72 212 426 75 203 310 ruberrimus 52 257 695 50 245 678 saxicola 29 141 240 73 159 358 semiclnctus 15 125 150 16 128 182 serranoides 60 235 469 70 229 528 500 450 E E X O z UJ 400 350 < O 300 - 250 - Female= ° Range a — a Male= • Range 12 13 14 15 16 OTOLITH LENGTH (mm) 17 18 Figure 2.— Linear regression of total length on otolith length of widow rockfish, Sebastes e7itomelas. The range of values for males (•-•) and females (o-o) at each whole millimeter of otolith length. 384 Table 2.— Results of linear regressions of total length (y) versus otolith length (x) for Sebastes. Measurements are in milli- meters. The F-test was run using the sums squared from the analysis of covariance comparing males and females; * - P = 0.05, ** -P = 0.01. Species of Sebastes Males Females r a b Sy. r a b Sy.x F crameri^-^ 0.926 19.270 24.629 22.054 0.988 -43.418 29.440 16.067 4.676* diploproa ^-^ 0.980 1.282 21.090 12.686 0.985 -22.120 23.717 13.390 6.601 •* entomelas^-^'* 0.871 -23.039 28.853 33.896 0.880 -51.835 32.452 28.968 6.243** flavidus^-^-^'^ 0.923 30.400 23.546 14.771 0.947 -12.604 26.901 18.389 25.637** goodei^ 0.975 1.696 23.866 12.741 0.987 -56.831 29.347 16.321 14.842** hopkinsi 0.868 20.734 20.172 9.567 0.951 -10.895 26.890 10.999 9.172** maliger 0.840 79.427 21.359 25.533 0.965 -105.649 33.479 19.259 4.221* melanops ^-^ 0.912 5.472 27.070 18.456 0.949 -124.076 35.784 19.480 25.338** melanostomus 0.907 -21.094 25.187 21.777 0.918 -21.713 26.211 23.405 8.441** miniatus 0.961 -42.615 28.385 23.399 0.971 -72.278 30.607 24.103 4.638* mystinus^'^ 0.910 -2.255 28.987 23.112 0.881 60.010 22.054 14.965 3.643** ovalis 0.903 26.185 24.964 13.268 0.963 -27.207 31.859 14.144 21.995** paucisplnis ' ^ 0.893 -51.911 38.441 55.433 0.931 -102.932 43.993 56.501 5.965** pinniger ' '^'^ 0.950 -61.508 27.663 25.655 0.967 -95.891 30.455 29.004 8.200** saxicola'' 0.928 -0.111 19.074 7.544 0.974 -19.565 22.495 12.797 11.412** semicinctus 0.880 34.331 16.015 4.061 0.938 12.022 21.372 4.981 20.081** serranoides ^^ 0.967 -5.307 25.898 13.462 0.969 -63.475 30.445 20.668 10.318** 'Westrheim and Marling 1975. "Lenarz 1987. 'Mille r and Geibel 1973. 2Shaw and Archibald 1981, 5Six and Morton 1977. 8Wilkins 1980. ^Boehlert and Kappenman 1980. eWyllie Echeverria 1986. ^Love and Westphal 1981. structed may also show sexually size-dimorphic growth. The 13 species with no difference noted between males and females consist primarily of two closely related taxonomic groups (Barsukov 1981). Few growth studies exist for these species. The first group of shallow, nearshore species is represented in this study by S. auriculatus, S. carnatus, S. caurinus, S. chrysomelas, and S. nebulosus. The growth curve for S. chrysomelas is the same for males and females (Zaitlan 1986). The second group is the subgenus Sebastomus (Chen 1971), repre- sented in this study by S. chlorostictus, S. con- stellatus, and 5. rosaceus. Growth curves exist for two members: S. helvomaculatus (Westrheim and Harling 1975) and 5. umbrosus (Chen 1971), which do not show sexual size dimorphism. The indications are relationships of otolith length to total length reflect the age-at-length relationship between the sexes. In food-habit studies, otoliths are often found but the sex and length of the fish are not known. Table 3 shows regressions for the combined sexes for those occasions when the sex is unknown or when the regressions were not significantly different between the sexes. Data analysis for S. entomelas shows a potential to derive estimates of age from otolith lengths (Fig. 3). The calculated total lengths for males and females for each 1 mm increment in otolith length are overlaid on the age-length curve (from Lenarz 1987). These relationships are species-specific and Table 3. — Results of linear regressions of total length (/) versus otolith length (x) for Sebastes for sexes com- bined. Measurements are in millimeters. Species of Sebastes a auriculatus 0.968 aurora 0.782 carnatus 0.945 caurinus 0.906 ctilorostictus 0.974 chrysomelas 0.919 constellatus 0.978 crameri 0.971 diploproa 0.980 elongatus 0.974 entomelas 0.898 flavidus 0.938 goodei 0.983 hopkinsi 0.957 jordani 0.985 levis 0.973 maliger 0.928 melanops 0.930 melanostomus 0.928 miniatus 0.962 mystinus 0.912 nebulosus 0.891 ovalis 0.952 paucisplnis 0.903 pinniger 0.957 rosaceus 0.902 ruberrimus 0.957 saxicola 0.977 semicinctus 0.924 serranoides 0.965 'y.x -53.032 33.159 17.729 15.124 19.910 24.818 -39.365 30.573 10.258 5.099 30.234 26.291 -18.537 24.113 14.898 -21.780 28.609 9.020 -37.484 25.266 13.123 -27.098 28.104 19.912 -12.854 22.635 14.020 -13.564 24.020 12.284 -6.890 33.113 32.247 -10.946 26.506 18.158 -57.996 29.129 17.819 -30.546 28.868 12.168 -2.313 22.096 7.353 -170.108 47.458 46.975 -53.107 29.967 23.862 -48.222 30.557 21.002 -47.070 27.362 23.590 -56.738 29.365 24.516 -18.175 29.765 23.204 32.970 25.181 16.131 -53.472 33.562 17.179 -77.089 41.089 59.143 -85.114 29.411 28.398 -83.484 22.533 10.908 -76.233 31.328 31.206 -32.765 23.399 12.663 -19.182 25.266 6.961 -51.013 29.350 18.965 385 11 12 13 14 15 520 I — \ — I — I 1 r OTOLITH LENGTH (mm) 16 480 E E - 440 X I- o 2 400 ILI o 360 320 FEMALE -r 17 — r— 280 I ^ ' ' I I I I I I 1 I I ' I I I I I L. 3 5 7 9 11 13 15 17 19 21 23 AGE (yr) Figure 3.— Age-length curve for widow rockfish, Sebastes ento- vielas (from Lenarz 1987). The calculated total length from otolith length is overlaid on the curve to obtain an estimate of age. should be used within well-defined limits. The scattergram (Fig. 2) with the mean and range of total length found at each 1 mm otolith length incre- ment indicates the ranges within which these data are useful. Some problems in relating otolith length to age include the increased range of fish lengths at older ages and the observed thickening-instead of lengthening of otoliths in Sebastes (Boehlert 1985). These results may be used to estimate total length from an otolith length as shown in the following ex- ample. If the otoliths are from fish of unknown sex, the regression statistics from Table 3 would be used to estimate fish length. If the otoliths are from fish of known sex, Table 2 would be consulted. If a species appears in Table 2, the regression statistics for the appropriate sex would be used to estimate fish length. If a species does not appear in Table 2, Table 3 (with regression statistics for males and females combined) would be used. For instance, to estimate fish length from otolith length (OL) for male S. auriculatus, the regression statistics from Table 3 are used. An otolith 10.0 mm long gives an estimated total length of TL = a + 6 (OL) TL = -53.032 + 33.159(10.0) TL = 279 mm. Tables have been constructed with the regression statistics presented here. The table for each species (and sex, where appropriate) represents otolith lengths measured in millimeters and the correspond- ing estimated total length. These tables are avail- able on request from the author. Acknowledgments I wish to thank Sharon Moreland for assistance in measuring the otoliths, and David Woodbury and Carol Reilly for editing the data and running the computer programs. The reviews by George Boeh- lert, Bill Barss, and an anonymous reviewer were very helpful. Literature Cited AiNLEY, D. G., D. W. Anderson, and P. R. Kelly. 1981. Feeding ecology of marine cormorants in southwestern North America. Condor 83:120-131. Barsukov, V. V. 1981 . A brief review of the subfamily Sebastinae. J. Ichthyol. 2(l):l-26. (Engl. Transl. Vopr. Ikhtiol.) Boehlert, G. W. 1985. Using objective criteria and multiple regression models for age determination in fishes. Fish. Bull., U.S. 83:103- 117. Boehlert, G. W., and R. F. Kappenman. 1980. Variation of growth with latitude in two species of rock- fish (Sebastes pinniger and S. diploproa) from the northeast Pacific Ocean. Mar. Ecol. Prog. Ser. 3:1-10. Chen, L.-C. 1971. Systematics, variation, distribution, and biology of rockfishes of the subgenus Sebastomus (Pisces, Scorpae- nidae, Sebastes). Bull. Scripps Inst. Oceanogr. Univ. Calif. 18, 115 p. Fitch, J. E. 1972. Fish remains, primarily otoliths, from a coastal Indian midden (SLO-2) at Diablo Cove, San Luis Obispo County, California. San Luis Obispo Cty. Archaeol. Soc. Occas. Pap. No. 7:101-120. Fitch, J. E., and R. L. Brownell, Jr. 1968. Fish otoliths in cetacean stomachs and their importance in interpreting feeding habits. J. Fish. Res. Board Can. 25:2561-2574. KiMURA, D. K., R. R. Mandapat, and S. L. Oxford. 1979. Method, validity, and variability in the age determina- tion of yellowtail rockfish (Sebastes flavidus), using otoliths. J. Fish. Res. Board Can. 36:377-383. Kleinbaum, D. G., and L. L. Kupper. 1978. Applied regression analysis and other multivariable methods. Duxbury Press, North Scituate, MA., 556 p. Lenarz, W. H. 1987. Aging and growth of widow rockfish. /w W. H. Lenarz and D. R. Gunderson (editors). Widow rockfish: Proceedings of a Workshop, Tiburon, California, December 11-12, 1980, p. 31-35. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 48. Love, M. S., and W. V. Westphal. 1981. Growth, reproduction, and food habits of olive rockfish, Sebastes serranoides, off central California. Fish. Bull., U.S. 79:533-545. Miller, D. J., and J. J. Geibel. 1973. Summary of blue rockfish and lingcod life histories; a reef ecology study; and giant kelp, Macrocystis pyrifera, ex- periments in Monterey Bay, California. Calif. Dep. Fish Game, Fish Bull. 158, 137 p. 386 Morrow, J. E. 1979. Preliminary keys to otoliths of some adult fishes of the Gulf of Alaska, Bering Sea, and Beaufort Sea. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 420, 32 p. Sen, a. R. 1984. Sampling commercial rockfish landings in California. U.S. Dep. Commer., NOAA Tech. Memo. NOAA-TM-NMFS- SWFC-45, 95 p. Shaw, W., and C. P. Archibald. 1981. Length and age data of rockfishes collected from B.C. coastal waters during 1977, 1978, and 1979. Can. Data Rep. Fish. Aquat. Sci. 289, 119 p. Six, L. D., and H. F. Horton. 1977. Analysis of age determination methods for yellowtail rockfish. canary rockfish, and black rockfish off Oregon. Fish. Bull., U.S. 75:405-414. Treacy, S. D., and T. W. Crawford. 1981. Retrieval of otoliths and statoliths from gastrointes- tinal contents and scats of marine mammals. J. Wildl. Manage. 45:990-993. Westrheim, S. J., and W. R. Harling. 1975. Age-length relationships for 26 scorpaenids in the northeast Pacific Ocean. Can. Fish. Mar. Serv. Tech. Rep. 565, 12 p. WiLKINS, M. E. 1980. Size composition, age composition, and growth of chili- pepper, Sebastes goodei, and bocaccio, S. paucispinis, from the 1977 rockfish survey. Mar. Fish. Rev. 42(3-4):48-53. W'VXLIE Echeverria, T. 1986. Sexual dimorphism in four species of rockfish genus Sebastes (Scorpaenidae). Environ. Biol. Fish. 15:181-190. Zaitlan, J. A. 1986. Geographical variation in the life history of Sebastes chrysomelas. M.A. Thesis, San Francisco State Univ., San Francisco, CA, 87 p. Tina Wyllie Echeverria Southwest Fisheries Center Tiburon Laboratory National Marine Fisheries Service, NOAA 3150 Paradise, Drive, Tiburon. CA 9U920 CRATER WOUNDS ON NORTHERN ELEPHANT SEALS: THE COOKIECUTTER SHARK STRIKES AGAIN A variety of wounds are observed on northern ele- phant seals, Mirounga angustirostris. We report a new type of wound observed on juveniles, primarily from the Mexican islands west of Baja California and rarely from off California. The form and shape of these wounds, and their similarity to wounds re- ported from other marine mammals, fishes, and squids, suggest that they were caused by a small, squaloid shark of the genus Isistius, commonly known as the cookiecutter or cigar shark. The shape of wounds, their location on the victim's body, the time of the year that the wounds are received, and the age of the seal provide a good in- dication of the cause. During the breeding season, for example, suckling seal pups bear bite marks on the snout, head, and rump, these having been in- flicted by adult females (Le Boeuf and Briggs 1977). Weaned pups and adult females bear fresh bite marks of varying severity caused by adult males biting their necks while attempting to mate with them, and breeding-age males inflict a variety of bite wounds on each other during fights to establish dominance (Le Boeuf and Reiter in press). During winter and spring, Mirounga angustirostris of both sexes and all ages exhibit fresh wounds inflicted by white sharks. Car char odon carcharias. The shape and serrated edges of those wounds are easily distinguished from the smooth-edged and halfmoon- shaped wounds caused by boat propellers (Le Boeuf et al. 1982; Tricas and McCosker 1984). The wounds that we discovered were round, hollowed-out craters, smooth edged at the margin, about the size of a tennis ball, and unlike any of the wounds described above. The similarity in appear- ance of these wounds to scars inflicted by Isistius upon cetaceans (Van Utrecht 1959) and fishes (Jones 1971) implicate the cookiecutter shark as the prob- able cause. The only reported eastern Pacific occur- rence of an Isistius is that of an /. hrasiliensis from off the Galapagos (Compagno 1984). However, we have examined additional eastern Pacific specimens of/, hrasiliensis, including a specimen from off Isla de Guadalupe. Background information. Northern elephant seals inhabit traditional island and mainland sites from mid-Baja California, Mexico, to central California. Their range at sea along the Pacific coast is from Isla Cedros, Mexico, to the southern Aleutians. Feeding occurs beyond the continental shelf in deep water (Le Boeuf et al. 1986). It is not known how far from shore they go to feed, but some animals have been seen as far as 3,000 miles away on Mid- way Island in the mid-Pacific (Condit and Le Boeuf 1984). Several islands are used regularly throughout the year (Guadalupe, San Benito, Cedros, and Co- ronoados in Mexico and San Miguel, San Nicolas, Ano Nuevo, and the Farallones); the sex and age composition of each colony varies with time of year (Le Boeuf and Bonnell 1980). Late August or early September, when most of the observations reported in this paper were made, is the end of the molt period for adult and subadult males and the begin- ning of the fall haul-out for juveniles, 1-4 years old. Breeding-age males, observed on land at this time, are completing the annual molt, a process that takes FISHERY BULLETIN: VOL. 85, NO. 2, 1987. 387 30-40 days; they are in the process of returning to sea to feed. As their number decUnes, juveniles of both sexes begin to haul-out in increasing numbers. They have been at sea for 4 months or more. Cen- sus counts of total northern elephant seals are lower at this time than at any other time of the year (Le Boeuf and Bonnell 1980). Observations and Methods Most of the observations were made during an ex- pedition to Mexican islands aboard the MV Mirage from 20 to 31 August 1986. Islands surveyed in- cluded Isla de Guadalupe, Islas San Benito, Cedros, San Martin, and Los Coronados. We censused north- ern elephant seals at all sites and, in doing so, recorded the incidence of fresh wounds. On 2 September 1986, we censused and recorded wounds on northern elephant seals at Ano Nuevo Island off central California. Similar observations and cen- suses were conducted weekly at Ano Nuevo Island during October and November, when peak numbers of juveniles are observed. Censuses and inspection for wounds were made from an inflatable 6 m boat, approximately 10 m from seals lying on sandy beaches near the water's edge, or from on foot to get closer to the animals. WTien possible, we inspected both sides of all seals; we made no attempt to turn animals over to inspect the ventrum or to arouse them to better inspect them for wounds. Approximately half of the animals counted were seen from only one side. Thus, the counts of wounded animals we present are clearly underestimates of the true figure. We noted the location of all wounds and estimated their size and freshness. We examined preserved specimens of Isistius housed in the Marine Vertebrates Division of the Scripps Institution of Oceanography (SIO). Results We observed fresh wounds on 20 juvenile northern elephant seals on three of the five island groups in- spected in August and September (Table 1); there were no elephant seals on Isla San Martin. Four ad- ditional wounded juveniles were observed on Ano Nuevo Island later in the year. All wounds were fresh, as indicated by their bloody color, and, with one exception, they were of similar size and shape (Fig. 1). The wounds were round and hollowed-out craters; the margin of each wound was smooth. Each wound was about 5-6 cm wide and 3-5 cm deep. One wound, although like the others in most respects, had a flap of skin and blubber still attached. No fresh crater wounds were observed on adult and subadult males. Most animals had one wound. Wounds were located on various parts of the body (Fig. 1): the side posterior to the flippers, on the ventrum or the back and to either side or on the midline, on the chest and neck, and just behind the ear. Two animals had two wounds and one had three. One animal had two fresh, identical wounds on the dorsal midline at the level of the foreflippers, separated by approximately 3 cm. Another had two wounds 0.3 m apart on its left side. One animal had three wounds: two on the abdomen and one on the ventral surface of the neck. The incidence of fresh wounds was highest on northern elephant seals inhabiting Isla de Guada- lupe (8.4% of the juveniles censused) followed by Isla Cedros and Islas San Benito (Table 1). No wounded Table 1 .—Proportion and percentage of fresh crater wounds on northern elephant seals censused on various Mexican and Caiifornlan islands during August and September 1986. Adult and Date Island Beach or Islet subadult males Juveniles 21-24 Aug. Guadalupe Pilot Rock Beach 0/34 = 0 7/95 = 7.37 Barracks Beach 0/19 = 0 3/35 = 8.57 Twin Canyons 0/10 = 0 6/61 = 9.84 Sum: all beaches 0/63 = 0 16/191 = 8.38 25-27 Aug. San Benito Este 0/10 = 0 1/57 = '1.75 Centre 0/38 = 0 1/127 = 0.79 Oeste 0/6 = 0 0/51 = 0 Sum: all islets 0/54 = 0 2/235 = 0.85 28-29 Aug. Cedros 0/13 = 0 2/45 = 4.44 31 Aug. Los Coronados 0/1 = 0 0/14 = 0 2 Sept. Ano Nuevo 0/23 = 0 0/200 = 0 ^Over 90 juveniles were counted but only 57 were observed close enough to document wounds. 388 animals were observed on Los Coronados during August or on Ano Nuevo Island during August and September. However, four juveniles with fresh wounds, among 700 juveniles present, were ob- served on Ano Nuevo Island during four censuses in November (1. 9, and 30 November). One animal, sighted on 1 November 1986, was marked; a 22-mo- old juvenile born on Ano Nuevo Point on 11 Feb- ruary 1984 and tagged 1 month later. As mentioned above, through examination of the holdings of the Scripps Institution of Oceanography Fish Collection, we uncovered additional Pacific specimens of Isistius brasiliensis. The eight speci- mens from seven lots included six males and two females. The largest, a 470 mm (standard length) female (SIO 69-345) with jaw width of 38 mm, was collected by IKMT (Isaacs-Kidd midwater trawl) be- tween the surface and 2,000 m from north of Easter Island (lat. 25°58.5'S, long. 108°50.7'W). Another eastern Pacific specimen (SIO 78-183) is from off Isla de Guadalupe (29°26.5'N, 119°44'W) and was collected by phytoplankton net. The other eastern Pacific specimen (SIO 52-413) is from west of the Galapagos (00°00", 100°00"W) and was captured at the surface by dip net. Discussion The fresh crater wounds we observed on juvenile northern elephant seals resemble those reported on beaked whales, sperm whales, several species of por- poises, and most of the baleen whales (Mackintosh and Wheeler 1929; Van Utrecht 1959), as well as those from a variety of pelagic fishes (Jones 1971) and a nuclear submarine (Johnson 1978). Jones (1971) and others have demonstrated conclusively that those wounds are the result of bites inflicted by the small squaloid shark, Isistius brasiliensis, or possibly by its congener, /. plutodus. To date, the only Pacific record of/, plutodus is from off Okinawa (Compagno 1984), so we therefore presume that the more wide-ranging and topotypical /. brasiliensis is the culprit. Isistius brasiliensis is epipelagic to bathypelagic and is known from all tropical oceans, extending northward to off Japan and Baja Califor- nia and southward to Lord Howe Island. It is typical- ly caught by midwater trawl at depths between 85 and 3,500 m; however, it is occasionally found at the surface at night. The shark is thought to be a diur- nal vertical migrator, perhaps traveling a distance as great as 2,000-3,000 m in each direction; in so doing, it apparently encounters feeding Mirounga. As noted by Compagno (1984, p. 94), Isistius is highly specialized as a facultative ectoparasite in its dentition, suctorial lips, and modified pharynx that allow it to attach to the side of large prey, drive its sawlike lower jaw teeth into the skin and flesh of its victim, cut a conical plug of flesh, and then pull itself free with the plug cradled by its scooplike lower jaw and held by the hooklike upper jaw teeth. The scar patterns of juvenile Mirounga support the scenario described above. A comparison of jaw width of Isistius of known size with the scar patterns observed on Mirounga suggests that the attacking sharks were at least 50-60 cm long. Northern elephant seals would appear to be easy prey for Isistius. They are slow swimmers, com- pared with large pelagic fishes, and they spend 85% of their time at sea underwater at depths of 400-650 m (LeBoeuf et al. 1985; Le Boeuf et al. 1986). Juvenile seals that use Isla de Guadalupe during the fall are evidently most prone to being parasitized. Juvenile seals hauling out on other islands, especially those to the north, are evidently not exposed to Isistius to the same degree. Until recently, no fresh crater type wounds were observed on seals at Ano Nuevo despite 16 years of observations by B. J. Le Boeuf. Some wounds observed may have been old, healing crater type wounds, suggesting that the animals bearing them may have been immigrants from the south, the predominant direction of disper- sal (Bonnell et al. 1979). Fresh crater wounds have not been observed on the northern elephant seals at the Farallones since their breeding began in 1972 (H. Huber, pers. commun.^). Le Boeuf never ob- served fresh crater wounds on northern elephant seals breeding on San Miguel and San Nicolas Islands during 1968-78, despite annual visits to these islands. Mexican northern elephant seals of juvenile age fall prey to Isistius shortly before they haul-out in late August. Le Boeuf and coworkers never ob- served crater wounds on seals at these Mexican rookeries during the winter breeding season (13 visits since 1968) or summer molt (4 visits). The juveniles may be exposed to Isistius while feeding or while returning to the island. The marked dif- ference in distribution of shark wounds is consistent with the observation that Mexican, southern Califor- nian, and central Californian juveniles feed in dif- ferent locations, and each "subpopulation" feeds north of its birthplace (Condit and Le Boeuf 1984). Juveniles, trapped in fishing gear, have been caught around 200 m below the surface and captured 16-224 km offshore. iR. Huber, Point Reyes Bird Observatory, 4900 Shoreline Highway, Stinson Beach, CA 94970. 389 Pf?i- w Figure 1.— Representative wounds on Mi- rounga angustirostris caused by Isistius at- tacks. A, B, C— taken at Isla de Guadalupe, Mexico; D, E— taken at Isla Este, Islas San Benito, Mexico; F— taken at Ano Nuevo Island, California. Not visible in F are two other healing wounds along the animal's right flank. (Photos A-E by B. J. Le Boeuf; photo F by P. Thorson.) 390 m ^^**- ■■^st-'--^. 391 That only juvenile northern elephant seals exhibit fresh crater wounds may be explained in several ways. It suggests that this age category is the only one exposed by depth or location to feeding Isistius; or, it may suggest that older age classes are able to avoid attack. Another interesting hypothesis con- cerns the common prey of both Isistius and Mirounga, midwater squid. It has been speculated that the bioluminescent pattern of Isistius might simulate the pattern of a large midwater squid and thereby attract squidophagus predators (Jones 1971) upon which it could prey. It seems unlikely that an Isistius could outswim a Mirounga; attacks by the shark would thus be accomplished either by attract- ing the seal, perhaps for a closer inspection of the shark, or by attacking the seal by stealth and sur- prise. The location of attack scars on the head region of Mirounga would indicate a frontal approach, whereas the scars on the back and flanks might in- dicate that the seal was unaware of the impending attack. Both scenarios are likely. It is also possible that juveniles seals are more readily attracted to the display of Isistius, not having learned yet to distinguish them from squid. Further insight into these hypotheses will be provided as more data con- cerning the spatial and temporal distribution of Isistius are collected. Acknowledgments We thank Al Giddings of Ocean Images and the National Geographic Society for financing this ex- pedition; Doc White of the Whitewater Enterprises; the crew of the MV Mirage for the logistics, camaraderie, and good food; R. H. Rosenblatt of the Scripps Institution of Oceanography for making available the specimens of Isistius; and an anony- mous reviewer for advice and comments. This pro- gram was supported in part by the National Science Foundation grant BSR-8605000 to B. J. Le Boeuf. Literature Cited BONNELL, M. L., B. J. Le Boeuf, M. 0. Peirson, D. H. Dett- MAN, AND G. D. FARRENS. 1979. Summary report 1975-1978. Marine Mammal and Seabird Surveys of the Southern Cahfornia Bight Area. Vol. III. Pinnipeds. Bur. Land Manage., Dep. Inter., Contract AA.550-CT7-36, 535 p. Compagno, L. J. V. 1984. Sharks of the world. An annotated and illustrated catalogue of shark species known to date. Part 1. Hexan- chiformes to Lamniformes. FAO Fish. Synop. 4(1):1- 249. CoNDiT, R. AND B. J. Le Boeuf. 1984. Feeding habits and feeding grounds of the northern elephant seal. J. Mammal. 65:281-290. Johnson, C. S. 1978. Sea creatures and the problem of equipment damage. U.S. Naval Inst. Proc, p. 106-107. Jones, E. C. 1971. Isistius brasiliensis, a squaliod shark, the probable cause of crater wounds on fishes and cetaceans. Fish. Bull., U.S. 69:791-798. Le Boeuf, B. J., and M. L. Bonnell. 1980. Pinnipeds of the California islands: abundance and distribution. In D. M. Power (editor). The California islands: Proceedings of a Multidisciplinary Symposium, p. 475-493. Santa Barbara Museum of Natural History, Santa Barbara, CA. Le Boeuf, B. J., and K. T. Briggs. 1977. The cost of living in a seal harem. Mammalia 41: 167- 195. Le Boeuf, B. J., and J. Reiter. In press. Lifetime reproductive success in northern elephant seals. In T. H. Clutton-Brock (editor). Reproductive Suc- cess. Univ. Chicago Press, Chicago. Le Boeuf, B. J., M. Reidman, and R. S. Keyes. 1982. White shark predation on pinnipeds in California coastal waters. Fish. Bull., U.S. 80:891-895. Le Boeuf, B. J., D. P. Costa, and A. C. Huntley. 1985. Diving behavior of northern elephant seals. 6th Bien- nial Conference on Marine Mammals, 22-26 November 1985, Vancouver, British Columbia. Le Boeuf, B. J., D. P. Costa. A. C. Huntley, G. L. Kooyman, and R. W. Davis. 1986. Pattern and depth of dives in northern elephant seals, Mirounga angustirostris. J. Zool. Lond. 208:1-7. Mackintosh, N. A., and J. F. G. Wheeler. 1929. Southern blue and fin whales. Discovery Rep. 1:257-540. Tricas, T. C, and J. E. McCosker. 1984. Predatory behavior of the white shark (Carcharodon carcharias), with notes on its biology. Proc. Calif. Acad. Sci., 43(14):221-238. Van Utrecht, W. L. 1959. Wounds and scars on the skin of the common porpoise, Phocaena phocaena (L.). Mammalia 23:100-122. BuRNEY J. Le Boeuf Institute of Marine Sciences and Department of Biology, University of California, Santa Cruz. CA 95061^ California Academy of Sciences, Golden Gate Park, San Francisco, CA 9Itll8 John E. McCosker John Hev^^itt 392 NOTICES NOAA Technical Reports NMFS published from March to December 1986. Technical Report NMFS 37. A Histopathic evaluation of gross lesions excised from commercially important North American marine fishes. By Robert A. Murchelano, Linda Despres-Patanjo, and John Ziskowski. March 1986, iii + 14 p., 13 figs., 4 tables. 38. Fishery atlas of the Northwestern Hawaiian Islands. By Richard N. Uchida and James H. Uchiyama (editors). September 1986, v + 142 p., 75 figs., 5 tables. 40. The potential impact of ocean thermal energy conversion on fisheries. By Edward P. Myers, Donald E. Hoss, Walter M. Matsumoto, David S. Peters, Michael P. Seki, Richard N. Uchida, John D. Ditmars, and Robert A. Paddock. June 1986, iii + 33 p., 11 figs., 8 tables. 41. A stationary visual census technique for quantitatively assessing com- munity structure for coral reef fishes. By James A. Bohnsack and Scott P. Bannerot. July 1986, iii + 15 p., 13 figs., 3 tables. 42. Effects of temperature on the biology of the northern shrimp, Pandalus borealis, in the Gulf of Maine. By Spencer Apollonio, David K. Steven- son, and Earl E. Dunton, Jr. September 1986, iii -i- 22 p., 25 figs., 7 tables. 43. Environment and resources of seamounts in the North Pacific. By Richard N. Uchida, Sigeiti Hayasi, and George W. Boehlert (editors). September 1986, v -i- 105 p. Foreword. By Richard N. Uchida, Sigeiti Hayasi, and George W. Boehlert, p. v. Seamounts: A biological concourse in the open sea. An introductory statement. By Richard S. Shomura, p. 1-2. Aspects of oceanic flow and thermohaline structure in the vicinity of seamounts. By Gunnar I. Roden, p. 3-12, 12 figs. Oceanographic studies of seamounts. By Hajime Yamanaka, p. 13-17. Session 1. Summary. By Richard N. Uchida and Sigeiti Hayasi, p. 19. Development and present status of Japanese trawl fisheries in the vicinity of seamounts. By Takashi Sasaki, p. 21-30, 8 figs., 3 tables. Review and present status of handline and bottom longline fisheries for alfonsin. By Michael P. Seki and Darryl T. Tagami, p. 31-35, 4 figs., 1 table. Albacore, Thunniis alalunga, pole-and-line fishery around the Emperor seamounts. By Minato Yasui, p. 37-40, 1 fig., 1 table. Session 2. Review of seamount fisheries. By Richard N. Uchida and Sigeiti Hayasi, p. 41. Precious corals: An important seamount fisheries resource. By Richard W. Grigg, p. 43-44, 1 table. Fish and crab populations of Gulf of Alaska seamounts. By Miles S. Alton, p. 45-51, 6 figs., 4 tables. Session 3. Summary. By Richard N. Uchida and Sigeiti Hayasi, p. 53. Review and current status of research on the biology and ecology of the genus Pseudopentaceros. By Robert L. Humphreys, Jr. and Darryl T. Tagami, p. 55-62, 5 figs., 2 tables. A seamount survey around Izu Islands. By Hiroyo Koami, p. 63-66, 4 figs. Zoogeographical features of fishes in the vicinity of seamounts. By Eiichi Fujii, p. 67-69, 2 figs., 1 table. Session 4. Summary. By Richard N. Uchida and Sigeiti Hayasi, p. 71. Problems in assessing the pelagic armorhead stock on the central North Pacific seamounts. By Jerry A. Wetherall and Marian Y. Y. Yong, p. 73-85, 7 figs., 1 table. 393 A review of the fishery and catch-per-cruise for alfonsin stocks in the vicinity of Izu Islands. By Koichi Yamamoto, p. 87-91, 6 figs. Session 5. Summary. By Richard N. Uchida and Sigeiti Hayasi, p. 93. Productivity and population maintenance of seamount resources and future research directions. By George W. Boehlert, p. 95-101, 1 fig., 1 table. Session 6. Summary. By Richard N. Uchida and Sigeiti Hayasi, p. 103. 44. Synopsis of biological data on the porgies, Calamus arctifrons and C. proridens (Pisces: Sparidae). By George H. Darcy. September 1986, iii -f 19 p., 26 figs., 8 tables. 45. Meristic variation in Sebastes (Scorpaenidae), with an analysis of character association and bilateral pattern and their significance in species separa- tion. By Lo-chai Chen. September 1986, ii -i- 17 p., 16 tables. 46. Distribution and relative abundance of pelagic nonsalmonid nekton off Oregon and Washington, 1979-84. By Richard D. Brodeur and William G. Pearcy. December 1986, iii -i- 85 p., 60 figs., 6 tables. Some NOAA publications are available by purchase from the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. 394 ERRATUM Fishery Bulletin: Vol. 85, No. 1 Cover page 2 should be corrected as follows: SCIENTIFIC EDITORS, Fishery Bulletin Dr. William J. Richards Dr. Andrew E. Dizon Southeast Fisheries Center Miami Laboratory Southwest Fisheries Center La Jolla Laboratory National Marine Fisheries Service, NOAA National Marine Fisheries Service, NOAA Miami, FL 33149-1099 La Jolla, CA 92038 The Scientific Publications Office congratulates the former Scientific Editor William J. Richards and his staff for a job well done! 698 fr 07 0 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 address, 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 American Fisheries Society Special Publication No. 12, A List of Com- mon and Scientific Names of Fishes fi^om the United States and Canada, Fourth Edition, 1980. Tfext footnotes should be typed separately from the text. 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 5% 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 Consistency in headings and format is desirable Vertical rules should be avoided, as they make the tables more expensive to print. Footnotes in tables should be numbered sequentially in arable numerals. Tb avoid confusion with powers, they should be placed to the left of numerals. 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 sur- name under the heading "Literature Cited." Only the author's surname and initials are required in the literature cited. The accuracy of the literature cited is the respon- sibility of the author. Abbreviations of names of periodicals and serials should conform to Serial Sources for the BIOSIS Data Base. (Chemical Abstracts 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), %, Voo, and so forth, should be used. Abbreviate units of measure only when used with numerals. Periods are only rarely used with abbrevia- tions. We prefer that measurements be given in metric units; other equivalent units may be given in parentheses. FORM OF THE MANUSCRIPT The original of the manuscript should be typed, double- spaced, on white bond paper. Please triple space above headings. We would rather receive good duplicated copies of manuscripts than carbon copies. The sequence 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 Notes BOROWSKY, BETTY. Laboratory studies of the pattern of reproduction of the isopod crustacean Idotea baltica 377 O'LEARY, JOHN, and DOUGLAS G. SMITH. Occurrence of the first freshwater migration of the gizzard shad, Dorosoma cepedianum, in the Connecticut River, Massachusetts 380 WYLLIE ECHEVERRIA, TINA. Relationship of otolith length to total length in rock- fishes from northern and central California 383 LE BOEUF, BURNEY J., JOHN E. McCOSKER, and JOHN HEWITT Crater wounds on northern elephant seals: the cookiecutter shark strikes again 387 Notices— NO AA Technical Reports NMFS published from March to December 1986. . . 393 • GPO 791-008 A^' °^Co Fishery Bulletin ^^ATES O^ [; Marino Bioiogical Laboratory — NOVil!^ 1007 Vol. 85, No. 3 ^ Woods Hole, Mass, ''"'^ ^^^'' MUGIYA, YASUO. Phase dfference between calcification and organic matrix 'for- mation in the diurnal growth of otoliths in the rainbow trout, Salmo gairdneri BEST, PETER B. Estimates of the landed catch of right (and other whalebone) whales in the American fishery, 1805-1909 HOLT, RENNIE S. Estimating density of dolphin schools in the eastern tropical Pacific Ocean by line transect methods HOLT, RENNIE S., TIM GERRODETTE, and JOHN B. COLOGNE. Research vessel survey design for monitoring dolphin abundance in the eastern tropical Pacific BUTLER, RICHARD W., WALTER A. NELSON, AND TYRRELL A. HEN- WOOD. A trawl survey method for estimating loggerhead turtle, Caretta caretta , abundance in five eastern Florida channels and inlets GROOT, C, and T. P. QUINN. Homing migration of sockeye salmon, Oncorhynch us nerka , to the Eraser River BECKER, D. SCOTT, and KENNETH K. CHEW. Predation on Capitella spp. by small-mouthed pleuronectids in Puget Sound, Washington HINCKLEY, SARAH. The reproductive biology of walleye pollock, Theragra chalcogramma , in the Bering Sea, with reference to spawning stock struc- ture KENDALL, A. W., JR., M. E. CLARKE, M. M. YOKLAVICH, and G. W. BOEH- LERT. Distribution, feeding, and growth of larval walleye pollock, Theragra chalcogramma , from Shelikof Strait, Gulf of Alaska ARMETTA, THERESE M., and BRADLEY G. STEVENS. Aspects of the biology of the hair crab, Erimacrus isenbeckii , in the eastern Bering Sea WENNER, ELIZABETH L., GLENN F. ULRICH, and JOHN B. WISE. Exploration for golden crab, Geryon fenneri, in the South Atlantic Bight: distri- bution, population structure, and gear assessment KORNFIELD, I., and S. M. BOGDANOWICZ. Differentiation of mitochondrial DNA in Atlantic herring, Clupea harengus AHRENHOLZ, DEAN W., WALTER R. NELSON, and SHERYAN P. EP- PERLY. Population and fishery characteristics of Atlantic menhaden, Breuoor- tia tyrannus (Continued on back cover) 395 403 419 435 447 455 471 481 499 523 547 561 569 J J Seattle, Washington U.S. DEPARTMENT OF COMMERCE Malcolm Baldridge, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Anthony J. Calio, Administrator NATIONAL MARINE FISHERIES SERVICE William E. Evans, Assistant Administrator Fishery Bulletin The Fisher;^' 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. Bruce B. Collette 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 Bulletin ( ISSN 0090-0656) is published quarterly by the Scientific Publications OfTice, 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 Orfice. 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. Fishery Bulletin ^ [ Marine Biological Laboratal f LIBRARY CONTENTS I ^ . '^OV 2 3 1987 Vol. 85, No. 3 H July 1987 Woods Hole Ma^Q MUGIYA, YASUO. Phase dfference between failiirii?afiQ|^ ^nd orgaiiic m^Yix formation in the diurnal growth of otoliths in the rainbow trout, !Salmo gdiidn&yt ^^95 BEST. PETER B. Estimates of the landed catch of right (and other whalebone) whales in the American fishery, 1805-1909 403 HOLT, RENNIE S. Estimating density of dolphin schools in the eastern tropical Pacific Ocean by line transect methods 419 HOLT, RENNIE S., TIM GERRODETTE, and JOHN B. COLOGNE. Research vessel survey design for monitoring dolphin abundance in the eastern tropical Pacific 435 BUTLER, RICHARD W., WALTER A. NELSON, AND TYRRELL A. HEN- WOOD. A trawl survey method for estimating loggerhead turtle, Caretta caretta, abundance in five eastern Florida channels and inlets 447 GROOT, C, and T. P. QUINN. Homing migration of sockeye salmon, Oncorhynchus nerka , to the Eraser River 455 BECKER, D. SCOTT, and KENNETH K. CHEW. Predation on Capitella spp. by small-mouthed pleuronectids in Puget Sound, Washington 471 HINCKLEY, SARAH. The reproductive biology of walleye pollock, Theragra chalcogramma , in the Bering Sea, with reference to spawning stock struc- ture 481 KENDALL, A. W., JR., M. E. CLARKE, M. M. YOKLAVICH, and G. W. BOEH- LERT. Distribution, feeding, and growth of larval walleye pollock, Theragra chalcogramma , from Shelikof Strait, Gulf of Alaska 499 ARMETTA, THERESE M., and BRADLEY G. STEVENS. Aspects of the biology of the hair crab, Erimacrus isenbeckii, in the eastern Bering Sea 523 WENNER, ELIZABETH L., GLENN F. ULRICH, and JOHN B. WISE. Exploration for golden crab, Geryon fenneri, in the South Atlantic Bight: distri- bution, population structure, and gear assessment 547 KORNFIELD, I., and S. M. BOGDANOWICZ. Differentiation of mitochondrial DNA in Atlantic herring, Clupea harengus 561 AHRENHOLZ, DEAN W., WALTER R. NELSON, and SHERYAN P. EP- PERLY. Population and fishery characteristics of Atlantic menhaden, Breuoor- tia tyrannus 569 {Continued on next page) Seattle, Washington 1987 For sale by the Superintendent of Documents. U.S. Government Printing Office, Washing- ton DC 20402 — Subscription price per year: $21.00 domestic and $26.25 foreign. Cost per single issue: $6.50 domestic and $8.15 foreign. Contents — Continued SOGARD, SUSAN M., DONALD E. HOSS, and JOHN J. GOVONI. Density and depth distribution of larval gulf menhaden, Brevoortia patronus , Atlantic croaker, Micropogonias undulatus , and spot, Leiostomus xanthurus , in the northern Gulf of Mexico 601 BOWERING, W. R. Distribution of witch flounder, Glyptocephalus cynoglossus , in the southern Labrador and eastern Newfoundland area and changes in certain biological parameters after 20 years of exploitation 611 DALLEY. E. L., and G. H. WINTERS. Early life history of sand lance (Am- modvtes), with evidence for spawning A. dubius (in Fortune Bay, Newfound- land 631 Notes MERCALDO-ALLEN, RENEE, and FREDERICK P. THURBERG. Heart and gill ventilatory activity in the lobster, Homarus americanus , at various tempera- tures 643 FOREMAN, TERRY. A method of simultaneously tagging large oceanic fish and injecting them with tetracycline 645 SCHAEFER, KURT M. Second record of the Kawakawa, Euthynnus affinis, from the eastern Pacific Ocean 647 COLLINS, MARK R., C. WAYNE WALTZ, WILLIAM A. ROUMILLAT, and DARYL L. STUBBS. Contribution to the life history and reproductive biology of gag. Mycteroperca microlepis (Serranidae), in the South Atlantic Bight 648 ROPES, JOHN. Age and growth, reproductive cycle, and histochemical tests for heavy metals in hard clams, Mercenaria mercenaria, from Raritan Bay, 1974- 75 653 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. PHASE DIFFERENCE BETWEEN CALCIFICATION AND ORGANIC MATRIX FORMATION IN THE DIURNAL GROWTH OF OTOLITHS IN THE RAINBOW TROUT, SALMO GAIRDNERI Yasuo Mugiya' ABSTRACT The relative role of calcium and organic matrix deposition in the formation of daily increments in otoliths was studied in in vitro preparations of otolith-containing sacculi of rainbow trout, Salmo gairdneri. Sacculi were incubated in a Ringer solution containing both ''•'^Ca and 3H-glutamic acid for 2 hours at 6-h intervals throughout a 24-h period and then the uptake of these isotopes was deter- mined for both otolith and saccular tissue fractions. Serum calcium and sodium concentrations were also analyzed for diurnal variations. Serum calcium concentrations varied diurnally by 87r in a single phasic pattern, reaching a peak at dusk (1600 hi and a nadir at night <2200 h), while sodium concentrations remained almost constant throughout a 24-h period. Diurnal variation in the otolith's uptake of calcium and glutamic acid showed discrete, antiphasic cycles. The rate of calcium uptake varied in a pattern closely resembling that of serum calcium (the peak at 1600 h and the nadir at 2200 h); glutamic acid uptake remained almost constant during the daytime and peaked at night (2200 h). The results indicate that in rainbow trout daily increments of otoliths are formed by the antiphasic deposition of calcium and organic matrix. Teleost otoliths consists of calcium carbonate in aragonite form and an organic matrix in which acidic amino acids dominate (Degens et al. 1969). Concentric rings within the microstructure of otoliths are commonly laid down on a daily basis fCampana and Neilson 1985; Jones 1986). A unit increment comprises one light and one dark ring when observed under transmitted light. These bi- partite structures are also observable by scanning electron microscopy. After etching with weak acids or decalcification with calcium-chelating agents, they usually appear as an alternating pattern of well-calcified zones with elongated crystals perpendicular to the otolith periphery (accretion zone) and narrow grooves which inter- sects the crystal development at right angles (dis- continuous zone). However, some recent studies (Mugiya and Muramatsu 1982; Watabe et al. 1982; Takahashi 1982; Morales-Nin 1987) showed that if the etching and subsequent treat- ments were carried out carefully, the organic ma- trix could be preserved in the discontinuous zone, appearing as a raised ridge. After complete decal- cification of the otolith, Dean et al. (1983) and Radtke and Targett (1984) observed incremental features in the remaining matrix. Thus, stated in 'Faculty of Fisheries. Hokkaido University, Minato-.3, Hako- date, 041 Japan. relative terms, the accretion and discontinuous zones appear to be alternatively calcium-domi- nant and matrix-dominant structures. However, Watabe et al. (1982) observed that morphologi- cally similar matrix material extended continu- ously between accretion and discontinuous zones, and proposed a possible mechanism for otolith in- crement formation. For their recently proposed model, Campana and Neilson (1985) also as- sumed continuous matrix formation in diurnal otolith growth. Based on these morphological studies, three hy- potheses might account for the formation of the bipartite structure of otolith increments; 1) both organic matrix and calcium deposition show diur- nal variations occurring in antiphase, 2) calcium deposition varies diurnally, while matrix deposi- tion does not, and 3) calcium deposits at a con- stant rate throughout a 24-h period, while matrix deposition varies diurnally. All these would re- sult in the formation of alternate zones where calcium or matrix deposition predominated. Of these, the last possibility can be excluded. Physi- ological studies indicate that the rate of calcium uptake by otoliths varies diurnally in goldfish and rainbow trout (Mugiya et al. 1981; Mugiya 1984). The present study was undertaken to investi- gate diurnal variation in matrix formation and to Manuscript accepted March 1987. FISHERY BULLETIN: VOL 8.5. NO. 3. 1987 395 FISHERY BULLETIN: VOL. 85, NO. 3 relate its phase, if apparent, to otolith calcifica- tion. Because diurnal variations in otolith calcifi- cation show seasonality (Mugiya 1984), an ideal way to examine such a phase relationship is to determine the rates of calcium and matrix deposi- tion on a single otolith simultaneously, using a double-tracer method and in vitro, isolated sac- culi from rainbow trout. Diurnal profiles of glu- tamic acid and calcium uptake were examined in the otoliths and the remaining saccular tissue. Serum calcium and sodium concentrations were also measured for diurnal variations. MATERIALS AND METHODS Rainbow trout, Sal mo gairdneri, 29-31 cm in standard length, were obtained from a commer- cial dealer and reared in a pair of outside ponds supplied with 14'C running water. They were fed trout food pellets once a day at around 0845 h. Two females showed maturing ovaries, so their data were omitted. No males were excluded be- cause maturing testes have little, if any, effect on the level of serum calcium. The experiment was carried out in December 1984. Dusk occurred at 1600 h and dawn at 0700 h. At each sampling time, five or six fish were gently netted one at a time. Blood was immedi- ately collected from the caudal vessels by cutting the tail of the fish and draining it into test tubes. After centrifugation, the separated sera were stored at -30°C for 6-24 hours and analyzed for calcium and sodium concentrations by flame pho- tometry using an atomic absorption spectrophoto- meter (Hitachi- 518). After blood collection, the head was severed, trimmed, and placed in an oxygenated Ringer so- lution kept at 14'C. The sacculi were dissected under a binocular microscope according to a pre- viously described technique (Mugiya 1984). The pair of sacculi were placed in the incubation medium, and the next fish was netted. Time was recorded to ensure that sacculi from each fish were incubated for the same length of time. Isolated sacculi were placed in a glass vessel and incubated in 50 mL of a Ringer solution (Mugiya 1986 1 containing ^''Ca and '^H-glutamic acid (New England Nuclear) at concentrations of approximately 0.17 fxCi mL and 0.33 jxCi/mL re- spectively. The incubation was carried out with oxygenation at 14°C for 2 hours. To determine proper incubation times, the uptake of "^H- glutamic acid by otoliths was plotted against time; although in this preliminary experiment, sacculi were incubated for periods of up to 3 hours, steady-state levels were obtained in less than 2 hours. After incubation, sacculi were rinsed several times in the radioisotope-free Ringer solution and separated into otolith and saccular tissue frac- tions under a binocular microscope. The sepa- rated otoliths were lightly rinsed in water, placed in individual counting vials, dried at 90°C overnight and then weighed. The saccular tissue was directly placed in the vial without a further rinse and air-dried. These samples were solubi- lized in a mixture of 0.2 mL perchloric acid and 0.2 mL hydrogen peroxide at about 80°C for 2 hours, and added to Scintisol EX-H (Wako) for counting (liquid scintillation spectrometer, Aloka LSC-673). Tritium and ^'^Ca activities were measured simultaneously using two channels with nar- rowed windows, 50-300 for '^H and 70-900 for ^■"^Ca. Although the amount of 'H activity enter- ing the Ca channel was found to be practically negligible, '^'^Ca would certainly affect counts on the H channel, despite the window conditions. Therefore, counts on the H channel were cor- rected by the equation: ^H activity = H - 1/a Ca ;i) where H and Ca represent counts on the H and Ca channels respectively, and « (contamination ratio) is experimentally defined as log a = 0.1386R - 0.0488 where R is a ratio deter- mined for the quenching level of each sample. The validity of this correction was further checked by another equation based on differences in the physical half-life of the isotopes: ^H activity CaoH2 - CaaHo Cao - Ca2 (2) -Reference to trade names doe.s not imply endorsement by the National Marine Fisherie.s Service, NOAA, where Hq and Cao ai'e counts on the H and Ca channels at time 0 respectively, and H2 and Ca2 are those recounted a few months later. Because these two methods of discrimination gave essen- tially the same results, the data from Equation (1) were presented in this study. Some rainbow trout have an aberrant otolith in either or both of their sacculi (Mugiya 1972). If 396 MUGIYA: DIURNAL GROWTH OF RAINBOW TROUT OTOLITHS the highly aberrant form was found by inspection after incubation, it was excluded from the data. RESULTS Serum calcium concentrations varied diurnally by approximately 89f in a single phasic pattern (Fig. 1). The maximum level (5.47 meq/L) oc- curred at dusk (1600 h), followed by a rapid de- crease (P < 0.05) to a nadir (5.04 meq/L) at night (2200 h). The level then gradually increased to- ward the next peak. In contrast, serum sodium concentrations showed a statistically insignifi- cant variation of only 0.67c iP > 0.05; 148.1-149.0 meq/L) throughout a 24-h period. When otolith-containing sacculi were incu- bated with '^H-glutamic acid, the saccular tissue (without otoliths) was almost saturated with the isotope within the first 30 minutes or hour of in- cubation (Fig. 2). Otoliths also showed a consider- able uptake (about 60'/^ of the total) of the isotope in the first 30 minutes, followed by a gradual increase in radioactivity until 3 hours, when the incubation was terminated. Tritium activities were always 6-8 times higher in the saccular tis- sue than in the respective otolith (Fig. 2). The 150- il49 f 148 147 56 55 5.4 i 5 3 5.1 5.0 4.9 time-related uptake of '^^Ca by these tissue frac- tions has been reported (Mugiya 1984). In that study, the saccular tissue was saturated with '^^Ca within the first hour of incubation, while otoliths showed an almost linear increase in "^^Ca uptake during the first 5 hours at which point the incubation was terminated. The uptake of calcium by otoliths varied diur- nally (Fig. 3), and the pattern was quite similar to that of diurnal variations in serum calcium con- centrations (Fig. 1). The rate of calcium uptake was intermediate at 1000 h, peaked at dusk ( 1600 h), and then decreased significantly (P < 0.02) by 37% to a nadir at night (2200 h). The low rate persisted through the night, increasing slightly at 0400 h. Clearly otolith calcification proceeded more actively during the daytime. The uptake of glutamic acid by the same otoliths also showed a diurnal variation, and its profile was almost an- tiphasic to that of calcium uptake (Fig. 3). The rate of the uptake remained rather low during the daytime with a small nadir at dusk ( 1600 h). Then the rate increased significantly (P < 0.05) to a peak at night (2200 h), followed by a return to the daytime level. Thus the most active deposition of otolith matrix (at least proteins) occurred during the first half of the nighttime period, when cal- cium deposition was at its lowest level. The uptake of glumatic acid by the saccular tissue showed significant (P<0.02), diurnal X 10 for sacculus 800 ^ 600 ^ 400 i 200 0 0.5 Hours in incubation 1000 1600 2200 0400 Time of day (hours) 1000 Figure 1. — Diurnal variations in serum calcium (•) and sodium fCi concentrations in rainbow trout. Each plotted value represents mean ± SE of 5 or 6 fish. P < 0.05 for 2200 h. Figure 2. — Time course for the in vitro uptake of ^H-glutamic acid by otoliths (•) and the saccular tissue (O) of isolated sacculi in rainbow trout. The radioactivity of the saccular tissue is expressed as dpm per otolith weight (mg) because the dry weight of the individual saccular tissue was too light to be determined accurately. Each plotted value represents mean ±SE of 8-10 samples. 397 FISHERY BULLETIN: VOL. 85, NO. 3 600 500 400- 300 200 1000 1600 2200 0400 Time of day (hours) 1000 FiGURK 3.— Diurnal variations in the in vitro uptake of ■'■"'Ca (•) and ^iH-glutamic acid (J) by otoliths in rainbow trout. Each plotted value represents mean ± SE of 8-10 samples. Dark horizontal bar indicates nocturnal and twilight peri- ods. *P < 0.02 for 2200 h; **P < 0.05 for 1600 h. variation with a single peak at 1600 h when ma- trix deposition on the otoliths was lowest (Fig. 4). Note that the active biosynthesis of matrix proteins in the saccular tissue is not necessarily followed by their instantaneous deposition on the otoliths, suggesting the presence of cyclic secre- tion activity in the cells of the sacculus. The rate of calcium uptake by the saccular tissue did not vary much throughout a 24-h period (Fig. 4). Ratios of counts of '^■^Ca and -^H-glutamic acid in the respective otoliths magnified the antiphasic relationship between '^^Ca and -^H uptake (Fig. 5). Significant variation (P < 0.01 ) between the peak (1600 h) and the nadir (2200 h) demonstrates the much greater deposition of calcium relative to glutamic acid during the daytime, which suggests that in December the accretion zone forms during the daytime with its peak at dusk. DISCUSSION Although previous studies (Mugiya et al. 1981; Mugiya 1984) showed that otoliths grew by the 14000 12000 10000 E 8000 E f 6000 ^ 4000 300 250 200 •■o 1000 1600 2200 0400 Time of day (hours) 1000 Figure 4.— Diurnal variations m the in vitro uptake of ■'■''Ca (•) and 'SH-glutamic acid CJi of saccular tissue in rainbow trout. Each plotted value represents mean ± SE of 8-10 sam- ples. *P < 0.02 for 1000 h. diurnal deposition of calcium, it remained to be determined whether matrix deposition on the otoliths was diurnal or not. Histochemically, otolith matrix consists of various kinds of sub- stances such as proteins, acid mucopolysaccha- rides, PAS-positive materials, and lipids (Mugiya 1968). Of these, proteins are the most dominant component and are characterized by a high con- tent of acidic amino acids (Degens et al. 1969). In the present study the diurnal deposition of otolith matrix was evident when examined in terms of the incorporation of glutamic acid into otoliths, showing a single peak at night. Interestingly, cal- cium deposition on the same otoliths proceeded most actively at dusk, followed by minimum de- position at night. Thus, it is concluded, in rain- bow trout kept under natural photoperiod, the pace of otolith calcification is almost antiphasal to the pace of matrix deposition on the otoliths. The present results, where both calcium and 398 MUGIYA: DIURNAL GROWTH OF RAINBOW TROUT OTOLITHS 1.8- 1.4 1.0- 0.6- 1000 1600 2200 0400 Time of day (hours) 1000 Figure 5. — Diurnal change in the ratio of ■'•^Ca to ^H-glutamic acid activity incorporated into the same otoliths in rainbow- trout. Each plotted value represents mean ± SE of 8-10 sam- ples. *P < 0.01 for 2200 h. matrix deposition on otoliths varied diurnally in antiphase, indicate the relative importance of these substances for daily increment formation in otoliths. The accretion zone is formed predomi- nantly by calcium deposition, while the discontin- uous zone results from reduced calcium and sub- stantially increased matrix deposition on the otoliths. These findings coincide with the morpho- logical observation that the accretion zone is a crystalline layer with organic materials and the discontinuous zone is a layer containing more or- ganic materials and less calcium (Mugiya and Muramatsu 1982; Watabe et al. 1982). Watabe et al. (1982) observed that the matrix fibers and their aggi'egates were morphologically similar in the two zones and continuous throughout in Tilapia and Fundulus otoliths. Based on these observations, they have suggested that the ma- trix materials are identical in the zones and their deposition might be an uninterrupted event dur- ing diurnal otolith growth. They also stated that this did not necessarily imply that the rate of organic matrix secretion was diurnally constant. In fact, the present study reveals that diurnal variations in the rate of the matrix deposition, coupled with variations in calcium deposition, play an important role in otolith increment for- mation. Although the sacculus contains the otolith, otolithic membrane, and endolymph, a high con- tent of acidic amino acids is characteristic of the calcified otolith (Degens et al. 1969). Therefore, variations in the uptake of glutamic acid by the saccular tissue should be closely related to the activity of the matrix formation of the otolith, even though glutamate is also used as a neuro- transmitter in the saccular macula (Potter et al. 1986). Otolith forming cells, yet to be positively identified (Mugiya 1974; Dale 1976; Dunkel- berger et al. 1980; Saito 1984), synthesize the pre- cursor of the matrix and secrete it into the lumen. The precursor may then deposit on the otolith after further biochemical modification. The present study showed the presence of a time-lag between the matrix biosynthesis in the saccular tissue and its deposition on the otoliths. However, this does not necessarily mean that these two processes are separated in phase. In this study, calcification and matrix formation were measured in terms of "instantaneous" gi'owth rates (Ottaway 1978). Therefore the max- imum deposition of organic matrix on otoliths at night must be accompanied by the active biosyn- thesis of the matrix in the cellular level and its consecutive secretion into the lumen, which may rather reduce the radioactivity in the saccular tissue. The high radioactivity in the tissue at dusk might result from the accumulation of the newly synthesized matrix owing to the reduction of its transport to the otoliths. These results sug- gest the presence of at least three different phases in otolith matrix formation: the active synthesis of matrix proteins on the cellular level with its reduced deposition on the otoliths, active synthe- sis with active deposition, and inactivity in both synthesis and deposition. Mugiya (1984) reported that the profile of diur- nal otolith calcification was antiphasal between the summer and winter solstices in rainbow trout. In the winter experiment, the peak and the nadir of calcium deposition on otoliths came at 1600 h and 0400 h, respectively; while in the present winter experiment the peak at 1600 h decreased to the nadir earlier, at 2200 h, followed by a slight increase at 0400 h. Although there is a difference in the time-related profiles, the results of both experiments mainly showed that otolith calcifica- tion slowed down after the onset of darkness and remained relatively inactive until the next sun- rise. Molluscan nacre shows a laminar structure re- sulting from alternate accumulation of organic 399 FISHERY BULLETIN: VOL. 85, NO. 3 matrix and calcium carbonate crystals. In the de- velopment of these bipartite structures, matrix deposition on growing crystals is known to inter- rupt further crystal growth (Wilbur 1980). This results in the alternate formation of calcium-rich and matrix-rich layers. In such cases, the matrix may have opposite functions in controlling crystal growth: as an inhibitor for the crystal growth along the C-axis and as a nucleator for the forma- tion of the next crystal layer (Crenshaw 1982). The matrix appears to play a key role in con- trolling the formation of these different layers. Although this sequence is likely to be the case for otolith increment formation (Wilbur 1980), the rate of calcium deposition on otoliths appears to be closely related to the level of serum calcium, which is regulated by the action of hypercalcemic and hypocalcemic hormones (Oguro and Pang 1982). Serum calcium has been suggested as a trigger for otolith calcification through the calcium-calmodulin system (Mugiya 1986). Al- though Pickford (1953) found that hypophysec- tomy resulted in no otolith growth in killifish, what exactly controls the rate of matrix deposi- tion on otoliths remains unknown. ACKNOWLEDGMENTS The author thanks David H. Secor, University of South Carolina, for his thoughtful review of the manuscript. This work was supported in part by Grant No. 61560202 from the Educational Min- istry of Japan. LITERATURE CITED CAMPANA, S E . AND J. D. Neilson. 1985. Microstructure of fish otoliths. Can. J. Fish. Aquat. Sci. 42:1014-1032. Crenshaw, M A. 1982. Mechanisms of normal biological mineralization of calcium carbonates. In G. H. Nancollas (editor). Biolog- ical mineralization and demineralization, p. 243- 257. Springer- Verlag, Berlin. Dale, T. 1976. The labyrinthine mechanoreceptor organs of the cod Gadus morhus L. (Teleostei: Gadidae). A scanning electron micro.scopical study, with special reference to the morphological polarization of the macular sensory cells. Norw. J. Zool. 24:85-128. Dean, J M., C A Wilson, P W Haake, and D W Beckman 1983. Microstructural features of teleost otoliths. In P. Westbroek and E. W. de Jong (editors), Biomineraliza- tion and biological metal accumulation, p. 353- 359. D. Reidel Publ. Co. Degens, E T , W. G. Deuser, and R L. Haedrich. 1969. Molecular structure and composition of fish otoliths. Mar. Biol. (Berl.) 2:105-113. Dunkelberger. D. G., J. M Dean, and N. Watabe 1980. The ultrastructure of the otolithic membrane and otolith in the juvenile mummichog, Fundulus heterocli- tus. J. Morphol. 163:367-377. Jones, C. 1986. Determining age of larval fish with the otolith in- crement technique. Fish. Bull., U.S. 84:91-103. 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. Mugiya, Y. 1968. Calcification in fish and shell-fish - VII. Histochem- ical similarities between the otolith and the macula re- gion of sacculus in juvenile rainbow trout, with special reference to the otolith formation offish. Bull. Jpn. Soc. Sci. Fish. 34:1096-1106. 1972. On aberrant sagittas of teleostean fishes. Jpn. J. Ichthyol. 19:11-14. 1974. Calcium-45 behavior at the level of the otolithic organs of rainbow trout. Bull. Jpn. Soc. Sci. Fish. 40:457-463. 1984. Diurnal rhythm in otolith formation in the rainbow trout, Salmo gairdneri : seasonal reversal of the rhythm in relation to plasma calcium concentrations. Comp. Biochem. Physiol. 78A:289-293. 1986. Effects of calmodulin inhibitors and other metabolic modulators on in vitro otolith formation in the rainbow trout, Salmo gairdnerii. Comp. Biochem. Physiol. 84A:57-60. Mugiya, Y., and J, Muramatsu. 1982. Time-marking methods for scanning electron mi- croscopy in goldfish otoliths. Bull. Jpn. Soc. Sci. Fish. 48:1225-1232. Mugiya, Y., N Watabe. J. Yamada, J M Dean, D G Dunkelberger, and M. Shimizu. 1981. Diurnal rhythm in otolith formation in the goldfish, Carassius auratus. Comp. Biochem. Physiol. 68A:659- 662. Oguro, C, and P. K. T. Pang (editors). 1982. Comparative endocrinology of calcium regula- tion. Jpn. Sci. Soc. Press, Tokyo, 245 p. Ottaway, E. M 1978. Rhythmic growth activity in fish scales. J. Fish Biol. 12:615-623. Pickford. G E. 1953. A study of the hypophysectomized male killifish, Fundulus heteroclitus (Linn.). Bull. Bingham Oceanogr. Collect., Yale Univ. 14:5-41. Potter, A J , M. J Drescher, and D. G Drescher. 1986. Potassium-stimulated efflux of radiolabeled prod- ucts formed from L-(i4C(U))-glutamine in vitro by the saccule of the rainbow trout (Salmo gairdnerii R). Comp. Biochem. Physiol. 84A:265-270. Radtke, R. L., and T E Targett 1984. Rhythmic structural and chemical patterns in otoliths of the Antarctic fish Notothenia larseni: their application to age determination. Polar Biol. 3:203- 210. Saito, S. 1984. Studies on the anatomy and microstructure of the inner ear in Sarotherodon niloticus, with special refer- ence to the otolithic organ. MS Thesis, Hokkaido Uni- versity, Hakodate, 48 p. 400 MUGIYA: DIURNAL GROWTH OF RAINBOW TROUT OTOLITHS Takahashi. M naeus). J. Exp. Mar. BioL EcoL 58:127-134. 1982. Daily growth increments preserved in fossil fish WILBUR. K. M otoliths. llnJpnl Aquabiology 4:308-309. 1980. Cells, crystals and skeletons, //i M. Omori and N. Watabk. N . K Tanaka. J. Yamada. and J. M. Dean. Watabe (editors), The mechanisms of biomineralization 1982. Scanning electron microscope observations of the in animals and plants, p. 3-11. Tokai Univ. Press, organic matrix in the otolith of the telcost fish Fundulus Tokyo. heteroclitus (Linnaeus) and Tilapia nilotica (Lin- 401 ESTIMATES OF THE LANDED CATCH OF RIGHT (AND OTHER WHALEBONE) WHALES IN THE AMERICAN FISHERY, 1805-1909 Peter B. Best' ABSTRACT Using a combination of the numbers of'bowhead, right, humpback, and gray whales listed for partic- ular voyages by C. H. Tovvnsend, and the declared returns of whale oil and whalebone from the same voyages as listed by A. Starbuck and R. B. Hegarty, mean oil and whalebone yields per whale are calculated and temporal trends in these yields investigated for each species. These are then used to obtain an e.stimate of the total landed catch for each 5-year period from 1805 to 1909, using the species composition from Townsend's lists and adjusting it upwards from the ratio of oil or bone production for Townsend's sample to the total known importation of these products to the United States for the same period. An alternative estimate is based on the catch per voyage in Townsend's sample, strat- ified by voyage-type (sperm, whalebone, or mixed), and prorated up by the number of whaling voyages of the same type as listed by Starbuck and Hegarty. The two methods produced estimates of the landed catch by American-registered vessels between 1805 and 1909 of 29,748-30,313 bowhead, 70.325-74.693 right. 14.164-18.212 humpback, and 2,665-3.013 gray whales. Between 1715 and 1928, whaling vessels from American ports are estimated to have made 13.927 voyages, mostly under sail, in their world- wide pursuit of oil and whalebone (Sherman 1965). In 1846, at the peak of the fishery, the American whaling fleet comprised over 735 ves- sels displacing 233,189 tons (Hohman 1928). Be- cause of the essentially unregulated and competi- tive nature of the enterprise, no systematic recording or collection of catch statistics was ever initiated for this very extensive fishery. In 1875. Alexander Starbuck began to compile a list of the returns of whaling vessels from Amer- ican ports from 1715, a task continued to the end of the fishery in 1928 by Hegarty (1959). These publications list for each voyage the vessel's name, class, tonnage, captain, managing owner or agent, destination, dates of sailing and arrival, and the results of the voyage in barrels of sperm or whale oil and pounds of whalebone. Numbers of whales taken are not given, but this did not prevent Starbuck (1878) from making his own calculations. In a footnote to his table J, which listed quantities of oil and whalebone imported into the United States from 1804 to 1876, Star- buck stated that Scammon estimate.s that sperm whales will average 'Mammal Research Institute. University of Pretoria. South .•\frica; mailing address: South African Museum. P.O. Box 61. Cape Town, 8000 South Africa. 25 and right whales 60 barrels of oil, and of the former 10 and of the latter 20 per cent of those killed are lost. Upon that basis the above amounts of oil would repre- .sent the slaughter of 225,521 sperm, and 193,522 right whales. The latter figure has frequently been quoted as representing the size of the historical take of right whales (sometimes incorrectly for the period 1804 to 1817, an error apparently originally per- petrated by Harmer 1 1928], who also inferred that the entire take was of southern right whales). It is clear however that the landings of oil not only included production from both northern and southern right whales, but also from bowhead, humpback, and gray whales, species for which Starbuck (1878) made no allowance in his origi- nal calculation. In this paper, an attempt has been made to revise Starbuck's calculations to account for the species composition of the catch, to extend his analysis forward in time using importation fig- ures provided by Hegarty (1959), and to use whalebone as well as oil production. An indepen- dent method of estimating the landed catch using the catch per voyage has also been developed. The motivation for this paper arose from the International Whaling Commission meeting on the past and present status of right whales, held in Boston m 1983, where the need for an improved estimate of the size of the American catch of right whales became apparent (Brownell et al. 1986). Manuscript accepted FISHERY BULLETIN: VOL. 85. NO ,3. 1987 403 FISHERY BULLETIN: VOL, 85, NO. 3 MATERIAL AND METHODS From logbook extractions, Townsend (1935) tabulated the numbers of sperm, bowhead, right, humpback, and gray whales taken per voyage by 744 whaleships (mostly American) between 1751 and 1925. These figures include not only the whales processed but also those killed and brought alongside but subsequently lost before processing; these statistics have thus been termed the "landed catch" in this paper. The numbers of right whales are listed by ocean (i.e., North and South Pacific, North and South Atlantic, and In- dian Oceans). In all, 53,877 whales are listed from 1,665 voyages. Excluding non-U. S. vessels, 16,837 baleen whales were taken in a total of 1.651 voyages, of which 636 were only sperm whaling voyages. The species composition of the baleen whale catch as extracted by Townsend has formed the basis of all the analyses performed in this paper, and is henceforward referred to as the "Townsend sample". Some of the original work sheets used by Townsend in his 1935 paper were discovered in 1978 in the library of the Osborne Laboratory of the New York Aquarium. These comprise voyage abstracts giving the date, ocean, geographical po- sition, number, and species of each whale landed, together with remarks such as "found dead", "cow and calf", etc.; about half of the work sheets are in the original handwriting of the compiler(s), while the remainder consist of typewritten copies. The abstracts cover voyages by most vessels whose names started with letters A through J (bark A. Houghton to brig Juno ). Because some errors ap- parently occurred between the original abstracts and the final printed version (Schevill and Moore 1983), the catch data for the 438 voyages on which baleen whales were landed and for which ab- stracts were available has been checked against the figures tabulated by Townsend (1935). Errors were found in 32 voyages (or about 7% of the total) and corrected. The abstracts examined rep- resent a landed catch of 6,982 baleen whales, or roughly 41''/^ of the total Townsend sample. Mean oil and whalebone yields per whale have been obtained by comparing the numbers of whales caught on a voyage (as listed by Townsend ) with the amount of whale oil or whale- bone landed for the same voyage (as listed by Starbuck ( 1878) or Hegarty ( 1959)). To avoid com- plications created when more than one baleen whale species was taken, only voyages where a single baleen whale .species was taken have been analyzed. Because of suspected differences in size (and so presumably in yields of products) between North Pacific right whales and those from other seas (Omura 1958), they have been considered as a separate "species" for the purposes of this sec- tion. In order to reduce the amount of variation in yield and to avoid situations where Townsend seems to have had access to only a partial log of the voyage, only voyages on which 10 or more animals of that species were taken have been used (or roughly 20% of Townsend's sample of voyages on which whalebone whales were taken). Oil or bone sent home or sold abroad has been included where it is known; as Starbuck (1878) has pointed out, that sold abroad was not always accounted for. Figures for the total annual importation of oil and whalebone into the United States have been taken from Starbuck (1878) and Hegarty (1959). For the catch per voyage analysis, the voyages in the Townsend sample have been stratified ac- cording to type, either sperm (when only that spe- cies was landed), whalebone (when no sperm whales were included in the catch), or mixed (when both sperm and whalebone whales were taken). The numbers of such cruises have then been adjusted upwards by the numbers of such voyages found in the Starbuck/Hegarty compila- tion, where vessels were identified as sperm whalers if they were reported as returning with or sending home only sperm oil, as whalebone whalers if they only reported whale oil and/or bone, and as mixed whalers if they returned with or sent home both whale oil/bone and sperm oil. Two additional classes were recognized in the Starbuck/Hegarty compilation: "clean" voyages and "incomplete" voyages. Clean voyages were those entered as such by Starbuck (1878), but as Hegarty (1959) did not continue this practice, any of the voyages he listed that were completed but for which no production was reported were scored as "clean". Both authors listed several voyages that were not completed owing to fire, shipwreck, the vessel being condemned, etc., and for which no production was reported. These voyages were scored as "incomplete", and half their number was allocated on a prorata basis as either sperm, whalebone, mixed, or clean whalers, based on the proportions of these categories in the sample of completed voyages. The other half of the incom- plete voyages was discarded, the assumption being that such voyages were on average proba- bly half as successful as those completed and that Townsend (1935) was unlikely to have had access 404 BEST; LANDED CATCH OF RIGHT WHALE to the logbooks of many of them. Incomplete voy- ages comprised 9.67( of those listed by Starbuck (1878) and Hegarty (1959). A "plus minus'" figure following any estimate refers to one standard error. Values given for the coefficient of variation (CV) have been obtained using variances calcu- lated by the jackknife method (using one voyage as the sampling unit). No attempt has been made to calculate coefficients of variation for the final estimates because 1) certain independent compo- nents of the variance could not realistically be assessed (e.g., variation in the proportion of a par- ticular species in the total catch to that in the Townsend sample from one 5-yr period to the next) and 2) any biases in the data are likely to be of greater magnitude than statistical errors re- sulting from sampling variation. ESTLVIATES BASED ON PRODUCTION The number of baleen whales landed by Amer- ican whalers as extracted by Townsend (1935) is listed by five yearly period in Table 1. If Townsend's data for a particular voyage covered more than one calendar year, the catch would be entered against the later date, as this was more likely to correspond to the importation figures used as a basis for reconstruction of the catch. A total landed catch of 4,963 bowhead, 8,293 right, 2,879 humpback, and 569 gray whales was recorded for the period 1805-1914. Average Oil Yield Per Whale Right WTiales There were 147 right whale cruises producing oil yields ranging from 22.5 to 219 barrels (Fig. 1). As expected, the 17 voyages that took North Pacific right whales had higher yields (83 to 219 barrels) than the 130 taking right whales else- where (22.5 to 150 barrels), and so have been con- sidered separately. There was no significant trend in oil yield per whale during the period in which North Pacific right whales were taken (6 = -1.15 ± 0.91, t = 1.27, P > 0.20), so the overall average oil yield per whale of 41,645/341 or 122 barrels (CV = 0.063) has been used. This compares with published averages of 125 barrels, males making 60 to 100 and females 100 to 250 barrels (Clark 1887a), and 130 barrels (Scammon 1874). There appeared to be a distinct decline in the oil yield of right whales on other grounds after 1882 ib = -1.09 ± 0.30, t = 3.60, P < 0.02). Oil yields after this date have therefore been calculated from the least squares estimating equation fitted to the data: y = 64.78 - 1.090 ix - 1882) Table 1. — Five-year compilation of whalebone whale catches from Townsend (1935). Northern right Southern ri ght Period (arrival) Bowhead Atl. Pac. Atl. Pac. Ind. Humpback Gray Total 1805-1809 — — — 43 — — — — 43 1810-1814 — — — — — — — — — 1815-1819 5 — — 3 22 — 2 — 32 1820-1824 — — — 81 — — — — 81 1825-1829 — — — 269 — — 1 — 270 1830-1834 — — — 940 71 72 5 — 1,088 1835-1839 — — — 761 356 477 96 — 1,690 1840-1844 — — 324 53 516 505 66 — 1,464 1845-1849 91 — 1.088 55 349 129 89 — 1,801 1850-1854 1,101 — 165 88 34 48 84 6 1,526 1855-1859 1,238 1 235 53 72 83 242 65 1,989 1860-1864 650 9 117 109 104 48 193 205 1,435 1865-1869 717 2 107 108 25 43 124 215 1,341 1870-1874 471 — 15 45 18 46 774 70 1,439 1875-1879 104 6 13 54 1 13 380 8 579 1880-1884 119 4 1 77 26 — 619 — 846 1885-1889 76 5 38 53 72 3 171 — 418 1890-1894 86 — 3 38 13 — 22 — 162 1895-1899 183 1 12 7 — 6 10 — 219 1900-1904 74 — 6 11 — 2 1 — 94 1905-1909 33 — 1 10 6 71 — — 121 1910-1914 15 — — 26 — 25 — — 66 Total 4.963 28 2.125 2,884 1,685 1,571 2,879 569 16,704 405 FISHERY BULLETIN: VOL. 85, NO. 3 200 < - I oj 160 - o 0 RIGHT WHALES o. •■■o Norfh Pacific • — • Elsewhere J L 1810 1830 1850 1870 1890 1910 YEAR FiGL'RE 1. — Mean oil yield per whale for right whales landed on U.S. voyages from 1822 to 1910, 160 - < - ^ 5 120 9o LU < ^ > o -7 ^ < 80- 40 0 BOWHEAD WHALES ^ Hudson's Bay, Cumberland Inlet •Other grounds , oo -.^ 1820 1840 1860 YEAR 1880 1900 Figure 2. — Mean oil yield per whale for bowhead whales landed on U.S. voyages from 1849 to 1901. 406 BEST: LANDED CATCH OF RICllT Wll Al.K where .r = year of arrival (>1882) and V = average oil yield (barrels). This produces a decline from 63.7 barrels in 1883 to 34.2 barrels in 1910. Prior to 1883 there was no significant trend with time (b = 0.16 ± 0.16, t = 1.01, P > 0.2), so the average overall oil yield of 206,328/3,080 or 67 barrels (CV = 0.024) per whale has been used. Right whales in the South Atlantic were said to yield (when full grown) from 40 to 60 barrels of oil if male and 60 to 80 barrels if female, or about 60 barrels on average. Those in the South Pacific and Indian Oceans were said to be smaller, aver- aging 40 barrels if male and 60 barrels if female (Clark 1887a). Eleven right whales taken in the Indian Ocean averaged 59 barrels, with a maxi- mum of 80 barrels (Wray and Martin 1983). In a sample of 29 right whales taken in the North Atlantic, Reeves and Mitchell (19861 found a range of oil yields from 6.5 to about 100 barrels with a mean of 58 to 59 barrels. These figures all agree fairly well with the calculated values used here: the decline in yield after 1883 may reflect market considerations and the relative value of oil and whalebone (see below). Bowhead AXTiales Oil yield data were available for 39 voyages on which 987 bowhead whales were taken; six voy- ages were to "Hudson's Bay" or '"Cumberland Inlet" (Fig. 2). The latter voyages had generally lower oil yields (22 to 79 barrels) than the other grounds (32 to 184 barrels), possibly reflecting differences in distribution of size groups, or the effects of greater depletion. However, as no dis- tinction was made in Townsend's (1935) tabula- tions between bowhead whales caught on differ- ent grounds, the data set has not been subdivided. Oil yields seemed to decline throughout the pe- riod of the fishery (6 = -1.19 ± 0.48, t = 2.46, P < 0.02). so for any particular year the mean oil yield has been calculated from the estimating equation: V = 105.11 - 1.1892 (A- - 1848) cline (6 = -0.58 ± 0.73, t = 0.79,P > 0.20). How- ever, in view of the decreasing oil yields vwith time found by other workers and the economic incen- tives after 1880 that favored the collection of whalebone rather than whale oil (see below), the regression coefficient shown in Figure 2 has been retained. This produces a decline from 103.9 barrels in 1849 to 31.4 barrels in 1910; yields before 1849 (for which no data exist) are taken as 103.9 bar- rels per whale. According to Scammon (1874), bowhead whales could be classified into three types, yielding on average 200 barrels, 100 barrels, and 75 barrels of oil. Bowhead whales in the Davis Strait were said to average about 120 barrels (males 100, fe- males 140), but had decreased in size "of late years". In the Okhotsk Sea, cows averaged about 130 barrels and the bulls about 90 barrels, but once again the whales had been much smaller "during recent years" (Clark 1887a). Oil yields for 333 whales from the Western Arctic stock listed by Bockstoce and Botkin (1983) averaged 112.4 barrels. These averages are all somewhat higher than the yield calculated here, but (with the ex- ception of the last) they referred principally to the commencement of the fishery. Bockstoce and Botkin (1983) have also documented a decline in the size of bowhead whales taken over time, but the regression of barrels per whale per year has a somewhat smaller slope (-0.61 barrels per year) than in the present case: the mean yield declined from about 119 barrels in 1850 to about 70 barrels in 1900 (cf 43 in the present data). This difference may simply reflect a different measurement — Bockstoce and Botkin apparently only considered the yield of animals for which a barrel-estimate was made by the whaling vessel so that animals may have been excluded if no oil was rendered from them. The present analysis however consid- ers all whales taken on a voyage (whether proc- essed into oil and whalebone or not), so that it is not surprising that its figures are somewhat lower than for previous estimates. An oil yield as low as 49 barrels per whale was calculated for bowhead whales in Hudson Bay between 1860 and 1890 (Ross 1974). where x = year of arrival 01848) and y = average oil yield (barrels). The slope of this regression is very sensitive to the three data points after 1890; their exclusion re- sults in a much slower, nonsignificant rate of de- Hunipback AXTiales Oil yield data were available for 29 cruises on which 1,137 humpback whales were taken (Fig. 3). There was no significant trend with time (6 = 0.31 ± 0.19, t = 1.58, P > 0.10), so the aver- 407 O 40 9o ^|30h ^ 0 FISHERY BULLETIN: VOL. 85, NO. 3 HUMPBACK WHALES 1810 1830 1850 YEAR 1870 1890 Figure 3. — Mean oil yield per whale for humpback whales landed on U.S. voyages from 1853 to 1887. age overall oil yield of 27,797/1,137 or 24.4 bar- rels (CV = 0.110) per whale has been used throughout. According to Scammon (1874), humpback whales varied more in their produc- tion of oil than all other rorquals. Some individu- als yielded only 8 or 10 barrels, whereas others gave up to 75 barrels; large females yielded on average 40 barrels. Mitchell and Reeves (1983) used an average of 25 barrels per whale, although individual whales yielded from 5 to 85 barrels each. The value calculated here is thus in reason- able agreement with previous estimates. Gray VHiales There were no cruises in Townsend (1935) on which gray whales were the only baleen whale species taken, on which at least 10 animals were landed, and for which production figures were available in Starbuck (1878). Consequently the average production figure of 35 barrels per whale estimated by Henderson (1972) has been used throughout. Average Yield of Whalebone Per Whale Average yields of whalebone have been calcu- lated essentially the same way as for oil. How- ever, as Starbuck (1878) pointed out, in the ear- lier years (before about 1844/45), reports of the amount of bone taken were only occasional: Most of that commodity was imported prior to 1840 in New London and Sag Harbor ships, its value being so low that captains of ves.sels from many of the other ports did not care to be encumbered with it. For this reason a large amount of bone was brought home which it is impossible to properly accredit. Figures for whalebone landings were listed for 94 to 959^ of the voyages on which bowhead or North Pacific right whales were taken, but for only 30 and 24'7( of the voyages taking other right and humpback whales respectively. Two alterna- tive (and probably extreme) assumptions can therefore be made: A) that only those vessels listed as landing whalebone actually did so, or B) that all vessels taking baleen whales retained the whalebone to the same extent as those for which whalebone production was reported. Aver- age whalebone yields per whale (and trends therein) have been calculated here under both assumptions A and B (Figs. 4 7). Right Uliales Of the 17 voyages on which 10 or more right whales were taken in the North Pacific, whale- bone production was reported for 16 (Fig. 4). There was no significant trend in bone yield per whale in all 17 voyages (/)- 6.36 ± 13.98, t = 0.46, P > 0.6) or in the 16 for which bone pro- duction was declared (6 = -12.84 ± 12.14, t ~ 1.06, P > 0.3). Consequently overall mean yields of 384,134/341 or 1,126 lb (CV = 0.098) whalebone (assumption A) and 384,134/323 or 1,189 lb (CV = 0.082) whalebone (assumption B) have been used. According to Clark (1887a), whalebone yield in North Pacific right whales av- eraged about 1,000 lb per 100 barrels (equivalent to a yield of 1,250 lb for an average whale), while Scammon ( 1874) gave a range of 1,000 to 1,500 lb. 408 BKST: LANDED CATCH OF RIGHT WHALE O o > or 2000- LLi < X o^ 1500 LU a. Q UJ >- ^1000 o CO LU < X -Z_ 500 < NORTH PACIFIC RIGHT WHALES \ i 1810 1830 1850 1870 YEAR 1890 1910 Figure 4. — Mean whalebone yield per whale for North Pacific right whales landed on U.S. voyages from 1841 to 1885 (A = assuming only those vessels listed as landing whalebone actually did so, B = assuming all vessels taking whalebone whales retained the whalebone to the same extent as those for which whalebone production was reported i. ^ SCO q3 600 2 O400 LU > X '^ z < 200 - RIGHT WHALES ^(excluding North Pacific) - • • • • • • • ..- *... :B..: ,••••»■••;.■■;■•■• - • • • • • - • • • - • 1 la. a JUHi • • 1810 1830 1850 1870 1890 1910 YEAR Fici'RE 5. — Mean whalebone yield per whale for right whales landed on U.S. voyages (other than in the North Pacific) from 1822 to 1910 (A = assumption A, B = assumption Bi. 409 FISHERY BULLETIN: VOL. 85, NO. 3 = 2 500 LU o < >- o > m2 000 < X - BOWHEAD WHALES o Hudson's Bay, Cumberland Inlet • Other grounds 1500 LU Q. Q I Z 1000 o CO LU < X Z 500 < o o o A - 0 i o J L J I I L ? 1810 1830 1850 1870 YEAR 1890 1910 FlCiURE 6. — Mean whalebone yield per whale for bowhead whales landed on U.S. voyages from 1849 to 1902 (A = assumption A, B = assumption Bi. LU a. 250 < X a: LU °" ^200 > Z o CO UJ < X Z < oioo 50 0 HUMPBACK WHALES B - ■^l»4-. 1810 1830 1850 1870 YEAR 1890 1910 Figure 7. — Mean whalebone yield per whale for humpback whales landed on U.S. voyages from 1853 to 1887 (A = assumption A, B = assumption B). 410 BEST: LANDED CATCH OF KIGHT WHALE Present calculations are therefore close to these estimates. Of the 127 voyages taking right whales on grounds other than the North Pacific, bone yields were available for only 37 (Fig. 5). For all 127 cruises there was a significant trend in mean yield with time it = 2.40, P <0.02), so annual values under assumption A were calculated from the least squares estimating equation: y = 92.51 +3.11 (.r - 1821) where x = year of arrival (>1821) and y = average whalebone yield (lbs). This equation produces a yield of 96 lb per whale in 1822 and a yield of 370 lb per whale in 1910: yields before 1822 (for which no data exist) are taken as 96 lb per whale. For the 37 cruises where bone production was declared, there was no significant trend in aver- age yields with time (6 = -0.67 ± 1.26, t = 0.53, P > 0.61, so the overall mean yield of 497,840/884 or 563 lb (CV = 0.043) of bone per whale has been used under assumption B. This agrees well with a calculated mean of 629 lb and adult range of 250 to 330 kg (550 to 726 lb) for right whales from South Africa (Best 1970) and the North Atlantic (Collett 1909). respectively. Clark (1887a) stated that right whales in the South Atlantic yielded on average about 300 lb of bone per 100 barrels of oil in the male and 400 to 600 lb per 100 barrels in the female (equivalent to an actual yield of about 180 lb per whale in the male and 240 to 360 lb in the female). Although he claimed that right whales in the Indian Ocean were smaller than those in the South Atlantic, average whalebone yields are given as 240 lb for males and 360 lb for females. It is not clear why Clark's figures are somewhat lower than the others quoted here. (assumption B). According to Clark (1887a), the yield of bone in bowhead whales from the Atlantic- Arctic grounds averaged about 1,300 lb to 100 barrels of oil (or about 1,560 lb per whale), whereas in the Okhotsk Sea the yield was about 1,500 lb to 100 barrels of oil (or about 1,650 lb per whale). In both areas, however, Clark commented that whales found there "during recent years" were much smaller than those taken at the begin- ning of the fishery; the values given above re- ferred essentially to the start of the fishery. Ross (1974) calculated the average yield for a bowhead whale from the Hudson Bay stock as 1,065 lb, later revised to 916 lb (Ross 1979). For whales from the Davis Strait stock, the average yield was calculated as 1,392 lb (Ross 1979). Present esti- mates are therefore within the range of those given previously. Humpback Whales Of the 29 voyages taking humpback whales that were analyzed, only 7 had associated whalebone production (Fig. 7). There was no sig- nificant trend in the mean yield of whalebone per whale, either under assumption A ib = -0.21 ± 1.20, ^=0.17, P>0.8) or B (6 = 1.95 ± 5.12, t = 0.38, P > 0.7), so the relevant overall means have been used, i.e., 15,116/ 1,137 = 13 lb (CV = 0.416) under assumption A and 15,116/345 = 44 lb (CV = 0.652) under assumption B. Scammon (1874) stated that humpback baleen was of inferior quality, but could be collected at a rate of about 400 lb per 100 barrels of oil; this would be equivalent in current calculations to a yield of 98 lb whalebone per whale. Mitchell and Reeves (1983) confirmed that baleen from humpback whales was generally considered of poor quality, but pointed out that it was occasionally marketed. Bowhead >XlTales Of the 39 voyages taking bowhead whales that were analyzed, 37 included reference to landings of whalebone (Fig. 6). There was no significant trend in the mean yield of whalebone per whale, either under assumption A (b = 0.03 ± 7.19, /= 0.005, P>0.9) or B (6 = -2.87 ± 6.66, / = 0.4305, P > 0.6). Overall mean yields can therefore be calculated as 1,060,911/993 = 1,068 lb iCV = 0.098) per whale (assumption A) or 1,060,911/949 = 1,118 lb (CV = 0.095) per whale Gray VHiales There were no voyages available on which 10 or more gray whales were taken and for which whalebone production was declared. According to Henderson (1972:84): Unlike the valuable baleen of the right and bowhead whales, whalebone from the gray never became an im- portant part of the catch . . . little bone wa.s recorded in the cargos of the gray whaling vessels. The few recorded cargos of gray whalebone to arrive in San Francisco and 411 FISHERY BULLETIN: VOL, 85. NO. 3 San Diego did not appear until gray whaling was in decline and the price of right and bowhead whalebone had risen considerably after the mid 1860s. Rather than adopting an arbitrary value for the average whalebone yield of gray whales, it has been taken as zero. This means that estimates of the size of the catch of other species using whale- bone production may be correspondingly overesti- mated by an unknown but probably small amount. However, as whalebone production is used only to estimate the landed catch from 1880 onwards (see below), and there are no gray whales in the logbook sample after this date, the practical effect of this assumption is minimal. Estimates of Total Landed Catch of Whalebone Whales Figures for the importation of whale products into the United States have been based on table J of Starbuck (1878), supplemented by data in Hegarty (1959). As pointed out by Starbuck, it would appear from a comparison of imports and exports from 1804 to 1817 that much oil and bone must have been imported which was not credited to any port, and thus did not appear in table J. After 1817 exports as listed by Starbuck totalled 0.373 of imports for whale oil and 0.697 for whale- bone. It was presumably these figures that led Starbuck (1878) to propose that exportation of whale oil and bone for 1804 to 1817 represented one-third and two-thirds respectively of the im- portation, and I have followed his proposal in ad- justing the figures for 1804 to 1817 upwards on a prorata basis. The validity of this assumption is of course unknown. Inspection of table K in Starbuck (1878) also shows that importation figures for whalebone from 1838 to 1842 were "estimated" or "as- sumed", apparently at a rate of 10 lb of whalebone per barrel of oil, and may not therefore be very reliable. The data, summed by five yearly periods, are shown in Table 2. In order to estimate the total landed catch for any 5-yr period, the catch of each species given in Table 1 has been multiplied by its mean yield of oil or whalebone (corrected for the relevant year of catch, if necessary, using the median year in any 5-yr period) and the resulting production fig- ures summed. Comparison of this total with that in Table 2 for the same period then provides a scaling factor by which the catches in Table 1 have to be multiplied to obtain the total landed catch for that period. These scaling factors are shown in Table 2. In two of the three data sets (those for oil and whalebone factor A), there was a tendency for the scaling factors to be particularly high at the be- ginning of the time series, indicating that logbook coverage (and hence the reliability of extrapola- tions) was poor in the earlier years. The great differences between the two scaling factors for whalebone before about 1845 suggests either that a lot of whalebone was not being collected from the whales taken, or that it was not possible to allocate imports of it to a particular port or vessel (Starbuck 1878). The low ratio of whalebone to whale oil imported from 1805 to 1834 (Table 2) would indicate that the former was the more likely. Given the unreliability of import figures for whalebone from 1804 to 1817 and between 1838 and 1842, this suggests that oil production figures would be a more appropriate measure of the landed catch before about 1845. All three factors converge closely from 1855 to 1879, presumably indicating that full utilization was being made of both whalebone and whale oil. During this period the ratio of whalebone to whale oil imported ranged from 7.7 to 11.0, with a mean of 9.1 lb to a barrel of oil (Table 2). After 1880 the factors tend to diverge again, but this time the divergence is mainly between Table 2. — Five-year compilation of imports of whalebone and wtiale oil into the United States (from Starbuck 1878 and Hegarty 1959). Total US- imports Ratio: Bone Sea! ling fac Bone torsi Whale oil Whalebone Bone Peru Dd (barrels) (lb) oil Oil A B 1805- -09 285,969 360,981 1.3 993 87,7 14.9 1810- -14 96,759 151,921 16 — — — 1815- -19 130,666 179,793 14 582 232 9 1 1820- -24 246,793 429,447 1 7 455 55,4 9.4 1825- -29 245,777 1,039,134 42 136 347 6.9 1830- -34 680,729 1,846,907 2 7 94 13,4 3.0 1835- -39 917,064 7,947,069 87 84 34,8 88 1840- -44 1,032.080 10,159,715 9.8 9,1 19,0 10.2 1845- -49 1,324,305 14.073,773 106 7,4 9.9 8.3 1850- -54 1,193,253 17,143,100 14.4 83 123 112 1855- -59 985,480 10,854,100 11,0 59 6.6 6.0 1860- -64 509,037 4,388,800 86 50 4,9 4.3 1865- -69 390,415 4,045,575 10.4 4 1 4,4 3.9 1870- -74 256,714 2,054,769 80 39 3.8 3,3 1875- -79 152,907 1,176,690 7.7 6.5 7.8 6.2 1880- -84 138,654 1,785,354 12.9 4.6 10.7 8.1 1885- -89 134.438 1,989,176 14.8 59 107 9,4 1890- -94 62,614 1,667,478 266 7.6 150 12.9 1895- -99 21,531 1.067,130 496 20 5,0 4.7 1900- -04 17,175 614,830 358 4.0 68 6.3 1905- -09 15,710 460,100 293 3.5 68 5.3 1 Rounded to one decimal place. 412 BKST: LANDED CATCH OF RICHT WHAI.K the oil factor and both whalebone factors. This is accompanied by a marked increase in the ratio of whalebone to whale oil imported, to a peak of 49.6 lb to a barrel of oil from 1895 to 1899 (Table 2). It is assumed that over this period whalebone was collected in preference to whale oil, as described by Ross (1974) for bowhead whales: . . . with a dramatic rise in the price of whalebone the oil diminished to less than 20'^^ of the value of a whale after 1890 .... As a result whaling masters intensified the search for bone; . . . the crews simply stripped away the baleen, which was readily transportable, and left the rest of the carcass, including the bulky blubber, to rot. Oil returns, therefore, do not accurately reflect the number of whales killed in the late decades of whaling. It is apparent that after 1880, whalebone produc- tion would be a more accurate measure of the total landed catch. The economic basis for these shifts in interest is clearly shown by the average prices of whale oil and whalebone imported into the United States each year from 1804 to 1909 (Starbuck 1878: Hegarty 1959). These have been used to calculate the relative contribution of whalebone to the total value of a right whale, assuming a ratio at maxi- mum utilization of 10 lb of bone to a barrel ( = 31.5 gal) of oil per whale (Table 3). Whalebone made a relatively minor contribution to the value of a whale (<20'7f) up to 1839, ranged from 20 to Table 3. — Prices paid for whale products imported into the United States and the relative value of whalebone from a nght whale. Average price (US$) °o Contribution Whale oil Whalebone whalebone in total value of Period (per gal) (per lb) adult right whale 1805-1809 048 0,08 5,0 1810-1814 0,64 0,09 4,2 1815-1819 0,59 0,11 56 1820-1824 0,32 0,12 106 1825-1829 0,29 0,20 180 1830-1834 0,29 0,17 15,7 1835-1839 037 021 15,3 1840-1844 033 028 21,2 1845-1849 035 029 20,8 1850-1854 0,44 0.38 21 5 1855-1859 065 076 27.1 1860-1864 0,75 1.14 32.5 1865-1869 1.05 1 30 282 1870-1874 0.64 1.02 336 1875-1879 050 2.09 57.0 1880-1884 0,53 2.35 585 1885-1889 0,37 296 71.7 1890-1894 0,41 4,20 76.5 1895-1899 034 3,10 74.3 1900-1904 0,37 388 76.9 1905-1909 033 438 80.8 1910-1914 0,37 0,63 35.1 34% between 1840 and 1874, but increased rapidly in value thereafter to a peak of 80.8% in 1905-09. To conclude, oil production is considered the more accurate measure of the landed catch from 1804 to 1879, but whalebone production there- after. With the high value of whalebone after 1879 (comprising more than half the total value of the whale), it is likely that it would be utilized whenever possible. Hence scaling factor B would be the more appropriate to use. Using these factors, the total landed catch from the data tabulated by Townsend (1935) for Amer- ican vessels only between 1804 and 1909 is esti- mated as 125,883 whales, comprising 30,313 bow- head, 74,693 right, 18,212 humpback, and 2,665 gray whales (Table 4). Of the right whales caught, 182 (0.2%) were taken in the North At- lantic, 15,374 (20.6%) in the North Pacific, 32,191 (43.1%) in the South Atlantic, 14,699 (19.7%) in the South Pacific, and 12,247 (16.4%) in the In- dian Ocean. ESTIMATES BASED ON CATCH PER VOYAGE The use of production figures to estimate catches masks certain fundamental problems. Ac- cording to R. C. Kugler (in lift. 6 March 1985), neither Starbuck nor Hegarty apparently made much effort to report a vessel's total take of oil. They relied primarily on newspapers, especially the Whalemen's Shipping List after it began pub- lication in 1843. These reports, however, seldom gave more than the amount of oil on board at the time of the vessel's arrival. Only sporadically and inconsistently was shipped oil added in. This fac- tor would mean that the mean oil yields per whale calculated here would be underestimated, and the total number of whales landed correspondingly overestimated. Nevertheless, the mean oil yields derived in this paper agreed reasonably well with contemporary opinion on how much a particular species should yield. A further problem with the use of whale oil production is that the term "whale oil" was used to designate not only that from right and other species of whalebone whale, but also elephant seal and walrus oil. At certain periods the amounts landed of the latter were not negligible (Bockstoce and Botkin 1982; Busch 1985; Kugler in lift. 6 March 1985). However, it is not clear how much and in which direction this factor would affect the present analysis, depending on whether 413 FISHERY BULLETIN: VOL. 85, NO. 3 Table 4. — Numbers of baleen whales landed by U.S. whalers. 1805-1909, based on oil production up to 1879 and whalebone production thereafter. Period Northern right Southern right (arrival) Bowhead Atl. Pac. Atl. Pac. Ind. Humpback Gray Total 1805-09 1810-14 1815-19 — — — 4,268 — — — — 4,268 291 175 1,281 116 1,863 1820-24 — — — 3,683 — — — — 3,683 1825-29 — — — 3,663 — — 14 — 3,677 1830-34 — — — 8,804 665 674 47 — 10,190 1835-39 — — — 6,394 2,991 4,008 807 — 14,200 1840-44 — — 2.957 484 4,709 4,608 602 — 13,360 1845-49 669 — 8,001 404 2,567 949 654 — 13,244 1850-54 9,103 — 1,364 728 281 397 694 50 12,617 1855-59 7,273 6 1,381 311 423 488 1,422 382 1 1 ,686 1860-64 3,250 45 585 545 520 240 965 1,025 7,175 1865-69 2,956 8 441 445 103 177 511 886 5,527 1870-74 1,815 — 58 173 69 177 2,983 270 5,545 1875-79 677 39 85 352 7 85 2.475 52 3,772 1880-84 958 32 8 620 209 — 4,985 — 6,812 1885-89 711 47 356 496 674 28 1,600 — 3,912 1890-94 1,108 — 39 490 168 — 284 — 2,089 1895-99 860 5 56 33 — 28 47 — 1,029 1900-04 468 — 38 70 — 13 6 — 595 1905-09 174 — 5 53 32 375 — — 639 Total 30,313 182 15.374 32,191 14,699 12,247 18.212 2,665 125,883 the whale oil production of the voyages listed by Townsend was diluted to a greater or lesser ex- tent with seal and other oil than the total produc- tion. The catch-per-voyage analysis attempts to avoid the problems created by the incomplete re- porting of the products of a voyage, and (at least partially) those arising from the dilution of whale oil with seal, walrus, and other oils. In order to make some further correction for voyages that were entirely devoted to sealing, all voyages recorded as returning only elephant oil, or as "skinning voyages", or voyages labelled as sealing by Starbuck (1878) and Hegarty (1959) have been excluded. In addition, all voyages from the Connecticut ports of New London, Stonington, or Mystic that were recorded as being bound for S. Shetlands, Desolation, Falklands, Hurds Island, or Crozettes and that returned with whale oil but no whalebone have been omitted on the grounds that these were probably sealing voyages. This has resulted in a total omission of 141 voyages between 1804 and 1921. Obviously this figure does not include all voy- ages on which seal oil was taken, as many seals were taken on combination sealing/whaling voy- ages. Between 1840 and 1890, an average 25*^ of "whaling" vessels leaving New London are said to have visited Desolation or Heard Island for ele- phant seals (Busch 1985), but of 110 voyages de- parting to these islands from New London during this period, 45 were reported as bringing back sperm oil and/or whalebone as well as "whale" oil (Starbuck 1878; Hegarty 1959). Starbuck (1878) and Hegarty (1959) also listed a number of mixed voyages from other ports in which small amounts of whale oil were landed but no whalebone. While some of these might repre- sent voyages on which whales with inferior whalebone (such as humpback or gray whales) were taken, other such small consignments of whale oil might have originated from seals or from "blackfish" (pilot whales Globicephala spp.). Pilot whales were sometimes taken by whalemen to supplement their cargoes, the oil being rated as common whale oil. Clark (1887b) listed 36 voy- ages on which from 2 to 200 barrels of blackfish oil was brought home, 33 (91.7%) of them bring- ing back 100 barrels or less. To investigate this further, mixed voyages on which 100 barrels or less of whale oil but no whalebone were landed (from Starbuck and Hegarty) were compared with the catch composition of the same voyages as given by Townsend (1935). Of 153 such voyages, baleen whales were reported as being taken on 55 (35.9'^ ) voyages, with the proportion approaching 1009^ as the amount of whale oil approached 100 barrels (Table 5). Consequently for each 5-yr time period the number of mixed voyages reporting 100 barrels or less of whale oil but no whalebone was adjusted by the proportion of such voyages in Townsend's sample that were reported as taking 414 BKST LAN'DEn CATCH OF KlClir WMAI.K Table 5. — Proportion of mixed voyages by US. whalers landing small consignments of whale oil (but no whale- bone) on which whalebone whales were taken. Table 6. — Breakdown of Townsend's (1935) sample into voyage- type, with scaling factors (A1, A2) derived from numbers of such voyages in Starbuck (1878) and Hegarty (1959). Proportion of Amount of Number of mixed voyages on whale oil mixed voyages which whalebone reported (barrels) examined whales taken 1-10 53 0.113 11-20 19 0.263 21-30 18 0.444 31-40 14 0500 41-50 11 0364 51-60 10 0.300 61-70 12 0.750 71-80 8 0.750 81-90 2 1.000 91-100 6 0833 baleen whales during the same period. This re- sulted in an effective conversion of 812 mixed voyages between 1805 and 1910 to sperm whaling voyages. Because Starbuck (1878) and Hegarty (1959) listed voyages by the year of departure, the data in the catch-per-voyage analysis has been com- piled against year of departure rather than (as was done for the production-based analysis) by year of arrival. The numbers of sperm, mixed, and whalebone voyages in Townsend's sample for each 5-yr period are given in Table 6, together with the scaling factors Al and A2 for mixed and whalebone whalers respectively. The latter repre- sent the ratio of the number of voyages of each type in the Townsend sample to the number of similar voyages in Starbuck/Hegarty for that pe- riod, after correction (as described above) for voy- ages believed to be sealing rather than whaling. These scaling factors are then applied to the total numbers of whales landed in the Townsend sam- ple for that period and voyage-type, and the re- sults for each voyage-type added to give the total number of each species for that period. This analysis provides an estimate of the landed catch of whalebone whales from American vessels between 1805 and 1914 as 117,308, com- prised of 29,788 bowhead, 70,343 right. 14,164 humpback, and 3,013 gray whales (Table 7). Of the right whales. 186 iO.S'/i) were taken in the North Atlantic, 14,480 (20.6'h in the North Pacific. 28.532 (40.6Vf) in the South Atlantic, 14,652 (20.8Vr) in the South Pacific, and 12,493 ilLS'/c) in the Indian Ocean. DISCUSSION The two methods used give somewhat similar Type ! of voyage Period Sperm n Mixed Whalebone n A2 (departure) n Al Total 1805-09 2 1 8.00 1 51.00 4 1810-14 1 0 — 0 — 1 1815-19 3 1 89.00 2 49.50 6 1820-24 6 3 23.67 1 117.00 10 1825-29 10 12 8.92 8 13.38 30 1830-34 29 47 6.00 11 18.09 87 1835-39 54 75 8.28 9 12.33 138 1840-44 56 88 8.64 4 28.75 148 1845-49 53 93 6.25 8 9.88 154 1850-54 51 102 5.74 9 16.56 162 1855-59 52 101 5.29 11 11.82 164 1860-64 49 54 4.76 16 3.06 119 1865-69 55 83 4.02 13 5.46 151 1870-74 30 31 4.77 4 6.75 65 1875-79 44 48 3.67 7 6.14 99 1880-84 29 28 4.32 11 11.91 68 1885-89 17 22 2.91 14 13.79 53 1890-94 18 16 2.56 8 18.50 42 1895-99 7 11 3.27 10 590 28 1900-04 22 9 2.67 10 4.70 41 1905-09 19 10 1.60 5 580 34 1910-14 11 2 2.50 3 2.67 16 Total 618 837 165 1.620 results, estimates of the landed catch differing by <10'^ in all cases except for South Atlantic right whales and humpback whales (where the production-based estimates exceeded the catch per voyage estimates by 13 and 297^ respectively) and gray whales (where the catch per voyage esti- mate exceeded the production estimate by 13%). Nevertheless, both sets of estimates are essen- tially derived from the same basic data (Townsend's sample), and a more fundamental problem with both analyses is how representative this sample was of the contemporary Yankee fish- ery. The 1,651 voyages examined by Townsend are equivalent to only 129^^ of the estimated total number of voyages made by American pelagic whalers (Sherman 1965), but the scaling factors in Tables 2 and 6 indicate that the coverage of voyages in some periods (particularly prior to 1830) was much less than this. Unless the log- book sample is truly random with respect to the species and numbers of whales taken, any simple reconstruction of the total catch therefrom is likely to be inaccurate. The catch per voyage analysis included some stratification of the Townsend sample, so that ex- trapolations to the total fleet might be more rep- resentative. To this extent, therefore, the catch per voyage method might seem the more reliable, 415 FISHERY BULLETIN: VOL. 85, NO, 3 Table 7, — Numbers of whalebone whales landed by U,S, whalers, 1805-1914. as calculated from the catch per voyage. Period (depart- ure) Bowhead Northern right Atl, Pac, Southern right Atl, Pac, Ind, Humpback Gray Total 1805-09 1810-14 1815-19 — — — 1,849 — — — — 1,849 248 149 1,958 178 2,533 1820-24 — — — 4,468 — — — — 4,468 1825-29 — — — 3,617 535 — 63 — 4,215 1830-34 — — — 8,902 390 600 384 — 10,276 1835-39 — — 149 5,662 5,190 5,561 285 — 16,847 1840-44 — — 5,728 598 3,542 3,939 657 — 14,464 1845-49 2,148 — 5,578 515 1,485 421 669 — 10,816 1850-54 10,260 6 951 511 161 775 565 281 13,510 1855-59 5,454 48 1,221 429 579 501 2,317 1.279 1 1 ,828 1860-64 2,583 — 152 516 280 96 447 665 4,739 1865-69 3,215 11 434 354 105 342 2,153 755 7,369 1870-74 744 — 52 81 14 29 2,561 33 3,514 1875-79 417 40 16 400 95 40 1,801 — 2,809 1880-84 604 9 48 216 186 — 1,881 — 2,944 1885-89 729 69 90 113 122 9 134 — 1,266 1890-94 2,098 — 26 44 — — 59 — 2,227 1895-99 715 3 16 23 — 26 10 — 793 1900-04 353 — 19 27 — — — — 399 1905-09 180 — — 40 10 154 — — 384 1910-14 40 — — 18 — — — — 58 Total 29,788 186 14,480 28,532 14,652 12,493 14,164 3,013 117,308 but other problems (allowance for sealing voy- ages, correct allocation of incomplete voyages) may not have been adequately solved. Further- more, even the stratification by voyage-type may have been insufficient to correctly portray the species composition of the total fleet; an alterna- tive procedure might be to stratify by home port, but this would probably involve too fine a stratifi- cation for the size of the sample available. A further problem identified with the Townsend sample is that there may be occasional misidentification or omission of catches (see Bockstoce and Botkin 1983). There is no indica- tion of the extent of this problem (which would require checking Townsend's tabulations against the original journals), but it means that the accu- racy of the extrapolations made in this paper may be adversely affected to an unknown degree. Some independent estimates of the landed catch of various stocks have been made. Hender- son ( 1972) estimated that 4,958-5,058 Californian gray whales were taken by pelagic whalers be- tween 1846 and 1874, whereas calculations from the Townsend sample are that 2,665 to 3,013 gray whales were taken over the period 1850 to 1879, Henderson's estimate, however, includes the catches of non-U. S. vessels. From data in Hender- son's table II it can be calculated that 64.9Vf of the 21,135 barrels of oil from Scammon's Lagoon be- tween 1858 and 1873 were taken by U.S.- registered vessels. If this proportion is applied to the total catch, it means that U,S, pelagic whalers may have accounted for a landed catch of 3,218 to 3,283 gray whales. On this basis, calculations from the Townsend sample may be about 6 to 19% too low. Bockstoce and Botkin (1983) calculated that 16,600 bowhead whales were taken from the Western Arctic population by the pelagic whaling industry between 1849 and 1914; this apparently included catches by non-U. S. vessels. Henderson has also estimated that the total catch of bowhead whales from the Okhotsk Sea stock between 1847 and 1867 was about 15,200 animals, with another 92 known to have been taken between 1867 and 1896 (Kugler 1984). About 90% of the voyages to the Okhotsk Sea between 1847 and 1867 were made by American whaleships, which if consid- ered applicable to the catch would mean that they took about 13,760 bowhead whales from 1847 to 1896, Ross ( 1979) has estimated the catch of bow- head whales by American whalers in the Davis Strait from 1847 to 1891 as 413, and that in Hud- son Bay from 1860 to 1912 as 532 animals; his figure for the Beaufort Sea of 794 bowhead whales between 1889 and 1908 is assumed to be included in Bockstoce and Botkin's calculations for the entire Western Arctic. Combining the data from Ross (1979), Bockstoce and Botkin (1983), and Kugler (1984) indicates a total bowhead 416 BEST: LANDED CATCH OF RI(!HT WHALE catch of 31.305 animals (some of which may have been taken by non-U. S. vessels). This is 3 to 5% higher than the total estimate of 29,788 to 30,313 bowhead whales from the Townsend sample. Reeves and Mitchell (1986) have attempted to reconstruct the American pelagic catch of right whales in the North Atlantic during the nine- teenth century. They document at least 116 right whales that were killed and processed by pelagic whales between 1855 and 1897. The present analysis indicates a total landed catch by U.S. whalers of 182 to 186 right whales over the same period. These comparisons suggest that, apart from gray whales, the estimates of landed catch ob- tained in this paper are not unduly biased. They are, however, clearly only first approximations. A much more detailed approach, including exami- nation of primary source material, is required be- fore a more reliable assessment of the American catch of right whales can be made. In particular, there needs to be more adequate sampling of log- books prior to 1830. It should also be stressed that the figures pro- duced here are estimates of the landed catch; fur- ther work is needed to determine the numbers of animals that were struck and lost, and the pro- portion of these that might have died, before an estimate of the total kill made by the American fishery can be made. Such research, requiring consultation of primary sources, is outside the scope of this paper. Nevertheless, a significant proportion of the landed catch of some species ap- parently consisted of whales found dead. In the Townsend abstracts examined here, there were records of 246 baleen whales processed that were found dead: 127 bowheads (6.39( of the landed catch). 103 right whales (2.97r of the landed catch), 5 humpback whales (0A9( of the landed catch), and 11 gray whales (or 4.4''7r of the landed catch). These figures might be underestimates if (as seems likely) not all the whales found dead were recorded as such in the logbooks or logbook abstracts. Most of these whales probably died as a result of whaling-related injuries. If so, this fact should be borne in mind when corrections are ap- plied to the landed catch to account for whales struck and lost that subsequently died. With no correction for animals dying after being struck and lost, the estimated number of right whales taken between 1805 and 1874 as calculated in this paper, 68.484 to 70,250 (of which 79*^ were southern right whales), is about one third of Starbuck's original estimate for the same period. This compares with an estimated total catch by French pelagic whalers between 1817 and 1868 of 11,000 right and bowhead whales (Du Pasquier 1986). Comparable figures for the British take are not yet available. ACKNOWLEDGMENTS I am indebted to M. A. Meyer for the labor of extracting relevant information from Starbuck and Townsend's tables. Access to some of Townsend's original tabulations was very kindly provided by H. E. Winn (University of Rhode Is- land) and W. E. Schevill (Woods Hole Oceano- graphic Institution), and J. H. Prescott (New Eng- land Aquarium) and W. F. Perrin (National Marine Fisheries Service) assisted in arrange- ments for transport of the data at very short no- tice. Valuable comments on an earlier draft of this paper were received from an anonymous re- viewer and J. M. Breiwick. Statistical advice was provided by D. S. Butterworth. Financial support for conducting this project came from the South African Department of Transport via the South African Scientific Committee for Antarctic Re- search. LITERATURE CITED Best. P B 1970. Exploitation and recovery of right whales Eubal- aena australis off the Cape Province. S. Afr. Div. Sea Fish. Invest. Rep. 80, 20 p. BOCKSTOCE. J R . AND D B BOTKIN 1982. The harvest of Pacific walruses by the pelagic whal- ing industry, 1848-1914. Arct. Alp. Res. 14 (3):183-188. 1983. The historical status and reduction of the western Arctic bowhead whale {Balaena mysticetus) population by the pelagic whaling industry, 1848-1914. Rep. Int. Whaling Comm. Spec. Issue 5, p. 107-141. Brownell. R L , P. B Best, and J H. Prescott (editors). 1986. Right whales, past and present status. Rep. Int. Whaling Comm. Spec. Issue 10, 289 p. Busch. B. C. 1985. The war against the seals. A history of the North American seal fishery. McGill-Queen's Univ., Kingston and Montreal. 374 p. Clark. A H 1887a. The whale-fishery: 1. — History and present condi- tion of the fishery. In G. B. Goode and staff of associates, The fisheries and fishery industries of the United States. Section V. History and methods of the fisheries. Vol. II, p. 1-881. Gov. Pnnt. Off., Wash., DC. 1887b. The blackfish and porpoise fisheries. In G. B. Goode and staff of associates. The fisheries and fishery industries of the United States. Section V. History and methods of the fisheries. Vol. II, p. 297-310. Gov. Print. Off, Wash., DC. 417 FISHERY BULLETIN: VOL. 85, NO. 3 COLLETT. R. 1909. A few notes on the whale Balaena glacialis and its capture in recent years in the North Atlantic by Norwe- gian whalers. Proc. Zool. Soc. Lond. (7):91-97. DuPasquIER, T. 1986. Catch history of French right whaling mainly in the South Atlantic. Rep. Int. Whaling Comm. Spec. Issue 10, p. 269-274. Harmer, S F 1928. The history of whaling. Proc. Linn. Soc. Lond. 140 (1927-281:51-95. Hegarty. R B 1959. Returns of whaling vessels sailing from American ports. A continuation of Alexander Starbuck's "History of the American whale fishery" 1876-1928. The Old Dart- mouth Historical Society and Whaling Museum, New Bedford, MA, 58 p. Henderson. D A 1972. Men & whales at Scammon's lagoon. Dawson's Book Shop: Los Angeles, CA, 313 p. HOHMAN, E. P 1928. The American whaleman. A study of life and labor in the whaling industry. Longmans, Green and Co., N.Y., Lond., Toronto, 355 p. Kugler. R C 1984. Historical survey of foreign whaling: North Amer- ica, /n Arctic whaling, p. 149-157. Proceedings of the international symposium, Arctic Whaling, February 1983. Arctic Centre, University of Groningen, Nether- lands. Mitchell. E D , and R R Reeves 1983. Catch history, abundance, and present status of Northwest Atlantic humpback whales. Rep. Int. Whal- ing Comm. Spec. Issue 5, p. 153-212. Omura. H 1958. North Pacific right whale. Sci. Rep. Whales Res. Inst., Tokyo 13:1-52. Reeves. R R . and Mitchell, E 1986. American pelagic whaling for right whales in the North Atlantic. Rep. Int. Whaling Comm. Spec. Issue 10, p. 221-254. Ross, W. G 1974. Distribution, migration, and depletion of bowhead whales in Hudson Bay, 1860 to 1915. Arct. Alp. Res. 6(l):85-98. 1979. The annual catch of Greenland (bowhead) whales in waters north of Canada 1719-1915: a preliminary compi- lation. Arctic 32(21:91-121. SCAMMON, C M 1874. The marine mammals of the North-western coast of North America, described and illustrated: together with an account of the American whale-fishery. John H. Car- many & Co., San Franc, 319 p. Schevill, W. E , AND K E Moore 1983. Townsend's unmapped North Atlantic right whales (Eubalaena glacialis) . Breviora 476:1-8. Sherman, S C 1965. The voice of the whaleman, with an account of the Nicholson Whaling Collection. Providence Public Li- brary, Providence, 219 p. Starbuck, a 1878. History of the American whale fishery from its ear- liest inception to the year 1876. U.S. Comm. Fish and Fish., pt. IV, Rep. Comm. 1875-1876, 768 p. TowNSEND. C H 1935. The distribution of certain whales as shown by log- book records of American whaleships. Zoologica (N.Y.) 19:1-50. Wray. P , AND K R Martin 1983. Historical whaling records from the Western Indian Ocean. Rep. Int. Whaling Comm. Spec. Issue 5, p. 213- 241. 418 ESTIMATING DENSITY OF DOLPHIN SCHOOLS IN THE EASTERN TROPICAL PACIFIC OCEAN BY LINE TRANSECT METHODS Rennie S. Holt ABSTRACT Data were collected from aerial and research ship surveys to estimate density of dolphin schools in the eastern tropical Pacific using line transect (LTi theory. The surveys were conducted from 1977 through 1983. Several assumptions of LT theory were investigated for both aerial and ship data. Factors were developed to alleviate eflects of suspected violations of the assumptions. I estimated densities from data stratified into an inshore area surveyed by planes and an offshore area surveyed by ships. The density estimate for the inshore area was 4.18 schools/1,000 km2 and 2.04 for the offshore area. For the entire area, the density estimate was 2.71 schools/1,000 km^. Adjustments for possible biases owing to adverse sea state and sun glare conditions increased the inshore estimate by B'.v and the total area estimate by 4%. The National Marine Fisheries Service (NMFS) is responsible for assessing the status of those dol- phin stocks taken incidentally by tuna purse sein- ers in the eastern tropical Pacific (ETP) Ocean. Techniques used to assess these stocks (Smith 1979") require estimates of school density, so den- sity estimates were made in 1975 (Smith 1975'^) and in 1979 iHolt and Powers 19821. Since 1979, NMFS has collected additional information to test the assumptions of its statistical methods and to further survey the areas inhabited by the dol- phins. In this paper, I present analyses of data collected from 1977 through 1983 to determine density estimates of dolphin schools in the ETP. In addition, I investigate several factors which may bias the estimates. To obtain estimates of density of dolphins (indi- viduals) it is further necessary to consider school size, the proportions of various species in mixed schools, and areas inhabited by the various stocks. Estimation of these factors is complex; they are to be dealt with elsewhere and are not addressed in this paper. MATERIAL AND METHODS Surveys Data used to calculate the density of dolphin schools were collected during several years. Aerial surveys were conducted in 1977 and 1979 (Fig. 1 ), and nine research ship cruises were made during 1977, 1979, 1980, 1982, and 1983 (Fig. 1). Most surveys were conducted between January and early April; one of the 1977 ship cruises was made in October and the two 1980 cruises were made from May through August. A two-engine PBY amphibious patrol bomber was used in the 1977 aerial survey (SWFC 1978^), and a four-engine PBY bomber was used in the 1979 aerial survey (Jackson 1980^). Operating and viewing conditions aboard the two aircrafts were similar. Both planes cruised at 148-240 km/ hour (80-130 kn) and had bubble-shaped waist windows. The PBY used in 1977 had a flat bow window which was shaped like an isosceles trape- zoid. The 1979 PBY had a round bubble-shaped bow window. The round bubble window allowed 'Southwest Fisheries Center La Jolla Laboratory. National Marine Fisheries Service. NOAA. P.O. Box 271. La Jolla. CA 92038. -Smith. T. 1979. Report of the status of porpoise stocks workshop (.August 27-31. ]979i. Southwest Fish. Cent. Adm. Rep. No. LJ-79-41. La Jolla, CA, 120 p. •'Smith. T. 197.5. Estimates of sizes of two populations of porpoise iStenclla) in the eastern tropical Pacific Ocean. Southwest Fish. Cent. Adm. Rep. No. LJ-7.5-65, La Jolla. CA. 88 p. Manuscript accepted February 1987. FISHERY BULLETLN: VOL." 85. NO 3. 1987 ■*SWFC iSouthwest Fisheries Centeri. 1978. Aerial survey trip report, January-June 1977. Southwest Fish. Cent. Adm. Rep. No. LJ-78-01, 73 p. National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. SJackson, T. 1980. Report: Porpoise population aerial sur- vey of the eastern tropical Pacific Ocean, January 22-April 25, 1979. Southwest Fish Cent. Adm. Rep. No. LJ-80-01, 74 p. National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. 419 FISHERY BULLETIN: VOL. 85, NO. 3 A. 30°N 20' 10= 0° - 10= 20°S Northern V Offshore Southern Offshore Area 160° 150° 140= 130° 120° 110° 100* 90= 80= 70° W B. 30° N 20= 10= 0= 10= 20°S 4 Northern Offshore Southern Offshore Area 160° 150= 140° 130° 120° 110° 100= 90= 80= 70° W Figure l.— Tracklines for 1977 and 1979 aerial (A) and combined 1977, 1979, 1980, 1982, and 1983 ship (B) surveys. 420 HOLT: DENSITY OF DOLPHIN SCHOOLS better lateral viewing, but both provided unob- structed forward and downward views. Two research vessels were used to collect the shipboard data. The NOAA ship David Starr Jor- dan was used during all years and the NOAA ship Toivnsend Cromwell joined it in 1977, 1979, and 1980. Both vessels were similar in length and cruising ability. Binoculars used to locate ani- mals were mounted approximately 10.7 m above the sea on the Jordan but were only 6.1 m above the sea on the Cromwell. In addition, observers aboard the Jordan used 20 x binoculars during the 1977 surveys and 25 x glasses on the rest of the surveys; observers aboard the Cromwell used only 20 X glasses during their surveys. Conse- quently, viewing conditions were generally much better on the Jordan. Study Area Survey efforts traversed the combined range of ETP dolphin stocks defined by Au et al. (1979)*^. The range was partitioned into "inshore" and "offshore" areas (Fig. 1). Airplanes were used to survey the inshore area, and ship surveys were conducted in both areas during each year, except during 1977 when ships surveyed only the off- shore area. Data Collection Aerial Data Data collecting procedures used during the aerial surveys are described by SWFC (fn. 4), Jackson (fn. 5), Holt and Powers (1982), and Cologne and Holt (1984)'. As the airplanes tra- versed predetermined tracklines (Fig. 1), the ob- servers recorded schools on and to either side of the lines. Observers searched through the bow window and from windows located on either side of the plane. The bow observer was responsible for detecting schools on the trackline (a path under- neath the plane 0.19 km wide). The searching mode was halted if environmental or oceano- graphic conditions restricted the observer's view of the trackline or when the plane was diverted from the trackline for closer examination of a 6Au. D.. W. Ferryman, and W. Perrin. 1979. Dolphin dis- tribution and the relationship to environmental features in the eastern tropical Pacific. Southwest Fisheries Center Status of Porpoise Stocks working paper SOPS/79'36, 59 p. ^Cologne, J., and R. Holt. 1984. Observer effects in ship- board sight surveys of dolphin abundance. Southwest Fish. Cent. Adm. Rep. No. L-J-84-.30, 42 p. National Marine Fish- eries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. school. Additional schools detected during these diversions were not included in the density analy- sis. Sea conditions were measured on the Beaufort scale (Bowditch 1966), which ranged from very flat, glassy seas (Beaufort 0 conditions) to rough seas with numerous large, white-capped waves (Beaufort 5 conditions). Sun location was de- scribed by horizontal and vertical position rela- tive to the bow observer (Holt 1983a). These were recorded for each segment of effort. Biological and environmental data were recorded at each sighting (Holt and Powers 1982). Data included species identification, school size estimates, sea state, sun position, and perpendic- ular distance to the school from the trackline. School size estimates consisted of an observer's "best" estimate plus an estimate of the minimum and maximum range. Ship Data Shipboard collection procedures are described in the various cruise reports (unpublished docu- ments available from the SWFC) and by Holt (1983b). Procedures and data recorded on ship- board surveys were similar to those for aerial sur- veys. Two observers used binoculars located on each side of the ship to search from directly ahead to abeam of their respective sides of the ship. Starting in 1979, sea state was recorded at the beginning of each effort segment (leg). Sun posi- tion was recorded during the 1982 and 1983 ship surveys. The bearing (0) and radial distance (r) to a school from the ship were recorded, and perpen- dicular distance (y ) was then calculated as y = r sin 9. In surveys conducted before 1980, observers rounded estimates of sighting angles to multiples of 5° or 10°, and radial distances to multiples of 185 m (0.1 nmi) within the first 1.85 km (1 nmi), and to 0.93 km (0.5 nmi) multiples at larger dis- tances (Fig. 2). During training, observers on the 1980 surveys were told of previous rounding inac- curacies and instructed to make estimates as pre- cise as possible. However, they were still unable to make precise visual estimates of angles and distances for schools recorded at great distances from the ship (Fig. 2). During the 1982 and 1983 surveys, estimates of bearing were recorded using a 360° graduated washer attached to the base of the binoculars, and the radial distances were measured using a graduated reticle enclosed in the right eyepiece of the binoculars (Holt 1983b). 421 FISHERY BULLETIN: VOL. 85, NO. 3 1979 1980 10 15 211 25 3D 35 40 4S 50 SS 60 65 SIGHTING PtHGLE IM DEGREES 70 75 BO 85 30 ttn [HIM i 10 IS 20 25 30 35 40 45 50 55 60 65 70 75 SIGHTING flMGLE IM DEGREES ml kfl Ik 0 0 0.5 l.O 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 G.O 6.5 7.0 7.5 8.0 RflDIflL DISTflriCE m nflUTICFIL HILES 15 14- 12- I JUl 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 S.S G.O 6.5 7.0 7.5 8.0 RFIDIflL DISTflnCE IM MHUTICflL nlLES 2.00 1.75- > 1. so- ul u l.ZS- f 1.00 -I m S o.so- p- 0.25 0.00 0.0 Hfn _ji_p ^ 4i^_ OS 10 l.'S 2^0 2.5 5.0 5.5 PERPEMDICULflR DISTflnCE In nflUTICFIL MILES f 1.00 _J 5 0.75- a: at o K 0.50 o- 0.25- M nn^ nn r^Hq 0.5 1.0 1.5 2.0 2.5 5.0 3.5 PERPEnDICULfiR DISTRnCE in nflUTICRL niLES FK'.URE 2.— Distribution of sighting angle, radial distance, and perpendicular distance from 1979, 1980, With this system, the rounding to convenient val- ues was not as evident (Fig. 2); hou'ever, mea- surements may still be inaccurate. ANALYTICAL METHODS Vessel data for area, sea state, sun glare, and observer performance strata were compared using rates of detection for all schools encoun- tered within 2.13 km perpendicular distance of the ship (schools/1,000 km searched) and esti- mates of density of schools (schools/1,000 km^). Similar comparisons of aerial data were com- pleted using rates of detection for all schools en- countered within 1.85 km perpendicular distance of the trackline, rates of detection for trackline schools, and estimates of school density. Density estimates were made using line tran- sect (LT) theory (Burnham et al. 1980). The basic equation (Seber 1973) is D n fiO) 2L 422 HOLT: DENSITY OF DOLPHIN SCHOOLS 1982 6 5^ 10 IS 2b 2S 3b 3S 40 5 were omitted from the analyses. The presence of whitecaps was important because an- imal splashes were used as sighting cues during calm conditions but could not be easily distin- guished from whitecaps during rough conditions. For aerial data and 1982-83 ship data, sun glare effects were investigated by classifying ef- fort at various sun positions into "good" and "poor" categories depending on the amount of sun glare on the trackline (see Holt* for method used *^Holt, R. 1984. Testing the validity of line transect theory to estimate density of dolphin schools. Southwest Fish. Cent. Adm. Rep. No. LJ-84-31, 56 p. National Marine Fisheries Ser- vice, NOAA, P.O. Box 271, La Jolla, CA 92038. 423 FISHERY BULLETIN: VOL. 85. NO. 3 to record position of sun relative to the platform and for criteria used to define sun categories for aerial data). Criteria used for ship data were based upon observations recorded during a subse- quent ship survey (Hohn^). Hohn found poor sun conditions on the trackline only when horizontal sun position was 12 and vertical position was 1, 2, or 3 or when clouds were accompanied by fog or rain. All other effort was defined as occurring during good conditions. In order to apply the Fourier series (FS) model to aerial and ship data, I structured the data by 1 ) selecting appropriate interval widths for grouping the perpendicular sighting distribu- tions (data cutpointsi, 2) choosing a maximum ob- servation distance perpendicular to the trackline (truncation point), 3) developing criteria to select the appropriate number of terms for the FS model, and 4) choosing the type of transformation to use in compensating for measurement error in the shipboard data. Based on a subset of the ship data (Holt^*^), I used an interval width of 0.37 km (0.2 nmi) and truncated the perpendicular distance distribu- tions at 3.7 km (2.0 nmi). Since perpendicular distance distributions for the ship data, and also to a lesser extent for aerial data, have very promi- nent modes or "spikes" at the origin, existing criteria to select the appropriate number of terms in the FS model were unsatisfactory. Therefore, I selected the model which provided the best visual fit to the distributions near the origin (Holt fn. 10). This technique was easily applied and was consistent among data sets. For use of the tech- nique I assumed that the sizes of the spikes near the origins of the perpendicular distance distribu- tions were indicative of relative density among the data sets. To minimize the effects of recording errors, the data were smoothed using the tech- nique "smearing" (Butterworth 1982; Hammond 1984). Based on previous investigations of aerial data (Holt and Powers 1982), I selected a truncation point of 1.94 km (1.05 nmi) and an interval width of 0.19 km (0.1 nmi) for the aerial data. I used the same technique as used for ship data to select the appropriate number of terms in the FS models; ^A. Holin, Southwe.st Fi.sheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038, pers. commun. January 1985. '"Holt, R. 1984. Estimation of density of dolphin schools in the eastern tropical Pacific Ocean u.sing line transect meth- ods. Southwest Fish. Cent. Adm. Rep. No. L.J-84-32. 72 p. National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. however, the aerial data were not smoothed be- cause there was no evidence that the data con- tained estimation errors as did the ship data. An estimate of density in the total area (Z)^.) was calculated by combining the aerial inshore (Z),) and ship offshore (Dq) density estimates weighted by the relative sizes of the inshore (A, ) and offshore (Aq) areas as D. D,A, + DqAq The estimate of variance of D^ is A,^Vdr(D, ) + Ao-VdriDo) Vdr(D,) (A, + Aq)" RESULTS Factors Affecting Density Estimates Aerial Data Density estimates for the aerial data in the in- shore area during calm seas or with minimal sun glare were more than twice the estimates for data taken during rough seas or poor sun conditions (Table 1). Differences in estimators were even greater for sea state and sun glare interaction effects. These differences may have occurred be- cause observers failed to detect trackline schools during poor conditions or because sea state condi- tions were spatially confounded with distance from shore. Therefore, these differences may be reflecting a decreasing onshore-to-offshore den- sity gradient. This was investigated by partition- ing the inshore aerial data into "coastal" and "offshore" bands for each Beaufort sea state (Fig. 3) and sun glare condition (Fig. 4). Sufficient data were not available in each band to stratify detection rates by eaqh sun and sea state interac- tion category. Sea conditions dufing the aerial surveys were rougher offshore than nearshore. More searching was done in the coastal band during low Beaufort states, whereas tnore searching was done in the offshore band at higher Beaufort states (Fig. 3). The rates of detecting dolphin schools were higher at each corresponding Beaufort state in the coastal band than in the offshore band (Fig. 5). The rates of detecting trackline schools were generally higher in the coastal band; how- ever, these rates were based upon very few 424 HOLT: DENSITY OF DOLPHIN SCHOOLS Table 1 . — Estimates of school density made during all conditions and during calm and rough seas using aerial and ship data: estimates made during good and poor sun condition using aerial data. Estimates are made for data in the Inshore, offshore and total areas. Estimates for all conditions were calculated using 1977 through 1983 data and estimates for sun and sea state conditions were calculated using 1979 through 1983 data. Estimates are also presented for data collected during an aerial experiment testing effects of sea state and sun glare. Number Density Distance schools 0) searched detected (schools/ SE CV Vanable (km) in) 1,000 km2) (D) (D) Inshore area Aerial data all data 34.006 152 4.18 0.902 0.216 calm seas 8,920 70 8.48 2.198 0.259 rough seas 25,086 82 2.71 0.611 0.255 good sun 1 1 ,994 74 6.57 1.504 0.229 poor sun 22,012 78 2.87 0.505 0.176 calm-good 3,026 30 12.64 5.290 0.418 calm-poor 5,894 40 6.24 2.311 0.370 rough-good 8,967 44 4.29 1.202 0.280 rough-poor 16,118 38 1.78 0.460 0.258 Ship data all data 27,840 379 4.47 0.514 0.115 calm seas 8,008 170 7.32 1.259 0.172 rough seas 14,668 149 4.05 0.772 0.191 Offshore area Ship data all data 46,567 322 2.04 0.263 0.129 calm seas 4,623 72 4.91 1.414 0.288 rough seas 20,976 99 2.01 0.435 0.217 All areas Ship data all data 74,407 626 2.95 0.253 0.086 calm seas 12,631 242 6.53 0.991 0.152 rough seas 35,644 248 3.02 0.445 0.147 Holt (text fn. 10) aerial experiment calm-good 1,414 37 29.18 7.357 0.252 calm-poor 3,014 81 23.78 5.888 0.248 rough-good 1,886 42 39.42 8.193 0.208 rough-poor 5,467 103 20.16 4.513 0.224 schools ( 18 trackline schools in the coastal and 10 schools in the offshore band were detected). Lower offshore estimates for data recorded under the same Beaufort state were consistent with a de- creasing onshore-offshore density gradient. Within each band, sea state conditions were also spatially stratified because the lower Beau- fort conditions occurred mostly in the nearshore and northern regions of each band (Fig. 3). Pre- dictably, detection rates for all schools within each band declined as the Beaufort condition in- creased. Because of the large variability inherent in small sample sizes and spatial stratification of searching effort at the various Beaufort condi- tions, comparisons of rates of detecting trackline schools did not yield consistent trends. For exam- ple, within both bands, the trackline detection rate for Beaufort 2 conditions was larger than for Beaufort 1 conditions. In the coastal band Beau- fort 5 conditions had higher trackline detection rates than Beaufort 4 conditions and rates for Beaufort 4 were higher than rates for Beaufort 3 (Fig. 5). Searching effort for aerial data during good and poor sun conditions was also confounded with dis- tance from shore (Fig. 4) and thus with sea condi- tions. Most good sun conditions (78%) occurred in the coastal band, whereas 59% of all poor sun 425 FISHERY BULLETIN: VOL. 85, NO. 3 BEAUFORT 1 ao-N 20* 10° o« 10° 20° S Northern V Offshore Southern Offshore Area Coastal band BEAUFORT 4 30° N 20° 20°S Northern Offshore Area Southern Offshore Area Coastal band 160° 150° 140° 130° 120° 110° 100° 90° 80° 70°W 160° 150° 140° 130° 120° 110° 100° 90° 80° 70°W BEAUFORT 2 30° N Northern 20° ] V Offshore Area 10° 0'- 10° 20° S Southern Offshore oftshote. Area band^ BEAUFORT 5 30° N 20°- 0° 10° 20°S ■*• Northern V Offshore Area Southern Offshore Area Coastal band 160° 150° 140° 130° 120° 110° 100° 90° 80° 70° W 160° 150° 140° 130° 120° 110° 100° 90° 80° 70° W BEAUFORT 3 30° N 20° 10° 20° S Northern v Offshore Southern Offshore Area Coastal band Figure 3.— Tracklines surveyed by airplanes during 1977 and 1979 in the coastal and offshore bands stratified by Beaufort state. 160° 150° 140° 130° 120° 110° 100° 90° 80° 70°W conditions occurred in the offshore band. This was because the general searching pattern was to begin searching on the westward, outbound leg in the morning, and then to turn the aircraft near noon and reach shore in late afternoon or night. Thus the sun was directly overhead or in front of the plane in the offshore reaches of the track and behind the plane in the nearshore areas. Detection rates during good and poor sun condi- tions were higher in the coastal band than in the offshore band (Fig. 5), which was consistent with a hypothesized decreasing density gradient. Within the coastal band, detection rates during good sun conditions were greater than during poor sun conditions, but most of the poor sun data was gathered in the westward portion of the band (Fig. 4). In the offshore band, trackline detection rates during good and poor sun conditions were similar, but the rate during good sun conditions was based upon three sightings and only 8% of the effort. Finally, I compared data collected by the ob- 426 HOLT; DENSITY OF DOLPHIN SCHOOLS A. GOOD SUN 30°N 20° 20°S 100° 90° 80° 70° W B. POOR SUN 30°N 20' 10° 0° 10° 20°S Southern Offshore Area I 1 1 1 1 1 1 1 160° 150° 140° 130° 120° 110° 100° 90° 80° 70°W Figure 4.— Tracklines surveyed by airplanes during 1977 and 1979 during (A) good and (B) poor sun glare conditions in the coastal and offshore density bands. 427 FISHERY BULLETIN: VOL. 85, NO. 3 10.0 9 0 - 8.0 O 7.0 o o i 6 0 o o £ u ■2 5.0 LU I- < z 4.0 o o liJ 3.0 2.0 10 00 (6) (12) (12) (2) (14) (5) (6) I I All Schools ^J Trackllne Schools 1-5 = Beaufort State GS^Good Sun Conditions PS- Poor Sun Conditions ( ); Percent of Total Effort (27) (11) (19) (( 8) (27) (8) (38) 12 3 4 5 Coastal Band 12 3 4 5 GS PS GS PS Offshore Band Coa&tal Offshore Band Band Figure 5. — School detection rates for aerial data in the coastal and offshore density bands for sea state and sun glare cate- gories. server teams to determine relative effects upon the density estimates. Team 1 and Team 2 searched approximately equal lengths of track- line (46% and 54% of the effort, respectively). No difference in performance of the two teams was evident: their rates of detecting schools, both on and off the trackline, and their estimates of school densities were approximately equal (Fig. 6). Ship Data The rates of detecting dolphins were greater during calm seas than during rough seas for the ship surveys from 1979 through 1983 (Fig. 7). The detection rate of dolphins during calm seas was more than twice the rate during rough seas in both the inshore and offshore areas. The ratio of calm sea to rough sea detection rates was larger in the offshore area than in the inshore area. The offshore area was surveyed during rougher seas more than the inshore area (Fig. 8); seas were calm in the offshore area during only 17% of the effort as opposed to 35% for the inshore area surveys (Fig. 7). Dolphin density was lower off- shore as indicated by lower offshore detection rates than inshore rates during either calm or rough seas (Fig. 7). The inshore-to-offshore-area detection ratios were 1.5 during calm seas and 2.0 during rough seas. Sun glare had little effect on the shipboard esti- mates during either year because poor sun condi- tions occurred only during 6% of the 1982 and 8% ^■° r □ All Schools s o o o ai < K o m (- a 4.0 3.0 2.0 1.0 0.0 t 1 Trackline Schools f^ t).0 s ^ o o 4.0 - o r- \ t- 2.0 - (/) Z UJ a 1 o 1.0 — o I o CO TEAM 1 TEAM 2 TEAM 1 TEAM 2 Figure 6. — School detection rates and density estimates for observer teams during the 1979 aerial survey. ( )= % calm sea effort in area UJ < CE 3.0 UJ o < UJ (E < UJ cc o X M < OC o z (/) z t- < OC 2.0 - 0.0 Figure 7. — Ratio of 1979-83 shipboard school detection rates for different sea states (calm sea versus rough seal and area (in- shore versus offshore). Detection rates computed with perpen- dicular distance data truncated at 2.1 km. 428 HOLT: DENSITY OF DOLPHIN SCHOOLS A. CALM SEAS 30° N 20' 10' 0° - 10° - 20° S NorthernjOff Shore ^rea Southern Offshore Area 160° 150° 140° 130° 120° 110° 100° 90° 80' 70° W B. ROUGH SEAS 30°N 20' 10' 0' 10' 20°S ^Northern Off^hoFe . ^)^:y\- Southern Offshore Area 160° 150° 140° 130° 120° 110° 100° 90° 80° 70° W Figure 8. — Distribution of searching effort for the 1979-83 ship surveys during (A) calm and (B) rough conditions. 429 FISHERY BULLETIN: VOL. 85, NO. 3 of the 1983 surveys. However, rates of detecting schools during good sun conditions were larger than during poor conditions (Fig. 9) and no schools were detected on the trackline during poor conditions. (92) E o o o o i 4 u (0 UJ H 3 < CE Z 2 2 I- o UJ I- tu . Q ' (94) (6) (8) I I Good sun all schools |:::;'-| Poor sun all schools ^^Good sun trackline schools ( ) Percent effort 1982 1983 Figure 9. — School detection rates and relative density esti- mates during good and poor sun glare conditions for 1982 and 1983 ship data. Density Estimates Inshore Area Aerial observers during the 1977 and 1979 sur- veys searched 34,006 km and detected 152 dol- phin schools in the inshore area (Table 1). The estimate of school density using aerial data was 4.18 schools/1,000 km^ with a standard error of 0.902. From 1977 to 1983, shipboard observers searched 27,840 km in the inshore area and de- tected 297 schools (Table 2). Ship data yielded an estimate of density for the inshore area of 4.47 schools/1,000 km^ with a standard error of 0.514 (Table 1). This was only slightly larger than the aerial inshore estimate. Oflfehore Area Observers aboard both vessels surveyed 46,567 km in the offshore area and detected 192 schools (Table 2). The estimate of density was 2.04 schools/1,000 km^ with a standard error of 0.263 (Table 1). Total Area From 1977 to 1983, observers on both vessels searched 74,407 km in all areas and detected 489 schools (Table 2). The density estimate for all shipboard data was 2.95 schools/1,000 km^ with a standard error of 0.253 (Table 1). The estimate of density using the aerial inshore estimate and the Table 2.— School detection rates for 1977-83 ship data and for 1979-83 ship data stratified by sea state category in the inshore, offshore and total areas. Data were truncated at 2.13 km perpendicular distance. Detection Distance Percent Number Percent rate SE Area/data searched (km) schools schools (schools/ (detection Number source (km) searched detected detected 1 ,000 km) rate) days searched Inshore area 77-83 all data 27,840 100.0 297 lOOO 10.67 0.82 173 79-83 calm seas 8,502 35.3 144 53.9 16.94 1.52 89 79-83 rough seas 15,609 64.7 123 46.1 7,88 0.92 124 Offshore area 77-83 all data 46,567 100.0 192 100.0 4.12 0.50 251 79-83 calm seas 4,129 17.1 44 36.7 10.66 2.30 58 79-83 rough seas 20,015 82.9 76 63.3 3.80 0.56 134 Total area 77-83 all data 74,407 100.0 489 100.0 6.57 0.47 417 79-83 calm seas 12,632 26.2 188 48.6 14.88 1.29 146 79-83 rough seas 35,624 73.8 199 51.4 5.59 0.54 256 430 HOLT: DENSITY OF DOLPHIN SCHOOLS ship offshore estimate was 2.71 schools/1,000 km^ with a standard error of 0.334. DISCUSSION Onshore-Offshore Density Gradients The onshore-to-offshore density gradient de- creased based on aerial data in the inshore area and comparison of inshore and offshore density estimates. Offshore density estimates were only about one-half the inshore estimates (Table 1). Although sea state and sun glare conditions were confounded with distance from shore, compari- sons of detection rates in the two inshore density bands for data stratified by Beaufort state or sun conditions indicated lower rates in the outer band (Fig. 5). Fit of Fourier Series Model Burnham et al. (1980) provided criteria for se- lecting the appropriate number of terms in the FS model. However, these criteria were not satisfac- tory for use with the aerial and ship perpendicu- lar distance distributions, which had pronounced modes at the origin. Instead, I selected models which had the fewest terms but provided a good fit near the origin. This resulted in models with large numbers of terms. However, to the degree that the modes are representative of school den- sity, my estimates of densities will be unbiased. Alternate statistical models need development which can fit data which lack a shoulder near the origin (i.e., data with pronounced modes at the origin). Buckland (1985) investigated several models but concluded that reliable estimation is not possible unless a shoulder exists. Line Transect Assumptions Aerial Data Confounding of aerial sea state and sun condi- tion data with distance from shore made it impos- sible to test the assumption that all trackline schools were detected during all viewing condi- tions. If viewing conditions had been homoge- neous throughout the area, the density estimate calculated for calm sea and good sun conditions (12.64 schools/1,000 km") could be used for the inshore area (Table 1). This estimate is over 7 times the rough sea and poor sun estimate (1.78 schools/1,000 km^). However, the calm seas and good sun condition effort occurred mostly in the northern nearshore region of the inshore area (Fig. 3, 4) where density may be high. Consequently, Holt (fn. 8) conducted an aerial experiment in a relatively small area to test sea state and sun effects upon LT density estimates. The results indicated that sun glare adversely affected estimates of school density. The density estimate was 39% larger during good sun condi- tions than during poor conditions. Although den- sity estimates were larger for calm sea data than for rough sea data, the differences were not signif- icant. The aerial experimental data (Holt fn. 8) may be used to estimate maximum bias for sun and sea state effects. The adjusted density estimate (D^ ) is Da=^^D,P,j 1=1 7=1 where D, P„ = D', Density estimate in survey area during ith sea state and jth sun condition, Proportion of effort in survey area with ith. sea state andj'th sun con- dition. Experimental density estimate during ;th sea state and jth sun condition determined from Holt (fn. 8). In addition, i equal 1 denotes calm sea states and i equal 2 denotes rough sea states, and j equal 1 denotes good sun conditions and J equal 2 denotes poor sun conditions. An estimate of the sampling variance (Var(D^ )) using the Taylor approxima- tion method is 2 2 Vdr(D^ ) = 2 S P'/ 1=1 j=i Vdr(D'u) + + Vdr{D\ The adjusted inshore density estimate is 4.51 schools/1,000 km^ with a standard error of 1.107. This is an 8% increase over the unadjusted esti- 431 FISHERY BULLETIN: VOL. 85, NO. 3 mate (Table 1). The adjusted combined estimate for the entire ETP was 2.81 schools/1,000 km^ with a standard error of 0.152, a 4% increase from the unadjusted estimate. Using the experimental results to adjust aerial estimates for sun glare (and possibly sea state), effects may be suspect because of differences in procedures followed and observational conditions encountered in the experiment and the surveys: 1) The wings on the aircraft used during the ex- periment were attached on the lower part of the fuselage, whereas wings on the 1977 and 1979 aircraft were attached to the upper part of the craft which allowed better lateral observation. 2) Procedures used to adjust for presence of sun glare during the surveys and the experiment dif- fered. Observers during the surveys were in- structed to stop searching if they believed condi- tions prevented their detecting trackline schools, but observers in the experiment searched during all conditions. 3) More rough seas were encoun- tered during the surveys (74%) than in the exper- iment (62%). Also, more (46% as compared to 15%) of the surveys' total effort occurred at ex- treme Beaufort 4 and 5 conditions. Because of these uncertainties, I used the unadjusted density estimate to determine school densities. Comparisons of the 1979 aerial observer teams' estimates did not indicate observers of either team missed dolphin schools on the trackline but both teams may have been equally affected by searching conditions. These results were consis- tent with results of the aerial experiment (Holt fn. 8) where comparisons of observer teams' per- formance also indicated no significant differ- ences. Ship Data The density estimates calculated from calm sea data were larger than estimates calculated from rough sea data (Table 1). The difference was prob- ably not due to missed trackline schools during rough seas. Schools on the trackline would proba- bly be detected as the ship approached unless the schools avoided the approaching ship. In a ship- helicopter experiment Hewitt (1985) investigated the reaction of dolphins to survey vessels and found that dolphin schools only occasionally react to the approach of a vessel before they are de- tected by shipboard observers (1 of 12 schools). The differences between calm and rough sea estimates may have resulted from actual differ- ences in densities in areas surveyed during calm and rough sea states (Fig. 8). Another possibility is that estimation errors resulted from observers detecting schools at greater radial distances dur- ing calm conditions (mean radial distance was 4.16 km) than during rough conditions (mean ra- dial distance was 3.55 km). Estimation of sight- ing angles and distances of schools at greater dis- tances from the ship may have been less accurate and may have increased the probability of schools being erroneously recorded near or on the track- line. Although sun glare was not shown to affect the shipboard density estimates, Cologne and Holt (fn. 7) found that shipboard observers tended to avoid searching areas with sun glare. However, because of the relatively slow speed of the ship and the dolphins and because sun glare at any specific time is usually concentrated in a small region of the observers' field of view, all regions may be observed without glare. The occurrence of errors in angle and distance estimations may have positively biased shipboard estimates. An inordinate proportion of dolphin schools (25% of all schools) was recorded as being on the trackline. Smearing the perpendicular dis- tance distributions helped alleviate the bias but may not have eliminated it. Comparison of Aerial and Ship Estimates The estimates of dolphin densities in the in- shore and the total areas using only ship data were slightly larger than estimates which used aerial inshore data (Table 1). This is logical be- cause ship surveys were designed to overlap with aerial coverage in the inshore area and to provide systematic coverage of the offshore area. There- fore, they spent disproportionately more of their effort in the inshore area compared to its relative size and, within the inshore area, they spent dis- proportionately more effort in the northern nearshore region (Fig. 1), which has relatively high dolphin density. Although the inshore area represented 31% of the total area, 37% of the ship's effort was in the inshore area. In addition, 61% of the inshore effort was in the northern in- shore region which represented approximately 44% of the inshore area. During the aerial sur- veys a systematic survey of the inshore area was conducted. Therefore, the best estimates of densi- ties in the inshore and total areas are estimates calculated using the unadjusted aerial inshore data. 432 HOLT: DENSITY OF DOLPHIN SCHOOLS Comparisons with Previous Density Estimates Density of ETP dolphin stocks have been esti- mated previously (SWFC 1976' '; Holt and Powers 1982). The methods I used to calculate estimates were similar to those used by Holt and Powers. Therefore, differences that they noted be- tween their assessment in 1979 and the SWFC 1976 assessment are also applicable to compari- sons between the SWFC 1976 assessment and this study. My estimates differ from the 1979 esti- mates in that mine include 1) schools where either the observers' "best" or "lowest" estimate of mean school size was more than 14 animals (the 1979 assessment included only schools with "best" estimates), 2) use of the 1977 aerial data in the inshore den- sity estimate, 3) ship data collected in 1977, 1979, 1980, 1982, and 1983 (the 1979 assessment included only 1979 ship data), 4) investigation of aerial and ship data for effects of sun, sea state, and observer performance, 5) application of LT methods to ship data to calcu- late density estimates. Density estimates calculated in this study were similar to those presented in the 1979 assessment (Holt and Powers 1982). My inshore and offshore estimates were 4.18 and 2.04 schools/1,000 km^, respectively, with standard errors of 0.902 and 0.263. Holt and Power's estimates were 3.51 and 1.89 schools/1,000 km^, respectively, with stand- ard errors of 0.590 and 0.766. CONCLUSIONS LT methods were used on 1977 and 1979 aerial survey data to estimate dolphin density in the inshore area at 4.18 schools/1,000 km^. LT meth- ods applied to 1977-83 ship data yielded an esti- mate of offshore dolphin density of 2.04 schools/ 1,000 km^. By weighting aerial inshore and ship offshore data by the respective size of the two areas, the total dolphin density was estimated at 2.71 schools/1,000 km^. 1 ISWFC (Southwest Fisheries Center). 1976. Report of the workshop on stock assessment of porpoises involved in the east- ern tropical Pacific vellowfin tuna fishery. Southwest Fish. Cent. Adm. Rep. No" LJ-76-29, 60 p. National Marine Fish- eries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. I investigated differences among densities at different visibility conditions for aerial data, but results were inconclusive owing to confounding of the factors with density gradient (area from shore). Adjusting the data for sea state and sun conditions increased the inshore aerial density estimate 8% and the total density estimate by 4%. ACKNOWLEDGMENTS I thank P. Hammond for providing the data smearing program. B. Vorndam provided exten- sive computer assistance and J. Cologne provided statistical consulting. K. Raymond and R. Allen drew the figures. I thank the many observers who participated in the surveys. J. Barlow, S. Buck- land, J. Michalski, W. Parks, S. Reilly, and S. Sexton provided many constructive comments for improvement of the paper. Members of SWFC Pre-Sops Panel C, including G. Broadhead, K. Burnham, D. Chapman, and P. Hammond, pro- vided many useful technical suggestions (i.e., re- view of procedures used to fit the Fourier series model to the data sets). LITERATURE CITED AU, D., AND W. Perryman. 1982. Movement and speed of dolphin schools responding to an approaching ship. Fish. Bull., U.S. 80:371-379. BOWDITCH, N. 1966. American practical navigator, an epitome of navi- gation. U.S. Naval Oceanogr. Off. H.O. Publ. No. 9, 1524 p. BUCKLAND, S. T. 1985. Perpendicular distance models for line transect sampling. Biometrics 41, 177-195. BuRNHAM, K , D Anderson, and J Laake 1979. Robust estimation from line transect data. J. Wildl. Manage. 43:992-996. 1980. Estimation of density from line transect sampling of biological populations. J. Wildl. Manage. Monogr. No. 72, 202 p. BUTTERWORTH, D. 1982. On the functional form used for g(y) for Minke whale sightings, and bias in its estimation due to meas- urement inaccuracies. Rep. Int. Whaling Comm. 32:883-888. Crain. B., K. Burnham, D Anderson, and J Laake. 1979. A Fourier series estimator of population density for line transect sampling. Utah State Univ. Press, Logan, 25 p. Hammond, P. 1984. An investigation into the effects of different tech- niques of smearing the IWC/IDCR Minke whale sight- ings data and of the use of different models to estimate density of schools. Rep. Int. Whaling Comm. 34:301- 308. Hewitt, R 1985. Reaction of dolphins to a survey vessel: effects on 433 FISHERY BULLETIN: VOL. 85, NO. 3 census data. Fish. Bull., U.S. 83:187-194. HOLT, R . AND J POWERS. Holt, R. 1982. Abundance estimation ofdolphin stocks involved in 1983a. Report of porpoise experiment testing detection of the eastern tropical Pacific yellowfin tuna fishery deter- on-track schools (PET DOTS), March 7-April 5, mined from aerial and ship surveys to 1979. U.S. Dep. 1981. U.S. Dep. Commer., NOAA-TM-NMFS-SWFC- Commer., NOAA-TM-NMFS-SWFC-23, 95 p. 27, 80 p. Seber, G 1983b. Report of eastern tropical Pacific research vessel 1973. The estimation of animal abundance. Hafner marine mammal survey. May 15-August 3, 1982. U.S. Publ. Co., Inc., N.Y., 506 p. Dep. Commer., NOAA-TM-NMFS-SWFC-29, 151 p. 434 RESEARCH VESSEL SURVEY DESIGN FOR MONITORING DOLPHIN ABUNDANCE IN THE EASTERN TROPICAL PACIFIC Rennie S. Holt.' Tim Gerrodette/ and John B. Cologne' ABSTRACT During 1986 the National Marine Fisheries Service began conducting long-term research ship sur- veys to determine status of spotted dolphin, Stenella attenuata, stocks in the eastern tropical Pacific. This is the main dolphin species taken incidentally by the yellowfin tuna, Thunnus albacares, purse seine fishery. We use research vessel survey data collected from 1977 to 1983 to investigate the annual changes in spotted dolphin population size that could be detected given various levels of research vessel survey effort during specified time periods for several levels of statistical error. We find that two research vessels each operating for 120 days per year for 5 years (six surveys) could detect a lO'J annual rate of decrease in dolphin abundance (a total 41% decrease over 5 years) with alpha and beta error levels of lO'^'f. Adding a third vessel would provide better coverage of the dolphins' range, but would allow only a slightly lower rate of decrease to be detected (an 11% annual rate, for a total decrea-se of 44%). These numbers point out the difficulty of detecting even major changes in spotted dolphin population size with present survey methods. Alternatives are discussed, but all either cost more money, require a longer time to detect a decline, or accept higher levels of statistical error. The National Marine Fisheries Service (NMFS) has the responsibility of determining the status of (dolphin stocks which are taken incicientally by the yellowfin tuna, Thunnus albacares, purse seine fishery in the eastern tropical Pacific (ETP) (Richey 1976'*). The status of spotted dolphins, Stenella attenuata , is of special concern since it is the major species taken by the fishery (Smith 1979^). Of the spotted dolphins, the northern off- shore stock is of more concern since it has been fished more frequently than the southern offshore stock. The spinner dolphin, S. longirostris , and the common dolphin, Delphinus delphis, are also taken. In addition, the striped dolphin, S. coeruleoalba, and the Eraser's dolphin, Lageno- delphis hosei , are occasionally caught but are dif- ficult to distinguish from the other three species ^Southwest Fisheries Center La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. ^Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 96822-2396. ■^Department of Biostatistics, SC-32, University of Washing- ton, Seattle, WA 9819.5. ^Richey, C. R. 1976. Memorandum of opinion. CA NO. 74-1465 and CA NO. 75-0227 U.S. District Court, District of Columbia, May 11, 1976. SSmith, T. D. 1979. Report of the status of the porpoise stock workshop (August 27-31, 1979, La Jolla, Califor- nia). Southwest Fish. Cent. Adm. Rep. No. LJ-79-41, 120 p. SWFC La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla. CA 92038 at a distance (Holt and Powers 1982). These 5 species are herein termed target species. The NMFS conducted assessments of popula- tion status in 1976 (SWFC 1976^) and again in 1979 (Smith fn. 5) based on estimates of absolute stock abundance. The validity of the absolute es- timates depended on several assumptions being met. Unfortunately, some assumptions, such as not allowing systematic errors in data recording or the assumption that dolphin schools do not move prior to being detected by shipboard observ- ers, may not have been met and thus the assess- ments were not entirely satisfactory. An alterna- tive approach for assessing stock status, therefore, is to use relative population estimates to detect trends in stock sizes over a long time period. Relative estimates can provide an assess- ment of stock condition as long as the biases in the abundance estimates are consistent over the sampling period. Therefore, the NMFS is presently considering using annual estimates of population abundance as relative estimates to de- tect declines in population size of spotted dolphins during a sampling period of at least 5 years. 6SWFC (Southwest Fisheries Center). 1976. Report of the workshop on stock assessment of porpoises involved in the east- ern tropical Pacific yellowfin tuna fishery. Southwest Fish. Cent. Adm. Rep. No. LJ-76-29, 60 p. SWFC La Jolla Labora- tory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. Manuscript accepted February 1987. FISHERY BULLETIN; VOL 85, NO 3. 1987. 435 FISHERY BULLETIN: VOL. 85, NO. 3 In this paper, we investigate the annual changes in the size of spotted dolphin populations that can be detected given various levels of re- search vessel survey effort within specified time periods. We investigate how many research ves- sels, assuming 120 days searching per vessel per year, would be required to survey the physical area inhabited by the major stocks. We also inves- tigate how many vessels would be required to de- tect various levels of population declines in spot- ted dolphins during 5 years or, given fixed number of vessels, how many years of survey ef- fort it would take to detect various population declines or, given fixed number of vessels for fixed number of years, the probability of detecting a decline (i.e., the power). We use historical data and current abundance techniques to predict variability of data which will be collected during the sampling period. AREA INHABITED AND DATA SOURCES For our analyses, the study area included the area described by Au et al. (1979)^ as being inhab- ited by the target species (Fig. 1). The area north of lat. 20°N was excluded because spotted dol- phins do not usually occur there. We partitioned the study area into four strata: the inside, middle, and west strata, which are located north of lat. 1°S, and a south stratum. The three northern strata were collectively termed the north area and all strata were termed the total area. In addi- tion, a calibration area was defined as including part of the inside stratum (Fig. 1). Data used in our analyses were collected from 1977 through 1983 by scientific observers aboard the NOAA ships David Starr Jordan and Townsend Cromwell. Survey coverage from the two ships for all years combined was thorough (Fig. 1). Data collected for each school included estimates of dolphin school size, species composi- tion, and line transect observations, which we used to calculate density estimates. SURVEY COVERAGE We investigated the physical coverage of the area that is possible when using 1, 2, or 3 ships for ■^Au, D., W. L. Perryman, and W. Perrin. 1979. Dolphin distribution and the relationship to environmental features in the eastern tropical Pacific. Southwest Fish. Cent. Adm. Rep. No. U-79-43, 59 p. SWFC La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. 130 H 120* H 110" LONGITUDE 70' H Figure 1. — Research vessel tracklines in each stratum during 1977 through 1983. 436 HOLT ET AL : MONITORING DOLPHIN ABUNDANCE 120 days each by plotting hypothetical tracklines. Approximately 370 km (200 nautical miles) of trackline could be covered in each survey day; with searching restricted to daylight hours, only about one-half of this distance would be searched. Approximately 40,700 km of trackline could be covered by each ship with less than 50^^ of this distance searched during daylight hours. Each ship's searching distance was allocated to each stratum by the square root of school density in the stratum. Effort of each ship was partitioned into 30-d segments between ports to meet logistical constraints of the vessels. We found that thor- ough coverage of the entire area was provided when three ships were used, two ships provided adequate coverage, and one ship provided very poor coverage with tracklines separated by large distances (Fig. 2). DETECTION OF CHANGES IN POPULATION SIZE Survey Design The relationship among the number of samples, the rate of change, the precision of the population estimate, and the levels of alpha (type I) and beta (type II) statistical errors for several models of change and sample variability was investigated by Gerrodette (in press). We assumed that popu- lation size would change exponentially (constant rate per year). From Gerrodette's equation 15, using slightly different notation, a(a + l)2(a + 2)[ln(l - r)]2 12(Z„ + Zp)2 ^\n (=0 _cv_ (1 -rY + 1 (1) where a - number of years in the survey period, r = annual rate of decrease, Za = percentile of standardized normal curve for one-tailed Type I error, Zp = percentile of standardized normal curve for Type II error, and CVq = coefficient of variation of the population estimate at the present population size. In this formulation, r is a positive number, and, since the first survey occurs at time 0, the total number of samples (i.e., number of annual sur- veys) is a + 1. Note that the null hypothesis is one-sided, namely, that spotted dolphin abun- dance is decreasing. In addition to the annual rate of decrease (r), the total population decrease which would occur over the entire survey period was calculated as Total decrease = [1 - (1 - r)°]. The survey design to detect changes in dolphin abundance was investigated in three ways. Using Equation (1), we computed 1) the minimum num- ber of years (a), given one to three ships per year and 120 searching days per ship per year, re- quired to detect various annual decreases in spot- ted dolphin abundance; 2) the minimum propor- tional annual change (r) that could be detected in 5 years given one to three ships per year at vari- ous levels of alpha and beta; and 3) power (1 - p) or the probability of detecting various decreases in population size in 5 years, given one to three ships per year. To use Equation (1), the relationship of CV (N), the coefficient of variation of the population esti- mate, and n , the number of schools detected must be determined. In addition, the rate per day at which dolphin schools are expected to be encoun- tered must be known. We used the 1977-83 re- search vessel data to investigate these factors as- suming these data would be representative of data that we will obtain during the proposed sam- pling period of 1986-91. Abundance Estimation Relative estimates of population abundance of spotted dolphins in the north and total areas were calculated using two methods, methods A and B. In method A, density and mean school size esti- mates were calculated in each stratum and abun- dance was determined (Holt and Powers 1982) as A^ Pt^D kStkPk-^k (2) k=i In method B, density and mean school size esti- mates were calculated for data pooled for the en- tire area (north area or total area) and abundance was determined as 437 FISHERY BULLETIN: VOL. 85, NO. 3 IB H ISO H lao II 110" a LONSITUre Figure 2. — Plot of hypothetical tracklines expected from use of one (A), two (B), or three (C) ships for 120 days each. N=P,DS^PkAk. (3) k=i where m = number of strata (3 for the north area and 4 for the total area), k = 1, 2, 3, or 4 denotes the inside, middle, west, or south stratum, respectively, N = estimated number of spotted dolphins in the survey area, D - density estimate of number of schools of all dolphin species in the survey area (schools/1,000 km^), Dk = density estimate of number of schools of all dolphin species in the ^th stratum (schools/1,000 km^), S = mean school size estimate for target species in the survey area (number of animals), Stk = mean school size estimate for target species in the /eth stratum (number of animals), Pi = proportion of all dolphins that were target species in the survey area, Pk = proportion of spotted dolphins in the target schools in the /eth stratum, and Ak = area inhabited by all dolphins in the kth stratum. The variance of N for Equation (2) was estimated using Taylor series expansion (Seber 1973) as Var (N) = 2 [iStk Pt Pk ^^ )^ Var (Dk ) k=i + (DkPtPkAk)^'^kr(Sik) + (D,S,,P,A,)2Var(A) + (DkSfkPtAk)^ War iPk)] . (4) The variance of iV in Equation (3) was determined using Equation (4), but density and school size estimates that were calculated for the entire area 438 HOLT ET AL.: MONITORING DOLPHIN ABUNDANCE so H 70 tt IX M 110 H LONOITUDE were substituted for the respective stratified esti- mates. Specific formulae to estimate variables and as- sociated theoretical variances in Equations (2) through (4) are from Burnham et al. (1980), Holt (1985,8 in press) and Barlow and Holt (1986). Variances for estimates of school sizes and school densities were calculated using jackknife tech- niques (Miller 1974). Since serial correlation among sampling units ^Holt, R. S. 1985. Estimates of abundance of dolphin stoclts taken incidentally in the eastern tropical Pacific yel- lowfin tuna fishery. Southwest Fish. Cent. Adm. Rep. No. LJ- 85-20, 32 p. SWFC La Jolla Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038. (days of effort) will yield biased estimates of standard errors using the jackknife method, we analyzed serial correlation of dolphin school de- tection rates among various combinations of suc- cessive days of effort. Analyses indicated that cor- relation was significant among successive single days but was not significant for periods of 2 or more days. Therefore, the data were grouped by 2-d increments for the jackknife analyses. Estimates of spotted dolphin population abun- dance and values used in Equations (2) and (3) to calculate the estimates are presented in Table 1. CV (N)s were smaller for estimates calculated using method B than for estimates using method A. 439 FISHERY BULLETIN: VOL. 85, NO. 3 Table 1 . — School density of all dolphin schools, proportion of all schools which were target schools, mean school size of target schools, proportion of target animals which were spotted dolphins, area of each stratum, abundance and K values for spotted dolphins. SE and CV denote standard error and coefficient of variation, respectively. Methods A and B refer to different ways of pooling data on school size and density (see text). Stratum Area Variable Inside Middle West South North Total School density (D) (Schools/1 ,000 km2) 5.33 3.42 0.82 1.93 3.20 3.03 SE(D) 0.87 1.13 0.30 0.39 0.54 0.51 CV(D) 16.3 33.1 37.2 20.2 17.0 16.8 Prop, target (P,)1 — — — — 0.775 0.775 Mean school size (S,) (Number animals) 108.59 113.89 121.06 157.65 111.62 118.21 SE (S,) 9.82 11.24 23.28 29.84 7.44 7.92 CV (S,) 8.6 9.9 19.2 18.9 6.7 6.7 Area (km2 . 106) 4.602 3.764 5.298 4,359 13.664 18.024 Prop, spotted {Pi,)^ 0.38 0.51 0.51 0.26 — — SE (P,)2 0.039 0.048 0.048 0.085 — — Abundance and K Values Method A N (Animals • 106) 1.571 1.839 SE{N) 0.283 0.294 CV(/V) 0.18 0.16 Sample size (n) 507 602 K 4.05 3.93 Method B N (Animals • 106) 1.761 2.081 SE(N) 0.240 0.250 CV (N) 0.14 0.12 Sample size (n) 507 602 K 3.06 2.94 'Source Holt (in press). 2Source Barlow and Holt (1986). Relationship Between Var (N} and Number of Schools Detected In order to minimize the number of ^ears re- quired to detect a specific trend, Var (N ) should be as small as possible (Gerrodette in press). Var (A^) depends on the variance of the estimates of school size, school density, and proportions of the various dolphin species, as shown in Equation (4). Each of the variances of these estimates, in turn, depends on n, the number of sighted schools. Therefore, the dependence of Var (N) on n must be known to calculate the number of sightings needed to attain a given level of precision (Var (A^)). We investigated the dependence of each of the individual variance terms on n. size estimates, its variance is Var iStk ) — Var {Stk )/n where Var iSfk ) is the variance of school size. The Var (P^) = Pfil - Pt)ln where Pf is the true proportion of target schools among all dol- phins. Var (Sfk) and P^(l - Pf) are both constant with respect to n, so Var (Sf) = Odin) and Var {Pf) = Odin), where Odin) means "of the same order as II n" and implies that as II n approaches zero, the variance approaches zero at the same rate. Similarly, Var (Pik), which is also a pro- portion, is equal to Odin). Dependence of Var (D ) on w The Var (D), based on replicate tracklines (Burnham et al. 1980), is Dependence of Var (5^^), Var (P,), and Var (P^ ) on m Because Stk is the mean of n individual school 440 Var (D) = i)2 Var(n) ^ Var[/'(0)] n2 (5) [/■(0)]2 where n is the number of sightings and/"(0) is the HOLT ET AL.: MONITORING DOLPHIN ABUNDANCE estimate of the probability density function of perpendicular distances extrapolated to the trackline. First, R i?2 ("' ~"^^ Var(n) i=l R where R is the number of replicate lines of equal length (/). For R of moderate size, R =(R - 1). Thus R Var in) - 2 ^"' " "^^ " ^^^^• 1=1 This is because Var (n) is the sum of the variances of i? independent values (n,, i = 1, 2, ..., R) each having the same expected variance. But R = n/Eirii), the total number of sightings divided by the expected number of sightings for a line of length /. Thus, R = Oin), and Var in) 0(n) Oa/n). (6) n' Second, fiO) was estimated using a Fourier series (FS) model (Burnham et al. 1980); therefore, Var[/(0)] = 2 2 Cov (d,,d^) ^ = 1, 2, 3, ..., and k >j > 1 where the a's are the coefficients in the series ak nw n V^ jk-nx, cos 1=1 w = 0(1) with jc, equal the perpendicular distance to the ith sighting and w equal the truncation point for the perpendicular distance. Therefore, we only need to know the dependence of cov(a,, Uk ) on n. If n is much larger than one, in - 1) = n and Gov (dj, dk ) = n (dk+j + dk-j) - dj dk = oain). Since /"(O) estimates a quantity which is constant with respect to n, Var[/(0)] [/(0)]2 Oilln). (7) Combining Equations (6) and (7) with Equation (5), [CV (Z))]2 = Oain). This confirms discussions presented by Burnham et al. (1980). In addition to investigating the theoretical de- pendence of [CV (D )]2 on n, we tested its empiri- cal dependence on n using the research vessel data which included 479 days of survey effort. Data were truncated at 3.70 km perpendicular distance from the ship. Paired days of shipboard searching effort were randomly selected using a uniform random number generator until the number of associated sightings (n) equaled or ex- ceeded a previously selected sample size. Sample sizes selected were 20, 30, 40, 50, 60, 80, 100, 200, 500, and 1,000. The resultant perpendicular dis- tance distributions were smeared (a data smooth- ing technique described by Butterworth 1982, Hammond 1984, and Holt fn.8), and density, vari- ance, and coefficient of variation estimates were calculated for each data set. The simulation was completed three times for each value of n . The relationship between CV (D) and l/Vn (Fig. 3) was linear (/^lack-of-fit = 0.83; P = 0.59) with intercept not significantly greater than zero (t = 1.56; P > 0.10). This confirms the analytical result above, that CV (D) = 0(1/Vn ); however, as n increased, the probability of randomly selecting data from each of the 240 pairs of days (479 sur- vey days) multiple times increased which may 0 7 r 2 06 z 2 0.5 = 04 > Z 03 UJ o iZ 0 2 u. UJ O O 0,1 0.0 -1—1 L_i I ' I i I I 0,00 0.05 0.10 0 15 1/Vn 0.20 0,25 Figure 3. — Comparison of number of dolphin sightings (l/Vn) and precision of the population estimate (CV(D)). 441 FISHERY BULLETIN: VOL. 85, NO. 3 have biased CV (D) if the distribution of sightings for the days were biased due to the effects of sea- son or area. If we had included more large sam- ples in our simulation, the linear relationship may not have been evident. Calculation of K Values Because all terms used to calculate Var {N) equal 0(l/n) and Var {N) is a linear sum of the terms, Var (N) = Oil/n) or CV (N) = Oil/Vn). Therefore, the relationship CY(N) ^KlVii (8) can be used to determine the change in CV (N) for various values of n, where /C is a constant. This relationship is true if the number of schools sighted is proportional to population size. This seems to be a reasonable assumption, although a more complicated relationship between density and school size, based on dolphin social structure and its interaction with the fishery process, is possible. K values for spotted dolphins in the north and total areas were calculated for methods A and B using the 1977-83 data (Table 1). These K values were then used to determine CV iN )s for specified values of n which would be expected assuming from one to three annual ship surveys. Detection Rates The number of expected sightings with use of one to three ships was calculated by computing detection rates as the average number of dolphin sightings per searching day. A day's searching effort generally consisted of searching from sunrise to sundown; therefore, we assumed most survey days covered approximately the same trackline distance. However, distance searched may vary inversely with rates of detecting dol- phin schools because effort is halted so that ob- servers can identify schools and make school size estimates. The number of survey days, and hence number of ships, required to obtain a specified CV {N) was determined by dividing the number of required sightings by the rate of detecting schools. Detection rates were caculated separately for data from the Jordan cruise and from the Cromwell cruise because of the wide disparity in detection rates of dolphins from the two vessels when operating simultaneously in the calibration area (Table 2). The Jordan has a much better platform from which to detect dolphins because its observation station was higher relative to the water and because the Jordan rode much smoother than the Cromwell. Pooled Jordan and Cromwell detection rates were calculated by stan- dardizing the Cromwell rates to Jordan rates (Table 2) as DR = RjTj + R(.T(.C (9) where DR = pooled standardized detection rate for all dolphin schools, Rj = dolphin schools detected per day by observers aboard the Jordan , Re = dolphin schools detected per day by observers aboard the Cromwell, T. = days searched aboard the Jordan , Table 2. — Detection rates of all dolphin schools from the Jordan and Cromwell in the calibration area and pooled standardized detection rates for both vessels combined calculated in each stratum. Standardized detection rates were calculated using the ratio of Jordan to Cromwell detection rates in the calibration area. Jordan (J) Cromwell (C) Stratum/area Number of Days schools searched (n) (D) nID Number of Days schools searched (n) (D) n/D J/C ratio of detection rates Calibration area 102 28 3.643 49 31 1.581 2.304 Pooled standardized Inside 237 106 2.24 87 56 1.55 n/D 1. 2.70 2. Middle 108 80 1.35 18 22 0.82 1.47 3. West 43 54 0.80 14 56 025 0.69 4. South 91 60 1.52 4 5 0.80 1.54 North area 388 226 1.72 119 128 0.93 1.87 (Pooled strata 1-3) Total area 479 282 1.70 123 132 0.93 1.84 (Pooled strata 1-4) 442 HOLT ET AL.: MONITORING DOLPHIN ABUNDANCE Tc = days searched aboard the Cromwell , and C = ratio of schools detected per day by observers aboard the Jordan in the calibration area during 1979 to schools detected per day by observ- ers aboard the Cromwell in the cali- bration area during 1979. The percent of searching days when one to three ships were used was allocated to each stra- tum (Table 3) by the square root of school density. The number of schools which would be expected to be detected based on the standardized detection rates then was calculated (Table 4). Table 3. — Percent of searching days allocated by square root of density to each stratum in the north and total areas. Stratum North area Total area Inside Middle West South 45.6 36.5 17.9 35.8 28.7 14.0 21.5 Table 4. — Number of days searched and number of schools de- tected per year of effort with use of 1 , 2, or 3 ships allocated to the various strata by square root of density. North Total Stratum Number Number days schools Number days Number schools 1 ship = 120 days Inside Middle West South Total 2 ships = Inside Middle West South 240 days Total 3 ships = Inside Middle West South Total 360 days 55 44 21 120 110 88 42 240 165 132 63 360 149 65 14 228 298 130 28 456 447 195 42 684 43 34 17 26 120 86 68 34 52 240 129 102 51 78 360 116 50 12 40 218 232 100 24 80 436 348 150 36 120 654 which uses pooled density and school size esti- mates, than when using method A, which uses estimates calculated for each stratum (Table 5). This is because large variances associated with the method A population size estimates occur due to small sample sizes in some strata. Therefore, method B was used in subsequent calculations. The same number of years is required to detect a specific trend if the north or total areas are surveyed (Table 5). This result is true only if the 1977-83 data, which contain small sample sizes in the south stratum, are representative of future data. However, the northern offshore spotted dol- phin stock occurs only in the north area and elim- ination of the south stratum will ensure better coverage of this north area, especially in the west stratum where sample sizes are minimal for ap- plying the Fourier series model (Table 4). There- fore, subsequent calculations were made only for the north area. Annual population estimates for the northern stock would be biased only if sub- stantial variation in the amount of dolphin mi- gration between the north area and south stra- tum occurred during survey years. Table 5. — Number of years required to detect an annual 5% de- crease in spotted dolphin population size using 1 , 2, or 3 ships and 2 different methods of pooling data. Method A utilized Equation (2) in text while method B utilized Equation (3). Alpha and beta levels equal 0.05, and effort was allocated to the various strata by square root of density. Number of schools expected to be detected each year determined using detection rates from Equation (9). K deter- mined using Equation (8). CV (N) denotes coefficient of variation of population abundance estimate. Stratum Number ships Number schools K CV (A/) Years required North area Method A Method B Total area Method A Method B 1 2 3 1 2 3 1 2 3 1 2 3 228 456 684 228 456 684 218 436 654 218 436 654 4.05 3.06 3.93 2.94 0.27 17 0.19 12 0.15 11 0.20 13 0.14 10 0.12 9 0.27 17 0.19 12 0.15 11 0.20 13 0.14 10 0.11 8 RESULTS For either the north or total area, the same decrease in spotted dolphin populations can be detected 2 to 4 years earlier using method B, At the 5% error level, only rates of change of 11% per year or greater can be detected in a 5-yr survey period, even using three ships per year (Table 6). This is a rather high rate of decrease. 443 FISHERY BULLETIN: VOL. 85, NO. 3 Table 6. — Minimum rates of annual decrease and minimum total decreases in spotted dol- phin population size which could be detected in 5 years under different conditions. Changes were calculated for several alpha and beta lev- els, with a one-tailed test, using 1, 2, and 3 ships, for CV {N) determined using jackknife formulae, and data in the north area pooled over all strata (method B). Number ships CV(A/) Decrease per year Total decrease a = p = 0.05 1 2 3 0.20 0.14 0.12 0.19 0.13 0.11 0.65 0.50 0.44 a= P = 0.10 1 2 3 0.20 0.14 0.12 0.14 0.10 0.08 0.53 0.41 0.34 a = p = 0.20 1 2 3 0.20 0.14 0.12 0.09 0.06 0.05 0.38 0.27 0.23 and would lead to a 44% reduction in population size over the 5-yr period. If two or one ship is used, however, the minimum detectable rates of de- crease are higher still, 13% and 19%, respec- tively. When the power of the survey design is considered, the same dilemma is evident (Table 7). Even when three ships are used, the power is acceptably high only if the rate of decrease is at least 10% per year. The probability of detecting a 5% per annum decrease at a 5% alpha level, for example, is only 0.51. This means that with a probability of 0.49 we would conclude that no Table 7. — Power, or the probability of detecting a decrease in spotted dolphin population size dunng a 5-yr penod. Power was calculated for surveys using 1, 2, or 3 ships, for various rates of annual and total population decrease, and for testing the regres- sion of population size against time at vanous significance levels («). Number of ships CV(A/) Rate of decrease per year Total decrease Power when a = 0.05 0.10 0.20 0.20 0.14 0.12 0.01 0.03 0.05 0.10 0.01 0.03 0.05 0.10 0.01 0.03 0.05 0.10 0.05 0.14 023 0.41 0.05 0.14 0.23 0.41 0.05 0.14 0.23 0.41 0.08 0.15 0.26 0.62 0.09 0.22 0.42 0.87 0.10 0.27 0.51 0.94 0.14 0.25 0.40 0.75 0.16 0.34 0.56 0.93 0.18 0.40 0.66 0.97 0.26 0.41 0.57 0.86 0.29 0.52 0.73 0.97 0.31 0.57 0.80 0.99 decrease had taken place, when in fact it had. Power is even less if only one or two ships are used. Alternatively, we may have either to conduct the surveys for more than 5 years and/or relax the acceptable alpha and beta error level (Table 8). With three ships and 5% error levels, 5 years is sufficient to detect a 10% per annum decline, but 9 years are required to detect a 5% per annum decline and 13 years are required to detect a 3% per annum decline. For alpha and beta levels equal 0.10 or 0.20 and use of three ships, a 5% decrease can be detected in 7 or 5 years, respec- tively. Table 8. — Number of years required to detect various annual de- creases and total declines of spotted dolphins calculated for sev- eral alpha and beta levels using 1, 2, and 3 ships. CV {N)s were calculated using jackknife formulae and using data in the north area pooled over all strata (method B). Number ships CV(A/) Decrease per year Number years required a = P = 0.05 1 a = p = 0.10 1 a = p = 0.20 1 0.20 0.14 0.12 0.20 0.14 0.12 0.20 0.14 0.12 Total decrease 0.01 39 0.32 0.03 19 0.44 0.05 13 0.49 0.10 8 0.57 0.01 30 0.26 0.03 14 0.35 0.05 10 0.40 0.10 6 0.47 0.01 27 0.24 0.03 13 0.33 0.05 9 0.37 0.10 5 0.41 0.01 32 0.28 0.03 15 0.37 0.05 11 0.43 0.10 7 0.52 0.01 25 0.22 0.03 12 0.31 0.05 8 0.34 0.10 5 0.41 0.01 23 0.21 0.03 11 0.29 0.05 7 0.30 0.10 5 0.41 0.01 24 0.21 0.03 11 0.29 0.05 8 0.34 0.10 4 0.34 0.01 19 0.17 0.03 8 0.22 0.05 6 0.26 0.10 3 0.27 0.01 17 0.16 0.03 8 0.22 0.05 5 0.23 0.10 3 0.27 444 HOLT ET AL.: MONITORING DOLPHIN ABUNDANCE DISCUSSION Our analyses indicate that our ability to detect changes in the size of spotted dolphin populations in the eastern tropical Pacific is not very great without substantial long-term ship time. This is not surprising given the vast area of ocean inhab- ited by the dolphins and the low sighting rate from ships. We feel our results represent a gener- ally accurate picture based on available data. However, the analyses must be qualified by not- ing that the data used to generate these results were accumulated during all seasons over 5 years. Data collected in the future will come from surveys conducted at the same time each year and may be less variable. In addition, more precise data gathering techniques or data fitting models may become available. If so, these factors would yield greater ability to detect lower rates of de- crease, greater power, and lower required number of years. On the other hand, the estimates of ex- pected variance have dealt with survey precision (measurement error) only. If environmental vari- ability is important, data collected in future surveys may be more variable than we have cal- culated. In long-lived animals with many year classes contributing to reproduction, however, en- vironmental variability will tend to be less impor- tant than survey imprecision (Gerrodette in press). The selection of appropriate alpha and beta errors levels depends on one's perspective. An alpha error would occur if we concluded that a decrease in dolphin abundance was occurring when in fact it was not. It is therefore of interest to the tuna industry to minimize this type of er- ror. A beta error would occur if we concluded that no decrease in dolphin abundance was occurring when in fact it was. It is in the interest of conser- vation groups to minimize this type of error. As is well known in statistical theory, however, there is a trade off between the two types of error, a decrease in one leads to an increase in the other. In our analyses we have balanced the two types of error by making alpha and beta equal. We have also used a range of equal alpha and beta levels (0.05, 0.10, and 0.20) to illustrate how choice of error level can affect sampling design. Higher tol- erance of error leads to lower rates of decrease which could be detected in shorter times, but, of course, one is less sure of the conclusions reached. Thus the choice of acceptable alpha and beta lev- els to use in detecting changes in spotted dolphin population size is a management decision based primarily on social rather than statistical crite- ria. At least two ships are required to provide repre- sentative coverage of the survey area. Although use of a third ship provides better coverage, it does not substantially improve detection of popu- lation decreases. For alpha and beta levels of 0.05, a 5% per year decrease can be detected in 9 years with use of three ships or 10 years with use of two ships (Table 8). For other alpha and beta levels, use of the third ship only increases our ability to detect specific decreases by about 1 year. Given the annual cost of each ship, it would be more cost-effective to conduct the sur- veys for an additional year using only two ships. Another strategy is to conduct surveys less fre- quently than annually. Gerrodette (in press) pro- vides a numerical example of this approach. For parameter values appropriate to spotted dol- phins, conducting surveys less frequently than annually (every second or third year, for example) could save substantial ship time, but more years would elapse before a trend was detected. If a 5% annual decrease in population size oc- curred, the number of spotted dolphins killed would have to be large. Assuming a spotted dol- phin population of 2.5 million animals (Table 1) and disregarding natural mortality and reproduc- tion, approximately 125,000 animals would be killed each year. The estimates of all dolphins taken by the fishery during each of the last few years are only about 40,000 animals per year (Hammond and Tsai 1983). It may be unreason- able to expect annual decreases at the 5% annual level; rather decreases of 3% or 1% per year would be more reasonable. If so, two ships would require at least 14 years to detect the decline (Table 8). Nonetheless, the number of dolphins actually killed may exceed 40,000 animals per year be- cause dolphin mortality aboard the unsampled trips of U.S. and non-U. S. registered vessels, which is assumed to be similar to that on the sampled trips, may in fact be substantially higher. In addition, the effects of chasing and cap- turing dolphins several times per year are not estimated in our analyses. Techniques and data are presented in our paper to determine the optimal number of ships and number of years required to detect decreases in spotted dolphin populations in the eastern tropi- cal Pacific. However, these techniques are appli- cable to investigate the amount of effort and time required to monitor changes in any appropriate population index for any species where sufficient 445 FISHERY BULLETIN: VOL. 85, NO. 3 data exists or can be collected to determine rea- sonable estimates of coefficient of variation. SUMMARY Use of three ships provides excellent physical coverage of the eastern tropical Pacific dolphin area. Coverage using two ships appears adequate while use of one ship yields very sparse cover- age. Assuming alpha and beta levels of 0.05, use of two ships for each of 5 years will only allow us to detect a 13% annual decrease in spotted dolphin abundance. This means that the population could decline by 50^% during the survey period before it could be detected. If three ships are used for 9 years, a 5% decrease per year could be detected. Use of two ships instead of three only decreases our ability to detect specific trends by about 1 year. For alpha and beta levels of 0.05, use of two ships will allow detection of a 5% annual decrease in 10 years, instead of 9 with three ships. The sampling period may be shortened if larger alpha and beta levels and larger annual decreases are acceptable. For alpha and beta levels of 0.10, use of two ships will allow detection of a 10% annual decrease after 5 years during which a 41% decrease in the population could occur. ACKNOWLEDGMENTS We thank J. Barlow, J. Michalski, W. Parks, S. Reilly, G. Sakagawa, S. Sexton, and T. Smith for their constructive reviews of the manuscript. The experimental design benefited greatly from discussions with members of the SWFC PreSOPS panel members including G. Broadhead, K. Burn- ham, D. Chapman, D. Goodman, and P. Patter- son. J. Joy, S. Sexton, and K. Wallace provided computer support, and B. Whalen contributed statistical support. LITERATURE CITED Barlow, J., and R S Holt. 1986. Proportions of species of dolphins in the eastern tropical Pacific. U.S. Dep. Commer., NOAA-TM- NMFS-SWFC-56, 44 p. BuRNHAM, K . D. Anderson, and J. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr. No. 72, 202 p. BUTTERWORTH, D 1982. On the functional form used for g(y) for Minke whale sightings, and bias in its estimation due to mea- surement inaccuracies. Rep. Int. Whaling Comm. 32:883-888. Gerrodette, T. In press. A power analysis for detecting trends. Ecology. Hammond, P. S. 1984. An investigation into the effects of different tech- niques of smearing the IWC/IDCR Minke whale sight- ings data and of the use of different models to estimate density of schools. Rep. Int. Whaling Comm. 34:301- 308. Hammond, P S., and K Tsai 1983. Dolphin mortality incidental to purse-seining for tunas in the eastern Pacific Ocean, 1979-1981. Rep. Int. Whaling Comm. 33:589-597. Holt, R S , and J E. Powers 1982. Abundance estimation of dolphin stocks in the east- ern tropical Pacific yellowfin tuna fishery determined from aerial and ship surveys to 1979. U.S. Dep. Com- mer., NOAA-TM-NMFS-SWFC-23, 95 p. Miller, R. G 1974. The jackknife-a review. Biometrika 61:1-15. Seber, G. a. F. 1973. The estimation of animal abundance. Hafner, N.Y., 506 p. 446 A TRAWL SURVEY METHOD FOR ESTIMATING LOGGERHEAD TURTLE, CARETTA CARETTA , ABUNDANCE IN FIVE EASTERN FLORIDA CHANNELS AND INLETS Richard W Butler. Walter A Nelson, and Tyrrell A. Henwood' ABSTRACT Five eastern Florida navigational channels were surveyed on a quarterly basis from November 1981 through August 1982. The purpose of the surveys was to provide estimates of loggerhead turtle abundance for each channel over all seasons of the year. Standard methods for estimating loggerhead turtle abundance from trawl samples were developed, and the probability of capture in a 30 m by 1,483 m substation (P) was estimated to be 0.28 ± 0.05 (95% confidence level). Abundance estimates based on this probability of capture were then developed for each channel and survey. Of the channels surveyed, only Port Canaveral harbored significant concentrations of loggerhead turtles; populations ranged from 701 ± 291 turtles in February to a low of 38 ± 26 turtles in August. A few loggerhead turtles were captured in the other channels, but infrequency of occurrence suggested random encoun- ters rather than areas of concentration. In the western Atlantic Ocean, loggerhead tur- tles, Caretta caretta, forage throughout the warm waters of the continental shelf from Argentina northward to Nova Scotia, including the Gulf of Mexico and the Caribbean Sea (Carr 1952). On a seasonal basis, loggerheads are common as far north as the Canadian portions of the Gulf of Maine (Lazell 1980), but during cooler months of the year distributions shift to the south (Shoop et al. 1981). Sporadic nesting occurs throughout the tropical and warm temperate range of distribu- tion, but the most important nesting areas are the Atlantic coast of Florida, Georgia, and South Car- olina (Carr and Carr 1978). The Florida nesting population of Caretta has been estimated to be the second largest in the world (Ross 1982). Although population levels of adult female log- gerheads can be estimated from counts on nesting beaches, the remaining animals (males, sub- adults, and nonbreeding females) do not come ashore and are not readily available for census. To estimate the total number of loggerheads in an area, all segments of the population should be considered. In the vicinity of Cape Canaveral, FL, logger- head turtles congregate in the Port Canaveral ship channel (Carr et al. 1980). Because turtles can be captured and studied in this unique area ^Southeast Fisheries Center Mississippi Laboratories, Na- tional Marine Fisheries Service, NOAA, P.O. Drawer 1207, Pascagoula, MS 39568-1207. Manuscript accepted February 1987. FISHERY BULLETIN; VOL.'85, NO. .3, 1987. throughout the year, the National Marine Fish- eries Service (NMFS) has conducted surveys to monitor population levels and estimate relative turtle abundance. This study is a continuation and expansion of research efforts which began in 1978. Presented are results of a 1-yr investigation conducted in response to requests from the U.S. Army Corps of Engineers (COE) and the U.S. Navy, to estimate sea turtle abundance and sea- sonality in five eastern Florida navigational channels. Animals captured in this study were subadults, adult males, and adult females. Popu- lation estimates of subadult turtles may prove to be particularly useful for management, as effi- cacy of conservation measures should be first evi- dent in the population levels of the youngest co- horts. Results of this study provide a reliable index of loggerhead turtle abundance for the study year and an alternative to population estimates based only on nesting females. Most importantly, for the first time, a standard method has been devel- oped that provides sea turtle abundance esti- mates with approximate standard errors. STUDY AREAS Five eastern Florida navigational channels were surveyed on a seasonal basis over the study period. A description of the survey sites follows (each site is diagramed in Figure 1): 447 FISHERY BULLETIN: VOL. 85, NO. 3 SO. CAROLINA €'•'■/: ,-■ '. -'i GEORGIA GULFOF MEXICO FLORIDA ST. MARY'S ENTRANCE PONCE DE LEON INLET CANAVERAL SHIP CHANNEL FORT PIERCE INLET ST. LUCIE INLET Figure 1. — Description of five eastern Florida navigational channels and inlets surveyed. 448 BUTLER ET AL.: ESTIMATING LOGGERHEAD TURTLES BY TRAWL SURVEY 1) St. Mary's entrance to King's Bay (lat. 30°43'N, long. 80°20'W) is divided by the state boundary between Georgia and Florida and in- cludes Cumberland Sound through which the In- tracoastal Waterway connects King's Bay with the entrance channel. Mud predominates inside of the jetties, and mud and rock bottom are found in the channel offshore. 2) Ponce de Leon Inlet (lat. 29°04'N, long. 80°53'W), on the northeast coast of Florida, is a small inlet accessible only to small craft. A jetty protects the inlet to the north; inside the inlet a narrow channel leads to the Intracoastal Water- way. The substrate is hard sand and silt with scattered rubble. 3) The Port Canaveral ship channel is located on the central east coast of Florida (lat. 28°23'N, long. 80°33'W). The ship channel allows naviga- tion from offshore, through a manmade inlet, into a protected harbor. A depth of 11 to 13 m is main- tained by dredging. Soft mud and detritus bottom is found in the channel and sand-clay in the sur- rounding areas. 4) Fort Pierce Inlet (lat. 27°28'N, long. 80°16'W) is located on the south-central east coast of Florida. The channel allows navigation from offshore, through the inlet that is protected by jetties, into the Intracoastal Waterway. The bottom is hard sand and rubble. 5) St. Lucie Inlet, also on the south-central east coast of Florida (lat. 27°09'N, long. 80°07'W), is another small inlet with use limited to small craft. A completed jetty protects the north side of the inlet and a second jetty was under con- struction to the south during the survey periods. The substrate offshore is sloping hard sand and silt. MATERIALS AND METHODS Quarterly trawl surveys of the navigational channels were conducted from November 1981 through August 1982. During each survey, the Port Canaveral ship channel was sampled twice and the remaining four sites (St. Mary's entrance. Ponce de Leon Inlet, Fort Pierce Inlet, and St. Lucie Inlet) were sampled once. A standard 18 m "mongoose" fish trawl, spread by 3 m x 1 m trawl doors and equipped with mudrollers, was used throughout the study period. Prior to the surveys, the boundaries of each channel were located using National Ocean Sur- veys charts and subdivided by a grid pattern for systematic sampling. Lengthwise, each channel was separated into 1,483 m stations which were divided into 30 m wide substations (Fig. 2). The number of substations in each station was dependent on channel width. A systematic sampling scheme was devised to sample each channel substation: every other sta- tion was sampled in leapfrog fashion in one direc- tion, and then the direction was reversed. The substation sampled within each station was de- termined by random drawing without replace- ment and sampling continued until all substa- tions were occupied. This approach avoided the "edge effect", but allowed samples to be statisti- cally treated as random (Milne 1959). Control sta- tions outside the channel were sampled at all sites during each survey period. In addition to standard survey procedures, ex- periments designed to estimate gear efficiency were conducted in the Port Canaveral ship chan- nel. Following each survey, a substation with abundant loggerhead turtles was selected and a series of repetitive tows performed. All logger- heads captured during these experiments were tagged and released on station prior to the next tow. As this was essentially a "removal" method, any recaptures of loggerhead turtles tagged dur- ing the experiment were not considered as part of the catch and were excluded from analysis. Tows were continued in rapid order until two consecu- tive samples yielded zero catches or the working day ended. ANALYTICAL PROCEDURES The efficiency of the sampling gear was estab- lished before population estimates were com- puted. The probability of loggerhead turtle cap- ture iP) was estimated for each repetitive towing experiment using the formula: P = C^/No where Ci - catch on the first tow in the substa- tion Nq = estimated number of loggerhead tur- tles in the substation. A regression of cumulative loggerhead turtle catch (Y) on catch per sample (X), expressed as y = 6o + biX, was used to estimate (A^o* based on the relationship: Nq = bo. The estimated variance of A^o was calculated according to procedures of Kleinbaum and Kupper (1978): 449 FISHERY BULLETIN: VOL. 85, NO. 3 CANAVERAL SHIP CHANNEL SURVEY AREA N 1 mile #SPOIL SITE Figure 2. — Description of the Port Canaveral ship channel survey area. The channel was separated into 1,483 m stations (7-14), which were divided into 30 m wide substations (A, B, C, and D). Var (Nq) = {SE)2 [l/« + X2/S(X. - X)^] where SE = standard error of the estimate pro- vided by the straight line fit n - sample size of the catch data set X, = observed catch per sample in the i^^ sample and The estimated variance of P was then calcu- lated using procedures of Mood et al. (1974): Var iP) - (Ci/iVo^)2 Var iVo- In one instance, the experiment was conducted 450 in an area larger than the standard substation and a ratio (standard area/larger area = 0.75) was used as a constant multiplier to standardize estimates. The mean probability of capture was calculated by combining all experimental P's using the for- mulae: P = P,/k, and Var iP) = Var iP.Vk'^ where k = number of estimates. Once the efficiency of trawling equipment had been determined, the number of loggerhead tur- tles present in a substation (N) was estimated using the following formula (Seber 1973): BUTLER ET AL.: ESTIMATING LOGGERHEAD TURTLES BY TRAWL SURVEY N = CIP where P = probability of capture C = number of animals captured. If more than one sample_tow was made in a substation, the mean catch (C ) was substituted in the above formula. To estimate the number of loggerhead turtles in a channel substation, sta- tion, or the entire channel, thejnean number cap- tured per substation sample (C) times the nuni- ber of substations (s) was substituted: N = sCIP. The estimated variance of this estimate is (Mood et al. 1974): Table 1 . — Estimated probability of loggerhead turtle capture in a Port Canaveral ship channel substation using an 18 m fish trawl. ( Datch on Population Probability first tow estimate of capture Approximate Date (Ci) (A/o) (P) SE(P) 95% C.I. (P) 11/6/81 8 20.19 0.40 0.09 ±0.17 12/5/81 6 19.54 0.31 0.01 ±0.02 12/7/81 13 51.48 0.28 0.03 ±0.07 2/28/82 7 30.31 0.23 0.02 ±0.04 3/2/82 15 72.85 0.21 0.07 ±0.13 5/23/82 2 (*) 5/28/82 2 (*) 6/1/82 1 (*) 8/6/82 3 11.82 0.25 0.05 ±0.10 mean (P) 0.28 0.03 ±0.05 ^ Data set discarded. Var (iV) = (sIP)'^ [Var (C) + (CIP)'^ Var (P)] . RESULTS Estimates of the probability of capture and as- sociated standard error estimates from nine repetitive trawl experiments are presented in Table 1. Estimated probability of capture within a substation based on six experiments ranged from 0.21 to 0.31 (P = 0.28; 95% confidence inter- val = ±0.05; estimated variance = 5.18 x 10""^). Three experiments were excluded from the analy- ses: two were discarded because the catch failed to decline due to low population levels, and a third was eliminated because of problems with the sampling trawl. Estimates of loggerhead turtle abundance by survey for the Port Canaveral ship channel ranged from 701 ± 291 turtles in late February 1982 to a low value of 38 ± 26 turtles in late Au- gust 1982 (Table 2). Port Canaveral channel sta- tions 9 through 11 (Fig. 2) exhibited the highest loggerhead turtle abundance during all seasons of the year. Mean catch for all samples in the chan- nel was 2.55 turtles/tow and 0.50 turtles/tow for control samples, supporting the hypothesis that loggerhead turtles congregate in the Port Canav- eral ship channel. Loggerhead turtle abundance estimates for the remaining four survey sites were low during all seasons of the year (Table 3). Over the study pe- riod, a total of 18 loggerhead turtles was cap- tured: 2 at St. Mary's entrance, 6 at Ponce de Leon Inlet, 3 at Fort Pierce Inlet, and 7 at St. Lucie Inlet. DISCUSSION Our estimates of the probability of capture Table 2. — Estimated number of loggerhead turtles (A/) at Port Canaveral ship channel by station and survey period (1981-1982). Nov. Dec. Feb. Feb. May May 28- Aug. Aug. Station 3-5 2-4 3-6 21-26 7-12 June 1 4-5 20-22 7 V) (') 20 20 221 20 214 20 8 (') (') 25 43 29 11 21 0 9 93 32 114 143 229 21 57 7 10 64 32 254 221 32 21 61 18 11 21 7 3157 146 21 36 8 7 12 21 4 43 89 7 21 4 4 13 0 0 0 11 210 0 0 0 14 0 0 0 4 20 20 0 0 Chanel 4200 475 632 701 152 122 168 38 Approx. 95% C.I. ±129 ±50 ±314 ±291 ±86 ±62 ±82 ±26 ■I Station not sampled 2statlon incompletely sampled. 3|ncludes 4 Kemp's ridley turtles, Lepidochelys kempi. 4Estimate is for stations 9-14, others are for 7-14, Table 3. — Estimated loggerhead turtle abundance dunng quarterly surveys of St. Mary's entrance — King's Bay, Ponce de Leon Inlet, Ft. Pierce Inlet and St. Lucie Inlet. Date St. Marys King's Bay Ponce de Leon Inlet Fort Pierce Inlet St. Lucie Inlet 11/81 2/82 5/82 8/82 9± 18 0 0 0 0 11 ± 15 0 0 0 4±7 4±8 0 0 4±7 11 ± 11 4± 7 were based on the supposition that catch-per-tow in a given substation will decrease as loggerhead turtles are removed. The regression of cumulative loggerhead turtle catch on catch per sample can then be used to estimate the orginal population size in the substation (Brownlee 1965) and using this estimate, the probability of capture can be computed. Assumptions associated with this pro- cedure are a closed population, the trawl fishes only within the defined bounds of the substation, each tow is an equal unit of effort and the proba- bility of capture remains constant. 451 FISHERY BULLETIN: VOL. 85, NO. 3 Although these assumptions may not be satis- fied in all cases, our estimates of probability of capture in a given substation were consistent ex- cept for the two discarded experiments conducted during periods of low loggerhead turtle densities. These findings suggest that some loggerheads en- countering the trawl were able to avoid capture, presumably by moving out of the trawl path. The results also indicate that a consistent percentage of loggerheads were captured by the trawl, facili- tating the estimation of turtle abundance based on number of turtles captured. It should be noted that the probability of capture in a given substa- tion (as presented in our results) is lower than the probability of capture in a given tow. To compute the probability of capture in a single tow, the width of the substation is divided by the width of the trawl and this factor multiplied by the proba- bility of capture in the substation. Loggerhead turtle abundance estimates in the Port Canaveral ship channel exhibited large sea- sonal variation (Table 2). The estimated popula- tion levels during the month of February were significantly higher than all other quarterly sur- veys indicating that loggerheads were most abun- dant during winter months. These findings are in agreement with other NMFS surveys in the Canaveral channel from 1978 to 1983 (Table 4) and support the contention of Carr et al. (1980) that loggerhead turtles may hibernate in the Port Canaveral channel in refuge from low water tem- peratures. The fact that the winter of 1981-82 was unusually mild, could account for the lack of an early winter peak in loggerhead turtle abundance observed in previous years. Data presented in Table 4, while of limited statistical value due to inconsistencies in sam- pling methodologies, are useful for comparisons between this study and other NMFS Canaveral channel surveys. It is worthy of note that mean catch per unit effort (CPUE) by month combining all years was in excess of 10 loggerhead turtles/ hour from November through March with peak concentrations in February and March. Lowest CPUE values and presumably population levels occurred from April through September, which is in agreement with our findings. It is evident that loggerhead turtle abundance estimates were highly variable between surveys made in the same quarter (Table 2). We speculate that these fluctuations in population levels were caused by short-term immigration and emigra- tion in response to local weather changes. We have observed daily changes in catch rates which appear to be correlated with passage of weather fronts. Distribution of loggerhead turtles within the Port Canaveral ship channel is also of interest. In every survey, stations 9, 10, and 11 exhibited the highest abundances, suggesting that they were preferred turtle habitat. Stations 7, 8, and 12 ex- hibited intermediate population levels and sta- tions 13 and 14 had low turtle abundance levels. Stations 7, 8, 9, and 10 were those where deepest cuts into the seabed have been made by dredging. The bottom was characterized by divers as clay- silt and detritus as opposed to the harder clay- sand bottom outside the channel (Carr et al. 1980). Interpretation of loggerhead turtle abundance estimates generated from this study is compli- cated by the fact that three different groups of Table 4. — Summary of catch per unit effort (CPUE) of loggerhead turtles in the Port Canaveral ship channel (1978-83). Values are in turtles per hour standardized to a single 100-ft net. N = number of tows. Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. 1978 37.74 A/ = 7 55.73 A/ = 7 18.99 A/ = 5 10.82 A/= 10 21.64 A/ = 14 1979 11.56 11.88 8.21 9.13 1.33 21.25 13.86 9.43 A/ = 17 A/ = 11 A/ = 11 A/=3 A/ = 5 A/ = 16 A/ = 32 A/ = 28 1980 24.82 19.61 28.57 3.38 3.77 3.31 3.29 262 5.44 11.81 5.11 A/ = 19 A/ = 40 N = 77 A/ = 22 N = 22 A/ = 60 A/ = 152 N = 189 N = 135 N = 105 A/ = 7 1981 15.89 A/ - 12 11.22 A/ = 16 7.88 A/ = 41 3.26 A/ = 51 22.06 A/ = 29 7.18 A/ = 42 1982 41.83 A/ = 99 58.53 A/ = 14 7.49 A/ = 96 4.24 A/ = 15 5.95 A/ = 83 1983 4.86 A/ = 20 2.35 A/ = 60 Totals 18.56 32.61 27.88 4.21 6.50 9.18 6.98 4.77 3.45 5.92 13.80 10.16 A/ = 36 N = 169 N = 129 A/ = 79 N = 123 A/ = 53 A/ = 92 N = 276 N = 268 N = 140 N = 144 A/ = 63 452 BUTLER ET AL.: ESTIMATING LOGGERHEAD TURTLES BY TRAWL SURVEY loggerheads (adult males, adult females, and subadults) utilize the channel at different times of the year (Henwood 1987). Adult males are dom- inant in April and May, adult females are most abundant from May through August and subadult turtles are predominant during the re- mainder of the year. For this reason, direct com- parisons between quarterly surveys may be inap- propriate. It is unfortunate that the three discarded repet- itive trawl experiments occurred in May and June when the population was comprised primar- ily of breeding adults. Low population levels at this time may reflect a reduced catchability coef- ficient in adult loggerhead turtles possibly associ- ated with behavioral changes. The ability of log- gerhead turtles to escape trawls may also be enhanced during periods of high water tempera- tures, but no evidence of this was noted during August or November. Loggerhead turtle abundance in the remaining four channels was low during all quarterly sur- veys. These findings confirm the presence of log- gerhead turtles along much of Florida's eastern coastline, but do not indicate any channel areas with turtle concentrations similar to Port Canav- eral. It is of special interest that only Port Canav- eral, a manmade habitat, harbors concentrations of loggerhead turtles throughout the year and particularly during winter months. The St. Mary's entrance to King's Bay survey area was by far the largest site investigated and may have been incompletely sampled relative to the total area involved. This location was of par- ticular interest to the U.S. Navy because of planned construction of a Trident submarine base in King's Bay. Although no concentrations of log- gerhead turtles were noted over the course of this investigation, future dredging of this channel could potentially result in a situation similar to Port Canaveral, with loggerhead turtles congre- gating in a deepwater manmade habitat. ACKNOWLEDGMENTS Funding for this investigation was provided by the U.S. Army Corps of Engineers, the U.S. Navy, and the National Marine Fisheries Service. The statistical approach was based on recommenda- tions of Al Rainosek. This manuscript benefited from the critical reviews of Steve Malvestuto, Ann Williams, Bill Davies, Bob Shipp, Andy Kemmerer, Warren Stuntz, and Brad Brown. LITERATURE CITED Brownlee, K. a. 1965. Statistical theory and methodology in science and engineering. 2d ed. John Wiley & Sons, Inc., N.Y., 590 p. Carr, a F. 1952. Handbook of turtles: The turtles of the United States, Canada and Baja California. Cornell Univ. Press, Ithaca, N.Y., 542 p. Carr, a , L Ogren, AND C McVea 1980. Apparent hibernation by the Atlantic loggerhead turtle, Caretta caretta, off Cape Canaveral, Flor- ida. Biol. Conserv. 19:7-14. Carr, D , and P. H Carr. 1978. Survey and reconnaissance of nesting shores and coastal habitats of marine turtles in Florida, Puerto Rico and the U.S. Virgin Islands. Report to the NMFS, St. Petersburg, FL, 34 p. Henwood, T A 1987. Movements and seasonal changes in loggerhead turtle, Caretta caretta, aggregations in the vicinity of Cape Canaveral, Florida (1978-84). Biol. Conserv. 40(3):191-202. Kleinbaum, D G , and L L Kupper 1978. Applied regression analysis and other multivari- able methods. Duxbury Press, Boston, MA. Lazell, J D 1980. New England waters: critical habitat for marine turtles, Copeia 1980:290-295. Milne. A. 1959. The centric systematic area-sampled treated as a random sample. Biometrics 15:270-297. Mood. A M , F A Graybill, and D C Boes. 1974. Introduction to the theory of statistics. McGraw- Hill, N.Y. Ross, J P 1982. Historical decline of loggerhead, ridley and leatherback sea turtles. In K. A. Bjorndal (editor). Biol- ogy and conservation of sea turtles, p. 185- 195. Smithson. Inst. Press, Wash. DC. Seber, G A F 1973. The estimation of animal abundance and related parameters. Hafner Press, N.Y., 506 p. SHOOP, C . T DOTY, AND N Bray 1981. Sea turtles in the region between Cape Hatteras and Nova Scotia in 1979. In A characterization of marine mammals and turtles in the mid- and north- Atlantic areas of the U.S. outer continental shelf: Annual report for 1979, p. IX 1-85. Univ. Rhode Island, Kingston. 453 HOMING MIGRATION OF SOCKEYE SALMON, ONCORHYNCHUS NERKA , TO THE ERASER RIVER C. Grooti and T. p. Quinn^ ABSTRACT Adult sockeye salmon, Oncorhynchus nerka, return to the Fraser River via either of two routes: a northern route through Queen Charlotte Strait, Johnstone Strait, and the Strait of Georgia between the mainland and Vancouver Island, and a southern route along the west coast of Vancouver Island and through Juan de Fuca Strait. The proportions of the total run of sockeye salmon using the two routes varies substantially from year to year. Understanding the factors influencing the migratory routes of Fraser River sockeye salmon provides a basis for forecasting the coastal migrations of salmon as they make the transition between oceanic and riverine environments. Our analysis of west coast troll catch and high seas tag-recovery data indicates that the salmon make landfall in different coastal regions from year to year. If the majority of Fraser sockeye approach the coast of Vancouver Island, then most will migrate via the Strait of Juan de Fuca. However, when landfall occurs north of Vancouver Island in the Queen Charlotte Sound area, most homeward migrating Fraser sockeye will travel through Johnstone Strait. Northern diversion rates of Fraser River sockeye salmon for the period 1953-77 were positively correlated with Fraser River discharge. For the period 1978-85 a strong positive correlation was evident with sea surface temperature (SST) along the northwest coast of Vancouver Island (Kains Island lighthouse). We conclude that Fraser River discharge and SST in the vicinity of Kains Island do not guide sockeye salmon in any direct way during their coastal approach, but that they reflect oceanographic conditions that affect salmon migrations directly or indirectly by acting on the feeding distribution, distance, or direction they must travel to reach home. The Fraser River in British Columbia, Canada, is among the most important producers of sockeye salmon, Oncorhynchus nerka, in North America. Forty to sixty separate stocks, inhabiting the dif- ferent lakes of its watershed, produce 2 to 20 mil- lion adults yearly (IPSFC 1954-1985). Sockeye salmon from the Fraser River system generally spend 1 year in nursery lakes after emergence and then migrate to sea as smolts. Most spend two winters in the ocean, returning to spawn in their home river as 4-yr-olds. To reach the Fraser River from their ocean feeding grounds they can take either of two routes around Vancouver Island (Fig. 1). From 1953 until 1977, the majority homed via the southern route through the Strait of Juan de Fuca (average 84%, range 65-98%). Since 1978, a larger proportion of sockeye have migrated via the northern route through John- stone Strait (average through Juan de Fuca Strait 56%, range 20-78%) (IPSFC 1954-1986). 1 Department of Fisheries and Oceans, Fisheries Research Branch, Pacific Biological Station, Nanaimo, British Columbia V9R 5K6, Canada. ^University of British Columbia, Department of Oceanogra- phy, Vancouver, B.C. V6T 1 Y4, Canada; present address: School of Fisheries WH-10, University of Washington, Seattle, WA 98195. Manuscript accepted April 1987. FISHERY BULLETIN: VOL. 85, NO. 3, 1987. In 1958 a relatively high proportion (35%) of Fraser River sockeye salmon returned via the northern route. A large number of fish did not make landfall off the west coast of Vancouver Is- land but rather arrived in the more northerly Queen Charlotte Sound area (Tully et al. 1960) (Fig. 1). This coincided with anomalously high water temperatures off the coast of British Co- lumbia. Tully et al. (1960) and Royal and Tully (1961) suggested that intrusion of warm water from the south in 1958 directed the homing sock- eye salmon northward and closer to the main- land. Moreover, the fish appeared 10 days later in the fishery around Vancouver Island and over a longer period than usual, suggesting that they might have detoured around the area of warm water and made their coastal approach in the cooler nearshore waters. Alternatively, they might have initiated their homeward migration later or from a more distant area than usual. Favorite (1961), on the other hand, took the view that the unusual extent of dilute sea water of Fraser River origin offshore from Queen Char- lotte Sound in 1958 determined the location where the migrating sockeye entered coastal water. He assumed that homeward migrating salmon are attracted to dilute seawater contain- 455 FISHERY BULLETIN: VOL. 85, NO. 3 PERCENT ERASER SOCKEYE USING NORTHERN PASSAGE 100 90 80 70 60 H o 50 40 30 20 10 0 Queen Charlotte' Sound Queen <^'t Charlotte . . ^ ^ -^ Strait Johnstone iStrait Juan de Fuca ■ Strait Salmon Banks 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 YEARS plOO -90 -80 -70 -60 50 o cr LlI MO ^ -30 -20 10 0 Figure l. — Migratory routes of adult sockeye salmon returning to the Fraser River around Vancouver Island. The bar graph indicates the proportion of the total run that was estimated to have used the northern route (data from IPSFC annual reports). ing homestream odors. Wickett (1977) extended this hypothesis by using indices of oceanographic processes to indicate the extent of dilute water plumes off Queen Charlotte Sound from 1953 to 1976. He concluded that the large proportion of Fraser River water discharged into the ocean northwest of Vancouver Island increased the rate of Fraser sockeye migrating through Johnstone Strait. The assumptions by Favorite (1961) and Wickett (1977) were influenced by the proposal of Hasler and Wisby (1951) that riverine odors learned during sensitive juvenile stages guide the homeward migration of adults in nearshore and river environments (Hasler 1966; Hasler and Scholz 1983). We evaluated the extent to which oceano- graphic conditions in offshore waters influence the migratory routes of returning Fraser River sockeye around Vancouver Island. Nine more years of data (1977-85) have become available since Wickett's (1977) publication and four of these show unprecedented (57% to 80%) diversion rates via Johnstone Strait. COASTAL MIGRATORY PATHWAYS OF FRASER RIVER SOCKEYE INPFC Data Under the auspices of the International North Pacific Fisheries Commission, Canada, the United States and Japan tagged 99,576 sockeye salmon in the North Pacific Ocean east of long. 165°E between 1956 and 1983. Of these, 4,842 were recovered, mostly (99.4%) along the coast of British Columbia and Alaska (INPFC 1984). We isolated data on Fraser River sockeye salmon from this larger data set to determine the migra- tory routes of these salmon. In the waters around Vancouver Island, 745 sockeye salmon were re- covered. Since sockeye salmon home accurately (Ricker 1972; Foerster 1968; Quinn 1985), the 456 GROOT and QUINN: HOMING MIGRATION OF SOCKEYE SALMON high seas tagging locations of the recovered fish give information on the distribution of Fraser River sockeye in the ocean. Southern British Co- lumbia sockeye (of which at least 90% are Fraser River fish) were distributed in the Gulf of Alaska southward to lat. 45°N and westward to long. 178°E, a distance of 6,600 km from the Fraser River. The monthly changes in distribution of tagged fish from April to August and recovered in the year of tagging suggest that during spring and summer there is first a shift northeastward in May and June and then southeastward in July and August along southeast Alaska and the Queen Charlotte Islands towards Vancouver Is- land (Fig. 2). The findings are in accordance with the migration model for southeast British Colum- bia sockeye salmon presented by French et al. (1976). Further indications of the coastal approach routes of Fraser River sockeye salmon can be derived from the rate of travel and the assumed direction of movement. From the positions and the dates of tagging and recovery, the rate of travel along the shortest route can be calculated for each fish. The rates ranged up to 98 km/day (4.1 km/hour or about 2 body lengths/second). Sonic tracking studies by Madison et al. (1972), Stasko et al. (1976), and Quinn and terHart (in press) showed that sockeye salmon travel at aver- age speeds of 1.8 to 2.2 km/hour (about 1 body length/second) when migrating, which approxi- mates the optimum sustained swimming speed for mature sockeye in endurance tests (Brett 1983). For 373 sockeye salmon that were tagged 1,000 km or more away from the Fraser River and re- covered around Vancouver Island, 86 (or 23%) travelled at speeds greater than 45 km/day (1.9 km/hour or about 1 body length/second) (Fig. 3). These estimates of swimming speed discount any effect of currents. Current direction and speed vary considerably in the regions through which 64°N I30°W Figure 2. — Distributions in the Gulf of Alaska of releases of sockeye salmon from April through August that were recovered during the year of tagging around Vancouver Island, 1953-85. 457 64°N FISHERY BULLETIN: VOL. 85, NO. 3 40° 130° 64° I30°W Figure 3. — Locations where maturing sockeye salmon were tagged and recovered in the Vancouver Island area. Only recoveries farther than 1,000 km away from the entrance of the Fraser River and with speeds over 45 km/day are illustrated. the sockeye migrate but are usually less than 0.4 km/hour (Favorite et al. 1976; Tabata 1984). We assume that these fish must have travelled on relatively direct courses, day and night, based on the optimum speed of sockeye salmon (~2 km/ hour) and the fact that substantial divergence from straight-line travel would have required the salmon to exceed their fatigue speed of 5 km/hour to accomplish the observed displacements (Quinn 1984; Quinn and Groot 1984). Connecting the tag- ging positions of these 86 sockeye salmon with the mouth of the home river shows that they must have approached Vancouver Island from a west- erly or northwesterly direction (Fig. 3). The sock- eye salmon that were about 3,000 km or more from the Fraser River (mostly tagged in April and May) were generally distributed farther to the south than those tagged later in the season at distances from 1,000 to 2,300 km. The former must have travelled almost due east, while most of the latter may have moved northeast first and then later in the season turned southeast towards Vancouver Island. Thus, sockeye salmon returning to the Fraser River from their ocean feeding grounds approach Vancouver Island from the west and the north- west and, depending on their homing course, gen- erally make landfall along the west coast of Van- couver Island or farther north in Queen Charlotte Sound (Fig. 1). West Coast Troll Fishery To derive information on areas of landfall for different years, we used records of troll fishing off the west coast of Vancouver Island. The Canadian west coast fishery has usually been open during the period that sockeye salmon arrive on the coast. Only during 1978, 1982, 1983, and 1985 were there short nonretention periods for sockeye salmon. We assume, therefore, that in general the catches reflect the migratory patterns of these 458 GROOT and QUINN: HOMING MIGRATION OF SOCKEYE SALMON fish in nearshore waters. The troll catches and boat efforts are recorded on a weekly basis for the different statistical areas by the Department of Fisheries and Oceans of Canada (F. Wong-^). The west coast sockeye salmon catches generally re- flect the annual variability in the total Fraser 3F. Wong, Pacific Biological Station, Nanaimo, B.C. V9R 5K6, Canada, pers. commun. 1984. River run (Henry 1961); therefore, most sockeye salmon captured along the west coast are consid- ered to be returning to this river. Small propor- tions of the catch are of Barkley Sound (Vancou- ver Island) and Lake Washington (USA) origin. In 1979 and 1982 peak catches of sockeye salmon occurred near the middle of Vancouver Island in areas 24-26 (Fig. 4). The relatively low diversion rates (27 and 22% respectively during < UJ >- oc. £ < o tr NORTH SOUTH " 40H Z H 20H X o < " 0 UJ >- UJ o o CO y^ 40- o 27 ^y^^^^^^^^^^^^y^^^^^^ NORTH SOUTH 40- 20--:: 67 25 K<^^^^^^^ 50 75 100% 1981 II 27 26 25 24 23 21 NORTH SOUTH 40- 20- 25 50 75 100% 1979 0- TF g;^^^^^^^^^^^^^^^^^^^^^^^^^^ 25 50 75 100% 1982 m II 27 26 25 24 23 21 NORTH SOUTH I 27 26 25 24 23 21 NORTH SOUTH LU - 20-1 < 0- 70 25 I 50 1^^^^;^^^^^ 75 100% 1980 fi I II m T^ 80 t^^5^^ 80- I 25 I 50 75 IOC 60-:;: 'an-'-- : 1983 20-:; :• \.A n-.ll CvH cn _ ^ r~^ 27 26 25 24 23 FISHING AREAS 27 26 25 24 FISHING AREAS 23 21 Figure 4. — Vertical bar graphs illustrate the proportions of the total troll catch of sockeye salmon off the west coast of Vancouver Island caught in different statistical areas (see map insert) for the years 1979-83. Horizontal bars indicate the proportions of sockeye salmon migrating to the Fraser River via the northern and southern routes. 459 FISHERY BULLETIN: VOL. 85, NO. 3 these years) indicate that most sockeye may have migrated southwest from a relatively southerly position offshore and returned to the Fraser River via the Strait of Juan de Fuca. In 1980 and 1981 the largest sockeye catches were made near the north of Vancouver Island in areas 11 and 27 (Fig. 4). In these years 70 and 67% of the Fraser sock- eye migrated through Johnstone Strait. In 1983 an extreme situation prevailed: most of the sock- eye salmon were caught in area 11 (Queen Char- lotte Sound) and 80% of the fish migrated via the northern route (Fig. 4). From these results we conclude that the propor- tions of Fraser River sockeye salmon returning via the northern and southern routes are gener- ally associated with the area where the fish make their landfall (see also Tully et al. 1960; Henry 1961; IPSFC 1979-1984). If the majority of Fraser sockeye approach the coast west of Vancouver Is- land, then most will continue to migrate via the Strait of Juan de Fuca. However, when landfall occurs north of Vancouver Island in the Queen Charlotte Sound area, most homeward migrating Fraser sockeye will travel through Johnstone Strait. North Coast Salmon Tagging Project During 1982 and 1983, Canada and the United States tagged sockeye salmon along the coast of northern British Columbia and southeastern Alaska to determine interception rates in the commercial fisheries of both countries near the boundary. The results of these studies provide ad- ditional information on migratory routes of this species along the North American coast. In 1982, 40,556 and in 1983, 23,052 maturing sockeye salmon were tagged in several places in south- eastern Alaska and northern British Columbia (Fig. 5) (B. RiddelH). Most of these fish were head- ing for spawning rivers in southeastern Alaska and the Nass and Skeena Rivers of northern British Columbia. However, a number of sockeye salmon, 24 in 1982 and 126 in 1983, were recov- ered in the commercial fishery around Vancouver Island. We assume that most of these were Fraser River fish because more than 90% of sockeye salmon captured in southern British Columbia belong to Fraser River stocks. Of the sockeye salmon tagged in the north, 9 4B. Riddell, Pacific Biological Station, Nanaimo, B.C. V9R 5K6, Canada, pers. commun. 1985. times more were recovered in the Vancouver Is- land area in 1983 than in 1982 (Fig. 5), despite the fact that the total run of sockeye to the Fraser River in 1982 (13,933,000) was more than twice as large as in 1983 (5,167,000; IPSFC 1983, 1984). This indicates that in 1983 a greater proportion of Fraser River sockeye made landfall north of Van- couver Island than in 1982. This was reflected in the diversion rates through Johnstone Strait of 80 and 22% respectively for the 2 years (Figs. 1,4). The results also show that relatively 9 (1982) to 13 (1983) times more sockeye were recovered in the Vancouver Island area from the outside (Noyes and Queen Charlotte Islands, and Cape Muzon) than from the inside (Clarence Strait and Areas 3, 4, and 5) tagging operations (Fig. 5). We suggest that the southern British Columbia sock- eye primarily migrated along the west coast of the Queen Charlotte Islands during their migra- tion south. However, some entered Dixon En- trance in 1983 and travelled through Hecate Strait towards the Fraser River, as indicated by recoveries from inside tagging locations. The findings from the North Coast Tagging Project support the evidence presented earlier that coastal migratory routes of Fraser River sockeye can vary considerably from year to year and that during years of high diversion through Johnstone Strait the returning sockeye make landfall farther north. Analysis of catches, run timing, and stock composition led the Pacific Salmon Commission to a similar conclusion sev- eral years ago (IPSFC 1983). In summary, our analysis of the INPFC, West Coast Troll, and North Coast Tagging data sets indicates that Fraser River sockeye salmon re- turning from ocean feeding grounds approach Vancouver Island from the west and northwest. The area of landfall varies yearly from the west coast of Vancouver Island to more northern re- gions in Queen Charlotte Sound. Moreover, the area where most salmon reach the coast is strongly correlated with the proportion that en- ters the Strait of Georgia via the southern or northern routes. Migratory Routes and Oceanographic Conditions A coastal approach north of Vancouver Island results in a higher proportion of fish moving through Johnstone Strait, while an approach far- ther south, along the west coast of Vancouver Is- land, directs the fish to the Fraser River through 460 GROOT and QUINN: HOMING MIGRATION OF SOCKEYE SALMON the Strait of Juan de Fuca. We propose that the coastal approach may be influenced by oceano- graphic conditions in the eastern Gulf of Alaska during the April-June period when the maturing sockeye perform their homing migrations from the high seas overwintering grounds to the coastal areas (French et al. 1976). The following environmental factors were ana- lyzed for correlation with diversion rates of Fraser River sockeye via the north: 1) Sea surface temperatures (SST) (average April-June) measured daily at four lighthouse stations along the British Columbia coast: Am- phitrite Point, Kains Island, Cape Saint James, and Langara Island (Fig. 6) (Dodimead 1984; L. F. Giovando^). SSTs can be used as indications 5L. F. Giovando, Institute of Ocean Sciences, Patricia Bay, B.C. V8L 4B2, Canada, pers. commun. 1985. 52°N 52°N Figure 5. — The numbers of sockeye salmon tagged in northern British Columbia and southeastern Alaska and recovered that year in waters around Vancouver Island. Tagging data are separated by location of tagging (inside vs. outside waters) and year (1982 vs. 1983). 461 54« 34° 132° 130 ^Langara I. <4j I28°W \ZS° FISHERY BULLETIN; VOL. 85, NO. 3 124° 122° 52°N 50° 48< Cape St James 5I°N-I3I°W PACIFIC OCEAN 54° 52°N Koins 50° Amphi trite Pt-* I 48° 134° 132° 130° I28°W 126° 124° 1220 FiC.URE 6. — Locations where environmental conditions were monitored for regression analysis with sockeye salmon migratory patterns: sea surface temperature and salinity at Amphitrite Point. Kains Island, Cape St. James, and Langara Island; sea level at Tofino; Ekman transport at lat. SIN, long. ISl'^W; and Fraser River discharge at Hope. of the extent of warm water intrusion from the south along the coast. 2) Sea surface temperatures (average April- June) at ocean station P (50 N, 145°W) (see Fig- ure 8) (S. Tabata'') as an indication of ocean condi- tions in the Gulf of Alaska. After sampling at this station was terminated in 1981, temperatures for this area in the Pacific Ocean were obtained from satellite and shipboard observations. 3) Monthly mean sea-levels (average April- June) recorded at Tofino, on the west coast of Vancouver Island (Fig. 6), as an indication of con- vergent and divergent conditions along the coast (A. Dodimead^). Also, high coastal sea levels indi- cate northward currents. 4) Ekman transport normal to the coast (aver- age April-June) at 5rN, 131°W (Fig. 6) calcu- lated from barometric pressure data (Dodimead 6S. Tabata, Institute of Ocean Sciences, Patricia Bay, B.C. V8L 4B2, Canada, pers. commun. 1985. ■'A. Dodimead, Pacific Biological Station, Nanaimo, B.C. V9R 5K6, Canada, pers. commun. 1984. 1984; Giovando fn. 5) to indicate the general pat- tern of circulation from wind-driven transport. 5) Fraser River discharge (average April- June) measured at Hope (Fig. 6) (LeBlond et al. 1983; Inland Waters Directorate^), as an indica- tion of coastal run-off and extent of Fraser River homewater along nearshore areas. Linear regression analysis of Fraser sockeye diversion rates from 1953 to 1985 with SST of lighthouse data from Amphitrite Point, Kains Is- land, Cape St. James, and Langara Island for the months of April to June showed significant corre- lations (Table 1). Since the correlation coefficient of diversion rates was highest with the data from the Kains Island lighthouse, these were selected for further analysis and averaged over April, May, and June. Regression analysis between the northern di- version rates and the environmental variables **Inland Waters Directorate, 1001 West Pender Street, Van- couver, B.C. V6E 2M9, Canada, 1985. 462 GROOT and QUINN: HOMING MIGRATION OF SOCKEYE SALMON Table 1 . — Correlations (R) of sea surface temperature at four light- hiouse stations along the British Columbia coast with the percent- age of sockeye salmon returning to the Fraser River via Johnstone Strait during the years 1953-83. Lighthouse stations tvlarch April t^ay June N N N N Amphitrite Pt. 0.65 31 0.69 31 0.59 31 0.59 31 Kains Island 0.65 31 0.65 31 0.65 31 0.63 31 Cape St. James 0.54 30 0.68 29 0.66 30 0.54 30 Langara Island 0.55 31 0.62 31 0.65 31 0.64 29 listed above for the years 1953-85 showed signifi- cant (P < 0.01) positive correlations with nearshore (Kains Island) SSTs, explaining 51% of the variance (Table 2; Fig. 7A). Table 2. — Relationship (R and R2) for linear regression analyses of average April-June SST at Kains Island lighthouse, sea level at Tofino, Ekman transport at lat. 50°N, long. 13rW, SST at Station P (50°N, 145°W), and Fraser River discharge at Hope with percent- age sockeye salmon returning to the Fraser River via Johnstone Strait for the years 1 953-85 (N = 33), 1 953-77 (N = 25), and 1 978- 85 {N = 8). The level of significance was set at P < 0.01. Factors 1953-85 R2 (%) F 1953-77 1978-85 R2 (%) R2 (%) SST Kains Island 51 0.71* 9 0.30 85 0.92* Sea level Tofino 0.19 19 0.44 3.9 0.20 Ekman transport 4 0.29 4 -0.20 5 0.23 SST Station P 15 0.39 11 0.33 11 0.33 Fraser River discharge 2 0.14 45 0.67* 30 -0.55 •P<0.01. YEARS 1953-1985 < IT 90 60 30 - ® • •• • • 8 8 9 6 10 4 112 12 0 TEMPERATURE °C KAINS ISLAND TEMPERATURE AV APR -JUNE 12 8 YEARS 1953-1977 2 X O 3 o o o V) o 45 n 30 ® 3200 4000 4800 5600 6400 DISCHARGE mVsec YEARS 1978-1985 FRASER RIVER DISCHARGE AV APR-JUNE -21.1 +0,0079x =2 = R' = 45% N = 25 7200 Figure 7. — Relationships between the proportion of Fraser River sockeye salmon migrating via the north- ern route (Johnstone Strait) and A) average sea sur- face temperature and salinity at Kains Island for April, May, and June for the years 1953-85; Bl aver- age Fraser River discharge for April, May, and June for the years 1953-77; C ) average sea surface temper- ature at Kains Island for April, May, and June for the years 1978-85. 90 60 30 j5 110 115 TEMPERATURE "C 12 0 KAINS ISLAND TEMPERATURE AV APR-JUNE y=-3l7 + 34.0x R^ = 85% N = 8 12 5 463 FISHERY BULLETIN: VOL. 85, NO. 3 Inspection of the diversion rates over the last 33 years suggested that after 1977 a change oc- curred in the migratory patterns of Fraser River sockeye approaching the coast. During the period 1978-85, unprecedented rates of 70 to 80% have used the northern route (Fig. 1). This change in migratory behavior coincided with a remarkable prolonged warming period in the northeast Pacific Ocean (Chelton 1984; McClean 1984). The trend culminated in the extended warm-water anomaly of 1983 along the coast of British Colum- bia, which was associated with the 1982-83 El Nino event that occurred in the equatorial Pacific Ocean. This event was one of the most extreme of the century (Mysak 1985). We therefore carried out separate analyses for the two periods, 1953-77 and 1978-85. The regression analyses for the 1953-77 period identified Fraser River discharge as the only sig- nificant iP < 0.01) factor, explaining 45% of the variance (Table 2; Fig. 7B). This positive relation- ship between northern diversion rates and Fraser River discharge was also suggested by Wickett (1977) for the same time period. For the period 1978-85, the regression analyses indicated that SST at Kains Island was the only helpful predictor, explaining 85% of the variance (Table 2; Fig. 7C). A strong positive relationship between northern diversion rates of Fraser River sockeye salmon and SST at Kains Island was also noted for the years 1973-83 by staff of the Pacific Salmon Commission (IPSFC 1984). DISCUSSION We suggest that sockeye salmon returning to the Fraser River may have been influenced by year-to-year changes in ocean conditions during and between the periods 1953-77 and 1978-85. The relationships of sockeye migration to sea sur- face temperature and river discharge will be dis- cussed separately. Sea Surface Temperature and Sockeye Salmon Migration Leggett's (1977) review of fish migration con- cluded that oceanic fish migrations largely repre- sent the continuous optimization of physiologi- cally important conditions. Temperature is an oceanographic feature whose importance in fish physiology is well established (Brett 1970). While evidence indicates that thermal conditions may be correlated with the timing of salmon migra- tions (Burgner 1980; Blackbourn in press) or the route of their return migration to coastal waters (this study), it is not clear how temperature af- fects salmon behavior. Temperature might di- rectly influence salmon in some way or it might merely correlate with some other oceanographic feature influencing them such as eddies and cur- rents (Mysak 1986), or the abundance or species composition of prey items (Fulton and LeBrasseur 1985). If so, a correlation of salmon behavior with temperature could mislead attempts to under- stand the control of migration. Alternatively, temperature may indeed have a direct impact on sockeye salmon. There is consid- erable evidence that temperature is correlated with the distribution of marine fishes (Brett 1970; Laurs and Lynn 1977; Laurs et al. 1977; Magnu- son et al. 1980). Manzer et al. (1965) and French et al. (1976) summarized the distribution of sock- eye salmon in relation to sea surface tempera- ture. While waters of certain temperatures were generally devoid of sockeye salmon, the apparent thermal preferendum was 3° to 5°C wide and changed seasonally. Manzer et al. (1965) reported that most sockeye were caught by research gill nets in the North Pacific Ocean and Bering Sea in waters of 4° to 6°C in May, 4° to 7°C in June, 8° to 12°C in July, and 9° to 12°C in August. Based on the occurrence of sockeye salmon in large areas of the North Pacific Ocean, French and Bakkala (1974) concluded that they are not exclu- sively associated with specific oceanic conditions. To determine the ways in which temperature might directly affect sockeye salmon, we must ascertain the horizontal and vertical distribution of temperatures which they experience at sea on their homeward journey. An oceanographic sur- vey of the North Pacific Ocean and Gulf of Alaska from 16 to 24 July 1959 (S.I.O.U.C. 1965) provide useful data to suit this purpose (Fig. 8). Tempera- tures at 0, 30, and 50 m depth were used to esti- mate the extent of horizontal and vertical gradi- ents experienced by salmon migrating to the northern tip of Vancouver Island along the path which Fraser River sockeye seem to take. If we assume that salmon swam 48 km/day on the surface along the route of the ship, they would have experienced total temperature changes of + 3.68°C or a daily average of +0.10°C/day (Table 3). Averaged over the stations, vertical excur- sions from 0 to 30 m would have caused the fish to experience changes of -0.81°C. Dives from the surface to 50 m would have been accompanied by changes averaging -5.00°C. Earlier in the sum- 464 GROOT and QUINN: HOMING MIGRATION OF SOCKEYE SALMON 160° 150° 64= 140"^ 130° 64°N 60= 56= 52= 48= Figure 8. — Map of the North Pacific Ocean showing the sites sampled by an oceanographic cruise in 1959 (S.I.O.U.C. 1965) and Station P, where sea surface temperatures and salinities were also recorded. Table 3. — Temperature at depth, recorded by the RV Brown Bear, 16-24 July 1959 (data from S.I.O.U.C. 1965). Latitude (N) Longitude (W) 0 m 30 m 50 m 53°56' 153°18' 10.52 9.98 4.80 53°42' 149°43' 10.54 9.49 4.60 53°30' 146°42' 10.82 9.94 4.80 52°58' 143°27' 11.20 10.90 6.82 52°18' 139°35' 10.90 10.70 6.79 51°40' 135°55' 12.39 12.12 8.58 50°40' 13r57' 13.59 10.60 8.18 49°30' 128°12' 14.20 13.96 9.59 48°44' 125°39' 113.18 18.90 18.23 iData not included In calculations of gradients. mer when the sockeye salmon migrate through these areas, the vertical and horizontal gradients are presumably smaller. Available information indicates that salmon in general and sockeye in particular do not restrict themselves to one depth but rather have a diel vertical movement pattern while at sea. Manzer (1964) reported that most sockeye were caught in gill nets at or near the surface during the night, but in the daytime they were caught in substan- tial numbers as deep as 48 to 60 m. Mishima and Shimazaki (1969) reported a more complex pat- tern: sockeye were most abundant on the surface at 13:00-15:00 h but a second peak of abundance occurred at 03:00-05:00 h. Whereas variations in diel movements and depth distribution may occur, it seems likely that sockeye experience temperature changes of 1°C during their daily movements, and may experience changes of 4° to 5°C if they dive below the mixed layer. The slight changes in temperature associated with horizontal movement relative to vertical movement make it unlikely that the long- distance migration of homing sockeye is deter- mined by physiological responses to temperature (Laevastu 1983). Moreover, the temperatures ex- perienced by sockeye salmon at sea do not seem to reflect physiological optima (Brett 1974, 1983). Nevertheless, there is a west-east gradient of in- creasing temperature over much of the homeward path of Fraser River sockeye salmon. Therefore, "predictive behavioural thermoregulation" (Neill 1979) may play a role in homing, though "reactive behavioural thermoregulation" (e.g., 011a et al. 1975) probably does not. However, gra- dients are an inefficient aid to migration unless 465 FISHERY BULLETIN: VOL. 85, NO. 3 coupled with an independent sense of direction. Moreover, as the salmon near the coast, they may experience a decrease in temperature (Table 3). We conclude that the relationship between sea surface temperatures at Kains Island and the di- version rate of sockeye salmon returning to the Fraser River via Johnstone Strait between 1978 and 1985 reflects the influence of ocean condi- tions on the behavior offish, either on the feeding distribution prior to homing (see also Mysak 1986) or on the homing migration itself. Fraser River Discharge and Sockeye Salmon Migration Most of the fresh water along the British Co- lumbia coast originates from the Columbia, Fraser, and Skeena Rivers and distinct tongues of dilute water (SSS of 32.6%c and less) extend sea- ward from the Strait of Juan de Fuca and Queen Charlotte Sound several hundred kilometers off- shore (Favorite 1961). Wickett (1977) suggested that it was the Fraser River water discharged in the ocean to the northwest of Vancouver Island that increased the percentage of Fraser River sockeye migrating through Johnstone Strait. Fraser River sockeye migrating from their ocean feeding grounds towards the British Co- lumbia coast pass through the areas of dilute sur- face water ("dilute domain") long before making landfall (Fig. 9). Interannual changes in river dis- charge and the resulting dilute extensions off- shore could affect the coastal approach routes of Fraser River sockeye by causing more northerly landfall than usual during years of high levels of runoff (Favorite 1961; Wickett 1977). What might be the mechanism that underlies a direct relationship between river discharge and migration route of adult Fraser River sockeye salmon? Two possibilities present themselves. First, returning sockeye salmon could prefer lower salinity water as they home, similar but 64°N I30°W Figure 9. — The estimated distribution of maturing Fraser River sockeye salmon in June, July, and August in relation to regions of dilute water and upwelling. 466 GROOT and QUINN: HOMING MIGRATION OF SOCKEYE SALMON opposite to the increasing saltwater preference documented by Baggerman (1960) and Mclnerney (1964) for juvenile salmon during the period of seaward migration. Mclnerney (1964) argued that the shift in salinity preference over time could gradually lead fish along the gradient of salinities found in coastal areas toward the open ocean. Salinity measurements during the 1959 cruise (S.I.O.U.C 1965) mentioned previously showed that from the middle of the Gulf of Alaska to the tip of Vancouver Island, sea surface salinities de- creased from 32.73 to 32.46%o over a distance of about 1,757 km (Table 4). Fish swimming at a rate of 48 km/day would have met changes aver- aging 0.00015%c/km or 0.0074%c/day when ap- proaching the coast. The threshold for recognition of salinity differences by sockeye salmon is un- known. However, if it is similar to that of min- nows iPhoxinus phoxinus) of 0.003S?f (Glaser 1966), then it is about 20 times higher than the average difference that wil be encountered during a km of travel. Table 4. — Salinity at depth, recorded by the RV Brown Bear. 16-24 July 1959 (data from S.I.O.U.C. 1965) Latitude (N) Longitude (W) 0 m 30 m 50 m 53°56' 153°18' 32.73 32.84 32.88 52°42' 149°43' 32.75 32.84 32.88 53°30' 146^42' 32.85 32.86 32.95 52°58' 143=27' 32.60 32.61 32.72 52°18' 139°35' 32.66 32.68 32.83 5r40' 135°55' 32.24 32.31 32.53 50°40' ^2,VbT 32.15 32.31 32.57 49°30' 128M2' 32.46 32.47 32.53 48^44' 125°39' 131.35 132.28 132.73 iQata not Included In calculations of gradients. Moreover, salinity changes towards the coast do not occur in a smooth gradient (Table 4). Water masses of different salinities and temperatures form a dynamic patchwork that is continuously changing under the influence of wind and cur- rents (Tabata 1984). We therefore consider it un- likely that a general preference for lower salinity water determines the approach direction of hom- ing Fraser River sockeye migration. Smith (1985) concluded that for fishes in general ". . .there is little evidence that salinity is a guiding mecha- nism." Second, the sockeye could react to home odors from the Fraser River in the offshore waters as suggested by Favorite (1961) and Wickett (1977). The sensitivity of salmonids to certain odors is high: 10"^ M for morpholine and 10"^ M for free amino acids (Brett and Groot 1963; Hara et al. 1984). However, it is questionable that, given the extensive mixing in the Fraser River and in the ocean, the already low concentrations of odors from the different nursery lakes would be above threshold. Moreover, odors generally act as re- leasers and not as directors of responses ( Johnsen and Hasler 1980). It is difficult to understand how salmon could change their migration routes far offshore in the ocean, even if they could sense the aroma of their home water. We therefore conclude that the relationship between Fraser River dis- charge and diversion rate of sockeye salmon re- turning via the northern route is not a direct, but probably an indirect one. CONCLUSION We hj^othesize that Fraser River discharge (1953-77) and SST at Kains Island (1978-85) pri- marily reflect certain atmospheric and related oceanographic conditions, which affect Fraser River sockeye salmon winter distribution and/or migration in the ocean. The weather conditions in the Gulf of Alaska are controlled by the locations and intensities of two major semipermanent at- mospheric pressure cells; the Aleutian low and the North Pacific high (Favorite et al. 1976; Thomson 1981; Emery and Hamilton 1985). The interannual variations of these pressure cells af- fect precipitation and the extent of the snow pack during the winter, as well as temperature, salin- ity, and circulation patterns in the ocean (Favor- ite et al. 1976; Thomson 1981; Emery and Hamil- ton 1985). Anomalous temperature conditions in the ocean, resulting from varying atmospheric condi- tions, may affect salmon migrations directly or indirectly by acting on their feeding distribution or on the distance or direction they must travel to reach home. When ocean conditions are warmer than usual, sockeye salmon tend to encounter the coast of British Columbia at the north of Vancou- ver Island. In such cases their approach to the Fraser River will be primarily through Johnstone Strait. Following cold winter conditions in the Gulf of Alaska, landfall usually occurs along the west coast of Vancouver Island and migration to the home river is primarily via the Strait of Juan de Fuca. ACKNOWLEDGMENTS We thank Lawrence Mysak, Kevin Hamilton, 467 FISHERY BULLETIN: VOL. 85, NO. 3 and David Blackbourn for valuable discussions. We also thank the following individuals for providing data used in this paper: Brian Riddell (North Coast Tagging), Fred Wong (West Coast Troll) and Colin Harris (updated information on INPFC tagging). Sherry Greenham and Laurie Mackie drew the figures. During the later stages of this study, T. P. Quinn was supported by NSERC Strategic Grant, G-1485. LITERATURE CITED Baggerman, B. 1960. Salinity preference, thyroid activity and the sea- ward migration of four species of Pacific salmon iOncorhynchus). J. Fish Res. Board Can. 17:295-322. Blackbourn, D. J. In press. Sea surface temperature and the pre-season pre- diction of return timing in Fraser River sockeye. In H. D. Smith, L. Margolis, and C. Wood (editors), Sockeye salmon iOncorhynchus nerka ) population biology and fu- ture management (Proc. Int. Sockeye Salmon Symp. 1985). Can. Spec. Publ. Fish. Aquat. Sci. 96. Brett, J. R. 1970. Temperature: fishes. In O. Kinne (editor). Marine ecology. Vol. I, Part 1, p. 515-560. Wiley-Interscience, Lond., N.Y., Sidney, and Toronto. 1974. Tank experiments on the culture of pan-size sock- eye iOncorhynchus nerka) and pink salmon (0 . gor- buscha) using environmental control. Aquaculture 4:341-352. 1983. Life energetics of sockeye salmon, Oncorhynchus nerka . In W. P. Apsey and S. I. Lustick (editors), Behav- ioral energetics: The cost of survival in vertebrates, p. 29- 63. Ohio State Univ. Press, Columbus, Ohio. Brett, J R , and C. Groot. 1963. Some aspects of olfactory and visual responses in Pacific salmon. J. Fish. Res. Board Can. 20:287-303. BURGNER, R L 1980. Some features of ocean migrations and timing of Pacific salmon. In W. J. McNeil and D. C. Himsworth (editors), Salmonid ecosystems of the North Pacific, p. 153-164. Oregon State Univ. Press, Corvallis, OR. Chelton, D. B. 1984. Commentary: Short-term climatic variability in the Northeast Pacific Ocean. In W. G. Pearcy (editor). The influence of ocean conditions on the production of salmonids in the North Pacific - A workshop, p. 87-99. Oregon State Univ. Sea Grant, Corvallis, OR. Dodimead. a J. 1984. A review of some aspects of the physical oceanogra- phy of the continental shelf and slope waters off the west coast of Vancouver Island, British Columbia. Can. MS Rep. Fish. Aquat. Sci. No. 1773, 309 p. Emery, W J . and K. Hamilton 1985. Atmospheric forcing of interannual variability in the Northeast Pacific Ocean: Connections with El Nino. J. Geophys. Res. 90:857-868. Favorite. F 1961 . Surface temperatures and salinity off the Washing- ton and British Columbia coasts, August 1958 and 1959. J. Fish. Res. Board Can. 18:311-319. Favorite, F., A. J. Dodimead, and K. Nasu. 1976. Oceanography of the subarctic Pacific region, 1960- 71. Int. North Pac. Fish. Comm. Bull. 33, 187 p. Foerster, R. E 1968. The sockeye salmon, Oncorhynchus nerka. Fish. Res. Board Can. Bull. 162, 422 p. French, R R , and R. G Bakkala. 1974. A new model of ocean migrations of Bristol Bay sockeye salmon. Fish. Bull., U.S. 72:589-614. French, R., H. Bilton, M. Osako, and A. Hartt. 1976. Distribution and origin of sockeye salmon iOncorhynchus nerka) in offshore waters of the North Pacific Ocean. Int. North Pac. Fish. Comm. Bull. 34, 113 p. Fulton, J. D., and R. J LeBrasseur 1985. Interannual shifting of the subarctic boundary and some of the biotic effects on juvenile sal- monids. /n Warren S. Wooster and David L. Fluharty (editors), El Nino North: Nino effects, in the eastern sub- arctic Pacific Ocean, p. 237-252. Washington Sea Grant, Univ. Wash., Seattle. Glaser, D 1966. Untersuchungen iiber die absoluten Geschmacks- schwellen von Fischen. Z. Vergl. Physiol. 52:1-25. Hamilton, K. 1985. A study of the variability of the return migration route of Fraser River sockeye salmon iOncorhynchus nerka ). Can. J. Zool. 63:1930-1943. Hara, T J , S MacDonald, R. E. Evans, T. Marui, and S. Arai. 1984. Morpholine, bile acids and skin mucus as possible chemical cues in salmonid homing: Electrophysiological reevaluation. In J. D. McCleave, G. P. Arnold, J. J. Dod- son, and W. H. Neill (editors). Mechanisms of migration in fishes, p. 363-378. Plenum Press, N.Y. Hasler, a. D 1966. Underwater guideposts. Univ. Wis. Press., Madison, 155 p. Hasler, A D , and A. T. Scholz. 1983. Olfactory imprinting and homing in salmon. Springer, Berlin, Heidelberg, N.Y., 134 p. Hasler, A. D , and W J. Wisby 1951. Discrimination of stream odors by fishes and its relation to parent stream behavior. Am. Nat. 85:223- 238. Henry. K A 1961. Racial identification of Fraser River sockeye salmon by means of scales and its application to salmon management. Int. Pac. Salmon Fish. Comm. Bull. 12, 132 p. International North Pacific Fisheries Commission (INPFC). 1984. Annual Report, 1983. Int. N. Pac. Fish. Comm. International Pacific Salmon Fisheries Commission (IPSFC). 1954-1986. Annual reports for 1953-1985. Int. Pac. Salmon Fish. Comm. Johnsen, P. B., and a. D. Hasler. 1980. The use of chemical cues in the upstream migration of coho salmon, Oncorhynchus kisutch Walbaum. J. Fish Biol. 17:67-73. Laevastu, T 1983. Numerical simulation in fisheries oceanography with reference to the Northeast Pacific and the Bering Sea. /n W. S. Wooster (editor). From year to year: inter- annual variability of the environment and fisheries of the Gulf of Alaska and eastern Bering Sea, p. 180- 468 GROOT and QUINN: HOMING MIGRATION OF SOCKEYE SALMON 195. Washington Sea Grant, Univ. Wash., Seattle. Laurs, R. M., H S. H Yuen, and J. H. Johnson. 1977. Small-scale movements of albacore, Thunnus alalunga, in relation to ocean features as indicated by ultrasonic tracking and oceanographic sampling. Fish. Bull., U.S. 75:347-355. Laurs. R M . and R J. Lynn 1977. Seasonal migration of North Pacific albacore, Thunnus alalunga, into North American coastal waters: distribution, relative abundance, and associations with transition zone waters. Fish. Bull., U.S. 75:795-822. LeBlond, p. H.. K Dyck, K Perry, and D Gumming 1983. Runoff and precipitation time series for the coasts of British Golumbia and Washington State. Univ. Br. Co- lumbia, Dep. Oceanogr., MS Rep. 39, 133 p. Leggett. W G 1977. The ecology of fish migrations. Ann. Rev. Ecol. Syst. 8:285-308. Madison. D M . R M Horrall, A. B. Stasko, and A. D. Hasler. 1972. Migratory movements of adult sockeye salmon (Oncorhynchus nerka ) in coastal British Columbia as re- vealed by ultrasonic tracking methods. J. Fish. Res. Board Can. 29:1025-1033. Magnuson, J J , S. B Brandt, and D J. Stewart. 1980. Habitat preferences and fishery oceanography. In J. E. Bardach, J. J. Magnuson, R. C. May, and J. M. Reinhart (editors), Fish behavior and its use in the cap- ture and culture of fishes, p. 371-382. Int. Cent. Living Aquat. Res. Manage., ICLARM Conf Proc. 5. Manzer, J I 1964. Preliminary observations of the vertical distribu- tion of Pacific salmon (Genus Oncorhynchus ) in the Gulf of Alaska. J. Fish. Res. Board Can. 21:891-903. Manzer, J I.. T Ishida, A E Peterson, and M G Hanavan. 1965. Offshore distribution of salmon. Int. North Pac. Fish. Comm. Bull. 15, 452 p. McInerney, J E 1964. Salinity preference: an orientation mechanism in salmon migration. J. Fish. Res. Board Can. 21:995- 1018. McLain. D R 1984. Coastal warming in the Northeast Pacific 1976- 83. In W. G. Pearcy (editor). The influence of ocean con- ditions on the production of salmonids in the North Pacific - A workshop, p. 61-86. Oregon State Univ. Sea Grant, Corvallis, OR. Mishima, S , and K Shimazaki 1969. On the diurnal change of the salmon catch by the gillnet in the Okhotsk Sea. Bull. Fac. Fish. Hokkaido Univ. 20:5-21. Mysak. L a 1986. El Nino, interannual variability and fisheries in the Northeast Pacific Ocean. Can. J. Fish. Aquat. Sci. 43:464-497. Neill, W H. 1979. Mechanisms of fish distribution in heterothermal environments. Am. Zool. 19:305-317. Olla, B. L., a. L Studholme, A J. Bejda, and A D Martin 1975. The effect of temperature on the behavior of marine fishes. In Combined effects of radioactive, chemical, and thermal releases to the environment, p. 299- 308. Int. Atomic Energy Agency, Vienna Austria. Quinn, T. P 1984. An experimental approach to fish compass and map orientation. In J. D. McCleave, G. P. Arnold, J. J. Dod- son, and W. H. Neill (editors), Mechanisms of migration of fishes, p. 113-123. Plenum Press, N.Y. 1985. Homing and the evolution of sockeye salmon, Oncorhynchus nerka . In M. A. Rankin (editor). Migra- tions: mechanisms and adaptive significance, p. 353- 366. Contrib. Mar. Sci., suppl. to Vol. 27. Quinn, T. P., and C. Groot. 1984. Pacific salmon (Oncorhynchus) migrations: orien- tation vs. random movement. Can. J. Fish. Aquat. Sci. 41:1319-1324. Quinn. T P , and B A terHart In press. Movements of adult sockeye salmon in British Columbia coastal waters in relation to temperature and salinity stratification: Ultrasonic telemetry results. In H. D. Smith, L. Margolis, and C. Wood (editors), Sockeye salmon (Oncorhynchus nerka) population biology and fu- ture management (Proc. Int. Sockeye Salmon Symp. 1985). Can. Spec. Publ. Fish. Aquat. Sci. 96. Richer. W E 1972. Hereditary and environmental factors affecting cer- tain salmonid populations. In R. C. Simon and P. A. Larkin (editors), The stock concept in Pacific salmon, p. 19-160. H. R. MacMillan Lectures in Fisheries, Seat- tle, Wash., April 1970. Univ. B.C. Vancouver, B.C. Royal, L. A., and J. P Tully. 1961. Relationship of variable Oceanographic factors to migration and survival of Eraser River salmon. Calif Coop. Oceanic Fish. Invest. Rep. 8, p. 65-68. S.I.O.U.C. (Scripps Institution of Oceanography, University of California). 1965. Oceanic observations of the Pacific: 1959. Univ. Calif. Press, Berkeley, Los Angeles. Smith, R. J. F. 1985. The control of fish migration. Springer, Berl., Heidelberg, NY., 243 p. Stasko, A B , R M Horrall, and A D. Hasler. 1976. Coastal movements of adult Eraser River sockeye salmon (Oncorhynchus nerka ) observed by ultrasonic tracking. Trans. Am. Fish. Soc. 105:64-71. Tabata, S 1984. Oceanographic factors influencing the distribution, migration, and survival of salmonids in the Northeast Pacific Ocean - A review. In W. G. Pearcy (editor). The influence of ocean conditions on the production of salmonids in the North Pacific - A workshop, p. 128- 160. Oregon State Univ., Corvallis, OR. Thomson, R E 1981. Oceanography of the British Columbia coast. Can. Spec. Publ. Fish. Aquat, Sci. 56, 291 p. TULLY, J. P., A J DODIMEAD, AND S. TABATA. 1960. An anomalous increase of temperature in the ocean off the Pacific coast of Canada through 1957 and 1958. J. Fish. Res. Board Can. 17:61-80. WiCKETT, W. Pearcy (MS). 1977. Relationship of coastal oceanographic factors to the migration of Fraser River sockeye salmon (Oncorhynchus nerka, W.). I.C.E.S. CM 1977/M 26, 18 p. 469 PREDATION ON CAPITELLA SPP. BY SMALL-MOUTHED PLEURONECTIDS IN PUGET SOUND, WASHINGTON^ D. Scott Becker^ and Kenneth K. Chew^ ABSTRACT This study examined the predation patterns of three flatfishes (English sole, Dover sole, and rex sole) on the opportunistic polychaetes Capitella spp. in disturbed soft-bottom habitats of Puget Sound, Washington. Sampling was conducted throughout the diel cycle during May and June 1981. All three fishes exhibited some degree of selective predation on Capitella spp. based on both number and size of these prey. Numerical dietary contribution by Capitella spp. was greatest at night for all three fishes, suggesting that these polychaetes become more accessible to predators at night. Predation on Capitella spp. allowed English sole to alter their normal diurnal feeding chronology and forage successfully at night. This study supports the hypothesis that some demersal fishes can exploit opportunistic prey in disturbed habitats. The composition of soft-bottom marine benthic in- vertebrate assemblages can be altered by a vari- ety of natural and anthropogenic disturbances, including salinity reduction (Boesch et al. 1981), storm-induced surge (Rees et al. 1977), hypoxia (Santos and Simon 1980), dredge-spoil dumping (Rhoads et al. 1978), sewage disposal (Pearson and Rosenberg 1978), and oil spills (Sanders et al. 1980). To predict the effects of these events on demersal fishes, predator-prey relationships be- tween benthic invertebrates and their piscine predators must be understood. Unfortunately, this kind of information is rare for marine ecosys- tems (Mills 1975). Frequently, benthic invertebrate assemblages in disturbed habitats are dominated by one or more opportunistic species (e.g., Grassle and Grassle 1974; McCall 1977; Pearson and Rosen- berg 1978; Rhoads et al. 1978). These oppor- tunists are adapted to rapidly colonize disturbed environments and often attain exceptionally high population densities. Because many of these spe- cies reside at or near the sediment-water inter- face, they represent a potential food bonanza to bottom-feeding demersal fishes. When fishes en- counter such an abundant and accessible food source, it seems likely that those species capable of modifying their foraging behavior to fully ex- ploit this windfall will do so. Such opportunistic predation on temporally or spatially variable superabundant prey has been found for a variety of fishes (e.g., Nilsson 1960; Ivlev 1961; Zaret and Rand 1971; Murdoch et al. 1975), and is one pre- diction of optimal foraging theory (review in Pyke et al. 1977). As an example of how a group of demersal fishes responds to a disturbed soft-bottom habitat domi- nated by opportunistic benthic invertebrates, we describe the foraging patterns of three flatfishes (Pleuronectidae) in Puget Sound, WA on Capitella spp., a well-known group of opportunis- tic polychaetes (Grassle and Grassle 1974; Pear- son and Rosenberg 1978). The flatfishes targeted for study were English sole, Parophrys uetulus; Dover sole, Microstomus pacificus ; and rex sole, Glyptocephalus zachirus. These fishes belong to the small-mouthed subgroup of pleuronectids identified by Moiseev (1953) and, as such, prey primarily upon small infaunal and epifaunal ben- thic invertebrates. These species also form a major component of demersal fish assemblages in Puget Sound (Miller et al. 1977; Wingert and Miller 1979; Becker 1984), as well as in most nearshore areas along the west coasts of the United States and Canada (e.g., Alverson et al. 1964; Day and Pearcy 1968; Hart 1973; Allen 1982). iContribution No. 723, School of Fisheries, University of Washington, Seattle. WA 98195. 2School of Fisheries, University of Washington, Seattle, WA 98195; present address: Tetra 'Tech, Inc., 11820 Northup, Bellevue, WA 98005. 3School of Fisheries, University of Washington, Seattle, WA 98195. Manuscript accepted March 1987. FISHERY BULLETIN: VOL. 8.5, NO. 3. 1987. MATERIALS AND METHODS Field Sampling The study was conducted on the delta of the 471 FISHERY BULLETIN: VOL. 85, NO. 3 Puyallup River, in Puget Sound's Commence- ment Bay (Fig. 1). This dynamic area receives a variety of anthropogenic and natural discharges. For example, the river discharges approximately 5,500 kg/year of sediments in a seasonally vari- able manner (Dexter et al. 1981). In addition, the City of Tacoma releases primary-treated sewage into the river at an annual flow rate of 0.9 m'^/sec- ond (20.5 MGD) approximately 2.4 km upstream from the river mouth (Tetra Tech 1981). A pre- liminary survey conducted by the authors showed that benthic invertebrate assemblages through- out much of the delta were dominated numeri- cally by Capitella spp. Field sampling was conducted from 26 May to 3 June 1981. All three target species have spawned by this time (Hart 1973) and, as typical of most adult pleuronectids, are presumably feeding in- tensely to replenish the energy used previously for migration, overwintering, and spawning (e.g., Moiseev 1953; Roff 1982). Sampling was conducted along two 300 m tran- sects located at a depth of 32 ± 2 m (Fig. 1). This depth corresponds to the upper boundary of the N /K CONTOURS IN METERS 100 200 100 J YARDS 1 METERS 200 SAMPLING TRANSECTS- / 50 10 PUGET SOUND TACOMA Puyallup River Figure 1. — Locations of sampling transects and benthic sampling points (i.e., large dots) along each transect. 472 BECKER and CHEW: PREDATION ON CAPITELLA SPP. intermediate faunal zone (i.e., 30-70 m depth) identified for Puget Sound demersal fish assem- blages by Wingert and Miller (1979). All three target species are sufficiently abundant in this depth zone to allow a quantitative food habits analysis to be conducted. Data from both tran- sects were pooled prior to analysis. Fishes were collected using a 7.6 m (headrope) otter trawl having a body mesh size of 3.2 cm (stretched) and a cod-end liner mesh size of 0.8 cm (stretched). Trawling was conducted along both transects at a constant vessel speed of approxi- mately 2.5 kn. All positioning was achieved using the LORAN-C navigation system. To assess diel variations in feeding behavior, hauls were made along each transect during four periods of the diel cycle: morning (0900-1030 h), afternoon (1300- 1500 h), evening (1900-2030 h), and night (2330- 0100 h). Each transect was sampled twice during each time period, yielding eight hauls per tran- sect or a total of 16 hauls for the study. At sea, the stomach contents of the target spe- cies were fixed by injecting, using a 50 cc syringe, a 10% solution of buffered formalin into the body cavity of each individual. These fishes were then brought to the laboratory and stored at 4°C. Benthic invertebrates along each transect were sampled within 2 days of trawling. Organisms were collected using a 0.1 m^ van Veen bottom grab, sieved through a 1.0 mm mesh screen, fixed with a 10% solution of buffered formalin, and transferred to 70% ethanol for storage. A single grab sample was taken during daytime at each of five sampling points positioned at approximately equal distances along each transect (Fig. 1). Laboratory Analysis Within 5 days of sampling, the total length (TL) of each fish was measured to the nearest 1.0 mm. The body cavity was then opened and the stomach was removed by severing the esophagus and py- lorus. Stomachs were stored in 70% ethanol prior to analysis. For food habits analysis, stomachs were sub- sampled from the total pool of available stomachs. To minimize within-species variation as a result of size-dependent foraging patterns (e.g., Gabriel and Pearcy 1981), only individuals within an 80 mm length range were selected for analysis. The ranges used for English sole, Dover sole, and rex sole were 240-320, 200-280, and 210-290 mm TL, respectively. Each length range bracketed the median length observed for each species. Identifications of all invertebrates in stomachs and benthic samples were made using a dissect- ing microscope. Sizes of all Capitella spp. were estimated using the width of the fifth setiger (cf. Tsutsumi and Kikuchi 1984). This measurement was used instead of body length because many of these polychaetes were fragmented during grab sampling or ingestion by the fishes. Setiger widths were measured to the nearest 0.1 mm using an ocular micrometer. The dietary contribution oi Capitella spp. to the total stomach contents of each target species was estimated using percentages based on numerical proportions. In addition, the total number of prey per stomach (i.e., Capitella spp. plus all other or- ganisms) was used as an index of feeding inten- sity for each species. Statistical Analysis Nonrandom predation on Capitella spp. (i.e., se- lection) was tested by comparing the numerical proportions of these polychaetes in the stomachs of the fishes with the proportion found in the ben- thos using a 2 X 2 contingency formulation and the chi-square criterion (Pearre 1982). Direction of selection was determined by inspecting the rel- ative proportions of prey in the stomachs and ben- thos. Nonrandom size selection of Capitella spp. was tested by comparing the size distributions of these polychaetes in the stomachs of the fishes with the size distribution found in the benthos using the Mann-Whitney U-test. In both of these analyses, four comparisons (i.e., one for each time period) for each species were made with a single set of benthic observations. Because these four comparisons lacked independence, significance levels were adjusted conservatively using Bonfer- roni's technique (Miller 1981). To examine how the foraging patterns of Eng- lish sole differed between habitats where benthic assemblages were dominated by Capitella spp. and habitats where assemblages did not include these polychaetes, the values of feeding intensity (i.e., numbers of prey per stomach) found in the present study were compared with those obtained at six other sites in Puget Sound by Becker (1984). These six sites were located at depths be- tween 12 and 32 m, and fishes were sampled and processed using methods identical to those de- scribed for the present study. Values of feeding intensity were compared during each period of the diel cycle using the Mann-Whitney U-test. Similar analyses could not be conducted for Dover 473 FISHERY BULLETIN: VOL. 85, NO. 3 sole and rex sole because these fishes were not sufficiently abundant at the six additional sites. RESULTS Prey Selection Throughout the diel cycle, the numerical pro- portion of Capitella spp. in the diets of all three fishes exceeded the proportion of these poly- chaetes in the benthos (Table 1). Selection of Capitella spp. was highly significant iP < 0.001) during all four time periods for English sole and rex sole, and during morning and night for Dover sole. Selection was significant atP < 0.01 during evening for Dover sole, but not significant (P > 0.05) during afternoon for this species. Table 1. — Comparisons of proportions of Capitella spp. in fish stomachs with the proportion in the benthos using a 2 x 2 contin- gency test. "P < 0.01 , ***P < 0.001 , ns = P > 0.05 (experiment- wise). Number of Capitella spp./Total number of prey 12 Species Morning Afternoon Evening Night English 508/1, 464*" 425/1,029*** 1,072/1,596**" 1,904/2,592* sole (50) Dover 329/861 *** sole (35) Rex 456/526*** sole (16) (40) 114/416 ns (21) 272/412*** (18) (32) 200/612* (36) 209/276* (15) (56) 218/461* (36) 603/671* (19) ■I Number of stomachs examined Is given In parentheses. ^Proportion of Capitella spp. in the benthos was 904/3,51 7. Percent numerical contribution by Capitella spp. to the total diet varied considerably among the three fishes (Fig. 2). Rex sole showed the greatest preference for these polychaetes, includ- ing them in 66-90% of the diet throughout the diel cycle. By contrast, Dover sole exhibited the least preference for Capitella spp., including them in only 27-47% of the diet. English sole showed mod- erate preference for these polychaetes, including them in 35-73% of the diet. Diel variation of feed- ing intensity closely paralleled dietary contribu- tions of Capitella spp. for English sole and rex sole, with both variables peaking at night (Fig. 2). For Dover sole, however, these two variables fol- lowed substantially different diel trends, with percent dietary contribution of Capitella spp. reaching its maximum and feeding intensity dropping to its minimum at night. Although the magnitudes of percent dietary contribution by Capitella spp. differed among the three fishes, several similarities existed in the diel variation of these values (Fig. 2). Minimum dietary contributions were found during morning (English sole) or afternoon (Dover sole and rex sole), whereas maximum contributions were found at night (all three fishes). < t- o Q. Cl •tr 100- 80 60- 40- 20- 0 English Sole 56 32 |— 1 50 ^° I- 100 80 60 40 20 0 O 100-1 O 80- Dover Sole m cc H Z o O < o cr LU z LU o cc LU Q. D 60- 40- 20- 100 80- 60- 40- 20 0 36 35 21 36 r— (0) 16 Rex Sole 15 18 19 Q 100 80 60 40 20 0 100 80 60 ■40 •20 0 X o < o I- w DC LU Q. CC Q. U- O OC UJ m 3 Z Z < Q LU MORNING EVENING AFTERNOON NIGHT SAMPLING PERIOD Figure 2. — Diel predation patterns of the target species. Num- ber of stomachs examined is presented above each pair of bars. Prey Size Selection Median size of Capitella spp. in stomachs ex- ceeded the median size of these polychaetes in the benthos during all four time periods for English sole and rex sole, and during morning, afternoon, and night for Dover sole (Fig. 3). Median prey size for Dover sole during evening was approximately equal to median size of Capitella spp. in the ben- thos. Size differences of Capitella spp. between diets and the benthos were highly significant iP < 0.001) during all four time periods for rex sole, during morning, afternoon, and night for English sole, and during morning and afternoon for Dover sole. Size differences were significant at 474 BECKER and CHEW; PREDATION ON CAPITELLA SPP. o (J) X DC 0) o CO CD > o Q 0) o w c LU E E CL LU O LU CO -^ X LL LL O X h- Q a z 2 ir o z z LU > UJ z 13 ,i C I. o< C X = 01 E ^ d => tn A go, « II 0 n J= C •*^ .So, m * N * i ° ^ V So, CO * u c — c 15 ■" .2 M T3 fl 0 00 1 S 3 ^ •r N -; ■S e g CO CC c 01 en C ■r -2 g K 3 c Co .tJ - — g m be J= — — c« Q. C '^ .2-;; "" r en Q. aj 0) - - |- 03 c o ■■a c -a a. a> c ^ cS § g .2 ■S -2 « 3 -g C £ ^ .2 to - ^ .- to ^ -a ^ ^ C J- O) o is c to ■— *J CO — i: CO — ^ ^ o c E o o ^ c 3 CO cci •J3 lU a; x> D C lN30d3d C Q. X O 3 475 FISHERY BULLETIN: VOL. 85, NO. 3 P < 0.01 during night for Dover sole. No signifi- cant size differences (P > 0.05) were found during evening for English sole and Dover sole. Of the three fishes, rex sole selected the largest Capitella spp. during every time period, with me- dian size ranging from 0.82 to 0.89 mm through- out the diel cycle. Median size of Capitella spp. selected by English sole and Dover sole ranged from 0.62 to 0.66 mm and 0.58 to 0.75 mm, respec- tively. Habitat Comparisons Differences in number of prey per stomach be- tween English sole captured in habitats where Capitella spp. were present and conspecifics cap- tured in habitats where these polychaetes were absent were highly significant (P< 0.001) at night, but not significant (P > 0.05) during morn- ing, afternoon, and evening (Fig. 4). The diel trends of feeding intensity in the two habitats were strikingly different. Where Capitella spp. were present, feeding intensity increased from afternoon to evening, and then peaked at night (median = 53.5 prey per stomach). By contrast, in habitats where Capitella spp. were absent, feed- ing intensity declined from afternoon to evening. 60 -| Capitella spp present * * * Capitella spp absent 56 1 I 50- O C 40- > UJ CL ns 50 32 IS o MBER 111 r^ = 20- z < o LU ^ 10- 145 135 (0) MORNING AFTERI^OON EVENING NIGHT SAMPLING PERIOD Figure 4. — Comparisons of values of feeding intensity for Eng- lish sole between habitats with and without Capitella spp. using the Mann-Whitney U-test. Number of stomachs examined is presented above each bar. Significance level is given above each pair of bars. ***P < 0.001, ns = P > 0.05. and reached a minimum (median = 0 prey per stomach) at night. DISCUSSION Although Capitella spp. accounted for only 25.7% of benthic individuals, their importance as prey to English sole, Dover sole, and rex sole was substantial. All three fishes exhibited significant [P < 0.05) numerical and size selection of these polychaetes during all or most of the diel cycle. Based on literature accounts of the food habits of these fishes, the observed importance oi Capitella spp. as prey could not have been predicted di- rectly. Most historical accounts of the food habits of the three fishes do not identify Capitella spp. as prey (e.g., Hagerman 1952; Kravitz et al. 1977; Hulberg and Oliver 1978; Pearcy and Hancock 1978; Gabriel and Pearcy 1981; Allen 1982; Hogue and Carey 1982). However, most of these studies were conducted in areas where Capitella spp. generally would not be expected to occur in large numbers in the benthos (i.e., the continen- tal shelf off Oregon and California). At least two studies have found that one or more of these fishes consume Capitella spp. Cross et al. (1984) examined the food habits of English sole (n = 13) and Dover sole (n =38) in areas influenced by sewage discharges off Los Angeles, CA. Although C capitata numerically accounted for 40-95% of benthic assemblages, the dietary contributions by this polychaete were small (i.e., 0% for English sole and <10% for Dover sole). Toole (1980) found that C. capitata was a major prey item of juvenile English sole (66-102 mm TL, n = 45) captured on an intertidal sand flat in Humbolt Bay, CA. How- ever, because benthic assemblages were not sam- pled, it is unknown whether these fish were prey- ing nonrandomly on C . capitata . Of the three fishes sampled in the present study, rex sole exhibited the greatest degree of selective predation on Capitella spp. This species was the only one to nonrandomly select Capitella spp. based on both prey number and prey size throughout the diel cycle. In addition, rex sole selected the largest Capitella spp. of the three fishes, and included these polychaetes in the largest percentage of total diet during all four time periods. The observed peak in feeding inten- sity at night agrees with past descriptions of rex sole as a nocturnal forager (Kravitz et al. 1977; Allen 1982; Becker 1984). The concomitant peak in percent dietary contribution oi Capitella spp. at 476 BECKER and CHEW: PREDATION ON CAPITELLA SPP. night indicates that when rex sole were feeding most intensely, selection of Capitella spp. was at its highest level. Dover sole was the least selective of the three fishes with respect to predation on Capitella spp. This species did not exhibit selective predation based on prey number during afternoon, nor based on prey size during evening. In addition, the percent dietary contribution by Capitella s^tp. for Dover sole was the smallest of the three fishes during three of the four time periods. The ob- served minimum level of feeding intensity at night is consistent with the description of Dover sole as a diurnal forager (Allen 1982; Becker 1984). The nighttime peak in percent dietary con- tribution by Capitella spp. suggests that even though this fish normally does not forage at night, Capitella spp. could be captured quite suc- cessfully relative to other benthic invertebrates. English sole was intermediate between rex sole and Dover sole with respect to degree of selective predation on Capitella spp. Although this species selectively consumed these polychaetes based on prey number throughout the diel cycle, prey size selection was not observed during evening. In ad- dition, dietary contribution by Capitella spp. for English sole was the smallest of the three fishes during morning, but intermediate in magnitude during the remainder of the diel cycle. The ob- served peak in feeding intensity at night is con- tradictory to the description of English sole as a diurnal forager (Allen 1982; Hogue and Carey 1982; Becker 1984). Because dietary contribution of Capitella spp. peaked at a high level of 73% at night, much of the ability of English sole to forage at night resulted from predation on these poly- chaetes. The influence of Capitella spp. on noctur- nal foraging by English sole was confirmed by the comparison of diel variation of feeding intensity in habitats with and without Capitella spp. The observed diel variations of predation on Capitella spp. could have resulted from behav- ioral differences of either the fishes or the poly- chaetes. Because the fishes were sampled throughout the diel cycle, much of the variation due to the predators was accounted for. However, because diel variation in behavior of Capitella spp. could not be evaluated using the sampling methods employed in this study, variation in prey availability is unknown. However, at least one pattern is suggested. Because dietary contribu- tion by Capitella spp. peaked at night for all three fishes, these polychaetes may become more active at the sediment surface and thus more vulnerable to predation at night. The ability of English sole to alter its normal diurnal feeding chronology to forage primarily on Capitella spp. at night further suggests that these polychaetes become more ac- cessible at night. Levinton (1971) found that the bivalve Macoma tenta foraged primarily at night and suggested that this periodicity was used, in part, to avoid diurnal predators (primarily winter flounder, Pseudopleuronectes americanus). Al- though this defense mechanism may succeed with obligate diurnal predators, it would not be effec- tive in avoiding nocturnal predators (e.g., rex sole) or species capable of modifying their normal diurnal feeding chronology (e.g., English sole). From an applied standpoint, results of this study have several implications regarding the concept of disturbance management described by Rhoads et al. (1978). Those authors suggested that by properly managing habitat disturbance (i.e., dredge-spoil disposal in their case), benthic invertebrate assemblages can be maintained in the early successional stages when they are dom- inated by pioneering species, including oppor- tunists such as Capitella spp. Because productiv- ity of these early successional stages generally exceeds that of later stages, Rhoads et al. (1978) hypothesized that benthic assemblages domi- nated by pioneering species represent an en- hanced food resource for demersal fishes. The ob- served importance of Capitella spp. as prey for the three fishes considered in the present study sup- ports this hypothesis. For example, all three fishes selectively preyed upon Capitella spp. throughout all or most of the diel cycle, and Eng- lish sole was able to modify its normal diurnal feeding chronology to prey primarily on these polychaetes at night. Although the hypothesis of Rhoads et al. (1978) is supported by the present study, enhancing the productivity of a food resource may not be benefi- cial to demersal fishes if the nutritional quality of their diet is reduced in the process. For example, a variety offish diseases have been attributed, in part, to dietary deficiencies or imbalances of specific nutrients (reviews in National Research Council 1977, 1981). In addition, the toxicity of chemical contaminants to fishes may be enhanced as a result of improper diets (e.g., Mehrle et al. 1977). Although Capitella spp. accounted for only 25.7% of the benthic invertebrates sampled in the present study, the dietary contributions of these polychaetes generally were much greater, espe- cially for rex sole. Given the influence of a bal- anced diet on fish health, it is possible that pro- 477 FISHERY BULLETIN: VOL. 85, NO. 3 longed dietary restriction to one or several oppor- tunistic prey could compromise the health of the fishes. In summary, all three fishes exhibited some degree of selective predation on Capitella spp. based on both number and size of these prey. Di- etary contribution by these polychaetes was greatest at night for all three fishes, suggesting that Capitella spp. may become more accessible to predators at night. Predation on Capitella spp. allowed English sole to alter its normal diurnal feeding behavior and forage successfully at night. Finally, this study supports the hypothesis that some demersal fishes can exploit opportunistic prey in disturbed habitats. ACKNOWLEDGMENTS We thank M. J. Allen, J. N. Cross, and E. W. Hogue for their comments on the manuscript. This study was supported by the Office of Marine Pollution Assessment of the National Oceanic and Atmospheric Administration (Contract NA80RAD00050). A. J. Mearns was the Project Officer. This study is part of a dissertation sub- mitted to the School of Fisheries of the University of Washington (Seattle, USA) in partial fulfill- ment of a Ph.D. degree for D. S. Becker. LITERATURE CITED Allen, M. J. 1982. Functional structure of soft-bottom fish communi- ties of the Southern California Shelf. Ph.D. Thesis, Univ. California, San Diego, 577 p. ALVER.SON, D L., A. T PRUTER, and L L RONHOLT. 1964. A study of demersal fishes and fisheries of the northwestern Pacific Ocean. H. R. MacMillan Lect. Fish., Inst. Fish., Univ. Br. Columbia, Vancouver, 199 p. Becker. D S 1984. Resource partitioning by small-mouthed pleuronec- tids in Puget Sound, Washington. Ph.D. Thesis, Univ. Washington, Seattle, 139 p. Boesch, D. F., R. J. Diaz, and R W Virnstein. 1981. Effects of Tropical Storm Agnes on soft-bottom mac- robenthic communities of the James and York estuaries and the Lower Chesapeake Bay. Chesapeake Sci. 17:246-259. Cross, J. N . J. Roney, and G. S Kleppel. 1984. Fish food habits along a pollution gradient. Calif Fish Game 71:28-39. Day, D. S , and W G, Pearcy. 1968. Species associations of benthic fishes on the conti- nental shelf and slope off Oregon. J. Fish. Res. Board Can. 25:2665-2675. Dexter, R N . D E Anderson, E A Quinlan, L S. Goldstein, R. M. Stickland, S. P Pavlou, J. R Clayton, R M Kocan, and M. Landolt. 1981 . A summary of knowledge of Puget Sound related to chemical contaminants. U.S. Dep. Commer., NCAA Tech. Memo. OMPA-13, Boulder, CO, 435 p. Gabriel, W. L., and W G. Pearcy. 1981. Feeding selectivity of Dover sole. Microstomas pacificus, off Oregon. Fish. Bull., U.S. 79: 749-763. Grassle, J. F., AND J p. Grassle 1974. Opportunistic life histories and genetic systems in marine benthic polychaetes. J. Mar. Res. 32:253-284. Hagerman, F. B 1952. The biology of the Dover sole, Microstomus pac- ificus (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. Hogue, E. W., and A G. Carey 1982. Feeding ecology of 0-age flatfishes at a nursery ground on the Oregon coast. Fish. Bull., U.S. 80:555- 565. Hulberg, L W., and J. S. Oliver. 1978. Prey availability and the diets of two co-occuring flatfish. In S. J. Lipovski and C. A. Simenstad (editors). Fish food habits studies, p. 29-36. Wash. Sea Grant, Univ. Wash., Seattle. IVLEV, V S 1961. Experimental ecology of the feeding of fishes. Yale Univ. Press, New Haven, CT, 302 p. Kravitz, M J , W G Pearcy, and M P Guin 1977. Food of five species of cooccuring flatfishes on Ore- gon's continental shelf Fish. Bull., U.S. 74:984-990. LEVINTON, J. S. 1971. Control of Tellinacean (Mollusca: Bivalvia) feed- ing behavior by predation. Limnol. Oceanogr. 16:660- 662. McCall, p. L. 1977. Community patterns and adaptive strategies of the infaunal benthos of Long Island Sound. J. Mar. Res. 35:221-266. Mehrle, p. M., F. L. Mayer, and W W. Johnson. 1977. Diet quality in fish toxicology: effects on acute and chronic toxicity. In F. L. Mayer and J. L. Hamelink (editors), Aquatic toxicology and hazard evaluation, p, 269-280. American Society for Testing and Materi- als, Philadelphia, PA. Miller, B. S., B. B. McCain, R C Wingert, S F. Borton, K. V. Pierce, and D T Griggs. 1977. Ecological and disease studies of demersal fishes in Puget Sound near Metro-operated sewage treatment plants and in the Duwamish River. Univ. Wash., Fish. Res. Inst., FRI-UW-7721, 164 p. Miller, R. G. 1981. Simultaneous statistical inference. Springer- Verlag, N.Y., 299 p. Mills, E L 1975. Benthic organisms and the structure of marine ecosystems. J. Fish. Res. Board Can. 32:1657-1663. MOISEEV, P A 1953. Cod and flounder of the Far Eastern Seas. Fish Res. Board Can. Transl. Ser., No. 119, 576 p. Murdoch. W. M , S. Avery, and M. E. B. Smyth. 1975. Switching in predatory fish. Ecology 56:1094 1105. National Research Council. 1977. Nutrient requirements of warmwater fishes. Natl. Acad. Sci., Wash., DC, 78 p. 1981. Nutrient requirements of coldwater fishes. Natl. 478 BECKER and CHEW: PREDATION ON CAPITELLA SPP. in 14 north-Swedish Drottningholm 41:185- Acad. Press, Wash., DC, 63 p. NiLSSON. N. A 1960. Seasonal fluctuations in the food segregation of trout, char, and whitefish lakes. Inst. Freshwater Res., 205. Pearcy, W G.. and D Hancock. 1978. Feeding habits of Dover sole. Microstomas pacift- cus ; rex sole, Glyptocephalus zachirus ; slender sole, Ly- opsetta exilis; and Pacific sanddab, Citharichthys sor- didus, in a region of diverse sediments and bathymetry off Oregon. Fish. Bull., U.S. 76:641-651. Pearre. S. 1982. Estimating prey preference by predators: use of var- ious indices, and a proposal of another based on chi square. Can. J. Fish. Aquat. Sci. 39:914-923. 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. Ann. Rev. 16:229-311. PYKE, G. H., H R PULLIAM, AND E L. CHARNOV 1977. Optimal foraging: a selective review of theory and tests. Q. Rev. Biol. 52:137-154. REES, E. I. S.. A. NiCHOLAIDOU, AND P. LASKARIDOU. 1977. The effects of storms on the dynamics of shallow water benthic associations. In B. F. Keegan, P. O. Ceidigh, and P. J. S. Boaden (editors). Biology of benthic organisms, p. 465-477. Pergamon Press, N.Y. Rhoads, D. C p. L. McCall. and J. Y Yingst. 1978. Disturbance and production on the estuarine seafloor. Am. Sci. 66:577-586. Roff, D, a. 1982. Reproductive strategies in flatfish: a first synthe- sis. Can. J. Fish. Aquat. Sci. 39:1686-1698. Sanders. H L.. J. F. Grassle, G. R. Hampson, L. S. Morse, S Garner-Price, and C C. Jones. 1980. Anatomy of an oil spill: long-term effects from the grounding of the barge Florida off West Falmouth, Mas- sachusetts. J. Mar. Res. 38:265-380. Santos. S. L., and J. L. Simon. 1980. Response of soft-bottom benthos to annual catastrophic disturbance in a South Florida estu- ary. Mar. Ecol. Prog. Ser. 3:347-355. Tetra Tech. 1981. Technical evaluation of City of Tacoma Central Treatment Plant Section 301(h) application for modifica- tion of the requirements of secondary treatment. Report prepared for U.S. Environmental Protection Agency, Washington, DC, 299 p. Toole, C. L. 1980. Intertidal recruitment and feeding in relation to optimal utilization of nursery areas by juvenile English sole iParophrys vetulus: Pleuronectidae). Environ. Biol. Fish. 5:383-390. TSUTSUMI. H.. AND T. KiKUCHI. 1984. Study of the life history o{ Capitella capitata (Poly- chaeta: Capitellidae) in Amakusa, South Japan includ- ing a comparison with other geographical regions. Mar. Biol. (Berl.) 80:315-321. Wingert, R. C. and B. S. Miller. 1979. Distributional analysis of nearshore and demersal fish species groups and nearshore fish habitat associa- tions in Puget Sound. Fish. Res. Inst. Univ. Wash., FRI- UW-7901, 110 p. Zaret. T M., and a. S. Rand. 1971. Competition in tropical stream fishes: support for the competitive exclusion principle. Ecology 52:3336- 3342. 479 THE REPRODUCTIVE BIOLOGY OF WALLEYE POLLOCK, THERAGRA CHALCOGRAMMA, IN THE BERING SEA, WITH REFERENCE TO SPAWNING STOCK STRUCTURE Sarah Hinckley' ABSTRACT The reproductive biology of walleye pollock, Theragra chalcogramma, in the Bering Sea was studied from collections of ovaries and observations of spawning in 1984. Spawning occurred in the Aleutian Basin from January through March, in the southeastern Bering Sea from March through June, and northwest of the Pribilof Islands from June through August. Spawning concentrations found in these areas showed significant differences in length at age and fecundity. Histological evidence indicated that the spawning period of an individual female probably lasts less than 1 month. Results indicate at least three separate spawning stocks of walleye pollock within the Bering Sea. These are located in the Aleutian Basin, over the southeastern continental shelf and slope and northwest shelf areas, and over the continental slope northwest of the Pribilof Islands. Mixing of stocks between widely separated spawning grounds due to extended spawning and migration is not likely. Walleye pollock, Theragra chalcogramma, a member of the gadid family found in the North Pacific Ocean, currently supports the largest single-species fishery in the world. In the eastern portion of its range, catches of pollock average 100,000 to 300,000 metric tons (t) per year in the Gulf of Alaska (Alton et al. in press) and 1 ,000,000 1 per year in the Bering Sea ( Bakkala et al. in press). The walleye pollock resource in the Bering Sea is presently managed as a single stock but there is increasing evidence that substocks may exist (Lynde et al. 1986'^). If substocks exist, this information should be considered in manage- ment strategy. If a stock is defined as a production unit or a group of fish showing similar responses to envi- ronmental conditions within a certain geographic area, then population characteristics such as growth rates, fecundity, and size or age at matu- rity may provide the most practical means of dif- ferentiating these units. As these parameters de- termine the yield of a stock to a fishery, identification of production units based on some ^Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. 2Lynde, C. M., M. Van Houghton Lynde, and R. C. Fran- cis. 1986. Regional and temporal differences in growth of walleye pollock (Theragra chalcogramma) in the eastern Bering Sea and Aleutian Basin with implications for manage- ment. NWAFC Proc. Rep. 86-10. Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. Manuscript accepted April 1987. FISHERY BULLETIN; VOL. 85. NO. 3, 1987. or all of them may improve the effectiveness of fisheries management. A clearer understanding of the reproductive process in walleye pollock may also aid in differ- entiating stocks or production units. Nishiyama and Haryu (1981) proposed that walleye pollock in the Bering Sea may migrate and spawn over extensive distances during the long spawning season, implying that spawning groups mix over large areas. Sakurai (1982) has shown in labora- tory experiments that walleye pollock spawn a single group of matured eggs in successive batches, possibly over a period of 1 month. It is not known, however, whether an individual fe- male is capable of repeating vitellogenesis with more than one separate group of eggs in 1 year (i.e., of rematuring the ovary). Evidence of batch spawning or rematuration in Bering Sea walleye pollock may indicate the duration of individual spawning and the potential for mixing between spawning concentrations found in widely sepa- rated areas. In this study, the spatial and temporal distribu- tion of spawning for 1984 has been documented and the length at age and fecundity of walleye pollock from spawning concentrations in different areas has been determined. Walleye pollock ovaries were examined histologically to clarify the process of oocyte development, to learn whether annual fecundity is determinate or inde- terminate and to learn the optimal stage for esti- mating fecundity, and to determine whether 481 FISHERY BULLETIN: VOL. 85, NO. 3 batch spawning or rematuration occur in Bering Sea pollock. METHODS Data and Sample Collection Data and specimens of walleye pollock were col- lected from December 1983 to October 1984 by National Marine Fisheries Service (NMFS) ob- servers aboard foreign commercial fishing vessels and by NMFS pesonnel aboard NOAA research vessels. Data collection was divided into three phases. First, fisheries observers logged the time and location of commerical hauls in which spawn- ing walleye pollock (those with running eggs or milt) were observed. A total of 1,538 observations was made between January and October. Second, otoliths were collected from walleye pollock in hauls where spawning was observed. Third, walleye pollock ovaries were collected for histol- ogy and anaylsis of fecundity. Three ovaries per maturity stage (Hinckley 1986) were collected and preserved in 10% neutral buffered formalin for histological analysis. A separate collection of late developing or mature ovaries was made for the fecundity analysis. Five ovaries per 5 cm length interval were collected over the entire length range encountered. These ovaries were preserved in modified Gilson's solution (Ito 1977'^). A total of 345 ovaries were collected for histology and 294 for fecundity analysis. Data and Sample Processing Location of Spawning and Length at Age of Spawners Spawning locations of walleye pollock were plotted by month. The distribution of fishing ef- fort by the foreign commercial fleet over the spawning season was examined and compared with locations where spawning was found. Water temperatures and depths of capture at spawning locations were also examined. For the length-at-age analysis, five areas were defined within the Bering Sea based on oceano- graphic features (after Lynde et al. fn. 2): the southeast continental shelf, the southeast conti- nental slope, the northwest shelf, the northwest slope, and the Aleutian Basin (Fig. 1). Northwest and southeast areas were divided at the Pribilof Islands and buffer zones were defined in order to clearly separate them. Ages were assigned to a maximum of 200 otoliths per area by readers at the Northwest and Alaska Fisheries Center (NWAFC) age and growth laboratory. Lynde et al. (fn. 2) found that vv^alleye pollock from the Aleutian Basin and the northwest slope were generally slower growing than pollock from the southeast shelf and slope areas. Based on this observation, R. Francis and A. Hollowed'* classi- fied walleye pollock as "northern" (slow-growing) or "southern" (fast-growing), and developed two corresponding growth curves. In the present study, the growth of walleye pollock from spawn- ing concentrations in different areas was com- pared to the two growth curves described in Fran- cis and Hollowed's unpublished study. The geographical distribution of the two growth types was then examined. To derive their "northern" and "southern" growth curves, Francis and Hollowed (unpubl. data) used age data collected by foreign fisheries observers from 1978 to 1983. Age samples were separated into 2° latitude by 1° longitude cells. The mean length at age was estimated for each cell by year, quarter, and sex. Estimates of mean length at age for each cell were used to classify the cell as "northern", "southern", or "unknown". The "northern" classi- fication indicated that the distribution of mean length at age in a given cell was similar to that seen by Lynde et al. (fn. 2) in the northwest slope and Aleutian Basin areas (Fig. 1). The "southern" classification indicated that growth was similar to that observed by Lynde et al. (fn. 2) in the southeast slope and shelf areas (Fig. 1). For this classification, the von Bertalanffy (1938) growth model was fitted to the weighted mean length-at-age data for the "northern" and "southern" areas for each quarter and sex over a period of 6 years (1978-83) (Lynde et al., fn. 2). The model was fitted using the BMDP PAR derivative-free nonlinear least squares estima- tion procedure (Dixon 1983) to produce predicted mean lengths at age for "northern" and "southern" fish. 3Ito, D. H. 1977. Fecundity of the copper rockfish, Sefcasfe.s caurinus (Richardson), from Puget Sound, Washing- ton. Unpubl. manuscr. Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115. ^Robert Francis and Anne Hollowed, Northwest and Alaska Fisheries Center, National Marine Fisheries Service NOAA, 7600 Sand Point Way N.E, Seattle, WA 98115, pers. commun. January 1985. 482 HINCKLEY: REPRODUCTIVE BIOLOGY OF WALLEYE POLLOCK ■ I — / \ \ N \. X.N Northwest / I "> slope Northwest shelf \ N I Southeast shelf sr% "200 mile imit ^ /'Y Southeast slope Aleutian Basin <5^ . 200 nV ^ ^ ^ 61 OON - 56 OON i;'9 OOE 176 OOW 171 OOW 166 OOW Figure 1.— Sampling areas within the Bering Sea. Shaded areas are buffer zones. 61 ■ 51 OON OOW The cells were then classified for each sex and quarter using an index of similarity, Pris ). This index was based on the ratio of the sum of squared deviations of the observed mean length at age from the 6-yr fitted average for the southern re- gion, to the sum of the squared deviations of the observed mean length at age from the southern and northern regions combined: Pr{s) = 1 -[SS{s)/SSis) + SSin)] where SS(s)^ Sum [X(i) - ES(i)f and SS(n) = Sum [X(i) - EN(i)]^. (=1 SS(s) = the sum of squared deviations for the southern regions, SSin ) = the sum of squared deviations for the northern regions, ES (i) = the 6-yr fitted average from the southern region, End) = the 6-yr fitted average from the northern region, I = age, X(i) = the observed mean length at age ii), k = the maximum age observed. Thus, Pris) is the likelihood of the cell being a southern cell. Arbitrary cut-off values of 0.75 and 0.25 were imposed to classify the southern and northern cells, respectively. Any value >0.25 and <0.75 defined an unknown cell. New 6-yr average growth curves were fitted for each region by quarter and sex using only those estimates of mean length at age from cells that were classified as northern or southern (i.e., ex- cluding all unknown cells). Partial F-tests showed (P < 0.001) that there are significant re- gional differences in growth between the "south- ern" (southeast shelf and slope) and the "north- ern" (Aleutian Basin and northwest slope) areas. Using the new 6-yr average growth curves, the 1984 data were divided into cells, compared to Francis and Hollowed's (unpubl. data) growth curve, and were classified as northern, southern, or unknown, based on the index of similarity [Pris)] for both sexes combined: 483 FISHERY BULLETIN; VOL. 85, NO. 3 Pr(s) = 1 - [SS(s)9 + SSis)s/SS{s)^ + SS(s)s + SSin)^ +SSin)j] The classification of the 1984 data was exam- ined to see whether the northern and southern cells fell into the same geographic areas as those which originally defined the growth types de- scribed by Lynde et al. (fn. 2) and the growth curves of Francis and Hollowed (i.e., the north- west slope and Aleutian Basin areas or the south- east shelf and slope areas). Analysis of Fecundity Collections of walleye pollock ovaries for fecun- dity analysis were subsampled by geographic area to provide at least four ovaries in each 5 cm length interval over the entire available length range. Ovaries were collected from walleye pol- lock in all areas except the northwest shelf, where observer coverage was inadequate. The oocytes were separated from the ovarian tissue by wash- ing them through successively finer meshed sieves; and a volumetric subsampling method, similar to that described in Gunderson (1977), was used to estimate the total number of oocytes in each ovary. In all ovaries examined, there was a clear separation in the sizes of unyolked (un- counted) and yolked (counted) oocytes. The histo- logical analysis (described below) indicated that ovaries showing this separation in egg size modes were at the stage of maturity suitable for fecun- dity counts. A minimum of five subsamples were counted for each ovary; if the coefficient of varia- tion between subsample counts exceeded 10%, more counts were done, to a maximum of 10 counts. It was discovered that recounts of individ- ual subsamples accounted for very little of the overall variation between counts, so these were not done on a regular basis. The average coefficient of variation between subsample counts for each ovary was 5.20% (range, 0.87 to 10.30%). The number of oocytes was estimated for a total of 115 ovaries. Fork length, ovary-free weight, and the mean of the subsample estimates of fecundity for each fish were used to derive three relationships for each area within the Bering Sea: length- fecundity, length-weight, and weight-fecundity. Nonlinear least-squares regression methods (Dixon 1983) were used to estimate parameters for each relationship. Unequal error variances were observed in the dependent variables of weight and fecundity. This was accounted for by the use of weighting factors in the nonlinear re- gression procedure. The weighting factor used was the inverse of the variance of the dependent variable. Because comparison of linear regressions is the most direct method and because the linear and nonlinear regressions showed the same relative positions for each area, Newman-Keuls multiple range tests (Zar 1974) were done on the slopes and intercepts of the equations resulting from the lin- ear regressions, to examine differences in the re- lationships by area. The regressions were fitted to a length range (38 to 60 cm) common to all areas before comparison. Tests performed included an overall test for co- incidental regression, Newman-Keuls multiple range tests on the slopes of the lines, and multiple range tests for equality of the intercepts of those relationships found to have equal slopes. The pro- cedures outlined in Zar (1974) were followed for all of these tests. If the tests indicated that the relationship under examination did not differ sig- nificantly between two or more areas, the data from the areas were combined and a curve fitted to the new set of data using the nonlinear proce- dure. Histological Analysis The ovaries of walleye pollock collected for his- tological analysis were classified by area, follow- ing the same geographic scheme used for the length-at-age and fecundity analyses. The collec- tion was subsampled to obtain ovaries in a com- plete range of development from each of the five geographic areas. Sections were removed from the central portion of the ovaries. Tanino et al. (1959) have shown that there is no difference in the size composition of oocytes throughout walleye pollock ovaries. All sections were 6 to 10 |jLm in thickness and were stained with Mayer's haemotoxylin and eosin. In all, 122 ovaries were examined histologically. Oocytes were classified into 12 categories of de- velopment (Hinckley 1986). The overall maturity of each ovary was based on the most advanced oocytes present, and each ovary was assigned to one of 10 maturity classes (Hinckley 1986). Egg-stage frequency counts, as determined from the histological slides, were used to examine the process of oocyte development. A transect grid was drawn on each slide, and oocyte counts for each stage of development were made at each in- tersection on the grid. Stage 8 (tertiary yolk) and 484 HINCKLEY: REPRODUCTIVE BIOLOGY OF WALLEYE POLLOCK stage 9 (nuclear migration) oocytes were called stage 8; stage 11 (maturation) and stage 12 (ovu- lation) were called stage 11, owing to the diffi- culty in consistently differentiating these stages. Counts from all points on the grid were combined into a total egg-stage frequency for each ovary. Development of the maturing oocytes could be followed by comparing the egg-stage frequencies for each level of maturity. Slides were also examined for evidence of ovarian rematuration. Rematuring ovaries were defined as those containing identifiable postovu- latory follicles (the remnant of the egg membrane left after ovulation and release of the oocyte); thick ovarian walls; and resorbing, unspawned, fully yolked oocytes in ovaries that also contained vitellogenic oocytes. RESULTS Spatial and Temporal Distribution of Spawning Observer information showed that walleye pol- lock spawning in the Bering Sea began in the Aleutian Basin in January. As the year pro- gressed, spawning was observed further inward over the continental slope and shelf (Fig. 2A through 2H). Spawning occurred between Janu- ary and March in the basin, between March and June over the southeastern Bering Sea slope and shelf, and between June and August over the northwest slope and shelf. Scattered spawning was noted in the northwestern areas as late as October. Spawning walleye pollock were caught at depths ranging from 46 to 360 m, most commonly between 100 and 250 m. Temperatures at these depths ranged from 1.8° to 6.0°C (x = 2.34°C). The monthly distribution of the commercial fishing fleet in the Bering Sea was examined to assess whether observer reports from the fleet represented the true distribution of spawning. If significant portions of the Bering Sea were not fished by the fleet, concentrations of spawning walleye pollock could have been missed. In Janu- ary and February, coverage of the continental shelf was scattered and in May most fishing oc- curred in the southeast portion of the Bering Sea. Coverage of the Aleutian Basin was minimal after March because harvestable concentrations of spawning walleye pollock could not be found in the area after this time (R. Nelson'^). Neverthe- less, the fishing fleet distribution appears to have been sufficiently extensive to detect the majority of spawning walleye pollock, and the reports of spawning obtained from the fleet appear to rea- sonably represent the true spawning distribution for 1984. Length-Age Characteristics of Walleye Pollock from Spawning Concentrations The plots of mean length at age for walleye pollock males (Fig. 3) and females (Fig. 4) suggest that length at age was similar for both sexes in the Aleutian Basin and over the northwest slope. Lengths for the older ages were smaller in these areas than in the other three. Although the data were widely scattered, length at age was similar in the southeast shelf, southeast slope, and north- west shelf, and was larger in these areas than in the basin and the northwest slope. Comparison of the length-at-age data from 5° by 1° cells with the growth curves generated by Francis and Hollowed showed that most of the cells assigned to the "northern" group were in the northern slope and buffer areas and in the Aleutian Basin (Fig. 5). Walleye pollock in these areas were characterized by a smaller mean length at age. Cells designated as "southern" were mostly from the southeast and northwest shelf areas and southeast buffer zone (Fig. 5) and contained walleye pollock with a larger mean length at age. The spawning concentrations of walleye pollock, therefore, show the same geo- graphic distribution of the two growth types as seen by Lynde et al. (fn. 2) and Francis and Hol- lowed (unpubl. data). Length-Fecundity Relationship In all areas of the Bering Sea, the length- fecundity relationship for walleye pollock (Fig. 6) was found to be curvilinear (F = aL^, Table 1), similar to that observed for walleye pollock from other regions. The overall test for coincidental regression indicated (F = 5.51, P < 0.001) that the length-fecundity relationships for the four areas were not identical. Multiple range test results (Table 2) indicated that the northwest and southeast shelf and slope area regression slopes did not differ significantly. 5R. Nelson, Observer Program, Northwest and Alaska Fish- eries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115, pers. commun. 1984. 485 FISHERY BULLETIN: VOL. 85, NO. 3 / JANUARY 1984 61 OON - 56 OON 51 OON FEBRUARY 1984 ^ ^ lii**"** 51 OON 55 OON 51 OON 179 DOE 1 76 OOW 171 OOW 166 OOW 51 OOW Figure 2.— Observed distribution of spawning walleye pollock in 1984, by month. Shaded areas indicate distribution of the 486 HINCKLEY: REPRODUCTIVE BIOLOGY OF WALLEYE POLLOCK MARCH 1984 51 OON 55 OON 51 OON APRIL 1984 f 61 OON t- 55 OON 51 OON 79 OOE 1 76 OOW 171 OOW 166 OOW 151 OOW Figure 2. — Continued — foreign commercial fishing fleet. Triangles indicate hauls in which spawning walleye pollock were caught. 487 FISHERY BULLETIN: VOL. 85, NO. 3 MAY 1984 - 56 OON 61 OON 51 OON 51 OON ■ 55 OON 51 OON 179 OOE 175 OOW 171 OOW 166 OOW 151 OOW Figure 2. — Continued. 488 HINCKLEY: REPRODUCTIVE BIOLOGY OF WALLEYE POLLOCK JULY 1984 - 56 OON 61 OON 51 OON / AUGUST- OCTOBER 1:984 :^ « - 55 OON 9-September 10 -October 61 OON 51 OON 179 OOE 176 OOW 71 OOW 166 OOW 61 OOW Figure 2.— Continued. 489 FISHERY BULLETIN: VOL. 85, NO. 3 o < < X o 90 n 80 70 60 50 - 40 30 - 20 90 MALES AlEUT 1 AN BASiN —; 1 1 1 r— 2 4 6 —I 1 — I — I 1 1 1 1 1 — I — r— I I 8 10 12 14 15 18 20 AGE fYEARS) 70 o - 60 LU O < < X I — o UJ 50 40 - 30 20 0 - FEMALES T 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 0 2 4 6 8 1012141618 20 AGE (YEARS) Figure 3. — Mean length at age for male walleye pollock from spawning concentrations in five areas of the Bering Sea. Squares indicate fish from the southeast shelf area; triangles, fish from the northwest shelf area; diamonds, fish from the southeast slope area; stars, fish from the northwest slope area; and circles, fish from the Aleutian Basin. Figure 4. — Mean length at age for female walleye pollock from spawning concentrations in five areas of the Bering Sea. Squares indicate fish from the southeast shelf area; triangles, fish from the northwest shelf area; diamonds, fish from the southeast slope area; stars, fish from the northwest slope area; and circles, fish from the Aleutian Basin. Table 1. — Length, weight (W), and fecundity (F) relationships for Bering Sea walleye pollock resulting from the nonlinear regression. n Length-fecundit> { Length-weight Weight-fecundity Region F R2 W R2 F fl2 Southeast shelf 25 0.1926L3 5439 0.865 0.01 20L2 8744 0.996 91.7096W1 1423 0.877 Aleutian Basin 20 469 2282L 15575 0.769 0.1257L2 2300 0.983 9,119.3100W04570 0.737 Southeast slope 28 4.6528L28066 0.944 0.0041 L3 1134 0.997 906.431 5W0 8534 0.935 Northwest slope 38 0.0872L3 7869 0.995 0.0001 L3 4862 0.999 118.6879W1 i'»39 0.995 Combined areas 91 '0.1 71 9L3 6046 0.907 20.0027L3 2743 0.996 1174.9222W10765 0.900 The multiple range test on the regression inter- cepts could not distinguish the intercepts of the three lines from the slope and shelf areas. The multiple range test on regression slopes also showed that the slopes of the regressions from the Aleutian Basin and from the southeast slope areas could not be distinguished; however, the intercepts of the regressions from these two areas differed significantly. Based on these analyses, the length-fecundity relationships from all shelf and slope areas ap- peared to be similar, and the relationship in these areas was different from that seen in the Aleutian Basin. Fecundity increased almost linearly with length in Aleutian Basin walleye pollock; how- ever, pollock larger than about 60 cm are not found in the basin (Okada 1977^ igsS'^; J. fiOkada, K. 1977. Preliminary report of an acoustic survey of the pollock stock in the Aleutian Basin and the adjacent waters in the summer of 1977. Document submitted to the annual meeting of the International North Pacific Fisheries Commission, September 1977. Fishery Agency of Japan. ''Okada, K. 1983. Biological characteristics and abun- dance of the pelagic pollock in the Aleutian Basin. Document 490 HINCKLEY: REPRODUCTIVE BIOLOGY OF WALLEYE POLLOCK \v. N N N ^ U N N N N KT \ N N JJ u u s is S N N N N ^ ^ 61 OON ■ 56 OON 51 OON 179 ODE 176 OOW 171 OOW 166 OOW 161 OOW Figure 5. — Locations of cells containing walleye pollock from spawning concentrations with growth rates similar to "northern" (N), "southern" (S), or "unknown" (U) groups of Lynde et al. (text footnote 2). Traynor^). The great increase in fecundity, seen in fish larger than 60 cm, resulted in the observed curvilinear relationships found in the other areas. Length-Weight Relationship The multiple range test results on the length- weight relationship for walleye pollock were in- conclusive (Table 3). The hypothesis that the length-weight relationship was the same in all areas was rejected (F = 3.4156, 0.0025 < P < 0.005), but the slopes of the regression lines did not differ significantly. A test of intercept equal- submitted to the annual meeting of the International North Pacific Fisheries Commission, 1983. Far Seas Fisheries Re- search Laboratory, Japan. ''^J. Traynor, Northwest and Alaska Fisheries Center, Na- tional Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., Seattle, WA 98115, pers. commun. 1984. Figure 6. — Observed and predicted relationships between fe- cundity and length for walleye pollock from four areas within the Bering Sea. Triangles indicate the number of oocytes per sampled fish. 1750 -1 1500 § 1250 CO a 1000 < V) 13 O X o 750 - 500 250 I / K SLOPE ALEUriAII SWIN 30 40 50 60 70 80 90 LENGTH (CM) 491 FISHERY BULLETIN: VOL. 85, NO. 3 Table 2. — Newman-Keuls multiple range tests on the slopes and intercepts of the linearized length-fecundity regressions from four areas in the Bering Sea (following the notation of Zar 1974). SLOPES Ranks of slopes: 1 2 Ranked slopes: 1.6554 2.7661 Area: Aleutian Basin SE slope 3 4.0967 SE shelf P 4 4.2659 NW slope Comparison Difference SE q q(0.05,78,p) 4 vs. 1 2.6105 4 vs. 2 1.4998 4 vs. 3 Do not test 3 vs. 1 2.4413 3 vs. 2 Do not test 2 vs. 1 1.1107 0.5320 0.5735 0.4926 0.5928 4.9070 2.6152 4.9559 1.8737 4 3 3 2 3.70 3.37 3.37 2.81 INTERCEPTS A. Ranks of intercepts: 1 Ranked intercepts: -4.211 Area: NW slope 2 -3.864 SE shelf 3 1.711 SE slope Companson q P q(0.05,60,p) 3 vs. 1 1.3870 B. Ranks of intercepts: 1 Ranked intercepts: 1.711 Area: SE slope 3 3.40 2 5.784 Aleutian Basin Comparison q(0.05,35,p) 2 vs. 1 3.6720 2.88 Table 3. — Newman-Keuls multiple range tests on the slopes and intercepts of the linearized length-weight regressions from four areas in the Bering Sea (following the notation of Zar 1974). SLOPES Ranks of slopes Ranked slopes Area 1 2.7927 Aleutian Basin 2 2.8883 SE shelf 3 3.1635 SE slope 4 3.3161 NW slope Comparison Difference SE q(0.05,107,p) 4 vs. 1 0.5234 1.7581 0.2977 3.70 INTERCEPTS Ranks of intercepts: Ranked intercepts: Area: 1 -6.3145 NW slope 2 -5.6965 SE slope 3 4 -4.5605 -4.1570 SE shelf Aleutian Basin Comparison q P q(0.05,107,p) 4 vs. 1 4 vs. 2 4 vs. 3 3 vs. 1 3 vs. 2 2 vs. 1 6.2685 4.1848 1.4944 4.3880 2.5906 0.4340 4 3 2 3 2 2 3.70 3.37 2.81 3.37 2.81 2.81 ity showed a significant difference (F = 3.608, 0.01

■ 8 c o u u c 0) 01 JS 0) CC 100 80 60 40 20 0 1 Class VIII [ Early spawning 1 1 1 1 1 .::::•,; K 2 3 4 5 6 7 8 9 10 11 12 Developmental stage of oocytes I 100 80 60 1 40 20 0- 80 60 40 20 0 JiU Class IX Late spawning Class IX Late spawning ' I ' I ; I Class X Spent n = 4 Figure 9.— Continued. 2 3 4 5 6 7 8 9 10 11 12 Developmental stage of oocytes 495 FISHERY BULLETIN: VOL. 85, NO. 3 pletely separated from the remaining reserve fund (Fig. 9E). This appears to be the best stage to estimate fecundity, as no more eggs will be re- cruited into the developing mode, and spawning has not yet begun. The final part of the maturation process ap- pears to occur in a synchronized manner, with successive groups of fully yolked oocytes proceed- ing through the last stages (homogenization of yolk, hydration, ovulation, and release) in dis- crete batches. During the spawning period, the proportion of nonhydrated (stage 8) oocytes grad- ually decreases, and that of hydrated oocytes (stage 11) increases, until only hydrated oocytes remain. This progression is visible in Figure 9H through L. The spawning of matured oocytes in discrete groups appears to represent the "batch spawning" process in walleye pollock. Spent ovaries contain oocytes in the early and late perinucleus stages (the reserve fund), post- ovulatory follicles, and, in some cases, yolked but unspawned oocytes undergoing resorption. Rede- veloping ovaries, i.e., those containing signs of prior spawning and vitellogenic oocytes, were found in small numbers (9 out of 122 ovaries ex- amined). Of the 18 ovaries examined for March, only one appeared to be developing new oocytes, possibly early enough for a second spawning that year. The rest of the redeveloping ovaries were collected from June to September from walleye pollock in the southeast shelf area. Redeveloped ovaries found in the summer with oocytes at the yolk vessicle stage may spawn in the autumn or during the next year (small numbers of walleye pollock have been observed in spawning condition throughout the year). Rematuring ovaries con- taining oocytes at the primary and secondary yolk stages were found only in August and Sep- tember. More than one-half of these ovaries (5 out of 8) showed signs of resorption of the developing oocytes, and would probably not spawn again that year. DISCUSSION The results of this study appear to indicate that at least three separate spawning stocks of walleye pollock exist in the Bering Sea. One is located in the Aleutian Basin, a second over the northwest continental slope, and a third in the southeast shelf, southeast slope, and northwest shelf areas. As noted by Ogawa (1956), geographical isola- tion or ecological separation of spawning concen- trations may indicate population separation. Overall, the spawning season in the Bering Sea lasts about 8 months, and spawning within the different areas is separated by 500 to 1,000 km. Within the different areas, spawning lasts 2 to 3 months. Dissimilarities in several population character- istics were observed between groups spawning in the different areas, supporting the concept of mul- tiple stocks. Length at age differed by area, with larger length at age seen in walleye pollock spawning over the southeast shelf, southeast slope, and northwest shelf, and smaller length at age seen in walleye pollock spawning in the Aleu- tian Basin and over the northwest slope. These results were also found by Lynde et al. (fn. 2). Fecundity relationships in all shelf and slope areas were similar, and differed from that seen in the Aleutian Basin. Aleutian Basin walleye pol- lock showed the lowest fecundity. Walleye pollock from the northwest and the southeast slope areas showed the highest fecundity. Fecundity esti- mates for walleye pollock in Shelikof Strait in 1982 (Miller et al. 1986^; F = 1.2604L32169) and in British Columbia waters in 1979 (Thompson 1981; F = 6.771L2^«^) are higher than the Bering Sea estimates from this study. A general trend of declining fecundity exists towards the northern range of walleye pollock. Due to possible interan- nual variability in fecundity, caution should be taken in comparing studies done in different re- gions and years. Further research is needed for walleye pollock, in the Bering Sea and elsewhere, to determine the proportion of annual fecundity actually realized, i.e., whether resorption of yolked oocytes during maturation and after spawning is significant. Preliminary histological analysis of walleye pol- lock ovaries from Shelikof Strait (Hinckley un- publ. data) suggests that resorption of yolked oocytes may not be significant. Based on similarities in growth, Lynde et al. (fn. 2) proposed that mixing of walleye pollock stocks occurs between the Aleutian Basin and the northwest slope; however, the findings of this study indicate the mixing does not occur during the spawning season. Spawning over the basin and the northwest slope is separated by about 5 months and 500 km, yet there was no sign of 9Miller, B. S., D. R. Gunderson, D. Glass, D. B. Powell, and B. A. Megrey. 1986. Fecundity of walleye pollock (Theragra chalcogramma ) from Shelikof Strait, Gulf of Alaska. Fish. Res. Inst. Rep. FRI-UW-8608. Univ. Washington, Seattle, WA 98195. 496 HINCKLEY: REPRODUCTIVE BIOLOGY OF WALLEYE POLLOCK rematuration in walleye pollock ovaries collected from the northwest slope late in the season, which could have indicated that these fish spawned first in the basin and then later on the northwest slope. The difference in fecundity between these areas also supports the theory that spawners found in the northwest slope area form a separate group from those '"ound in the Aleutian Basin. The similarities in gi'owth found by Lynde et al. (fn. 2) may be a result of mixing occurring at other times of the year. A reduced food supply may produce the smaller length at age and lower fecundity of Aleutian Basin walleye pollock. The reduced growth of walleye pollock in the basin is probably due to the lack offish, particularly juvenile walleye pollock, in the diet (Okada fn. 7; Traynor and Nelson 1985; Dwyer et al. in press). Dwyer (1984) also found that the mean weight of stomach contents of basin-caught fish was low compared with that of fish caught over the shelf and slope. Reduced food supplies have been shown to lower fecundity in several species (Scott 1962; Hester 1964; Bage- nal 1969; Leggett and Power 1969; Wootton 1973, 1977; Hislop et al. 1978). Histological examination of walleye pollock ovaries further supports the theory that spawn- ing concentrations found in widely separated areas do not mix extensively. Walleye pollock spawning in the Bering Sea can be classified as partially synchronous, with one discrete group of oocytes brought to maturation and then spawned in successive batches. This is similar to the pro- cess described for walleye pollock from Japan (Sakurai 1977). The maturation of a second group of oocytes from vitellogenesis to spawning within 1 year does not appear to be common. Maximum walleye pollock fecundity is therefore annually determinate, and the duration of an individual female's spawning period is limited by the dura- tion of the batch spawning process. If an individ- ual does batch spawn a group of matured oocytes over 1 month, as Sakurai's (1982) laboratory studies suggest, it seems unlikely that it would migrate any great distance over this time while actively spawning. To infer stock separation from the results of this study (i.e., that fish return to the same dis- crete areas each year to spawn) requires assum- ing that the timing and distribution of spawning, the dynamics of ovarian maturation, and the dif- ferences in growth and fecundity remain rela- tively constant from year to year. The timing and location of spawning have been similar from 1982 to 1986 (Hinckley unpubl. data; R. Nelson fn. 5). The process of maturation is basically the same for walleye pollock found in the Bering Sea, in the Gulf of Alaska (Miller et al. fn. 9), and in Japanese waters (Sakurai 1977, 1982). Differ- ences in mean length at age represent cumulative differences in growth over the life of a fish, and systematic variation by area such as that seen in this study probably reflects separation over a pe- riod of years. The study by Lynde et al. (fn. 2) documented the same differences in growth by area over a period of 8 years as was seen in this study for 1984. It is not known at what point annual fecundity is determined in walleye pol- lock, but as egg production is influenced by food supply in many species, and as walleye pollock feed mostly during the spring and summer and less during the winter (Dwyer et al. 1986) and the spawning season, yearly fecundity may be deter- mined about 1 year before spawning. The results of this study suggest that the assumption of stock separation over a period of years is reasonable. This study has outlined the timing and distri- bution of walleye pollock spawning in the Bering Sea for 1984, and postulates the existence of at least three separate spawning stocks. Further re- search is needed on the biological and oceano- graphic conditions occurring in the different spawning areas in order to understand the rea- sons for the apparent separation of stocks, and to clarify differences in recruitment and production by these stocks. ACKNOWLEDGMENTS I would like to express my thanks to Kevin Bailey and Robert Francis for their guidance and support. Russell Nelson and the U.S. foreign fish- eries observers made possible the collection of data and samples. Anne Hollowed assisted in the length at age analysis, Linda Rhodes and Beverly Macewicz gave advice on histology, and Berne Megrey assisted with the analysis of fecundity. Arthur Kendall, William Karp, and Gary Stauf- fer made useful comments on the manuscript. LITERATURE CITED Alton, M. S.. M. O. Nelson, and B. A. Megrey. In press. Changes in the abundance, composition, and dis- tribution of pollock {Theragra chalcogramma) in the western Gulf of Alaska (1961-1984). Fish. Res. (Amst.) Bagenal. T. B. 1969. The relationship between food supply and fecundity in brown trout Salmo trutta L. J. Fish. Biol. 1:167-182. 497 FISHERY BULLETIN: VOL. 85, NO. 3 Bakkala, R. G.. V Wespestad, and L-L. Low. In press. Historical trends in abundance and current con- dition of walleye pollock in the eastern Bering Sea. Fish. Res. (Amst.) Bertalanffy. L von 1938. A quantitative theory of organic growth. Human Biol. 10;81-113. Dixon, W J (editor). 1983. BMDP Statistical Software. Univ. Calif. Press, Berkeley. CA., 735 p. DWYER. D A 1984, Feeding habits and daily ration of walleye pollock (Theragra chalcogramma) in the eastern Bering Sea, M,S, Thesis, Univ. Washington, Seattle, 102 p. DwYER, D. A., K Bailey, and P. Livingston In press. Feeding habits and daily ration of walleye pol- lock (Theragra chalcogramma) in the eastern Bering Sea, with special reference to cannibalism. Can. J. Fish. Aquat. Sci. DwYER. D. A , K Bailey, P Livingston, and M Yang, 1986, Some preliminary observations on the feeding habits of walleye pollock {Theragra chalcogramma) in the eastern Bering Sea, based on field and laboratory studies, Int, North Pac. Fish, Comm, Bull, 45:228-246. Foucher, R P . AND R J Beamish 1980, Production of nonviable oocytes by Pacific hake (Merluccius productus ). Can. J. Fish. Aquat. Sci. 37:41- 48. Gunderson, D R. 1977. Population biology of Pacific ocean perch, Sebastes alutiis , stocks in Washington-Queen Charlotte Sound re- gion, and their response to fishing. Fish. Bull., U.S. 75:369-403. Hester. F J 1964. Effects of food supply on fecundity in the female guppy, Lebistes reticulatus (Peters). J. Fish. Res. Board Can. 21:757-764. Hinckley, S 1986, Spawning dynamics and fecundity of walleye pol- lock [Theragra chalcogramma) in the eastern Bering Sea. M.S. Thesis, Univ. Washington, Seattle, 104 p. HiSLOP. J R. G.. A P. ROBB, AND J A GaULD 1978. Observations on effects of feeding level on growth and reproduction in haddock, Melanogrammus aeglefinus (L.) in captivity. J. Fish. Biol. 13:85-98. Leggett. W C , AND G Power 1969. Differences between two populations of landlocked Atlantic salmon (Salmo salar) in Newfoundland. J. Fish. Res. Board Can. 26:1585-1596. Nishiyama, T., AND T Haryu 1981. Distribution of walleye pollock eggs in the upper- most layer of the southeastern Bering Sea. In D. W. Hood and J. A. Calder (editors). The eastern Bering Sea shelf: oceanography and resources, Vol. 2, p, 993- 1012, Univ, Wash, Press, Seattle, Ogawa. T 1956, Studies on fisheries and biology of important fish Suketo-dara (Pollock), Bull, Jpn. Sea Reg, Fish, Res, Lab, 4:93-140. Sakurai, Y (formerly known as T. H. Yoon) 1977, Sexual maturity and spawning process of Alaska pollock, Theragra chalcogramma (Pallas) in Funka Bay, Hokkaido, M.S. Thesis, Hokkaido Univ., Hakodate, Hokkaido, Japan. (Translation by A. Shiga and A. M. Shimada, National Marine Fisheries Service. Northwest and Alaska Fisheries Center, 7600 Sand Point Way NE, Seattle, WA 98115.) 1982. Reproductive ecology of walleye pollock Theragra chalcogramma (Pallas). Ph.D. Thesis, Hokkaido Univ., Hakodate, Hokkaido. Japan. (Translation by T. Nishida, N. D. Davis, and T. Nishiyama for National Marine Fisheries Service, Northwest and Alaska Fish- eries Center, 7600 Sand Point Way NE. Seattle, WA 98115.) ScOTT, D P. 1962, Effect of food quantity on fecundity of rainbow trout, Salmo gairdneri. J, Fish, Res, Board Can, 19:715-731. Tanino, Y , H TsujiSAKi, K Nakamichi, and K Kyushin 1959, On the maturity of Alaska pollock, Theragra chalcogramma (Pallas). (In Jpn.) Hokkaido Reg, Fish, Res, Lab. 20:145-164. Thompson, J M 1981. Preliminary report on the population biology and fishery of walleye pollock (Theragra chalcogramma ) off the Pacific coast of Canada, Can, Tech, Rep. Fish. Aquat. Sci. 1031. Traynor, J. J , AND M O Nelson 1985. Results of the U.S. hydroacoustic survey of pollock on the continental shelf and slope. In R. G. Bakkala, and K. Wakabayashi (editors). Results of cooperative U.S. -Japan groundfish investigations in the Bering Sea during May-August 1979, p. 192-199. Int. North Pac. Fish. Comm. Bull. 44. Wootton, R J 1973. The effect of size of food ration on egg production in the female three-spined stickleback, Gasterosteiis aculea- tus L. J. Fish. Biol. 5:89-96. 1977. Effect of food limitation during the breeding season on the size, body components and egg production of fe- male sticklebacks (Gasterosteus aculeatus). J. Anim. Ecol, 46:823-834. Zar.J H 1974, Biostatistical Analysis, Prentice-Hall, Inc, Engle- wood Cliffs, NJ, 620 p. 498 DISTRIBUTION, FEEDING, AND GROWTH OF LARVAL WALLEYE POLLOCK, THERAGRA CHALCOGRAMMA , FROM SHELIKOF STRAIT, GULF OF ALASKA A. W. Kendall, Jr .^ M. E. Clarke,^ M. M. Yoklavich,^ AND G. W. BOEHLERT"* ABSTRACT Walleye pollock in the Gulf of Alaska have recently been found to form an intense spawning aggregation in late winter in Shelikof Strait. This produces a dense patch of planktonic eggs in early April, and later in spring a patch of larvae that can be followed as it drifts to the southwest. The density of larvae observed in 1981 indicated that density-dependent effects on feeding may be important for larval survival. In May 1983 we conducted a field study to investigate spatial and vertical distribution, feeding, and growth of larvae from this spawning. During this study we found, in an area of maximum concentration (-1 larva m"3) located by an initial survey, larvae averaged 11.1 mm SL, and were similar in size to those found elsewhere. The larvae in 1983 were larger, and less abundant than at the same time in 1981. Larval growth was estimated from the number of otolith daily growth increments at size of larva and was similar in the area of maximum concentration and in other areas. Larvae were concentrated vertically between about 15 and 50 m and showed a crepuscular pattern of increased density at 14-28 m during twilight. Neither the vertical nor horizontal patterns of larval occurrences seemed closely associated with particular values of temperature or salinity. Most larvae were found in a temperature range of 7.0°-5.5°C and a salinity range of 31. 5-32. 29??. Guts of larvae collected during darkness contained less food than those from daytime. Copepod nauplii were largely replaced by Pseudocalanus spp. copepodids in the diet of larvae larger than 14 mm. At the densities of walleye pollock larvae observed in this study, it appears that zooplankton production in the area did not impact larval growth, even in the area of maximum density. A large spawning concentration of walleye pol- lock, Theragra chalcogramma, was discovered in 1980 in Shelikof Strait, and subsequently a 220,000 metric ton/year fishery developed. Shel- ikof Strait, a 50 by 200 km body of water in the northern Gulf of Alaska, between the Kodiak Archipelago and the Alaska Peninsula, is appar- ently the major spawning center for Gulf of Alaska walleye pollock. Ichthyoplankton surveys in 1981 and 1982 showed that spawning occurs primarily in a restricted area within Shelikof Strait and over a short period of time, producing a dense patch of eggs. Thereafter, larvae drift southwest with prevailing currents (Fig. 1). The iNorthwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Pomt Way, N.E., Building 4, BIN C15700, Seattle, WA 98115. 2Rosenstiel School of Marine and Atmospheric Sciences, University of Miami, Miami, FL 33149. ^Cooperative Institute for Marine Resources Studies, Oregon State University, Marine Science Center, College of Oceanography, Newport, OR 97365. ■^Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, 2570 Dole Street, Honolulu, HI 96822-2396. densities of walleye pollock eggs and early larvae found in Shelikof Strait in 1981 exceeded 50 m"*^ (Dunn et al. 1984^), far greater than their densi- ties in the Bering Sea (Kim and Kendall 1983^) or Funka Bay, Japan (Hayashi et al. 1968); more- over, these densities significantly exceed those reported for larvae of any other fish (Hempel 1979). Energetic requirements of larvae in high densi- ties may exceed production of food and possibly lead to density-dependent effects on larval growth and survival (Jones 1973). At larval densities fre- quently found (ca. 1 m""^), density-dependent ef- fects are not considered important (McGowen and Miller 1980; Gushing 1983). Laboratory studies, however, have demonstrated effects of stocking 5Dunn, J. R., A. W. Kendall, Jr., and R. D. Bates. 1984. Distribution and abundance patterns of eggs and larvae of walleye pollock (Theragra chalcogramma) in the western Gulf of Alaska. NWAFC Proc. Rep. 84-10, 66 p. 6Kim, S., and A. W. Kendall, Jr. 1983. The numbers and distribution of walleye pollock eggs and larvae in the southeastern Bering Sea. U.S. Dep. Commer., Natl. Mar. Fish. Serv., NOAA, NWAFC Proc. Rep. 83-22, 35 p. Manuscript accepted March 1987. FISHERY BULLETIN: VOL 85, NO 3, 1987. 499 FISHERY BULLETIN; VOL. 85, NO. 3 ''W^—T^' r'^ Cape ^/~ Kekurnoi ^ .*'**, ■ ( "yY/ V, il' EGGS * *-^ _i 30 March -8 April ^'4^ (no larvae found) I I I Figure 1. — Distribution and abundance of walleye pollock eggs and larvae, spring 1981. Based on Bates and Clark (text fn. 9). Cape Kekurnoi is blackened as a point of reference. Area shown in Figures 2 and 6 is outlined is upper panel. ^ LARVAE 20 -28 May >S9 OOw < ^e OD ■16 00 ■ 5S 00 . OC 'iJ OC 500 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK density upon growth (O'Connell and Raymond 1970; Houde 1975), and recent studies on patches of larvae co-occurring with prey suggest that en- hanced growth may be observed at high prey con- centrations in the field (Govoni et al. 1985). Search volumes of 5 mm larvae morphologically similar to walleye pollock are about 10 L per day (Laurence 1982); thus with densities of 1 larva in 25 L found in Shelikof Strait in 1981, it is possible that density-dependence is important in larval feeding rate and growth. Walleye pollock is widely distributed in the subarctic North Pacific. Larval feeding habit studies have been conducted in Uchiura Bay, Hokkaido, Japan (Kamba 1977) and in the south- eastern Bering Sea (Clarke 1978) where the prin- cipal prey has been found to be copepod nauplii with Pseudocalanus spp. becoming increasingly important as the larvae grow. Larval growth in the same areas has been studied by Hayashi et al. (1968) and Nishimura and Yamada (1984) for Hokkaido and by Walline (1985) and Clarke (1984) for the Bering Sea. Growth rates in field collections and laboratory rearing studies have been shown to be quite variable, from about 0.16 to 0.37 mm d"i (Bailey and Stehr 1986). We conducted a field study to investigate the ecology of larval walleye pollock in Shelikof Strait in May 1983 by locating and sampling the densest patch of larvae. Here we report on growth, feeding habits, and depth distribution of larval walleye pollock we collected. METHODS AND MATERIALS Field Collections An ichthyoplankton survey of 63 stations on a 15 nmi (27.8 km) grid southwest of Kodiak Island, AK, was conducted aboard the NOAA ship Chap- man from 21 to 28 May 1983 (Fig. 2). At each station a MARMAP double oblique bongo tow (Posgay and Marak 1980) was made from the sur- face to 200 m (as water depths permitted) with a 60 cm bongo net equipped with 505 jxm mesh nets. Flowmeters were mounted in the net mouths and a bathykymograph was used to determine the maximum tow depth and to evaluate the tow pro- file. A neuston net (Sameoto and Jaroszynski 1969) with 505 ixm mesh was also towed for 10 minutes at each station. The neuston net sample and one of the bongo net samples at each station were preserved in 5% sodium borate buffered for- malin in seawater. Most walleye pollock larvae from the other bongo net were rough sorted at sea and were immediately preserved in buffered 90% ethanol for otolith examination. The results of the sorting of larvae at sea were used to choose a location of high larval density for the diel feeding/distribution study (Fig. 2). An- other oblique tow, after this survey, confirmed the presence of high concentrations of larvae. Several preliminary tows with four 20 cm bongo nets on the towing wire fixed at 10 m depth inter- vals (between 5 and 91 m) were made to find depths of maximum larval concentrations. A tow was then taken every 4 hours for 48 hours during 28-30 May 1983 with 20 cm bongo nets equipped with 253 |xm mesh nets on one side and 333 ixm mesh nets on the other. Four nets were fished simultaneously for 10 minutes at a ship speed of approximately 100 cm/second. The nets were placed on the wire to fish at four depths within the region of larval abundance (nominally 20, 30, 40, and 50 m). Flowmeters were mounted in the mouths of the nets, and a bathykymograph was deployed with the deepest net to record actual tow depths. During setting and retrieving, the ship maintained reduced speed to minimize fishing outside the chosen depth strata. Thus, although no closing devices were used, nearly all of the water passing into the nets was at the chosen depth (Kendall and Naplin 1981). Tows were made at 1030, 1430, 1830, 2330, 0230, and 0630 local time (sunrise was at 0455 and sunset 2138 h). During the sampling of the stations at 1430 and 0230 on both days, a 1 m^ mechanical Tucker trawl with 505 |xm mesh was fished for 10 min- utes at 35 m to investigate escapement from the 20 cm bongo nets. Also during the second 24-h period a 60 cm bongo net with 505 jxm mesh was fished about 2 m below the deepest 20 cm bongo net to stabilize the wire and allow further catch comparisons. Expendable bathythermograph (XBT) casts were done at each survey grid station and at the 1430 and 0230 vertical distribution study sta- tions. Conductivity-temperature-depth (CTD) casts (Ocean Data Equipment Corporation'^ Model 302) were made at 15 of the survey grid stations selected to provide three sections across the major southwesterly setting flow field in the area (Fig. 2). "Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 501 FISHERY BULLETIN: VOL. 85, NO. 3 KODIAK ISLAND - 57 OON - 56 OON SAMPLING ACTIVITIES (Station numbers to right of activity symbols) -f- 60 cm bongo, neuston. XBT ^ 60 cm bongo, neuston, CTD ^ 60 cm bongo, neuston, XBT otolith analysis B Vertical/diel stations 60 cm bongo, neuston, XBT, series of 20 cm bongos. Tucker trawls LARVAL POLLOCK ABUNDANCE Number/10 m' 0-100 100-1000 1000 55 OON 159 OOW 158 OOW 157 OOW 156 OOW 155 OOW 154 OOW Figure 2. — Distribution, abundance (numbers per 10 m2) and lengths of walleye pollock larvae from 60 cm MARMAP bongo tows superimposed on sampling pattern used in the northern Gulf of Alaska, May 1983. Hydrographic sections are labelled A, B, C (see Figure 4). Laboratory Procedures Fish eggs and larvae were identified to the low- est taxon possible at the Polish Plankton Sorting Center in Szczecin, Poland. Fish larvae were measured to the nearst 0.1 mm standard length (SL); when more than 50 larvae of a taxon C':- curred in a sample, a random subsample of 50 was selected for measurement. Identifications were verified at the Northwest and Alaska Fisheries Center. For distributional analysis of fish eggs and larvae from the survey, numbers per tow for each taxon were converted to numbers under 10 m^ of water surface using volumes of water fil- tered and maximum tow depth (see Smith and Richardson 1977). To compare relative abun- dances of various taxa, an estimate of the total number of eggs or larvae of each taxon present in the entire survey area was derived by summing the catches at each station and multiplying by the 502 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK area of sea surface represented by that station (the Sette-Ahlstrom method, see Smith and Richardson 1977). This estimate was thought to make best use of all the available data. A chi-square test analyzed differences in the numbers of walleye pollock larvae caught as a function of time of day and depth at the diel sta- tion. For this test, numbers of larvae in the two sides of the 20 cm bongo nets were combined. Also, catches at the same time and depth but on different days were combined when complete depth series were collected. Four out of six times two complete depth series were collected; at two times only one complete depth series was col- lected. Zooplankton were sorted, identified, and enu- merated from subsamples of collections made with the 253 ixm mesh net. The subsample was chosen such that at least 500 organisms were sorted from each sample. For larval feeding analysis, 20 walleye pollock larvae (or the total sample when <20 were caught) were selected to represent the size range in the total sample from each of the 333 fxm mesh, 20 cm bongo net samples. The guts were dissected from the larvae, and all food items in the foregut, midgut, and hindgut were teased out, identified, and counted. Lengths and greatest widths were measured for all food items in the larvae collected at 0630, 29 May. Lengths used were carapace length for cope- pod nauplii, metasome length for copepodids, and total length for all other prey. These measure- ments were used to estimate volumes of prey or- ganisms, which were applied to the rest of the samples. Mensuration formulae were used to cal- culate the volume of copepod eggs, copepod nau- plii, copepodids of Pseudocalanus spp., Acartia spp., and Oithona spp. (Nishiyama and Hirano 1983; Table 1). Pseudocalanus spp. mensuration formulae were used to estimate the volumes of unidentified copepodids. The volumes of other food items were not estimated since their low abundance did not allow adequate measurement of body proportions. Samples used for age and growth analysis were selected from one station within the area of highest larval density (Station 37, Fig. 2), and from four stations located outside of this dense patch. Standard lengths of larvae from as broad a size range as possible within each sample were measured to the nearest 0.1 mm using an ocular micrometer. Both sagittal otoliths were removed and cleaned using a pair of fine needles under a dissecting microscope fitted with polarizing fil- ters. Whole otoliths were affixed to microscope slides with clear histological mounting medium and increments read in the sagittal plane under a compound microscope with transmitted light at 1000 X magnification. Most of the otoliths had a distinct distal-proximal curvature and readabil- ity was enhanced when the otolith was mounted with the concave side up. Increments were identified as a pair of adjacent light and dark bands, formed concentrically around the focus. A prominent dark band sur- rounding the focus was observed on each otolith (Fig. 3). Since mean otolith diameter at this band (16.0 ± 0.13 |jLm SE) was similar to the diameter of otoliths from 1-day-old, laboratory-reared lar- val pollock (18.97 ± 0.37 ^JLm, Nishimura and Ya- mada 1984; 16-20 jxm, Walline 1983; 15.3 ± 1.2 Table 1. — Mensuration formulae (Nishiyama and Hirano 1983),^ length to width ratios, metasome to whole body ratios, metasomal lengths, mean lengths, and mean diameters used to calculate volumes of copepodids, copepod nauplii, and copepod eggs in guts of larval walleye pollock in Shelikof Strait. K m Lm Metasome: whole body Metasomal Length:width (Nishiyama and length Species (this study) Hirano 1983) (this study) Pseudocalanus spp. 2.239 0.97 0.870 Oithona spp. 1.824 0.93 0.189 Acartia spp. 3.213 0.95 0.916 Unidentified copepodids 2.539 0.97 0.870 'Mensuration formulae (Nishiyama and Hirano 1983): Volume of copepodids = [Lm/6(Lm/K)2TT]/m. Volume of copepod nauplii = (^ carapace length^ j; mean carapace length = 0.187. Volume of copepod eggs = (? egg diameters J; mean egg diameter = 0.1 10. 503 FISHERY BULLETIN: VOL. 85, NO. 3 |jLm, Bailey and Stehr^), this band was presumed to be the hatching check and was counted as the first increment. Nishimura and Yamada (1984) found that increments were formed daily, begin- ning with the day of hatching, on the otoliths of laboratory-reared larval pollock. Similar obser- vations have been made on larval walleye pollock otoliths viewed with both light and scanning elec- tron microscopy by Bailey and Stehr (fn. 8). Incre- ments were, therefore, considered to be deposited daily and increment counts were equated with the age of the fish in days after hatching. The mean of two independent increment counts was used in growth rate analysis. Age-at-length data from each station were fitted separately with simple linear regressions. Analysis of covariance 8K. Bailey and C. Stehr, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way, N.E., Seattle, WA 98115, pers. commun. 7 February 1985. was used to compare growth rates and Dunnett's test of multiple comparisons identified signifi- cantly different rates (Zar 1974). RESULTS Hydrographic Observations Temperature in the survey area varied from just above 7°C at the surface at some stations to slightly <5°C in deeper shelf water (Fig. 4). Tem- perature gradually decreased with depth, and temperatures in the upper 50 m (where most of the walleye pollock larvae were found) were gen- erally between 7.0° and 5.5°C. At the diel-vertical distribution station, temperatures were similar to those found throughout the area, although the temperature gradient was more uniform than elsewhere. In the upper 60 m at this station, tem- FlGURE 3.— Otolith from a 11.57 mm SL walleye pollock larva showing 27 daily growth increments. The arrow near the focus indicates the first increment. Scale bar indicates 20 nm. 504 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK STATION NUMBERS Section A 0 50 100 -: 150 -; 200 ; 250 300 35 17 11 6.0- 32.0 6.5 -V- 5.5 33 0 8 35 17 11 25.0 25.5 si 26.0' Section B 49 47 39 50 s 100 H -B 150 a. Q 200 250 i- 6.5 6.0' 32.0 30 22 r 300 f.' ■...■■ ■ ■■. 49 47 39 25.0 30 22 26.0 26.5 Section C 27 25 42 Temperature (C), Salinity (%o) 25.5 Density (o, ) 27 25 27.0 Figure 4. — Temperature (solid contours), salinity (broken contours), and density observed in three sections across the sampling area. May 1983. Station and section locations are shown in Figure 2. 505 FISHERY BULLETIN: VOL. 85, NO. 3 perature steadily decreased from 6.9°C to 6.0°C, and at the bottom (120 m) the temperature dropped to 5.5°C. Among the four XBT casts (taken at 12-h intervals), the temperature at a given depth varied as much as 0.5°C: an isotherm depth varied vertically by as much as 50 m. Salinity varied from 31.5 to >33.5%<- in the sur- vey area. Lowest salinities were found at the sur- face toward the Alaska Peninsula, and high salin- ities were found offshore in deeper waters. Isohalines generally sloped from offshore to in- shore. This slope was most pronounced at Section A, the one closest to Shelikof Strait. Most larvae were in water between 31.5 and 32.2%f. The salin- ity profile at the CTD station closest to the diel station showed a slight and steady increase in salinity with depth starting from a surface value of 31.8%(: and ending with a bottom (142 m) value of32.1%o. Density sections (ct^) show the same sloping pattern as the salinity sections but are even more pronounced (Fig. 4). Values ranged from <25.0 at the surface near the Alaska Peninsula to >26.4 in deeper waters near the edge of the continental shelf. No sharp pycnocline was observed but rather a gradual increase in density with depth and distance from the Alaska Peninsula. Most walleye pollock larvae were in water with densi- ties between 25.0 and 25.4 a^. The density profile observed near the diel station closely paralleled the salinity profile, with a gradual increase with depth from a, = 24.9 at the surface to ct< = 25.4 at the bottom (142 m). Relative Abundance of Eggs and Larvae Neuston tows and bongo tows captured eggs of 13 and 14 taxa, respectively (Fig. 5). Rank orders of abundance, based on estimated total numbers offish eggs in the neuston catches showed Micro- stomus pacificus (Dover sole) to be in greatest abundance, followed by Glyptocephalus zachirus (rex sole) and Theragra chalcogramma. In bongo catches, unidentified pleuronectid (righteye flounders) eggs were most abundant, followed by those of M. pacificus, G. zachirus, and T. chalco- gramma. Larvae of 29 and 42 taxa were identified in neuston and bongo catches. Rank order of abun- dance of fish larvae in neuston tows, based on estimated total numbers, showed Ammodytes hexapterus (Pacific sand lance) to be most abun- dant followed by Hexagrammos decagrammus (kelp greenling), Lyconectes aleutensis (dwarf wrymouth), Bathymaster spp. (ronquils), and T. chalcogramma. In bongo catches T. chalco- gramma larvae were most abundant followed by those of Bathymaster spp., A. hexapterus, Hip- poglossoides elassodon (flathead sole), and unidentified gadids (codfishes). Distribution and Abundance of Walleye Pollock Eggs and Larvae Eggs of walleye pollock were taken in 26% of the bongo tows and in 27% of the neuston tows but in low abundance. Only 262 eggs were collected. Some early stage eggs were collected, indicating recent spawning, but older eggs were also present. Eggs were found mainly in water over the deeper part of Shelikof Strait, with decreasing abundance to the southwest (Fig. 6). Larvae of walleye pollock were found in 89% of the bongo catches and 24% of the neuston catches. The center of larval concentration was near the middle of the survey pattern (Fig. 2). Mean stan- dard length of the larvae throughout the survey was 10.63 mm (range 3.8-21.3 mm, SD = 1.81 mm), with no differences in mean length by area. At each of five stations near Sutwik Island and the Semidi Islands, more than 1,000 larvae/10 m^ were encountered. At 44 of the 64 stations, more than 100 larvae/10 m^ were found. A total of over 10^^ larvae was estimated to be present in the survey area. Vertical Distribution of Walleye Pollock Larvae In preliminary tows with the 20 cm bongo nets most larvae were caught above 60 m. During the vertical distribution study actual depths of sam- pling based on bathykymograph records covered the ranges of 14-20, 21-28, 28-38, and 39-47 m (Table 2). The mean length of the larvae during our diel vertical distribution study was 11.1 mm SL. The range of mean lengths among the individual sam- ples was 10.0-12.2 mm SL, and the range of stand- ard deviations was 0.8-2.3 among hauls with more than 10 larvae. No patterns of size of larvae with depth or time of day were seen by visual inspection of the data, and since the range of mean lengths was so narrow, and the confidence intervals overlapped, no further analysis was per- formed. There were no diel differences in catch rates 506 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK NEUSTON BONGO LOG OF NUMBER IN SURVEY AREA CD 5 L 6 _i_ MlCROSIOnUS PACIFICUS CLIPIOCEPMALUS ZACHIRUS IHERAORO CH»I.C0(!R*ni1A HIPPOGIOSSOIDES eiASSOOON PLEURONECIIOAE lELEOSI ItPE A CNIROLOPHIS NUSATOR DISIN1ECRAIE0 UNIOENItFIED PAROPHRIS VEIUIUS PLEURONECIES OUAOR I IU8ERCUL AlUS ISOPSEIIA ISOLEPIS PLAIlCHtMIS SIELIAIUS 8 9 10 8 PLEUROHECTIOAE nicROSioNus PACirtcus CLtPIOCEPHALUS ZACHIRUS IHERACRA CHALCOCRArWA HIPPOGLOSSOIDES ELASSOOON MACROURIOAE ISOPSEtlA ISOLEPIS lELEOSI ITPE A OISINTECRATEO CHIROLOPHIS MUCAIOR EMSASSICHIHrS BAtHYBlUS PSEIIICHIMTS MELAHOSI ictus PLEURONECIES OUADRI TUBERCUL ATUS littlDENIlFlEO 10 11 12 13 _i I I I 3 LU < > < 6 _L 8 10 8 9 10 11 12 13 AMMOOTIES HE«APIERUS HEXAORAnnOS OECACRAriMUS LtCONECIES ALEUIEHSIS BATHTf1ASI£R SPP. IHERACRA CHALCOCRArWIA HEMILEPIDOIUS HEnUEPlDOIUS AROPLOPOMA FlnSRIA riTOXOCEPMALUS SPP. ZAPRORA SUCNUS SE8ASIES SPP. HEM[LEP100IUS JOROAMl HALIOIUS VILLOSUS SIICHAEIOAE HEXACRAMMOS SIELLERI HIPPOCLOSSUS SIENOLEPIS DELCLEPIS ClCANtEA CADUS MACROCEPHALUS CAOIDAE HEMILEPIDOIUS SPINOSUS nrOXOCEPHALUS B DISINUCRATED HIPPOGLOSSOIDES ELASSOOON AGON I DAE COIIIDAE CtClOPTERIDAE NANSENIA CANDIDA PI EUHOGRAHMUS HONOP T f RT C I US LtPIOOPSEIIA BILINEAIA HEXACRAHnOS OCIOGRAHnUS IHERACRA CHALCOGRAHflA BAIHTMASIER SPP. AMHOOTIES HEXAPIERUS HIPPOGLOSSOIDES ELASSOOON GAD I DAE SEBASTES SPP. ANOPLARCHUS SPP. CADUS MACROCEPHALUS COITIOAE LEPIDOPSETIA 81LIKEATA SIENOBRACHIUS LEUCOPSARUS AGON] DAE CICLOPIERIDAE LEUflOCLDSSUS SCMMIOII LUMPENUS MACULATUS PDROCLINUS ROIHROCKI AIHERESIHES SIOnlAS OLIPIOCEPHALUS ZACHIRUS PLAIICHTHYS SIELLAIUS LUMPEHELLA LONCIROSIRIS PSEITICHTHTS MELAHOSIICIUS DISINIEGRAIED LTCONECIES ALEUIENSIS PROTOMTCIOPHUr IHOMPSONl ZAPRORA SIlENUS PHOLIS SPP. ARIEDIUS MARRINGIONI HEXAGRAMMOS OECACRAMMUS PLEURONECIES QUADRl IU6EBCUL AIUS HIPPOCLOSSUS SIENOLEPIS MALLOIUS VILLOSUS BAIHTLAGUS PACIFICUS IRIGLOPS MACELLUS ANOPLOPOMA FIMBRIA ARIEDIUS MEANT 1 RADULINUS SPP. OPHIODON ELONCAtUS AIHERESIHES SPP. SIENOBRACHIUS SPP. SIICHAEIOAE OASTCOIIUS SEIICER BTOXOCEPHALUS SPP. Figure 5. — Rank order abundance, as log of total numbers in the survey area, offish eggs and larvae in neuston and bongo tows during the survey, May 1983. Table 2. — Chi-square test of numbers of walleye pollock larvae with time and depth from the vertical distribution study, May 1983. Time 1840-1920 2235 0320-0238 0629-0635 No. samples 1047 1432-1515 Depth (m) 16 Obs. Exp. 8 Obs. Exp. 16 16 Obs. Exp. Obs. Exp. 8 Obs. Exp. 16 Obs. Exp. 14-19 21-28 28-38 39-47 76 76 110 136 (94) (122) (95) (87) 98 138 11 83 (78) (101) (79) (72) 133 202 201 140 (159) 100 (207) 68 (161) 46 (148) 35 (59) (76) (59) (55) 61 108 87 33 (68) (89) (69) (63) 37 (47) 65 (62) 56 (48) 43 (44) Total 398 330 676 249 289 201 X2 52.19 78.89 14.73 39.46 23.77 3.63 Total x2 = = 212.69 (15 df. P < 0.005) 507 FISHERY BULLETIN: VOL. 85, NO. 3 KODIAK ISLAND CHIRIKOF I, POLLOCK EGG ABUNDANCE Number/10 m I 4* I 1-100 ^IP >100 ^ ^ ^ 57 OON - 56 OON - 55 OON 159 OOW 158 OOW 157 OOW 156 OOW 155 OOW 154 OOW Figure 6. — Distribution and abundance of walleye pollock eggs, May 1983. (Table 2). The means of the catch per 10 m-^ at each time-depth combination expressed as a per- centage of the total catch at that time of day sug- gests a pattern of limited diel vertical migration (Fig. 7). The chi-square test was highly signifi- cant iP < 0.005) indicating that the null hypothe- sis, that the larvae were distributed at each depth in the same proportions among the different times, should be rejected. Examining the relative abundances within each time period, the larvae appeared to be concentrated above 20 m at 0630 h, and at 28-47 m by 1830 h (Fig. 7). They were most evenly distributed in the early afternoon and most abundant in the 21-28 m stratum during darkness (2230 and 0230 h) and at 1030 h. The lowest percent abundance at each time period shows a complementary pattern, with relatively small catches at 39-47 m from 0230-1430 h. This pattern was observed on both days during the 48-h sampling. In summary, it appears that some larvae grad- ually move up in the water column from a depth of 30-50 m in the evening to above 20 m in early morning. They gradually descend during day- time, and are most evenly distributed in the early afternoon. 508 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK PERCENTAGE OF LARVAE AT EACH DEPTH 10 20 30 Q. Q 40 1830 2230 Time 0230 0630 1030 1430 • 50 I I — I — I — I • Actual sampling depth q 20 40 Percent scale Figure 7. — Relative abundance as percent of larvae at each depth for each time interval, of walleye pollock larvae from the vertical distribution study, May 1983. Comparison of Catches of Walleye Pollock Larvae by Different Gears The sizes of larvae in the different bongo nets at the vertical distribution station were similar (Table 3). Mean lengths of larvae in the 20 cm bongo nets varied from 11.00 in the 333 ixm mesh net to 11.10 mm in the 253 fxm mesh net (SD = 1.76 and 1.69 mm respectively); mean length in the 60 cm bongo nets was 11.07 mm iSD = 1.77 mm). The Tucker trawl, however, caught larvae that had a mean length of 9.64 mm (SD = 1.67 mm). The overall mean abundance of larvae in the 20 cm bongo in the 28-38 m depth stratum nets (11.82 larvae/10 m-^) was similar to that in all the Tucker trawls (11.66 larvae/10 m^) which were towed at 35 m. The mean of the catches in the 60 cm bongo nets, towed just below the deepest 20 cm bongo, was not notably differ- ent from the mean of those 20 cm bongo catches taken at the same times. Variations in overall catches in the 20 cm bongo nets at the vertical distribution stations seemed to reflect the patchy nature of the concen- tration of larvae and not net avoidance related to time of day. The largest catches occurred during daylight, at 1030 h, while the smallest catches occurred during the time intervals immediately preceding (0630 h) and immediately following (1430 h) the largest catches. Since we sampled one geographic site rather than following a drogue, we probably sampled water with different concentrations of larvae as it drifted past us dur- ing the 48-h sampling. It appears that the larvae decreased from a concentration greater than 1 m '^ during the first 24 hours to <0.5 m "^ dur- ing the second 24 hours. The size of larvae did not change during the study again indicating that increased daytime net avoidance was not signifi- cant. Hydrography in Relation to Distribution of Walleye Pollock Larvae No obvious hydrographic features were associ- ated with larval distributions. At the diel station, larvae were concentrated between 14 and 47 m where temperature within the upper 50 m was 509 FISHERY BULLETIN: VOL. 85, NO. 3 Table 3. — Comparisons of catches and lengths (mm SL) of walleye pollock larvae with gear, time of day, and depth from the vertical distribution study, May 1983. mm SL (SD) number mm SL (SD) number Depth (m) 14-19 333 fjim mesh 253 |jim mesh 11.13 (1.80) 333 10.86 (1.77) 333 21-28 11.07 (1.74) 405 11.07 (1.87) 427 28-38 11.24 (1.68) 232 11.04 (1.73) 290 39-47 11.19 (1.60) 249 11.17 (1.46) 229 Time (local) 1830 11.03 (1.67) 210 10.86 (1.65) 118 2230 10.98 (1.58) 189 10.76 (1.56) 185 0230 11.50 (1.83) 315 11.39 (1.75) 361 0630 11.12 (1.42) 133 11.31 (1.82) 116 1030 11.01 (1.75) 331 10.61 (1.81) 301 1430 10.73 (1.57) 183 11.13 (1.72) 182 Gear 20 cm bongos, 253 M.m mesh 11.10 (1.69) 1,388 20 cm bongos, 333 (xm mesh 11.00 (1.76) 1,381 All 60 cm bongos 11.07 (1.77) 578 All Tucker trawls 9.64 (1.67) 2,180 Stable at 6.90°-6.15°C (Fig. 8). Nearby, salinity and density showed a very gradual increase with depth. The pattern of water movement in the survey area, derived from the temperature and salinity observations, indicated a general southwest flow of water at all depths. In the area of larval concen- tration, most of the flow tended southward, fol- lowing the deep trough from Shelikof Strait across the continental shelf between the Semidi Islands and Cherikof Islands (S. Kim^). Diet and Feeding of Walleye Pollock Larvae Eighteen different food items were identified in walleye pollock larvae sampled at the diel sta- 9S. Kim, School of Fisheries, University of Washington, Seat- tle, WA 98195, pers. commun. September 1985. 21 22 23 24 25 26 27 1 29 r I 30 ■■"T r 31 1 r 32 33 34 I 35 1 25 50 75 100 125 150 Temperature - Salinity Temperature at diel station Density Density (a ) Salinity (%o) Temperature (C) Figure 8. — Temperature, salinity, and sigma-? profiles at Station 47 with temperature profile from the diel station (Stations 63-77) shown for compari- son. See Figure 2 for station locations. 510 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK tions. Copepod nauplii 90-600 |xm in length and between 40 and 250 jxm in width were the most abundant food item for larvae <14.0 mm (Fig. 9). The length- and width-frequency distributions of nauplii indicate that Pseudocalanus spp. and Oithona spp. are the likely prey taxa. Nauplii became numerically less important in diets of larger larvae. Pseudocalanus spp. copepodids, copepod eggs, and Oithona spp. copepodids made up a larger numerical portion of the food of these larger larvae (Fig. 10). The increased importance of Pseudocalanus spp. copepodids in the diet of larger larvae is evident in terms of gut volume (Fig. 10). Examination of the contribution of the copepod eggs to the diet indicates that while eggs increase numerically from 1.5% in the 8.0-8.9 mm size group to 55% in the 14.0-14.9 mm size group, their volume increases only from 0.5 to 3.6%. The number of food organisms and mean prey volume per larva show a diel feeding pattern (Fig. 11). The maximum gut contents were observed during the afternoon. Collections at 2235, 0230, 0238, and 0629 have the lowest mean number of food items per larva, which suggests that feeding during darkness is reduced. The mean gut vol- umes show a similar pattern (Fig. 11). Although gut volume at 1047 h on 29 May was unusually low, it was not an indication of reduced feeding. Here the gut contents consisted of numerous cope- pod nauplii and few adult copepods. Thus the cal- culated prey volume is quite low, while the num- ber of food items per gut is high. Zooplankton Composition Thirty-nine groups of zooplankters were identi- 500 400 300 200 100 I r I I _1_ _1_ N = 1191 700 500 300 100 80-110 120-150 160-190 200-230 240-270 280-310 320-350 360-600 Length of nauplii (/jm) B _L N= 1181 30-50 60-80 90-110 120-140 150-170 180-200 210-230 240-260 Width of nauplii (*jnn) Figure 9.— Length (Al and width (B) size-frequency distribution of copepod nauplii from guts of walleye pollock larvae. 511 E C (J Q- E o > CD O \_ Q. N =1 10 100 80 60 40 20 0 100 80 - 60 - 40 - 20 - 0 FISHERY BULLETIN: VOL. 85, NO. 3 9 39 96 176 184 177 56 35 13 2 2 0 1 -\-^ iA L.l t.l Ll. 11 \"' V-^Y> ^ Ll. T7 '■■M t..l I ^ rv^ F^ rm SS^ m: ■.■ < L*lit ■ 1 A ^ B 10 15 20 Size (mm) -/V- Unidentified copepodids '■; Copepod nauplii pH^^ Pseudocalanus spp. copepodids 0/f/70/7a spp. copepodids Copepod eggs 1 1 1 1 1 Acartia spp. copepodids Other food items Figure 10. — Gut contents of walleye pollock larvae by 1 mm length intervals (5-20 mm) from northern Gulf of Alaska, May 1983. A — percent number; B — percent volume. 512 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK O O J2 E 1200 C3 0.600 r 0.500 0.400 1200 Figure 11. — Numbers of prey (A) and gut fullness (B) of walleye pollock larvae by time of day and depth from the vertical distribution stations, May 1983. fied from samples taken with the 253 [xm mesh net (Table 4). Pseudocalanus spp. were generally the most abundant taxon. Abundances of Pseudo- calanus females ranged from 11 m~'Ho a peak of 1,398 m'^ at 1515 h on the first day at a depth of 47 m. The overall mean abundance of Pseudo- calanus females was 224 m"-^. These females com- prised between 9 and 23*^ of the total zooplankton at all depth strata both during the daylight and darkness (Fig. 12). The highest percent contribu- tion by this stage (22-23%) was during daylight hours at the two deepest strata. Peak abundances usually occurred at depth strata below 25 m. Copepodid stages of Pseudocalanus also reached a peak of 1,890 m"-^ at the same station. These copepodids contributed the greatest percentage (21-33%) of the total zooplankton at all depths and times. There were no obvious patterns associ- ated with depth or time of day in either abun- dance or percent composition of the total zooplankton. Oithona spp. were abundant in the samples even though they were not collected quantita- tively in the 253 jxm mesh net. The peak abun- dance observed was 1,323 m ■^. Oithona spp. were most abundant in the surface stratum comprising 18% of the total zooplankton during the day ver- sus <10% of the zooplankton at deeper strata dur- ing the day and at all depths during the night (Fig. 12). Acartia spp., Neocalanus spp., and Calanus spp. were the only other abundant cope- pods. The Neocalanus spp. and Calanus spp. in- 513 FISHERY BULLETIN: VOL. 85, NO. 3 Table 4. — Species composition of zooplankton samples from the vertical distribution stations. Taxon Number (See Fig. 11) Name Maximum No./IO m3 Minimum No./IO m3 Pseudocalanus spp. adult female Pseudocalanus spp. adult male Pseudocalanus spp. copepodids 1-5 Neocalanus spp. and Calanus spp. Oithona spp. Acartia spp. Centropages spp. Metridia spp. Eucalanus bungli Paracalanus sp. Clausocalanus sp. Torianus spp. Unidentified calanoid Euphausiid furcilia Euphauslid calyptopsis Euphausiid crytopsis Thysanoessa Inermis furcilia T. inermis calyptopsis T. inermis crytopsis Euphausiid juveniles T. inermis juvenile Chaetognatha Appendicularia Hypendae Gastropoda Decapoda Balanidae Gammaridae Evadne spp. Limacina spp. Pontellidae Thecosomata Brachyura Gymnocomata Echinodermata Hydrozoa Pelecypoda Medusa Siphonophore Mean No./IO m3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 13,981 118 2.245 4,711 0 567 18,915 46 3,592 12,163 0 1,545 13,234 0 1,167 2,476 101 794 380 0 0 0 50 192 7 46 0 5 25 0 1 7 0 — 201 0 9 3,431 0 511 97 0 0 0 0 0 0 0 7 1,307 517 97 5 73 5 592 0 94 1,088 0 290 183 0 31 484 0 62 262 0 24 253 0 39 7 0 — 97 0 3 1,569 0 364 19 0 — 255 0 21 292 0 34 38 0 1 1,511 0 117 311 0 18 28 0 1 14 0 — 8 0 — eluded N. plumchrus , C. pacificus, and C. mar- shal lae. Together these species contributed be- tween 13 and 259r of the total zooplankton (Fig. 12). Euphausiids were the only other abundant group of zooplankters; furcillae were the most abundant stage. The abundance of furcillae was generally <200 m ■^; however, at 1047 h on 29 May, the numbers of furcillae exceeded 1,300 m -3 Age and Growth Ages were determined for 109 walleye pollock larvae, including 40 individuals collected from Station 37 within the area of highest larval abun- dance and 69 specimens from four stations well removed from this area. Standard lengths ranged from 6.0 to 14.6 mm and mean increment counts (days since hatching) ranged from 7 to 45.5 (Table 5). The average gi'owth rate varied from 0.12 to 0.25 mm/day among the five stations (Table 5). When compared pairwise with growth rates from all other stations, growth rates from all stations located outside the dense patch, ex- Table 5. — Average growth rates, statistics on the linear regression, comparison lengths, and ages of larval walleye pollock used in growth analysis from five areas in the Shelikof Strait. Station 37 is in the area of highest larval density (Fig. 1). No. Y- of Length Age intercept b-growth fish range range Station (mm) (mm d ') r^ N aged (mm) (days) 37 37 0.24 0.82 40 40 6.8-13.1 12-42 53 4.2 0.23 0.81 20 20 7.6-13.0 20-41 9 3.5 0.25 0.81 20 20 6.0-14.6 7-43 6 4.7 0.18 0.61 20 20 6.6-12.3 10.6-42.2 12 6.6 0.12 0.90 9 9 8.1-12.6 14-45.5 514 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK DAY 35 30 - 25 - 20 15 10 5 0 NIGHT 20 m LL J r-ll L-M-n->r- n L iCUn-, 20 m 1__ C o \— a. 35 30 25 ' 20 - 15 I- 10 5 0 30 m ■J I .. ■■■ . n dfc_- m nhlmr-^^m 30 m JL JLjl^ J nm-r-i. 40 m ■ xl 1 . •*■ -n 1 II 40 m JL -ID LL-OL- 35 30 25 20 15 10 5 0 .11 Lm n . _nnw: 50 m ■ n - - . XO nmnn^-*. 50 m -.11 » n_,- 13 17 21 25 29 33 37 15 Taxon number 13 17 21 25 29 33 37 Figure 12. — Zooplankton distribution (percent total zooplankton) for day and night by depth from the vertical distribution stations, May 1983. Taxon numbers: 1: Pseudocalanus spp., adult female; 2: Pseudocalanus spp., adult male; 3: Pseudocalanus spp., copepodids 1-5; 4: A'^eoca/ani/s spp. and Ca/anMS spp.; 5: 0(7/io^a spp.; 6: Acartja spp.; 7: Centropages spp.; 8: Metridia spp.; 9: Eucalanus bungii; 10: Paracalanus sp.; 11: Clausocalanus sp.; 12: Tortanus spp.; 13: Unidentified calanoid; 14; Euphausiid furcilia; 15: Euphausiid calyptopsis; 16: Euphausiid crytopsis; 17: Thysanoessa inermis furcilia; 18: T. inermis calyptop- sis; 19: r. i>!ermi.s crytopsis; 20: Euphausiid juveniles; 21: T. (nerm/s juvenile; 22: Chaetognatha; 23: Appendicularia; 24: Hyperidae; 25: Gastropoda; 26: Decapoda; 27: Balanidae; 28: Gammaridae; 29: Evadne spp.; 30: Limacina spp.; 31:Pontel- lidae; 32: Thecosomata; 33: Brachyura; 34: Gymnosomata; 35: Echinodermata; 36: Hydrozoa; and 37: Pelecypoda. 515 FISHERY BULLETIN: VOL. 85, NO. 3 cept Station 12, were not statistically different from growth measured within the dense patch (Station 37). Larvae collected at adjacent Stations 6 and 12, located to the northeast of the densest area (Fig. 2), exhibited the lowest growth rates and were not statistically different from each other. Estimated age-at-length data from all stations were combined to describe early growth in walleye pollock from the Shelikof Strait region as follows: SL = 4.29 + 0.21 (age, d) n = 109, r^ = 0.75 where SL = standard length (Fig. 13). This rela- tionship suggests a mean growth rate of 0.21 mm/ day and an intercept of 4.29 mm, which corre- sponds with the known size of newly hatched larvae (Walline 1983; Nishimura and Yamada 1984). The distribution of dates of hatching was esti- mated by back-calculating from the ages deter- mined for larval walleye pollock in the Shelikof Strait. The median birthdates from all stations were similar and thus all 109 samples were com- bined (Fig. 14). The hatching period ranged from early April to mid-May with a mode in the last 15 r 10 T3 C 10 20 30 40 50 Age (days) Figure 13. — Estimated age at length, fitted with hnear regression, for all walleye pollock larvae analyzed from the northern Gulf of Alaska, May 1983. lOr o n 10 15 20 April 25 30 10 ■ May 15 20 Hatching date Figure 14. — Distribution of hatching dates for walleye pollock determined by back- calculation using age and date of collection. 516 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK week of April. Incubation time for walleye pollock eggs held at 5°-6°C in the laboratory is estimated to be 14 days (Haynes and Ignell 1983; Nakatani and Maeda 1984; Nishimura and Yamada 1984); spawning of walleye pollock occurs primarily in late March and early April in Shelikof Strait (Dunn et al. fn. 5), supporting our estimated hatching dates distribution. DISCUSSION Relative Abundance of Eggs and Larvae Neuston collections in spring in the northern Gulf of Alaska have been reported only for 1978, mainly over the shelf, south and east of Kodiak Island, a month earlier than the present study (Kendall and Dunn 1985). Eggs of Theragra chalcogramma were most abundant; the rest of the identified eggs were of several pleuronectids. The greater abundance of T. chalcogramma in relation to the pleuronectids in the earlier cruise, when compared with the present data (Fig. 5), probably reflects the seasonal difference in spawning times. Pleuronectids spawn mainly in late spring and early summer in the Gulf of Alaska while T. chalcogramma is mainly a late winter-early spring spawner (Kendall and Dunn 1985). Eggs have been reported from bongo catches from three other cruises in the northern Gulf of Alaska in May (Bates and Clark 1983^0. Kendall and Dunn 1985); eggs of pleuronectids and those of T. chalcogramma were most abundant. How- ever, the rank order of abundance of the various pleuronectids varied considerably among the cruises. Usually eggs of Glyptocephalus zachirus, Hippoglossoides elassodon, and Microstomas pacificus were among the five most abundant taxa. Among larvae in the neuston tows during the present study, mainly spring spawning taxa (e.g., Ammodytes hexapterus, Bathymaster spp., T. chalcogramma) were represented; whereas lar- vae of fall-winter spawning taxa (e.g., three hexa- grammids, and Hemilepidotus spp. [Irish lords]) were abundant during the earlier cruise (Kendall and Dunn 1985). In bongo catches during other May cruises in lOBates, R. D.,andJ.Clark. 1983. Ichthyoplankton off Ko- diak Island and the Alaska Peninsula during spring 1981. U.S. Dep. Commer., Natl. Mar. Fish. Serv., NOAA, NWAFC Proc. Rep. 83-09, 105 p. the area, larvae of T. chalcogramma, Hippoglos- soides elassodon, Bathymaster spp., and A. hexapterus have always occurred frequently, as in the present study (Bates and Clark fn. 9; Kendall and Dunn 1985). Distribution of Walleye Pollock Eggs The few eggs collected during the present study represented a very late part of the spawning, which occurs in Shelikof Strait mainly in early April (Dunn et al. fn. 5) (Fig. 1). Recently hatched larvae (<5 mm SL), which were collected during our survey (Fig. 2), also indicate prolonged spawning but probably at a low level after mid- April. The eggs we found were mainly over the deep waters at the southwest end of Shelikof Strait, and it would be expected that they were also farther to the northeast. This is the same area of occurrence of eggs during the height of spawning (Dunn et al. fn. 5), indicating that the adults spawn mainly in Shelikof Strait through- out this period, although individual spawning fish probably migrate in and out of the area. Comparisons of Distribution of Walleye Pollock Larvae In 1981, several sequential cruises to Shelikof Strait mapped a large concentration of walleye pollock eggs in early April; and in late April and again in mid-May, a concentration of larvae was found progressively further to the southwest of the area where the eggs had been (Bates and Clark fn. 10) (Fig. 1). The size of the larvae in the concentration increased between the cruises. Sampling in 1981 and 1982 for walleye pollock larvae was at the same area and time (24-28 May) as the present study (Dunn et. al. fn. 5). Compari- sons of distribution, abundance, and size of the larvae among these 3 years reveal remarkable differences (Table 6). Spawning time in 1981, based on ages of eggs caught in early April, and presence of newly hatched larvae in late April, centered around 5-8 April. In 1983, based on birthdate distributions presented here, modal spawning time was also in the second week of April. Sampling in subsequent years has shown a remarkable consistency in spawning place and time (Kendall unpubl. data). By 24-28 May the patch of larvae in 1981 and 1983 had drifted to the same area, just north of Sutwik Island (Figs. 517 FISHERY BULLETIN: VOL. 85, NO. 3 Table 6. — Concentrations and lengths of pollock larvae collected 24-28 May 1981, 1982, and 1983 In the patch resulting from the Shelikof Strait spawning (1981 and 1982 data from Dunn et al. text fn. 5). Lengths Date (mm SL) Year Cruise (f\^ay) Stations Larvae m 2 Mean SD 1981 4MF81 24 73, 74 2,318 7.36 1.11 3SH81 25 225 1,285 (mean = 2,040) 7.77 1.18 1982 2DA82 24-28 108, 115, 117, 123, 124, 126 14-38 (mean = 23) 7.74 1,10 1983 1CH83 24, 28 31, 32, 36, 37. 63 104-214 (mean = 151) 11.23 1.65 1, 2). However, the abundance of larvae in 1981 was about 2,000 m^ while in 1983 it was only about 150 m"". Mean length in 1981 was about 7.5 mm while in 1983 it was 11.2 mm. Since the spawning dates were the same, this would indi- cate a much slower growth rate (about 0.09 mm d-^) in 1981 than in 1983 (0.21 mm dM. In 1982 there were fewer larvae (20 m"'^) and they were distributed further southwest than in the other 2 years (Dunn et al. fn. 5). These larvae were not different in length from those in 1981 (7.7 mm). The position of the larvae in 1982 suggests a much faster drift than in 1981 and 1983. Although most of the spawning occurs in the deep trench (>200 m) in Shelikof Strait, the lar- vae are in the upper part of the water column. Southwest of Shelikof Strait the trench runs south between the Semidi Islands and Chirikof Island (Fig. 2). Early larvae drift in the Alaska Coastal Current which flows southwest, parallel to the Alaska Peninsula (Schumacher and Reed 1980). At the time of our surveys in May 1981 and 1983, larvae had drifted to the area between Sutwik Island and the Semidi Islands. In 1982 they were further to the southwest in water over the deeper trough and the continental shelf. Vertical Distribution of Walleye Pollock Larvae Vertical distribution of walleye pollock larvae has been addressed in detail in several other stud- ies (Kamba 1977; Cooney et al. 1978; Walline 1981^1; Dagg et al. 1984). Although the areas of study, procedures, and gear have varied, a consis- tent pattern of diel-vertical distribution emerges from these studies and is supported by the present study. Haryu (1980) summarized his and earlier work by stating "larvae inhabit the mid-layer rather than the surface layer and perform diurnal vertical migration in search of food." Most larvae have been found between 10 and 60 m, and within this depth range, some larvae generally move to shallower depths at night. Vertical movement is not pronounced in any of the studies but is evi- dent by comparing proportions of larvae at vari- ous sampling depths at different times of day. In general it appears that larvae <15 mm are most concentrated vertically at 10-15 m at twilight, both in the evening and morning. During night- time and daytime the larvae are more dispersed vertically, and during daytime their distribution is deeper than at night. Samples confined to day and night periods do not show the crepuscular nature of the distribution. This pattern is seen in the present study but is less pronounced than in some others, possibly because we conducted our sampling only in the vertical range of high con- centration. Larvae larger than about 15 mm ap- pear able to avoid plankton nets to some extent, particularly during daytime. The available data, however, suggest that these larger larvae remain concentrated in a shallow depth stratum (5-15 m) except at night when they are more dispersed ver- tically and may rise closer to the surface (Walline fn. 11). Fish larvae of most other species that have been studied also migrate upward in the water column at night (Kendall and Naplin 1981). Some species undergo a much more pronounced vertical migration than is apparent with walleye pollock larvae and may cross much greater temperature gradients than observed here. Similar to walleye pollock, larvae of other fish species are visual feeders, and their vertical movements are proba- bly associated with a diel feeding periodicity. Walleye pollock larvae may move to shallower depths at night to allow more feeding in reduced light. They then may spread downward in the water column during daytime in response to in- creased light penetration and the distribution of their prey. Too little is known about predation on fish larvae to assess the importance of vertical movements on predator avoidance (see Incze et al. 1984). iiWalline, P. D. 1981. Hatching dates of walleye pollock iTheragra chalcogramma) and vertical distribution of ichthy- oplankton from the eastern Bering Sea, June-July 1979. U.S. Dep. Commer., Natl. Mar. Fish. Serv., NOAA, NWAFC Proc. Rep, 81-05, 12 p. 518 KENDALL ET AL.: GROWTH OF LARVAL WALLEYE POLLOCK Feeding of Walleye Pollock Larvae The diet composition of larval walleye pollock in Shelikof Strait is similar to that described for walleye pollock larvae collected in the southeast- ern Bering Sea (Clarke 1978) and Uchiura Bay, Hokkaido, Japan (Kamba 1977). Copepod nauplii and copepodids of Pseudocalanus spp. were the dominant food items in the guts of 6-20 mm lar- vae in all these studies. As in this study, copepod eggs were also abundant food items. It is difficult in any of these studies to judge if the eggs were captured as individual food items or along with adult female copepods. Feeding by larvae in the Gulf of Alaska is highest during daylight hours, as observed in other studies (Kamba 1977; Clarke 1978). Clarke (1978) reported that the few collections made at sunrise had larvae with the lowest feeding inci- dences. Kamba (1977) also reported that the low- est feeding incidences and the lowest abundance of food in the gut occurred near sunrise. The high densities of larvae in the Shelikof Strait seem to have little effect on their food habits. Oithona spp. are abundant in the Bering Sea, Gulf of Alaska, and Uchiura Bay, Japan, and are intermediate in size between Pseudocalanus spp. copepodids and copepod nauplii. Oithona are an important component of the diet of pollock lar- vae in the Bering Sea, accounting for more than 25% of the total number of food items for larvae between 11.8 and 17.7 mm (Clarke 1978), but are rare in guts of larvae collected near Hokkaido (Kamba 1977), and represent <167c of the food items for all size groups in the present study. Kamba (1977) cited the low incidence of occur- rence of this food item in larvae collected in Uchi- ura Bay as evidence of selective feeding by walleye pollock larvae. The zooplankton species composition in the oceanic and outer shelf regions of the Bering Sea (Cooney and Coyle 1981; Smith and Vidal 1984) is similar to that described for the northern Gulf of Alaska and Ocean Station P (Le Brasseur 1965; Damkaer 1977; Fulton 1983; Miller et al. 1984). The Shelikof Strait species composition is similar to these areas. Our zooplankton sampling did not include copepod nauplii so we cannot assess their abundance. The size distribution of copepod nau- plii ingested, however, indicates that Pseudo- calanus spp. and Oithona spp. are the probable sources of the copepod nauplii ingested by larval walleye pollock in Shelikof Strait. Daily production of copepod nauplii at a single station in the Bering Sea has been estimated to be 27,094 m^^, of which more than 95% was Pseudo- calanus spp. (Dagg et al. 1984). The abundance of Pseudocalanus females ranged from 9.9 to 258.9 m''^ (x = 87.7). The mean abundance of Pseudo- calanus females in Shelikof Strait was 244 m"^, or 2.6 times greater than the mean abundance in the Bering Sea. Assuming the same rate of daily production, about 69,000 nauplii m"^ would be produced in Shelikof Strait. Mean abundance of walleye pollock larvae where Dagg et al. (1984) performed their study was 6.3 larvae m"^, whereas at the diel station in Shelikof Strait the abundance was 156 m^, about 25 times greater than in the Bering Sea study. If these larvae ate nauplii at the same rate as those in the Bering Sea, 18.3 per day, they would eat about 24% of the production, as opposed to the <1% in the Bering Sea. Other factors such as the relationship be- tween size of larvae and daily ration need to be investigated before more precise estimates of the impact of larval feeding and the possibility of food limitation can be made. It appears that enough nauplii were being produced to preclude density dependent food restrictions at the larval densities observed in the present study. Growth of Walleye Pollock Larvae Growth rates were similar in areas of both high and low density (Table 5, Fig. 13). It cannot be determined from our study whether density- dependent factors modified larval growth; growth variations could be produced by patchy distribu- tions of prey. Walleye pollock larvae have been shown to grow faster in the laboratory at higher food densities (Bailey and Stehr 1986), further, where lower or constant larval densities interact with variable prey density, field studies have shown variability in growth (Govoni et al. 1985). Without knowledge of prey availability at each location, however, it is difficult to discern if high densities of prey coincide with dense patches of larvae. The relatively low growth rates found at two adjacent stations outside the patch (Stations 6, 12; Table 6) might indicate an area of less than adequate prey availability. Growth rates for fishes can be influenced by environmental factors such as temperature, as well as availability of adequate food supplies (Boehlert and Yoklavich 1983). Within species. 519 FISHERY BULLETIN: VOL. 85, NO. 3 growth rates are usually positively correlated with temperature over the normal temperature range. Growth rate for larval walleye pollock in the Gulf of Alaska (0.21 mm d"^) is considerably lower than that determined for larvae of the same size range (4-25 mm SL) collected in the south- eastern Bering Sea in June-July 1979 (0.35 mm d"\ Walline 1985). Water temperatures were similar in both studies (about 6°-8°C). Maximum growth determined for larvae collected in the Bering Sea in March-June 1980 while water tem- perature was much cooler (2°-6''C), was 0.22 mm d~^ (Clarke 1984). Thus differences in growth rates of larval walleye pollock in the Bering Sea could be due to differences in water temperature, food availability, or a variety of other factors which may affect growth in larvae (Bailey and Stehr 1986). Future research on growth variabil- ity in the field should take into account prey availability, temperature, and size-specific mor- tality rates. SUMMARY 1. Walleye pollock in the Gulf of Alaska form an intense spawning aggregation in Shelikof Strait in late winter that produces a dense patch of planktonic eggs in early April. Larvae from this spawning can be followed as they develop and are carried by currents to the southwest during spring. 2. In late May 1981, the density of larvae in this patch OlO m"'^) suggested that density- dependent effects on growth and survival might be expected. A field study of larvae in late May 1983 found maximum densities of only 1 larva m '^, and investigated growth and vertical distribution, and feeding in the patch. 3. The larvae were concentrated vertically be- tween about 15 and 50 m, and tended to be in the upper part of this range during night and early morning, whereas they were deeper dur- ing the afternoon and evening. 4. Larvae <10 mm fed primarily on copepod nauplii, with copepodids becoming more im- portant in larvae up to 20 mm. Copepodids of Pseudocalanus spp. made up a large fraction of the diet of larvae >10 mm. Most feeding oc- curred during daylight. 5. The copepods Pseudocalanus spp., Neocalanus spp., Calanus spp., Oithona spp., and Acartia spp. dominated the net zooplankton samples (253 |jLm mesh net). 6. Growth, based on otolith increments counted on 109 larvae (6.0-14.6 mm) was linear (0.21 mm/day, intercept = 4.29 mm, r^ = 0.75). Growth rates in the area of high abundance were generally not significantly different from those elsewhere. 7. While at the larval densities observed in 1981, density-dependent effects are possible, at the lower densities we observed in 1983 no such effects were expected or indicated in growth rates or diet. Future studies should include direct measurement of copepod naupliar pro- duction rates in the areas inhabited by the larvae. ACKNOWLEDGMENTS We wish to thank the officers, crew, and scien- tific staff aboard the NO A A ship Chapman for making the field work associated with this study both pleasant and productive. We thank Beverly Vinter, Jay Clark, and Darlene Blythe of NWAFC for their help in various stages of this study: converting animals in vials to data in the computer and our scribblings into typescript. Suam Kim, University of Washington, kindly helped analyze the hydrographic data. J. J. Gov- oni, NMFS Beaufort, NC; R. T. Cooney, Univer- sity of Alaska, Fairbanks; and A. J. Paul, Univer- sity of Alaska, Seward, provided excellent, thought-provoking reviews of an earlier draft of this paper. LITERATURE CITED Bailey, K M , and C L Stehr 1986. Laboratory studies on the early life history of the walleye pollock, Theragra chalcogramma (Pallas). J. Exp. Mar. Biol. Ecol. 99:233-246. BOEHLERT, G. W., AND M M. YOKLAVICH 1983. Effects of temperature, ration, and fish size on growth of juvenile black rockfish. Sebastes melanops. Environ. Biol. Fishes 8:17-28. Clarke. M E 1978. Some aspects of the feeding ecology of larval walleye pollock, Theragra chalcogramma in the southeastern Bering Sea. M.S. Thesis, Univ. Alaska, Fairbanks, 44 P- 1984. Feeding behavior of larval walleye pollock, Theragra chalcogramma (Pallas) and food availability to larval pollock in the southeastern Bering Sea. Ph.D. Thesis, Univ. California, San Diego, 208 p. COONEY, R T., AND K. O COYLE. 1982. Trophic implications of cross-shelf copepod distributions in the southeastern Bering Sea. Mar. Biol. 70:187-196. CooNEY, R T., T S. English, and T. Nishiyama 1978. Upper trophic level ecology with emphasis on juvenile pollock in the southest Bering Sea. In PROBES: Processes and resources of the Bering Sea shelf, p. 241-405. Prog. Rep. 1978. CUSHING, D H 1983. Are fish larvae too dilute to affect the density of 520 KENDALL ET AL.: GROWTH OK LARVAL WALLEYE POLLOCK their food organisms? J. Plankton Res. 5:847-854. Dago. M . M E Clarke. T Nishiyama, and S L Smith 1984. The production and standing slock of copepod nau- plii, food items for larvae of the walleye pollock Theragra chalcogramma in the southeastern Bering Sea. Mar. Ecol. Prog. Ser. 19:7-16. Damkaer. D M 1977. Initial zooplankton investigations in Prince William Sound, Gulf of Alaska and Lower Cook Inlet. NOAA' OCSEAP Annu. Rep. 137-220. Fulton. J D 1983. Seasonal and annual variations in net zooplankton at Ocean Station P, 1956-1980. Can. Data Rep. Fish. Aquat. Sci. 374, 65 p. GovoNL J J . A. J Chester. D E Hoss. and P B Ortner 1985. An observation of episodic feeding and giowth of larval Leiostomus xanthurus in the northern Gulf of Mex- ico. J. Plankton Res. 7:137-146. Haryu, T 1980. Larval distribution of walleye pollock, Theragra chalcogramma (Pallas), in the Bering Sea with special reference to morphological changes. Bull. Fac. Fish. Hokkaido Univ. 31:121-136. Hayashi K . H Kitahama, H Suzuki and U Endo 1968. Suketodara yogyo-ki no setei-do nan taihuyo 1964 nenkv ugen no kotatsu. lln Jpn.j J. Hokkaido Fish. Sci. Inst. 25:394-403. HAYNES, E G . AND S E IGNELL. 1983. Effect of temperature on rate of embryonic develop- ment of walleve pollock, Theragra chalcogramma. Fish. Bull., U.S." 81:890-894. Hempel, G 1979. Early life history of marine fish: the egg stage. Univ. Washington Press, Seattle. Houde, E D. 1975. Effects of stocking density and food density on sur- vival, growth and yield of laboratory-reared larvae of sea bream Archosargus rhomboidalis (L.) (Sparidae). J. Fish Biol. 7:115-127. INCZE. L S . M E Clarke, J. J Goering. T Nishiyama, and A J Paul 1984. Eggs and larvae of walleye pollock and relationships to the planktonic environment. In D. H. Ito (editor). Proceedings of the workshop on walleye pollock and its ecosystem in the eastern Bering Sea, p. 109-159. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-62. Jones. R 1973. Density-dependent regulation of the numbers of cod and haddock. Rapp. P. -v. Reun. Comm. int. Explor. Mer 164:165-173. Kamba, M 1977. Feeding habits and vertical distribution of walleye pollock, Theragra chalcogramma (Pallas), in early life historv stage in Uchiura Bay. Hokkaido-Res. Inst. North Pac. Fish., Hokkaido Univ., Spec. Vol., p. 175-197. Kendall. A W., Jr , and J R Dunn 1985. Ichthyoplankton of the continental shelf near Ko- diak Island. Alaska. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 20, 89 p. Kendall. A W . Jr , and N A. Naplin. 1981. Diel-depth distribution of summer ichthvoplankton in the Middle Atlantic Bight. Fish. Bull., U.S. 79:705- 726. Laurence. G. C. 1982. Nutrition and trophodynamics of larval fish — a re- view, concepts, strategic recommendations and opin- ions. MARMAP Contrib. MED'NEFC 82-50, 42 p. Le Bras.seur, R J 1965. Seasonal and annual variations of net zooplankton at Ocean Station P, 1956-1964. Fish. Res. Board Can. Manuscr. Rep. Ser. (Oceanogr. Limnol.) 202, 153 p. McGowan. J. A , and C B. Miller 1980. Larval fish and zooplankton community struc- ture. Calif Coop. Fish. Invest. Rep. 21:29-36. Miller. C B , B W. Frost, H. Batchhelder. M Clemens, and R E Conway. 1984. Life histories of large, grazing copepods in a subarc- tic ocean gyre: Neocalanus plumchrus. Neocalanus cris- tatus, and Eucalanus bungii in the northeast Pacific. Prog. Oceanogr. 13:201-243. Nakatani. T.. and T Maeda 1984. Thermal effect on the development of walleye pol- lock eggs and their upward speed to the surface. Bull. Jpn. Soc. Sci. Fish. 50:937-942. Nishimura. a , and J Yamada 1984. Age and growth of larval and juvenile walleye pol- lock, Theragra chalcogramma (Pallas), as determined by otolith dailv growth increments. J. Exp. Mar. Biol. Ecol. 82:191-205. Nishiyama. T , and K Hirano 1983. Estimation of zooplankton weight in the gut of lar- val walleye pollock {Theragra chalcogramma). Bull. Plankton Soc. Jpn. 30:159-170. O'Connell. C P , and L P Raymond 1970. The effect of food density on survival and growth of early post yolk-sac larvae of the northern achovy (En- graulis mordax Girard) in the laboratory. J. Exp. Mar. Biol. Ecol. 5:187-197. POSGAY. J. A., and R R MaRAK 1980, The MARMAP bongo zooplankton samplers. J. Northwest Atl. Fish. Sci. 1:91-99. Sameoto. D D , AND L O JAROSZYN-SKI 1969. Otter surface sampler: a new neuston net. J. Fish. Res. Board Can. 26:2240-2244. Schumacher, J D , and R K Reed 1980. Coastal flow in the northwest Gulf of Alaska: the Kenai Current. J. Geophys. Res. 85:6680-6688. Smith. P , and S Richard.son 1977. Standard techniques for pelagic fish egg and larva surveys. FAO Tech. Publ. 75, 100 p. Smith, S. L., and J, Vidal 1984. Spatial and temporal effects of salinity, temperature and chlorophyll on the communities of zooplankton in the southeastern Bering Sea. J. Mar. Res. 42:221-257. Walline. P D 1983. Growth of larval and juvenile walleye pollock re- lated to year-class strength. Ph.D. Thesis, Univ. Wash- ington, Seattle, 144 p. 1985. Growth of larval walleve pollock related to domains within the SE Bering Sea. "Mar. Ecol. Prog. Ser. 21:197- 203. Zar.J H 1974. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, NJ, 620 p. 521 ASPECTS OF THE BIOLOGY OF THE HAIR CRAB, ERIMACRUS ISENBECKII, IN THE EASTERN BERING SEA Therese M. Armetta' and Bradley G. Stevens-^ ABSTRACT The distribution and relative abundance of the hair crab, Erimacrus isenbeckii . were determined from data collected during annual summer trawl surveys conducted by the National Marine Fisheries Service (NMFS) in the eastern Bering Sea, 1979-84. The estimated population was about 23 million crabs from 1979 to 1981. but declined sharply to 4.4 million by 1984. The majority (67'-^) of the population occurred in the Pribilof Islands area. Male crabs occurred at a mean temperature of 3.4°C and depth of 66 m, whereas females occurred at a mean of 2.4°C and 64 m. Females comprised < W/r of the catch in NMFS surveys. Over 99'^J of the females caught were mature, but only eight were ovigerous with from 34,000 to 160,400 eggs. Length-width and length-weight relationships were calculated for males and females. The majority (77*7^ ) of£ . isenbeckii caught during an independently conducted study in May 1983 were found on a mixed sand and shell substrate. Scientific literature (mostly Japanese) was reviewed to provide information on larvae, reproduction, molting, growth, feeding habits, predation, migration, behavior, fishing, and marketing. The hair crab, Erimacrus isenbeckii (Brandt) (Fig. 1 ). is a medium-sized brachyuran in the fam- ily AtelecycHdae. Hair crab have been fished in Japanese and Korean waters for over 60 years (Kawakami 1934), and much hterature is avail- able on the biology, distribution, and abundance of the species in those waters. In contrast, fishing for hair crab in U.S. waters began in 1979 (Griffin and Dunaway 1985^). The recent development of a U.S. fishery for hair crab and the substantial decline of the eastern Bering Sea (EBS) popula- tion from 1981 to 1984 prompted an analysis of hair crab data collected by the National Marine Fisheries Service (NMFS) during the summers of 1979-84. This report presents data on the distri- bution and abundance of hair crab in the EBS during those years, as well as aspects of ecology, reproduction, molting, and growth. Additionally, we have summarized the literature concerning this species, since most of it is published in 'Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Wav NE, BIN CI 5700, Seattle, WA 98115. -Kodiak Facilitv, Northwest and Alaska Fisheries Center, National Marina Fisheries Service, NOAA, P.O. Box 1638, Ko- diak. AK 99615. ^GrifTin, K., and D. Dunaway. 1985. Bering Sea area shell- fish management report to Alaska Board of Fisheries. In Westward region shellfish report to the Alaska Board of Fish- eries, p. 179-245. Alaska Department of Fish and Game, P.O. Box 308, Dutch Harbor. AK 99692. Manu.scnpt accepted April 1987. FISHERY BULLETIN: VOL. 8.5, NO. 3, 1987. Japanese and not easily accessible to English- speaking readers. REVIEW OF PUBLISHED LITERATURE ON ERIMACRUS ISENBECKII The hair crab has a quadrangular carapace slightly longer than it is wide, and is densely covered with short bristles and sharp granular projections; seven teeth are present on each lat- eral margin. Chelipeds and walking legs are stout and spiny. The epistome has a nearly straight anterior margin. Rathbun (1930), Sakai (1939), and Kobyakova (1955) described the morphology of adult E . isenbeckii in detail. Five zoeal stages and one megalopa stage in the development of this crab are described by Kurata (1963). Accord- ing to Kurata, the zoeae are relatively large, ranging from 2.7 to 6.5 mm in body length (orbit to midpoint of posterior edge of telson), depending on zoeal stage, and are equipped with a long dor- sal spine (1.2-2.8 mm) and prominent lateral spines that are about one-fourth of the length of the dorsal spine. Abdominal margins of the cara- pace are fringed with setae; abdominal segments possess knobs, spines, and spinules. The mega- lopa is about 7.2 mm long. The rostrum is short and wide and ends anteriorly in three short teeth. Abdominal segments lack spines. In the EBS, hair crab occur from the northern 523 FISHERY BULLETIN: VOL. 85, NO. 3 Figure 1. -Male hair crab. Erimacrus isenbeckii , from eastern Bering Sea. Dorsal view. shore of the Alaska Peninsula to the Pribilof Is- lands and St. Matthew Island (Fig. 2). Hair crab are also found along the Aleutian Archipelago from Unimak Island as far west as long. 170°E (west of Attn Island; NMFS unpubl. data). In the western Pacific, hair crab occur along the eastern coast of Korea, the western and eastern coasts of Japan, and southern Sakhalin Island (Rathbun 1930; Tanikawa 1971). They are particularly abundant around the island of Hokkaido and along the Kurile Islands to southern Kamchatka, and are common along west Kamchatka to lat. 54°40'N (Vinogradov 1947). They are unknown from the western Bering Sea. Dall reported hair crab from Kachemak Bay and Cook Inlet, AK (Rathbun 1930). These, how- ever, were probably Telmessus cheiragonus, a similar atelecyclid that commonly occurs in the northern Gulf of Alaska (Calkins 1978), since no verified observations of hair crab have been reported east of Unimak Island despite numer- ous inquiries to commercial fishermen and biol- ogists working in the northern gulf in recent years. Sakurai et al. (1972) (Fig. 3, bottom) reported that primiparous (first time breeders') female hair crab off Hokkaido mate from December to Febru- ary and multiparous (have bred more than once) females mate from August to November (the lat- ter in deeper waters than the former). According to Sakurai et al., mating occurs immediately after molting when females are in the soft-shell condi- tion. When the female is ready to molt, a male crab grasps her chelipeds and holds on to them until after ecdysis. While the female is still soft, the male inserts his copulatory processes into her genital openings and fills the spermathecae with seminal fluid containing spermatophores. The male then secretes a mucoid, proteinaceous sub- stance from his seminal glands which congeals immediately into hard plugs that firmly close the female's genital apertures. The male may then mate with other receptive females. Primiparous Figure 3.— Average molting (top, Abe 1984, see text footnote 13) and breeding (bottom, adapted from Sakurai et al. 1972) cycles oi Erimacrus isenheckti (through age 8) offshore of south- eastern Hokkaido. It is uncertain when eggs hatch after the second spawning period. Carapace lengths were measured from the notch between the rostral spines. "C" numbers indicate post- larval instars. 524 ARMETTA and STEVENS: BIOLOGY OF THE HAIR CRAB Alaska Peninsula Unimak I. PACIFIC OCEAN m° N 50° 40" 120- E 150" 180° W 150° Figure 2. — Known world distribution of Ertmacrus isenbeckii. Areas of high and low density are indicated by crosshatching and parallel lines, respectively. Question marks indicate areas where hair crab are believed to occur but are not well documented. 120 110 100 90 80 70 60 50 40 30 20 10 0 C 14 C7 C9 J C 10 C 11 C 12 Male C 11 c Female breedincj cycle Molt and breed 111 Spawn Eggs hatch C 13 Female Molt and breed Eggs Spawn hatch .Ovigerous»^_ C 12 ( I .J Settle Hatch T V ^ ^ _' ' '_ "' •* — Ovigerous ♦ • ^ S , ~f-- -r— ^- , 1 1 1 1 I 1 I I I I I 1 1 I 1 I n I I I I 1 1 I I I I I I 1 1 I 1 III I I I I I I 1 1 II 1 1 I II ] I I I! I 1 1 1 I I I I I I 1 1 I 1 1 1 1 I Aprii Apr^, April April April April April April Apr-' 0 12 3 4 5 6 7 Age (years) I'll April 8 525 FISHERY BULLETIN: VOL. 85. NO. 3 females extrude eggs in October and November while muciparous females extrude eggs from March to May (Fig. 3) (Sakurai et al. 1972). This indicates a 4-10 mo interval between copulation and extrusion. Yoshida (1940) suggested that spawning occurs only in alternate years. Oocytes are yellowish white and immature at the time of copulation in both primiparous and multiparous females. During the interval be- tween copulation and extrusion, the oocytes ma- ture into dark orange ova 0.6 mm in diameter (Sakurai et al. 1972). After extrusion, the em- bryos are carried for at least a year during which time they become dark brown; hatching occurs from about March to May. Sakurai et al. reported that an average of 40,000-50,000 eggs and a max- imum of 160,000 eggs were produced and that external embryos were 0.8-0.9 mm in diameter. The distribution and timing of the occurrence of hair crab larvae near Hokkaido, the Kamchatka Peninsula, and in the eastern Bering Sea have been the focus of several studies. Near Hokkaido, Takeuchi ( 1969) and Abe ( 1977) found stages I-III in May and June; stages IV and V in June and July; and megalops from June to August. Takeuchi determined that stages were roughly 2 weeks apart. Stages I and II were only found in the surface layer (0-19 m), whereas stages III-V occurred in the surface and middle (20-50 m) lay- ers. Temperatures ranged from 6^^ to 11°C at the surface and from 2° to 10°C in the middle layer. Along the West Kamchatka Shelf, stage I larvae were found late- April to May and stages II- V in June and July (Makarov 1967). In the BBS, stages I and II were found late-April to June; stage III, May and June; and stages IV and V and megalops in June (Armstrong et al. 1983"*). Hair crab larvae in all three geographical areas were concentrated over bottom depths of 20-200 m, al- though some have been found in waters outside that range. Abe (1977) reported that settlement of hair crab larvae near Hokkaido occurred in July, in waters 20-50 m deep, 5''-7''C, on sandy mud or fine sand. Juvenile crabs remained in that same gen- eral area for the next 1.5 years as they grew from 5.1 mm to about 44.5 mm RL""' (40.2 mm CL) in eight successive molts. Ovigerous females were found in that habitat during the spring. Adult hair crabs moved offshore during July through September, as nearshore water temperatures gradually increased from 6^ to 15°C (Abe 1977). Matui (1970) found adults at depths of 20 m in April to 130 m in autumn, offshore of eastern Hokkaido, but hair crab have been found at depths of 5-364 m in other areas around Hok- kaido, the Kamchatka Peninsula, and Korea (Kawakami 1934; Sakai 1939). Hair crab were found on a variety of substrates, including sand, mud, gravel, rock, and broken shells, but sandy mud seemed to be most common (Kawakami 1934; Sakai 1939; Matui 1970; Abe 1977). After settlement in July, hair crab metamor- phose to first postlarval crab instars (CD with a mean size of 5.2 mm RL (Abe 1977, 1982) (Fig. 3, top). External sex characteristics are evident at stage C2 and a mean size of 7.0 mm RL. By the following April, 12 months after hatching, the crab reach stage C7 at a mean length of 27.4 mm RL. Approximately 33 months after hatching, the crab reach maturity at CIO with a length of 55-60 mm RL (50-54 mm CL) (Abe 1977, 1982); how- ever, hair crab males do not mate until 4 years of age and 70 mm RL (64 mm CL) (Sakurai et al. 1972). The smallest recorded male with mature spermatozoa was 41 mm RL (37 mm CL) (Hirano 1935). Molting frequency and mean carapace length are the same for both sexes through stage C9 (Abe 1977, 1982), however, after maturity males molt more frequently (Sakurai et al. 1972) and show greater growth per molt (Abe 1982) than females. Males begin to molt annually at about 55 mm RL (51 mm CL), once every 1-2 years (tending toward 2) in the size range 89-95 mm RL (81-87 mm CL), and biennially at sizes >100 mm RL (91 mm CL) and growth rate de- creases with age ( Yamamoto 1971 ). Males 65-105 mm RL (59-96 mm CL) experience a 10-25% ^Armstrong, U. A., L. S., Incze, D. L. Wencker. and J. L. Arm.strong. 1983. Di.stribution and abundance of deca- pod cru.stacean larvae in the .southeastern Bering Sea with em- phasis on commercial specie.s. Final Rep. to Natl. Oceanic Atmos. Admin., OCSEAP contract no. NA81-RAC- 00059. Office of Marine Pollution Assessment, Alaska Office RD'MPF24. P.O. Box 1808, Juneau. AK 99802. •Japanese scientists have traditionally measured crab lengths from the notch between the rostral spines ("rostral- length", RL), whereas NMFS scientists measure from the right orbit ("carapace length", CL). We converted rostral lengths to orbit lengths with the following equations, determined for crab in the size range of 40-100 mm CL: Males: CL = -0.81 + 0.921 RL /?2 = 0.982 N = 122 Females: CL= -2.10 + 0.943 RL ft^ = 0.998 N=8 These equations have similar slopes but significantly different intercepts (P < 0.05). All NMP^S crab measurements in this re- port are carapace lengths (CL). Japanese data are reported in original units (RL) and corresponding carapace lengths are also given for crabs '40 mm CL. 526 ARMETTA and STEVENS: BIOLOGY OF THE HAIR CRAB growth rate for carapace length ( Yamamoto 1971) and females of 50 mm RL (45 mm CD an 8-17% growth rate (Sakurai et al. 1972). Molting periods for adult hair crab vary with sex and locality. In general, males distributed along the coasts of Hokkaido and Korea molt be- tween the months of January and July (Yoshida 1940; Demon et al. 1956; Matui 1970; Sakurai et al. 1972 > and females molt during the periods of April to June (Yamamoto 1966) or August to February (Sakurai et al. 1972). Amphipods, anomurans, and isopods are impor- tant food items of the hair crab and peak feeding occurs at midday (Hirano 1935; Sakurai et al. 1972; Abe 1973). Hair crab are prey to fish species including various cottids (Sakurai et al. 1972; Abe 1973, 1982), salmon (Takeuchi 1972), and cod (June*'), and are occasionally eaten by red king crab, Paralithodes camtschatica , (Cunningham 1969). Hair crab migrate between shallow and deeper waters for mating purposes or in response to tem- perature changes (Yamamoto 1966; Sakurai et al. 1972). Primiparous females mate nearshore dur- ing winter, whereas multiparous females mate in deeper waters during autumn. Juveniles remain nearshore in water temperatures up to 15°C in late summer but adults move offshore. Hair crab also migrate along shore possibly to avoid in- creased densities (Hirano 1935; Abe 1977). Hirano reported that the longest straight-line mi- gration of a tagged crab was 18 km over a 16-d period and the greatest migration speed was 1.39 km day; however, the remaining 180 crabs recov- ered (442 tagged crabs released) within a 48-d period were at the site of release or within 7 km. Hair crab have been fished in Japanese and Korean waters with the use of conical pots (Fig. 4), trawls, and gill nets (Matui 1970; Yamaha Fishery Journal 1981). In any month of the year fishing occurs at some location around Hokkaido; it occurs from about November to April offshore of southern Hokkaido (Kawakami 1934; Yamaha Fishery Journal 1981) and from March to Decem- ber offshore of northern Hokkaido (Kawakami 1934; Matui 1970; Tanikawa 1971). Management measures have included area closures, total catch limits, pot limits, legal-size restrictions, and male-only restrictions (Matui 1970; Yamaha Fishery Journal 1981). Hirano (1935) and Kawakami (1934) believed that hair crab are es- pecially vulnerable to fishing pressures owing to "localized" migratory behavior, low number and fecundity of females, and the extended breeding period. By 1980, about 10 t of hair crab were har- vested offshore of Hokkaido every day, with 909^ of the harvest transported live to fish markets throughout Japan (Iversen^), and the remaining 10% sold frozen. U.S. fishermen began to land hair crab from the EBS in 1979 (Table 1). The majority of the com- mercial harvest has occurred incidental to snow (Tanner) crab (Chionoecetes sp.) fishing in the Bering Sea during the months of March through June, however, fishing season is not restricted. Only male crabs are landed. The Pribilof District (see Results for description of district) contributed 94-98% of the total Bering Sea catch during 1980- 84. Harvested crabs averaged 105.7 mm CL and 0.91 kg in 1984. Landings ranged from 2 t in 1979 to a peak of 1,108 t in 1981. Modified, baited king and Tanner crab pots are normally used. Pribilof Islanders, however, conducted an experimental •y. June, Fisheries Research Biologist, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Wav N.E.. Seattle, WA 9811.5, pers. commun. December 1982. "R. Iverscn, Regional Fisheries Attache, U.S. Embassy, Tokyo, Japan, APO San Francisco, CA 96503, pers. commun. August 1982. Table 1— Statistics of the US. commercial fishery of Erimacrus isenbeckii in the eastern Bering Sea (modified from Griffin and Dunaway, see text footnote 3). Mean length determined from port sampling, mean weight from landing records. Crab catch Pots Crabs Crab size Mean weight Mean number Metric length Year Vessels Landings (■ 1000) tons lifted pot (kg) (mm) 1979 11 16 2 2 9,908 0.2 0.95 111.8 1980 9 17 25 24 14,506 1.7 0.95 114.5 1981 67 192 1,127 1.108 172.695 6.5 1.00 104.8 1982 48 159 467 423 117.518 4.0 0.91 103.1 1983 52 161 575 550 84,346 68 0.95 1032 1984 19 74 398 364 42,806 9.3 0.91 105.7 527 FISHERY BULLETIN: VOL. 85, NO. 3 85 cm n B 50 Pots Figure 4. — Construction design and fishing method of Japanese hair crab pots. A, Side View: 1, lower ring of steel, upper ring and lateral bars of bamboo; 2, baitcan; 3, polyethylene cylinder entrance (28 cm top diameter, 25 cm bottom diameter, 19 cm high); 4, branch line attached to lateral bars and lower ring. Mesh shown only on top and one side. B, Top view: C, Longline method of pot fishing (fm = fathoms). fishery for hair crab during the summer of 1980 with the use of small conical crab pots (Mer- culieff^). All hair crab harvested in the United States have been exported to Japan as live or whole-boiled product, and prices to fishermen have ranged from $0.50 to $1.60/lb. MATERIALS AND METHODS Hair crab were caught by NMFS during annual summer trawl surveys (primarily designed to as- sess the abundance of king crab, Tanner or snow crab, and ground-fish species) in the EBS from 1971 to the present, but detailed data on hair crab have been collected only since 1979. Fishing was conducted with a 400-mesh eastern otter trawl in ^Merculieff L. 1981. Final report on the Pribilof hair crab project. Unpuhi. manuscr., 18 p. Tanadgusix Corp., St. Paul Lsland, AK 99660. 1979-80 and with an 83-112 eastern otter trawl in 1981-84; effective widths were 12.2 and 15.2 m, respectively (both nets were described by Wathne (1977)). Studies comparing the two nets showed no differences in size selection for king and Tan- ner crabs. We assumed the same for hair crab, which were too scarce for comparison. These dif- ferences in net widths have very minimal effect on the presentation of crab abundance, which is by order of magnitude (0-1, 1-10, 10-100 crab/nmi towed). In all years, the survey area extended from the Alaskan coast out to approximately the 200 m isobath and included Bristol Bay and the Pribilof Islands area, where hair crab densities are usu- ally greatest (Fig. 5). Only the northern limit of the survey area varied annually. Hair crab were also collected during NMFS cruises to the EBS in February of 1983 and 1985 and during an Outer Continental Shelf Environmental Assessment 528 ARMETTA and STEVENS: BIOLOGY OF THE HAIR CRAB - 64' N ■ 1979 only • 1982 only ▲ 1979 and 1982 0 1979 - 1980 A 1979 - 1981 D 1979 - 1982 • 1979 - 1984 179" E 1 175" W 169' 163" 157- Figure 5. — Standard station pattern for NMFS eastern Bering Sea summer surveys, 1979-84. Actual stations occupied may have differed slightly from this pattern in any year. Program (OCSEAP) cruise to the Pribilof Islands in May 1983. The seasons and durations of cruises are shown in Figure 6. Additional data concern- ing the NMFS summer cruises is contained in Otto et al. (1985, for the 1984 survey; similar documents are produced annually). Eastern otter trawl gear was used on all vessels throughout the 6-yr study period, except during the May 1983 OCSEAP cruise. On that cruise, either a 3.0 m beam trawl, 7.2 m try-net, or 1.2 m rock dredge were used to collect crabs, depending on bottom type determined from sediment samples taken with a Shipek bottom grab at each station. Fish- ing during both February cruises and the May 1983 cruise was conducted round-the-clock, whereas that during the summer surveys oc- curred only during daylight hours. Data collected during the February cruises and May 1983 cruise were not used in this report to determine distribu- tion and abundance of hair crab in the EBS be- cause of the limited area surveyed; only 1979-84 summer survey data were used for that purpose. Because different techniques were involved and no comparative fishing was conducted, catches of hair crab during the May 1983 cruise cannot be compared statistically with those from the sum- mer surveys. Tows were made in a systematic grid pattern with stations located 37 km (20 nmi) apart. Dur- ing several years, extra tows were made in areas of higher hair crab abundance around the Pri- bilofs and in Bristol Bay, which increased the pre- cision of population estimates during those years. Each tow lasted 0.5 hour and most were 2.2-3.3 km (1.2-1.8 nmi) long. Bottom temperatures were recorded with an expendable bathythermograph 529 FISHERY BULLETIN: VOL. 85. NO. 3 Sampling periods 111 MMFS 983' ILJ. NMFSr 985! I I .i.i,i|i|'.i.i.i.i|i| ' NMFS 1979 '.' ' 'iiiVi, M!'iTiTi;i 1,1.1,1.1,1,1,1,1, 1 I I 1 I I I NMFS 1980 1 I I II I M I I, M 1,1,1 I, ;|||X'1VNMFS 1981 ■rri'n'i'i'i','.'i',i,i.i,i|i|i|i|i|i|i|| ,' Ji'i'iNMFS 1982i|i|'|iiVi'i'i OCSEAPII ','1983:1:1 I 1,1 I 1,1 I 1,1 1,1,1 i,i,i,i|i|'i'|i,i|'i'l ' I i'i|''iNMFS1983;i|iJi;i|iXi| II I I I I I I I, NMFS 1984 ',',,i,'|i|i,J •'■'•''•'''' I . I I .,1 1, 1. 1,1, 1, i« January February March April May Months of year June July August FuiUKK 6. — Seasons and durations of sampling periods in the eastern Bering Sea, from 1979 through February 1985. at as many stations as possible. After the catch was brought aboard, all species including hair crab were removed, counted, and weighed. Carapace length of each crab was measured with steel vernier calipers to the nearest 1.0 mm from the rear of the right orbit to the middle of the posterior edge of the carapace (carapace length, CL). Carapace width (CW) was measured to the nearest 1.0 mm across the widest part of the cara- pace, excluding the lateral spines. Crabs were weighed on a triple-beam balance, and weights recorded to the nearest 1.0 g for crabs selected from a stratified size distribution. Shell condition was recorded as follows: molting (Drach's stages D2 through E; Passano 1960), softshell (stages Aj through Bj), new hard shell, old hard shell (prob- ably skipped one annual molt), and very old hard shell (probably skipped several annual molts). Hard-shell conditions were graded subjectively according to the amount of epifauna on the cara- pace, color of carapace, and wear on the spines. A new hard-shelled crab carapace was relatively clean with no epifauna, reddish to yellowish brown, with sharp spines. A very old hard-shelled carapace, however, was usually darker brown in color and almost always had epifauna, and spines that were rounded or worn smooth. An old hard shell was intermediate between these two condi- tions, but in practice it was difficult to distinguish between new and old hard shell. The presence or absence of external embryos was recorded for all female crabs. Six ovigerous females caught by NMFS in the EBS, 1979-85, were preserved in 10% formalin and returned to the Kodiak NMFS laboratory for determination of fecundity. The en- tire clutch was removed from the crab, and the embryos dried, sieved to remove debris, and weighed to the nearest 0.1 mg. Three subsamples of embryos from each crab were weighed and counted. The total number of embryos was esti- mated by dividing the total clutch weight by the average embryo weight. For each of three crabs caught in 1980, diameters of 30 fixed embryos were measured to the nearest 0.1 mm under a stereomicroscope with an ocular micrometer, and average embryo diameters were calculated. Em- bryos were nearly spherical so no distinction was made between length and width. Other data were analyzed to determine distribution and abun- dance, sex composition, length frequency, molt- ing periods, relative age according to shell condi- tion, distribution by temperature and depth, and reproductive condition of females. Population estimates were derived from trawl data using the area-swept technique (Alverson and Pereyra 1969) as described in Otto et al. (1985). The sampling variable was crab density, expressed as crabs caught per unit area swept, the latter equalling the product of net width and distance fished (determined with loran). High- and low-density strata were defined using the cu- mulative square root of frequencies method (Cochran 1963). Mean, total, and variance of crab density was determined within each stratum, and these combined for extrapolation to the survey area. 530 ARMETTA and STEVENS: BIOLOGY OK 11 IK IIAIK CKAB RESULTS Distribution and Abundance In the EBS, hair crab range from Bristol Bay west to about long. 174°00'W and north to St. Matthew Island at lat. 60"30'N (Fig. 7). Because so few juvenile and female hair crab were caught in NMFS surveys, the following information on distribution and abundance primarily concerns large males. Since fishery landings consist pri- marily of male crabs >89 mm CL, these are called "large", whereas "small" refers to male crabs <90 mm CL. Within the survey area, the crabs are divided into eastern and western centers of abundance. The western group occurs primarily in the Pri- bilof District (Alaska Department of Fish and Game [ADF&G] statistical district; south of 58°39'N, and west of 168°00'W) and is most dense (>10 crabs/nmi trawled) immediately adjacent to the Pribilof Islands. Moderately dense concentra- tions (1-10 crabs/nmi trawled) surround the Pri- bilof high-density region, especially to the north- east and south. The eastern gi'oup occurs in the Bristol Bay District (south of 58°39'N and east of 168°00'W) and is centered along the northern shore of the Alaska Peninsula from western Uni- mak Island to about 160°00'W. This group is mod- erately dense, with areas of high density (10-100 crabs/nmi trawled) located near the western end of the Alaska Peninsula in 1979, and offshore of Unimak Island in 1981. Hair crab are scattered across the continental shelf between these two major population centers and in the Northern District (north of 58°39'N) in low densities (<1 crab/nmi trawled). As with large males, small males and females displayed distinct east- ern and western concentrations, but very few were scattered between these two regions. Be- cause of the more-or-less continuous distribution of hair crab across the EBS, we subsequently treat them as belonging to a single widespread population. Population estimates have been made for hair crab only since 1979 (Table 2), and as previously mentioned, these reflect primarily the abundance of large males. From 1979 to 1981, the estimated population of EBS hair crab remained fairly sta- ble between 22 and 24 million crabs. The popula- tion dropped 607( between 1981 and 1982, 35'7r from 1982 to 1983, and 307f more from 1983 to 1984, to a low of only 4.4 million crabs. {Note added in proof: Hair crab abundance has contin- ued to decline to a total of 2.5 million crabs in 1986.) From 1979 to 1984, an average of 67Vr of the EBS hair crab occurred in the Pribilof Dis- trict, 279^ in the Bristol Bay District, and 67( in the Northern District. Although the total popula- tion size did not vary greatly from 1979 to 1981, the proportion of the population in the Pribilof District increased from 51 to 81'7( , while it de- creased from 40 to 18% in the Bristol Bay District, and from 9 to 1% in the Northern District. By 1984, the population distribution was again simi- lar to that of 1979. The population was very densely concentrated around the Pribilofs in 1981; however, since that time, the densities and range of hair crab in the EBS have declined greatly. Females comprised only 8% (248) of the total catch of about 3,091 hair crab during the 1979-84 NMFS summer surveys. In contrast, females ac- counted for 40% (48) of the 120 hair crabs >40 mm CL caught during the survey conducted in the Pribilof Islands in May 1983, when fishing was conducted both day and night around the Pribilof Islands, with dredge, try-net and beam trawl. Habitat Male E. isenbeckii collected during the summer Table 2. — Population estimates for Erimacrus isenbeckii. in the eastern Bering Sea, and proportions of the total population present in each statistical district. See text and Figure 7 for description of districts. Numbers are millions of crabs', M = Male, F = Female. Pribilofs Bristol Bay Northern Tota Is Grand Year M F All % M F All % M F All % M ±%2 F ±% total 1979 11.9 0.3 122 51 88 0.9 9.7 40 1.8 04 22 9 22.5 29 1.6 35 24.1 1980 15.1 23 17.4 77 3.6 0.7 4.3 19 0.8 0.1 0.9 4 19,5 47 3.1 138 22.6 1981 18.1 03 18.4 81 3.7 0.5 42 18 0.2 0.0 02 1 220 19 0.8 42 22.8 1982 63 0.1 64 67 2.3 0.2 25 26 0.5 0 1 06 6 9,1 22 0.4 56 9.5 1983 2.8 0.3 3 1 48 1.9 05 24 39 0,7 0.1 08 13 5,4 15 0.9 38 6.3 1984 23 02 2.5 57 1-2 0 1 1.3 30 0.5 0.1 0.6 13 40 16 04 48 4,4 1 Numbers represent only crabs within the survey area and those large enough to be retained by the trawl, i e mostly large males 2Two standard errors expressed as a percentage of the mean ■89 mm CL) 531 FISHERY BULLETIN: VOL. 85, NO. 3 ALASKA N. limit of survey Crabs per n. mi. trawled □ l-lO ho-100 • 100 I- 62° N 58° 54° 175°W 170° 165° 160° 157° N. limit of survey Crabs per n. mi. trawled 1 E3io-ioo 10 m ■ 100 ALASKA \ 62° N 58° 175°W FiGUKE 7. — Relative abundance of Erimacrus isenbeckii in the eastern Bering Sea, 1979-84. Depth contours are shown at 50, 100, and 200 m. Dotted Hne indicates the northern hmit of the survey in each year. Dashed lines demarcate the Bristol Bay (south of lat. 58°39'N, east of long. 168°00"W), Pribilof (south of lat. 58°39'N, west of long. 532 ARME'ITA and STKVKNS: BIOLOGY OK THE HAIR CRAB ALASKA N. limit of survey Crabs per n. mi. trawled • -.: 1 llO-lOO 100 62° N 58° 54° 175°W 165° 157° ALASKA t 62° N N. limit of survey Crabs per n. mi. trawled • -. ) EDlO-100 CIDl-IO E!3:-ioo 58° 1- 54° 175° W 170° 165° 160° 157° Figure 7 .—Continued— 168''00'W), and Northern (north of lat. 58°39'N» statistical districts. Densities expressed as crab per nautical mile trawled. 533 FISHERY BULLETIN: VOL. 85, NO. 3 62° N ALASKA N. limit of survey Crabs per n, ml. trawled • < 1 [I]l~10 110-100 100 58° 54° 175° W 170° 165° 160° 157° ALASKA N. limit of survey Crabs per n. mi. trawled • < 1 □ l-lO ho -100 100 62° N 58° 175°W Figure l.—Contmiwd. 170° 165° 160° ■^ 54° 157° 534 ARMETTA and STEVENS: BIOLOGY OF THK HAIR CRAB surveys occurred at a mean temperature (weighted by crab abundance) of 3.4°C and depth of 65.6 m, although they ranged from -0.9° to 10.1°C and from 22 to 249 m depth. One male hair crab was found outside this range, at 401 m. The mean values for females were 2.4''C (range -C.9°- 7.3°C ) and 63.8 m depth (range 26-243 m). Results of a 2-sample ^-test with unequal variances (Minitab "Twosample T" test; Ryan et al. 1976) indicated there was no significant difference (^ = 1.52, df=213, P=0.13) between mean depths at which male and female hair crab were found; however, there was a significant difference it = 5.82, df = 219, P < 0.01) between mean tem- peratures. Data from the Pribilof Island study of May 1983 indicate that hair crab appear to prefer a het- erogenous substrate as early juveniles, switching to sandy bottoms with increasing age. Among 120 juveniles <20 mm CL, 41% were on a substrate of gi-avel (less than about 1 cm diameter), poly- chaete tubes, and shell fragments, and small numbers were found in areas of large rocks, mud, or large shells. A substrate of medium or fine sand (usually containing shell fragments) was oc- cupied by 589f of crabs in the size range <20 mm CL, 70% of 10 crabs in the range 20-40 mm CL, 94% of 73 males >40 mm CL, and all 48 females >40 mm CL. Reproduction The scarcity of juvenile and female hair crabs in NMFS collections prevented a thorough study of reproductive characteristics of the EBS popula- tion; only eight ovigerous females were caught from 1979 to 1985. The size at maturity of these crabs in the EBS is unknown, however, the smallest mature female caught by NMFS was 38 mm CL and had spermathecae filled with a vis- cous liquid, indicating it had been mated. The smallest female with empty egg cases caught by NMFS was a 42 mm CL old hard-shell crab. We follow Abe ( 1977) in assuming that the mean size at maturity for female hair crab is above 55 mm RL (50 mm CL). Some female hair crabs collected during NMFS summer surveys were found with hard, proteina- ceous plugs in the gonopores. The plugs were root- like in appearance and formed a large, whitish, irregular-shaped protuberance outside the aper- ture (Fig. 8a). Each plug had a white, tapered stem that extended inward to the spermatheca (Fig. 8b). Some gonopores without plugs were closed by a flexible, swollen membrane (Fig. 8c) similar to the arthrodial membrane and continu- ous with the lining of the canal leading to the spermatheca. Some gonopores were open (Fig. 8d), owing to the flexible membrane having be- come flaccid. Although the presence of closed pores was not associated with any particular shell condition of the female, plugs and open pores were. Plugs were present only in recently molted soft-shell crabs, while most females with open pores (28 of 30, or 93%) were new or old hard-shell crabs. Some females had only one plug, and 96'7( of these also had the other pore closed. During the May 1983 OCSEAP cruise, 40% (19) of the 48 large females O40 mm CL) caught had plugged gonopores, 1 had new uneyed embryos, 4 carried eyed embryos that were in the process of hatch- ing, and 11 carried empty egg cases. Of the 19 females with plugged gonopores, 89'}^ (17) were new hard-shell crabs and 2 were newly molted soft-shell crabs. Most female hair crabs caught carried no exter- nal embryos (Table 3A). Although few crabs with Table 3. — Seasonality of egg bearing and molting in Erimacrus isenbeckii Uom the eastern Bering Sea. A) Percent of female crabs (actual number in parentheses) with embryos in each of 4 developmental stages. B) Percent of molting or soft-shelled crabs for each sex (total number of males or females caught shown in parentheses). February surveys 1983 1985 Summer surveys, 1979-1984, years combined (ylay June July August 89(16) 0 11(2) 0 88(109) 0 1(1) 11(14) 73(54) 1(1) 0 26(19) 38(12) 3(1) 0 59(19) 1 .3(234) 0.0(18) 2.0(1240) 5.6(124) 3.1(1215) 18.9(74) 0.0(154) 3.1(32) A. Condition of external embryos None 86(26) 93(14) Uneyed 0 7(1) Eyed 7(2) 0 Hatched 7(2) 0 B. Frequency of molting or soft-shell crab Male 20.0(136) 0.0(56) Female 30.0(30) 6.7(15) 535 FISHERY BULLETIN: VOL. 85, NO. 3 r ^f «s^^v^ Wv-I^A/ •^'- 0^' V.% O cm / 1 1 1 1 2 3 4 5 1 1 2 1 Figure 8.— Spermathecal plugs in formalin preserved female Erimacrus isen- beckii. a, ventral view of third thoracic sternite, with anterior towards the top of page, showing both gonopores with plugs; b, plugs removed from female crab; c, both pores closed (no plugs); d, both pores open. Plugs in live animals were whitish in color. 536 ARMETTA and STEVENS: BIOLOGY OF THE HAIR CRAB I cm 1 r 2 3 1 I "i »50 mm RL) L„+i = -0.40+ 1.336 (L„) Ln^i = 1L68 + L036 (L„) L„ + i = 9.49 + 0.998 (L„) Abe (1982) plotted the regression of percent growth per molt on length, and estimated maxi- mum lengths to be 125 mm RL (116 mm CL) for females and 177 mm RL (162 mm CL) for males. The largest hair crabs observed by NMFS in the EBS (Table 5) were smaller than these projec- tions, as were those caught near Hokkaido by Abe (1982), who reported maximum lengths of 152 mm RL (139 mm CL) for males and 105 mm RL (97 mm CL) for females. Reportedly, female hair crabs from Hokkaido rarely reach a carapace length >80 mm RL (73 mm CL) (Sakurai et al. 1972). In contrast, over 20*^/^ of the females caught in the EBS were >80 mm CL. However, NMFS trawl gear caught few hair crab <40-50 mm CL, and juvenile and female crabs may occupy rocky nearshore habitat which cannot be adequately sampled by such gear. The mean age of hair crab in the fishable popu- lation can be estimated from available data. Abe ( 1982) concluded that male crabs mature in their 10th postlarval instar, about 33 months after hatching, at about 60 mm RL (54 mm CL; Fig. 3). According to Yamamoto (1971), they would re- quire one more annual molt to reach stage Cll in their fourth year. At this size, crabs may molt annually or biennially. Male crabs landed in the EBS fishery averaged 106 mm CL (116 mm RL) in 1984, or about stage C14 (Tables 1, 7). To attain this size would require 3 molts from Cll, and these crabs would range in age from 7 to 10 years depending upon whether their last 3 molts were annual or biennial. If any failed to molt more than 1 year in a row, they would be age 11 at this size. Abe (fn. 13) (Fig. 3) indicated that male hair crab in Hokkaido waters reach similar sizes at the age of 6 years (assuming none skip molted). Resource Potential and Management Because of the great declines in abundance of the Bering Sea populations of E. isenbeckii from 1979 to 1984, this fishery will probably not be of great economic importance in the near future. If abundance increases in the future and prices re- main adequate, this fishery might become lucra- tive, albeit on a small scale relative to other Bering Sea crab fisheries. The species could then probably support a small boat fishery in the Pri- bilof Islands. Hair crab are still in high demand in Japan. The EBS hair crab fishery is not intensively managed. Fishing may occur year-round and is not limited by quotas. However, only males may be landed and gear is restricted to crab pots. There is no minimum size limit since the mar- ketable size is large relative to the probable size of male maturity (about 54 mm CL), although the latter has not been adequately determined. As a result of distribution and habitat differ- ences as well as gear selectivity, the size- frequency distributions of NMFS collections have been largely unimodal with few juveniles and fe- males in the catch. Thus, a thorough study of EBS hair crab reproduction and recruitment has not been feasible. Much useful information could probably be gained by systematic, year-round sampling of rocky heterogeneous habitats around the Pribilof Islands with appropriate gear such as rock dredges and beam trawls. Some data of this sort have already been collected by Armstrong et al. (fn. 11) and during other NMFS surveys, but are too limited to allow an improved understand- ing of growth rates, or the seasonality of molting and spawning of hair crab in the EBS. Further information on maturity, growth, and mortality is critical for informed management and will be necessary if this fishery gains importance. ACKNOV^XEDGMENTS This research would not have been possible without the cooperation and assistance of the masters and crews of the vessels used, including the RV Alaska, RV Chapman, RV Miller Free- man, RV Oregon, FV Paragon II, FV Discovery Bay , FV Ocean Harvester, and FV Pat San Marie. We must also acknowledge the significant efforts of the biological staff at the Kodiak Research Lab- oratory, who collected most of this data, and the secretarial and graphics staffs at the Kodiak and Seattle NMFS offices who assisted in the prepara- tion of this manuscript. We thank R. S. Otto, NMFS, Kodiak, AK, for reviewing the manu- 543 FISHERY BULLETIN. VOL. 85, NO. 3 script and providing additional data on distribu- tion. We extend appreciation to D. A. Armstrong, University of Washington, for allowing T. Armetta to participate in the 1983 NOAA/OC- SEAP survey, Contract No. 83-ABO-00066, and for the use of larval distribution data. Photo- graphs were taken by D. Kessler, NMFS, Ko- diak. LITERATURE CITED Abe. K. 1973. On the daily periodic behavior of hair crab Eri- macrus isenbeckii (Brandt). J. Hokkaido Fish. Exp. Sta. 30:1-14. 1977. Early life history of hair crab in the eastern Pacific waters of Hokkaido. Bull. Jpn. Soc. Fish. Ocean. 31:14- 19. 1982. The frequency of molting and growth of the horse crab. Bull. Jpn. Soc. Sci. Fish. 48:157-163. ALVERSON, D L . AND W T Pereyra 1969. Demersal fish explorations in the northeastern Pacific Ocean - an evaluation of exploratory fishing methods and analytical approaches to stock size and yield forecasts. J. Fi.sh. Res. Board Can. 26:1985-2001. Calkins. D G 1978. Feeding behavior and major prey species of the sea otter, Enhydra lutris. in Montague Strait, Prince William Sound, Alaska. Fish. Bull., U.S. 76:125-131. Cochran. W G 1963. Sampling techniques. John Wiley and Sons, New York, 413 p. Cunningham, D T 1969. A study of the food and feeding relationships of the Alaskan king crab, Paralilhodes camtschatica. MS The- sis, San Diego State Coll., San Diego, CA, 78 p. Do.MON. T . H Suzuki. M Yamamoto, T Mori, A Harada, and S Tachioka 1956. Survey for the hair crab stock in the Okhotsk Sea. J. Hokkaido Fish. Exp. Sta. 13:8-23. Hahtnoll. R G 1969. Mating in the Brachyura. Crustaceana (Leiden) 16:161-181. Hirano, Y 1935. Survey for the hair crab. Ten-day report, Hok- kaido Fish. Exp. Lab. No. 296, 10 p. Kawakami.S 1934. Survey for hair crab. Ten-day Report, Hokkaido Fish. Exp. i.ab. Nos. 258. 259, 16 p. Kobyakova. Z I 1955. Order Decapoda. //; E. N. Pavlovskii (editor), Atlas bespozvonochnyka dal'nevostochnykh morei S.S.S.R. (Atlas of the invertebrates of the Far Eastern seas of the U.S.S.R.). Izd. Akad. Nauk S.S.S.R., Moscow-Leningrad. In Russ. (Transl. by Israel Program Sci. Transl., 1966, p. 200-215, available U.S. Dep. Com- mer., Natl. Tech. Inf Serv., Springfield, VA, as TT66- 51067). KUR.'VTA. H 1963. Larvae of decapoda crustacea of Hokkaido 1. Atele- cyclidae (Atelecyclinaei. Bull. Hokkaido Reg. Fish. Res. Lab. 27:13-24. MakarOV. R. R. 1967. Lichinki krevetok, rakov-otshel'nikov i krabov za- padnokamchatskogo shelTa i ikh raspredelenie (Larvae of the shrimps and crabs of the West Kamchatkan shelf and their distribution). Izd. Nauka, Moscow, 163 p. In Russ. (Transl. by B. Haigh, Natl. Lending Lib. for Sci. and Technol., Boston Spa, Yorkshire, Engl. 1967, 199 p.) Matui, M. 1970. Pot fishery of the hair crab. In An illustration book of the fishing gears and techniques in Hokkaido area, p. 392-395. Hokkaido Assoc. Fish., Nemuro. Otto, R. S., B. G. Stevens, R A. Macintosh, K. L Stahl- JoHNsoN, and S. J. Wilson. 1985. United States crab research in the eastern Bering Sea during 1984. Int. North Pac. Fish. Comm., Annu. Rep. 1984:47-59. Passano, L M 1960. Molting and its control. In T. H. Waterman (edi- tor). The physiology of Crustacea, Vol. 1, Chap. 15. Acad. Press, NY., 670 p. 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, 609 p. Ryan, T A , Jr , B L Joiner, and B. F Ryan. 1976. Minitab Student Handbook. Duxbury Press, Boston, 341 p. Sakai.T, 1939. Studies on the crabs of Japan. IV. Brachygnatha. Brachyrhncha. Yokendo, Ltd., Tokyo, 741 p. Sakurai, M , S. Yamashiro, S. Kawashima, H. Omi, and K. Abe. 1972. Fish seen in the fishing industry. In The fish and fishing industry of Kushiro, p. 146-152. Kushiro City Publ. Kushiro Ser. No. 13. Takeuchi, I 1969. On the distribution of the larval stage of "Okuri- gani", Enmacrus isenbeckii and "Zuwai-gani", Chionoe- cetes opilio elongatus in the northeastern and the eastern regions of Hokkaido in 1958. Bull. Hokkaido Reg. Fish. Res. Lab. 35:20-43. 1972. Food animals collected from the stomachs of three salmonid fishes iOncorhynchus ) and their distribution in the natural environments in the northern North Pacific. Bull. Hokkaido Reg. Fish. Res. Lab. No. 38., 119p. Tanikawa, E 1971. Marine products in Japan - size, technology, and research. Koseisha-Koseikaku Co., Tokyo, 26, 507 p. Vinogradov, L G 1947. Dezyatinogiye rakoobrazniye Okhotskogo Morya (Decapod crustaceans of the Okhotsk Sea). Izv. Tik- hookean. Naukno-Issled. Inst. Rybn. Khoz. Okeanogr., Tom XXV, Vladivostok. (Transl. by S. Harder, Bureau of Translations, Dept. of Secretary of State, Canada, 1964, 100 p.) Wathne, F. 1977. Performance of trawls used in resource assess- ment. Mar. Fish. Rev. 39(6):16-23. Yamaha Fishery Journal 1981. Coastal crab fishery. 15:4 Dec. 1981. Published by Yamaha Motor Co., Ltd., Shizuoka-ken, Japan. Yamamoto. M. 1966. On the hair crab stock in the Abashiri re- gion. J. Hokkaido Fish. Exp. Sta. 23(12):15-33. 1971. On the growth of hair crab revealed by tagging ex- 544 ARMETTA and STEVENS: BIOLOGY OF THE HAIR CRAB periment. J. Hokkaido Fish. Exp. Sta. 28(4):8-13. Suisan Kenkyushi 35:192-200. YOSHIDA. Y 1941. On the reproduction of useful crabs in North Korea 1940. On the breeding of edible crabs produced in North (II). Suisan Kenkyushi 36:116-121. Korea. 1. Hair crab, Erimacrus isenbeckii (Brandt). 545 EXPLORATION FOR GOLDEN CRAB, GERYON FENNERI, IN THE SOUTH ATLANTIC BIGHT: DISTRIBUTION, POPULATION STRUCTURE, AND GEAR ASSESSMENT^ Elizabeth L Wenner.^ Glenn F. Ulrich,' and John B. Wise- ABSTRACT Exploratory trtipping for golden crab, Geryon fenneri. was conducted from 5 August 1985 to 21 February 1986 ofT South Carolina and Georgia. A buoyed system with strings of six traps (three side-entry Fathoms Plus and three top-entry Florida traps) was fished in six depth strata: 274-366 m, 367-457 m, 458-549 m, 550-640 m, 641-732 m, and 733-823 m. A total of 3,152 G. fenneri (2,661.9 kg) were collected at sampled depths between 296 and 810 m. The only other numerically important species caught was the jonah crab. Cancer borealis (864 individuals, 227.5 kg). Catches of golden crab were highly variable between strata. Catch per trap increased from 1.6 crabs (1.67 kg) in the shallowest stratum sampled to a maximum abundance of 22.3 crabs/trap (18.04 kg/trap) in the 458-549 m depth zone. Catches abruptly declined in the deeper strata sampled. Number of golden crab per trap (1.7:1) and weight per trap (1.6:1) in the Florida trap exceeded that in the Fathoms Plus trap for all completed sets. Traps yielded golden crab as small as 85 mm CW but the greatest proportion of crabs was >100 mm CW. Over 90^^ of all individuals exceeded 114 mm CW which is the minimum size of red crab, G. quinquedens , accepted for commercial utilization. Male golden crab were more numerous and larger than females. Crabs of the genus Geryon (Brachyura: Gery- onidae) are deepwater inhabitants of the At- lantic, Indian, and Pacific Oceans (Rathbun 1937; Monod 1956; Christiansen 1969; Manning and Holthuis 1981). Species reported off the United States in the western Ath^ntic and Gulf of Mexico include the red crab, G. quinquedens Smith, and the golden crab, G. fenneri Manning and Holthuis. At the time G. fenneri was described (Manning and Holthuis 1984), its geographic and bathymetric distribution included the continen- tal slope off eastern Florida, the Florida Straits, and the Gulf of Mexico. An exploratory fishing effort in 1984 collected the first known specimens of golden crab off South Carolina'*, and it is now known that golden crab occur in waters off Bermuda (Luckhurst in press). Both G. quinquedens and G. fenneri have been 'Contribution No. 232, South Carolina Marine Resources Center, Marine Resources Research Institute. '■^Marine Resources Research Institute, South Carolina Wildlife and Marine Resources Department, P.O. Box 12559, Charleston, SC 29412. •^Office of Fisheries Management. South Carolina Wildlife and Marine Resources Department, P.O. Box 12559, Charleston, SC 29412. ''South Carolina Wildlife and Marine Resources Department, unpubl. data, courtesy Charles Wenner, Marine Resources Re- search Institute, Charleston, SC. the target of limited and sporadic commercial fishing efforts off the east coast of the United States (Gerrior 1981), in the Gulf of Mexico. (Otwell et al. 1984; National Marine Fisheries Service 1986''), and off Bermuda (Luckhurst in press). Although much information is available concerning the biology and commercial fishery of red crab (summarized by Gerrior 1981), biological information on golden crab is more limited. Otwell et al. (1984) demonstrated exploratory trapping and processing techniques for golden crab from the Gulf of Mexico. The initiation of a small commercial crabbing enterprise during 1984 in South Carolina yielded promising quantities of golden crab*'. We began the present study to determine the fishery poten- tial, compare trap designs, delineate bathymetric distribution, and describe the biology of golden crab in the South Atlantic Bight. This report doc- uments results on catch rates, size and sex compo- ■''National Marine Fisheries Service. 1986. Species profile: deep red crab, Geryon quinquedens, Smith and golden crab, Geryon fenneri, Manning and Holthius, 1984 from the south- eastern U.S. south of Cape Hatteras, N.C. U.S. Dep. Commer. Natl. Mar. Fish. Serv., NOAA, Pascagoula Lab., Latent Resour. Rep., 17 p. 6H. Holley, commercial fisherman, Charleston, SC, pers. com- mun. 1985.' Manuscript accepted March 1987. FISHERY BULLETIN: VOL. 85. NO 3, 1987. 547 FISHERY BULLETIN: VOL. 85, NO. 3 sition of G. fenneri as a function of depth and trap type, and examines aspects of adult life history and reproductive biology of this species in the South Atlantic Bight. METHODS Cruises were made during the period from 20 June 1985 to 21 February 1986 on board the South Carolina Wildlife and Marine Resources Department (SCWMRD) research vessels Oregon and Lady Lisa , and the NOAA ship Chapman. All vessels were equipped with large capacity hy- draulic systems and a heavy duty pot hauler. Two commercially available trap designs were used to sample crabs. The Fathoms Plus^ traps are oval (85 cm long x 66 cm wide x 30 cm high) and constructed of injection molded plastic. The trap has two side-entry funnels that can be en- larged by removing more of the plastic funnel's inner lip. The original, oval funnel opening is 10 cm X 20 cm. Both funnels were cut out to a maxi- mum opening size of 14 cm x 22 cm. Traps were weighted with chain, making the total weight of each trap 11 kg. The Florida trap is an injection molded, high-impact plastic version of a Florida spiny lobster trap (82 cm long x 61 cm wide x 45 cm high). The top of the trap is constructed of wood lathing to provide a biodegradable escape panel. The top entrance funnel has adjustable panels and is 20 cm x 25 cm in the most open position, as fished throughout the study. Two strips of poured concrete in each end of the trap provided ballast, making the total weight of the trap about 22.7 kg. Traps were baited with 1.2-1.6 kg of clupeids. Three Florida and three Fathoms Plus traps were alternately attached at 61 m intervals to 365.6 m of groundline. The groundline was constructed of 8 mm diameter Iceline, a dacron, polyethylene line that has a high tensile strength relative to its diameter. A small weight consisting of —9.0 kg of chain was attached to one end of the groundline and an anchor ( -25 kg) was attached to the buoy- line end of the gear. Buoy lines were 366 m sec- tions of 8 mm Iceline joined together to achieve at least a 2:1 ratio of line to water depth. Four inflat- able net buoys and a spar buoy with radar reflec- tor were attached to the buoyline. Six depth strata were sampled between lat. 29°53.1'-32°20.0'N and long. 78°01.5'-79°24.8'W: 274-366 m (stratum 1), 367-457 m (stratum 2), 458-549 m (stratum 3), 550-640 m (stratum 4), 641-732 m (stratum 5), and 733-823 m (stratum 6). Three sets of six traps each were made approx- imately 1-2 km apart within a depth stratum over a 24-h period. Sampling locations for each set were selected by making fathometer transects of the potential fishing area to determine depth and bottom type. Because of bad weather and logisti- cal constraints all strata did not receive equal effort (Table 1). The first trap type on the groundline was ran- domly selected with trap type alternating until six traps (three of each type) were attached. The exception to this arrangement occurred in the deepest stratum (733-823 m) where only the Fathoms Plus trap was used. Fishing duration was standardized at 20 hours; however, poor weather conditions and logistical considerations altered this. Average fishing du- ration within strata exceeded 17 hours (Table 1). Bottom temperature was determined in each depth stratum by reversing thermometers. Bot- tom sediments were sampled by a geological rocket gi'ab for each group of three sets made in an area. Sediments retrieved were frozen on board and examined under a microscope for gross characterization in the laboratory. Sampling depth and location were recorded at deployment of the anchor. Decapod crustaceans in each trap were identi- fied, counted, and weighed. Catches from dam- aged traps or those sets that moved due to cur- rents were excluded from analyses of distribution and abundance, but were included in biological studies of size and sex composition. Each golden crab was individually sexed, measured to the nearest millimeter (carapace width, CW, distance between the tips of the fifth lateral spines; cara- pace length, CL, distance from the diastema be- Table 1. — Mean, standard deviation, mini- mum, and maximum fishing duration of trap sets, for Geryon fenneri, within six strata sam- pled from August 1985 to March 1986. 'Reference to trade names does not inply endorsement by the National Marine Fisheries Service, NOAA. No. sets Fishing di uration (h) Stratum y (s) Min. Max. 1 8 17.5 3.75 12.4 21.2 2 32 18.8 2.36 14.0 23.2 3 16 20.5 2.07 16.2 23.7 4 6 22.9 1.15 21.5 24.7 5 4 17.2 0.73 16.4 18.1 6 4 20.2 5.69 11.7 23.3 548 WENNER ET AL.: EXPLORATION FOR GOLDEN CRAB tween the rostral teeth to the posterior edge of the carapace, along the midline), and most were weighed to the nearest gram. The number of missing chelae and pereopods was recorded for each crab, as was molt condition and presence of chitinolysis and poecilasmatid barnacles, Trilas- mis inaequilaterale , on the exoskeleton. Molt con- dition of G. fenneri was modified from criteria established by Beyers and Wilke (1980) for G. quinqiiedens (probably G. maritae Manning and Holthius) and consisted of five categories: 1) Hard - carapace at maximum strength, fouling by barnacles or chitinolytic bacteria mini- mal, 2) Hard old - carapace strong but heavily fouled by barnacles and abraded or blackened by chitinolytic bacteria, 3) Soft old - resorptive line along posterolateral sides of the carapace is weak; carapace heavily fouled as with hard-old condition, 4) Soft new - carapace soft or jellylike with no fouling, and 5) Hard new - carapace cracks under pressure and is not fouled. Female G. fenneri were examined for evidence of egg extrusion and mating. Presence of eggs or egg remnants on pleopods and the size, shape, and physical condition of vulvae, as described by Haefner (1977), were noted. We examined semi- nal receptacles for presence of sperm or sperma- tophores and for relative size. Ovaries from 72 of the 166 female G. fenneri captured were initially classified by relative size and color following the scheme described by Haefner (1977) for G. quinquedens. After gross classification of ovaries, tissues were removed for histological preparation and examination in order to describe ovarian structure and validate assigned ovarian stages. Tissues were fixed for at least 48 hours in 10% seawater formalin. After fixation, tissues were dehydrated, cleared, and embedded in paraffin. Sections were cut at 6-9 (xm and were stained with Gill's hematoxylin and counterstained with eosin-Y. Oocytes from G. fenneri were measured using an ocular microm- eter. Testes and vas deferentia from three G. fenneri were fixed for 24 hours in 2.57^ glutaraldehyde, rinsed in cacodylate buffer, and dehydrated in ethanol. Tissues were then critical-point dried, sputter coated, and examined using a Jeol JSM- 35C scanning electron microscope (SEM). RESULTS Distribution and Relative Abundance The 70 valid sets (416 individual trap observa- tions) caught 3,152 G. fenneri (2,661.9 kg) at sam- pled depths between 296 and 810 m. The only other numerically important species caught was the Jonah crab, Cancer borealis (864 individuals, 227.5 kg). Catch per trap increased from 1.6 crabs (1.67 kg) in the shallowest stratum to a maximum abundance of 22.3 crabs/trap (18.04 kg'trap) in the 458-549 m depth zone (Fig. 1). Catches then abruptly declined with increasing depth. The ab- sence of golden crabs in traps fished between 550 and 640 m appears to be related to unsuitable Geryon fenneri Tt CO com N-l^ COCO CVJCSi rr 050) Tj-T}- »-»- T- T- c\j 14.0- 0. < DC 10.0- I- ~^ DC UJ m 2 6.0- D Z 2.0- i NS — r: 21 1 I I 0 0 0 NS a. < DC 1- ~^ O 10.0- 6.0- 2.0- I NS 1 - '/A I I Fathoms Plus ^ Florida Traps 0 0 274- 366 367- 458- 550- 641- 733- 457 549 640 732 823 DEPTH (m) Figure 1. — Catch per trap of Geryon fenneri for six depth strata sampled. Effort (number of traps) is shown in parentheses. Statistical significance of catches between trap types, as deter- mined by two sample /-test, is indicated by * (P <0.05). NS indicates no significant difference in catch rates between the two trap types. 549 FISHERY BULLETIN: VOL. 85, NO. 3 sediments at sites in this stratum since grab sam- ples contained coral fragments and rubble. At shoaler locations where golden crabs were abun- dant, sediments were a mixture of soft silt-clay, molluscan shell fragments, and foraminiferan tests. Temperatures at sites where golden crabs were collected ranged from 7.14° to 9.15°C. The number of golden crab per trap (11.4) and the weight of golden crab per trap (9.37 kg) in the Florida trap exceeded that in the Fathoms Plus trap (7.0 individuals, 6.16 kg) for all combined sets (Table 2). Statistical results by strata using the two-sample ^-test or an approximate f-test when variances were heterogeneous (Sokal and Rohlf 1983), indicated significantly more crabs were collected with the Florida trap than with the Fathoms Plus trap from 367 to 457 m (stratum 2) and from 458 to 549 m (stratum 3) (Fig. 1, Table 2). Weight per trap was significantly differ- ent for the 367-457 m stratum only. Size and Sex Composition Male G. fenneri were significantly more numer- ous than females, outnumbering them by -18:1. No ovigerous females were collected during the sampling period. Dominance of males was statis- tically significant for strata 1-3 (Table 3). In these depth strata, males were 20 times as numerous as females. In depths of 550-732 m, a male was the ranged from 85 to 193 mm in carapace width and weighed from 100 to 2,109 g. Average weight of male golden crab collected during the study was 927 g (s = 373.448, n = 1,640) while average weight of females was 443 g (s = 289.385, n =86). Carapace width-frequency distribution for G. fenneri gave modes at 155 mm for males and 100 mm for females (Fig. 2). The largest crab collected measured 193 mm and weighed 2,091 g. Linear least-squares and functional regression equations (Ricker 1973; Sokal and Rohlf 1983) relating carapace length and live wet body weight with width are in Table 4. Width-weight relation- ships were calculated from data on individuals that were not missing appendages. Of the 3,183 golden crabs examined for missing appendages, 2.4% were missing one or both chelae. Pereopods were missing from 307 individ- uals (9.6%). Examination of carapace width and weight statistics for each depth stratum showed that mean size of male G. fenneri was greatest for the shallowest (274-366 m) and deepest (733-823 m) strata sampled (Table 5). For females, however, mean carapace width and weight were greatest in the deepest zone. At depths of peak abundance, mean carapace width it^ = 4.70, P < 0.001) and mean body weight it, = 2.70, P < 0.01) of male crabs were significantly greater in the 367-457 m than in the 458-549 m depth stratum. No signifi- Table 2 — Results of /-test (T^) comparisons of mean number and weight (kg) per trap for two trap types (FM + and FLA) fished in each depth stratum for Geryon fenneri. Standard deviation is noted in parentheses; ' indicates significance at 0.05 level. Number 'trap Weight'trap Stratum FM + FLA Ts FfVl+ FLA Ts 1 2 3 4 1.6(1.92) 7.5(596) 10.4(5.08) 0 1.8(2.21) 11.9(9.95) 15.2(7.08) 0 0.14 2.16* 2.22* 1.66(1.912) 1.80(2.157) 6.67(5 097) 10 02(7.651) 8 82(4 164) 11.84(4,752) 0 0 0.14 2.06* 1.91 5 0 1 0 — 007 0 — 6 Total 0.5(0.28) 7.0(5.98) 11 4(9.38) 8.51* 0.43(0.321) — 616(5.052) 9.37(7.078) 6.18* only crab collected. In the deepest stratum sam- pled (733-823 m), females significantly outnum- bered males 2.9:1. Although the Florida trap caught significantly more crabs than the Fath- oms Plus trap overall, no significant difference was noted in the number of female crabs between those two trap types (x^ test, P > 0.5). The 3,217 golden crabs which were measured Table 3. — Frequency of male and female Geryon fenneri within each depth stratum. Asterisks denote significant deviation (P ■ 0.05) from 1:1 by Chi-square analysis Strata (m) Sex 274-366 367-457 458-549 550-640 641-732 733-823 IVIale Female 84* 1,790* 1,165* 0 1 11 3 91 41 0 0 32* 550 WENNER ET AL.: EXPLORATION FOR GOLDEN CRAB Geryon fenneri cant differences, however, were noted in mean carapace width (tg = 0.85, P > 0.05) and mean body weight itg = 1.48, P > 0.05) of females from these same strata. Of the two traps used, the Fathoms Plus trap caught larger and heavier golden crabs than did the Florida trap. Mean carapace width (y = 143 mm, s - 19.69, n = 1303) of crabs in the Fathoms Plus trap was significantly larger than that of crabs in the Florida trap (y = 139, s = 20.21, n = 1914) Us = 5.478, P < 0.001]. A statistically significant difference was also noted for mean weight (Fathoms Plus: y - 928, s = 366.77, n = 775; Florida: y = 881, s = 377.69, n = 951) [t, = 2.598, P < O.obl]. Figure 2. — Width-frequency distribution.s of male and female Geryon fenneri caught in traps, y = mean; s = standard devia- tion; n = number of individuals. O z ai Z) o m DC LU O cc LU Ql 20- 15- 10- 0- MALE y = 143inm s =18.93 n=3051 "I 5- 1—' ~u ^ I 1 1 "T 1 1 1 1 1 1 1 1 'l FEMALE y = 1 13 mm s= 19.99 n =166 E 1 — \ — \ — \ — I — I — \ — \ — r I I 1 90 100 110 120 130 140 150 160 170 180 190 200 CARAPACE WIDTH (mm) Table 4. — Least-square linear and geometric mean functional regression equations of carapace length (CL) and live body weight (WT) on carapace width (CW) for each sex of Geryon fenneri. Length and width units are millimeters while weight units are kilograms. All least square regressions were signifi- cant at a = 0 05. Sex Least squares equation GM functional equation Male CL log 10 Female CL log,o WT 9 5 + 0.9 CW WT = -4.74 + 3,54 (logig CW) 4.0 + 0.8 CW -3.97 + 3.14 (logio CW) 3,042 0.95 CL= -11.9 -H 0.9 CW 1 ,453 0.94 logio WT = -4.99 + 3.66 (logio CW) 141 0.92 CL = 0.7 + 0.8CW 74 0.91 logio WT= -4.27 -^ 3.29 (logio CW) Table 5. — Size and weight statistics of male and female Geryon fennen horn sampled depth strata y = mean: s = standard devi- ation, n = number of individuals. Stratum (m) Carapace width (mrr y Mm. Max. s >) n Weight (g) Sex y s n Male 274-366 156 117 186 14.4 84 1,064 339.15 84 367-457 144 100 190 18.1 1 790 937 354 03 983 458-549 140 88 193 19.9 1 165 884 373.47 561 550-640 — — — — — — — — 641-732 139 — — — 1 809 — 1 733-823 161 135 181 117 11 1,112 225.22 11 Female 274-366 105 92 113 11.2 3 189 80 88 3 367-457 105 85 145 86 91 265 103.13 35 458-549 104 85 137 9.7 41 228 7006 16 550-640 — — — — — — — — 641-732 — — — — — — — — 733-823 149 117 170 13.6 31 768 201 09 32 Reproductive Biology We obtained satisfactory histological sections from 39 of 72 female golden crabs examined. From histological and gross examination we de- scribed four ovarian developmental stages: Dearly, 2) intermediate, 3) advanced, and 4) mature. In the early stage of development, the slightly lobate ovary is very small, transparent to white in color, and bounded by fibrous connective tis- sue. Oocyte diameter ranged from 58 to 92 ixm with a mean of 75 fxm (Fig. 3A). Nuclei and nucle- oli are readily apparent in the early oocytes, as are follicle or accessory cells which surround each 551 { FISHERY BULLETIN: VOL. 85, NO. 3 j>«^' ■ «- ht^ Figure 3. — Ovarian and testicular tissue of Geryon fenneri. A. Ovarian tissue showing early (EOC) to intermediate oocyte (IOC) development. Oocytes range in size from 30 to 100 jjim. Scale bar 60 jim. B. Oocytes at the intermediate stage of development. Nucleus (N), nucleolus (NU), cyto- plasmic yolk globules (CYG), follicle cells (FC). Oocyte size extremes are 100-125 |j.m. Scale bar 50 |j.m. 552 WENNER ET AL.: EXPLORATION FOR GOLDEN CRAB V* '"*Aff*'y'/^' ^: .«-"i.; .- -^V "1^' ^ "' ;^.-'''j -^ ^l ST. l'^''^ A AN _^ % -W . •» . . <• r ^ V K^ V « -* ''**•" .1'"" Figure 3.— Continued. C. Oocytes in the advanced stage of development. Nucleolus (NU), cytoplasmic yolk gran- ules (CYG). Oocytes 200-300 jim in size. Scale bar 30 jim. D. A portion of a mature testis showing the seminiferous duct (SD) and testicular lobes containing spermatocytes (SO, spermatids (ST) and sperm (S). Accessory cell nuclei (AN). Scale bar 100 ^Jlm. 553 FISHERY BULLETIN: VOL. 85, NO. 3 cell. In the larger oocytes, cytoplasmic vitellin globules indicative of vitellogenesis are pre- sent. The intermediate stage ovary is yellow in color, has more pronounced lobation, and is larger than the early stage ovary. The diameter of oocytes ranged from 112 to 175 ixm with a mean diameter of 145 |xm. Most oocytes were undergoing vitello- genesis in this stage (Fig. 3B). As the ovary matures to the advanced stage, the ovarian lobes become enlarged and the color becomes light orange to orange-red in color. The anterior portion of the ovary obscures the ante- rior hepatopancreas from dorsal view. Oocytes were 175-300 |j.m in diameter (y = 240 p.m) and enlarge as vitellogenesis continues (Fig. 3C). The mature ovary, brown to purple in color, is the dominant visible organ and obscures the hep- atopancreas in dorsal view. Oocytes are filled with yolk globules and average 300-400 fjim in diameter as vitellogenesis nears completion. Size at sexual maturity was difficult to assess because of the small number of females collected. Overlap existed in the size of female G. fenneri in each stage of development.The carapace width of females in early ovarian development ranged from 85 to 116 mm (j = 104 mm, n = 27). Inter- mediate ovaries were present in females measur- ing 105-169 mm CW ^ = 127 mm, n = 13) while advanced ovaries occurred at sizes from 110 to 136 mm CW (y = 123, n = 2). The 30 females with mature ovaries ranged from 97 to 169 mm CW (y = 141). Five vulval forms were identified among the 142 females examined. Most of the females had immature vulvae (types a and b) suggesting that these crabs had not mated. The observed ovarian condition in a subsample (n = 26) of these fe- males indicated that all had ovaries in an early stage of development (Table 6). Only one female (111 mm CW) with immature vulvae contained sperm in the seminal receptacles, indicating cop- ulation had occurred. Type c vulvae were noted on two females, one with ovaries in early develop- ment and lacking sperm in the seminal recepta- cles while the other crab had mature ovaries and sperm present. Type e and f vulvae were found on the largest females collected, all of which had at least intermediate stage ovaries. Eight of the fourteen females with these vulval types whose seminal receptacles were examined had been in- seminated. Three male G. fenneri examined exhibited typi- cal brachyuran reproductive morphology. The testes, which are dorsal to the hepatopancreas, were tubular and highly lobate. The testicular lobes, adjacent to the central seminiferous duct, contained spermatocytes, spermatids, and sper- matozoa, suggestive of asynchronous develop- ment (Fig. 3D). In mature individuals, ripened spermatozoa were found in the seminiferous duct. Examination of the testes and vas deferentia by SEM revealed germ cells at various stages of de- velopment. Spermatids (Fig. 4A), surrounded by supportive tissue, were composed of a central nu- cleus framed in c3^oplasm. With spermiogenesis, multiple projections or spikes form which are characteristic of developed sperm (Fig. 4A). An- other portion of the same testis yielded a more advanced germ cell displaying well-defined cj^to- plasmic spikes (Fig. 4B). A sagittal section through the vas deferens revealed stellate sper- matozoa (Fig. 4C), which had previously been em- bedded in this complex of supportive tissue (Fig. 4D). Table 6. — Incidence of vulval type (after Haefner 1 977) in relation to carapace width and gonadal condition of female Geryon fen- neri. n = number of individuals examined. Carapace Gonadal Type n width (mm) n condition a 112 85-119 22 early b 4 98-116 4 early c 2 97-109 1 1 early mature d 0 — 0 — e 19 105-156 8 1 9 intermediate advanced mature f 5 124-169 2 3 intermediate mature Molt Condition and Fouling Most (80%) of the 3,183 male and female G. fenneri were in the intermolt stage. Less than 1% of the 3,041 male golden crab showed evidence of having recently molted. The incidence of immi- nent or recently molted female golden crab was higher than that observed for males, with four individuals classified as premolt (soft-old) and two in the newly molted (soft-new) condition. Most (95%) of the 3,183 G. fenneri examined for molt condition had blackened abraded areas on the exoskeleton, indicative of damage by chiti- nolytic bacteria. Exoskeleton damage was most prevalent on individuals in the intermolt (75%) and premolt (19%) condition. 554 WENNER ET AL.: EXPLORATION FOR GOLDEN CRAB DISCUSSION Although the results of this study suggest that G. fenneri has a wide bathymetric occurrence in the South Atlantic Bight, the depth extremes for the species probably extend beyond those encom- passed by our sampling design. Records of Geryon sp. and G. affinis (which were probably G. fen- neri ) from the Gulf of Mexico indicate a depth distribution of 365-1,455 m (Pequegnat 1970), while Luckhurst (in press) reported golden crabs from 786 to 1,462 m near Bermuda. Although a broad bathymetric range for the species is likely, maximum abundance occurs be- tween 367 and 549 m in our study area. This depth range coincides with that reported by Stone and Bailey (1980) for maximum trap catches of G. quinqiiedens along the Scotian Shelf and ap- proximates the limits (320-530 m) determined by Wigley et al. (1975) by trawl and photographic methods to be most productive for that species off the northeastern United States. Information on sediment composition taken co- incidentally with fishing activities suggests that abundance of both G. fenneri and G. quinquedens is influenced by sediment type at these optimum depths. Our catches were highest on substrates containing a mixture of silt-clay and fora- miniferan shell. In contrast, no golden crab were collected on rock and coral rubble bottom such as was encountered in the 550-640 m stratum. Other studies have described an association of G. quin- quedens with soft substrates. Wigley et al. (1975) noted that bottom sediments throughout the area surveyed for red crab from offshore Maryland to Corsair Canyon (Georges Bank) consisted of a soft, olive-green, silt-clay mixture. If golden crabs preferentially inhabit soft substrates, then their zone of maximum abundance may be limited within the South Atlantic Bight. Surveys by Bullis and Rathjen (1959) indicated that green mud occurred consistently at 270-450 m between St. Augustine and Cape Canaveral, FL (30''N and 28°N). This same depth range from Savannah, GA to St. Augustine was generally characterized by Bullis and Rathjen (1959) as extremely irregu- lar bottom with some smooth limestone or "slab" rock present. Our study indicates, however, that the bottom due east between Savannah and St. Catherines Island, GA at 270-540 m consists of mud and biogenic ooze. Further north from Cape Fear, NC to Savannah, bottom topography be- tween 270 and 450 m is highly variable with rocky outcrops, sand and mud ooze present (Low and Ulrich 1983). Additional information on sedi- ment type during future fishing efforts will be necessary before any validation of sediment pref- erence by golden crab can be made. The catch data for golden crab in our survey compares favorably with catch rates reported by Otwell et al. (1984) in the Gulf of Mexico. Al- though their study was not intended to assess the resource, they reported mean catch per trap val- ues of 7.4-8.4 for the nested design fished between 210 and 340 fathoms. Information on catch rates of red crab from trap surveys and the fishery is perhaps more relevant to our study. Ganz and Herrmann (1975) reported an overall unculled mean catch per pot of 40-93 red crabs off southern New England; their study used four types of dou- ble parlor offshore lobster pots. An average catch of 26.8 red crabs per trap (conical-top entry) was reported in 360-540 m depths on the Scotian Shelf by Stone and Bailey (1980). The only available information on weight per trap was provided by Gerrior (1981) who found seasonal catch rates that ranged from a low of 8.4 kg in March to a high of 11.1 kg per pot in June. Although com- parison of catch per unit of effort between these studies is questionable because trap type and fishing duration, as well as physical features of the sampling areas differ, catch per trap of golden crab in depths of maximum abundance off South Carolina and Georgia appears promising. Comparison of catches (no. /trap) between the Fathoms Plus trap and the Florida trap clearly indicate superiority of the latter for golden crab. These two traps also differed in the size and weight of individuals caught, with larger and heavier golden crab occurring in the Fathoms Plus trap. Advantages of the Fathoms Plus traps for commercial fishing operations would include their lighter weight, ease of handling, and stack- able configuration which conserves deck space. Differences observed between traps may be re- lated to trap design which affects success of entry and maximum catch (Miller 1980) or behavioral interactions which affect probability of capture (Richards et al. 1983). Although no studies have been done to evaluate behavior of G. quinquedens or G. fenneri in regard to traps, responses of the spider crab, Hyas araneus, and the rock crab, Cancer irroratus, to top and side entry traps were reported by Miller (1980). He found success of entry by C. irroratus was greater, escapement was reduced, and fewer agonistic encounters oc- curred in top entry traps. In a complementary study, however. Cancer productus had highest 555 FISHERY BULLETIN: VOL. 85, NO. 3 Figure 4. — Scanning electron micrograph of testis and vas deferens from male Geryon fenneri. A. Testis; Maturing germ cells (spermatids, ST) surrounded by sustentacular tissue. De- veloping sperm (D-S), cytoplasmic spike (SP); x 3200. Scale bar 3 |xm. B. Testis: A developing sperm (D-S) possessing partial to fully formed cytoplasmic spikes (SP); X 3200. Scale bar 3 jim. 556 WENNER ET AL.: EXPLORATION FOR GOLDEN CRAB Figure 4.— Continued. C. Vas deferens:- Mature multiple stellate sperm (S) showing cytoplasmic spike (SP); X 3840. Scale bar 2 \i.m. D. Vas deferens: Pockets (P) within the vas deferens previously occupied by the mature stellate sperm; x 1600. Scale bar 10 (xm. 557 FISHERY BULLETIN. VOL. 85, NO. 3 success in entering a side entry trap whose en- trances were parallel to the current (Miller 1978). Although our traps were deployed parallel to sur- face current, their orientation on the bottom rela- tive to bottom current is unknown. We are assum- ing that golden crabs were successful in locating the entrance and were retained longer in the top entry Florida trap than in the Fathoms Plus trap. It is possible, however, that golden crab were equally or more successful in locating the side entrances of the Fathoms Plus trap but that es- capement, especially of smaller golden crab, was higher. This would explain the capture of few- er but larger individuals by the Fathoms Plus trap. The overwhelming dominance of males in this study contrasts with results reported in other geo- graphic areas for golden crab. Luckhurst (in press) noted that sex ratio in his sample (n = 244) of G. fenneri from Bermuda waters was approxi- mately 1:1. Otwell et al. (1984) noted that males tended to be more abundant at greater depths (>540 m) in the Gulf of Mexico; however, they cautioned that trap design may influence the per- centage of male crabs caught. Commercial crab- bers noted a decline in catch rates and number of male G. fenneri with increasing depth on the slope in the eastern Gulf of Mexico (National Marine Fisheries Service fn. 5). We also found increased abundance of females at greater depths, although our results are limited due to the small number of females collected. This is apparently not an artifact of sampling with only the Fathoms Plus trap in the deepest stratum since more fe- males were collected in the Florida trap than with the Fathoms Plus trap when only strata 1-3 were considered. Segregation of the sexes by depth has been observed in several studies of G. quinque- dens. Wigley et al. (1975) collected more female red crabs than males, but this dominance was limited to intermediate depths (320-503 m). Ganz and Herrmann (1975) similarly noted dominance by male red crab at depths >685 m off Rhode Island. This same pattern was noted for red crab in the vicinity of Norfolk Canyon where females were more abundant than males from depths <600 m (Haefner and Musick 1974; Haefner 1978). In Canadian waters, however, female red crabs were reported by Stone and Bailey (1980) to be considerably less abundant than males. Al- though they attributed this discrepancy to trap bias, another study in the same general area found females were present but highly contagious in distribution. Whether seasonal migrations re- lated to mating or spawning occur as hypothe- sized by Wigley et al. (1975) for G. quinquedens remains to be substantiated. What is evident from our results is that male G. fenneri are domi- nant in depth strata where catch per unit of effort is highest. Size-related distribution of G. fenneri with depth, similar to that reported for red crab, may occur in the South Atlantic Bight. We found the largest crabs in the shallowest (274-366 m) and deepest (733-823 m) strata. A clear trend of size- related up-slope migration such as Wigley et al. (1975) reported for G. quinquedens is not appar- ent, however, because of trap bias for capture of larger crabs of both sexes. Otwell et al. (1984) also noted no pattern in size of golden crab by depth for either sex. Tagging studies of red crab off southern New England provided no evidence for migration patterns and indicated instead that tagged crabs seldom moved more than 20 km from their site of release (Lux et al. 1982). The size composition of golden crab from our study showed that crabs become trappable as small as 85 mm CW but that the greatest propor- tion of trapped individuals is >100 mm CW. Over 90% of all individuals collected exceeded 114 mm CW which is the minimum size of red crab ac- cepted for commercial utilization (Wigley et al. 1975). A much smaller proportion (52%) of golden crab >114 mm was indicated in size-frequency distributions of trap-caught golden crab near Bermuda (Luckhurst in press). Although Otwell et al. (1984) did not present size and weight- frequency data for golden crab in the Gulf of Mex- ico, they found mean size of male crabs ranged from 155 to 163 mm with mean weight extremes of 1.07-1.15 kg, while females were smaller with mean CW ranging from 119 to 135 mm and mean weight extremes of 0.45-0.50 kg. These data and those from our study suggest that the average size of golden crab from the South Atlantic Bight and Gulf of Mexico is larger than the average size of red crab reported along the eastern United States and Canada. Wigley et al. (1975) reported average width of male G. quinquedens was 99 mm with an average weight of 413 g. Average width of all females from their study was 90 mm with a mean weight of 244 g. Comparisons of size compo- sition between the two studies must be qualified, however, by a caveat that differences in sampling methods probably influenced sample statistics. The apparent larger size of golden crab may be better substantiated by maximum width and weight measurements, which for our study were 558 WENNER ET AL.: EXPLORATION FOR GOLDEN CRAB 193 mm and 2,109 g, respectively. These values were markedly larger than those reported for red crab in the vicinity of Norfolk Canyon (Haefner 1978), off northeastern United States (Wigley et al. 1975), or the Scotian Shelf (Stone and Bailey 1980; McElman and Elner 1982). The small number of females collected during the first year precludes any definitive statements regarding ovarian cycles or spawning patterns. Ovarian developmental stages are similar to those reported by Haefner (1977) for G. quinque- dens. We also found his use of vulvae condition as an external indicator of copulation to be fairly reliable, but examination of the seminal recepta- cles for sperm or spermatophores provided the only true indication of mating. Tentative inter- pretations on ovarian development, vulval condi- tion, and presence of seminal products suggest that females may become sexually mature at 97 mm CW. Haefner (1977) suggested that female G. quinquedens become sexually mature within the intermolt size of 80-91 mm CW. A lack of ovigerous females in our first-year sampling effort could be indicative of a restricted spawning season similar to that reported for red crab (Haefner 1977; Wigley et al. 1975). Absence of ovigerous females from our samples, however, may be related to the small number of female golden crab collected. Observations on molting and mating of a fe- male (110 mm CW), which had been held in a refrigerated aquarium since February 1986 and had completed ecdysis in late May 1986, con- firmed that female golden crab molt just before mating occurs. This behavior, as well as the ob- served premolt embrace, has been described for G. longipes (Mori and Relini 1979), although it has not been reported previously for either G. quin- quedens or G. fenneri. Stage of ecdysis is an important factor affecting meat condition and yield in golden crab. Crabs which have recently molted generally have a very poor meat yield and are not marketable^. Since most golden crab in the intermolt stage had blackened abraded areas or poecilasmatid barna- cles on the exoskeleton, their presence was useful in distinguishing premolt from postmolt crabs which were brighter in color and had few abra- sions. 8W. Lacy, Seafood Marketing Section, South Carolina Wildlife and Marine Resources Department, Charleston, SC 29412, pers. commun. 1985. ACKNOWLEDGMENTS This project could not have been completed without the assistance of numerous individuals and organizations. We are most grateful to H. R. Beatty and D. Oakley of SCWMRD who helped extensively with all aspects of the project. Initial phases of this project benefited from our discussions with commercial fishermen H. Holley and J. Busse; C. A. Wenner (SCWMRD); W. Rathjen (Florida Institute of Technology); R. Manning (U.S. Na- tional Museum of Natural History); and S. Otwell (University of Florida). We also appreciate the hospitality and information on the Bermuda fish- ery provided by J. P. Ingham (Pathfinder Fish- eries) and B. Luckhurst (Bureau of Agriculture and Fisheries). We thank all colleagues who as- sisted on cruises and the captains and crews of the RV Oregon and RV Lady Lisa. Technical assis- tance was provided by W. Roumillat for photo- graphic preparations; B. Ashcraft (Medical Uni- versity of South Carolina) for SEM preparations; K. Swanson who drafted the figures; and M. Lentz who typed the manuscript. Financial support for this project was provided by the Gulf and South Atlantic Fisheries Development Foundation under contract 27-08-44394/79489, the Mar- quette Foundation, the South Carolina Wildlife and Marine Resources Department, and the South Carolina Sea Grant Consortium. LITERATURE CITED Beyers, C. J. de B , and G G. Wilke 1980. Quantitative stock survey and some biological and morphometric characteristics of the deep-sea red crab Geryon quinquedens off southwest Africa. Fish. Bull. S. Afr. 13:9-19. BuLLis, H. R., Jr., and W F. Rathjen. 1959. Shrimp explorations off southeastern coast of the United States (1956-1958). Comm. Fish. Rev. 21(6):1- 20. Christiansen, M. E 1969. Crustacea Decapoda Brachyura, Mar. Invertebr. Scand. (Oslo) 2:1-143. Ganz, a R., and F J. Herrmann 1975. Investigations into the southern New England red crab fishery. Rhode Island Dep. Nat. Resour., Project Rep., New England Fish. Devel. Prog., Natl. Mar. Fish. Serv., 78 p. Gerrior. P 1981. The distribution and effects of fishing on the deep sea red crab, Geryon quinquedens Smith, off southern New England. M.S. Thesis, Southeastern Massachu- setts Univ., North Dartmouth, MA, 130 p. Haefner. P. A., Jr., and J A Musick. 1974. Observations on distribution and abundance of red 559 FISHERY BULLETIN: VOL. 85. NO. 3 crabs in Norfolk Canyon and adjacent continental slope. Mar. Fish. Rev. 36(l):31-34. 1977. Reproductive biology of the female deep-sea red crab, Geryon quinquedens, from the Chesapeake Bight. Fish. Bull., U.S. 75:91-102. 1978. Seasonal aspects of the biology, distribution and relative abundance of the deep-sea red crab Geryon quin- quedens Smith, in the vicinity of the Norfolk Canyon, western North Atlantic. Proc. Natl. Shellfish. Assoc. 68:49-62. Low, R. N., AND G F ULRICH 1983. Deep-water demersal finfish resources and fisheries off South Carolina. S.C. Mar. Resour. Cent. Tech. Rep. No. 57, 24 p. LUCKHURST, B. In press. Discovery of deep-water crabs (Geryon spp.) at Bermuda - a new potential fishery resource. Proc. 37th Annu. Gulf Carribb. Fish. Inst. Lux, F. E.. A R. Ganz, and W. F. Rathjen 1982. Marking studies on the red crab Geryon quinque- dens Smith off southern New England. J. Shellfish Res. 2(l):71-80. Manning, R. B . and L B Holthuis 1981. West African Brachyuran Crabs (Crustacea: Deca- poda). Smithson Contrib. Zool. 306, 397 p. 1984. Geryon fenneri , a new deep-water crab from Florida (Crustacea: Decapoda: Geryonidae). Proc. Biol. Soc. Wash. 97:666-673. McElman, J F . AND R W Elner. 1982. Red crab (Geryon quinquedens) trap survey along the edge of the Scotian shelf, September 1980. Can. Tech. Rep. Fish. Aquat. Sci. 1084, 12 p. Miller, R J 1978. Entry of Cancer produc^us to baited traps. J. Cons. Int. Explor. Mer 38:220-225. 1980. Design criteria for crab traps. J. Cons. Int. Explor. Mer 39:140-147. MONOD, Th. 1956. Hippidea et Brachyura ouest-africains. Mem. Inst. Fr. Afr. Noire 45:1-674. Mori, M.. and G Relini 1979. Mating behaviour of Geryon longipes A. Milne Ed- wards 1881 (Crustacea: Decapoda: Brachyura) in captiv- ity. Quod. Lab. Tecnol. Pesca 3:173-178. Otwell, W. S.. J Bellairs. and D Sweat 1984. Initial development of a deep-sea crab fishery in the Gulf of Mexico. Fla. Sea Grant Coll. Rep. No. 61, 29 p. Pequegnat, W. E 1970. Deep-water brachyuran crabs. In W. E. Pequag- nat and F. A. Chace, Jr. (editors), Contributions on the biology of the Gulf of Mexico, p. 171-204. Texas A&M Univ. Oceanogr. Stud. 1. Rathbun. M. J 1937. The oxystomatous and allied crabs of Amer- ica. U.S. Natl. Mus. Bull. 166, 278 p. Richard, R A . J S Cobb, and M J Fogarty 1983. Effects of behavioral interactions on the catchabil- ity of American lobster, Homarus americanus, and two species of Cancer crab. Fish. Bull, U.S. 81:51-60. Richer. W E 1973. Linear regression in fishery research. J. Fish. Res. Board Can. 30:409-434. SOKAL, R R , AND F J ROHLF 1983. Biometry. 2d ed. W. H. Freeman and Co., San Francisco, 850 p. Stone. H . and R F J Bailey. 1980. A survey of the red crab resource on the continental slope, N. E. Georges Bank and western Scotian shelf. Can. Tech. Rep. Fish. Aquat. Sci. 977, 9 p. WiGLEY, R. L., R. B. THEROUX, AND H. E. MURRAY. 1975. Deep-sea red crab, Geryon quinquedens , survey off northeastern United States. Mar. Fish. Rev. 37(8):l-27. 560 DIFFERENTIATION OF MITOCHONDRIAL DNA IN ATLANTIC HERRING, CLUPEA HARENGUS I. KORNFIELD AND S. M. BOGDANOWICZ' ABSTRACT To investigate genetic relationships among spawning stocks of Atlantic herring, Clupea harengus, in the Gulf of Maine and Gulf of St. Lawrence, mitochondrial DNAs from ripe females at three localities were examined by restriction endonuclease analysis. Using seven variable restriction enzymes, mtDNAs from 69 completely characterized individuals produced 26 composite digestion patterns. The majority of individuals (65%) possessed composites which were common to two or more spawning localities; the other individuals displayed locality specific "unique" composites. Analysis of relation- ships among these unique composites suggested that some may have been derived from other areas. These results are not consistent with the idea that separate genetic stocks of Atlantic herring exist in the Gulf of Maine. The relationships among discrete spawning stocks of Atlantic herring, Clupea harengus , are problematical. A large number of stocks and stock complexes are recognized throughout the eastern and western North Atlantic; these delineations are based largely on meristic characters, spawn- ing time, and spawning location. Tagging studies in the western North Atlantic have shown exten- sive migration and mixing of stocks during nonre- productive periods (Creaser et al. 1984). More limited studies of spawning fish have demon- strated that some tagged individuals returned to their spawning locations (Wheeler and Winters 1984). Recent work has advanced the hypothesis that specific environmental attributes essential for growth and survival of larval herring largely determine where Atlantic herring will spawn (lies and Sinclair 1982). The notion that spawn- ing occurs near areas suitable for larval retention could explain the discontinuous or patchy distri- bution of spawning areas. Similarly, the occur- rence of fall spawning and spring spawning At- lantic herring stocks may be a function of completion of larvae growth and metamorphosis constrained by resources within the larval reten- tion area (Sinclair and Temblay 1984). There is thus a reasonable model to explain the existence of geographically or temporally discrete spawn- ing stocks. However, the genetic structure among these different spawning groups is unresolved. Implicit in the Atlantic herring stock concept is 'Department of Zoology and Center for Marine Studies, Uni- versity of Maine, Orono, ME 04469. Manuscript accepted March 1987. FISHERY BULLETIN: VOL. 85, NO. 3, 1987. the idea that individual fish belong to defined groups by virtue of returning to specific spawning sites. If this is the case, there should exist high genetic continuity among individual Atlantic herring within stocks and relatively lower conti- nuity among stocks. That is, genetic differences should be observable among stocks. Unfortu- nately, meristic characters useful for stock defini- tion are under environmental influence and have a complex genetic basis. Electrophoretic charac- terization of allozyme variation should poten- tially permit identification of genetic discontinu- ities among stocks. However, despite the availability of a large number of polymorphic markers and adequate sample sizes, significant genetic heterogeneity among Atlantic herring stocks has not been demonstrated (Anderson et al. 1981; Kornfield et al. 1982; Grant 1984; Riv- iere et al. 1985). The inability of allozyme analy- sis to differentiate among herring stocks could occur for two alternative reasons. Herring stocks could have originated so recently that there has been insufficient time for stock specific allozyme variation to accumulate. Further, natural selec- tion may be acting to homogenize allele frequen- cies that may characterize stocks. Thus, standard allozyme analyses may not be sufficiently sensi- tive to detect genetic variation which distin- guishes stocks. Alternatively, herring stocks could be largely composed of individuals that do not return to natal spawning sites. Under this explanation, herring stocks would not represent discrete genetic groups but rather random assem- blages of spawning individuals. Management of 561 FISHERY BULLETIN: VOL. 85, NO. 3 exploited herring stocks could differ dramatically depending upon which alternative is correct (MacLean and Evans 1981). Restriction endonuclease analysis of mitochon- drial DNA (mtDNA) has, in recent years, uncovered substantial genetic variation in natu- ral populations (Brown 1983). The technique is potentially much more sensitive than conven- tional allozyme analysis for characterizing popu- lation structure and has been successfully ex- ploited to discriminate groups not detectable with allozymes (e.g., Avise et al. 1986; R. W. Chapman unpubl. data). To further examine genetic rela- tionships among herring stocks, restrictive en- zyme digestion patterns of mtDNA were exam- ined in individuals from three spawning localities in the northwestern Atlantic. MATERIAL AND METHODS Samples of fall spawning Atlantic herring were obtained from two discrete localities in the Gulf of Maine: Jeffries Ledge, MA (lat. 42°50'N, long. 66°30'W; 9 September 1984) and Trinity Ledge, NB, Canada (lat. 45°20'N, long. 65°30'W, Sep- tember 1984; 30 August 1985). A sample of spring spawning Atlantic herring was collected from off Pt. Escuminac, Gulf of St. Lawrence, NB (lat. 47°01'N, long. 64°40'W; 12 May 1985) (Fig. 1). All Atlantic herring were collected during peak re- production. Samples were frozen in the field and stored at — 80°C for up to 6 months prior to anal- ysis. Lansman et al. (1981) provided a useful review of the application of mtDNA to population stud- ies. mtDNA was prepared from egg tissue (11-15 g/female) by the rapid phenol extraction proce- dure of Chapman and Powers (1984). After the final chloroform extraction, mtDNA in the aqueous phase was precipitated in 95% ethanol in the presence of 3 M sodium acetate, dried under vacuum and dissolved in 10 mM Tris, pH 7.5. Figure 1. — Collection localities for Atlantic herring. (1) Jeffries Ledge, MA; (2) Trinity Ledge, NB; (3) Point Escuminac (St. Lawrence), NB. KORNFIELD AND BOGDANOWICZ: MITOCHONDRIAL DNA IN ATLANTIC HERRING Samples were digested with 16 six-base restric- tion endonucleases (Table 1 ) under conditions rec- ommended by suppliers (Bethesda Research Labs, New England Biolabs). Just prior to addi- tion of restriction enzymes, samples were incu- bated with Ribonuclease A (RNase) at 60°C for 5 minutes and allowed to cool to 37°C. Restriction fragments were separated by horizontal elec- trophoresis in 1% agarose gels. Hindlll digests of lambda DNA were used as molecular weight standards on all gels. Gels were stained for 60 minutes with 0.5 g L~^ ethidium bromide and destained for 30 minutes in 5 mM MgS04 prior to photography. The relative mobilities of mtDNA and lambda fragments were measured from pho- tographs with a stereomicroscope. Molecular weights of restriction fragments were calculated from least squares third order polynomial regres- sions of log- transformed lambda fragment mobili- ties. RESULTS Restriction digests of mtDNAs prepared by the rapid phenol extraction procedure well resolved, repeatable digestion patterns (Fig. 2). Two en- zymes. Bam HI and Sail did not digest the her- ring mtDNA molecule. Variant digestion pat- terns were noted for the majority of enzymes examined (Table 1) and were common both within Figure 2. — Ethidium bromide stained agarose gel of mtDNA digestion patterns of Atlantic herring, Clupea harengus (lanes 2-8). Samples were digested with BstYAl llanes 2, 3; phenotypes B, C), £co RI (lanes 4, 5; phenotypes A, C), and fig/ II (lanes 6, 7; phenotypes A, B). Standard (lanes 1, 8) is a //mdlll digest of lambda DNA. Table 1. — Digestion patterns of Atlantic herring mtDNA produced by six-base restriction endonucleases^. Superscripts denote homolo- gous fragments, measured independently. Apa\ BstEW BgtW B H I B B 8,800a 8,800a 7,200e 6,100= 900ti 960d 900b 7,200e 5,300 3,450 900b 6,350 6,100c 2,450^1 960d 900b 8,800a 7,200e 1,100 7,350 6,100= 2,450b 900b 1 1 ,830d 4,670a 550b 10,820d 4,670a 1,1 60C 550b 15,160 1,150= 550b 12,200 4,700a 1 1 ,700d 3,650 1,340 550b 9,940 4,1 30a 1 ,990b 1,270= 6,590d 14,000 4, 1803 1,990b 3,230e 1,270= 1 ,960b 1 ,220= 16,900 16,760 16,850 16,760 17,100 16,800 17,050 17,200 16,860 16,900 17,240 17,330 Dra\ EcoRI EcoRV B B B 7,600 3,700 2,570a 2J310b 11,370 2,530a 2,290b 9,460a 4,220 3,010b 14,020 3,130b 9,400a 7,250 16,170 16,190 16,690 17,150 16,650 PvuW 8,680a 6,000 2,230b 16,910 Sac\ 8,200 6,490 2,030 1,000 17,720 8,600a 8,200 16,800 Xba\ 14,500 2,270b 16,770 H/ndlil A 13,500 2,820 1,000 17,320 Xho\ 17,180 17,260 Kpn\ A 8 15,200 17,000 1,910 17,110 17,000 ^Two additional enzymes, Bamh\ and Sail produced no (or one) cuts. 6,600d 3,900 3,250e 1,970b 1 ,220= 16,940 Pst\ A 10,790 5,850 16,640 Xmn\ A B C A A A B A 7,480 12,840 13,600 9,800 5,350 14,000 16,500 7,940 5,590 1,550a 1 ,560a 7,300 4,880 2,690 3,780 1,580a 1,420b 1 ,400b 2,890 2,330 1,420b 850= 2,060 1,810 900= 2,050 17,230 1,320 16,970 16,660 16,560 17,100 16,690 16,500 17,180 563 FISHERY BULLETIN: VOL. 85, NO. 3 and among population samples. For polymorphic restriction enzymes, all digestion profiles of vari- ants were consistent with the hypothesis of single nucleotide substitutions. No mtDNA size vari- ants, resulting from additions or deletions of DNA, and recently found in a number of fish groups (Bermingham et al. 1986; R. W. Chap- man^), were observed. The mean mtDNA genome size, found by aver- aging the sums of all digestion patterns (Table 1), was 16,990 bp (base pairs) ± 620 bp (SD). In quantifying molecular size from ethidium bro- mide stained agarose gels, two constraints must be noted. First, variation associated with mea- surement of fragment mobilities is inevitable. Because of the nonlinear relationship between fragment mobility and molecular size, slight measurement errors can produce large variations in estimated sizes, particularly for fragments with low mobilities. As a consequence, ho- mologous cleavage fragments (those which con- sistently exhibit the same mobility on a single 2R. W. Chapman, Chesapeake Bay Institute, The Johns Hop- kins University, Shandy Side, MD 20764, pers. commun. De- cember 1985. Table 2. — Distribution of mtDNA composite digestion patterns in samples of Atlantic herring. Composite Composite mtDNA digestion Jeffries Trinity Ledge St. designation pattern! Ledge 1984 1985 Lawrence 1 AAAAAAA 5 2 5 2 AABAAAA 2 2 2 6 3 BAAAAAA 2 3 6 4 AABAAAB 1 1 3 1 5 ABAABAA 1 1 2 1 6 BABAAAA 1 7 AABACAA 1 1 8 AAAADAB 1 9 lAAAAAA 2 10 DAAAAAA 1 11 CAAAABA 1 12 BAEAAAA 1 13 HABAAAA 1 14 AAAAABB 1 15 CAAAAAA 1 16 BABAAAB 1 17 ABABBAA 1 18 ABBACAA 19 AAACCAA 20 lABAAAA 21 AACAAAA 22 BADAAAA 23 CABAAAA 24 AADAAAA 25 ACDAAAA 26 ADDABAA Ts 8 22 26 1 Letters (from left to right) are digestion patterns for Apa\, BglW, BstEW, EcoRI, EcoRV, Kpn\, and Xho\ (Table 1). agarose gel) may yield different molecular size estimates, e.g., Apal fragment "a". Table 1. Sec- ond, mtDNA cleavage fragments less than 500 bp could not be routinely scored on ethidium bro- mide gels because of their low absolute staining intensities and fluorescent background in this re- gion (Fig. 2). Regardless of the above constraints, individual cleavage fragment phenotypes could be consistently determined. Seven polymorphic restriction endonucleases (Apal, Bglll, BstEll, EcoRI, EcoRV, Pvull, and Xhol) were used to generate composite digestion patterns for individual Atlantic herring. Twenty- six unique composite digestion patterns were ob- served in 69 completely characterized individual Atlantic herring. The distribution of these com- posites with respect to spawning locality is given in Table 2; five common composites (nos. 1-5) were observed to occur at all three spawning lo- calities. Shared fragment similarity was calculated pairwise for all composites and was used to generate estimates of p, percent sequence diver- gence (Upholt 1977). Estimated sequence diver- gence among composites varied considerably, mean = 1.66% +/- 0.91 (SD), range 0.19% - 4.37% (Table 3). Phenetic relationships among composites were examined by UPGMA (Un- weighted Pair Group Method of Arithmetic aver- aging) clustering (Sneath and Sokal 1973) of se- quence divergence (Fig. 3). Two major clusters were noted: one involving three composites (5, 17, 26) and the other including all other composites. Composites from both clusters were present in all spawning populations. A network of relationships among composites was constructed by connecting composites in in- crements of single site gains or losses to minimize the total number of restriction site changes re- quired (Fig. 4). Sixteen equally parsimonious net- works requiring 29 steps were generated. Com- posite number 1 can be considered central because it is the most common pattern observed and also occurs in the Eastern Atlantic (S. M. Bogdanowicz unpubl. data). DISCUSSION Based on the occurrence of geographically dis- junct spawning groups and homing of some tagged individuals, it has been tacitly accepted that Atlantic herring stocks are reproductively isolated. The basic analytical premise of this study was that restriction endonuclease analysis 564 KORNFIELD AND BOGDANOWICZ: MITOCHONDRIAL DNA IN ATLANTIC HERRING 05 c c < c tn 0) in O Q. E o o < Q CD O c CD cr 0) c 0) o k_ a> Q. « E to LU C\J CO CO r~-c\ja)cocotDr--cD'--cDco(£)(DLor~-'^a)co T;rcoO'^r--r^'^'— cDOJor^ocoO'— c\ir^-^t-^05'^c\joqr-- cNic\icocr)Ocnoocococri'^(rjco-^n-^'-->-cr)cocJc\icoT-^T- OT^-'^r^-ootDr^cooocD-^O'Tco-'rcot-^-cqcopnojcp-'j; ina)t-c\jr-tDC\ioococoa)0-^oor-cocj>cocoof^ir)(0 , •^Cvja)t--r^Trcocoir)cO'-omooO'Jc\i''-^T-: c7)CMcO''-ojr-~'--r--M--'-cD'-r^i~~^O'*'^'cr)-*CO'<3-OOCO-a)'- •r-d-'-t-'-'-dcjT-cvJoJoJ'^i-^'-^'-cNJ '3-OOCMh-C\J'>a-r^'^C75'3-I---'^CnCOCOC3) , or^3-r-~'^'*'d-->-coocO''--^ir) | •^T-ddcodT-T^-^'^'oO'-'-'-'-^ COr^h-COC\lC\IOCJ)t~-OC\JCO^-C7) , r--ooa)i---coco'5}--a5coooco , ■■-c\jLOi^a)Ot-^->l;'r-coh-;cq dddd-'-'^dT^'^T^-'-^'r^ I^COOCVJOCOOJ'^OCOCVJ , co(D'*cncj>''-ppP'-co d'-d'-CNJi-'r^cO-'-^T-cvJ ■t-OOr^OJCM-^'tCDCJi , OJ'^'^-^ocnocDO--- O'-'-CvicO-.-cJcJi--^ r^cooc\joc\jc\jTto , coco' , h~ooLn'^cocD'q-oo t^ o> O) r^ 05 ■^ -T , CD -^ ■* O) Tt o t^ d »- -^ d T- cvj ■■- ■■- C\J O -^ C\J CO , CT> •* ■^ en CO LD o di -^ di r C\J CM UO (J5 CO 1 '- ■^ ■.- C\J ■^ C\J o , CJ) TJ- Tt 1 o O y- Cvj ■>3- ^ 1 d d ' C\J 1 o ■■-CMco'^LOcor--cocJ>0'-c\jcoTrLncDt^cocDO'-c\ic^'^ioco ■■-•i-,-i--t-i-i->-T-T-C\JC\IC\J00C\JC\JOJ 565 FISHERY BULLETIN: VOL. 85, NO. 3 -c {1 -c 20 10 3 6 22 \7 1 13 • 2 24 • 21 -25 ■ 15 23 11 8 14 4 16 ■ 7 ■ 18 ■ 19 • 5 17 26 3.0 2.0 1.0 DIVERGENCE Figure 3. — Phenetic relationships among mtDNA composite cleavage patterns of Atlantic herring. Estimates of sequence divergence were clustered by UPGMA. 12 > 1 14 3 ) ^ t 22 2^ • 1 - 16 25 21 Figure 4. — Cladistic relationships of 26 composite cleavage pat- terns of Atlantic herring mtDNA. Composites are connected parsimoniously to minimize the number of restriction site changes required. Shaded numbers refer to composites observed at all three spawning locations. Crossbars on connecting lines indicate minimum number of site changes required to connect adjacent composites; arrows indicate direction of site losses. Lo- cality symbols: square - Trinity Ledge; triangle - Jeffries Ledge; diamond - St. Lawrence. of mtDNA should have been able to differentiate among such reproductively isolated populations. However, the spawning groups studied were not fully distinguishable by composite mtDNA diges- tion patterns generated by six-base restriction en- donucleases; no absolute stock markers were present. Six of the twenty-six composite designa- tions, representing more than 65% of all individu- als, were shared by at least two geographically distinct spawing groups. The occurrence of common composites in all spawning populations could occur for at least two reasons: First, commonality could reflect recent and/or ongoing gene exchange among popula- tions. Consistent with this idea, there is no associ- ation between frequencies of common composites (nos. 1-5) and spawning locality (G = 6.29, p 0.5, Sokal and Rohlf 1981). As is generally acknowl- edged, small numbers of individuals migrating among populations are sufficient to homogenize different groups (Allendorf 1983). The absence of two common composites (1 and 3) in the 1984 Trinity Ledge sample might be due to the stochas- tic effect of small sample size, though this obser- vation could also imply some element of temporal instability in composition. Second, and alterna- tively, these common composites could represent ancestral mtDNAs which were widespread prior to any genetic isolation of populations. That is, 566 KORNFIELD AND BOGDANOWICZ: MITOCHONDRIAL DNA IN ATLANTIC HERRING the occurrence of common composites need not imply current gene exchange (Avise et al. 1984; Neigel and Avise 1986); population sizes are suffi- ciently large to support the co-occurrence of com- mon ancestral composites and their derivatives. The presence of 20 composites which were specific to spawning groups suggests that there may be some degree of genetic isolation among stocks. Given the limited number of individuals sampled, it is difficult to know whether "unique" composites are actually restricted to specific stocks. For example, rather than increasing the abundance of previously observed unique com- posites, the second sample from Trinity Ledge generated additional composites. There is thus little indication that "unique" composites may be useful in defining stocks. The great composite di- versity displayed in the samples of Atlantic her- ring most probably reflects the very large popula- tion sizes involved. In the absence of gene flow among spawning populations, we would expect a unique composite to be found in the same population as its most probable precursor. In three out of seven in- stances, precursors of unique composites occurred in diff'erent spawning populations (this result holds for all other equally parsimonious net- works). For example, composite 9, which only oc- curs in the St. Lawrence sample, is the immediate ancestor of composite 20 from Trinity Ledge. In addition, the two unique composites which were maximally divergent (12 steps) occurred in the same population (Trinity Ledge 1985). These con- siderations, as well as the absence of any consis- tent geographic pattern of unique composites are consistent with the idea of gene flow. Evidence for the ability of mtDNA analysis to detect subtle population differentiation is com- pelling (Avise et al. 1979; Lansman et al. 1981; Wilson et al. 1985; Bermingham and Avise 1986). However, since differentiation of mtDNA restric- tion patterns is a time dependent process (Kessler and Avise 1985), it is possible that there has been insufficient time to accumulate population specific differences in Atlantic herring (Grant 1984, 1985). Atlantic herring stocks, as they cur- rently exist, can not predate the origin of the Gulf of Maine following glacial withdrawal 18,000 years ago (Kellogg 1980). In addition, since the effective population sizes of Atlantic herring stocks are very large, they would be expected to diverge only very slowly by lineage sorting (Neigel and Avise 1986). Consistent, significant genetic differences among spawning groups of Atlantic herring is a sufficient, but not a necessary, condition to regard populations as discrete stocks. Our results do not support the hypothesis that discrete Atlantic her- ring stocks exist throughout the Gulf of Maine; however, the absence of such differences does not allow us to rigorously conclude that there is gene flow among the populations in question. More comprehensive sampling of mtDNA composites within and among populations in the western North Atlantic may better allow resolution of this problem. Regardless, for the sake of preserving variability, resources like the Atlantic herring should be managed under the assumption that every spawning group is a semi-discrete genetic entity. ACKNOWLEDGMENTS We thank David Pierce (Massachusetts Fish Division), Susan Safford (University of Massa- chusetts), Clarence Bourque and Sandy Seagle (Fisheries and Oceans, Canada), and the captain and crew of the FV Barnegat (Gloucester, MA) for assistance in field collections. Thomas Dowling kindly provided computer programs for mtDNA analysis. Robert Chapman critically commented on an early draft of this manuscript; Eldredge Bermingham and an anonomyous reviewer pro- vided comments which greatly improved the final version. Our work was supported by BRSG 507 RR07161 from the National Institutes of Health and by funds from Sea Grant and the Migratory Fish Research Institute, Orono ME. LITERATURE CITED Allendorf. F W 1983. Isolation, gene flow, and genetic differentiation among populations. In C. M. Schonewald-Cox, S. M. Chambers, B. MacBryde, and W. L. Thomas (editors), Genetics and conservation, p. 51-65. Benjamin Cum- mings Publ., Menlo Park, CA. Andersson. L., N. Ryman, R. Rosenberg, and G. Stahl. 1981. Genetic variability in Atlantic herring (Clupea harengus harengus ): description of proten loci and popu- lation data. Hereditas 95:69-78. Avise. J. C C Giblin-Davidson. J. Laerm, J C Patton, and R A. Lansman 1979. Mitochondrial DNA clones and matriarchal phy- logeny within and among geographic populations of the pocket gopher, Geomys pinetis. Proc. Natl. Acad. Sci. 76:6694-6698. Avise, J. C, G. S Helfman, N. C. Saunders, and L. S Hales. 1986. Mitochondrial DNA differentiation in North At- lantic eels: population consequences of an unusual life history pattern. Proc. Natl. Acad. Sci. 83:4350-4354. 567 FISHERY BULLETIN: VOL. 85, NO. 3 AvisE. J C , J. E Neigel, and J. Arnold 1984. Demographic influences on mitochondrial DNA lin- eage survivorship in animal populations. J. Mol. Evol. 20:99-105. Bermingham, E., and J C. AVISE. 1986. Molecular zoogeography of freshwater fishes in the Southeastern United States. Genetics 113:939-966. Bermingham. E., T. Lamb, and J C Avise 1986. Size polymorphism and heteroplasmy in the mito- chondrial DNA of lower vertebrates. J. Hered. 77:249- 252. Brown. W. M. 1983. Evolution of animal mitochondrial DNA. In M. Nei and R. K. Koehn (editors), Evolution of genes and proteins, p. 62-88. Sinauer Assoc, Sunderland, MA. Chapman, R W . and D A Powers 1984. A method for the rapid isolation of mitochondrial DNA from fishes. Maryland Sea Grant Program, UM- SG-TS-84-05, 11 p. Greaser, E. P., D A Libby, and G. D. Speirs. 1984. Seasonal movement of juvenile and adult herring, Clupea harengus L., tagged along the Maine coast. J. Northwest Atl. Fish. Sci. 5:71-78. Grant, W. S. 1984, Biochemical population genetics of Atlantic her- ring, Clupea harengus. Copeia 1984:357-364, 1986, Biochemical genetic divergence between Atlantic Clupea harengus, and Pacific, Clupea pallasi, her- ring, Copeia 1986:714-718, ILES, T D., and M Sinclair. 1982. Atlantic herring: stock discreteness and abun- dance. Science 215:627-633. Kellogg, T. B 1980, Paleoclimatology and paleo-oceanography of the Norwegian and Greenland seas: glacial-interglacial con- trasts. Boreas 9:115-137. Kessler, L G , and J C Avise 1985, A comparative description of mitochondrial DNA differentiation in selected avian and other vertebrate genera. Mol. Biol. Evol. 2:109-126. Kornfield, I , B D Sidell, and P S Gagnon. 1982. Stock definition in Atlantic herring (Clupea haren- gus harengus): genetic evidence for discrete fall and spring spawning populations. Can. J. Fish. Aquat. Sci. 39:1610-1621. Lansman, R. a , R. O. Shade, J. F. Shapiro, and J. C Avise. 1981. The use of restriction endonucleases to measure mi- tochondrial DNA sequence relatedness in natural popu- lations. II. Techniques and potential applications. J. Mol. Evol. 17:214-226. MacLean, J. A , AND D O Evans. 1981. The stock concept, discreteness of fish stocks, and fisheries management. Can. J. Fish. Aquat. Sci. 38:1889-1898. Neigel, J. E., and J C Avise. 1986. Phylogenetic relationships of mitochondrial DNA under various demographic models of speciation. In S. Karlin and E. Nevo (editors). Evolutionary processes and theory, p. 515-534. Acad. Press. N.Y. Riviere, D., D Roby, A. C Horth, M Arnac, and M. F. Khalil. 1985. Structure genetique de quatre populations de hareng de I'estuaire du Saint-Laurent et de la Bale des Chaleurs. Nat. Can. 112:105-112. Sinclair, M., and M. J. Tremblay 1984. Timing of spawning of Atlantic herring (Clupea harengus harengus) populations and the match- mismatch theory. Can. J. Fish. Aquat. Sci. 41:1055- 1065. Sneath, P H a, and R R Sokal. 1973. Numerical Taxonomy. W. H. Freeman, San Franc, 359 p. Sokal, R. R , and F J. Rohlf 1981. Biometry. W. H. Freeman, San Franc, 859 p. Upholt, W B. 1977, Estimation of DNA sequence divergence from com- parison of restriction endonuclease digests. Nucl. Acids Res. 4:1257-1265. Wheeler, J. P., and G H Winters. 1984. Homing of Atlantic herring (Clupea harengus harengus ) in Newfoundland waters as indicated by tag- ging data. Can. J. Fish. Aquat. Sci. 41:108-117. Wilson, A C , R L Cann, S M Carr, M. George, U B Gyllen- sten, K M. Helm-Bychowski, R. G. Higuchi, S R. Palumbi, E M Prager, R. D. Sage, and M Stoneking. 1985. Mitochondrial DNA and two perspectives on evolu- tionary genetics. Biol. J. Linn. Soc 26:375-405. 568 POPULATION AND FISHERY CHARACTERISTICS OF ATLANTIC MENHADEN, BREVOORTIA TYRANNUS Dean W. Ahrenholz,' Walter R Nelson,^ and Sheryan P Epperly' ABSTRACT A stock assessment analysis of the Atlantic menhaden, Brevuortia tyrannus, fishery was conducted with purse seine landings data from 1940 to 1981 and port sampling data from 1955 to 1981. Virtual population (cohort) analysis was used to estimate historical stock sizes, rates of fishing, and numbers of recruits. The population exploitation rate (age 1 and older) ranged from 0.29 to 0.51 and averaged about 0.38 for the 1955-79 period. Recruitment at age 0.5 during the 1955-79 period ranged from 1.5 to 18.6 billion fish, with a mean of 5.1 billion. Classical spawner-recruitment relationships describe the data poorly. Growth and mortality data were used to examine yield per recruit for temporal and geographic fishing areas and for the entire fishery. Size at age data, while supporting an earlier hypothesis of density-dependent growth, show a trend toward slower apparent growth in the 1970's than is explained by this hypothesis alone. Yield per recruit of Atlantic menhaden dropped from 107 g for the 1970-72 period to 57 g for the 1976-78 period. A Graham-Schaefer production model estimate of maximum sustainable yield (MSY) for the 1955-79 period was 414,000 metric tons. A modified Pella-Tomlinson production model provided a MSY estimate of 557,000 metric tons. The latter esti- mate is probably unattainable given current temporal and geographic fishing patterns. Results of these analyses indicate that the Atlantic menhaden fishery suffers from growth overfishing. Some fishing activity has been conducted on At- lantic menhaden, Brevoortia tyrannus, since colo- nial times, but the purse seine fishery and factory reduction activities began in New England about 1850 (Reintjes 1969). The geographic range of the modern reduction fishery was established by the 1930's (Nicholson 1971a) and the fishery under- went substantial expansion following World War II. Good discussions of the actual fishing op- erations and types of gear involved are contained in Reintjes (1969) and Nicholson (1971a). With the exception of the 1950 fishing season, the Atlantic menhaden fishery has dominated total U.S. fishery landings in volume since 1946, when the Pacific sardine, Sardinops sagax, fish- ery was declining. This dominance continued until 1963, when, during its own decline, Atlantic menhaden landings were surpassed by the gulf menhaden, Brevoortia patronus, purse seine fish- ery. Gulf menhaden landings have dominated U.S. fishery landings since, and Atlantic men- haden currently account for about one-third of the total menhaden landings. The U.S. Fish and Wildlife Service, Bureau of 'Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Service, NCAA, Beaufort, NC 28516. ^Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Service, NOAA, Beaufort, NC; present ad- dress: Southeast Fisheries Center Miami Laboratory, National Marine Fisheries Service, NOAA, 75 Virginia Beach Drive, Miami, FL 33149. Commercial Fisheries"^ began biological investi- gations on Atlantic menhaden in 1952. Studies were initiated during what were, in retrospect, peak landing years with the goal to determine the nature of population fluctuations and variability in geographic abundance (June and Reintjes 1959). Following the marked reduction in stock abundance that occurred in the late 1960's, stud- ies were initiated to determine probable causes for the decline and to develop management op- tions to avert a second decline. The fishery for this migratory clupeid takes place primarily within states' jurisdictional waters (<3 miles from shore), and managerial authority rests with the individual states. Coast- wide management plans are cooperatively formu- lated under the auspices of the Atlantic States Marine Fisheries Commission (ASMFC), but the implementation requires separate legislative or regulatory action by each member state. Individ- ual states are not obligated to act upon coopera- tively derived plans or management actions from the ASMFC. Stock assessment studies provide fundamental scientific information required to formulate coastwide management actions. An early evalua- tion of the stock status of Atlantic menhaden, Manuscript accepted April 1987. FISHERY BULLETIN: VOL. 85, NO. 3, 1987. 3Presently the National Marine Fisheries Service (NMFS), NOAA, U.S. Department of Commerce. 569 FISHERY BULLETIN: VOL. 85, NO. 3 covering the 1955-68 fishing seasons, was pre- pared by Henry (1971). Applying virtual popula- tion methods (number of fish alive that will be caught in the future (Ricker 1958)) to landings from 1955 to 1969, Schaaf and Huntsman (1972) conducted additional analyses of the population dynamics of this resource. Stock status was again examined with production models using adjusted effort and landings through 1973 (Schaaf 1975a). Population dynamics and potential yield of At- lantic menhaden were further examined by Schaaf (1979), using estimates of numbers landed through the 1976 season; he also employed cohort analysis with Pope's (1972) approximation and the Leslie matrix (after Leslie 1945). In response to a request from the State/Federal Fishery Man- agement Program, Atlantic Menhaden Scientific and Statistical committee, a population dynamics subcommittee was formed (Federal, state, and in- dustry membership). Their report (ASMFC^) con- tained an indepth stock assessment (conducted by the Southeast Fisheries Center (SEFC) Beaufort Laboratory, NMFS) based on landings data through the 1977 season, and was the basis for the Atlantic menhaden management plan adopted by the ASMFC (ASMFC 1981). A com- puter simulation model of the fishery was devel- oped in an independent analysis and was based on the 1965 through 1978 seasons (Reish et al. 1985; Ruppert et al. 1985). The general concensus of the earlier studies (Henry 1971; Schaaf and Hunts- man 1972; Schaaf 1979; ASMFC fn. 4) was that the Atlantic menhaden stock was being overex- ploited and concern was expressed regarding the reduced spawning stock and/or the high rate of harvest of immature fish. The primary objective of this report is to evalu- ate the stock status of Atlantic menhaden through the 1981 season. The more recent 1970- 78 fishing seasons are emphasized, notably in presentations of yield per recruit. Effort, land- ings, and biological sampling data from 1955 through 1981 are used to estimate historic popu- lation sizes, age-specific rates of fishing mortal- ity, actual and potential fishery yield, and to ex- amine the spawner-recruitment relationship. The secondary objective is to determine what historical series of events led to recent conditions 4ASMFC (1980). Report of the Atlantic Menhaden Population Dynamics Subcommittee to the Atlantic Menhaden Scientific and Statistical Committee on the status of the Atlantic men- haden stock and fishery. Unpubl. Rep., 68 p. Atlantic States Marine Fisheries Commission, 1717 Massachusetts Ave., N.W., Washington, D.C. 20036. in the fishery. This objective can be reasonably met by examining the geographic patterns of har- vesting rates and the relative amount of effort expended in each geographic area. The final objective is to generate some informa- tion on the relative abundance and age structure of the menhaden stock during the earlier, pre- sampling period of 1940-54. This is accomplished by comparing, with inferences, the geographic patterns of harvest and effort distribution from the time period with port sampling data (and thus estimates of age-specific exploitation rates) to the patterns of an earlier period when only landings and effort data are available. OVERVIEW OF LIFE HISTORY AND STOCK STRUCTURE Hypotheses of the seasonal distribution and mi- gration patterns of adult menhaden were formu- lated from observations of fish schools (June and Reintjes 1959; Roithmayr 1963) and analysis of age-length distributions (Nicholson 1971b). These hypotheses were later supported by results of tagging studies (Dryfoos et al. 1973; Nicholson 1978). Much of the population is believed to over- winter south of Cape Hatteras to northern Flor- ida, and in late winter begins moving north. By summer, adult menhaden are normally found in dense schools in open coastal waters, bays, and sounds from northern Florida to Maine. These fish schools are stratified by age and size, with the average length and weight increasing with in- creasing latitude. In September, the most northerly portion of the population begins a southerly movement. During November, most of the adult population that summered in waters north of Chesapeake Bay move south around Cape Hatteras. These larger fish are followed in early December by a southward migration of young of the year that have emigrated from estu- arine systems north of Cape Hatteras. Atlantic menhaden spawning occurs to some degree during virtually the entire year, but not over the entire range at any given time. Evidence for this comes from an ovarian maturation study (Higham and Nicholson 1964) and observed dis- tributions of menhaden eggs and larvae on the continental shelf (Reintjes 1969; Chapoton^; Kendall and Reintjes 1975; Judy and Lewis SChapoton, R. B. 1972. On the distribution of Atlantic menhaden eggs, larvae, and adults. Unpubl. manuscr., 76 p. Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Service, NOAA, Beaufort, NC 28516. 570 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY 1983). These authors inferred that menhaden spawn in waters north of Long Island from May to September, in the Middle Atlantic Bight south of Long Island from March through May and again in September and October, but primarily spawn in the South Atlantic Bight from October through March. Menhaden are believed to spawn in neritic waters over most of the continental shelf, as well as in bays and sounds in the Long Island waters and northward (Reintjes and Pacheco 1966; Nel- son et al. 1977; Ferraro 1980). Higham and Nicholson (1964) concluded that menhaden do not C'pawn inside Chesapeake Bay. The buoyant eggs normally hatch in about 2 days, and the larvae have absorbed their yolk sac in about 4 days at a length of about 5 mm (Kuntz and Radcliffe 1917). The larvae subsequently enter estuarine systems when they are 10-34 mm long (June and Cham- berlin 1959; Reintjes and Pacheco 1966; Lewis and Mann 1971). The seasonality of larval immi- gration varies among geographical sites within years and also among years at the same site, owing at least, in part to environmental condi- tions. Immigration of larvae has been observed from November through April in the South At- lantic Bight, October to June in the Middle At- lantic and Chesapeake Bay areas, and May to October in waters of Long Island and northward (see Reintjes and Pacheco 1966 for literature summary; Lewis and Mann 1971). Following entry into estuarine waters, larvae progress to lower salinity waters and are fre- quently found in high abundance in smaller trib- utary estuaries, where they metamorphose into juveniles at a length of about 34 mm (June and Chamberlin 1959; Wilkens and Lewis 1971; Lewis et al. 1972). Young of the year generally remain in estuaries until the fall when most mi- grate downstream to larger rivers, bays, sounds, and into the ocean. The range in length of juve- niles in the fall has been observed from 40 to 185 mm (Kroger et al. 1974). After their estuarine emigration in the late fall and early winter, juve- nile menhaden from New England to the north- ern portion of the South Atlantic Bight migrate southward in dense schools often very close to shore (surf zone) (Kroger et al. 1971; Kroger and Guthrie 1973). There are debates whether the Atlantic men- haden population is composed of more than one stock (June 1958, 1965; Sutherland 1963; June and Nicholson 1964; Nicholson 1972, 1978; Ep- perly 1981). However, the apparently similar movement patterns make what potentially may be genetically different spawning groups insepa- rable in the fishery. Hence, the menhaden popu- lation is currently treated as one exploited and managerial stock. FISHERY DATA BASE Records of landings of Atlantic menhaden taken by purse seine and the level of effort ex- pended (vessel weeks) have been collected and maintained back to 1940. Early summaries (through 1971) are available from Nicholson (1971a, 1975) and more recently (through 1984) from Smith et al. (in press) (Fig. 1). Fishery landings have been sampled and num- bers at age landed estimated since 1955. The basic sampling methodology used was described by June and Reintjes (1959), with some recent modifications of sample size (Chester 1984). The efficacy and statistical design of the sampling methods are treated in detail by Chester (1984). The aging technique is described by June and Roithmayr (1960). Estimates of numbers at age landed for the 1955-64 seasons are available by fishing area on an annual basis only and are sum- marized by Nicholson (1975). Estimates of num- bers at age landed used in this report for the 1965- 81 seasons were available on a subseasonal basis (weekly) but varied slightly from some published annual summaries (Nicholson 1975 [through 1971]; Schaaf 1979 [through 1976]; and ASMFC 1981 [through 1980]) because of rounding differ- ences, biostatistical error corrections, and use of a different estimation methodology^. Smith et al. (in press) gave landings at age values for 1965-84. Values we used were similar except that the 1970-81 values given in Smith et al. (in press) reflect minor corrections to the data file which were made subsequent to these analyses. Size at age data were available from the port sampling data. Summaries of length and weight by age and geographic area are in Nicholson (1975) and more recently, Smith et al. (in press). Detailed treatment of seasonal length and age distributions by geographic area are contained in Nicholson (1971b, 1972). eSome of the landing estimates (1965-1976) used in the devel- opment of ASMFC 1981 were obtained from an estimation pro- cedure which used standardized, noncalendar weeks. While this procedure had some advantages relative to accuracy, it was not fully adaptable to the NMFS port sampling design. It was deemed more desirable to maintain a continuous set relative to estimation procedure. The differences are not large enough to alter analytical conclusions. 571 FISHERY BULLETIN: VOL. 85, NO. 3 1940 44 48 52 56 60 64 68 72 76 80 YEAR OF FISHING Figure 1. — Catch of Atlantic menhaden in thousands of metric tons from 1940 to 1981 (solid line) and fishing effort in vessel weeks from 1941 to 1981 (dashed line). Effort data from 1941 to 1968 are from Nicholson (1971a), catch and effort data from 1968 to 1981 are from Smith et al. (in press). DESCRIPTION OF THE FISHERY Geographic Fishing Areas For purposes of summarization and analysis, June and Reintjes (1959) divided the U.S. At- lantic coast into four geographic fishing areas and one temporal fishing area (Fig. 2). With only a change in the boundary line between the south Atlantic and Chesapeake Bay areas (Nicholson 1975), these divisions have continued to be useful to date. North Atlantic Area: Waters along the south- ern coast of Long Island, east of a line due south of Moriches Inlet, and waters northward. Middle Atlantic Area: Waters west of a line run- ning due south of Moriches Inlet (lat. 40°46'N, long. 72°44'W) on the southern coast of Long Island, southward to Great Machipongo Inlet, VA. Chesapeake Bay Area: Chesapeake Bay proper and coastal waters south of Great Machipongo Inlet, VA (lat. 37°22'N, long. 75°43'W) to 36°20'N on the North Carolina coast. South Atlantic Area: Coastal waters of North Carolina south of lat. 36°20'N to Cape Canav- eral, FL. North Carolina Fall Fishery: A temporal fish- ing area consisting of waters from Cape Hat- teras south to the southern border of North Carolina, beginning some time between the last week of October and the second week of November, depending on the arrival of migra- tory menhaden from more northerly waters, to the end of February of the next calendar year (fishing usually stops by mid-January). For standardized data summary, the week of each season that ends between 8 and 14 November is taken to be the first week of the fall fishery. Geographic Fishing Seasons With the exception of state jurisdictional waters in the Chesapeake Bay area, the begin- ning and ending of seasonal fishing activities were determined by weather and the abundance of fish. Hence, the seasons were somewhat vari- able. Fishing normally began earlier and ended later in the year in the south Atlantic area, with progressively later beginnings and earlier end- ings proceeding northward. The south Atlantic (summer) fishery usually began in late March or April and normally ceased in late October or early November. Fishing in waters adjacent to Chesapeake Bay usually began about mid-May and ceased in November, but it occasionally per- sisted to early December. In the middle Atlantic 572 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY 1955 1981 Figure 2.— Geographic fishing areas for the Atlantic menhaden purse seine fishery, and landing ports for 1955 and 1981 seasons. The number of plants operating at each port is given in parentheses. area, fishing usually began about middle to late May and continued into late October. Fishing in the north Atlantic area usually commenced in late May to mid June and continued until mid- September or early October. Current state regu- lation opens the fishing season in Virginia waters of Chesapeake Bay on the third Monday in May and closes the season on the third Friday in November. All territorial waters of Maryland are closed to purse seine fishing the entire year. 573 FISHERY BULLETIN: VOL. 85, NO. 3 Location and Number of Reduction Plants and Number of Vessels in Purse Seine Fishery During 1955, 23 plants operated at 16 ports along the U.S. Atlantic coast from Maine to Flor- ida. By 1981 this number had been reduced to 11 plants operating from 8 ports (Fig. 2). The num- ber of vessels landing fish declined from 150 dur- ing the 1955 season to 64 by 1967 (Table 1). Dur- ing 1981, 57 purse seine vessels landed menhaden. Table 1. — Number of purse seine vessels that landed Atlantic menhaden during the fishing year by area, 1955-81 (from ASMFC 1981). North Middle Chesapeake South Fall Year Atlantic^ Atlantic Bay2 Atlantic3 TotaH fishery 1955 39 48 20 34 150 51 1956 40 47 24 30 149 63 1957 33 46 25 31 144 64 1958 23 44 28 26 130 63 1959 34 45 31 25 144 59 1960 19 47 22 20 115 37 1961 21 47 23 20 117 44 1962 20 47 29 15 112 49 1963 10 46 36 16 112 46 1964 9 37 38 16 111 51 1965 6 13 38 19 84 46 1966 5 10 36 16 76 43 1967 0 4 32 16 64 46 1968 2 4 25 16 59 45 1969 3 4 22 16 51 36 1970 4 1 18 11 54 37 1971 5 2 20 11 51 32 1972 9 4 19 11 51 5 1973 10 6 23 11 58 4 1974 12 6 22 12 63 12 1975 9 5 22 14 61 17 1976 12 4 21 12 62 13 1977 12 5 24 10 64 16 1978 13 5 22 11 53 18 1979 11 4 22 13 54 18 1980 5 6 24 12 51 19 1981 8 7 23 13 57 19 'Vessels fishing from New England ports in recent years are all trawlers that convert to purse seme in summer Some fish regularly and others sporadically ^Vessels that fished only in regular season. Does not include vessels added In October and November ^Includes only vessels that landed regularly In the summer fishery. "•includes all vessels that landed fish dunng the year. Trends in Nominal Effort, Landings, and Age Composition Since the early 1940's, the Atlantic menhaden fishery has displayed a somewhat classical har- vest pattern with an historic increase to a record high, fluctuations, decline, and a secondary slower regrowth (Fig. 1). After an initial slight decline, landings of Atlantic menhaden steadily increased from 167,200 t in 1942 through 1947. Nominal effort generally paralleled landings, but slightly lagged, from a low in 1943 to a minor peak in 1951 and rose to the record high of 712,100 t in 1956. Effort levels increased again in 1953 and reached a secondary peak in 1956 as well. Effort reached its highest level in 1959 with landings at their second highest level of 659,100 t. Landings dropped precipitously from 1962 to a record low 161,600 t in 1969, while effort dropped from 1964 and bottomed in 1971. Although fluctuating, landings showed a net in- crease from 1970 through 1981. Effort slowly in- creased up to 1977 and then began a declining trend. All of the five fishing areas showed a net in- crease in catches from the 1940's to the peak year in 1956 (Fig. 3). But, the increase was dispropor- tionately distributed between fishing areas. The middle Atlantic area showed the greatest relative increase, followed by the north Atlantic and Chesapeake Bay areas, with the south Atlantic and North Carolina fall fishery only showing slight increases (Fig. 4). While 1956 represented the year of peak landings for the fishery, only the middle Atlantic catches peaked during this year. The increase in fishing effort expended during the 1940's to the mid-1950's was also dispropor- tionately distributed between fishing areas, with the middle Atlantic showing the greatest in- crease, followed by the north Atlantic (Fig. 5). The Chesapeake Bay area showed a slight in- crease, while the south Atlantic and North Caro- lina fall fishery areas showed little, if any, actual increase in nominal effort. Following 1956, the proportion of the catch taken in the middle Atlantic area decreased rela- tive to the proportion of effort expended in that area. After 1962, effort (and catches) rapidly de- creased in the middle Atlantic area, with both the effort and catches seemingly shifted to the Chesapeake Bay area (Figs. 4, 5). This was a rel- atively dramatic shift, considering that the mid- dle Atlantic catch was generally dominated by age-2 and -3 fish while Chesapeake Bay catches were dominated by ages 1 and 2. Additionally, middle Atlantic fish were generally larger for any given age than those from Chesapeake Bay. Catches and effort also decreased in the north Atlantic area during the same time frame. The Atlantic menhaden stock probably had its strongest and broadest age structure in 1955 and 1956, which also represent the first years when port sampling covered the full geographic range 574 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY O X CO z o I- o oc I- UJ :e < 800 700 600- 500- 400 Z 300 X o 200- 100 kt'2 Fall Fishery South Atlantic r ' ^ Chesapeake Bay '-v-X^-:;:;:;,-:-,--?-!---;- 1940 44 48 52 56 60 64 68 72 76 80 J L Middle Atlantic North Atlantic YEAR OF FISHING Figure 3. — Contribution to landings of Atlantic menhaden by fishing area and season in thou- sands of metric tons for years 1940-81. The middle Atlantic and Chesapeake Bay area landings are combined after 1972 due to confidentiality restrictions. I O I- < o u. o UJ > I- < Zi o 100 90 80 70 60 50 40 30 20 10 0 - / ^ /X/ ..J ^■' \ Kf ■\/ "V- I I I Fall Fishery \/ V/\ South Atlantic Chesapeake Bay Middle Atlantic \-, ,-■ •■_/ ""^Z --• North Atlantic \' ^-^—^■- 1940 44 48 52 56 60 64 68 72 76 80 YEAR OF FISHING Figure 4. — Cumulative contribution by percent total landings of Atlantic menhaden by fishing area for years 1940-81. The middle Atlantic and Chesapeake Bay area values are combined after 1972 due to confidentiality restrictions. of the fishery (Fig. 6). The number of older fish in the population was somewhat reduced by 1959, but showed a slight recovery in 1960. The age structure became markedly constricted by 1965, and drastically truncated by 1967. The age struc- ture appeared to begin broadening slowly about 1972, and appears to be continuing this trend through 1981. Although a few individuals aged 9 and 10 years were taken by the fishery during the late 1950's, the maximum age class with adequate represen- tation for computational purposes is age 8. (Hence, the age category of 8+ is taken in this report to be age 8, which for most years contained only age-8 fish.) Individuals age 8-H were repre- sented in landings from 1955 through 1966 fish- ing seasons. From 1967 through 1969 the oldest fish were age 6, and likewise; 1970 through 1974, 1976, and 1978, age 5; and again, 1975 and 1977, and 1978 through 1981, age 6. 575 FISHERY BULLETIN: VOL. 85, NO. 3 O li. u. UJ O LU > < 3 O 100 -•■ ' '. 90 ...,,'\ / ■■■-■•.. •, Fall Fishery 80 70 60 ,\A. '■' --''■\ South Atlantic 50 >^/ ■ . ,. -•••■■-. /-■•■, -• •' ■. .•' • Chesapeake Bay 40 ..•■'' '"•■""■' 30 -■'■~-' 20 '■ -•' " '. ,'. 10 0 y^ ^^x ^_ ■.. .•■•.,-■ ^.-^.-.^ " Middle Atlantic ~'^-'' '^ " ' '""'\ .^-/ ""' North Atlantic ' r— 1 1 r— 1 1 ! 1 1 1 1 1 r--"'T r— 1 1 1 1 1 — ■ 1 1942 46 50 54 58 62 66 70 74 78 YEAR OF FISHING Figure 5. — Cumulative fishing effort on Atlantic menhaden as a percent of total, by fishing area for years 1941-81. (Data for 1941 through 1968 were adjusted (reduced) by Nicholson (1971a) to compensate for the small size of vessels that frequently fished in that area. The data for the middle Atlantic area were also adjusted, but to a lesser degree. Data from 1969 to 1981 are unadjusted. North Atlantic area data, adjusted by the Nicholson (1971a) criteria, would probably be less than half the amount shown, while the middle Atlantic area values may only be slightly reduced.) n o X « z o I- o cc I- LU 5 X o < o 800 700 600 500 400 300- 200 1001- ./■-.\ k 'V \/ / A aC^- — Age 5 + ir~:^~^' Age 4 ^ ' ... Age 3 -Age 2 /-vA./' \ ' ' .' ' '. I ' I ;•'■;•-'■■' ■^--•i'-v---i--i-i -■'—■■-■'' ■<-■ 1940 44 48 52 56 80 64 68 72 76 80 -Age 1 -Age 0 YEAR OF FISHING Figure 6. — Catch of Atlantic menhaden in thousands of metric tons by age group for 1955 to 1981. SIZE AT AGE AND GROWTH ANALYSIS It is necessary to derive two types of size at age and growth estimates because of the seasonal dif- ferential distribution of Atlantic menhaden by age and size (Nicholson 1972, 1978). Parameters are estimated by area to characterize the seg- ments of the population normally harvested within that area and are estimated for the entire fishery to characterize the harvested population. Because relative harvest rates may vary among areas between seasons, apparent area-specific growth parameters are estimated for yield-per- 576 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY recruit analyses. On the other hand, parameters generated in each area are inappropriate (biased) for describing the entire population (or fishery). For example, growth rates and average size at age values for the north Atlantic area will be greater than those for the population, and simi- larly values estimated for the south Atlantic will be less than true population values. (The only exception is the North Carolina fall fishery, which apparently harvests a reasonably well- mixed migratory population.) Size at age and growth estimates for the entire stock are needed for yield-per-recruit analysis for the entire fishery and to ascribe an average size at age for the spawning stock for each year. These estimates are obtained by appropriately weight- ing the sampling results from each fishing area, as will be shown. Additionally, since growth of Atlantic men- haden has been shown to be inversely related to year class size (density-dependent) (ASMFC fn. 4), a condition predicated during the estuarine portion of the life cycle (Reish et al. 1985), growth equations must be computed for each year class. The fishing season is divided into quarterly incre- ments for these analyses (Table 2). The analytical steps taken to obtain these size estimates and how they are used follows. Table 2. — Quarterly time increments used in stock assessment analysis of Atlantic men- haden. Beginning week Ending week Quarter ending date ending date 1 2 3 4 s 3/01 > 5/31 > 8/30 >11/29 = 5/30 i 8/29 s 11/28 = 114/29 'February of next calendar year, but same season. Area-Specific Mean Size at Age and Growth Rates Mean lengths at age by area by quarter for each of the 1965-81 seasons were estimated directly from the port sampling data as unweighted arith- metic means. These results were in turn arranged by specific year class and fitted to the von Berta- lanffy growth equation using the computer pack- age BGC3 (Abramson 1971). It was assumed that each mean length estimate was representative of the middle of the quarterly interval, i.e., for the first quarter (age X.0-X.25) the mean value is assigned to age X.125, etc. These fitted area- specific von Bertalanffy parameters were used to derive estimates of length at age for the begin- ning of each time interval. These estimates were in turn converted to estimates of weight at age for the area-specific yield-per-recruit analysis. Mean Size at Age and Growth for the Entire Fishery Predictive equations for growth which are rep- resentative of the population as a whole (entire fishery) are needed to estimate size at age for the spawning stock and for yield-per-recruit analysis of the entire fishery. For years 1965 to 1981, mean lengths at age by quarter for the entire fishery were obtained by weighting each area's estimate of mean length by its corresponding catch in numbers at age by quarter. Age 0.875 (fourth quarter age 0) was the youngest age for which mean length was calculated. These values were arranged by year class and were fitted to the von Bertalanffy growth equation. Estimates of weighted mean length by quarter could not be calculated for the fish caught before 1965 because estimates of numbers at age landed by quarter by area were not available even though size at age is available weekly. Since much of the fourth quarter catch of the North Carolina fall fishery is composed of migratory stocks, it was presumed that a representative es- timate of length at age for the entire population might be obtained from the fourth quarter values from this area alone. To test this hypothesis, mean lengths for the 1965 to 1978 year classes from the fourth quarter in the North Carolina fall fishery were fitted to the von Bertalanffy growth equation. The resultant curves were compared vi- sually with results when all weighted mean length values were used. The results were quite similar when five or more data points were avail- able and dissimilar to relative degrees when <5 data points were available. Because all year classes from 1955 to 1964 had at least 5 data points which met the above criteria, von Bertalanffy curves were fitted to these values (Table 3). Weight-Length Relationship The predictive growth equation used in this re- port uses length at age. Weight-length relation- ships were derived to estimate weight at age val- ues. The greatest potential within year variation in weight-length parameters is expected among 577 FISHERY BULLETIN: VOL. 85, NO. 3 Table 3. — Estimated von Bertalanffy growth parameters for Atlantic menhaden, year classes 1955-78. Year n Age range class Lx K '0 (means) (years) 19551 339.49 0.5401 0.1234 7 0-7 1956 343.67 0.4598 0.0245 7 0-8 1957 324.49 0.6260 0.0707 7 0-7 1958 363.73 0.3637 -0.1163 7 0-7 1959 355.64 0.3631 -0.4709 6 0-6 1960 354.69 0.4009 -0.1481 5 0-5 1961 340.52 0.4514 -0.3506 5 0-5 1962 376.35 0.4012 -0.1122 5 0-5 1963 370.04 0.3494 -0.4652 6 0-6 1964 331.17 0.6138 0.0606 5 0-5 19652 404.13 0.3187 -0.3529 17 0-6 1966 367.15 0.4575 -0.0645 18 0-6 1967 375.81 0.4539 0.1815 16 0-5 1968 415.22 0.2813 -0.7318 16 0-5 1969 356.19 0.5868 0.0530 18 0-6 1970 348.00 0.5351 0.0034 16 0-5 1971 356.82 0.4103 -0.2772 18 0-6 1972 316.74 0.6058 0.0767 17 0-5 1973 341.53 0.3884 -0.2947 20 0-5 1974 325.60 0.4329 -0.1280 22 0-6 1975 420.86 0.1779 -1.1445 21 0-6 1976 393.73 0.2410 -0.4372 20 0-5 1977 528.99 0.1455 -0.8354 16 0-4 1978 246.54 0.5807 -0.3399 12 0-3 'Year classes 1955-64 represented by fitted values for 4th quarter, area 5 (see text). 2Year classes 1965-78 represented by fitted values for weighted quar- terly mean lengths VIRTUAL POPULATION ANALYSIS Virtual population analyses (VPA's) were con- ducted to reconstruct population sizes and esti- mate rates of fishing mortality. Analyses were conducted on all age groups of the 1947-78 year classes which were represented in the 1955-81 landings. The backward sequential computations were performed using the computer program MURPHY written by Tomlinson (1970). Instantaneous Rate of Natural Mortality The estimate of the annual instantaneous rate of natural mortality (M) used in this report was 0.45. Early estimates were from catch statistics, 0.37 (Schaaf and Huntsman 1972); from prelimi- nary tag-recovery analysis, 0.52 (Dryfoos et al. 1973); and from a more extensive tag-recovery analysis, 0.50 (mean of age-specific rates for ages 2 and 3) (Reish et al. 1985). The 0.45 value represents a mean of the range of available esti- mates. The implications of the selection of 0.45 for M are addressed. quarters. Because each area (except perhaps area 5) contains a limited portion of the range of sizes extant in the menhaden population, the weight-length relationship is estimated across areas, but within quarters. Annual variation was also assumed to exist, thus parameter estimates were calculated for each fishing season where subsequent yield-per-recruit analysis was in- tended (1970-78) using logg transformed data and least squares regression (Table 4). Annual Mean Weights Weighted by Catch Estimates of annual weighted mean weight by age of Atlantic menhaden in purse seine catches were calculated to permit computations of age- specific and year class-specific biomass contribu- tions to landings. Weighted mean weight for the entire fishery was calculated from the average weight by age by season for each of the five recog- nized areas of the fishery and then weighted by the estimated numbers caught by age in each re- spective area. These data were derived directly from the port sampling data and are not from von Bertalanffy derived lengths converted to weights. Temporal Organization of Analyses The time periods used in these analyses (Table 2) closely correspond to critical life history and fishery events. The birth date for a year class, and the beginning of a new fishing season for Atlantic menhaden, is 1 March. Because of the protracted spawning season of menhaden, young of the year may have been spawned as early as the previous August and as late as the following May or June, but most of the spawning takes place in the fall and winter (Nelson et al. 1977). The beginning of the fourth quarter (week begin- ning > Nov. 29) is used as a finite date to estimate spawning stock size. Thus, the spawning stock in the fall (beginning of the fourth quarter) of the previous calendar year (age X.75) is defined to be the parental stock for a subsequent (1 March) year class. Recruitment is examined at age 0.5 (beginning of third quarter, age 0) and at age 1.0 (beginning of first quarter, age 1). Three sets of VPA's were conducted with the length of the time intervals varied between sets. The first set provided the basic estimates for re- construction of the historical population and the estimates of rates of fishing mortality. This series was done on an annual basis and involved all subject year classes. These estimates are used pri- 578 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY Table 4. — Weight-length regression parameters for Atlantic menhaden, by quarter and year, 1970-81 seasons (\n W = a + b\r\ L). Fishing years Quarter 1 a b Quarter 2 Quarter 3 Quarter 4 Annual a b a b a b a b 1970 -11.9324 3.1924 -11.5760 3.1224 - 1 1 .3909 3.0971 -11.5124 3.1087 -11.6666 3.1421 1971 -10.8692 2.9838 -11.0121 3.0135 -11.4336 3.0941 -12.0821 3.2044 -11.3620 3.0786 1972 -13.0384 3.3809 -11.5388 3.1072 -11.5989 3.1131 -11.5413 3.1028 -11.6553 3.1264 1973 — — -10.6360 2.9401 -10.7756 2.9741 -11.4291 3.0889 -10.9727 3.0060 1974 -12.8892 3.3503 -11.3321 3.0695 -11.2386 3.0552 -12.1803 3.2310 - 1 1 .4423 3.0896 1975 -10.3727 2.9027 -11.9798 3.1950 -11,7856 3.1555 -11.6995 3.1388 -11.8524 3.1703 1976 -10.9908 2.9972 -12.5123 3.2945 -12.4698 3.2933 -11.9663 3.1873 -12.3503 3.2642 1977 -12.6133 3.3110 -12.9689 3.3856 - 1 1 .9884 3.2043 -12.0643 3.2109 -12.5865 3.3137 1978 -12.3304 3.2666 -12.5737 3.3096 -12.0319 3.2124 -11.8477 3.1642 -12.3324 3.2653 1979 -11.9230 3.1950 -11.8170 3.1701 -12.0982 3.2243 -12.5516 3.2900 -12.4043 3.2793 1980 -12.2804 3.2580 -12.8763 3.3696 -11.9109 3.1884 -11.7750 3.1547 -12.3969 3.2792 1981 -12.2049 3.2386 -12.4031 3.2765 -12.6597 3.3215 - 1 1 .9663 3.1747 -12.5365 3.3004 marily in discussions of the impact of nominal effort and for modification of input variables for surplus production analysis. The second series, with shorter time intervals, permitted a more precise apportioning of fishing mortality between fishing areas within a fishing season and provided within season estimates of numbers at age present in the population. This series was conducted on a quarterly basis and included the 1965-78 year classes. These esti- mates permitted a reconstruction of the fishery for the 1970-78 seasons which forms the basis of the subsequent yield-per-recruit analyses. The quarterly estimates of numbers at age are also used to estimate numbers of recruits for the 1965-78 year classes and the numbers of spawn- ers that were ultimately derived from these year classes. The third series was conducted to estimate numbers of recruits and their parental spawning stock for the 1955-64 fishing seasons. This series included the 1947-64 year classes and used mixed length time intervals. A V2-yr interval, which in- cluded ages 0.50 and 0.75 (quarters 3 and 4), was used to provide estimates of numbers of age-0.5 individuals present in the population. The time interval for age-1 fish was annual. Intervals for age-2 and older individuals were alternately three quarters of a year (quarters 1-3), corre- sponding to ages X.0-X.50, and one quarter of a year (quarter 4), corresponding to age X.75. This temporal construction of the numbers at age data required several adjustments which are dis- cussed. Additional VPA's were conducted to examine the sensitivity of results to the value of natural mortality used (0.45) and to the initial estimates of fishing mortality rates. A series of annual VPA's were conducted for the 1955-78 year classes with M = 0.35 and M = 0.55, which en- compass the range of available estimates dis- cussed earlier. Additional annual runs were con- ducted with varied starting F's for the 1955 year class, which for reasons discussed later, poten- tially represents a worst case situation relative to rates of convergence of estimates. VPA Numbers at Age Landed Data Sets The annual estimates of numbers at age caught were rearranged from a seasonal format to a year- class format. For the quarterly runs, the weekly catch at age estimates were summed to quarters and rearranged by year class. The mixed time interval VPA data sets were derived from the an- nual set, and thus required some approximations and adjustments to obtain a subyear format. An- nual catches of age 0 (1955-64 seasons) were as- sumed to be in quarters 3 and 4 since the bulk of the age-0 catches occurred after 30 August (be- ginning of the third quarter). The major portion, if not all, of the catches of age-2 + fish (spawning ages) that were made during the fourth quarter time interval, was in the North Carolina fall fish- ery. During the 1965-69 fishing seasons, an aver- age of 58% of the age-2 + fish that were landed in the fall fishery were landed during the fourth quarter. Hence, 58% of the fall fishery landings of age-2 + fish of the 1947-64 year classes were as- sumed to have been taken during the fourth quar- ter. The remainder of the total annual catch was assigned to the single three-quarter time interval (quarters 1-3). The time period used for the age-1 fish remained annual, so no adjustments were required. 579 FISHERY BULLETIN: VOL. 85, NO. 3 Estimates of Initial Annual Rate of Instantaneous Fishing Mortality Although a single method of estimation for the starting instantaneous annual fishing mortality rate (F ) for backward calculations in VPA is de- sirable, a few year classes required alternate ap- proaches. Year class-specific catch curves were examined visually. All years of age from the old- est to youngest that lie within a reasonably straight portion of a semilogarithmic catch plot were logg transformed and regressed against age. The slope was taken as an estimate of -Z (total instantaneous mortality) and by subtracting M a "general trend" F was obtained. Starting F's were obtained by this method for all year classes except 1947, 1948, 1949, 1954, 1966, 1971, 1974, and 1978. Since only 2 years of landings after full recruitment were available for the 1978 year class, an estimate of Z was made: logg (catch in numbers of age 3/catch in numbers of age 2). The estimate for 1949 was obtained similarly using ages 7 and 8. The 1954, 1966, 1971, and 1974 year classes experienced an apparently higher fishing rate during their last year in the fishery com- pared with that experienced 1 year earlier. Thus, starting F's were obtained from average VPA results from several other age classes caught in the same year. Starting F for age 8 of the 1954 year class was estimated as the mean F for ages 5-7 (year classes 1955-57) caught in the same 1962 season. Initial F for age 5 of the 1966 year class was the annual VPA estimate of age 4 from the 1967 year class. Similarly, starting values of F for the 1971 and 1974 year classes were the means of the same fishing season age-4 and age-5 values for the 1972 and 1973, and 1975 and 1976 year classes. Starting F's on age 8 for the 1947 (one age class represented) and 1948 (two age classes represented) year classes were means of similar fishing season VPA F's for ages 5-7. The initial annual F values and their sources for VPA are summarized in Table 5. Conducting the VPA computations for the an- nual series was straight forward, as the trial F's were annual. The quarterly series and mixed time interval series required trial and error runs. Trial starting F's for these VPA's were adjusted downward until the sum of the F's within the last year (oldest fish of the year classes) were ±0.5% of the initial annual F estimate for each year class (Table 5). Except for the sensitivity computations, the general results of the VPA's are presented and discussed where they are subsequently used. Using a relatively wide range of starting values for the 1955 year class, estimates of age-specific F and numbers at age present at the beginning of a season (Fig. 7) converged quite rapidly to similar values for the (younger) ages which dominate the fishery in both numbers and biomass (Figs. 8, 9). This is expected because of the relatively high rates of fishing mortality exerted on the stock (Ulltang 1977). The lower the exploitation rate, the slower the values will converge. The 1955 year class had the lowest starting F of any year class (Table 5) (other than 1947 which had only one age class represented) and relatively low to moderate rates of exploitation on all age classes. Thus the estimates for the other year classes should converge more rapidly than the one shown. Since it is highly unlikely that the initial estimates of F differ from true values to the ex- treme degrees tested, the VPA estimates used Table 5. — Estimates of initial Ffor annual virtual population analy- ses (VPA's) of Atlantic menfiaden, source of estimate, and ages involved, by year class. Year Initial VPA's Regression Mean from class F ages ages VPA results! 1947 0.4862 8 yes 1948 1.3134 7-8 — yes 1949 1.34981 6-8 — — 1950 1 .6504 5-8 6-8 — 1951 0.9243 4-8 4-8 — 1952 0.8590 3-8 5-8 — 1953 0.8620 2-8 2-8 — 1954 1.1645 1-8 — yes 1955 0.7006 0-8 2-8 — 1956 08557 0-8 2-8 — 1957 1.0651 0-8 5-8 — 1958 1.8497 0-8 4-8 — 1959 1.4325 0-7 3-7 — 1960 1 .7620 0-6 2-6 — 1961 30108 0-6 4-6 — 1962 1.9074 0-6 2-6 — 1963 20482 0-6 2-6 — 1964 2.1116 0-5 2-5 — 1965 2.7386 0-5 3-5 — 1966 1.6194 0-5 — yes 1967 1.1191 0-5 2-5 — 1968 1.6677 0-5 2-5 — 1969 1.9585 0-6 3-6 — 1970 22554 0-5 2-5 — 1971 1.5437 0-6 — yes 1972 1,7143 0-5 2-5 — 1973 1 .5403 0-5 2-5 — 1974 1.4321 0-6 — yes 1975 1 3185 0-6 2-6 — 1976 1.0167 0-5 2-5 — 1977 1 .3079 0-4 2-4 — 1978 1.4391 0-3 — — 'See text 580 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY O z X CO u. z z o < > UJ o 1- z UJ o a: UJ Q. cr o -40- 0.25 F 100 CO 80 Ui 00 2 3 60 z Z 40 z o I h- < u. 20 ■> U- UJ O Q 0 t- z UJ o -20 a. Ui Q. -40 -60 0.25 F 0 50 F 0.67 F -J 8 AGE IN YEARS Figure 7. — Deviations in annual VPA (virtual population analysis) estimates of numbers at age present in the population and age specific fishing mortality rates for Atlantic menhaden resulting by varying the initial rate of annual F by the multiples shown. The estimates are for the 1955 year class (initial F = 0.7006). here should be considered reasonably precise (and stable) relative to the initial F values. Since F and M are additive with respect to Z, errors in these computations resulting from an incorrect selection of the rate of M are additive with respect to subsequent estimates of F, and estimates of numbers at age present in the popu- lation will differ in a proportional fashion. In other words, if our selection of the estimate of M is too great, numbers at age are overestimated, and if M is too low, the reverse is true (Fig. 10). Similarly, estimates of year class size vary by a nearly constant proportion (Fig. 11). The range of available estimates of M is relatively narrow, hence it is unlikely that conclusions reached in this report would be altered even if the estimate of M for these analyses were allowed to vary ran- domly within these bounds between years. 581 FISHERY BULLETIN: VOL. 85, NO. 3 100 90 CO tr 80 lU m S 70 3 Z 1 60 I 50 O < 40 O li. 30 o # 20 10 0 y 5*'s 4's 3's 02's 1's O's 56 58 60 62 64 66 68 70 72 74 76 78 80 82 YEAR Figure 8. — Contribution in percent of total numbers of Atlantic menhaden landed by age group, from 1955 to 1981. y s 5+'s 4's 3's 2's 1's O's Figure 9. — Contribution in percent of total biomass of Atlantic menhaden landed by age group, 1981. from 1955 to SPAWNER-RECRUITMENT RELATIONSHIP Estimates of Numbers of Spawners Higham and Nicholson (1964) concluded that a few age-1 Atlantic menhaden and most age-2 fish are sexually mature by the end of the season. The simplifying assumption for purposes of estimat- ing spawning stock in these analyses is that no age-1 fish and all age-2 fish are mature by the beginning of the fourth quarter. This assumption was also used in other studies (Nelson et al. 1977; ASMFC fn. 4; Schaaf and Huntsman 1972). Estimates of the number of fish age 2.75 and greater alive at the beginning of the fourth quar- ter of any given year n comprised the parental spawning stock for year class n + 1. The esti- mates of number of spawners resulting from the 1947-64 year classes were obtained from the mixed time-interval VPA series. The number of spawners from the 1965-78 year classes were ob- 582 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY 9r (r m 2 01 O o AGE IN YEARS Figure 10. — VPA (virtual population analysis) estimates of numbers at age present in the Atlantic menhaden population for the 1955 year class for three levels of natural mortality (lower left), and natural logarithms of these same numbers at age (upper right). X CO tr o LU tr a. ui m 3 Z 18 16 14 12 10 , Mr 0.55 U-- 0.45 M= 0.35 J — I — I — I — I — I — I — I — I — I r I 'y^i. I I i I I III' 1956 58 60 62 64 66 68 70 72 74 76^ 78 YEAR CLASS Figure 11. — Annual VPA (virtual population analysis) estimates of numbers of Atlantic menhaden recruits at age 1 for three levels of natural mortality. tained from the quarterly VPA series. These esti- mates were rearranged to correspond to fishing seasons (Table 6). There were no estimates for numbers at age landed in the 1954 North Caro- lina fall fishery. Hence, the parental spawning stock for the 1955 year class was obtained by using 1955 estimates for age X.O and back- calculating numbers at age to the beginning of the 1954 fourth quarter (age (X - 1).75) with mean fourth quarter survival rates at age for fish- ing years 1955-57. An additional adjustment was made to complete the estimates of spawning stock size. If an age group was represented during quar- ters 1-3 but not in quarter 4 of the last year it 583 FISHERY BULLETIN: VOL. 85, NO. 3 Table 6. — Estimates of spawning stock size in thousands by age for Atlantic menhaden, 1954-81 seasons, at start of the fourth quarter of each fishing season. Age Year 2.75 3.75 4.75 5.75 6,75 7.75 8.75 + 19541 720,068.8 966.532.5 158,421.1 44,416.7 10,592.6 2,383.0 NE 1955 750,253.7 216,441.3 325,276.2 50,554.4 12,061.1 3,463.6 932.1 1956 285,976.8 213,151.0 102,272.1 98,950.9 11,254.9 2,541.6 593.9 1957 321,399.1 99,134.2 72,526.3 35,097.5 19,882.5 1,250.7 3620 1958 1,060,499.7 145,248.4 46.161.4 29,036.6 11,021.2 5,504.1 82.2 1959 330,845.0 366,056.3 65.942.6 19,506.9 7,873.1 2.094.9 1,339.2 1960 2,662.230.0 144,140.8 125.048.6 20,387.2 5,552.2 1.337.9 693.7 1961 432,466.5 736,311.5 69.834.8 46,372.7 6,667.1 1.300.3 144.1 1962 215,894.3 84,439.3 97.991.5 18.987.4 6,579.9 1,795.1 258 7 1963 173,757.5 43,433.2 17.968.5 15.709.5 2,756.5 991.8 539 0 1964 151,149.8 26,764.3 4.358.7 2.199.5 1,112.9 157.6 194.2 1965 101,340.2 12,848.1 1.419.3 164.1 180.0 53.7 13.2 1966 194,222.2 18,780.6 1,401.8 15.5 18.0 27.4 5.4 1967 133.362.5 36,015.8 2.858.6 207.1 0.5 0 0 1968 122,033.7 16,235.3 1.227.0 199.0 8.1 0 0 1969 125,131.1 26,492.8 761.0 16.4 1.1 0 0 1970 175,837.2 34,303.4 6,017.2 48 1 0 0 0 1971 265.058.8 29,169.5 4.107.8 619.4 0 0 0 1972 64,386.9 14,863,7 1.187.3 766.8 0 0 0 1973 80,116.3 5,210.6 2,021.8 142.9 0 0 0 1974 94.3483 7,760.3 241.5 153.0 0 0 0 1975 140,817.4 12,947.3 356.3 12.9 13.8 0 0 1976 212,735.5 37,466.2 2,306.5 164.1 0 0 0 1977 498,1358 59,415,2 5,630.1 2456 22.3 0 0 1978 451,889.2 80.9629 13,390.0 927.6 0 0 0 1979 486,903.4 155.932.5 26,789.6 2.772.6 47.3 0 0 1980 424.606.5 103,707.0 42,254.5 4.995.9 909.8 0 0 1981 -No. est- 57,691.0 14,868.6 7.335.7 397.5 0 0 'Derived from 1 March 1955 estimates (see text). appeared in the fishery, it was assumed that some representatives were still present in the popula- tion at the beginning of quarter 4 (age X.75). Es- timates of numbers present were obtained by a forward calculation using the mortality estimates of the previous interval obtained from the VPA's. not available, means from the nearest three year classes were used. Estimates of egg production were obtained from the expression used by Nelson et al. (1977), which was derived from data of Higham and Nicholson (1964): Estimates of Potential Egg Production The age structure of the Atlantic menhaden population varied substantially during the time period under study. Therefore, an alternate method of examining spawning stock (i.e., poten- tial egg production) as used by Nelson et al. (1977) was employed. Sizes at age X.75 for spawners derived from the 1955-78 year classes were calculated from the von Bertalanffy growth parameters derived earlier for the entire fishery (Table 3). However, because of insufficient data, growth curves were not fitted to the 1947-54 year classes, so observed mean lengths at age during the fourth quarter were used for the spawners derived from these year classes. If the fourth quarter length at age was ln(£;) = 0.3149 + 0.0176L where, E = thousands of eggs produced per fe- male, and L = estimated fork length. This equation was used with estimated mean length at age X.75 for fish 2 years and older. As- suming a 50/50 sex ratio, potential egg production by age by season was obtained by multiplying the values per female by number of females at age (Table 7). Estimates of Numbers of Recruits Estimates of the number of recruits for each year class were obtained from the results of the VPA's. The year class-size estimates for 1955-64 are from the mixed time-interval analyses and 584 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY Table 7. — Estimated number of recruits by year class at age 0.5 and 1.0, estimated number of spawners thiat produced the year class, and estimated egg production from the spawning stock, for Atlantic menhaden. Number of recruits No. of No. of Year X lU-' spawners eqqs class Age 0.5 Age 1 .0 X 103 j3 X 1012 1955 7,888,342 5,621,258 1,902,414.7 219.659 1956 8,999,656 7,153,549 1,358,982.4 147.047 1957 4,419,989 3,263.196 714,741.2 83.977 1958 18,612,316 14,767,294 549,652.3 57.768 1959 2.722,999 2,164,428 1,297,553.6 143.822 1960 3,786,692 2,958,923 793,658.0 76.642 1961 2,769,147 2,210,534 2,959,390.4 156.058 1962 2,841,268 2.222,880 1,293,097.0 106.781 1963 2.304,564 1,754,140 425,946.2 37.508 1964 2.764,796 1.938,001 255,156.0 21.466 1965 2.072,852 1.430,539 185,937.0 13.806 1966 2,879,544 2,001,871 116,018.6 7.552 1967 1,522,438 1,209,954 214,470.9 17.017 1968 2,319,215 1,710,666 172,444.5 13.053 1969 3,448,326 2.611.940 139,703.1 1 1 .240 1970 1,755,217 1,382,032 152,402.4 12.056 1971 4,513,962 3,539,073 216,205.9 17.594 1972 3,516,016 2,760,443 298,955.5 31.279 1973 3,908,494 3,085,954 81,204.7 8.044 1974 5,197,484 3,866,593 87,491.6 6.076 1975 9,024,340 6,932,136 102,503.1 6.591 1976 6.953.329 5,297,439 156,147.7 7.575 1977 6,619.024 4,827,413 252,672.3 1 1 .966 1978 6,040,678 4,404,267 563,449.0 18.864 19791 10,322,177 6,890,589 547,169.7 18.389 1980 NE NE 672,445.4 26.045 1981 NE NE 576,473.7 22.294 1 Preliminary estimates. those for 1965-78 are from the quarterly analyses. Estimates of recruitment were computed for both age 1.0 and age 0.5 (Table 7). Estimates at age 1.0 are provided for comparative purposes, as this age has been frequently used for studies on Atlantic menhaden (ASMFC fn. 4). Age 0.5 is used here to appropriately credit a year class with the num- bers of juvenile fish removed from the population by the fishery during the fall and early winter. Although perhaps underestimates because the value of M (0.45) may be too low for fish younger than 1-yr old, these estimates are relatively con- sistent. The degree of dependency of the number of re- cruits on the size of the parental stock has been examined by Schaaf and Huntsman (1972), Nel- son et al. (19771, Schaaf (1979), and Reish et al. (1985). All earlier workers employed the Ricker ( 1954) model, but Reish et al. (1985) also used the Beverton and Holt model as well as the unnor- malized gamma function. The published results indicate weak relationships, with substantial variability about both fitted models. Nelson et al. (1977) developed a multiple regression model to explain observed deviations from the Ricker model attributable to several annually varying environmental parameters, primarily Ekman transport, which would affect the oceanic larval stage. The spawner-recruitment data (Table 7) were fitted with both the Ricker and Beverton-Holt models using a nonlinear least squares method (Marquardt's (1963) algorithm). Both models fit the data poorly (Fig. 12). The Beverton-Holt model is slightly better than the Ricker model if residual sum of squares is used as a goodness of fit criterion. The Beverton-Holt residual is only slightly less than that about a mean value, which assumes no relationship between numbers of spawners and numbers of recruits. Residual sum of squares in the Ricker model was slightly greater than results for the mean. POTENTIAL AND ACTUAL YIELD Production Models The application of production models to the At- lantic menhaden purse seine fishery is hampered on theoretical grounds by two major conditions: 1) the fishery has not been operating under equi- librium conditions, and 2) fishing effort is not proportional to fishing mortality (F). The catcha- bility coefficient {q ) is inversely related to popula- tion size (Schaaf 1975b). Schaaf and Huntsman (1972) and Schaaf (1979) circumvented this latter problem by adjusting effort to a base year. The effects of this problem were reduced in this analysis by using an estimate of population F for the independent variable instead of adjusting ef- fort (Nelson and Ahrenholz 1986). To estimate a population rate of fishing mortality (Fpop), esti- mates of the population sizes (excluding 0-age fish) at the beginning of each fishing season from 1955 to 1979 were reconstructed from annual VPA estimates. These were in turn divided into the estimated catch in numbers (excluding age 0), to obtain an estimate of population rate of ex- ploitation (C/pop). By trial and error, estimates of Fpop were obtained for each fishing season from Food =U Z/(1 P"P pop e-^), assuming M = 0.45 - F^K" " po ^ (Table 8). A Graham-Schaefer curve was fitted to the catch and population fishing mortality data by Marquardt's (1963) algorithm. This procedure produced an MSY (maximum sustainable yield) estimate of 414,000 t at Fpop = 0.574. Recent pop- ulation fishing mortality values have been slightly above and below this value, and yield has 585 FISHERY BULLETIN: VOL 85, NO. 3 If) d lij < I z o _l _i o I UJ t a. o u 20 40 60 80 100 120 140 160 180 200 220 240 NUMBER OF EGGS IN TRILLIONS Figure 12. — Numbers of Atlantic menhaden recruits (R) in millions plotted against esti- mated egg production IP) in trillions for year classes 1955-1978. Curve a represents the fitted Beverton and Holt function, R = 1/(0.00019 + 0.00027/P). Curve b represents the fitted Ricker function, R = 325..35 P exp (-0.0158 P). Table 8. — Estimates of Atlantic menhaden population size and catch in numbers in thousands (age 1 to maximum observed age), population exploitation rates, population F. and catch in thousands of metric tons, by year. Population Popu- Population Catch in exploit. lation Catch Year size numbers rate {U) F (t) 1955 6,955,987.3 2,357,430.0 0.3389 0.532 641.4 1956 8,305,282.6 3,528,450.0 0.4248 0.721 712.1 1957 9,829,843.1 3,212,0900 0.3268 0.508 602.8 1958 7,125,101.7 2,613,150.0 0.3668 0.590 5100 1959 17,630,370.1 5,342,240.0 0.3030 0.462 659.1 1960 9,309,904.0 2,702,940.0 0.2903 0.438 529.8 1961 6,843,711.6 2,598,060.0 0.3796 0.618 575.9 1962 4,587,482.6 2,048,280.0 0.4465 0.774 537.7 1963 3,602,666.4 1,667,620.0 0.4629 0.816 3469 1964 2,770,398.2 1,426,470.0 05149 0.961 269.2 1965 2,593,453.3 1,260,362.0 0.4860 0.878 273.4 1966 2,127,951.0 991,157.0 0.4658 0.824 219.6 1967 2,592,784.2 977,2550 0.3769 0.612 193.5 1968 2,098,231.9 993,717.0 0.4736 0.845 234.8 1969 2,281,471.7 710,048.0 0.3112 0.478 161.6 1970 3,507,601.0 1,379,039.0 0.3932 0.648 259.4 1971 2,555,012.9 896,250.0 0.3508 0.556 250.3 1972 4,466,635.5 1,664,286.0 0.3726 0.602 365.9 1973 4,294,409.0 1,780,916.0 0.4147 0.697 346.9 1974 4,433,007.7 1,671,675.0 0.3771 0.612 292.2 1975 5,401,009.0 1,857,415.0 0.3439 0.542 250.2 1976 8,927,821.9 3,005,106.0 0.3366 0.527 340.5 1977 8,652,084.7 3,181,191.0 0.3677 0.592 341 2 1978 7,855,446.4 2,622,620.0 0.3339 0.522 344.1 1979 7,370,472.2 2,353,757.0 0.3193 0.493 375.7 remained slightly below the MSY value (Fig. 13). Catch and F^^^ values were also fitted to a mod- ified version of the Pella-Tomlinson (1969) model, which compensates for nonequilibrium conditions (PRODFITXFox 1975), assuming two significant year classes. This technique resulted in an MSY estimate of 557,000 t at Fpop = 0.336 (Fig. 13). The scatter plot of yield/Fpn^ contains three pop 586 AHRENHOLZ ET AL.: ATLANTIC MKNUADEN POPULATION AND FISHERY clusters of yield values: high abundance gears (1955-62), low abundance (high Fp^p) years of 1963-68 (with the exception of 1967), and years of low to moderate population size (1969-79). These clusters are less distinct when yield is measured in numbers of fish rather than biomass. A Graham-Schaefer curve, fitted to fishing mortal- ity and yield in numbers of fish, resulted in a numbers MSY of 2.383 billion fish at an optimum Fpop = 0.522 (Fig. 14). Each year's yield in num- bers as compared to biomass (Fig. 13, 14) suggests that the fishery in the late 1970's was not produc- 0) z 70C o H o 600 a. y- UJ S 500 u. O CO - , Q 400 z. < tn D 300 o I 1- z 200 I o < 100 o i ^ TY~^ t p 1 :8 — % — i^ — 1.1 ^ 1.2 POPULATION F Figure 13. — Catch of Atlantic menhaden in thousands of metric tons plotted against estimates of population F for years 19-5.5-79. Curve a is the result of fitting the Pella- Tomlinson's ( 1969) generalized yield function with adjustments for nonequilibrium con- ditions iPRODFIT, Fox 1 1975)1. Curve b is a nonlinear least squares fit of the parabolic (Graham-Schaefer) production model. X CO u. O CO z o m z I o < o POPULATION F Figure 14.— Catch of Atlantic menhaden in billions offish against estimates of population F for years 1955-79. The curve is the result of a nonlinear least squares fit of the parabolic (Graham-Schaefer) production model. 587 FISHERY BULLETIN: VOL 85, NO. 3 ing the biomass in landings in proportion to num- bers of fish caught when compared to earher years. Catchability coefficients ((/p^p) for the popula- tion were estimated directly by dividing the Fpop estimates by nominal effort (vessel weeks) for years 1955-79. A plot of these estimates on popu- lation size indicates a pronounced inverse rela- tionship similar to that shown by Schaaf (1975b), who estimated q differently (Fig. 15). Addition- ally, there is a pronounced historical trend in the data. There appear to be at least two families of points and thus two functional curves in the fig- ure, 1955-69 and 1970-79. Beginning in 1959, the catchability coefficient progressively increased, the stock size was decreasing, and the fleet was becoming more efficient due to modernization and increased vessel size, coupled with technological innovations in the fishery operations themselves. This trend in efficiency probably made the ascent from the earlier, lower ^pop series of years excep- tionally rapid. As population size began to in- crease after 1971, the catchability coefficient re- flected a steady decline in magnitude, but it was at a level almost twice as great (hence a doubling in killing power or efficiency") as from the late 1950's and early 1960's. Therefore, the reduction ■^Given modern day work weeks, real time .spent fishing, and intervessel competition, the killing power has probably more than doubled. in the number of vessels from 1955 to the present did not represent a proportional reduction in po- tential effective fishing effort. In spite of the com- pounding effect of a true increase in efficiency (fishing technology), q, and thus F for given lev- els of effort, appear to be responding in inverse fashion to population size. The computed F^^p values are derived inde- pendently of nominal effort. Hence this set of val- ues was used to determine if effort would be use- ful to verify trends in F at age or provide supportive information for improving the VPA estimates by employing more sophisticated models (see Deriso et al. 1985). A scatter diagram of the differences between estimates of Fp^p and their mean on nominal effort demonstrated that no useful information is available from nominal effort when considering the entire time span in- volved (Fig. 16). The progi-essive increase in effi- ciency of the purse seine fleet, plus problems asso- ciated with estimating abundance with CPUE data from purse seine fisheries for schooling fishes (see Clark and Mangel 1979) makes nomi- nal effort useful only for relatively short-time span comparisons between adjacent years, and then only when changes are pronounced. Yield Per Recruit Yield-per-recruit calculations for the 1970-78 fishing seasons were performed using the com- o X I- z UJ o u. u. tu o o m < I o I- < o 0.8 07 0.6 0.5 0.4 0 3- 0.2 0.1 .73 72 ■84 .65 •77 .78 76 I I L. 1 2 3 5 6 7 8 9 10 1 1 ^"^8 POPULATION SIZE (ages 1 to 8* ) IN BILLIONS Fl<;URE 15.— Estimates of the Atlantic menhaden population catchability coefficient ((/p„pl for years 19.5.'S-79, plotted against the estimated population size (excluding age-0 fish). 588 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY LU O z LLI .3 - .64 .65 .2 " •" .66 •63 •62 .1 " .73 5*6 0 .70 • 72 .'< .67 .77 .61 .58 -.1 -.2 .7 1 .69 .7 5 76. .78 . 79 60 55 .57 .59 -.3 1 1 1 1 1 1 1 1 1 1 1 io6o 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 EFFORT Figure 16. — Differences between observed Atlantic menhaden population fishing mortality rates (Fp,ipl and their mean plotted against nominal effort by fishing season. puter program MAREA (Epperly et al. 1986) as- suming exponential growth of the biomass in the population (Epperly and Nelson 1984). This pro- gram, modified for Atlantic menhaden from MGEAR (Lenarz et al. 1974), uses the Ricker- type yield-per-recruit model (Ricker 1975) and permits estimation of yield in each of the Atlantic menhaden fishing areas, with area-specific growth and fishing mortality rates. Each com- puter run of the model generates a matrix of yield per recruit at varied age of entry and multiples of F for each of the fishing areas (5) and one for the entire fishery. Yield for the entire fishery can be obtained either by summing yield from each area or by calculations based on input from the entire fishery estimated as a unit. A summation of the five matricies should be similar (but not equal) to the independently calculated entire fishery ma- trix. Input for the model includes estimates of area specific proportional fishing mortalities, es- timates of weight at age by area and for the entire fishery, and an estimate of M (0.45). Estimates of area-specific quarterly F at age were obtained by apportioning F at age for the entire fishery (obtained from the quarterly VPA, 1965-78 year classes) with the ratios of the num- ber at age caught in a given area to the total number at age landed in the entire fishery. CThe resulting proportional F's are not equivalent to true area specific F's, but are the correct values for the yield-per-recruit model used). Area-specific length at age estimates for the beginning of each quarter for each year class were rearranged to correspond to fishing season, and converted to weight at age using season-specific weight-length equations (from Table 4). Parallel conversions were done on lengths at age for the entire fishery to estimate weights at age for the fishery as a whole. Annual trends in historic yield per recruit were examined with the fishing mortality, age and size, and geographic pattern extant during each year from 1970 to 1978. Results from these com- putations indicate a severe decline in actual yield per recruit for the entire fishery and Chesapeake Bay area from 1971 to 1978 (Fig. 17). Estimates of potential changes in yield per re- cruit under regimes of varied age at entry and (multiple) changes in fishing mortality rate were obtained by averaging parameters reflecting con- ditions during 3-yr intervals, i.e., 1970-72, 1973- 75, and 1976-78 (Table 9). Attainment of the max- imum potential yield from Atlantic menhaden in the purse seine fishery would have required a very high rate of fishing at a substantially de- layed age at entry of about 3 years of age (Figs. 18, 19). More practically, yield could have been increased by reducing F's. For example, with a F multiplier of 0.6 and the current age of entry, the gain would have been 6.99f for the con- ditions of 1976-78 (Fig. 19, Table 9). With an in- crease in age of entry to 1.0 (eliminate the harvest of age-0 fish) the gain would have been 10.2%. The patterns of potential gain for conditions under the 1970-72 and 1973-75 time periods are similar, but of lesser magnitude (Fig. 18, Table 9). 589 FISHERY BULLETIN: VOL 85, NO. 3 (0 £ (0 QC O UJ ir 0. Q _i UJ 120 110 100 90 80 70 60 50 40 30 20 10- 0 A. ^ Entire Fishery Q Chesapeake Bay ^'^-,--°- '^--O-- V J I L J L _i North Atlantic I? Middle Atlantic 20 10 . South Atlantic ° Fall Fishery 1970 71 72 73 74 75 76 77 78 YEAR OF FISHING Figure 17. — Estimated yield per recruit of Atlantic menhaden for fishing pat- terns and growth prevalent during years 1970-78. Table 9. — Estimates of percentage change in yield per recruit of Atlantic menhaden with varied age at entry and rates of F for 1970-72, 1973-75, and 1976-78. The F multiple of 1.0 represents the Fat age vector extant during each of the three time periods. Time Age at entry F - multiple period 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1970-72 4.0 -35.8 -23.3 -16.1 -11.8 -9.2 -7.5 -6.3 3.5 -21.9 -8.9 -1.7 2.6 5.3 7.3 8.7 3.0 -12.7 0.2 6.2 9.8 12.0 13.4 14.4 2.5 -4.6 6.8 12.1 14.9 16.5 17.5 18.1 2.0 0.5 9.9 13.6 15.1 15.8 16.0 16.0 1.5 -0.7 7.0 9.3 9.6 9.1 8.3 7.3 1.0 -3.4 2.1 2.6 1.3 -0.6 -2.6 -4.7 0.5 -3.9 1.4 1.5 (107.33)1 -2 1 -4.3 -6.5 1973-75 4.0 -7.7 4.6 10.5 13.6 15.4 16,5 17.3 3.5 1.1 12.5 17.7 20.4 22.1 23.1 23.8 3.0 10.5 20.2 242 26.2 27.3 279 28.4 2.5 13.3 19.4 20.9 21.1 20.8 20,4 199 2.0 12.8 15.2 14.8 14.0 13.2 12,5 12,0 1.5 10.3 11.2 9.6 78 6.0 4,5 3,1 1.0 82 8.1 5.6 2.9 0.5 -1.7 -3.7 0.5 6.9 6.2 3.1 (84,07)1 -2.9 -5.5 -7.9 1976-78 4.0 -5.9 12.7 235 30.0 34.1 36.8 388 3.5 1.5 18.8 27.8 32.7 35.5 37.1 38.1 3.0 9.8 25.6 33.2 37.0 39.0 40.0 40.7 2.5 12.1 22.7 26.0 265 26.0 25 1 24.2 2.0 11.7 17.8 18.2 17.0 15.6 14.3 13.2 1.5 9.6 14.0 13.2 11.0 8.8 6.7 4.9 1.0 7.2 10.2 8.3 5.2 2.1 -0.7 -3.2 0.5 5.0 6.9 3.9 (56.84)1 -3.8 -7.2 -10.3 1 Estimated yield per recruit in grams for conditions of the time period, which is the base value for calculation of percentage change 590 AHRENHOLZ ET AL : ATLANTIC MENHADEN POPULATION AND FISHERY 160i CO < cc C3 O -i UJ > «,/i 1 1 1 1 1 r O- 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 F-MULTIPLE Figure 18. — Hypersurface representation of potential yield per recruit for Atlantic menhaden with varied age of recruitment and fishing mortal- ity for conditions extant during 1970-72. The X denotes estimate of actual yield per recruit for the time period. 160 (0 < o O^O 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 F-MULTIPLE Figure 19. — Hypersurface representation of potential yield per re- cruit for Atlantic menhaden with varied age of recruitment and fish- ing mortality for fishing conditions extant during 1976-78. The X denotes estimate of actual yield per recruit for the time period. Results from these analyses indicate a general reduction in the maximum potential as well as actual yield per recruit from the early to the late 1970's, but the maximum potential total yield is greater in the later time period due to increased numbers of recruits (Table 10). Thus, estimation of potential changes in yield by the Atlantic men- haden fishery from yield-per-recruit models be- comes a function of density-dependent growth rates (lower yield per recruit with larger year 591 FISHERY BULLETIN: VOL. 85, NO. 3 Table 10. — Estimates of yield per recruit (grams) and mean yield (thousands of metric tons) of Atlantic menhaden for three, 3-yr intervals, maximum possible yield per recruit (Y/R) and yield with the existing fishing pattern. f^ean Esti- Mean Maxi- Maxi- no. of mated yield mum mum In- recruits Y/R estimate Y/R yield crease Years X106 (g) (t) (g) (t) % 1970-72 2,711.8 107.33 291.1 127.56 345.9 18.8 1973-75 3,778.2 84.07 317.6 108.40 409.6 29.0 1976-78 6,340.5 56.84 360.4 80.49 510.3 41.6 classes), rate of fishing, geographical pattern of fishing, age of recruitment, and numbers of re- cruits. The general conclusion reached with these analyses is that the stock suffers from growth overfishing. To determine if a different initial choice of a constant rate of M would alter this conclusion, the relative biomass of a hypothetical year class was estimated at specific ages with M equal to 0.35, 0.45, and 0.55 and F equal to zero. The growth equation for the 1970 year class (from Table 3) and the annual weight-length expression for 1972 (from Table 4) were used in these compu- tations. The age of maximum biomass decreases with increasing rates of M, as expected, but even at M = 0.55 the age of maximum biomass exceeds 2.5 years (Fig. 20). This decrease (from about age 2.8 for M = 0.45) is insufficient to affect the con- clusion of growth overfishing. However, if the ini- tial choice of M is too high, the analyses are un- derestimating potential gains in total yield that could be realized by decreasing fishing pressure on younger ages. Similarly, if the choice of M is too low, the analyses are overestimating potential gains, but a net gain would still be realized within the range of available estimates of M. Actual Yield by Year Class Using the estimates of numbers caught by age and annual weighted mean weights at age, an- nual landings were apportioned into biomass at age landed and then summed by year class through age 5. The calculations provide estimates of yield by each year class. A plot of yield against year-class size reveals lower than expected yields from the 1975 and 1976 year classes, given their magnitude (Fig. 21). This trend appears to have started about 1973. Comparisons of growth and mortality patterns were made of similar-sized year classes, 1955 and 1976, and 1956 and 1975, in search of causes of the observed decrease in yield. Differences in fish- ing mortality rates at age do not explicitly ac- count for the dramatic differences in yield be- tween the two pairs of similar-sized year classes (Table 11). The 1955 year class was harvested at a greater (less desirable) rate during the critical 160 M=0.35 AGE IN YEARS Figure 20. — Age-specific relative biomass estimates of a hypothetical year class of Atlantic menhaden in the absence of fishing, exposed to three rates of natural mortality. If the year class was harvested instantaneously at any given age, the corresponding ordinate value would represent yield per recruit in grams. 592 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY Ul O < X a D o a. X 1200 1000 ° 800 c o ^ e < I- o 600 400 200 •6S •SS 89 S7 \.71 64 . s« \ '."eb 74 .: 6r>2 " •7- 1" ° 66 J 1_ J I I I L 10 J I L 12 14 16 J I L. 18 YEAR CLASS SIZE (numbers x 10') Figure 21. — Estimated yield contribution of Atlantic menhaden in thousands of metric tons, by the 1955-76 year classes through age 5, plotted against year class size. Table 11. — Estimates of annual Fat age for the 1955 and 1976. and 1956 and 1975 year classes of Atlantic men- haden, through age 5. Year Age class 0 1 2 3 4 5 1955 1976 0.10 0.04 0.58 0.27 1.56 1.51 0.39 0.57 0.43 1.04 0.90 1.02 1956 1975 <0.01 0.03 0.32 0.34 088 1.60 0.74 1.48 0.61 0.60 0.48 1.82 ages of 0-2 than was the 1976 year class. In fact, yield from the 1955 year class could have been markedly increased with a reduced fishing rate on these younger fish. The harvest rates of the 1956 and 1975 year classes were similar for ages 0-1, but markedly greater at age 2 for the 1975 cohort. While the high age-2 fishing mortality rate probably contributed to the lower yield, it did not fully explain the marked difference. The gen- erally higher rates of F for ages 3-5 exhibited for both the 1975 and 1976 year classes were proba- bly inconsequential for this comparison of yield. While an increase in the true value of natural mortality could cause a decrease in total yield by year class, differences in growth provide a more obvious explanation for the differences within these two pairs of year classes. Growth curves (in length and weight) for the 1956 and 1975 year classes and the 1955 and 1976 year classes show that individuals were much smaller during the dominant harvest ages (1-3) for the two most re- cent year classes (Figs. 22, 23). These differences could be great enough to account for most of the differences observed in yield. Slower relative growth in post age-1 fish has been apparent for year classes 1973-78. To determine if age of maximum theoretical biomass had changed owing to the different (flat- ter) shaped growth curves displayed by the year classes in the later 1970's, the relative biomass at age of an unfished hypothetical year class was estimated with the growth equation for the 1975 year class with M equal to 0.45 and the annual weight-length expression for 1972. The results display an increase in maximum age (to about 3.25 years), a decrease in total biomass, and a much slower ascent and even slower descent from maximum biomass than results for the 1970 year class growth curve (Fig. 24). These results indi- cate that age of entry to the fishery could have been greatly delayed in the later 1970's with little chance of losing yield. Given the progressive decrease in average size at age of fish in the age classes which dominate landings (Fig. 25), the decline in yield per recruit following 1971 (Fig. 17) is expected. However, the rapid decline in size at age is not entirely ascrib- able to density-dependent growth. More impor- tantly, the potential for increased yield with re- ductions in F is probably greater than the results from the MAREA yield-per-recruit model indi- cate owing to the likelihood of size selective fish- ing and the potential for differential stock- specific growth rates. Additionally, the relatively high MSY estimate obtained from the PRODFIT 593 FISHERY BULLETIN: VOL. 85, NO. 3 z LU O 350- 300 250 I 200 150- 100 CO < cr o X LLJ Figure 22. — Comparative fitted von Bertalanfiy growth curves of two similar-sized (numbers offish) Atlantic menhaden year classes, 1955 (solid curves) and 1976 (dashed curves). Upper curves are fork length in millimeters, and lower curves are weight in grams. a. o 350 300- 250 S 5 2 200- I t- O uj 150 100- < cr o I g LU AGE IN YEARS Figure 23. — Comparative fitted von Bertalanffy growth curves of two similar-sized (numbers offish) Atlantic menhaden year classes, 1956 (solid curves) and 1975 (dashed curves). Upper curves are fork length in millimeters, and lower curves are weight in grams. 594 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY 160- 140 120 w < 100 o m > 80 < 60 40 20- 1970 1975 _L _L J I L _L J I L _J 0 0 5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6 0 AGE IN YEARS Figure 24. — Age-specific relative biomass estimates of a hypothetical year class of Atlantic menhaden in the absence of fishing, with growth parameters estimated for the 1970 and 1975 year classes. If harvesting occurred instanta- neously at any given age, the corresponding ordinate value would represent yield per recruit in grams. w < o I z < 600 500 400 300 200- 100- /^ /\. :, Age 3 Age 2 \ / \.-v /^./\ A_ge I'^Xxv A • — /•■ \ Age 0 "^---^ I ■ I ■ I ■ I ■ I ■ I ' I ■ I ' I ' I ' "^ 1955 57 59 61 63 65 67 69 71 73 75 77 79 81 YEARS Figure 25. — Weighted mean annual weight of purse seine landed Atlantic menhaden, ages 0-3, for years 1955-81. solution shown earlier may be more realistic than first impressions would indicate. However, to at- tain that level of harvest would require a restruc- turing of the fishery, and continued moderate to high levels of recruitment to sustain it. GENERAL DISCUSSION With estimates of year-class sizes, exploitation rates, and an understanding of the interaction of population size and effort relative to rates of fish- ing mortality, the trends observed in the fishery since 1955 are more readily explained. Addition- ally, this information permits inferences to be drawn about the earlier presampling period of the 1940's and early 1950's. The major premise of this discussion is that harvest levels and effort of the earlier presampling period were probably near 595 FISHERY BULLETIN: VOL 85, NO. 3 the maximum that the population could support, and thus the rather rapid growth of the fishery in the early to mid-1950's was a response to a sub- stantial increase in abundance of Atlantic men- haden and not simply increased effort applied to an underexploited stock. Changes in stock abundance due to fluctuations in spawning success were prevalent in the 1940's. This conclusion is drawn from the statement of purpose for study given by June and Reintjes (1959). They noted changes in abundance among geographic areas and seasons and some poor catches. This condition is expected when recruit- ment fluctuates in a fishery that has a stock which is differentially distributed by age and size. Data from early years indicated a limited re- source. Catch closely paralleled effort for the 1941-47 seasons, but catches during 1948 and 1949 (about 350,000 t) were less than expected given the effort expended (Fig. 1). Apparently the stock subsequently underwent a marked increase in abundance, noticeable first about 1952, but even more pronounced in 1953 and 1954. The fishery responded, as effort again began to rise, but lagged for the next two or three seasons. In 1959 catches were dominated by the 1958 year class, which continued to provide significant biomass to the fishery through 1962. It appears that the 1950's marked a period with above average recruitment, and this was accentu- ated with the apparently very large 1951 year class, followed by the three relatively large year classes of 1953, 1955, and 1956, and finally the largest documented year class of 1958. Recruit- ment did not return to the level of the 1955 year class (the lesser of the three documented large year classes) until 1975. To obtain some idea of relative sizes, a reconstruction of these earlier year classes was made using arbitrary, but con- servative values for F (F = 0.25 for age 1 and 0.50 for age 2 + ) (Table 12). Given the (older) larger sized fish which were taken during the peak land- ing years, these speculative year class estimates are less than or nearly equal to a size necessary to support the large catches of the mid-1950's (see Figure 14 and Table 7). The age structure of the Atlantic menhaden population has undergone at least two periods of expansion and contraction since about 1950, and has shown signs of expanding again by 1980 (Fig. 6). Inferences on early age composition were derived primarily from the early 1952-54 sam- pling of the fishery in the middle Atlantic area (June and Reintjes 1959, their appendix table 2). Age-5 and older fish were 1.0%, 0.8%, and 1.1% by number in the samples from that area for 1952, 1953, and 1954 as compared to 1.8% and 2.9% for 1955 and 1956. The 1951 year class dominated the catch during this period, comprising 84.59% of the samples as age 1, 98.05% as age 2, and 59.02% as age 3 in the middle Atlantic area and 87.1 1%> as age 3 in the newly sampled north At- lantic area. Hence, as noted earlier, the stock probably had its strongest age structure in 1955 and 1956, which were coincidentally the first years for which port sampling covered the full geographic range of the fishery. This strength was due to the increased population size, subse- quent decrease in the catchability coefficient, and thus a reduced fishing mortality on most age groups. Higher rates of survival led to more indi- viduals in the older age groups. The landings were sustained above 500,000 t in the mid-1950's by contributions of the 1955 and 1956 year classes, but these year classes were too small to prevent an increase in mortality because of in- creased effort and the number of older fish were reduced by 1959. The large 1958 year class aided the replenishment of the older age groups, by about 1960. Without another large year class, Table 12. — Estimates from virtual population analyses (mixed time interval, see text) of number at age in tfiousands on 1 March for the 1950-59 year classes of Atlantic menhaden. Bracketed values represent estimates obtained from back calculations using arbitrary values of F(0.25 for age 1 and 0.50 for ages 2 + ) Year class Age 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1 [3,943,235] (10,782,529] [3,162,907] [4,504,191] 3,040,558 5,621,258 7,153,549 3,263,196 14,767,294 2,164,428 2 [1,958,152] [5,354,445] [1,570,653] 2,236,715 1,410,646 1,976,767 3,306,709 1,412,179 6,627,079 1,158,337 3 [757,298] [2,070,784] 607,436 642,482 240,748 274,476 893,865 275,897 2,357,641 361,743 4 [292,878] 800,857 189,410 169,952 80,038 119,036 270,347 116.411 602,146 71,280 5 113,268 281,543 86,255 53,482 37,304 48,835 92,494 59,258 75,460 11,589 6 43,677 63,025 23,309 21 ,243 14,284 12,269 35,962 14,698 8,894 1,616 7 5,980 1 1 ,726 7,639 3,717 2,765 5,618 4,883 1,629 901 161 18 600 3,564 1,446 478 1,162 1,331 640 92 48 — 'This age group may contain a small number of age 9 or 10 individuals, see text 596 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY mortality apparently again increased, as the age structure became markedly constricted by 1965, and drastically truncated by 1967. Recruitment began to improve by 1971, culminating with rela- tively large year classes in 1975 and 1979. Mor- tality rates for some ages began to decline during the late 1970's, and the older age groups began to strengthen, consistent with the decline of the catchability coefficient with increased population size. With respect to yield per recruit, there appears to be no recorded period of Atlantic menhaden fishing when an ideal harvesting regime existed in the purse seine fishery. Age-0 fish have been harvested since at least 1955 (Figs. 8, 9). Except for influences of the exceptionally large 1951 and 1958 year classes, most of the catches sampled for age have been dominated by age 2 relative to biomass, and ages 1 and 2 relative to numbers. Major numerical but minor biomass contribu- tions have been evident for age-0 fish. Inferences of the fishery's high dependency on younger age groups can be traced back to 1940. Given that the population distributes itself by age and size along the Atlantic coast, the quantity of landings and degree of effort expended in areas where younger and smaller fish predominate suggests a similar age composition for total catches during the pre- sampling period (Figs. 3, 5). Landings in the middle Atlantic area domi- nated the fishery in earlier years, but a shift had occurred by 1964, at which time Chesapeake Bay landings began to dominate (Fig. 4). Responding to a reduced population of larger and older fish, the industry increased the proportion of fishing effort exerted in areas closer to the large nursery areas of Chesapeake Bay and the south Atlantic area. Further, the fishery shifted from one that harvested the larger age I's and 2's, and older fish, to one that harvests the smaller and younger fish. The larger, older fish were and still are vul- nerable to the fishery during their fall migra- tions, but effort on these fish appears to be re- duced within the north Atlantic and middle Atlantic areas. A comparison of age-specific estimates of ex- ploitation rates supports the earlier discussion on age dependency, population sizes, and age struc- ture (Fig. 26). In the mid-1950's, the exploitation rates, although varying, were lower than those g I- < 0. X m < IT 1.00- .90 .80 .70 .60 .50 .40 .30- .20 .10 0 .80- .70 .60 .50 .40 .30 .20 .10 Age 5 Age 4 / Age 2 V / \° ' / • ?^ — ■ o -Age 3 /\ Age 1 \. j^l^i. /, ^— Age 0 _L 1956 60 64 68 72 76 80 YEAR OF FISHING Figure 26. — Estimates of annual rates of exploitation of Atlantic men- haden, ages 0 through 5. 597 FISHERY BULLETIN: VOL. 85. NO. 3 for the subsequent years when population size was decreasing. The rates reached their lowest points for ages 1-3 during 1960, when the 1958 year class was fully recruited. A low point fol- lowed for ages 4 and 5 during 1961. All exploita- tion rates generally increased during the low re- cruitment years of the 1960's, and began decreasing during the 1970's as the population size began to increase due to higher recruitment. The rate of reduction noted for nominal effort (Fig. 1) lagged behind that of the stock and was apparently too slow to prevent the observed rise in exploitation rates during the 1960's. The exploitation rates of age-2 fish appear to have progressively declined during the later 1970's, following a disproportionately large in- crease after 1971. This increase apparently was a product of a shift in the pattern of fishing that occurred during the regrowth of the fishery and that pattern still exists. Although slightly lag- ging behind that of the age-2 fish, exploitation rates on age-0 fish began increasing by 1974, and reached an alarming rate of about 15% in 1979 (preliminary estimate). This rate of exploitation occurred in virtually one quarter of fishing and slightly exceeded the rate of exploitation for age-1 fish for the entire season. The disproportionately low exploitation rate on age I's is probably due to their increasingly smaller size, a consequence of which would be a more southerly distribution. Additionally, many would remain in or near estu- arine nursery areas and surrounding smaller bays and sounds, and thus be less available to the fishery. ACKNOWLEDGMENTS We are grateful to the many individuals who contributed to the countless tasks involved in the development and maintenance of the complex 29- yr Atlantic menhaden data set. Donnie L. Dudley, Joseph E. Hightower, John E. Hollingsworth, David E. Hopkins, Charles W. Krouse, and Larry L. Massey made significant contributions to vari- ous aspects of data collection and processing and analytical computer model development and im- plementation. Gary T. Sakagawa provided help- ful suggestions for analytical approaches. Alex- ander J. Chester, Charles S. Manooch III, John V. Merriner, William R. Nicholson, William E. Schaaf, and Douglas S. Vaughan provided excel- lent reviews and helpful comments during vari- ous stages of report preparation. We thank Her- bert R. Gordy (figure preparation) and Beverly W. Harvey (word processing) for their patience dur- ing the course of the many changes we made. A special thanks is due to Robert B. Chapoton (de- ceased) for ideas, review, and the needed encour- agement during most of the development of this report. LITERATURE CITED Abramson, N. J. 1971. Computer programs for fish stock assessment. FAO Fish. Tech. Pap. 101:1-154. ATLANTIC States Marine Fisheries Commission 1981. Fishery management plan for Atlantic menhaden. Atl. States Mar. Fish. Comm., Fish. Manage. Rep. 2, 134 p. Chester. A J 1984. Sampling statistics in the Atlantic menhaden fish- ery. U.S. Dep. Commer., NOAA Tech. Rep. NMFS 9, 16 p. Clark. C W , and M Mangel 1979. Aggregation and fishery dynamics: a theoretical study of schooling and the purse seine tuna fisheries. Fish. Bull., U.S. 77:317-337. DERISO, R. B . T J. QUINN II, AND P R NEAL 1985. Catch-age analysis with auxiliary information. Can. J. Fish. Aquat. Sci. 42:815-824. Dryfoos, R L . R P Cheek, and R L Kroger 1973. Preliminary analyses of Atlantic menhaden, Bre- voortia tyrannus, migrations, population structure, sur- vival and exploitation rates, and availability as indicated from tag returns. Fish. Bull., U.S. 71:719-734. Epperly, S P 1981. A population investigation of Atlantic menhaden: a meristic, morphometric and biochemical approach. MS Thesis, Univ. South Florida, Tampa, 107 p. Epperly, S. P.. W H Lenarz, L L Massey, and W R Nelson. 1986. A generalized computer program for yield per re- cruit analysis of a migrating population with area specific growth and mortality rates. U.S. Dep. Com- mer., NOAA Tech. Memo. NMFS-SEFC-180, 26 p. Epperly, S. P., and W R Nelson 1984. Arithmetic versus exponential calculations of mean biomass. Fish. Bull., U.S. 82:446-448. Ferraro, S. p. 1980. Daily time of spawning of 12 fishes in the Peconic Bays, New York. Fish. Bull., U.S. 78:455-464. Fox. W W , Jr 1975. Fitting the generalized stock production model by least squares and equilibrium approximation. Fish. Bull., U.S. 73:23-37. Henry. K A 1971. Atlantic menhaden (Brevoortia tyrannus) resource and fishery - analysis of decline. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF 642, 32 p. Higham, J R , AND W R Nicholson 1964. Sexual maturation and spawning of Atlantic men- haden. Fish. Bull., U.S. 63:255-271. Judy. M. H , and R M Lewis 1983. Distribution of eggs and larvae of Atlantic men- haden, Brevoortia tyrannus. along the Atlantic coast of the United States. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF 774, 23 p. 598 AHRENHOLZ ET AL.: ATLANTIC MENHADEN POPULATION AND FISHERY June, F. C. 1958. Variation in meristic characters of young Atlantic menhaden, Brevoortia tyrannus. Rapp. P. -v. Reun. Cons. int. Explor. Mer 143:26-35. 1965. Comparison of vertebral counts of Atlantic men- haden. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 513, 12 p. June, F C. and J. L. Chamberlin. 1959. The role of the estuary in the life history and biol- ogy of Atlantic menhaden. Proc. Gulf Caribb. Fish. Inst. 11:41-45. June, F C, and W R Nicholson 1964. Age and size composition of the menhaden catch along the Atlantic coast of the United States, 1958. U.S. Fish Wildl. Serv. Spec. Sci. Rep. Fish. 446, 40 p. June, F. C , and J W Reintjes. 1959. Age and size composition of the menhaden catch along the Atlantic coast of the United States, 1952-55; with a brief review of the commercial fishery. U.S. Fish. Wildl. Serv., Spec. Sci. Rep. Fish. 317, 65 p. June, F C . and C M Roithmayr 1960. Determining age of Atlantic menhaden from their scales. Fish. Bull., U.S. 60:323-342. Kendall, A W , Jr , and J W Reintjes 1975. Geographic and hydrographic distribution of At- lantic menhaden eggs and larvae along the middle At- lantic coast from RV Dolphin cruises, 1965-66. Fish. Bull., U.S. 73:317-335. Kroger. R L , R L. Dryfoos. and G R Huntsman 1971. Movement of juvenile Atlantic menhaden tagged in New England. Chesapeake Sci. 12:114-115. Kroger, R L , and J F Guthrie 1973. Migrations of tagged juvenile Atlantic menhaden. Trans. Am. Fish. Soc. 102:417-422. Kroger. R L , J F Guthrie, and M H Judy 1974. Growth and first annulus formation of tagged and untagged Atlantic menhaden. Trans. Am. Fish. Soc. 103:292-296. Kuntz, a . AND L. Radcliffe. 1917. Notes on the embryology and larval development of twelve teleostean fishes. Bull. U.S. Bur. Fish. 35:87- 134. Lenarz, W H , W W Fox, Jr. G T Sakagawa, and B J Rothschild 1974. An examination of the yield per recruit basis for a minimum size regulation for Atlantic yellowfin tuna, Thunnus albarcares. Fish. Bull., U.S. 72:37-61. Leslie, P H 1945. On the use of matrices in certain population mathe- matics. Biometrika 33:183-212. Lewis. R M , and W C Mann 1971. Occurrence and abundance of larval Atlantic men- haden, Brevoortia tyrannus , at two North Carolina inlets with notes on associated species. Trans. Am. Fish. Soc. 100:296-301. Lewis. R M . E P H Wilkens. 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. Marquardt, D W 1963. An algorithm for least-squares estimation of non- linear parameters. J. Soc. Ind. Appl. Math. 11:431-441. McHugh, j L . R T Oglesby, and A L Pacheco 1959. Length, weight, and age composition of the men- haden catch in Virginia waters. Limnol. Oceanogr. 4:145-162. Neuson. W. R., and D. W. Ahrenholz. 1986. Population and fishery characteristics of gulf men- haden, Brevoortia patronus. Fish. Bull., U.S. 84:311- 325. Nelson, W. R., M. C. Ingham, and W E. Schaaf. 1977. Larval transport and year-class strength of Atlantic menhaden, Brevoortia tyrannus. Fish. Bull., U.S. 75:23- 41. Nicholson, W. R 1971a. Changes in catch and effort in the Atlantic men- haden purse-seine fishery 1940-68. Fish. Bull., U.S. 69:765-781. 1971b. Coastal movements of Atlantic menhaden as in- ferred from changes in age and length distributions. Trans. Am. Fish. Soc. 100:708-716. 1972. Population structure and movements of Atlantic menhaden, Brevoortia tyrannus, as inferred from back- calculated length frequencies. Chesapeake Sci. 13:161- 174. 1975. Age and size composition of the Atlantic menhaden, Brevoortia tyrannus, purse seine catch, 1963-71, with a brief discussion of the fishery. U.S. Dep. Commer., NOAA Tech. Rep. NMFS SSRF 684, 28 p. 1978. Movements and population structure of Atlantic menhaden indicated by tag returns. Estuaries 1:141- 150. PELLA, j j , AND P K TOMLINSON. 1969. A generalized stock production model. Inter-Am. Trop. Tuna Comm. Bull. 13:421-496. Pope, J G. 1972. An investigation of the accuracy of virtual popula- tion analysis using cohort analysis. Int. Comm. North- west Atl. Fish. Res. Bull. 9:65-74. Reintjes, J W 1969. Synopsis of biological data on Atlantic menhaden, Brevoortia tyrannus. U.S. Fish. Wildl. Serv., Circ. 320, 30 p. Reintjes, J W , and A L Pacheco 1966. The relation of menhaden to estuaries. Am. Fish. Soc. Spec. Publ. 3:50-58. Reish, R L , R B Deriso. D Ruppert, and R J Carroll 1985. An investigation of the population dynamics of At- lantic menhaden {Brevoortia tyrannus). Can. J. Fish. Aquat. Sci. 42 (Suppl. 1):147-157. Richer. W E 1954. Stock and recruitment. J. Fish. Res. Board Can. 11:559-623. 1958. Handbook of computations for biological statistics of fish populations. Bull. Fish. Res. Board Can. 119, 300 p. 1975. Computation and interpretation of biological statis- tics offish populations. Bull. Fish. Res. Board Can. 191, 382 p. Roithmayr. C M 1963. Distribution of fishing by purse seine vessels for Atlantic menhaden, 1955-59. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 434, 22 p. Ruppert, D , R L Reish, R B Deriso, And R J Carroll 1985. A stochastic population model for managing the At- lantic menhaden {Brevoortia tyrannus ) fishery and as- sessing managerial risks. Can. J. Fish. Aquat. Sci. 42:1371-1379. Schaaf. W E. 1975a, Status of the Gulf and Atlantic menhaden fish- eries and implications for resource management. U.S. Natl. Mar. Fish. Serv., Mar. Fish. Rev. 37(9):l-9. 599 FISHERY BULLETIN: VOL. 85, NO. 3 1975b. Fish population models: potential and actual links to ecological models. In C. S. Russell (editor), Ecological modeling in a resource management framework, p. 211- 239. Resources for the Future, Washington, D.C. 1979. An analysis of the dynamic population response of Atlantic menhaden, Brevoortia tyrannus, to an intensive fishery. Rapp. P.-v Reun. Cons. int. Explor. Mer 177:243-251. ScHAAF, W. E , .^ND G. R Huntsman. 1972. Effects of fishing on the Atlantic menhaden stock: 1955-1969. Trans. Am. Fish. Soc. 101:290-297. Smith, J. W, W R Nicholson, D S Vaughan, D L Dudley, and E. A. Hall In press. Age and size composition of the Atlantic men- haden, Brevoortia tyrannus, purse seine catch, 1972- 1984, with a brief discussion of the fishery. NOAA Tech. Rep. NMFS. Sutherland, D. F. 1963. Variation in vertebral numbers of juvenile Atlantic menhaden. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 435, 21 p. TOMLINSON, P K. 1970. A generalization of the Murphy catch equation. J. Fish. Res. Board Can. 27:821-825. Ulltang, O. 1977. Sources of errors in and limitations of virtual popu- lation analysis (cohort analysis). J. Cons. Int. Explor. Mer 37:249-260. WiLKENS, E P H . AND R M. LEWIS. 1971. Abundance and distribution of young Atlantic men- haden, Brevoortia tyrannus, in the White Oak River Es- tuary, North Carolina. Fish. Bull., U.S. 69:783-789. 600 DENSITY AND DEPTH DISTRIBUTION OF LARVAL GULF MENHADEN, BREVOORTIA PATRONUS, ATLANTIC CROAKER, MICROPOGONIAS UNDULATUS, AND SPOT, LEIOSTOMUS XANTHURUS, IN THE NORTHERN GULF OF MEXICO Susan M. Sogard,' Donald E. Hoss,- and John J. Govoni^ ABSTRACT Densities of larval gulf menhaden, Brevoortia patronus; Atlantic croaker, Micropogonias undulatus; ■ and spot, Leiostomus xanthurus. compared among three transects in the northern Gulf of Mexico, indicated that all three species were more abundant at inshore ( 18 m isobath) than offshore stations (91 and 183 m isobaths). Gulf menhaden and Atlantic croaker were most abundant off Southwest Pass, Louisiana, a major outlet of the Mississippi River into the Gulf of Mexico. Gulf menhaden larvae caught at inshore stations were larger than those collected at offshore stations. Of the three species, only gulf menhaden showed any consistent pattern in vertical distribution. At inshore stations, gulf menhaden were concentrated near the surface at midday, but distributed across sampling depths (1 m, 6 m, and 12 m) at dawn, dusk, and midnight, a pattern opposite to that typically reported for larval fish. At offshore stations (with sampling depths of 1 m, 30 m, and 70 m), gulf menhaden larvae were present at 70 m, but most were caught near the surface. A concentration in surface waters was again most pronounced at midday. Gulf menhaden, Brevoortia patronus; spot, Leios- tomus xanthurus; and Atlantic croaker, Micropo- gonias undulatus, are thought to spawn offshore in winter months in the northern Gulf of Mexico (Nelson 1969; Fore 1970; Diaz 1982; Christmas et al. 1982). Larvae of the three species are trans- ported inshore to nursery grounds in marshes and estuaries along the northern coast. One passive mechanism suggested for movement of gulf men- haden includes longshore advective transport, en- trainment into the coastal boundary layer, and eventual transport into the estuary effected by the seasonal rise of sea level in spring (Shaw et al. 1985a). The passage of winter cold fronts can also be expected to influence transport. Spawning of gulf menhaden occurs in shelf waters out to at least 91 m (Guillory et al. 1983), but is concentrated around the Mississippi River Delta (Fore 1970). Atlantic croaker apparently spawn in waters <54 m in depth (Diaz 1982), while spot spawn in waters >27 m (Dawson 1958; Nelson 1969). Gulf menhaden larvae spend 3 to 5 weeks at sea before entering estuaries 'Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Service, NOAA, Beaufort, NC 28516; present address: Rutgers Marine Field Station, P.O. Box 41. Tuckerton, NJ 08087. ^Southeast Fisheries Center Beaufort Laboratorv, National Marine Fisheries Service, NOAA, Beaufort, NC 28516. Manuscript accepted March 1987. FISHERY BULLETIN: VOL. 85. NO. .3, 1987. when they are 12 to 25 mm in length (Reintjes 1970; Christmas and Etzold 1977; Guillory et al. 1983). Fish larvae are nonrandom in their spatial dis- tribution in both the vertical and horizontal di- mensions. One primary influence on the vertical distribution of larvae is their diel vertical move- ment (migration) in the water column; larvae of many species rise to the surface by night and de- scend by day (e.g.. Smith et al. 1978; Kendall and Naplin 1981; Sameoto 1982, 1984). Horizon- tal distribution is also dynamic, with dispersion and aggregation of larvae affected by such factors as adult spawning behavior, water mass move- ments, localized larval mortality, and larval be- havior (Smith 1981; Houde 1982; Jahn and Lavenberg 1986). In this study we examined the density and depth distribution of larval gulf menhaden, spot, and Atlantic croaker at three locations in the northern Gulf of Mexico, with emphasis on the area around Southwest Pass, LA, the main dis- charge of the Mississippi River among the delta distributaries. Size distributions of gulf men- haden were compared to determine if inshore lar- vae were older than offshore larvae, the expected pattern if adults are spawning primarily offshore and larvae are moving inshore. 601 FISHERY BULLETIN: VOL. 85, NO. 3 METHODS Larvae were collected along three inshore- offshore transects (off Southwest Pass, LA; Cape San Bias, FL; and Galveston, TX) at stations posi- tioned over the 18 m (10 fm), 91 m (50 fm), and 183 m (100 fm) isobaths (Fig. 1). Sampling took place on four cruises in December 1979, February 1980, December 1980, and February 1981. Collec- tions were made with a Multiple Opening/Closing Net and Environmental Sensing System (MOC- NESS, Wiebe et al. 1976). The MOCNESS con- sisted of nine 505 |xm mesh Nitex'^ plankton nets with mouth openings of 1 m by 1.4 m. Due to equipment problems, only the inshore Southwest Pass station was sampled on the first cruise. Galveston stations were added to the sampling program on the February 1981 cruise. MOCNESS nets were deployed in the following manner: Net 1 remained open as the MOCNESS descended from the surface to the deepest depth to be sampled. Nets 2 and 3 sampled at that depth, one at a time, and net 4 opened as the MOCNESS was raised to an intermediate depth, where nets 5 and 6 sampled. Net 7 was open while the MOC- NESS was brought to the surface, where nets 8 and 9 fished. Discrete depth nets generally fished from 2 to 3 minutes before deployment of the next net. Sampling depths were approximately 12, 6, and 1 m at inshore stations and 70, 30, and 1 m at the offshore stations. At each station, MOCNESS casts were made at 0600, 1200, 1800, and 2400 h, with a towing speed of approximately 2 nmi/hour. Sensors on the MOCNESS provided continuous recording of temperature and depth. Two flowme- ters, one mounted on top of the MOCNESS and one within the net opening, were used to calculate the volume of water sampled by each net and to detect net clogging. The mean volume filtered by each discrete depth net was 140 m'^ (SD = 101.2, n = 529). The collection of one net at each discrete depth was preserved in 5% buffered formalin-seawater and the collection of the other was preserved in 70% ethanol. Formalin-preserved larvae were used in gut content analysis (Govoni et al. 1983), and alcohol-preserved larvae were used in otolith analysis of age and growth (Warlen in prep'*). In the laboratory all fish larvae were removed from ■^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 4S. M. Warlen. Manuscr. in prep. Southeast Fisheries Center Beaufort Laboratory, National Marine Fisheries Serv- ice, NOAA, Beaufort. NC 28516. 30° 00' - 28° 00' 26 00' - 24° 00' 96° 00' 94° 00' 92° 00' 90° 00' 88° 00' 86° 00' 84° 00' Figure 1. — Location of sampling transects and stations in the northern Gulf of Mexico. 602 SOGARD ET AL : LARVAL OLU.F MENHADEN. ATLANTIC CROAKER. AND SPOT the samples and counted. Gulf menhaden were measured (standard or notochord length) to the nearest 0.01 mm with an ocular micrometer. When more than 30 menhaden occurred in a sam- ple, 30 individuals were randomly selected and measured (length measurements were not cor- rected for shrinkage). Due to the scarcity of lar- vae at outer stations, menhaden from both the 91 and 183 m Southwest Pass stations were com- bined in the comparison of offshore size distribu- tion with inshore (restricted to the Southwest Pass station to allow valid comparison). Spot and Atlantic croaker were too rare at all offshore sta- tions to allow size comparisons. Analysis of vertical movement was based on the mean percentage of larvae caught at each of three discrete depths on each MOCNESS cast, allowing comparison among casts with widely varying densities of larvae. Gulf menhaden caught at the inshore stations were divided into three size classes to determine if vertical distribu- tion varied with age. Because of the low number of larvae at the 183 m stations (see above), analy- sis of offshore vertical distribution was based on MOCNESS casts from both the 91 and 183 m sta- tions. Mean densities at each depth were also cal- culated for each time of sampling. RESULTS Overall densities (number x 100 m •^) of gulf menhaden, spot, Atlantic croaker, and the total of all species (these three species plus all others, including damaged and unidentifiable clupeids that may have been gulf menhaden) varied widely among cruises, stations, times, and depths. The majority of the 529 net tows did not catch any gulf menhaden (67%), spot (83%), or Atlantic croaker (82%). Smaller individuals of all three species, however, were probably not re- tained by the 505 jjim mesh nets. In all but four cases, gulf menhaden were more abundant than spot or Atlantic croaker (Table 1). The density of all three species was generally greatest at the inshore (18 m) stations and declined offshore, with low or zero densities common at both off- shore stations of Cape San Bias and Galveston (Table 1). Gulf menhaden were most abundant at the Southwest Pass stations, except on the December 1980 cruise, when they were most abundant at the 18 m Cape San Bias station (Table 1). Atlantic croaker larvae were most abundant at the inshore Southwest Pass station in December 1980 and February 1981, but not in February 1980. Spot Table 1 — Mean densities (SD in parentheses) of ichthyoplankton (larvae ■ 100 m 3) collected at three stations at three sites in the northern Gulf of IVIexico. Densities are averaged over three discrete depths and four times of day. Station 1 was over the 18 m isobath, station 2 over the 91 m isobath, and station 3 over the 183 m iso- bath, n = number of net tows. "Total larvae" includes the three target species and all others. Total Gulf Atlantic Cruise Station n larvae menhaden Spot croaker Dec 1979 SW Pass 1 24 78.5 (88.9) 43.7 (66 9) 1.3 (2.3) 3.5 (3.9) Feb. 1980 SW Pass 1 24 228.7 (281.4) 79.3 (122.5) 1.1 (0.3) 0.1 (0.3) SW Pass 2 23 32.1 (43.0) 7,4 (16.0) 0 0 SW Pass 3 7 7.5 (5.9) 0,4 (0.7) 0 0 San Bias 1 24 20.0 (14.4) 0.2 (10.5) 0.6 (0 9) 0 6 (0.8) San Bias 2 23 114.3 (85.3) 0 1 (0.2) 0.1 (0.3) 0 1 (0.4) San Bias 3 16 18.4 (13.2) 0 0 0 Dec 1980 SW Pass 1 35 38.4 (30.4) 6 0 (8 0) 0.1 (0.2) 9.1 (15.2) SW Pass 2 24 56.0 (55.5) 88 (18.3) 0.2 (0.6) 3.1 (7 8) SW Pass 3 24 43 7 (38 4) 0.7 (2.1) 0 0 San Bias 1 24 330 5 (618.6) 14.1 (42.4) 98.1 (259.3) 0.1 (0.1) San Bias 2 24 112.8 (83.2) 0 0.5 (1.2) 0 1 (0.3) San Bias 3 24 55.1 (61.0) 0 0 0 Feb 1981 SW Pass 1 23 36.7 (49.7) 26 8 (43 8) 0 6 (0.9) 1.8 (2.6) SW Pass 2 22 65.8 (69.1) 13.6 (17.5) 0.6 (0.9) 0 6 (1.2) SW Pass 3 24 18.6 (150) 6.2 (11.8) 0.1 (0.1) 0 San Bias 1 23 28.7 (15.8) 0 0 0 San Bias 2 24 10.0 (9.0) 0 0 0 San Bias 3 24 9.3 (6.8) 0 0 0 Galveston 1 23 53.2 (26.1) 9.3 (6.8) 0 0 Galveston 2 24 87 1 (650) 0.1 (0.1) 0 0 Galveston 3 24 22.1 (16.4) 0 0 0 603 FISHERY BULLETIN: VOL. 85, NO. 3 larvae were never very abundant except at the inshore San Bias station on the December 1980 cruise (Table 1), when a single dense patch of larvae was encountered on two successive MOC- NESS casts (Govoni et al. 1985). Densities were as high as 993 larvae x 100 m"'^, an abundance not observed for other species and not approached by spot densities in any other sample. On the Table 2. — Mean standard length (mm, SE in parentheses) of gulf menhaden caught on three cruises at inshore (18 m isobath) and combined offshore (91 and 183 m isobaths) stations off Southwest Pass, LA. F values are results of a two-way ANOVA comparing lengths among cruises and between stations. Station F Cruise Cruise x Cruise Inshore Offshore Station Station Feb. 1980 15.2(0.3) 8.9(0.2) 21.8" 81.8" 62.6" Dec. 1980 13.5 (0.3) 8.7 (0.3) Feb. 1981 12.3 (0.2) 12.3 (0.2) "P< 0.001. Inshore 2/80 >- o z UJ ID o LU cc LL LU O < z UJ o cr lU a. Offshore 2/80 Inshore 12/80 Offshore 12/80 Inshore 2/81 n= 383 Offshore 2/81 LENGTH (mm) Figure 2. — Length-frequency distributions of gulf menhaden larvae collected from inshore (18 m isobath) and offshore (91 and 183 m isobaths combined) stations off Southwest Pass, Lou- isiana, on three cruises (February 1980, December 1980, and February 1981). n = number of larvae measured. February 1981 cruise, no spot larvae were col- lected at this station. The occurrence of this ag- gregation, then, appeared to be a reflection of patchiness rather than geography. An inshore-offshore comparison (Southwest Pass stations) of menhaden lengths revealed that on two of three cruises (February 1980 and De- cember 1980) larvae were larger at the inshore station (Table 2). This was not the case on the third cruise, however, when mean lengths were similar. Results of a two-way ANOVA (Table 2) indicated significant differences in mean length for both main effects of station and cruise and for their interaction (P < 0.001). Thus, the pattern to the data is more complex than can be summarized by main effects alone. Length-frequency distribu- tions indicated a bimodal pattern at the inshore station with most larvae in the larger size mode (Fig. 2). Larger larvae were not as common at the offshore stations except on the final cruise, when a bimodal pattern occurred offshore as well as inshore. At the inshore stations, total larvae were gen- erally distributed evenly among the three sam- pling depths at all times of day (Fig. 3), although Table 3. — Mean density (larvae x lOO m^S) at each dis- crete depth during each time period. MOCNESS casts in which no larvae of the target species were caught at any of the discrete depths were not included in calculation of means. Time Station/species 0600 1200 1800 2400 Inshore Total larvae 1 m 6 m 12 m Gulf menhaden 1 m 6 m 12 m Spot 1 m 6 m 12 m Atlantic croaker 1 m 6 m 12 m Offshore Total larvae 1 m 30 m 70 m Gulf menhaden 1 m 30 m 70 m 68.23 120.97 341.10 76.36 56.01 25.76 144.20 86.82 42.44 25.24 69.20 173.78 31.72 4.54 13.16 1.36 0.76 0.82 2.05 4.08 12.65 103.90 35.41 23.38 23.70 7.47 0.22 37.35 48.83 0.14 101.06 0.60 29.93 1.99 1.04 0.39 1.07 4.05 1.32 57.37 21.30 8.35 35.60 0.00 0.39 154.35 2.02 1.79 2.47 3.61 4.57 57.64 63.37 14.93 14.36 7.77 0.42 12.44 22.74 22.37 8.13 5.12 34.63 1.62 1.66 6.43 49.65 42.37 14.29 14.19 3.76 0.29 604 SOGARD ET AL : LARVAL GULF MENHADEN, ATLANTIC CROAKER. AND SPOT Spot 1 6 12 J L. J I 1 1 B- I I I I Croaker E X 1- CL LLI 1 6 12 -f- "^ U —^- 1 1 1 1 _i I I I 13- j I I I Total Larvae 1 6E 12 J L I I -I J I .25 .75 .25 .75 .25 .75 .25 .75 0600 1200 1800 2400 TIME Figure 3. — Mean percentage of Atlantic croaker, spot, and total larvae collected at three discrete depths at inshore stations (18 m isobath). Error bars are standard errors. mean densities were greater in surface waters at 1200 and 1800 h and in deeper waters at 2400 h (Table 3). The high mean densities of spot in sur- face waters at 1800 h and at 12 m at 2400 h (Table 3) were related to the encounter with the previ- ously mentioned patch of larvae at Cape San Bias. When mean relative proportions were con- sidered, however, these trends were moderated (Fig. 3). Although mean densities suggested a propensity for Atlantic croaker larvae to occur in deeper waters (Table 3), this trend also weakened when relative proportions were considered (Fig. 3>. Gulf menhaden larvae in all three length groups were highly concentrated at the surface at 1200 h, but showed inconsistent patterns at other times (Fig. 4). At the offshore stations (91 and 183 m), where there was a broader scale for vertical distribution, total larvae were generally less abundant at the deepest sampling depth (70 m, Table 3), but mean relative distributions indicated only slight trends (Fig. 5). Gulf menhaden larvae at offshore sta- tions had gi'eater densities at the surface at all times, with few larvae present at 70 m (Table 3). They again occurred almost exclusively in surface samples at 1200 h (Fig. 5). (Spot and Atlantic croaker were too rare at the offshore stations to allow examination of vertical distribution.) Comparison of MOCNESS casts in thermally stratified versus isothermal water columns indi- cated that the presence of a weak thermocline did not inhibit vertical movement by any of the three target species or total fish larvae. Depth distribu- tions were similar regardless of the thermal structure of the water. In most cases where a ther- mocline occurred, it was reversed, with colder water overlying warmer water, and a tempera- ture difference of <5 C, the result of the Missis- sippi River plume. DISCUSSION High densities of gulf menhaden larvae at the Southwest Pass stations support the conclusions of Fore (1970) and Christmas and Waller (1975^) that spawning is concentrated around the Missis- sChristmas, J. Y., and R. S. Waller. 1975. Location and time of menhaden spawning in the Gulf of Mexico. Unpubl. manuscr. Gulf Coast Laboratory, Ocean Springs, MS 39564. 605 FISHERY BULLETIN: VOL. 85, NO. 3 E I H Q. UJ Q 1 63- 12:^ 1 63- ^2~ 6]- 12 3-10 mm J I u 3- 15-21 mm 13- j I I I 3- .J I 3- .25 .75 0600 .25 .75 .25 .75 1200 1800 TIME I ' .25 .75 2400 Figure 4. — Mean percentage of gulf menhaden larvae, divided into three size groups, collected at three discrete depths at inshore stations (18 m isobath). Error bars are standard errors. Menhaden E 70 I I- a. UJ a .25 .75 0600 Total Larvae 1 -t- 3U 70 + ■ + ■ 111 3- J L 3- 3- _l I L .25 .75 .25 .75 1200 1800 TIME J- 3- 3^ J I L .25 .75 2400 Figure 5. — Mean percentage of gulf menhaden and total larvae collected at three discrete depths at offshore stations (91 and 183 m isobaths). Error bars are standard errors. 606 SOGAHl) KT AL : LARVAL GULF MENHADKN. ATLANTIC CROAKER, AND SPOT sippi River Delta. In addition, Atlantic croaker larvae were rarely caught except at the inshore Southwest Pass station. High levels of nutrients (Riley 1937) and the resultant high plankton biomass in this region (Bogdanov et al. 1968) may make conditions exceptionally favorable for fish larvae. Densities of all three target species showed a clear decline from inshore to offshore waters (Table 1). Shaw et al. 11985b) found a similar pattern for gulf menhaden larvae farther west along the Louisiana coast; densities were greatest in waters between the 14 and 24 m isobaths, with a shift in concentration to very nearshore waters by the end of the spawning season. The major spawning efforts of gulf menhaden, spot, and At- lantic croaker appear to occur in a relatively nar- row band along the coast. Size-frequency distributions of gulf menhaden larvae along the Southwest Pass transect showed that offshore stations were populated with smaller larvae on two of three cruises (Fig. 2, Table 2), but off western Louisiana, Shaw et al. ( 1985a I detected no difference in the size distribu- tion of gulf menhaden from the 183 m isobath to inshore waters, except at stations immediately adjacent to shore (approximately 9 m in depth). Our observed pattern of decreasing size with dis- tance from shore could arise either by adults spawning offshore and larvae growing as they move toward estuarine nursery grounds, or from serial spawning as adults move offshore during the protracted spawning season. The latter pat- tern is corroborated by Roithmayr and Waller (1963) and Fore (1970). Only gulf menhaden showed clear evidence of a diel pattern in vertical distribution; they were concentrated almost exclusively at the surface at midday, but were more vertically dispersed at night at inshore stations. Size did not determine which larvae descended by dusk, because the ver- tical distribution was similar across all three size groups. In contrast, vertical migration of yellow- tail flounder, Limanda ferruginea, and Atlantic herring, Clupea harengus, larvae varies with size, with smaller individuals remaining closer to the water surface (Smith et al. 1978; Wood 1971). Depth distributions of northern anchovy, En- graulis mordax, and white croaker, Genyonemus lineatus. also vary with age, with older larvae concentrating in deeper waters (Brewer and Kleppel 1986). Gulf menhaden larvae >12 mm SL have de- flated swimbladders bv dav and inflated swim- bladders at night, achieved by swallowing air at the surface (Hoss and Phonlor 1984). This behav- ior, common among clupeoids (Hunter and Sanchez 1976; Uotani 1973), is thought to allow passive depth maintenance during nonfeeding hours at night (Hunter and Sanchez 1976). The observed depth distribution of gulf menhaden in- dicates that larvae must actively swim to stay at the surface during daylight hours. Apparently, the larvae slowly sink at night despite having gas in their swimbladders, and are, therefore, dis- tributed at various depths. Data from offshore stations (Fig. 4) suggests, however, that most lar- vae are able to maintain their position within the upper 30 m of the water column. The pattern of vertical distribution of gulf men- haden is opposite of that reported for numerous other species, in which larvae rise toward the sur- face at night and descend by day (e.g., Seliverstov 1974; Smith et al. 1978; Kendall and Naplin 1981; Sameoto 1982, 1984). A reversed pattern has also been observed for Gadus macrocephalus (Boehlert et al. 1985) and Ammodytes personatus (Yamashita et al. 1985). Yamashita et al. (1985) suggested that diurnal feeding requirements and nocturnal avoidance of upwardly migrating predators influence the vertical migration of Am- modytes. The behavior of Atlantic menhaden, Brevoortia tyrannus, is probably similar to gulf menhaden, as they are also reported to be more concentrated in surface waters by day than night (Thayer et al. 1983). The presence of a weak thermocline with a gra- dient of <5°C did not appear to influence the ver- tical movement offish larvae in this study. Other studies have reached conflicting conclusions. Ahlstrom (1959) and Loeb (1980) found thermal stratification with a temperature difference of 8' to 10 C very important in determining vertical distribution. Smith et al. (1978), Kendall and Naplin (1981), and Sameoto (1982), however, found that thermal gradients of 8 to 14 C did not inhibit vertical migration. The depth of the water column, the intensity of temperature change at the thermocline, and behavior of the species in question likely influence migration patterns. In relatively shallow water (Smith et al. 1978; Kendall and Naplin 1981; this study), thermal stratification appears less of a barrier than in deeper water. In this study larvae of gulf men- haden, spot, and Atlantic croaker largely re- mained within the upper 30 m, even when the water column was well-mixed to a depth of over 100 m. As we found for gulf menhaden. Brewer 607 FISHERY BULLETIN: VOL. 85, NO. 3 and Kleppel (1986) noted significant diurnal depth stratification of larvae in the absence of thermal stratification. Absolute depths may be more important than thermal layers in determin- ing vertical distribution when temperature dif- ferences are small. ACKNOWLEDGMENTS The collection of the data presented in this paper would have been impossible without the assistance of Shailer Cummings (Atlantic Oceanographic and Meteorological Laboratories, NOAA). We thank Stanley Warlen for assistance in larval identifications. Edward Houde and William Hettler provided valuable comments and suggestions on earlier versions of the manuscript. We also thank the crew of the NOAA ship Ore- gon II for assistance in sampling. This research was supported by a contract with the Ocean Assessments Division, National Ocean Service, NOAA. LITERATURE CITED Ahlstrom, E H 1959. Vertical distribution of pelagic fish eggs and larvae off California and Baja California. Fish. Bull., U.S. 60:107-146. BOGDANOV, D. V , V A SOKOLOV, AND N S KHROMOV. 1968. Regions of high biological and commercial productivity in the Gulf of Mexico and Caribbean Sea. Oceanologia 8:371-381. BOEHLERT, G. W., D. M. GADOMSKI, AND B. C. MUNDY 1985. Vertical distribution of ichthyoplankton off the Oregon coast in spring and summer months. Fish. Bull., U.S. 83:611-622. 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. Christmas, J Y , and D. J Etzold. 1977. The menhaden fishery of the Gulf of Mexico, United States: A regional management plan. Gulf Coast Res. Lab., Tech. Rep. Ser. 1, 53 p. Chrkstmas, J. Y, J T Mcbee. R. S. Waller, and F. C. Sutter III 1982. Habitat suitability index models: Gulf menhaden. U.S. Fi.sh Wildl. Serv., Biol. Serv. Prog. FWS/OBS-82/10.23. Dawson, C E 1958. A study of the biology and life history of the spot, Leiostomus xanthurus Lacepede, with special reference to South Carolina. Contrib. Bears Bluff Lab. 28, 48 p. Diaz, R J 1982. Habitat suitability index models: Juvenile Atlantic croaker. U.S. Fish. Wild. Serv., Biol. Serv. Prog. FWS/OBS-82/10.21. Fore, P. L 1970. Oceanic distribution of the eggs and larvae of the gulf menhaden. U.S. Fish. Wildl. Serv., Circ. 341, p. 11-13. GtovoNi, J. J.. D. E. Hoss, and A. J. Chester. 1983. Comparative feeding of three species of fish 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 Ortnes. 1985. An observation of episodic feeding and growth of larval Leiostomus xanthurus in the northern Gulf of Mexico. J. Plankton Res. 7:137-146. Guillory, v., J Geaghemand, and J Roussel 1983. Influence of environmental factors on gulf menhaden recruitment. La. Dep. Wild. Fish., Tech. Bull. 37, 32 p. Hoss. D E.. AND G Phonlor. 1984. Field and laboratory observations on diurnal swim bladder inflation-deflation in larvae of gulf menhaden, Brevoortia patronus. Fish. Bull., U.S. 82:513-517. HOUDE, E. D 1982. Micro- and fine-scale biology. In B. J. Rothschild and C. G. H. Rooth (editors). Fish ecology III, p. 99-125. Univ. Miami Tech. Rep 82008, Coral Gables, FL. Hunter, J R , and C Sanchez. 1976. Diel chang« s in swim bladder inflation of the larvae of the northern anchovy, Engraulis mordax. Fish. Bull., U.S. 74:847-855. Jahn, a E., and R J Lavenberg. 1986. Fine-scale distribution of nearshore, suprabenthic fish larvae. Mar. Ecol. Prog. Ser. 31:223-231. Kendall, A. W., and N. A. Naplin. 1981. Diel-depth distribution of summer ichthyoplankton in the middle Atlantic bight. Fish. Bull., U.S. 79:705- 726. LOEB, V J 1980. Patterns of spatial and species abundance within the larval fish assemblage of the North Pacific central gyre during late summer. Mar. Biol. 60:189-200. Nelson. W. R. 1969. Studies on the croaker, Micropogonias undulatus Linnaeus, and the spot, Leiostomus xanthurus Lacepede, in Mobile Bay, Alabama. J. Mar. Sci. 1:4-92. Reintjes, J. W. 1970. The gulf menhaden and our changing estuar- ies. Proc. Gulf Caribb. Fish. Inst. 22:87-90. Riley, G. A. 1937. The significance of the Mississippi River drainage for biological conditions in the northern Gulf of Mex- ico. J. Mar. Res. 1:60-74. Roithmayr, C M , AND R S Waller 1963. Seasonal occurrence of Brevoortia patronus in the northern Gulf of Mexico. Trans. Am. Fish. Soc. 92:301- 302. Sameoto, D 1982. Vertical distribution and abundance of the Peru- vian anchovy, Engraulis ringens. and sardine, Sardinops sagax. larvae during November 1977. J. Fish Biol. 21:171-185. 1984. Environmental factors influencing diurnal distri- bution of zooplankton and ichthyoplankton. J. Plank- ton Res. 6:767-792. Seliverstov. a. S 1974. Vertical migrations of larvae of the Atlanto- Scandian herring (Clupea harengus ). In J. H. S. Blaxter (editor), The early life history of fish, p. 253- 262. Springer-Verlag, N.Y. 608 SOGARD ET AL.: LARVAL GULF MENHADEN, ATLANTIC CROAKER. AND SPOT Shaw. R F . W J Wiseman. R E. Turner. L. J. Rouse, R. E. Con- drey. AND F J Kelly 1985a. Transport of larval gulf menhaden Brevoortia patronus in continental shelf waters of western Loui- siana: A hypothesis. Trans. Am. Fish. Soc. 114:452- 460. Shaw. R. F . J H Cowan Jr.. and T L Tillman 1985b. 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. Smith. P E. 1981. Fisheries on coastal pelagic schooling fish. In R. Lasker (editor), Marine fish larvae. Morphology, ecol- ogy, and relation to fisheries, p. 1-31. LIniv. Wash. Press, Seattle. Smith. W. G . J D Sibunka. and A Wells 1978. Diel movements of larval yellowtail flounder, Li- manda ferriiginea. determined from discrete depth sam- plmg. Fish. Bull . U.S. 76:167-178. Thayer. G W . D R. Colby, M. A. Kjelson, and M P Weinstein. 1983. Estimates of larval fish abundance: diurnal varia- tion and influences of sampling gear and towing speed. Trans. Am. Fish. Soc. 112:272-279. UOTANI, I 1973. Diurnal changes of gas bladder and behavior of postlarval anchovy and other related species. Bull. Jpn. Soc. Sci. Fish. 39:867-876. 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. Wood. R. J. 1971. Some observations on the vertical distributions of herring larvae. Rapp. P. -v. R6un. Cons. int. Explor. Mer 160:60-64. Yamashita. Y, D Kitagawa. and T Aoyama 1985. Diel vertical migration and feeding rhythm of the larvae of the Japanese sand-eel, Ammodytes person- atus. Bull. Jpn. Soc. Sci. Fish. 51:1-5. 609 DISTRIBUTION OF WITCH FLOUNDER, GLYPTOCEPHALUS CYNOGLOSSUS, IN THE SOUTHERN LABRADOR AND EASTERN NEWFOUNDLAND AREA AND CHANGES IN CERTAIN BIOLOGICAL PARAMETERS AFTER 20 YEARS OF EXPLOITATION W R BOWERING ABSTRACT Witch (lounder were distributed throughout the study area from Hamilton Inlet Bank to the Northern Grand Bank. The main concentrations were located in Hawke Channel, the channel around Funk Island Bank, and the north slopes of the Grand Bank. While these are believed to be the main locations of three separate stocks, there was no apparent discontinuity in the distribution among the three NAFO Divisions investigated. It is clear, however, that the stock located in NAFO Div. 3K is considerably larger than the combined stocks of NAFO Div. 2J and 3L. Stocks showed minimal variations in depth and temperature preference. Depth and temperature preferences were demon- strated for different size and age-classes offish. There were substantial reductions in the number of age gi'oups composing the stocks; this was complemented by increases in mean sizes at age for each stock although the magnitude of this increase varied from one stock to another. There was evidence of reduced size and age at sexual maturity in some instances, however, in most cases the results are difficult to explain. These changes in population dynamics are discussed in relation to changes in exploitation over the past 20 years. Prior to the early 1960's, fishing for witch floun- der, Glyptocephalus cynoglossus , in the area of southern Labrador and eastern Newfoundland was practically nonexistent. When a significant fishery began in the early 1960's, catches were taken from the accumulated virgin stock in NAFO (Northwest Atlantic Fisheries Organiza- tion) Div. 2J, 3K, and 3L (Fig. 1) primarily by large offshore otter trawlers from Canada, Poland, and the Soviet Union. Significant catches were also taken by Newfoundland gill net fisher- men in the deepwater bays of northeastern New- foundland (Bowering and Pitt 1974) (Fig. 1). Annual landings increased dramatically from < 1,000 t in 1963 to peak at nearly 24,000 t in 1973 (Fig. 2). It should be noted, however, that catch statistics prior to 1973 were based upon a formula for breaking down catches of unspecified flounder catches into species and may not be to- tally accurate. Subsequent to 1973, landings de- clined nearly as dramatically as they had risen until they stabilized at about 3.000-5.000 t annu- ally over the period 1980-85. In 1973, ICNAF (In- ternational Commission for the Northwest At- iFisheries Research Branch, Department of Fisheries and Oceans, P.O. Box 5667, St. John's, Newfoundland AlC 5X1, Canada. Manu.>icnpt accepted April 1987. FISHERY BULLETIN: VOL. 8.5. No. 3, 1987. lantic Fisheries) decided to place catch quota regulations on witch flounder in this area; for management purposes witch flounder in Div. 2J, 3K, and 3L was treated as a single unit (Fig. 1). The first TAC (total allowable catch) was placed on this stock in 1974 at a level of 22,000 t, which was subsequently reduced to 17,000 t for 1975-80, based upon an assessment by Bowering and Pitt (1974). An updated assessment by Bowering and Baird (1980) advised a TAC of 8,000 t for 1981, and this TAC level was in effect up to 1986. The TAC for 1987 was further reduced to 4,000 t. Although, for management purposes, witch flounder in Div. 2J, 3K, and 3L is considered a single population (stock), stock delineation stud- ies have shown this not to be the case. Fairbairn (1981), using biochemical systematics (electro- phoresis), distinguished two separate breeding stocks in this area, one in Div. 3K and one in Div. 3L. No data were available from Div. 2J. Bowering and Misra (1982), employing a new multivariate analysis technique on meristic data, corroborated Fairbairn's (1981) findings and also identified a separate stock in Div. 2J. The purpose of this paper is to describe the dis- tribution of witch flounder throughout this man- agement zone during recent years and to examine age, growth, and sexual maturity patterns by di- 611 FISHERY BULLETIN: VOL. 85, NO. 3 7I* 70* 69' «8" 67' 66* «5* 64* 83* 62* 6r SO" 59' M* 57* 56* 5S* 54- 53* 52" 51* 50- 49' 4»' 47- *t' 48* 44" 4S" 42* -. **" 72* 7r 70* W M* «7* S6* SS" •4* M« 62* Sr 60' 99* 5»' 57' 5«* 5S- 54" »)• 52" 51' SO* 49* 4»* 47* 46* 49* 44" 4S' 42* Figure 1. — Map of major place names mentioned in the text. vision (since they contain different stocks) dur- ing, and subsequent to, years of heaviest exploita- tion. MATERIALS AND METHODS All data were collected during research vessel 612 surveys for groundfish, carried out on an annual basis by Newfoundland-based research vessels or chartered vessels used for research purposes. All vessels used otter trawls with small mesh (12.7- 28.1 mm) nylon liners in the cod end to prevent the escape of juvenile fish. Catch records of sets which experienced enough damage to the gear BOWERING: DISTRIBUTION OF WITCH FLOUNDER Figure 2.— Nominal catches of witch flounder in NAFO Div. 2J, 3K, and 3L from 1963 to 1984. that overall catch might be affected were not in- cluded in the analyses. All fishing tows were of 30-min duration. At the end of each set, bottom temperature was measured using an expendable bathythermograph (XBT). The geographic distribution and relative abun- dance is shown by indicating, in 1/2° latitude and 1° longitude rectangles, the average numbers of fish caught per 30-min set. A preliminary exami- nation of distribution of witch flounder by year (all divisions) and season (Div. 3L only where enough seasonal data were available) showed no differences in geogi'aphic distribution; therefore, all trips were combined for the period 1977-83. These years were chosen for the distribution study since surveys during this period were based upon random distribution of sets throughout the fishing area whereas prior to this time some areas were surveyed using fixed station line transects not covering the whole area. The number of sets used in the presentation of the distribution is shown in Figure 3. Ages were determined from otoliths (Powles and Kennedy 1967). Age composition is presented for males and females separately by division. Comparisons of age composition were made for Div. 2J for the periods 1973-78 and 1979-83, Div. 3K for the periods 1970-78 and 1979-83, and Div. 3L for the periods 1968-78 and 1979-83. For Div. 2J and 3K most data were collected during the last quarter, whereas for Div. 3L most data were collected over the last half of the year. How- ever, considering the extent of the data and the very slow growth rate of witch flounder (Dower- ing 1976), slight differences in the timing of data collection are not sufficient to invalidate compari- sons among divisions. Growth (cm) was expressed in terms of log-log regressions (Log^, Length = a + b log^. Age (years)). Growth curves were computed for each of the age compositions stated above using data for each fish and not mean length at age. Differences in weight at age were then calculated between the earlier period and later period for each divi- sion by applying the length-weight relationship of Bowering and Stansbury (1984) to the observed mean length at age. The maturity rates were calculated as the length (cm) and age (years) at which 507f offish were mature (M50) as determined by probit analy- sis according to the method of Bliss (1952) as ap- plied to witch flounder in the Gulf of St. Lawrence by Bowering and Brodie (1984). The results are presented with 957^ fiducial limits. Males for Div. 2J were not included in the calculations due to too low numbers of immature fish in the samples to allow for a significant probit analysis. RESULTS Geographic Distribution and Relative Abundance Witch flounder was caught throughout the stock management zone of Div. 2J, 3K, and 3L from the northern tip of Hamilton Bank to the northern half of the Newfoundland Grand Bank (Fig. 4). Catches were insignificant in the Hamil- ton Bank area of Div. 2J with mean numbers per 30-min set generally <1.0. The general area of highest abundance in this division was to the south in the Hawke Channel. Here mean num- bers per set ranged from about 2.0 in rectangle Q16 to 11.0 in P17, the highest in Div. 2J (Fig. 4). In Div. 3K, mean numbers per set per rectangle were considerably higher than those in Div. 2J. For many rectangles the mean numbers per set were in excess of 10.0. The general areas of highest abundance occurred along the deep waters around Funk Island Bank where mean catches ranged from about 13 to 53 fish/set. The highest density occurred in rectangles P21, P22, Q22, Q23, and R24 where mean catches ranged from 43 to 53 fish/set. In Div. 3L the only signifi- cant catches occurred along the northern slope and northeast cape of the Newfoundland Grand Bank. In this area, catches ranged from about 2 to 613 FISHERY BULLETIN: VOL. 85, NO. 3 61° eO" 59° 58° 57° 56° 55« 54° 53' 52° 51' 50« 49" 48° 47 ' 46° i"'i[miniiiii[iiiii|iiiiimiiniiiii|iiiii|iiiiiiiiiii|iiiii|iii|i|iiiii |lllll lll|ll|IIMI 111111)11111 II 23 ! II r-^■ 31 : 45 ! 6 I --T I ! 41 ^ — l- Zl^'i 31 12 [\ 37 . 31 - 1- I ■J- 4 ! i9 1 46 i 29 6j^^2 j 36 1 26 y-^;-^--^ ! '\ 19 I 36 ! 36 '22 i 30 I 29 <^-^ — r ?n 1M8 I 32 : 34 2 J 24 I 28 i 35 ^f^ n 33 Pi ;/_.__ , 34 20 I 41 |S5*Z0'li 2 28 32 30 29 36 24 38 44 71 43 41 38 8- r 15 43 28 24 30 51 68 38 39 31 30 34 32 ii'ii n 4 3 2 2 I 13 S_^l Y- 34 40 I 16 54 41 42 20 22 32 ■ ■I.....I i...,.i,....i...,.i,... i,..,,i i,,...i,..,,i.. 2J X 3K 49* 15 N "1 3L 6 47 50 53 /5I 3 I 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 :|28 29 .■■■I., ,..li....iiiiii" - 30 M N P^IGURE 3.-Number of successful 30-mm bottom trawl sets per (rectangle) in NAFO Div. 2J, 3K, and 3L by Canadian research vessels from 1977 to 1983. 15 fish. The highest mean catch was 23 and oc- curred in rectangle W27. This was based upon only 1 set however (Fig. 3). On the other hand, the second highest mean catch per set was 15 fish in the adjacent rectangle V27, based upon 47 sets (Figs. 3, 4). It should be noted also that some of the highest mean numbers per set occurred due east of Cape Freels along the dividing line be- tween Div. 3K and 3L. Catches in the southern portion of Div. 3L were insignificant with most rectangles having mean catches of about 0.1 fish/ set. Distribution by Depth and Temperature Distribution by 100 m depth intervals of witch flounder is presented in Figure 5. For Div. 2J, witch flounder were caught in depths ranging from 101 to 900 m. However, only the depth range 614 BOWERING: DISTRIBUTION OF WITCH FLOUNDKR 61° 60° W° 58° 57° 56° 55° 54° 53° 52° 5l° 50° 49° 48° 47° 46° Figure 4.— Mean number of witch flounder caught per 30-min set per (rectangle) in NAFO Div. 2J. 3K, and 3L during 1977-83 by Canadian research vessels. of 301-600 m showed any significant numbers with a peak at 501-600 m after which numbers declined rapidly to near zero. In Div. 3K witch flounder were caught in depths ranging from 201 to 900 m. Only one set was conducted in depths <201 m which yielded no catch (Fig. 5). The mean catch per set increased very sharply from this depth to peak at 401-500 m. The mean catch per set of about 47 fish at that depth interval was significantly different than mean catches in all other depths (Fig. 5). Mean catches beyond 500 m dropped off dramatically to near zero beyond 600 m. In Div. 3L witch flounder were caught in depths ranging from < 101 to 800 m. No sets were conducted beyond 800 m (Fig. 5). Mean catches in depths <201 m were close to zero, based upon 864 sets. The increase in mean catch per set was sim- ilar to that of Div. 3K and peaked at 301-500 m, however, there was considerable overlap in confi- dence limits (Fig. 5). Unlike Div. 2J and 3K there 615 FISHERY BULLETIN: VOL 85, NO. 3 '5 - in 55 2 35 :25 DIVISION 2J (204) I'lOj. (121 (131 (II (13) — * • _l 1 I 1 I I I I I 1 1- (231 _J L_ _1 1 1 L_ DIVISION 3K •i* (10) (13) (4) (12) _l 1 . I I I L. 10 25 45 35 DIVISION 3L 1286) 1578 - 30g -25 2 20 - 10 -° ^°° ^^ .^ .^ .°° ^°° ^- .- •f «P ^0 \0 V° V° yO vO o x°^ ^0^ ^O^ ►O^ ^° fcO^ ^o^ ^o^ ^o^ DEPTH INTERVAL (M) O O O O O 1 ^^ TEMPERATURE (°C) Fl(;i!KK 5. — Mean numbers of witch flounder caught per 30-min set per depth and temperature interval in NAFO Div. 2J, 3K, and 3L from 1977 to 19S3 with 95'; confidence limits, (Numbers in parentheses are number of sets observed.) was not a marked decline in mean catch beyond the peak with substantial numbers of witch floun- der being caught in depths up to 700 m. Distribution of witch flounder by bottom tem- perature is also presented in Figure 5. Because of the very low catches in Div. 2J, it is difficult to identify discrete optimum temperatures. There was a slight increasing trend, however, from l.T- 2.0°C to peak at 3.r'-4.0"C and decline beyond that. In Div. 3K witch flounder were caught in bottom temperatures from <0.0° to 6.0°C. There was a marked increasing trend from the 0.1°- 1.0°C interval to peak at a temperature range of 2.r'-3.0''C beyond which the mean catch per set declined. The highest mean catch per set occurred at 5.1°-6.0°C. However, this resulted from one large catch in one of the two sets at this interval giving extremely wide 957r confidence limits. In Div. 3L an increasing trend was evidenced simi- lar to that shown for Div. 3K, however, it was 616 BOWERING: DlSTRIBl'TION OK WITCH FLOUNDER slightly less marked and peaked at 3.1°-4.0°C and then declined. The data were too few at the higher temperatures to evaluate the significance of the decline. Since depth and bottom temperature are gener- ally related, a mean catch per 30-min set by depth and temperature is presented in Figure 6. When depth and bottom temperature are combined the range of best catches for Div. 2J is about 201-600 m and 2.1°-5.0°C. For Div. 3K it is about 201- 500 m and 2.1°-4.0°C. For Div. 3L the range of best catches appears to be about 301-600 m and 2.1'-4.0 C. Despite the variations in depth and temperature distribution throughout the three divisions shown in Figures 5 and 6, the depth and temperature relationships among the divisions did not differ greatly or at least not among the locations fished (Fig. 7). Mean length (cm) and mean age (years) for each depth and temperature interval are pre- sented in Figures 8 and 9 respectively. Mean lengths and ages by depth interval for the three divisions showed similar trends in that they were highest in the shallower depths and declined to reach a low at some intermediate depth then in- creased and stabilized. For the three divisions combined there was a significant decrease in mean length from about 52 cm at < 100 m to about 43 cm at 401-500 m. Mean lengths increased be- yond this to near 47 cm and stabilized in the 600- 900 m range. The mean age decreased from near 11 years old at <100 m to about 8 years old at 401-500 m. It stabilized at about 10 years old be- yond that. For all three divisions there was a general de- clining trend in mean length and age from the lower to intermediate temperatures (Fig. 9). In all cases (except Div. 2J where catches were gener- ally low), the lowest mean size and age occurred at the 3.1°-4.0°C range. For the three divisions combined, there was a significant decline in mean length from about 50.5 cm and 10.0 years old in the 0.1°-1.0°C temperature range to about 45.5 cm and 9 years old in the 3.1-4.0 temperature range. Beyond this range the mean length and age increased again. Age Composition The age compositions of male and female witch flounder by division and time period are pre- sented in Figure 10. For Div. 2J during 1973-78, BOTTOM TEMPERATURE (•C) DEPTH INTERVAL 01 1 1 <0 lo lo 10 20 21 lo 30 3 1 to 40 4 1 to 50 51 to 60 01 II 21 31 41 51 <0 to to to to to to 10 20 30 40 50 60 01 II 21 31 41 SI <0 lo 10 M to lo to 10 20 30 40 50 60 6 1 to TO .©3 Ol ©2 ©17 501-600- O40 •5 ©7 ©4 ©3 ©.8 ©3 601-700- •a •3 0^ 0„ •, ©, ©,o#, 701-800- 0,0 • , Oe ; ©, •. 801-900- o„ • 1 0, •„ •, 901-1000- 01 V 2 J •, DIV 3K #4 DIV 3L >l000- : •n 0 ZERO CATCH Q < I FISH O I -5 FISH © II- 15 FISH 16-20 FISH 21 -30 FISH >30 FISH Figure 6— Mean number of witch flounder caught per .30-min set by depth and temperature intervals in NAFO Div from 1977 to 198.3. (Numbers in parentheses are number of sets ob.served.i 2.J. 3K, and 3L 617 4-0 30 2-0 z o > o z < a: a. 3-0 ui o a: a. uj 2-0 < a: UJ I 1-0 4-0 3-0 Z-0 1-0 FISHERY BULLETIN: VOL. 85, NO. 3 (82) (45) (11) (II) (12) -nJ" (13) OIV- _2J (305) / (204) f'"' T (13) DIV- 3K (20) (24) (12) 10-05 (2) f^Us. OIV- 3L (281) (285) \, (562) _i_ -L- _!_ _l_ _1_ _1_ 0°'' ..^°° .^^o"" ..^°^ ^'>^'^ ..fc°° ..1°° .„«>°° ..<»^ vo^"" ,xocP ^^* _\ \° «^^° ^x'P _\\° _\ ^° NO -^°' QVNO- ^0^^■' ^ON^" ^0^^^ ^O^^- ^O^^- feO^^- iO^^- %0^ <,. DEPTH INTERVAL (Ml Figure 7. — Mean temperature with 95'7r confidence limits per depth interval in NAFO Div. 2J, 3K, and 3L from 1977 to 1983 as determined from research vessel surveys. The number of sets used in the calculations are shown in brackets for each depth interval. males and females in the catches ranged from 6 to 18 years old and 4 to 21 years respectively. More than 50*^ of the males v^^ere older than age 10 and more than 509^ of the females were older than age 12. In Div. 2J during 1979-83 the males and fe- males ranged from 4 to 11 years old and 6 to 13 years old respectively. Less than 27( of the males were older than age 10 and less than 2% of the females were older than age 12. For Div. 3K during 1970-78, males and females in the catches ranged from 3 to 16 years old and 3 to 22 years old respectively. About 509( of the males were older than 8 years while more than 50% of the females were older than 10 years. In Div. 3K during 1979-83 the males and females ranged from 2 to 13 years old and 2 to 14 years old respectively. About 30'7f of the males were older than 8 years with about 309f of the females older than 10 years. In Div. 3L during 1968-78, males and females in the catches ranged from 2 to 18 years old and 4 to 23 years old respectively. About 509^^ of the males were older than 11 years and about 50% of the females were older than 12 years. During 618 BOWERING: DISTRIBUTION OF WITCH FLOUNDER 55 50 45 40 60 55 — 50 — 45 — ii 40 35 — LENGTH 1977-1983 AGE DIV 2J '^"-^ (13) Div, 2J (269) ,,9g, (275) (8) I I I 2) ♦ 69 41 *I8 59 J L (1129) DIV 3K J I L 55 50 45 40 55 50 45 40 J L J J L DIV 2J3KL COMBINED 275) J I I DIV 3K (8) (1129 (2) I4 85t 2l5i 12 II 10 9 8 7 6 - 13 12 - II 10 9 (1109) J \ I I I (27) — II 10 9 — 8 DIV 3L J I \ L J L DIV 2J3KL COMBINED (3) t5694 »3706 (1510) I I I I I I I I I I, I I I I I J L 8 ? 7 " z -O'^ oO^ .0° .O^ J=^ j5^ .O^ n.O^ .O^ csN^" ^x^° .X- ^ox- ^o^- ^oN' ^o^ ^o^ \" xN° x'P -xO xo^ tP^ .o- <^ fc^ A^ %•" q,' ^O VO »0 vO DEPTH INTERVAL (M) cP cP O^ o*^ cP rP cP cs^ cP \^ iP ^ ^ vP liP aP (jP cj^ V- v"^"" v"^° kX° vX° \*-° x'-^ x^° x^^ ^^ ^^ iP «,^ fc^ ^<^ 7.0°C (higher temperatures were not encoun- tered throughout the study area). Bowering (1976), for the Newfoundland Region as a whole during 1958-74, did not report catches of witch flounder beyond a depth of 869 m, but they were caught at bottom temperatures up to 10°C. Those catches at high temperatures were due mainly to the inclusion of catch data from the southern Grand Bank area (Div. 3N and 30), where water temperatures are highly influenced by the Gulf Stream. Markle (1975), in studying young witch flounder on the slope off Virginia, caught them down to a depth of 1,408 m and a temperature of 11.3°C. He also found that they were caught in significantly deeper and cooler water in Novem- ber compared to June. Such a comparison was not possible here. It should be pointed out, however, 625 FISHERY BULLETIN: VOL. 85, NO. 3 Table 2. — Changes In mean size at age of witch flounder from 1973-78 to 1979-83 for Division 2J; 1970-78 to 1979-83 for Division 3K; and from 1968-78 to 1979-83 for Division 3L. Only ages common to both pehods are included in the compansons. Mean weight (g) Mean weight (g) Male % Age Female Age % (yr) Early Late Difference (yr) Early Late Difference Division 2J 6 137.53 363.07 164.0 6 215.90 369.73 71.3 7 368.39 486.10 32.0 7 283.25 522.90 84.6 8 430.66 620.16 44.0 8 410.45 723.18 76.2 9 480.01 781.58 628 9 528.07 884.41 67.5 10 579.47 907.10 56.5 10 708.68 1,113.32 57.1 11 708.15 1,128.11 59.3 11 860.30 1,355.22 57.5 12 751.14 1,739.31 131.6 12 1,101.59 1,535.95 39.4 13 882.54 1,842.59 108.8 13 1,211.18 1,885.11 55.6 Division 3K 3 71.05 44.58 -37.3 3 84.24 47.46 -43.7 4 112.53 99.48 -11.6 4 102.27 99.48 -2.7 5 178.11 199.25 11.9 5 183.34 193.28 5.4 6 257.01 302.54 17.7 6 252.89 300.23 18.7 7 389.19 425.08 9.2 7 402.56 450.79 12.0 8 492.65 602.41 22.3 8 513.94 633.84 23.3 9 654.76 797.71 21.8 9 660.32 842.68 27.6 10 791.92 1,017.10 28.4 10 869.52 1,116.27 28.4 11 872.61 1,202 60 37,8 11 1,003.37 1,416.30 41.2 12 723.18 1,342,57 85,6 12 1,165.69 1,644.10 41.0 13 714.56 2,340,84 227,6 13 1,362.86 1,889.40 38.6 14 1,277.29 2,062.18 61.5 15 1,217.44 1,950.18 60.2 Division 3L 4 87 79 58.23 -33.7 4 93.95 5968 -36.5 5 93,95 119.42 27.1 5 215.90 123.66 -42.7 6 179.51 213.39 18.9 6 208.90 209.79 0.4 7 326.43 324.28 -0.7 7 303.71 316.72 4.3 8 451.56 471.19 4.3 8 429.91 466.03 84 9 572.11 617.26 7.9 9 626.00 607.17 -3.0 10 632.86 822.34 29.9 10 729.69 845,70 15.9 11 767.38 990.45 29.1 11 920.52 1,168,74 27.0 12 821.16 1,032.36 25.7 12 1,046.37 1,477.51 41.2 13 870.14 1,151.31 32.3 13 1,196.39 1,632.47 36.4 14 962.30 1,367.11 42.1 14 1,211.96 1,628,61 34.4 15 1,227.67 1,803,88 46.9 16 1,265 96 2,362,17 86.6 17 1,445.34 2,299.83 59.1 that Markle (1975) referred only to fish <5 years old, whereas this study has few fish <5 years old. Witch flounder preferred depths and tempera- tures at intermediate levels among these samples in Div. 2J, 3K, and 3L. Preferred depth in Div. 3K is more clearly defined than preferred tempera- ture; this is associated with the occurrence of smaller younger fish at intermediate values of the observed ranges. Powles and Kohler (1970) suggested that juveniles in the Gulf of St. Lawrence were in deeper water, separate from the adult, a built-in conservation mechanism for the young. S. J. Walsh (pers. comm.)^ on the other hand, in examining the distribution of juvenile versus adult witch flounder in the Gulf of St. Lawrence also found that juveniles (<30 cm) had a well-defined preferred depth range whereas the adults (>30 cm) were distributed over a much wider depth range. For demersal fish it is more common for younger fish to be found in shallower water with most of the larger fish in deeper water (e.g., See Bowering [1984] for Greenland halibut). There is some indication that young American plaice in the Newfoundland-Labrador area may also occupy some intermediate depth over the 2S. J. Walsh, Juvenile Flatfish Biologist, Department of Fish- eries and Oceans, P.O. Box 5667, St. John's, Newfoundland AlC 5X1, Canada, pers. commun. 1986. 626 BOWERINC: DISTRIBUTION OF WITCH FLOUNDER 26 28 SO —I — LENGTH (CM) 12 — I— X AGC (YRSI 6 T T- MALE -t 5K (ST»-e3) -t JK it9TO-7e) 3L (1979-83) •- 14 49 -I 5L (I9«e-TB) UAU » < 3K (1979-83) I t I 3K (1970-781 -559 -I 3L (1979-83) I I 1 iL (1968-78) fiMAL£ f£MM£ I ) 2J (1979 - 83) -»- 2J (1973-78) I I I ■»— I 3K (1979 -83) I 3K (1970-78) ( ( 2 J (1979-83) 2J (1973-78) I t ) I ( ( 3« (1979-83) I 1 \ 3K (1970-78) > < 3L (1979-83) I ( 1 JL (19*8-78) ( ( 3L (1979-83) 3L (1968-78) ' I ' _!_ _l_ _!_ _1_ _l_ _l_ 38 40 42 44 LENGTH (CM) 46 48 8 9 »C€ (YRS) 10 Figure 14. — A comparison of lengths (cm) and ages (years) at which 50';f of male and female witch flounder are mature from NAFO Div. 2J, 3K, and 3L from earlier and later time periods. range of its distribution (T. K. Pitt-^), although the size-depth distribution is not as well defined as presented here for witch flounder. Walsh (1984) concluded, on the other hand, that juvenile plaice occupy the same depth ranges as the adult plaice on the Grand Bank. It should be noted that the results presented in the present paper reflect the depth and temperature preference of young ver- sus old adult witch flounder and not juveniles versus adults as in the studies mentioned. Age and Growth The age compositions of witch flounder have changed substantially for all three NAFO Divi- sions over the study period, with a much shorter life span experienced in recent years. The impact of the reduced life span appears to be greatest on the Div. 2J stock. This may very well be the result of heavy exploitation on prespawning concentra- tions in the late 1960's when catches were double recent levels of estimated biomass (Bowering 3T. K. Pitt, Section Head, Flatfish Research, Department of Fisheries and Oceans, P.O. Box 5667, St. John's, Newfoundland AlC 5X1, Canada, pers. commun. 1986. 1985). Bowering and Brodie (1984) showed a sim- ilar reduction in the life span of witch flounder in the Gulf of St. Lawrence. However, that reduction was more dramatic because it occurred over a shorter time. In 1976, the commercial catches in the Gulf of St. Lawrence comprised fish up to 26 years old compared with a maximum age of 16 years old by 1981. Bowering and Brodie (1984) attributed the sudden change in the population age structure to the fact that almost the entire fishable stock is located in a small area during the winter months when the fishery is most intense. This is particularly true for small stocks such as that in Div. 2J, where it may not be economical to direct effort when the fish are not densely concen- trated. In Div. 3K, where the stock biomass is relatively high in comparison to that of Div. 2J and 3L, the fishery is spread more through the year, although the main effort is still directed towards prespawning concentrations. Therefore, the reduction in age groups could be over a longer time period and therefore less dramatic, which seems to be the case here. However, this argu- ment can only be true if, because of such a pattern of fishing, the fishing mortality exerted in areas such as Div. 2J is much higher than for areas such 627 FISHERY BULLETIN: VOL. 85, NO. 3 as Div. 3K. Precise information on fishing mortal- ity is unavailable. The age composition for the 1970-78 period in Div. 3K indicates that little more than 109^ of the population was older than 15 years. However, more than 40^/c of the commercial otter trawl catches in 1976 by both Canada and Poland in Div. 3K were comprised offish 15 years and older, and more than 309f were 17 years and older (Dow- ering and Pitt 1977). Thus the impact of the fish- ery was greatest on the older age groups, and this could explain the rapidity of the disappearance of the population. Accompanying the reduced age span was an overall increase in size at age for both sexes in all three divisions. The mean size at age for older fish showed a considerable increase from the earlier to later periods examined, whereas the mean size at age for the younger fish was not very different between the two periods. The substantial increase in size of older, commercially exploited fish may be the result of reduced abundance as indicated by the reduced age span. While there is no direct evidence here of density dependent growth, there have been studies published which show that it does occur. Bowering and Brodie (1984) showed that there was a systematic increase in mean size at age of witch flounder in the Gulf of St. Lawrence from 1976 to 1981 for age groups fully recruited to the commercial fishery accompanied by a significant reduction in the age span of the stock. They suggested it was likely the result of increased exploitation and subsequent reduction in abundance. They also showed that because of the increase in growth rate in particular the stock biomass remained relatively stable despite the fact that the stock abundance had been reduced. Unfortunately, estimates of stock abundance and biomass are not available for the earlier periods of this study for comparison. Bowering (1976) ruled out temperature as a major contributing factor for changes in growth of witch flounder in the Canadian Northwest At- lantic for two reasons: 1) they mainly inhabit depths that are not usually subjected to wide fluc- tuations in bottom temperature and 2) the growth rates of witch flounder in the more southerly re- gions are much slower than in the more northerly regions, the opposite of what one would expect if temperature were considered to have a signifi- cant influence on its growth rate. Bowering and Brodie (1984) suggested that given the feeding behaviour of adult witch flounder as described by Rae (1969), competition with other species is un- likely to be a significant factor in changes in growth rate, and within species competition is likely to be a more important factor. Sexual Maturity It should be pointed out that immature males are not particularly well sampled by the survey gear, and, therefore, the maturity rates may be slightly biased. However, any bias would be con- sistent for males in all divisions. On the other hand, although younger females are also not well sampled, the first occurrence of mature fish in the samples is reasonably well established. Most of the data here were collected in late summer and early autumn, several months after spawning, and one could have some concern as to the inter- pretation of immature versus fully recovered gonad condition. I do not feel, however, that witch flounder in these areas studied present signifi- cant cause for concern in this regard. The Green- land halibut, a flatfish whose gonad condition is more difficult to interpret, was sampled from northern Labrador in the autumn for visual inter- pretation of gonad condition complemented by histological analysis by Walsh and Bowering (1981). They found no significant error in the vi- sual (at sea) interpretation of gonad condition, at least for females. Due to the time of year when sampling oc- curred, a slight bias in true size and age and ma- turity may occur; however, it should be consistent throughout the division and time periods exam- ined. Therefore, comparisons should not be bi- ased. Despite such concerns, for the data pre- sented here, it would appear that there was a significant reduction in both mean length and mean age at 509^ maturity for both Div. 2J and 3L females and Div. 3L males (for mean age only) over the study period. For Div. 3K there was no significant change in either mean length or mean age at 50% maturity for either males or females over the time period. Molander (1925) found that for plaice and flounder in the Baltic, maturity appeared at a lower age but at a higher length with increased growth rate, and suggested that when growth rate was poor, the age at maturity was higher. Pitt (1975) found that for American plaice on the Newfoundland Grand Bank, the faster growing fish matured at an earlier age but at approxi- mately the same size. Bowering and Brodie ( 1 984 ) found similar results for witch flounder in the Gulf of St. Lawrence. It has been suggested. 628 BOWERING: DISTRIBUTION OF WITCH FLOUNDER therefore, that sexual maturity may be more de- pendent upon body size than on age; the hor- mones that stimulate sexual maturity may only be produced when the fish and gonads reach a certain size or stage of development. Such a con- clusion would not be reached from the results pre- sented here for Div. 2J and 3L, and there are other exceptions in the literature such as those of Fleming (1960) and Pinhorn (1966) for cod in the Newfoundland and Labrador area. It should be noted, however, that although there are statisti- cally significant reductions in mean sizes at maturity presented here, they may not be bio- logically significant; the actual relationships between size and age at sexual maturity is still unclear although some physiologists, such as Aim (1959), believe that it is closely related to initial growth rate. Since little is known about the early life history of witch flounder in the Newfoundland-Labrador area, such relationships between maturity, size, and age are difficult to evaluate conclusively. ACKNOWLEDGMENTS The assistance of C. Butt who drafted all the illustrations is gratefully acknowledged. LITERATURE CITED Alm. G 1959. Connection between maturity, size and age in fishes. Rep. Inst. Freshwater Res. Drottningholm 40, 14.5 p. BLI.S.S. E I 1952. The statistics of bioassay with special reference to the vitamins. Acad. Press. Inc., N.Y.. 184 p. B()\vp:rin(;, W R 1976. Distribution, age and growth, and sexual maturity of witch flounder iGlyptocephalus cynoglossus) in New- foundland waters. J. Fish. Res. Board Can. 33:1574- 1584. 1984 Distribution and relative abundance of the Labrador-eastern Newfoundland stock complex of Green- land halibut iRemhardlius hippoglossoides). NAFO SCR Doc. 84 61. Ser. No. N850, 14 p. 1985. Witch flounder in the Eastern Newfoundland Area, NAFO Divisions 2J3KL. CAFSAC Res. Doc. 85'38. 14 p BOWERlNdW R,.-\ND.J BAIRD 1980. Estimates of stock biomass and long-term mortality of the northern witch flounder stock (Divisions 2J + 3KL1. NAFO SCR Doc. 80 108, Ser. No. N164, 14 p. BOWKRINC. W R , AND W B Brodie 1984. Distribution of witch flounder in the northern Gulf of St. Lawrence and changes in its growth and maturity patterns. North. Am, J. of Fish. Manage. Vol. 4. No. 4A: 399-413. BOWERING. W R . AM) R K Ml.SRA 1982. Comparisons of witch flounder iGlyptncephalus cynoglossus) stocks of the Newfoundland-Labrador area, ba.sed upon a new multivariate analysis method for meristic characters. Can J. Fish. Aquat. Sci. 39:564- 570. BOWERING, W R , AND T. K PiTT 1974. An assessment of witch (Glyptocephalus cyriglos- sus ) for ICNAF Divisions 2J-3KL. ICNAF Res. Doc. 74/ 48, Ser. No. 3255, 7 p. 1977. An evaluation of the status of witch flounder (Glyp- tocephalus cynoglossus) from ICNAF Divisions 2J, 3K, and 3L. ICNAF Res. Doc. 77 VI 10, Ser. No. 5030, 9 P- BOWERING, W. R.. AND D E STANSBURY 1984. Regressions of weight on length for witch flounder, Glyptocephalus cynoglossus , of the eastern Newfound- land area. J. Northwest Atl. Fish. Sci. 5:105-106. Fairbairn. D J 1981. Which witch is which? A study of the stock struc- ture of witch flounder {Glyptocephalus cynoglossus) in the Newfoundland region. Can. J. Fish. Aquat. Sci. 38:782-794. Fleming, A M 1960. Age, growth and sexual maturity of cod iGadus morhua L.) in the Newfoundland area, 1947-1950. J. Fish. Res Board Can. 17:775-809. Markle. D G 1975. Young witch flounder, (Glyptocephalus cynoglos- sus ), on the slope off Virginia. J. Fish. Res. Board Can. 32:1447-14,50. Molander, a. R. 1925. Observations on the witch flounder and its growth. Publ. Circ. Cons. Explor. Mer 85, 15 p. Pinhorn. AT, 1966. Fishery and biology of Atlantic cod (Gadus morhua > off the southwest coast of Newfoundland. J. Fish. Res. Board Can. 26:3133-3164. PITT T K 1975. Changes in abundance and certain biological char- acteristics of Grand Bank American plaice, Hippoglos- soides platessoides. J. Fish. Res. Board Can. 32:1383- 1398. Powle.s, P M . AND V S Kennedy 1967. Age determination of Nova Scotian greysole (Glyp- tocephalus cynoglossus) from otoliths. Int. Comm. Northwest Atl. Fish. Bull. 4:91-100. Powle.s. P M , and A C Kohi.er 1970. Depth distribution of various stages of witch floun- der (Glyptocephalus cynoglossus) off Nova Scotia and in the Gulf of St. Lawrence. J. Fish. Res. Board Can. 27:2053-2062. Rae. B B 1969. The food of the witch. Dep. Agri. Fish. Scotland, Edinburgh, Scotland, Mar. Res. No. 2, 23 p. Templeman. W 1966. Marine resources of Newfoundland Bull. Fish. Res. Board Can. 154, 170 p. Walsh. S J 1984. Distribution and abundance of pre-recruit and commercial-sized American plaice on the Grand Bank. J. Northwest Atl. Fish. Sci. 3:149-158. Walsh, S J . and W R Bowering, 1981. Histological and visual observations on oogenesis and sexual maturity in Greenland halibut off northern Labrador. Northwest Atl. Fish. Organization, Sci. Council Studies 1, p. 71-75. 629 EARLY LIFE HISTORY OF SAND LANCE {AMMODYTES), WITH EVIDENCE FOR SPAWNING OF A. DUBIUS IN FORTUNE BAY, NEWFOUNDLAND E. L Dalley and G H Winters' ABSTRACT Ichthyoplankton surveys in Fortune Bay, Newfoundland, indicate that sand lance (Ammodytes sp.) larvae occur annually in Fortune Bay from February, when recently hatched yolk-sac larvae occur, until July August when, it is assumed, the larvae have grown to the size of metamorphosis and have taken up a demersal existence. Length-frequency data indicate the spawning season to extend from December to May-June, and this extended spawning season probably accounts for the consistent polymodality in length-frequency distribution of sand lance larvae from the Newfoundland area. Meristic development is shown to be complete by the time a length of 35-40 mm is reached and analyses of meristic counts indicate that the large ( >20 mm) sand lance larvae caught in Fortune Bay belonged to the offshore species Ammodytes dubius. Further, analyses of pre-anal melanophore counts and oceanographic features of the area indicate that yolk-sac larvae taken in Fortune Bay in February were also A. dubius. This is the first record of the occurrence and spawning oi A. dubius in coastal Newfoundland waters. This finding is significant in view of the current confusion regarding the appropriate taxonomy of sand lance populations in the Northwest Atlantic. Sand lance, Ammodytes sp., are widely dis- tributed in the Northwest Atlantic from Green- land south to Cape Hatteras, NC (Liem and Scott 1966). Although presently commercially unim- portant, they hold a strategic niche as a major food organism for numerous commercial fish spe- cies, and Winters ( 1981, 1983) listed sand lance as a prey of haddock, Atlantic cod, silver hake, yel- lowtail flounder, American plaice, and Atlantic salmon. They are also fed heavily upon by certain large marine mammals (Overholtz and Nicolas 1979), and it has been postulated (Winters 1983) that their importance as a prey species is en- hanced during times of low capelin abundance. Taxonomy of the Northwest Atlantic sand lance has received considerable attention (Richards et al. 1963; Richards 1982; Scott 1968, 1972; Winters 1970). Generally two species ai-e recognized in the Northwest Atlantic, i.e., Am- modytes americanus (= Ammodytes hexapterus), which is the deep-bodied, inshore form, and Am- modytes dubius, the slender-bodied, offshore form. Their taxonomy, generally, is confused by the presence of two clines in their meristic char- acter frequencies: one north to south cline, the other inshore to offshore cline, with frequent 'Science Branch, Department of Fisheries and Oceans, P.O. Box 5667, St. John's, Newfoundland AlC 5X1, Canada. Manuscript accepted April 1987. FISHERY BULLETIN: VOL. 85, NO. 3. 1987. overlap in the ranges of meristic numbers (Reay 1970). The validity of the two species is also ques- tioned due to correlations of meristic numbers with environmental conditions (Scott 1972). In the Newfoundland area, however, the two species exhibit quite distinct meristic counts (Winters 1970). For the purposes of this paper we use the species classification of Liem and Scott (1966). Winters (1970) has described the meristics and morphometries of both species from the New- foundland area and described A. dubius in the offshore and A. americanus inshore. Winters (1981, 1983) has described aspects of the biology of A. dubius from the Newfoundland Grand Banks. Little information exists on their early life history in the Newfoundland area. Dannevig (1918) provided length and distribution informa- tion for 89 specimens of sand lance captured off southern Newfoundland in surface and vertical hauls in early summer 1915. He assigned the specimens to A. tobianus (Linnaeus), a European inshore species. Frost (1938) presented informa- tion on the distribution of sand lance larvae around Newfoundland from 1931 to 1935. No size information was provided, and she assigned all the specimens from both near shore and the edge of the Grand Banks to A. americanus . In spite of its importance as a forage fish to commercially important species, no other information exists for sand lance larvae from the Newfoundland area. 631 FISHERY BULLETIN: VOL. 85, NO. 3 Information in the literature supports the hy- pothesis that the two species are allopatric. Win- ters ( 1970) found, based on adult samples, that A. dublus occurred exclusively in offshore areas and A. americanus exclusively in inshore bays around Newfoundland. Reay (1970) pointed out that A. dubius is exclusively an offshore species although Richards (1982) indicated that there are offshore components to A. americanus in the New England area. This paper presents information on sand lance in Fortune Bay, Newfoundland, as a further test of the hypothesis that the two species are allopatric. Size and distribution (seasonal and di- urnal) information on Ammodytes larvae both within and at the mouth of Fortune Bay is pre- sented as well as length-frequency information for other parts of the Newfoundland-Labrador area. The development of the definitive number of meristic characteristics is examined to identify the Fortune Bay larvae and, in conjunction with this, aspects of the developmental biology of Am- modytes sp. in the area are also described. MATERIALS AND METHODS Fortune Bay is a three still fjord located on the south coast of Newfoundland (Figs. 1, 2). Typi- cally it has a two-layered structure with rela- tively warm (1.9°C) deep water in the outer por- tion in winter and cold ( -0.25"-0.50°C) deep water in summer (de Young 1983). Annual sur- face temperatures typically range from 0.0°-1.0°C in February to 12°-16'C in August and Septem- ber. From June 1979 to February 1981, several ich- thyoplankton surveys were carried out annually in Fortune Bay. The target species for the surveys was herring, Clupea harengus , but due to low numbers of herring larvae during the first three years the study was relocated, and during 1982- 83 only one survey (July) was carried out in each year in Fortune Bay. Sand lance data examined here were collected during these surveys (1979- 83). Larvae collected m 1979-80 were taken by standard oblique plankton tows (Smith and Richardson 1977) using a 60 cm diameter bongo frame with 333 ixm mesh netting on one side of the frame and 505 p.m on the other. Tows were to a maximum depth of 200 m where possible. Nets were payed out at a speed of 0.77 m/second and retrieved at a speed of 0.38 m/second. After 1980 samples were caught in nets when both sides of the bongos were equipped with 333 \j.m netting. In February 1981, collections were made using stepped oblique tows (5 minutes at each of 200, 150, 100, 50, 20, and 0 m) with a nonclosing N.I.O. rectangular midwater trawl (RMT-8) (Baker et al. 1973). During this survey all stations were fished during 6 hours of daylight with sets on the same stations being repeated after dark. A 10- min surface tow (3/4 m conical plankton net, 333 [xm mesh) was carried out during each of the oblique tows with the RMT-8. In June 1981, sand lance larvae were collected using bongo nets during the regular survey and also from special stations to investigate diurnal distribution. During each of these special sta- tions, bongos were fished obliquely (to 200 m), and 3/4 m conical plankton nets (333 ixm) were fished for 10 minutes at the surface during day- light hours and again (at the same positions) after dark of the same day. Catches from all trips were preserved in a 5% formalin solution buffered with sodium borate. Fish larvae were later sorted, identified, counted, and measured. Total length was recorded to the nearest millimeter. A Macdonald and Pitcher (1979) mixture analysis was performed on the length-frequency distribution of the large sample from June 1979 to investigate the fit of the data to mixtures of normal distributions approximat- ing the data. Species designation was determined using meristic characters where possible (June 1979, June 1981), namely, vertebral, anal fin ray, and dorsal fin ray counts. Due to their small size, the specimens were stained for cartilage using Alcian blue and counterstained with alizarin red (Dingerkus and Uhler 1977). Counts were then done under a dissecting microscope (20-40 x mag- nification) with the aid of the camera lucida. Additional sand lance length-frequency data are presented, which have been on file at the Sci- ence Branch, Northwest Atlantic Fisheries Cen- ter in St. John's. These sand lance were collected incidentally during tows for other target species around Newfoundland using fishing and plank- ton gear (see Figures 1 and 4). RESULTS Seasonal and Diurnal Distribution Table 1 lists the cruises in Fortune Bay from 1979 to 1983. Sand lance larvae were encountered during 9 of the 17 cruises in Fortune Bay from 1979 to 1983 632 DALLEV and WINTERS EAKLY LIFE HISTORY OF SAND LANCE Hamilton "\<% ] Inlet 1% !\ Bank '--1l 51° „ - 51' 54" 53° 52° 50° 49° 48° 47= 46° 45° 51° Fkhkk 1 ^Map of Newfoundland indicating places referred to in the text. 633 FISHERY BULLETIN: VOL. 85, NO. 3 o- -3-0 O o o 634 DALLEY and WINTERS: EARLY LIFE HISTORY OF SAND LANCE (Table 1), and the spatial distribution of these catches is shown in Figure 2. Sand lance larvae have a fairly wide distribution within and at the mouth of Fortune Bay; however, the incidence of positive catches and the size of the catches were greatest in the area between Miquelon and For- tune, north to Brunette Island. The highest and most consistent catches of sand lance larvae were made in June and July, and only two specimens were caught in the August- November period. Yolk-sac larvae were taken only in surveys carried out in February, but larger larvae apparently remained in the water column until July-August at which time it is as- sumed that metamorphosis is complete, and the postlarvae take on a demersal existence. Macer (1966) gave the length of metamorphosis as 30-40 mm. Surveys in August of 3 consecutive years caught only one larva (43 mm), and surveys later than August caught only a single 82 mm post- larva in December 1980. Catches of sand lance larvae from two periods in 1981 illustrate day-night variability in catches, especially at the surface (Table 2). Dur- ing February, 44 stations were fished twice the same day, once during daylight and once after dark using stepped-oblique tows with the RMT-8 and surface tows with conical nets. Thirty-two larvae (94%) (mean length of 7.7 mm) were taken in night sets: 2 in oblique tows and 30 in surface tows. Of the two caught during daylight hours, one was in the surface tow and one in the oblique. In June, 12 stations were fished using standard oblique tows with bongo nets and horizontal sur- face tows with conical nets during daylight and darkness of the same day. Twenty-five larvae Table 1. — List of ichthyoplankton surveys carried out in Fortune Bay. Newfoundland, 1979-83, indicating gear fished, number of stations, numbers, and lengtti information of Ammodytes sp. caught. No. No. No. No. stations posi- No. No. positive larvae Length Mean and fished during tive larvae extra catches in extra range (mm) SD of total Cruise Date Gear' survey sets caught sets (extra sets) sets (all larvae) length (mm) Marinus 15 June 1979 B 71 28 302 10-43 22.9 (15.7) Shamook 51 Aug 1979 B 60 1 1 43.0 Marinus 21 Nov. -Dec. 1979 B 53 Shamook 57 Feb. 1980 B, R 25, 25 1 1 8.0 Marinus 25 June 1980 B 52 8 12 8-40 18.3 (9.5) Marinus 28 Aug 1980 B 42 Shamook 64 Sept. 1980 R 50 Shamook 67 Nov 1980 B 48 Shamook 69 Dec. 1980 B 52 1 1 228B,28S 82.0 Shamook 71 Feb. 1981 R, S 44, 44 12 34 6-12 7.7(1.0) Shamook 76 June 1981 B 52 4 8 224B,24S 8 27 12-61 38.4 (14.2) Marinus 39 Aug. 1981 B 52 222B,20R Shamook 79 Oct. 1981 B 52 Shamook 81 Dec 1981 R 35 Shamook 82 Feb. 1981 R 29 Shamook 88(2) July 1982 B 52 11 11 14-46 22.1 (10.4) Shamook 98(2) July 1983 B 52 5 5 11-22 14.8 (4.2) IB = BONGO, R - Rf\/IT-8, S - Surface net. 2Half of sets with each gear type during daylight; half at same station during darkness of same day. Table 2. — Numbers of sand lance larvae caught in oblique'' and surface tows during investigations into day-night catch variability, February-June 1981. Day Night February June February June Type of tow Oblique Surface Oblique Si jrface Oblique Surface Oblique Surface No. of sets 44 44 12 12 44 44 12 12 No. of positive catches 1 1 1 0 2 8 3 4 No. of Ammodytes larvae caught 1 1 2 0 2 30 9 16 'Oblique tow with RMT-8 in February, bongo in June ^Includes 1 set at dusk in which 6 larvae were captured. 635 FISHERY BULLETIN; VOL. 85, NO. 3 (93%) of 27 were taken in night sets: 16 at the surface and 9 in obhque tows. Two were caught in the oblique tows, and none, at the surface during daylight. Size and Length-Frequency Distributions Mean length data (Table 1) illustrate the an- nual variation in mean size during the late June- July period. In 1979, 1980, and 1982, the mean lengths of sand lance were 22.9 mm, 18.3 mm, and 22.1 mm, repectively. The 34 larvae caught in June 1981 were larger with a mean length of 38.4 mm, and the 5 caught in early July 1983 were smaller with a mean length of 14.8 mm. Length-frequency distribution of the largest samples from Fortune Bay are shown in Figure 3. The distributions from June of each year indicate a wide range of lengths, and although samples from June of 1980 and 1981 are not large, the extended range in both suggests that the distribu- tion is not unimodal and that there is more than one spawning cohort present. This extended length range may also result from delayed hatch- ing or a combination of both these processes. In June 1979 there is a distinct mode at 23 mm with others probable at 14-15 mm and another at 29 mm. The length frequency of the June 1979 sam- ple was subjected to modal analysis as described by MacDonald and Pitcher (1979). Three modes were interpreted from the length-frequency his- togram. The results of the analysis indicated nor- mal distributions with means at 13.4, 22.2, and 28.8 mm with standard deviations of 1.76, 3.02, and 4.36 mm, respectively (Table 3). The June 1981 sample of 33 larvae has a distinct mode at 45 mm, which is widely separated from a group of fish ranging in length from 12 to 17 mm. The wide range and polymodality in the length distribu- tions of the June samples is good evidence of mul- tiple spawning cohorts although differential hatching rates of the same cohort cannot be ruled out (S. Richards^). Given that sand lance larvae take 3-5 months (Reay 1970) to attain metamor- phic sizes (30-40 mm, Macer 1966), the above ob- servations suggest a spawning season extending from December through to April. In addition, the occurrence of larvae as small as 11 mm in July and 8 mm in June 1980 (Table 1) suggests that spawning may occur as late as May or June in certain years. n 10 r n JUNE 1979 , jjm N = 302 0 >]'"" TT-n n J~H w JUNE 1980 10 0 TTi rn -1 N=1 1 >- O40 =3 30 o ^20 10 0 J FEBRUARY 1981 N = 34 In JUNE 1981 N = 33 10 0 n n TTTinn n n 5 10 15 20 25 30 35 40 45 50 55 60 65 LENGTH (MILLIMETERS) Figure 3. — Length-frequency distributions from Fortune Bay sand lance samples. Table 3. — Results of the MacDonald and Pitcher (1979) method of analyzing length distribution mixtures of sand lance, June 1979, assuming a mixture of three (spawning) components. Results as- suming only one component are also shown. Percent of total Standard Compo- population in Mean length deviation nent component (SE) (SE) (SE) 13.4(0.77) 1.76(0.52) 22.2 (0.65) 3.02 (0.53) 28.8 (9.2) 4.36 (3.46) 22.4 (0.27) 4.65 (0.19) K = 3 1 8.2 (2.9) X2 = 25.32 2 78.1 (28.2) df 19 P = 0.1504 3 13.7 (27.1) K = 1 X2 = 49.55 1 100 df 25 P -- = 0.0024 This tendency for protracted length ranges is also evident in the historical samples collected around Newfoundland (Fig. 4). Samples are small, but the two from Labrador range in length from 22 to 55 mm with a break between the larger and smaller groups. The 14 fish from Colliers Bay, Conception Bay exhibit a protracted length range from 28 to 65 mm. The two samples col- lected off southern Newfoundland (Penguin Is- 2Sarah W. Richards, Little Harbor Laboratory, Inc., 69 An- drews Road, Guilford, CT 06437, pers. commun. October 1986. 636 UALLEY and WINTERS: EARLY LIFE HISTORY OK SAND LANCE (J z 3 O 10 u z 3 O 10 u z 3 O 10 a: 5< u z 3 O 10 (J z 3 O 10 LAKE MELVILLE, OCT. 1952 N=14 HAMILTON INLET BANK, OCT. 1952 N = 3J nmn COLLIERS BAY.SEPT,1965 N=14 PENGUIN ISLANDS, AUG, 1969 N = 23 OFF HERMITAGE BAY, AUG, 1969 N=34 010. 15 20 25 JO 35 40 45 50 55 60 65 LENGTH(MILUMETERS) FlGURF> 4. — Length-frequency distributions of historical sand lance samples from the Newfoundland area. lands and Hermitage Bay) 22.5 173 71-77 1 66 73.98 Dorsal fin rays 22.5 69 22-45 3.66 30 04 27.5 69 28-47 3.93 3572 32.5 16 39-65 9.27 49.31 37.5 5 44-66 9.63 61.20 42.5 10 64-68 1.27 65.50 47.5 5 65-67 1,10 65.80 52.5 4 64-67 1.26 65.25 ■42.5 19 64-68 1.17 65.53 Anal fin rays 22.5 67 25-33 1.79 30.66 27.5 72 27-35 1 20 32.17 32.5 18 31-35 1,13 32.89 37.5 6 33-36 1,17 34.17 42.5 8 33-36 1 06 34.38 47.5 4 33-35 0.96 34.25 52.5 4 32-34 0.96 33.25 >37.5 22 32-36 1.06 34.09 count myomeres because the staining process re- sulted in the fleshy parts of the body being cleared. Matarese et al. (1980) pointed out that consid- erable variation occurs in the development of meristic structures of Pacific tomcod, Microgadus proximus. because the size at which bone ossifies varies from specimen to specimen. Figure 5 illus- trates this variation in the mean counts of dorsal and anal fin rays at the smaller length classes. Species Identification Sand lance larvae collected in Fortune Bay in June 1979 had a mean vertebral count of 73.98 (SD = 1.66) (Table 4, Fig. 5). From Figure 5 and Table 4, the definitive number of dorsal fin rays varies from 64 to 68 with a mean of 65.53, and similarly the definitive number of anal fin rays varies from 32 to 36 with a mean of 34.09 (Table 80 01 70 >- < a: ^ 60 ^ 50 (/) a: § 40 30 20 36 >- < a: 32 30 . 28 o 2 26 24 78 LlJ < q:. S 76 h 74 72 70 L- DORSAL FIN RAYS I 1 ANAL FIN RAYS VERTEBRAE 20 30 40 LENGTH (MILLIMETERS) 50 60 Figure 5. — Development of definitive number of dorsal fin rays, anal fin rays, and vertebrae with increasing size of sand lance, showing mean and standard deviation for each length class. 638 DALLEY and WINTERS: EARLY LIFE HISTORY OF SAND LANCE 4). Winters (1970) gave mean vertebral counts of approximately 73.8-73.9, mean dorsal fin ray counts of 64.3-64.6, and mean anal fin ray counts of 33.2-33.6 for A. dubius from the Newfoundland Grand Banks. Comparable values for A. ameri- conus from inshore areas of Newfoundland are 66.2-68.2, 55.7-57.8, and 27.8-30.6. We conclude, therefore, that the larger larvae collected in For- tune Bay in June 1979 belong to the offshore form, A. dubius . The small larvae from February 1981 (lengths 6-12 mm) were examined for the presence of lat- eral or pre-anal melanophores. Counts of other melanophores (e.g., pectoral and subdorsal) were not possible, due either to the early stage of devel- opment of the larvae or to bleaching by the preservative. According to Richards (1982), counts of pre-anal melanophores can be used to distinguish between the larvae of A. dubius and A. americanus. From 6.0 to 8.9 mm in length, A. americanus has 0-16 pre-anal melanophores, and A. dubius has 10-21. Full or partial counts were possible for 11 of the larvae. Ten were 6-7 mm in length. Of these, two had 15 pre-anal melanophores, four had 16, one had 17, one had 18, and two had 20. One 9 mm laiva had 15 pre-anal melanophores. According to Richards' ( 1982 ) criteria, four larvae are definately A. dubius because the counts are out of range for A. americanus. The other seven are in the upper extreme of the overlap range for the two species suggesting that these too are A. dubius. Thus, not only do larger larvae of the offshore species occupy Fortune Bay, but yolk-sac larvae are also present. This suggests that spawn- ing of the offshore species occurs within the bay. Because A. dubius is present on St. Pierre Bank (Winters 1970), we have considered the possibil- ity that the yolk-sac larvae collected from For- tune Bay in February 1981 were transported from St. Pierre Bank. Smigielski et al. (1984) gave times for yolk-sac absorption from 5 to 14 days, depending on temperature. Using minimum and maximum speeds (0.05-0.20 ms M of the Labra- dor Current along the south coast of Newfound- land (Petrie and Anderson 1983), it would take yolk-sac larvae 9-36 days to be carried from St. Pierre Bank to inner stations in Fortune Bay where they were collected. De Young ( 1983), how- ever, described a seasonal cycle of water exchange for Fortune Bay in which the flow of Labrador Current water over St. Pierre sill (between Miquelon and the tip of the Burin Peninsula (Fig. 2)1 is minimal in winter months and pre- dominates in the summer. Under normal current conditions, it appears unlikely that these yolk-sac larvae were transported over St. Pierre Bank into Fortune Bay with the Labrador Current during the winter period. A persistent wind event from the south could transport larvae in the surface layers at a much faster rate than would the in- shore branch of the Labrador Current. However, the prevailing wind direction at St. Lawrence on the southern part of the Burin Peninsula for Jan- uary and February 1981 was from the west with wind from the south only 17f of the time in Janu- ary and 67f in February 1981 (Anonymous 1981). A peak wind event from the SSE on February 9 did not persist into the next day when winds were again from the west. We conclude, therefore, that it is unlikely that the yolk-sac larvae found within Fortune Bay in February 1981 were trans- ported from St. Pierre Bank. More likely, these larvae were spawned in Fortune Bay. Such spawning is consistent with evidence from hydro- graphic charts that indicate many areas of gravel and sand mixtures in Fortune Bay. According to Reay (1970), this is the preferred spawning sub- strate for sand launce. Yolk-Sac Absorption Sixty-two percent of the larvae collected in February 1981 contained yolk sacs. The mean length of those in which the absorption of yolk-sac (+ oil globule) was complete was 8.1 mm while those with absorption incomplete had a mean length of 7.4 mm. This is consistent with pub- lished records. Smigielski et al. (1984) gave yolk- sac absorption lengths of 7.2-7.41 for laboratory reared A. americanus , and Richards (1965) found that oil globule absorption was complete between 5 and 7.5 mm. DISCUSSION AND CONCLUSIONS The conclusion from the data collected during the day-night investigations is that smaller yolk- sac larvae in February and larger larvae in June are more abundant or more available to the gear at night. It is difficult to discern a particular pat- tern of diurnal and vertical migratory behavior for sand lance larvae from the literature because documentation on the subject has not been consis- tent. The observation that larger larvae in June are more abundant at night is not unusual and may be attributable to net avoidance (Norcross et al. 1961; Richards and Kendall 1973; Potter and Lough 1986). The observation that yolk-sac lar- 639 FISHERY BULLETIN: VOL. 85, NO. 3 vae are more abundant in surface night sets ap- pears unusual since Richards and Kendall (1973) found that in winter larvae 8-17 mm were more abundant in deep tows at night and surface tows during the day. Avoidance behavior does not ap- pear to develop until a size of greater than 10 mm (Norcross et al. 1961; Potter and Lough in press). Assuming the same for larvae in Fortune Bay, it is not likely that yolk-sac larvae were avoiding the gear during the day. Although it is not possi- ble to make a definite conclusion on vertical mi- gration with the relatively small number of lar- vae, the data suggest that yolk-sac larvae in February are also capable of vertical migration. Our analyses of sand lance larvae from Fortune Bay have demonstrated for the first time the oc- currence and probable spawning of the slender- bodied A. da hi us in coastal waters in Newfound- land. Previous studies by Winters (1970) have indicated the occurrence of only the deep-bodied form A. americanus in Newfoundland bays with A. dubius being found exclusively on the offshore banks. This finding is significant in light of the current confusion as to the appropriate taxonomy of the sand lance populations in the Northwest Atlantic. Both A. americanus and A. dubius ap- pear to resemble A. nutrinus which is currently considered to occur only in European waters iReay 1970), and the characteristics used to sepa- rate the two Northwest Atlantic types from each other and from A. marinus are sometimes tenu- ous particularly in southern parts of the range. In the Newfoundland area, however, A. americanus and A. dubius maintain distinct meristic counts. The occurrence and probable spawning of A. du- bius in a coastal area, formerly considered to be inhabited exclusively by A. americanus (Winters 1970), indicates sympatry. This provides evidence that the two forms are reproductively isolated and therefore separate species. This is substanti- ated by the fact that the meristics described for A. dubius larvae in Fortune Bay and those offshore (Winters 1970) are identical. We have also demonstrated that the spawning season of sand lance in Fortune Bay, Newfound- land, is protracted and probably extends from De- cember to May or June. This spawning period is much longer than in southern areas of the North- west Atlantic where the spawning season in in- shore waters is from the period December to February (Richards 1982). It is possible that this extended spawning season is also a result of the mixture of the two species in Fortune Bay; how- ever, the polymodality in the length frequency of large A. dubius (>20 mm) larvae in Fortune Bay suggests that an extended spawning season may be characteristic of this species in coastal New- foundland waters. ACKI^JOWLEDGMENTS The authors acknowledge the technical staff of the Pelagic Section of Science Branch who partic- ipated in the ichthyoplankton surveys in Fortune Bay. In particular we wish to thank B. W. Slaney for carrying out the laboratory procedures to double-strain the larvae. We also thank S. A. Akenhead, I-H. Ni and J. Carscadden for review- ing a draft of the manuscript. LITERATURE CITED Anonymous. 1981. Monthly record, meteorological observations in Eastern Canada. Atmosphere Environ. Serv., Environ. Can. 66(2(3)):l-69. Baker, A de C, M R Clarke, and M J Harris 1973. The N.I.O. combination net (RMT 1 + 8) and further developments of rectangular midwater trawls. J. Mar. Biol. Assoc. U.K. .53:167-184. Dannevig. a 1918. Biology of Atlantic waters of Canada. Canadian fish eggs and larvae. Can. Fish. Exped. 1914-1915:3-74. DE YOUNC. B 1984. Deep water exchange in Fortune Bay, Newfound- land. M.S. Thesis, unpubl., Memorial University of Newfoundland, 146 p. Dingerkus, G and L Uhler 1977. Enzyme clearing of alcian blue stained small verte- brates for demonstration of cartilage. Stain Technol. 52:229-232. Fritzsche. R a , and G D Johnson 1980. Early osteological development of white perch and striped bass with emphasis on identification of their lar- vae. Trans. Am. Fish. Soc. 109:387-406. Frost. N 1938. Some fishes of Newfoundland waters (with notes on the distribution of eggs and larvae). Nfld. Gov. Res. Bull. 4:1-16. Lelm. A H.andW B Scott 1966. Fishes of the Atlantic coast of Canada. Fish. Res. Board Can. Bull. No. 155, 485 p. Ma( Donald, P D M . and T J Pitcher 1979. Age groups from size frequency data: a versatile and efTicient method of analyzing distribution mix- tures. J. Fish. Res. Board Can. 36:987-1001. Macer C T 1966. Sand eels ( Ammodytidaei in the southwestern North Sea: their biology and fishery. Fish. Invest., U.K., Ser. II, 24(6), 55 p. Matarese. a. C. S. L Richardson, and J R Dunn 1980, Larval development of Pacific tomcod, Microgadus proximus, in the northeast Pacific Ocean with compara- tive notes on larvae of walley pollock. Thrrogra chalco- gramma. and Pacific cod, Gadus macrocepholus (Gadi- dae), Fi.sh Bull, 78:92.3-940, 640 DALLEY and WINTKRS EARLY I, IKK HISTORY OK SAND LANCE MOSEK. H G 1972. Development and geographic distribution of the rockfish. Sebastes macdonaldi (Eigenmann and Beeson, 1893), family Scorpaenidae. off southern California and Baja, California. Fish. Bull., U.S. 70:941-958. NORCROSS. J J . W H MaSSMAN. AND E R JCSEPH 1961. Investigations of inner continental shelf waters ofT lower Chesapeake Bay. Part II. Sand Launce, Aw- modytes americanus. Chesapeake Sci. 2:49-59. OVERHcn.TZ, W J , AND J R NlCOLAS. 1979. Apparent feeding by the fin whale, Bolaenoptcra physalus. and humpback whale, Megaptera novaengliae . on the American sand Xsince, Ammodytes americanus. in the Northwest Atlantic. Fish. Bull., U.S. 77:285-286. Petrie, B., AND C. Anderson 1983. Circulation on the Newfoundland Continental Shelf. Atmos. Ocean 21:207-226. Potter, D C and R G Lough In press. Vertical distribution and sampling variability of larval and juvenile sand launce, Ammodytes sp., on Nan- tucket Shoals and Georges Bank. J. Northw. Atl. Fish. Sci. ReaY. P J 1970. Synopsis of biological data on North Atlantic sand eels of the genus Ammodytes (A. tohianus , A. duhius. A. americanus, and A. mannus). FAO Fish Sym. No. 82, 28 p. Richards. S W 1965. Description of the post larvae of the sand lance lAmmodytes) from the east coast of North Amer- ica. J. Fish. Res, Board Can. 22:1313-1317. 1982. Aspects of the biology of Ammodytes americanus from the St. Lawrence River to Chesapeake Bay, 1972-75, mcluding a comparison of the Long Island Sound post- larvae with Ammodytes duhius. J. Northw. Atl. Fish. Sci. 3:93-104. Richards, S. W . and A W Kendall 1973. Distribution of sand lance. Ammodytes sp.. larvae on the continental shelf from Cape Cod to Cape Hatteras from R. V. Dolphin surveys in 1966. Fish. Bull, U.S. 71:371-386. Richards, S W , A Perlmutter. and D C McAvery 1963. A taxonomic study of the genus Ammodytes from the east coast of North America (Teleosteri: Ammodytes), Copeia 1963:358-377. Scott. J S 1968, Morphometries, distribution, growth, and maturity of offshore sand lance {Ammodytes duhius) on the Nova Scotia banks. J. Fish. Res. Board Can. 25:1775-1785. 1972. Eggs and larvae of northern sand lance {Am- modytes duhius) from the Scotian Shelf J. Fish. Res. Board Can. 29:1667-1671. Smigielski, a S., F a Halavik, L J Buckley. S M Drew, and G C Lawrence 1984. Spawning embryo development and growth of the American sand lance, Ammodytes americanus. in the laboratory. Mar. Ecol. Prog. Ser. 14:187-292. Smith. P . and S Richard.son 1977. Standard techniques for pelagic fish egg and larva surveys, FAO. Fish. Tech. Pap. 175, 100 p. Steedman, H F 1976. General and applied data on formaldehyde fixation and preservation of marine zooplankton. In H. F. Steed- man (editor), Zooplankton fixation and preservation, p. 103-154. Unesco Press, Paris. STEED.MAN, H F , AND M Omorl 1976. Cell products: calcium salts. /;; H F. Steedman (editor). Zooplankton fixation and preservation, p. 209- 221. Unesco Press, Pans. Winters, G H 1970. Meristics and morphometries of sand launce in the Newfoundland area. J. Fish. Res., Board Can. 27:2104- 2108. 1981. Growth patterns in sand lance, Ammodytes duhius. from the Grand Bank. Can. J. Fish. Aquat. Sci. 38:841- 846. 1983. Analysis of the biological and demographic parame- ters of northern sand lance, Ammodytes duhius . from the Newfoundland Grand Bank. Can. J. Fish. Aquat. Sci. 40:409-419. 641 NOTES HEART AND GILL VENTILATORY ACTIVITY IN THE LOBSTER, HOMARUS AMERICANUS, AT VARIOUS TEMPERATURES Heart rate and gill ventilatory activity have been suggested as useful measures of the physiological condition of decapod crustaceans and their re- sponse to various environmental conditions. Sev- eral authors have described altered ventilatory and heart rates in response to such variables as temperature, salinity, and dissolved oxygen (Uglow 1973; Cumberlidge and Uglow 1977; Tay- lor 1977; Hagerman and Uglow 1979). Price and Uglow (1980) also discussed the applicability of these measures in studies of pollutant stress where they described the effects of copper, cad- mium, and zinc on the heart and ventilatory rates of Crangon crangon . The mechanics of ventila- tory reversals of decapod crustaceans have also been described; for example, the reverse ventila- tory pulses (coughs) produced by the American lobster, Homarus americanus , during a muscular compression of the branchial chamber probably provide irrigation to the posterior area of the gills or help to clear detritus from gill surfaces (Wilkins and McMahon 1972; Bill and Thurberg 1985). The frequency of the lobster cough re- sponse increases after exposure to a variety of waterborne chemicals, and it has been suggested that this response might be a useful measure for detecting aquatic pollutants (Bill and Thurberg 1983, 19851. Before these heart and ventilatory measures can be employed as monitoring tools, however, baseline information should be collected on their seasonal variability under normal, un- polluted conditions, against which to interpret any stress-induced change. This study addresses the relationship between heart, gill-bailer, cough rate, and seasonal water temperature. Methods Adult American lobsters (61.4-91.2 cm cara- pace length) were trawl-collected in Long Island Sound off Milford, CT, and held in running sea- water at ambient temperature. Seawater for this building is taken from Milford Harbor, a harbor with good tidal flushing and no industrial devel- opment. The pollutant content here is very low; for example, seawater cadmium is <0.5 ppb, mer- cury <1 ppb, lead <5 ppb, and copper 2-4 ppb. The PCB levels (0.67 ppm, wet weight) in blue mus- sels, Mytilus edulis, from Milford Harbor are typ- ical of levels found in molluscs along the U.S. east coast (Farrington et al. 1983; Greig and Sen- nefelder 1985). Although no area of Long Island Sound can be considered "pristine", this area has excellent water quality for holding and rearing marine animals as evidenced by a 50-yr labora- tory history of marine invertebrate culture. The salinity range is 26-28 ppt with occasional brief low salinity episodes during extreme rains (not during this study, however) and the dissolved oxygen levels remain at or near saturation at the temperatures in this study. The lobsters were fed chopped clams, fish, or crabs daily. Heart, gill- bailer, and cough rates were monitored with 6 mm silver disc electrodes, an impedance con- verter, and an amplified polygraph recorder, fol- lowing the methods described in Bill and Thurberg (1985). Measurements were made at 2°, 6°, 10°, 14°, and 18°C over a 1-yr period. Each lobster was allowed to acclimate to temperature for at least 2 weeks befoi'e testing. Between 9 and 23 lobsters were monitored at each temperature for a 1-h period. Rates were calculated on a per- minute basis. Results and Discussion Crustacean metabolism varies with tempera- ture (Wolvekamp and Waterman 1960; Taylor et al. 1977, 1973). Aiken (1980) observed that ele- vated temperatures accelerate the metabolic processes in lobsters, although the parameters were not defined. The data presented here con- firm this increase in metabolism using three physiological parameters. Figure 1 shows the in- creasing frequency of heart and gill-bailer rates as the temperature rose from 2° to 18°C. Cough rate also increased with increasing temperatures (Fig. 2). Bill and Thurberg (1983) reported a cough rate of 0.4 coughs/minute at 10°C in this species, a rate similar to that reported in this study at 10°C (0.32 coughs/minute). The data re- ported here present a full seasonal profile of three important metabolic measures. They provide a FISHERY BULLETIN: VOL. 8.5. NO. 3, 1987. 643 TEMPERATURE VS. HEART AfJD GILL BAILER TEMPERATURE (t) FiC.URK 1. — Temperature ("C) versus heart ('■ and gill bailer (D) rate (counts per minute) of the American lobster, Homarus amcncaniis, at 2 , 6\ 10°, 14°, and 18 C. Each point is the mean value of 9-23 lobsters and the vertical line is ± 1 standard error. TEMPERATURE VS. COUGH RATE 1 09 - 08 - 0 7 1 P 0 6 1 ) 0 b 0 4 0 3 1 1 0 2 0 1 * + TEMPERATURE CC) Figure 2.— Temperature (°Cl versus cough (□) rate (counts per minute) of the American lobster, Homariis americanus . at 2°, 6°, 10 . 14\ and 18'C. Each point is the mean value of 9-23 lobsters and the vertical line is ± 1 standard error. comprehensive and easy-to-interpret baseline against which to compare similar measures made on lobsters from areas suspected of pollutant im- pact. Literature Cited 91-163. Acad. Press, N.Y. Bill. R G . and F P Thurberg 1983. Effects of pollutants on heart and gill bailer activities of the lobster, Homarus americanus. ICES Mar. Environ. Qual. Comm. CM. 1983/E58. 1985. Coughing; A new description of ventilatory reversals produced by the lobster, Homarus americanus. Comp. Biochem. Physiol. 80A(3):333-336. CUMBERLIDGE. N , AND R. F UOLOW 1977. Heart and scaphognathite activity in the shore crab, Carcinus maenas (L). J. Exp. Mar. Biol. Ecol. 28:87-107. Farrington. J W , E D Goldberg, R W Risebrough. J H. Martin, and V T Bowen 1983. U.S. "Mussel Watch" 1976-1978: An overview of the trace-metal, DDE, PCB, hydrocarbon, and artificial radionuclide data. Environ. Sci. Technol. 17:490-496. Grieg, R A , and G Sennefelder 1985. Metals and PCB concentrations in mussels from Long Island Sound. Bull. Environ. Contam. Toxicol. 35:331-334. Hagerman, L . and R F Uglow 1979. Heart and scaphognathite activity m the shrimp Palaemon adspersus (Rathke). Ophelia 18:89-96. Price. R J K . and R F Uglow 1980. Cardiac and ventilatory responses of Crangon crangon to cadmium, copper and zinc. Helgol. Meeresunter. 33:59-67. Taylor. A C 1977. The respiratory responses of Carcinus maenas (L.) to changes in environmental salinity. J. Exp. Mar. Biol. Ecol. 29:197-210. Taylor, E. W., P. J. Butler, and A Al-Wassia, 1977. Some responses of the shore crab, Carcinus maenas (L.), to progressive hypoxia at different accHmation tem- peratures and salinities. J. Comp. Physiol. 122:391-402. Taylor, E W , P, J Butler, and P J Sherlock 1973. The respiratory and cardiovascular changes associ- ated with the emersion response of Carcinus maenas (L.) during environmental hypoxia, at three different temper- atures, J. Comp. Physiol. 86:95-115. Uglow, R F 1973. Some effects of acute oxygen changes on the heart and scaphognathite activity in some portunid crabs. Neth. J. Sea Res. 7:447-454. WiLKlNS, J L , AND B R McMaHON 1972. Aspects of branchial irrigation in the lobster Homarus americanus . I. Functional analysis of scaphognathite beat, water pressures and currents. J. Exp Biol. 56:469-479. WOLVEKAMP. H P. ANDT H WATERMAN 1960. Respiration. In T. H. Waterman (editor), The physiology of Crustacea, Vol. 1 (Metabolism and growth), p. 35-100. Acad. Press, N.Y. Renee Mercaldo-Allen Frederick P Thurberg Northeast Fisheries Center Milford Laboratory National Marine Fisheries Service, NOAA 212 Rogers Avenue Milford. CT 06460 AlKKN, D E 1980. Molting and growth. In J. S. Cobb and B. F. Philips (editors). The biology and management of lobsters, Vol 1, (Physiology and behavior), p. 644 A METHOD OF SIMULTANFX)USLY TAGGING lARGE OCEANIC FISH AND INJECTING THEM WITH TETRACTCLINE A simple method of marking large oceanic fish such as yellowfin tuna, Thunnus albacares, and wahoo, Acanthocybium solandri, for age determi- nation studies is described. The method, devel- oped for tagging and injection with tetracycline of yellowfm tuna >45 kg from the deck of sport fish- ing vessels, is easily adaptable to other species, including billfish and possibly marine mammals. The use of calcium-specific markers, such as oxytetracycline (OTC), to validate the temporal significance of natural marks in hard parts has become increasingly widespread. Validation is recognized as a basic requirement of age and growth studies (Beamish and McFarlane 1983). OTC is usually administered orally, intraperi- toneally, or intramuscularly (Weber and Ridg- way 1967; Wild and Foreman 1980; Campana and Neilson 1982). Boating and restraining while OTC is injected causes stress and trauma to the fish and may result in injury to the tagger when large, powerful pelagic fish are tagged. Neverthe- less, biologists from the Inter-American Tropical Tuna Commission (lATTC) have successfully tagged and injected medium-sized (up to 36 kg) yellowfin tuna (Anonymous 1982) where, using multiple poles (two pole method, described in Godsil 1938), the crew pulled the fish onto the padded aft deck (Bayliff and Holland 1986) of a dedicated tuna baitboat. Although this method would probably suffice when tagging even larger fish, dedicated vessels are costly and there is no guarantee of locating adequate-sized fish during the charter period. Opportunistic tagging by the crews of long-range sportfishing boats, which fre- quently capture large tuna but lack gear to han- dle live fish on deck, was an attractive prospect. Methods and Materials A device used for administering drugs to zoo animals (Extend-0-Jector\ model A, Kay Re- search Products, Hyde Park Bank Bldg., Suite 503, 1525 East 53rd St., Chicago, IL 60605 USA) was modified (Fig. 1) by adding a stainless steel dart tag applicator held at an appropriate dis- tance with 13 mm thick PVC sheet press-fitted to the distal end of the injector head. The applicator 'Reference to trade names doe.s not imply endorsement by the National Marine Fishers Service, NOAA. was then fastened to the grooved base of the injec- tor head with a hose clamp, stabilizing it during use. Other types of applicators, such as those used for metal anchor tags (Bayliff and Holland 1986) can be easily substituted. Depending on the tag type, a rubber band may be used to hold the tag in place during application. The device utilizes either a 3 to 5 cc disposable syringe and a 2-in (51 mm) needle. For tunas, a 17 gauge needle provided the best combination of sufficiently high delivery rate and minimum puncture diameter. As the decks of most long-range sportfishing boats are quite far off the water (2 m), the device was bolted inside a 25 mm ID telescoping tubular aluminum pole, the type normally used with swimming pool cleaning equipment. The length may then be adjusted to suit individual situa- tions. When a large tuna was captured by an angler, the tag and injection was administered below the second dorsal fin while the fish was still in the water. Through trial and error, application was found most efficient when the device is con- tinuously pushed toward the fish's body after ini- tial insertion of the applicator needles. Best re- sults are obtained when the device is kept as near to perpendicular to the body of the fish as possi- ble, preventing the needles from bending, damag- ing the fish, or both. The fish is released by removing the hook by jerking the bend of the hook with a gaff forwards while pulling the line back- wards, or cutting the leader as close to the hook as possible. Under certain conditions, free- swimming fish found at the surface may be tagged and injected. A previous experiment (unpubl. data, Inter- American Tropical Tuna Commission, La Jolla, CA) determined that doses as low as 10 mg/kg body weight formed readable marks in the verte- brae of mackerel. Scomber japonicus. Wild and Foreman (1980) and Foreman (1987) determined that a dosage of 27 mg OTC/kg body weight formed a brilliant mark in the otoliths of yel- lowfin and skipjack, Katsuwonus pelamis , tunas and in otoliths and vertebrae of bluefin tuna, Thunnus thynnus . For large fish (near 45 kg), the volume required using standard veterinary in- jectable (100 mg/mL) OTC would be unmanage- able; a more concentrated form (200 mg/mL; Pfizer Liquamycin LA-200) was substituted. Be- cause the marks formed in smaller fish were suf- ficiently bright, the dose was reduced by half, to about 13.5 mg/kg. FISHERY BULLETIN: VOL. 8.5, No. 3. 1987. 645 SYRINGE NEEDLE INJECTOR HEAD (3cc SYRINGE FITS INSIDE) HOSE CLAMP PLUNGER FOR SYRINGE IS INSIDE THIS SECTION RUBBER BAND 1" (2.5 cm) ALUMINUM TUBING T-Jl. I I I I ^ TAG APPLICATOR DART TAG (INSIDE APPLICATOR) SPRING-LOADED INSIDE FiCURE 1. — A tag and injection device for large pelagic fish. Results and Discussion On the night of 27-28 January 1986, the crew and passengers of a 34 m long-range sportfishing boat. Royal Polaris , tagged, injected, and released 36 yellowfin tuna, all estimated to be >45 kg. Six of these fish were recaptured from 14 to 83 days later, indicating that the tagged fish were active and feeding. Differences in return rates between this method and the padded deck method (Table 1 ) are possibly due to such factors as differences in size and age of the fish tagged, fishing effort in the tagging area, or the amount of stress caused by different fishing methods, rather than some char- acteristic of the pole method. The otoliths from recaptured fish displayed the yellow-green fluorescent mark when viewed under ultraviolet light, but the marks appeared much fainter than those on otoliths returned from the program which used a padded deck. Since the dosage each fish received was monitored by the amount (if any) left in the syringe after applica- tion, failure of the device to deliver the full amount of OTC was ruled out. Similarly, the needle size was nearly the same for both treat- ments, and pore seepage is assumed equal. There were, however, differences in the type of OTC used. Liquamycin LA-200, used with the pole de- vice, was found to contain a slow-release agent (2-pyrrolidone) which extends the antibiotic ef- fect over time. Evidently the agent also slows de- Table 1. — Comparison of tag return rates from the padded deck method (unpubl. data, Inter-American Tropical Tuna Commission, La Jolla, CA) and pole injection device method for yellowfin 100 cm (20.5 kg) at release. Releases Returns Percent Padded deck Pole device 49 36 16 6 32.7 16.7 position of the fluorophors such that their concen- tration in the area of osteogenesis and hence the brilliance of the mark is diminished. I recommend that an OTC solution without slow-release agents be used, e.g., Anchor Oxy-Tet 100, as in previous experiments. Acknowledgments I appreciate the assistance of J. Allen and his staff at the San Diego Wild Animal Park, and T. Dunn for input into the design of the injection equipment, and the crew and passengers of the Royal Polaris, especially J. Heyn, W. Lang, F. LoPreste, and C. Miller, for their assistance and dedication to the project. Literature Cited Anonymous 1982. Annual Report of the Inter-American Tropical Tuna Commission, 1981. Inter-Am. Trop. Tuna Comm., p. 27. Bayliff. W H . AND K N. Holland 1986. Materials and methods for tagging tunas and bill- fishes, recovering the tags, and handling the recapture data. FAG Fish. Tech. Pap., 279; 36 p. Beamlsh. R J . AND G A McFarlane 1983. The forgotten requirement for age validation in fisheries biology. Trans. Am. Fish. Soc. 112:735-743. 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. Foreman, T J In press. An assessment of age determination techniques from hard parts of northern bluefin tuna iThunnus thyn- niis L.) from the Pacific Ocean. Inter-Am. Trop. Tuna Comm., Bull., vol. 19. GODSIL, H C 1938. The high seas tuna fishery of California. Calif. Div. Fish Game, Fish. Bull. 51, p. 3-41. Weber. D . and G J Ridgway. 1967. Marking Pacific salmon with tetracycline antibi- otics. J. Fish. Res. Board Can. 24:849-865. 646 Wild. A., and T. J. Foreman. 1980. The relationship between otolith increments and time for yellowfin and skipjack tuna marked with tetra- cycline. Inter-Amer. 'Prop. Tuna Comm., Bull., 17:509- 560. Terry Foreman Inter-American Tropical Tuna Commission Scripps Institution of Oceanography La Jolla. CA 92093 SECOND RECORD OF THE KAWAKAWA, EUTHYNMS AFFIMS. FROM THE EASTERN PACIFIC OCEAN Although the kawakawa, Euthynnus affinis (Cantor 1849), is widely distributed throughout the warm waters of the Indo-West Pacific (Yoshida 1979), it is replaced by the black skip- jack, E. lineatus Kishinouye, in the eastern Pacific. There is only one previous record of E. affinis in the eastern Pacific. That specimen, 361 mm fork length (FL), was reported from Los An- geles Harbor, CA, in 1952 (Fitch 1953). The sec- ond documented occurrence of £■. affinis from the waters of the eastern Pacific is recorded in this note. The specimen, E. affinis , 920 mm FL and 13.15 kg, was caught by Ronald Nakamura using hook and line from the long-range San Diego-based sport-fishing boat. Royal Polaris , on 17 December 1986, off Clarion Island (lat. 18°22'N, long. 114°44'W) in the Revillagigedo group. The speci- men has been deposited in the Scripps Institution of Oceanography fish collection (SIO 87-70). The morphometric and meristic characters for the specimen are given in Table. 1. The measure- ments were taken according to the methods of Godsil and Byers (1944) and Gibbs and Collette (1967). The external characters of this specimen agree with Godsil's (1954a) description of the spe- cies. The wavy oblique markings on each side of the dorsal surface, no dip in the lateral line below the second dorsal fin, and the several black to gray spots scattered over a relatively wide area between the pectoral and pelvic fins are charac- teristic of most specimens of this species. Further- more, the morphometries for this specimen are within the ranges for those body proportions re- ported by Godsil (1954b) and are closer to the morphometries for E. affinis from Hawaii, rather than from Japan. The internal characters also appear to agree with Godsil's (1954a) description of the species. High-quality radiographs produced by computer- assisted tomography (C.A.T.) scanning equip- ment were utilized for examing skeletal charac- ters. The vertebral count is 20 + 19 = 39, and the radiographs showed no bony protuberances on any of the caudal vertebrae. However, no vomer- ine teeth were present. Although there was no indication of their previous presence, their ab- sence could be explained by wear in this pre- sumably old specimen. Nevertheless, the primary characters distinctive of £". affinis, 39 vertebrae, the total gill raker count of 31, and the absence of bony protuberances on the caudal vertebrae, leave no doubt on the identity of this specimen. The occurrence of E. affinis, as well as the first documented occurrence of this specimen in the eastern Pacific, should be considered extremely rare events. No specimens of E. affinis were noted, during 1980-82 while personally examin- ing a few thousand specimens of E. lineatus landed by commercial tuna vessels operating in the eastern Pacific. One of the remarkable fea- Table 1. — Summary of morphometric and meristic data. The measurements are in millimeters. Character Measurements (mm) Fork length 920 Head length 263 Snout-first dorsal 301 Snout-second dorsal 552 Snout-anal 590 Snout-ventral 291 Max. body depth 232 Max. body width 156 First dorsal-ventral 225 First dorsal-anal 385 Ventral-vent 310 Base first dorsal 238 Base second dorsal 72 Base anal 61 Pectoral length 138 Anal length 65 Diameter of iris 29 Maxilla length 97 Snout-posterior of eye 106 Counts Dorsal spines 15 Second dorsal rays 13 Dorsal finlets 8 Anal rays 14 Anal finlets 7 Pectoral rays 26 Gill rakers 8 + 1 + 22 = 31 FISHERY BULLETIN; VOL. 85. NO. 3, 1987. 647 tures ofE. affinis is its extremely large size, par- ticularly its weight of 13.15 kg, as this specimen represents the heaviest E. affinis documented. The previous documented record of maximum size for E. affinis was 11.79 kg based upon a specimen captured in Merimbula, NSW, Australia in 1980 (Anonymous 1986). It is interesting that the maximum size records established for the black skipjack, E. lineatus, and the yellowfin tuna, Thunnus albacares , are based upon sport-caught specimens from the Re- villagigedo group of islands. The fact that many species tend to be longer lived and reach maxi- mum sizes in the northern latitudinal ranges of their distributions, apparently pertains to the aforementioned species of tunas, as well. In the case of this record specimen of E. affinis, al- though found outside its normal geographical dis- tribution, the maximum size was attained in this same region of the Pacific Ocean. Acknowledgments I thank William H. Bayliff, Bruce B. Collette, and Witold L. Klawe for their reviews of this note. Literature Cited Anonymous. 1986. World record game fishes. Int. Game Fish Assoc. FlTfH, J. E. 1953. Extensions to known geographical distributions of some marine fishes on the Pacific coast. Calif. Fish Game .39:.539-5.52. GiBBS, R. H., JR , AND B. B. Collette. 1967. Comparative anatomy and systematics of the tunas, genus Thunnus. Fish. Bull., U.S. 66:6.5-130. G0D.SIL, H. C. 1954a. A descriptive study of certain tuna-like fi.shes. Calif Dep. Fish Game, Fish Bull. 97, 185 p. 1954b. A comparison of Japanese and Hawaiian specimens of the black skipjack, Euthynnus yaita . Calif Fish Game 40:411-413. GoD.siL. H. C, AND R. D. Byers. 1944. A systematic study of the Pacific tunas Calif Dep. Fish Game, Fish Bull. 60, 131 p. Yo.siiiDA, H. O. 1979. Synopsis of biological data on tunas of the genus Euthynnus. U.S. Dep. Commer., NOAA Tech. Rep. NMFS Circ. 429, 57 p. Kurt M. Schaefer Inter-American Tropical Tuna Commission CO Scripps Institution of Oceanography La JoUa, CA 92093 CONTRIBUTION TO THE LIFE HISTORY AND REPRODUCTIVE BIOLOGY OF GAG, MYCTEROPERCA MICROLEPIS (SERRANIDAE), IN THE SOUTH ATLANTIC BIGHT^ The gag, Myceteroperca microlepis, is a demersal serranid found along the southeastern coast of the United States and in the Gulf of Mexico (Smith 1971; Fischer 1978). Throughout its range the gag is of both commercial and recreational impor- tance. Because of its relatively slow growth rate (Manooch and Haimovici 1978) and desirability, overfishing is of wide concern. The gag is a protogynous hermaphrodite, and McErlean and Smith (1964) suggested that sex- ual transformation occurs during the 10th or 11th year. Spawning occurs from January to March off the west coast of Florida (McErlean 1963), and the maximum reported age is 13 years in both the Gulf of Mexico (McErlean 1963) and the South Atlantic Bight (SAB) (Manooch and Haimovici 1978). Microscopic examination of the gonads is necessary for definite sexual identification, but gonad morphology has not been specifically de- scribed. The purpose of this study is to provide new information on the age, growth, and repro- ductive biology of this important species, includ- ing a description of the morphology of gag ovaries and testes. Methods Most samples were obtained from the commer- cial hook and line fishery, and others were col- lected on research cruises aboard the RV Dol- phin , RV Oregon , and RV Lady Lisa from 1976 to 1982. Specimens were measured (total and stand- ard lengths), weighed, and sagittae removed from the otic capsule through the branchial chamber. Otoliths were stored dry and later viewed in a dish of cedar wood oil with reflected light over a dark background using a binocular microscope. Since opaque bands in larger otoliths were thin and often too crowded near the edge to permit accurate counting, cross sections (approximately 0.5 mm thick) were made on the dorsoventral plane of the otoliths through the center with a diamond dicing wheel mounted on an ISOMET' low speed saw. Sectioned otoliths were viewed in 'Contribution No. 226 of the South Carolina Marine Re- sources Center, South Carolina Wildlife and Marine Resources Department, Charle.ston, SC 29412. '■^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. 648 FISHERY BULLETIN: VOL 8.5, NO 3, 1987. the same manner as whole otoliths. If two read- ings by a single observer did not agree, otoliths were deleted from analyses. Additional verifica- tion of counts was then obtained from a second observer who read 200 otoliths (359r ) selected randomly. Because monthly samples did not contain simi- lar porportions of large and small fish, calcula- tions of monthly mean marginal increments of sagittae were biased. For instance, if primarily small fish are sampled in one month followed by mostly large fish in the next month, the mean marginal increment will decrease regardless of the time of year. To alleviate this bias, marginal increments were standardized by converting each measurement to a proportion of the maximum recorded for that age group. Thus, a measurement of 2.5 ocular units in an age group for which the maximum is 10.0 becomes equivalent to a mea- surement of 0.5 ocular units in an older group for which the maximum is 2.0. Sex and reproductive conditions were deter- mined from histological sections of gonads, which were preserved in lO'yf formalin, and later pro- cessed through an Autotechnicon Duo Model 2A automatic tissue processor, then embedded in paraffin, and sectioned with a rotary microtome at approximately 7 |xm (Humason 1972). Tissues were then stained with Harris' hematoxylin and counterstained with eosin-Y. Sexes were identi- fied as male, female, and hermaphroditic female or "transitional" (gonad primarily ovarian, with some traces of active testicular tissue present). Maturity was described following the synopses listed in Waltz et al. (1982). Terminology used in histological descriptions follow Hyder (1969), Combs (1969), and Wallace and Selman (1981). Results A total of 1,039 gag ranging in total length (TL) from 153 to 1,150 mm was examined for life his- tory information. Of the 652 otoliths on which age determinations were attempted, 87'/ showed dis- cernible rings verified by two readings. No otoliths were deleted from analyses because of disagreement between primary and secondary readers. Marginal increment measurements from the outer edge of the last opaque band to the dor- sal margin of whole sagittae indicate that these bands are laid down in late spring to midsummer. Bands are apparently laid down earlier and over a longer time period (May-August) in ages age IX on which marginal increments were measured are small in relation to those of younger gag, it appears that ring formation is concentrated in August (Fig. 1). Twenty-two age groups were identified (Table 1). The gag ovary is a hollow, bilobed organ sus- pended in the posterior region of the body cavity from the swimbladder by mesenteries. Blood ves- sels and nerves enter the gonad at the anterior point of each lobe's suspension and course medi- ally to the mesenteries along the dorsomedial sur- face of each lob. The lobes fuse posteriorly, their N^ D LU CC DC < O Q Z z — < _J I- < CO z z o < DC m < .70 -I .50- .30- .10- o < age VIII • > age IX 6 3 >\27 66 5 7 MONTH ~l 1 11 Figure l. — Mean standardized marginal increments by month for gag sage VIII and ^age IX, and sizes of monthly samples. 649 Table 1 — Observed mean lengths (mm), weights (kg), and sample size by age for Mycleroperca microlepis . Total Standard length length Weight Age Number (mm) (mm) (kg) 0 14 187 153 0087 1 17 347 293 0605 2 18 466 383 1,263 3 26 575 470 2.104 4 49 677 555 3881 5 80 767 641 5.558 6 102 824 686 6040 7 91 865 729 7.612 8 51 895 751 8,454 9 21 936 789 9810 10 19 959 807 10.491 11 23 993 826 1 1 .399 12 7 1,004 841 13.327 13 4 1.048 884 14.950 14 2 1.066 905 13 000 15 2 1.096 927 14.748 16 9 1.064 905 14,495 17 10 1.076 904 14,767 18 10 1,068 887 14310 19 5 1,087 924 14,074 20 5 1,071 896 15 150 21 1 1,125 950 15 400 22 1 1,124 946 — Figure 2. — Schematic representation of a cross-sectioned gag ovary (diameter = 3 cm). HS = halar stroma, L = lumen, OL = ovarian lamellae, OW = ovarian wall, T = "Typhlosole- type" invagination of the dorsal wall. lumena forming a common oviduct. The lumena are incompletely lined with folded germinal ep- ithelium (ovarian lamellae) within which oocytes develop and mature. The ventral regions of the lumena remain void of lamellae, and these alamellar areas are contiguous with the alamel- lar oviduct. In addition to the dorsal and lateral walls of the lumena showing lamellar develop- ment, there is a "typhlosole-type" continuation of the dorsal gonad wall projecting into each lumen. This projection of connective tissues into the cen- ter of the lumen apparently allows additional sur- face area for attachment of ovarian lamellae (Fig. 2). During the sexual transition phase, testicular growth fills the existing ovarian lamellae, dis- placing and possibly dislodging the already de- generating oocytes. Transitional gonads were rarely found, and there were no cases of simulta- neous development of gonad tissues. Male gonads retain the somatic morphology of the ovary. Tes- ticular tissue is arranged in "false lamellae", pri- marily suspended from the "typhlosole-type" structure. Sperm sinuses form peripherally in what was previously the ovarian wall and con- tinue posteriorly becoming the vas deferens within the oviduct wall. The vestigial ovarian lu- mena and oviduct remained in all testes exam- ined. All testes also possessed many residual oocytes, some as large as 500 ixm in diameter. Females made up 849f of the gag which were sexed. Examining the percentage in each age class, we found that 28% of age III, 5V7c of age IV, and all older female gag had mature ovaries. Im- mature gag ranged from 290 to 680 mm TL, whereas the smallest mature female was 600 mm TL. Male gag accounted for 15% of the animals sexed and were found in ages V through XX (no sex available for age XXI and XXII fish). No males were found smaller than 790 mm TL (Fig. 3) and no juvenile males were found. Gag with transitional gonads made up 1.25% of all the groupers sexed and occurred in ages V through XI. The size range for fish undergoing sex succes- sion was from 750 to 950 mm TL (Fig. 3). The gag spawns once a year in late winter-early spring. Analysis of the relative abundance of de- veloping, ripe, and postspawned gonads indicated that peak spawning activity was reached in late March and early April (Fig. 4) in the SAB. Discussion Use of whole sagittae in aging gag has been validated (McErlean 1963), and Matheson et al. (1986) successfully validated the use of sectioned sagittae to age the congeneric scamp, Myctero- perca phenax . These studies together with the present data provide good evidence that sagittal 650 100 • females « transitional <700 1 — r BOO 900 1000 1100 1200 TOTAL LENGTH (mm) Figure 3. — Percent female gag by length class, and occurrence of transitional gag. rings detected in the present study are annual in nature. While the importance of adequate valida- tion has been well documented (Beamish and McFarlane 1983), it has become increasingly clear that annulus formation in many species can take place over an extended period, making it difficult to pinpoint this event in time. Peak an- nulus formation covered a 3-mo period for the groupers Epinephelus drummondhayi (speckled hind) and E. niveatus (snowy grouper) (Matheson and Huntsman 1984). Thus, the May-August pe- riod of peak ring formation for younger gag in the present study is not unusually long. Reasons for differences between ages IX in tim- ing of peak annulus formation are not apparent. However, that the differences are at least par- tially based on sex is probable since the younger fish are predominately female while approxi- mately 60% of the older group are male. A com- parison strictly between sexes was not possible since samples of males available for measure- ment were obtained in only 5 months. The use of sectioned sagittae greatly enhanced clarity among the higher age groups and allowed for greater distinction between rings in compari- son to whole otoliths. Beamish (1979) found that sectioned otoliths of the Pacific hake, Merluccius productus , gave a more accurate account of age, especially when thick otoliths with poorly defined annuli were encountered. This appears to be true for the gag, as well. Twenty-two age groups were distinguished in the present study, similar to the 21 age groups reported for the scamp (Matheson et al. 1986), compared with previous reports of 13 age groups for the gag (McErlean 1963; Manooch and Haimovici 1978). The nine groups not de- tected previously do not represent just an in- crease in the percentage of readable otoliths (present study: 87%; McErlean 1963: 87%; Manooch and Haimovici 1978: 79%). Rather, it appears that additional annuli are present in the otoliths of the larger size classes that were not detected in previous studies using whole otoliths. For instance, the oldest fish collected by Manooch and Haimovici (1978) was age XIII and 1,201 mm TL, longer than any gag aged in the present study and 77 mm longer than the age XXII fish (Table 1). Moe (1969) described the reproductive biology of the red grouper, Epinephelus morio, and many aspects of the development and sex succession schedules are similar to those found in the gag. I00-, < 80- UJ u. UJ 60- tr. < z UJ u UJ a. 40- 20- GONAD CONDITION O Oeveloping - running ripe • Post -spawned "T" M n I M J MONTH Figure 4. — Maturity stages of female gag by month of capture, illustrating the late winter-early spring spawning season. 651 The actual morphology of the gonad, however, differs from that described for red grouper. Moe (1969) cited Smith's (1965) description of an E. fulvus ovary, but Smith did not mention the "typhlosole-type" structure from which ovarian lamellae are suspended that was found in gag gonads in the present study. This structure is also found in M. phenax, M. inter stitialis , E. adscen- cionis. E. drummondhayi , E. flavolimhatus , and E. niveatus (Roumillat, unpubl. data) and may be present in other gi'oupers. Although Moe (1969) found only 1.43Vr of red groupers undergoing sexual succession, the per- centage offish with transitional gonads was even lower in the gag (1.257^). These frequencies are much lower than the transitional frequencies of such sympatric species as Centropristis striata (14'7f; Wenner et al. 1986), Calamus leucosteus (10-139r; Waltz et al. 1982), Pagrus pagrus (10%; Roumillat, unpubl. data), and Hemanthias vi- vanus (9%; Hastings 1981). A rapid rate of sex succession is probably the reason for the low fre- quency of transitional gonads found. Smith ( 1965) and Moe (1969) suggested that other gi-oupers have a very quick rate of succession, and Fishel- son (1970) and Shapiro (1981) have shown that Anthias squamipinnis can change sex within a few weeks. Despite suggestions that gag transform to males during their 10th or 11th year (McErlean and Smith 1964), it is evident that sexual succes- sion occurs at younger ages. Only seven transi- tional gag were collected, and of the five that were aged, there was one each in age gr-oups V, VI, VII, VIII, and XI. However, age X was the group in which the sex ratio approximated unity. The age of first maturity for females was lower than the previously speculated fifth or sixth year (McEr- lean and Smith 1964), and 289^ of age III, 517r of age IV, and all older femal gag had mature go- nads. Thus, there may be significantly more gag (all females because of protogyny) producing gametes than indicated in the literature, suggest- ing a greater ability to rebound from intensive overfishing than previously suspected. Acknowledgments Appreciation is expressed to C. Wenner (South Carolina Marine Resources Research Institute) for his considerable contribution to this work, and to C. Manooch (National Marine Fisheries Serv- ice, Beaufort, NC) for his helpful review of the manuscript. The study was sponsored by the Na- tional Marine Fisheries Service (Southeast Fish- eries Center) under Contract 6-35147, and by the South Carolina Wildlife and Marine Resources Department. Literature Cited Beamish. R J 1979. Differences in the age of Pacific hake iMerluccius productus ) using whole otoliths and sections of otoliths. J. Fish. Res. Board Can. 36:141-151. Beamish, R. J., and G. A. MacFarlane. 1983. The forgotten requirement for age validation in fisheries biology. Trans. Am. Fish. Soc. 112:735-743. Combs, R. M. 1969. Embryogenesis, histology and organology of the ovary of Brevoortia patronus. Gulf Res. Rep. 2:333 -434. Fischer. W (editor). ~ 1978. FAO species indentification sheets for fishery pur- poses. Western Central Atlantic (fishery area 31 ). Food Agric. Organ. U.N., Rome, Vol. 5, pag. var. Fishelson, L. 1970. Protogynous sex reversal in the fish Anthias squamipinnis (Teleostei: Anthiidae) regulated by pres- ence or absence of male fish. Nature (Lond.) 227:90-91. Hastings, P. A. 1981. Gonad morphology and sex succe.ssion in the proto- gynous hermaphrodite Hemanthias vivanus (Jordan and Swain). J. Fish. Biol. 18:443-454. Humason, G. L. 1972. Animal tissue techniques. 3d ed. W. H. Free- man and Co., San Franc, 641 p. Hyder, M. 1969. Histological studies on the testis of Tilapta leuco- sticta and other species of the genus Tilapia (Pisces: Teleostei). Trans. Am. Microsc. Soc. 88:211-231. Manooch, C. S., and M. Haimovici. 1978. Age and growth of the gag Mycleroperca microlepis, and size age composition of the recreational catch off the Southeastern United States. Trans. Am. Fish. Soc. 107:234-240. Matheson, R H . III. AND G R Huntsman 1984. Growth, mortality, and yield-per-recruit models for speckled hind and snowy gi-ouper from the United States South Atlantic Bight. Trans. Am. Fish. Soc. 113:607- 616. Matheson. R HIII.G R Huntsman, and C S Manooch III 1986. Age, growth, mortality, food and reproduction of the scamp, Mycteropcrca phenax, collected ofT North Carolina and South Carolina. Bull Mar. Sci. 38:300-312. McErlean. A J 1963. A study of the age and growth of the gag, Myeterop- erca microlepis Goode and Bean (Pisces: Serranidae) on the we.st coast of Florida. Fla. Board Conserv. Mar. Lab., Tech. Ser. No. 41, 29 p. McErlean, A J. AND C L Smith 1964. The age of sexual succession in the protogynous hermaphrodite Mycteroperca microlepis. Trans. Am. Fish. Soc. 93:301-302. Moe. M a , Jr 1969. Biology of the red grouper, Epinephelus mono (Va- lenciennes), from the eastern Gulf of Mexico. Fla. Dep. Nat. Resour. Mar. Res Lab., Prof Pap. 10, 95 p. 652 Shapiro. D Y 1981. Size, maturation and the social control of sex rever- sal in the coral reef fish Anlhuis sqiianiipinnis. J. Zool., Lond. 193:105-128. Smith. C L 1965. The patterns of sexuality and the classification of serranid fishes. Am. Mus. Novit. 2207:1-20. 1971. A revision of the American groupers: Epinephelus and allied genera. Bull. Am. Mus. Nat. Hist. 146:67- 242. Wallack. R a . AND K Selman 1981. Cellular and dynamic aspects of oocyte growth in teleosts. Am. Zool. 21:211-219. Waltz. C W . W A Roumillat. and C A Wenner 1982. Biology of the whitebone porgy. Calamus leucos- teus, in the South Atlantic Bight. Fish. Bull., U.S. 80:863-874. Wenner, C A , W A Roumillat, and W Waltz 1986. Contributions to the life history of black sea bass. Centropristis striata, off the southeastern United States. Fish. Bull., U.S. 84:723-741. Mark R Collin.s C Wayne Waltz William A Roumillat Daryl L Stubbs South Carolina Wildlife am! Marine Resources Department Marine Resources Research Institute P.O. Box 125.59 Charleston, SC 29412 AGE AND GROWTH, REPRODUCTIVE CYCLE, AND HISTOCHEMICAL TESTS EOR HEAVY METALS IN HARD CLAMS, MERCENARIA MERCEXARIA, FROM RARITAN BAY, 1974-75. Raritan Bay has historically supported an abun- dant hard clam, Mercenaria mercenaria L., re- source. It was considered the most important com- mercial species in the bay with an estimated total value of 34 million dollars in 1963 (Jacobsen and Gharrett 19671. Campbell (1967) reported a total standing crop of 4.8 million bushels of clams for the Bay for the same year (3.4 million bushels in New York waters and 1.4 million bushels off New Jersey). More recent estimates are unavailable. Raritan Bay waters have historically received various domestic and industrial wastes, some of which have had adverse effects on its shellfish resources and fisheries. Raritan Bay was closed to harvesting of hard clams on 1 May 1961, after an epidemic of human infectious hepatitis was traced to the consumption of raw clams from the bay (Campbell 1967). The closure remains in ef- fect to the present time. Zoellner (1977) reviewed the nationwide water quality problems related to shellfish and included Raritan Bay as one of the case studies in the report. Bivalves accumulate various biological and chemical contaminants by mechanisms related to their filter-feeding habits and transport across their mucous-covered, semipermeable soft body tissues (Goldberg 1957; George 1982). The accu- mulation of heavy metals, pesticides, polychlori- nated biphenyls (PCB's), oil and dispersants, dis- ease causing bacteria, viruses, fungi, parasites, and toxic phytoplankton have serious public health implications and may also adversely affect bivalve resources. Zoellner (1977) has reviewed natural and manmade conditions affecting bivalve populations, including specific studies of the effects of heavy metals, pesticides, and PCB's. McCormick et al. (1984) reviewed physical and sediment characteristics of Raritan Bay, studies of benthic organisms, plankton, and fish, and im- pact of pollution from sewage, organic chemicals, and heavy metals. The present study was undertaken to assess potential impacts of contaminants in Raritan Bay on the spawning potential of hard clams. Monthly samples were collected from three study areas within the bay to obtain measurements of the shells, soft body tissues for observations of gen- eral condition, and gonadal tissues for observa- tions of the reproductive cycle. Selected speci- mens were chosen to determine age and growth, and special tissue samples were collected for his- tochemical tests of certain metals. Published hy- drographic conditions and assessment of pollu- tants in Raritan Bay are discussed in relation to sample results. Methods Campbell (1967) described the distribution of hard clams in Raritan Bay and, based on his find- ings, sites were chosen for repeated collections. The sites were Ward Point, New Dorp Beach, and Horseshoe Cove (Fig. 1). Each was sampled at about monthly intervals beginning on 21 f^ebru- ary 1974 and ending on 7 April 1975. The clams were collected by towing a drag-type, non- hydraulic dredge with a 12-in (30 cm) wide knife from the U.S. National Marine Fisheries Service (NMFS) RV Rorqual. Tows were made at each site until 30 or more clams larger than 50 mm in shell length were caught. Special collections were made at Ward Point and New Dorp Beach on 1 November 1978 to obtain tissues for histochem- FISHERY BULLETIN VOL 8.5. NO .3, 1987. 653 Figure 1. — Locations sampled for hard clams, Mcrcenaria mercenana . in Raritan Bay during 1974 to early 1975 and 1978. ical tests of some metals, using the same dredge operated from the NMFS RV Kyma. Following sampling operations at each site, the length (longest anterio-posterior dimension), height (deepest dorso ventral dimension from the umbo to ventral shell margin) and width (thickest lateral dimension) of each clam shell was meas- ured to the nearest millimeter with vernier calipers. About 20 clams were then opened and the meat and shell liquor packaged for immediate freezing. Later measurements were made of the drained meat weight, dry meat weight, and per- cent solids by methods outlined by Ropes (1971a) for other bivalves. Specimen shells for age and growth observa- tions of hard clams were chosen from Ward Point and New Dorp Beach samples, since clams from these sites exhibited extremes in shell and weight measurements. The selection included five clams in a sample having the smallest mean size, five in a sample having the largest mean size, and those in a sample having a mean size nearly equal to the grand mean for all samples taken at the par- ticular site. The shells were radially sectioned and the cut edge polished to a high luster, as described by Peterson et al. (1983) to facilitate detection of annual growth lines. Ten additional clams were opened and the soft body tissues were removed for preservation in Bouin's fixative. Methods used for dehydrating, embedding, sectioning, and staining gonadal tis- sues to prepare slides for microscopic examina- tion of the reproductive cycle were as outlined by Ropes and Stickney (1965) and Ropes (1968). Stages in the development of gonadal tissues were established. The progressive development of sex cells through early to ripe condition and even- tual expulsion by spawning activity was a basis for evaluating reproductive viability. Failure to complete all stages of a cycle was considered an indication that the clams were being impacted by environmental conditions. Specimens were collected for histochemical tests on 1 November 1978 (10 hard clams and 2 oysters from Ward Point and 10 hard clams from New Dorp Beach). The soft body mass of each 654 specimen was removed on the vessel within an hour of capture. One cm-thick sHces were dis- sected from the central body mass of each speci- men and immersed in a vial of fixative appropri- ate for the histochemical test. Samples were then transported to the NMFS laboratory at Oxford, MD, where slides were prepared for microscopic examination. The following histochemical tests were performed: 1. Inorganic iron. Perl's (1867) reaction, re- ported in Casselman (1959), to produce ferric ferrocyanide from ferric ions (Fe^^^). 2. Arsenites and arsenates. The method of Castel (1936), reported in Pearse (1972), to precipitate the cupric salts of arsenites and arsenates (As*^^, As^^"^^'- 3. Lead. The method of Cretin (1929), called the chromate method from a reaction with neutral potasssium dichromate, was used, to- gether with the rhodisonate method (Pearse 1972). Both methods are reported in the latter paper. 4. Copper. A method is reported by Uzman (1956) to localize copper by direct treatment of the tissues with rubianic acid. After examin- ing the results of this treatment, the intensifi- cation technique to release copper "bound" proteins suggested by Uzman (1956) was also tried. Results Hard clams were generally more easily col- lected at the New Dorp Beach and Horseshoe Cove sites than at Ward Point. Bottom substrata in the dredge with the clams was predominantly a black silty-sand at New Dorp Beach and a rela- tively clean sand with some shell at Horseshore Cove, but at Ward Point, shell (mostly from oys- ters) was a major component of the black mud substrata. Live oysters were taken only at Ward Point and their soft body tissues were decidedly green. Shell and Weiglit Measurements Table 1 summarizes the mean, standard devia- tion, and number of clams for all measurements of clam shell dimensions, weights, and percent solids taken during the study. Values for Ward Point were consistently lower than those for Horseshoe Cove or New Dorp Beach where the values were similar to one another. Student's '7" test revealed no significant difference in mea- surements between Horseshoe Cove and New Dorp Beach (P > 0.05), but a significant differ- ence in all measured parameters was found be- tween clams sampled in these areas and at Ward Point (P <0.01) (Table 2). Age Observations For Ward Point, age determinations were made for the 5 smallest (collected 13 January 1975), the 5 largest specimens (collected 19 June 1974), and 8 specimens (collected 24 April 1974) which ap- proximated the overall mean shell length. For New Dorp Beach, similar determinations were made for the 5 smallest (collected 21 February 1974), the 5 largest (collected 22 May 1974), and 13 specimens (collected 1 October 1974) which approximated the overall mean shell length. Some shells in the 24 April 1974 Ward Point and 1 October 1974 New Dorp Beach samples were not prepared for age determination, because severe shell erosion at the umbonal area indi- cated tha. early age lines would be incomplete. Two differences were apparent. Firstly, none of the clams from Ward Point was older than 14 years, whereas clams at New Dorp Beach were as old as 20 years. Secondly, clams at Ward Point on Table 1 . — Measurements of ttie shell and meats of Raritan Bay hard clams, Merce- nana mercenaria , collected in 1974 and 1975. SampI e locations Ward Point New Dorp Beach Mean SD No. Horseshoe C Mean SD ;ove Measurements Mean SD No No. Shell length (mm) 77.3 9.1 221 882 12.7 225 87.0 13.2 230 Shell height (mm) 646 98 221 74.5 11.0 225 72.6 12.7 230 Shell width (mm) 44.8 5.4 221 50.3 7.5 225 50.6 10.1 230 Shell dry wt (g) 95.0 29.5 201 136.2 48.9 225 138.1 48.7 230 Underwater wt (g) 51.4 15.0 221 74.6 26.7 225 77.6 31.3 230 Drained meat wt (g) 1.3 14.1 201 56.9 18.1 205 55.0 20.9 210 Dry meat wt (g) 5.9 2.1 201 9.0 3.5 205 9.1 4.0 210 Percent solids 14.3 3.2 201 158 3.7 205 16.5 3.9 210 655 Table 2. — Results of "t" test comparing measurements of Raritan Bay hard clams. Mercenarla mercenaria, collected in 1974 and 1975. Sample locations compared New Dorp Horseshoe Cove Beach and Horseshoe Cove Measurements and Ward Point Ward Point New Dorp Beach Shell length (mm) 9.0313" 10.3795** 0.9857 Shell height (mm) 7.4520** 10.0069** 1 .7005 Shell width (mm) 7.5447** 8.8542** 0.3583 Shell dry wt (g) 10.8928** 10.3523** 0.4143 Underwater wt (g) 11.2380** 11.2120** 1.1330 Drained meat wt (g) 7.7373*' 9.6513** 0.9866 Dry meat wt (g) 10.0644** 10.7691** 0.2701 Percent solids 6.2215** 4.3548** 1.8704 Highly significant (P < 0.01). average were smaller at each age than those at New Dorp Beach (Fig. 2). mining five stages of gonadal condition, as fol- lows: Reproductive Cycle This part of the study focused on a microscopic examination of the cellular events leading up to an accumulation of sex cells in the gonads and their disappearance during the spawning act. Pe- riodic sampling identified the time and duration of spawning. Cellular structures within the sex cells and alveoli walls were important in deter- Fentales Early stage. Alveoli semicontracted, base- ment membrane thickened, and small oocytes embedded in the alveoli walls. Late stage. Oocytes more numerous at the basement membrane, some of larger diameter ex- tended into the lumina of alveoli, but attached to 120 . WARD POINT 100 E e eo z 60 X 40 20 . NEW DORP BEACH t \*' * -I I I I l_ 4 6 8 AGE 10 12 14 10 12 AGE 16 18 20 Figure 2. — Mean, standard deviation, and range of shell lengths (mm) vs. age (yr) relationships of hard clams, Mercenaria mercenaria , at Ward Point and New Dorp Beach, Raritan Bay, 1974-75. 656 the thickened basement membrane by a stalk. A few of the largest oocytes appear free in the lu- mina of alveoli. Amphinucleoli prominent in the largest oocytes. Ripe stage. Numerous large oocytes in the la- mina of alveoli. Basement membrane thin, con- taining only a few developing oocytes. Most oocytes appear free in the lumina of alveoli. Partially spent stage. Some small oocytes em- bedded in the basement membrane, with reduced numbers of large oocytes in the alveoli lumina. Spent stage. A very few of the largest oocytes in some of the alveoli lumina. Basement mem- brane somewhat thickened and containing small embedded oocvtes. earliest part of this stage. Secondary spermato- cytes and spermatids proliferating into the cen- ters of the lumina. Differentiated spermatozoa ar- ranged in dense radiating bands at the centers of the lumina. Ripe stage. Lumina of alveoli densely packed with spermatozoa. Fewer spermatocytes and spermatids occurring than in the preceding stage. Partially spent stage. Few to no spermato- cytes or spermatids at the basement membrane. Relatively reduced numbers of spermatozoa in the lumina of alveoli compared to the ripe stage. Spent stage. Alveoli nearly empty of sperma- tozoa, but a few near the basement membrane and in the lumina. Males Early stage. Alveoli semicontracted, base- ment membrane thickened, and follicle cells prominent in the lumina. A few spermatogonia or spermatocytes occur at the periphery of the lu- mina in most clams, but some with a few sperma- tozoa scattered in the lumina of some alveoli. Late stage. Alveoli expanded, basement membrane thin and attached follicular cells less apparent. Primary spermatocytes numerous at the basement membrane, especially during the For all three sites, gametogenesis progressed from the early to late stages from 21 February to 24 April 1974 (Table 3). Ripening was earliest at Horseshoe Cove, with 109f of the clams in this condition by 22 May and TO'/r by 19 June. At this latter date, 509r of the Ward Point clams were ripe. Clams at all the sites had ripened by 23 July, and some (20'^f ) at Ward Point and many (70'';/ ) at New Dorp Beach were in the partially spent con- dition, an indication that spawning had begun at these two sites. Spawning was later at Horseshoe Cove, with 679^ in the partially spent condition by 21 August. Ripe clams were observed in the sam- Table 3^ — Percent occurrence of developmental stages during the reproductive cycle of Raritan Bay hard clams, Mercenana mercenana. collected in 1974 and 1975. 1974 1975 Sample site and gonad condition 2 21 3 28 4 24 5 22 6 19 7'23 821 10 1 11 5 1 13 26 47 Horseshoe Cove early 60 70 30 40 10 40 100 late 40 30 70 50 20 ripe 10 70 100 33 40 30 part spent 67 20 60 60 spent 40 10 40 60 New Dorp Beach early 70 70 70 50 80 90 late 40 20 30 50 100 ripe 30 30 part spent 70 70 50 90 20 spent 10 50 10 80 20 10 Ward point early 40 40 50 40 60 50 late 50 40 50 60 50 20 ripe 50 80 30 30 part spent 20 30 50 100 spent 10 20 40 20 100 40 30 657 pies from Horseshoe Cove as late as 5 November, but only until 1 October and 21 August at Ward Point and New Dorp Beach, respectively. Par- tially spent clams were collected from Horseshoe Cove and New Dorp Beach as late as 13 January 1975, but only until 5 November 1974 at Ward Point. The spent and early gametogenic stages identified in clams from all sites by 6 February 1975 indicated that the 1974 reproductive cycle had been completed by all clams and that a new cycle had begun for some. These observations sug- gest that the spawning period was shortest at Ward Point (5 months) and longest at New Dorp Beach (7 months), with Horseshoe Cove interme- diate (6 months). At all three sites, gametogenesis progressed through morphologically normal stages resulting in the complete spawning of most ripe gametes. Cytolysis of unspent cells was not observed. No hermaphrodites were seen in any of the samples. The sex ratios of hard clams in the samples were as follows: at Ward Point, 59 males and 60 females; at New Dorp Beach, 56 males and 64 females; and at Horseshoe Cove, 74 males and 45 females. The hypothesis of 1:1 sex ratio was tested by chi-square for all three populations; re- sults indicated a significant (P < 0.01) deviation for Horseshoe Cove. Histochemical Tests for Metals Histochemical tests were performed on four male and six female hard clams and one male and one female oysters collected at Ward Point; and three male and seven female hard clams collected at New Dorp Beach. The expected histochemical reactions for metals in clam tissues were not ob- served, i.e., deep Prussian blue for inorganic iron, green granular precipitate for arsenites and arse- nates, yellow opaque crystals or scarlet red pre- cipitate by the chromate or rhodizonate methods, respectively for lead, and deep greenish-black precipitate for copper. Thus, the tests for metals in hard clam tissues proved to be negative. Fe- male gonadal tissues tested for arsenites and ar- senates from both collecting sites had 1-2 \xm di- ameter granules in the oocyte cytoplasm but the color could not be determined. No similar gran- ules were seen in male tissues. For the two oys- ters from Ward Point, the deep greenish-black precipitate for copper was evident in connective tissue cells beneath the body wall and palps (a similar reaction was also seen in the epidermal cells of the palps), around the digestive divertic- ula (Fig. 3), and near the base of cells lining the gills and gut. The connective tissue cells sur- rounding the male oyster gonadal tubules also tested positive. No similar reaction was seen in the connective tissue cells surrounding the fe- male oyster gonadal tubules, which contained large and apparently normal oocytes. Although not seen in New Dorp Beach hard clam tissues, some gill tissues of hard clams from Ward Point tested for copper showed a slight darkening, but a precipitate was not clearly evident, such as was seen in oysters. The darkening was also absent in underlying connective tissue cells and other tis- sues. Modification of the technique to intensify the copper reaction was negative for all hard clam and oyster tissues from the two collection sites. Discussion Shell dimensions and shell and body weights clearly indicated a smaller size for hard clams at Ward Point than at New Dorp Beach and Horse- shoe Cove, which was reflected in the age esti- mates. Clams at Ward Point were younger (none >14 years) and smaller {X = 77 mm; none >97 mm) than clams at New Dorp Beach (none >20 years; X = 88 mm; none >113 mm). These are values much lower than the 111 mm mean and maximum length of 144 mm reported for Nan- tucket Sound hard clams by Ropes and Martin (1960) and recent, almost 60-yr longevity esti- mate for the species (Ropes pers. obs.). Determination of the percentage solids for the meats of bivalves is a measure of condition (Engle and Chapman 1953). The following mean values have been reported: 18.4% for soft-shell clams, Mya arenaria (Harriman 1954), 17.09^ for oysters, C. virginica (Engle 1958), 21.4% for surf clams, Spisula solidissima (Barker and Merrill 1967), and 18.5% for ocean quahogs, Arctica islandica, (Ropes 1971a). These compare favorably with 16.5% and 15.8% for Horseshoe Cove and New Dorp Beach hard clams, respectively (Table 1). The low value of 14.3% for Ward Point hard clams is an indication of poor condition. Reported age and growth determinations for hard clams suggest that the Ward Point portion of the Raritan Bay population was being adversely affected. Ansell (1968) has extensively reviewed the literature on annual and seasonal growth of hard clams from various investigations in Canada, the United States, and Europe. Length- on-age observations of the growth of hard clams at sites in Florida, North Carolina, New Jersey, 658 1 < •• ^ * » \ • t> * • * • •• ft • % < « • « • ••. -p • d ft 4: ♦ • • 4- 1 * * # ft ■ ♦ 9 • * * • « If • • t - • • * 1 '50]ims 4 o •• . «(^^ "^ w^ ♦ 1^ 10 pm Figure 3. — Digestive diverticula of a copper "sick" oyster, Crassostrea virginica, from Raritan Bay. Connective tissue cells surrounding the diverticula were greenish black, a positive reaction of copper by the rubianic acid method. A scale of magnification appears in the lower right-hand corner of each photomicrograph. 659 New York, Massachusetts, Maine, and Canada were compared with the data for the hard clams sampled at the New Dorp Beach and Ward Point, Raritan Bay sites. Growth of Raritan Bay clams was about midway between the fastest (Florida) and the slowest (Canada). At Ward Point, growth to age 4 was about equal to New Jersey growth, but at New Dorp Beach, gi-owth was consistently greater. After age 4, clams at Ward Point grew much slower than New Jersey clams, and even slower than Maine clams after age 5. These obser- vations support the conclusion by Ansell ( 1968) of extreme local variations in the annual growth of hard clams. The absence of large, old hard clams in the present samples may be the result of pollution effects, as Jefferies (1972) found for Providence River, RI, hard clams stressed by hydrocarbons. A high Ci5, hydrocarbon value of 3,672 ppm in sed- iments at Ward Point was reported by Koons and Thomas (1979). Nevertheless, mortalities (e.g., paired valves containing dead bodies) were not evident at any site. Food availability is considered an important factor for growth of hard clams by Ansell (1968). Jefferies ( 1962) considered the nutrient content of Raritan Bay to be rich and the environment capa- ble of supporting dense biotic communities, be- cause of a sluggish circulation pattern. Patten (1962) found that phytoplankton species diversity decreased up bay (towards Ward Point) from the higher values in the Lower Bay. A lower diversity of phytoplankton in the vicinity of Ward Point may have resulted in lower amounts of food or- ganisms being available for the nutritional needs of the clams, and was probably reflected in their slower growth and poor meat condition. Gonadal development culminated in spawning at three Raritan Bay sample sites. This suggests that the reproductive capacity of hard clams in Raritan Bay was not being affected by pollutants. However, some differences were noted in a com- parison of the results based on available informa- tion about the time and duration of spawning and larval production at several northwestern At- lantic coast locations. At more northern locations, Belding (1912) and Deevey (1948) in Massachu- setts, Landers (1954) in Rhode Island, and Car- riker (1959, 1961) in Long Island, NY, and Little Egg Harbor, NJ, observed that spawning was ini- tiated 1 to 2 months earlier than was observed in Raritan Bay during 1974. Similarly, at more southern locations. Keck et al. (1975) in Dela- ware, Sieling (1956) in Maryland, Ropes (1971b) and Chanley and Andrews (1971) in Virginia, Porter (1967) in North Carolina, and Eversole and Michner (1980) in South Carolina observed that spawning was initiated 1 to almost 3 months earlier. A spawning beginning about three- fourths of a month earlier than the present study was observed by Jefferies (1962) in Raritan Bay, but Loosanoff (1937) in Long Island, NY, ob- served that spawning began at the same time as in Raritan Bay during 1974. No particular trend in the time of peak spawning was evident in the several studies, except that the peak spawning in 1974 at all Raritan Bay sample sites was later than reported in any of the other studies. Spawn- ing ceased somewhat earlier at more northern locations, and was not as prolonged at more southern locations as was observed in Raritan Bay during 1974. The project was initiated under the premise that heavy metal pollution in Raritan Bay could be affecting the viability of adult hard clams. Studies indicated that tests for copper and lead should be specifically included, because high con- centrations of both have been found in Raritan Bay sediment and water samples (Greig and Mc- Grath 1977; Waldhauer et al. 1978). The negative results of histochemical tests for heavy metals in Raritan Bay hard clams are not readily explained. Eisler (1981) has listed studies that found 15 heavy metals (including those tested for in the present study) in field collected hard clams. However, heavy metals can occur in several forms (Waldichuk 1979; Fayi and George 1985), suggesting that the histochemical tests may not have been specific for those occurring in Raritan Bay hard clams. Pringle et al. (1968) re- ported lower levels of copper in field collected hard clams than oysters iCrassostrea virginica and C. gigas ). The positive result for copper in the oysters from Ward Point is probably related to the species greater sensitivity and accumulation of more of the metal in their tissues than hard clams. Copper may have been at a level too low for detection by the histochemical test, although lim- its for detection of copper or other metals were not given in Pearce (1972). Hydrographic conditions (not specifically sam- pled for during the present study) probably influ- enced the growth and survival of hard clams in Raritan Bay. Based on current flow observed by Jefferies (1962) and Patten (1962), the New Dorp Beach area is influenced principally by water from the Hudson River and the ocean; the Ward Point area is influenced by an eddy formed from 660 the westward flow of water from the ocean, and flows from the Arthur Kill and Raritan Rivers; the Horseshoe Cove area is affected most strongly by water flowing from the Shrewsbury River. Ansell (1968) analyzed data from throughout the geogi'aphical range of adult hard clams to de- velop a relationship between temperature and gi'owth rate. Shell growth occurred between tem- peratures of 9°-3rC and ceased at lower and higher temperatures; the optimum was 20°C. Castagna and Chanley (1973) reported a salinity tolerance range of 12.5-46'/ff for survival of adult hard clams, with an optimum of 24-28'^ff. The above temperature and salinity limits for adult hard clams were compared with hydro- gi'aphic results reported by Jefferies (1962). He listed mean surface and bottom water data begin- ning in summer 1957 to summer 1958 at two loca- tions (Stations 1 and 6) in Raritan Bay near the Ward Pomt and New Dorp Beach sample sites, respectively. Throughout Raritan Bay no growth of adult clams would be expected during winter due to low bottom temperatures (2.3"-3.0°C); slow growth would occur during the increasing and decreasing temperatures of the spring and fall. Near normal growth probably occurs at New Dorp Beach and Ward Point during the summer when temperature means appeared to be near optimum conditions. Lowest bottom salinities were re- corded during the spring at Jefferies' (1962) sta- tion number 1 near Ward Point. These values indicate that the minimum salinity tolerance limit for adult hard clams may occasionally be reached in the area. Salinities near the New Dorp Beach area (Jefferies 1962, station 6) were all within the tolerance limits for adult hard clams. Jefferies (1962) reported dissolved oxygen measurements and found relatively low concen- trations in the water near Ward Point in both summer periods. Slightly higher values occurred in the fall of 1957 and following spring near Ward Point, but the confidence intervals were greater than for any other period, indicating more vari- able conditions. Dissolved oxygen levels near New Dorp Beach were consistently higher than near Ward Point. Literature Cited A.NSKLL A D 1968. The rate of growth of the hard clams Merccnana mercenaria (L.) throughout the geographic range. J. Cons. 31:364-409. Barker. A M . and A S Merrill 1967. Total .soHd.s and length-weight relation of the surf clam, Spisula soUdissima. Proc. Natl. Shellfish. Assoc. 57:90-94. Belding. D L 1912. A report upon the quahaug and oyster fisheries of Massachu.setts, including the life history, growth and cultivation of the quahaug [Venus mercenaria ), and ob- servations on the set of the oyster spat in Wellfleet Bay. Wright and Potter Printing Co., Boston, 134 p. Campbell. R 1967. A report on the shellfish resources of Raritan Bay, New Jersey. In Proceedings of the Conference on Pollu- tion of Raritan Bay and Adjacent Interstate Waters, Vol. 2, p, 6.53-681. U.S. Dep. Inter., Fed. Water Pollution Control Admin., Third Sess. Carriker. M R 1959. The role of physical and biological factors in the culture of Crassoslrea and Mercenaria in a salt-water pond. Ecol. Monogr. 29:219-266. 1961. Interrelation of functional morphology, behavior, and autecology in early stages of bivalve Mercenaria mer- cenaria. J. Elisha Mitchell Sci. Soc. 77:168-241. Casselman. W G B 1959. Histochemical technique. Methuen's Monogr. Biol. Subj., John Wiley and Sons, Inc., N.Y., 205 p. Castagna. M , and P Chanley 1973. Salinity tolerance of some marine bivalves from in- shore and estuarine environments in Virginia waters on the Western Mid-Atlantic coast. Malacologia 12:47-96. Chanley. P . and J D Andrews 1971. Aids for identification of bivalve larvae of Vir- ginia. Malacologia 11:45-119. Deevey. G B 1948. The zooplankton of Tisbury Great Pond Bingham Oceanogr. Inst. Coll. Bull. 12:1-44. ElSLER, R 1981. Trace metal concentrations in marine organ- isms. Pergamon Press, N.Y., 687 p. EngleJ B 1958. The seasonal significance of total solids of oysters in commercial exploitation. Proc. Natl. Shellfish. Assoc. 48:72-78. Engle.J B.andC R Chaf'.man 1953. Oyster condition affected by attached mus- sels. Proc. Natl. Shellfish. Assoc, 1951 Conven. Add. p. 70-78. Eversole. a G . and W K Michner 1980. Reproductive cycle of Mercenaria mercenaria in a South Carolina estuary. Proc. Natl. Shellfish. Assoc. 70:22-30. Fayl L . AND S G George 1985. Purification of very low molecular weight Cu- complexes from the European oyster. In F. J. Vernberg, F. P Thurburg, A. Calabrese, and W Vernberg (editors). Marine pollution and physiology: recent advances, p. 145-155. Univ. South Carolina Press. George, S G 1982. Subcellular accumulation and detoxification of metals in aquatic animals. In W. B. Vernberg, A. Cal- abrese, F. P. Thurburg, F. J. Vernberg (edi- tors). Physiological mechanisms of marine pollutant toxicity, p. 3-52. Acad. Press, N.Y. Goldberg, E D 1957. Chap. 12. Biogeochemistry of trace metals. In J. W. Hedgpeth (editor). Treatise on marine ecology and paleoecology, p. 345-357. Geol. Soc. Amer. Mem. 67, 661 Vol. 1. Ecol. GREIG, R. a , AND R A McGRATH 1977. Trace metals in sediments of Raritan Bay. Mar. Pol- lution Bull. 8:188-192. Harriman, D M 1954. Variations in total solids of the soft clam (Mya are- naria i. Maine Dep. Sea Shore Fish., Res. Bull. No. 23, 14 p. JACOBSEN, F L , AND J T Gharrett 1967. Joint Statement at Conference on Pollution of Rari- tan Bay and Adjacent Interstate Waters. U.S. Dep. In- ter., Fed. Water Pollution Control Admin., Third Bess., Vol. 2, p. 683-698. Jefferies, H P 1962. Environmental characteristics of Raritan Bay, a polluted estuary. Limnol. Oceanogr. 7:21-31. 1972. A stress syndrome in the hard clam, Mercenaria mercenaria. J. Invert. Pathol, 20:242-251. Keck, R T , D Maurer. and H Lino 1975. A comparative study of the hard clam gonad devel- opmental cycle. Biol. Bull. (Woods Hole) 148:243-258. KooNS. C, B , AND J P Thomas 1979. Ci5+ hydrocarbons in the sediments of the New York Bight. Proc. 1979 Oil Spill Conference (Preven- tion, Behavior, Control, Cleanup), p. 625-628. Am. Pet. Inst., Washington, DC. Landers, W. S. 1954. Seasonal abundance of clam larvae in Rhode Island waters, 1950-52. U.S. Dep. Inter., Bur. Comm. Fish., Spec. Sci. Rep. Fish. 117, 29 p, LOOSANOFF, V L , 1937. Seasonal gonadal changes of adult clams, Venus mercenaria (L), Biol, Bull, (Woods Hole) 72:406-416. McCORMICK, J M , H G MULTER, AND D. M StAINKEN 1984, A review of Raritan Bay research. Bull, N,J. Acad. Sci, 29:47-58. Patten, B C 1962. Species diversity in net phytoplankton of Raritan Bay. J. Mar, Res, 20:57-75, Pearse, a G E 1972, Histochemistry theoretical and applied, 3d ed. Williams and Wilkins Co,, Baltimore, Vol, 2, 1518 p, Peterson. C H, P B Duncan, H C Summerson, and G W Safrit.Jr 1983, A mark-recapture test of annual periodicity of in- ternal growth band deposition in shells of hard clams, Mercenaria mercenaria , from a population along the southeastern United States, Fish, Bull. U,S, 81:765- 779, Porter, H J 1967, Seasonal gonadal changes of adult clams, Merce- naria mercenaria (L.), in North Carolina, Proc. Natl. Shellfish. Assoc. 55:35-52. Pringle, B H . D E Hissong, E. L Katz. and S T Mulawka 1968. Trace metal accumulation by estuarine mol- lusks. J. Sanit. Eng. Div., Proc. Am. Soc. Civil Eng. Proced. Pap. 5970 (SA3):455-475, Ropes. J W 1968, Reproductive cycle of the surf clam, Spisula solidis- sima, in offshore New Jersey, Biol, Bull. (Woods Hole) 135:349-365. 1971a. Percentage of solids and length-weight relation- ship of the ocean quahog. Proc. Natl. Shellfish, Assoc. 61:88-90, 1971b. Maryland's hard clam studied at Oxford Labora- tory, Comm, Fish, News, Md, Fish, Wildl, Admin. 4(6):2-3. Ropes, J W . and C E Martin 1960. The abundance and distribution of hard clams in Nantucket Sound, Massachusetts, 1958. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. No. 354, 12 p. Ropes. J W , and A P Stickney 1965. Reproductive cycle of Mya arenaria in New Eng- land. Biol. Bull. (Woods Hole) 128:315-327. Sibling, F W 1956. The hard shell clam fishery of Maryland wa- ters. Md. Tidewater News 12(10), suppl. 9, 2 p. UZMAN. L. L 1956. Histochemical localization of copper with rubeanic acid. Lab. Invest, 5:299-305, Waldhauer, R . A Matte, and R E Tucker, 1978, Lead and copper in the waters of Raritan and lower New York Bays, Mar, Pollutants Bull, 9:38-42, Waldichuk. M 1979, Review of the problems, //! H, A. Cole (editor). The assessment of sublethal effects of pollutants in the sea, p. 1-26. Royal Soc. Lond. Zoellner, D R 1977. Water quality and molluscan shellfish: An over- view of the problems and the nature of appropriate Fed- eral laws. U.S. Dep. Commer., NOAA, NMFS, 106 p. (with appendices). John Ropes Northeast Fisheries Center Woods Hole Laboratory National Marine Fisheries Service, NOAA Woods Hole, MA 02543 662 ERRATA Fisbeiy Bulletin Vol. 85, Vo. 1 Utter, Fred, David Tell, George Milner, and Donald Mclsaac, "Genetic estimates of stock compositions of 1983 chinook salmon, Oncorhynchus tshawytscha, harvests off the Washington coast and the Columbia River," p. 13-23. Page 16. Table 1, Stock group, correct to read as follows: California 6 Puget Sound and British Columbia 4 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), 7(, '/<(, 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 5'^/4 inches (for single-column figures, allowing for 509^ 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 arabic 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 arabic numeral and heading provided). LIST OF FIGURES (Entire figure legends) FIGURES (Each figure should be numbered with an arabic 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 SOGARD, SUSAN M., DONALD E. HOSS, and JOHN J. GOVONI. Density and depth distribution of larval gulf menhaden, Breuoortia patronus , Atlantic croaker, Micropogonias undulatus, and spot, Leiostomus xanthurus, in the northern Gulf of Mexico 601 BOWERING, W. R. Distribution of witch flounder, Glyptocephalus cynoglossus , in the southern Labrador and eastern Newfoundland area and changes in certain biological parameters after 20 years of exploitation 611 DALLEY, E. L., and G. H. WINTERS. Early life history of sand lance {Am- modytes), with evidence for spawning A. dubius (in Fortune Bay, Newfound- land 631 Notes MERCALDO-ALLEN, RENEE, and FREDERICK P. THURBERG. Heart and gill ventilatory activity in the lobster, Homarus americanus, at various tempera- tures 643 FOREMAN, TERRY. A method of simultaneously tagging large oceanic fish and injecting them with tetracycline 645 SCHAEFER, KURT M. Second record of the Kawakawa, Euthynnus affinis, from the eastern Pacific Ocean 647 COLLINS, MARK R., C. WAYNE WALTZ, WILLIAM A. ROUMILLAT, and DARYL L. STUBBS. Contribution to the life history and reproductive biology of gag, Mycteroperca microlepis (Serranidae), in the South Atlantic Bight 648 ROPES, JOHN. Age and growth, reproductive cycle, and histochemical tests for heavy metals in hard clams, Mercenaria mercenaria, from Raritan Bay, 1974- 75 653 GPO 791-008 *^''' ?"^o^ \ r I s n 6 XM^S Hall ® * ' ^ ^^ATts 0< ^ / LIBRARY MAR 2 2 1988 Woods Hole, Mass. Vol. 85, No. 4 October 1987 QUAST, JAY C. Morphometric variation of Pacific Ocean perch, Sebastes alutus, off western North America 663 REISENBICHLER, R. R., and S. R. PHELPS. Genetic variation in chinook, Oncorhynchus tshawytscha, and coho, O. kisutch, salmon from the north coast of Washington 681 KLEIBER, P., and B. BAKER. Assessment of interaction between North Pacific albacore, Thunnus alalunga, fisheries by use of a simulation model 703 LAI, HAN-LIN, DONALD R. GUNDERSON, and LOH LEE LOW. Age determina- tion of Pacific cod, Gadus macrocephalus , using five ageing methods 713 ABLE, K. W., D. C. TWICHELL, C. B. GRIMES, and R. S. JONES. Sidescan sonar as a tool for detection of demersal fish habitats 725 SMITH, ERIC M., and PENELOPE T. HOWELL. The effects of bottom trawling on American lobsters, Homarus americanus, in Long Island Sound 737 KUDO, GEORGE, HAROLD J. BARNETT, and RICHARD W. NELSON. Factors affecting cooked texture quality of Pacific whiting, Merluccius productus, fillets with particular emphasis on the effects of infection by the Myxosporeans Kudoa paniformis and K. thyrsitis 745 LEIS, JEFFREY M., BARRY GOLDMAN, and SHOJI UEYANAGI. Distribution and abundance of billfish larvae (Pisces: Istiophoridae) in the Great Barrier Reef Lagoon and Coral Sea near Lizard Island, Australia 757 BRILL, RICHARD W., ROBERT BOURKE, JAMES A. BROCK, and MURRAY D. DAILEY. Prevalence and effects of infection of the dorsal aorta in yellowfin tuna, Thunnus albacares, by the larval cestode, Dasyrhynchus talismani 767 FABLE , WILLIAM A. , JR. , ALL YN G. JOHNSON, and LYMAN E . B ARGER. Age and growth of Spanish mackerel, Scomberomorus maculatus, from Florida and the Gulf of Mexico 777 O'NEIL. STEVEN P., and MICHAEL P. WEINSTEIN. Feeding habitats of spot, Leiostomus xanthurus, in polyhaline versus meso-oligohaline tidal creeks and shoals 785 APPELDOORN, RICHARD S. Assessment of mortality in an offshore population of queen conch, Strombus gigas L., in southwest Puerto Rico 797 {Continued on back cover) Seattle, Washington 20-1 U.S. DEPARTMENT OF COMMERCE C. William Verity, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION NATIONAL MARINE FISHERIES SERVICE William E. Evans, Assistant Administrator 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. Bruce B. Collette 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 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 Cppimercehas determined that the publication of this periodical is necessary in the transaction of the public business required by law of this Department. 20-2 Fishery Bulletin CONTENTS Vol. 85, No. 4 October 1987 QUAST, JAY C. Morphometric variation of Pacific Ocean perch, Sebastes alutus, off western North America 663 REISENBICHLER, R. R., and S. R. PHELPS. Genetic variation in chinook, Oncorhynchus tshawytscha, and coho, O. kisutch, salmon from the north coast of Washington 681 KLEIBER, P., and B. BAKER. Assessment of interaction between North Pacific albacore, Th annus alalunga, fisheries by use of a simulation model 703 LAI, HAN-LIN, DONALD R. GUNDERSON, and LOH LEE LOW. Age determina- tion of Pacific cod, Gadus macrocephalus , using five ageing methods 713 ABLE, K. W., D. C. TWICHELL, C. B. GRIMES, and R. S. JONES. Sidescan sonar as a tool for detection of demersal fish habitats 725 SMITH, ERIC M. , and PENELOPE T. HOWELL. The effects of bottom trawling on American lobsters, Homarus americanus, in Long Island Sound 737 KUDO, GEORGE, HAROLD J. BARNETT, and RICHARD W. NELSON. Factors affecting cooked texture quality of Pacific whiting, Merluccius productus, fillets with particular emphasis on the effects of infection by the Myxosporeans Kudoa paniformis and K. thyrsitis 745 LEIS, JEFFREY M., BARRY GOLDMAN, and SHOJI UEYANAGI. Distribution and abundance of billfish larvae (Pisces: Istiophoridae) in the Great Barrier Reef Lagoon and Coral Sea near Lizard Island, Australia 757 BRILL, RICHARD W., ROBERT BOURKE, JAMES A. BROCK, and MURRAY D. DAILEY. Prevalence and effects of infection of the dorsal aorta in yellowfin tuna, Thunnus albacares, by the larval cestode, Dasyrhynchus talismani 767 FABLE, WILLIAM A., JR., ALLYN G. JOHNSON, and LYMAN E. BARGER. Age and growth of Spanish mackerel, Scomberomorus maculatus, from Florida and the Gulf of Mexico 777 O'NEIL, STEVEN P., and MICHAEL P. WEINSTEIN. Feeding habitats of spot, Leiostomus xanthurus, in polyhaline versus meso-oligohaline tidal creeks and shoals 785 APPELDOORN, RICHARD S. Assessment of mortality in an offshore population of queen conch, Strombus gigas L., in southwest Puerto Rico 797 (Continued on next page ) Seattle, Washington 1987 For sale by the Superintendent of Documents, U.S. Government Printing OfTice, ton DC 20402— Subscription price per year: $16.00 domestic and $20.00 foreign Marine Biological Laboratory LIBRARY MAR 2 2 1988 'ashing- ost per ion u\^ ,£U4u^ — ouDscription price per year: 3>lb.UU domestic and ^ZU.UD loreign. Jost per u I kAa»» single issue: $9.00 domestic and $11.25 foreign. I WOOuS "016, WlaSS. t 2-1 Contents — Contin ued BOTTON, MARK L., and JOHN W. ROPES. Populations of horseshoe crabs, Lim - ulus polyphemus, on the northwestern Atlantic continental shelf 805 Notes HENWOOD, TYRRELL A., and WARREN E. STUNTZ. Analysis of sea turtle captures and mortalities during commercial shrimp trawling 813 RENDER, JEFFREY H., and ROBERT L. ALLEN. The relationship between lunar phase and gulf butterfish, Peprilus burti, catch rate 817 FISHER, J. P., and W. C. PEARCY. Movements of coho, Oncorhynchus kisutch, and Chinook, O. tshawytscha, salmon tagged at sea off Oregon, Washington, and Van- couver Island during the summers 1982-85 819 YOKLAVICH, MARY M., and GEORGE W. BOEHLERT. Daily growth incre- ments in otoliths of juvenile black rockfish, Sebastes melanops: An evaluation of autoradiography as a new method of validation 826 Index 833 Notices 846 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. MORPHOMETRIC VARIATION OF PACIFIC OCEAN PERCH, SEBASTES ALUTUS, OFF WESTERN NORTH AMERICA Jay C. Quast' ABSTRACT Pacific ocean perch, Sebastes alutus, vary in body form over the eastern Pacific Ocean and southeast- ern Bering Sea. When related to a 260 mm standard length, a small adult size, most of 18 body measurements change from east to west as V-shaped clines. Belly size, however, lengthens as a single cline from the Vancouver vicinity westward, and the lengthening is coupled with shortening of measurements complementary to the belly measurement, from head to belly and belly to tail. Most adult body dimensions are sexually dimorphic, but the dimorphism is slight. Growth inflections may occur but, if so, are hidden in the data. Body form does not change markedly with growth except for the symphyseal knob, which becomes relatively larger, and the 3d anal-fin spine, which becomes relatively shorter. Putative subspecies of S. alutus probably are premature because supportive mor- phometric comparisons and criteria on the subspecies seem based on too few data, and do not ade- quately consider complex clinal variation and growth allometry evidenced in the eastern part of the species range. Also, significant morphometric variation may be phenotypic. Pacific ocean perch, Sebastes alutus, a commer- cially important rockfish (Scorpaenidae) in the North Pacific Ocean, range ft-om northern Hon- shu, Japan, to California. To date, the species' taxonomy has been based on relatively limited local representation, on preserved material, and on analyses without probabilistic interpretation or attention to allometric growth. Matsubara (1943) identified a Honshu representative as a new species {Sebastes paucispinosus), but Bar- sukov (1964), after examining specimens taken across the North Pacific Ocean, suggested that the variation he found indicated, at most, possible eastern and western Pacific subspecies, with both possibly occurring off the Aleutian Arc. Chen (1971) demonstrated that growth rate, can influ- ence body proportions in a Sebastes species, and Westrheim (1973) found a cline of increasing growth rate in S. alutus from the northern Gulf of Alaska to Washington (Quast^ found that the cline may be more related to latitude than tem- perature). Because of the commercial importance of Pacific ocean perch and the lack of definitive in- formation on possible subspecies or genetic stocks at the onset of the study, the National Marine 'Northwest and Alaska Fisheries Center Auke Bay Labora- tory, National Marine Fisheries Service, NOAA, P.O. Box 210155, Auke Bay, AK 99821; present address: 1565 Jamestown Street S.E., Salem. OR 97302. 2Jay C. Quast. Annual growth in Pacific ocean perch, Se- bastes alutus: variation, stanzas, compensation, and simula- tion. Manuscr. in prep. Manuscript accepted June 1987. FISHERY BULLETIN: VOL. 85, NO. 4, 1987. Fisheries Service gathered morphometric data on representatives from the Gulf of Alaska and east- ern Bering Sea from 1968 to the mid-1970's. We sought evidence of disjunct geographic variation that might indicate genetic stocks, and analyzed for characteristics of growth and sexual variation of possible taxonomic significance. Sampling was in the shallow range of the species distribution, at <200 fathoms (366 m). METHODS The prohibitively large volume needed for spec- imen storage for an extensive statistical study and the shrinkage, bending, and other distortions caused by preservation were avoided by pho- tographing freshly trawled Pacific ocean perch. A portable stand supported a 35 mm camera and flash unit 145 cm above a V-shaped easel, which helped restrain the fish from vessel motion. The long focal distance minimized foreshortening in the photographs, and a 100 mm telephoto lens reduced field size to include only the specimen, a centimeter scale, a numbered theater ticket, and a plastic card with pencil-inscribed catch infor- mation. Specimens were flattened, straightened, and centered on the easel; fins were placed as erect as possible; and the lower jaw was propped closed with handheld forceps. Later, in the labo- ratory, body measurements were taken from im- ages projected from the color transparencies onto the back of a ground-glass screen. The images 663 FISHERY BULLETIN: VOL 85. NO. 4 were brought to natural size by reference to the centimeter scale in each photograph. Accuracy was ensured for body-depth measurements by using a right-angle scale on the body axis. The study used 18 linear measurements'^ (nearest millimeter) in addition to standard length (SL) of approximately 1,500 specimens sampled from the Bering Sea and Gulf of Alaska during summers of 1968-70 (Fig. 1, Table 1). The measurements gave major dimensions of the body and accessory structures, similar to those stand- ard in taxonomic studies (Hubbs and Lagler 1949): Standard length . — From tip of the premaxillary to the skin deflection (evidenced by a change in shading) at end of the hypural plate. Nape. — From tip of premaxillary to the anterior insertion (junction of anterior outline with body profile) of the first spinous ray in the dorsal fin. Spinous dorsal-fin length. — From anterior inser- tion of the spinous dorsal fin to the posterior 3Data for the measurements, including additional measure- ments not given in this paper, are available on tape as JQUAST/ MORFl on file at the Northwest and Alaska Fisheries Center, National Marine Fisheries Service, 2725 Montlake Blvd. East, Seattle, WA 98112 insertion (junction of posterior outline of spine with body profile) of the 13th spinous fin ray. Hind-trunk dorsal. — From posterior insertion of the 13th spinous fin ray to end of the hypural plate. Hind-trunk ventral. — From posterior insertion of the second anal-fin spine to end of the hypural plate. Belly. —From posterior insertion of the pelvic fin to posterior insertion of second anal-fin spine. Pelvic insertion. — From tip of the premaxillary to posterior insertion of the pelvic fin. Head. — From tip of the premaxillary to posterior edge of the opercular fiap. Body-depth pelvic. — Dorsoventral distance, tak- en perpendicular to the longitudinal axis of the fish, from the posterior insertion of pelvic fin to dorsal body outline. Body-depth anal. — Dorsoventral distance, taken perpendicular to the longitudinal axis of the fish, from the anterior insertion of first anal spine to dorsal body outline. Caudal peduncle. — Shortest distance across cau- dal peduncle, taken perpendicular to its longi- tudinal axis. Orbit.— Greatest diameter between opposite sides of the orbit, taken parallel to longitudinal axis k^ 60- 55- 50- ATKA-GULF GULF OF ALASKA VANCOUVER -J L ■■ I I I I I I I I I I I I I I I I I I I I I I I I I I 'III 1_L I I I 170 165 160 155 150 145 140 135 130 Figure 1. — Geographic regions and statistical areas (dark rectangles) where Pacific ocean perch were obtained for this study. 664 QUAST: MORPHOMETRIC VARIATION ON PACIFIC OCEAN PERCH Table 1. — Standard-length (SL) frequencies of Pacific ocean perch used in this study. Geo- graphic zones are shown in Figure 1 . M = male; F = female. Standard Atka- Atka - South- length (mm) Beri M ng F Gulf Kodiak M F Yakutat M F eastern M F Vancoi M uver Class Midpoint M F F 115-134 125.5 5 5 — — — — — — — 1 — — 135-154 145.5 3 4 — 1 — — — — — 1 — — 155-174 165.5 4 5 1 — 4 4 — — 1 2 — — 175-194 185.5 16 14 2 4 14 15 2 5 4 5 — — 195-214 2055 20 15 13 9 6 7 4 11 15 7 — — 215-234 2255 13 26 51 55 10 6 16 24 13 8 — — 235-254 245.5 36 22 49 52 6 7 14 19 7 5 5 — 255-274 2655 22 19 25 12 7 8 9 10 12 17 4 2 275-294 2855 16 22 15 13 19 9 4 10 16 9 10 4 295-314 3055 10 13 26 21 65 33 6 7 23 5 13 3 315-334 325.5 17 25 13 22 16 35 4 8 11 5 9 4 335-354 345.5 3 26 1 7 4 18 4 4 — — 2 8 355-374 365.5 3 17 — 1 — 2 1 5 — 1 2 1 375-394 3855 — 2 — — — — 2 — 1 — — 1 395-414 405.5 — — — — — — — — 1 — — 1 Total by sex: 168 215 196 ■ 197 151 144 66 103 104 66 45 24 Region totals: 383 393 295 169 170 69 1 Average SL, by region: 260.5 2579 281.4 261.5 261.2 308.4 Average SL, all datai 266.41 {N = ■ 1 ,479) 'Class midpoints weighted by frequency. of the fish. Longest pectoral -fin ray.— From tip of the longest ray to its origin. Upper-jaw length. — Greatest distance from tip of the premaxillary to posterior edge of the maxil- lary. Upper-jaw width. — Greatest width of this part taken perpendicular to axis of the mouth line in a closed mouth. Symphyseal knob. — From tip of symphyseal knob to its posterior insertion. 6th dorsal spinous fin ray. — From tip of this fin ray to its posterior insertion. 13th spinous ray in dorsal fin. — From tip of this ray to its posterior insertion. 3d anal-fin spine. — From tip of this spine to its posterior insertion. The data were divided into six geographic re- gions (Atka-Bering, Atka-Gulf, Kodiak, Yakutat, Southeastern, and Vancouver (Fig. 1)), and mea- surements regressed on SL by sex after all vari- ates were logio transformed. (Transformation al- lowed measurements to be expressed as linear functions of SL and stabilized the variance.) Geo- graphic and sexual variation were tested by anal- ysis of covariance (ANCOVA), with statistical tests "significant" when P < 0.05. In "allometric" growth in fishes, the size of one variate bears a power relationship to an index of body length. The power exponent can be ex- pressed as the slope of a simple linear regression fit to log-transformed variates, the "allometric equation", as used in this study. A slope greater than unity, "positive allometry" (Simpson et al. 1960), indicates that growth in a body dimension relative to growth in fish length (SL in the present paper) increases as fish length increases; a slope less than unity, "negative allometry", in- dicates that growth in a body dimension de- creases; and a slope of unity, "isometry", indicates that it stays the same. A preliminary survey of the measurement data disclosed that most body dimensions probably grow allometrically, although not strongly so be- cause slopes are near unity (Table 2). Because body proportions that relate size of a character to SL are specific to body length when growth is allometric, both proportions and dimensions must be referred to a size standard if either are to be objectively compared. For this reason, measure- ments and proportions used here are usually re- ferred to a hypothetical fish of 260 mm SL, a size near the average for all specimens in the study (Table 1), and likely a common collection size. In most instances, body proportions represented by estimated dimensions at the standard size need not be limited to exactly that size for purposes of general comparison because of the proximity of allometric slopes to unity (Table 2). The data 665 FISHERY BULLETIN: VOL. 85, NO. 4 Table 2. — Parameters for linear regressions (data log ,0 transformed) of measurements on standard length (mm) for Pacific ocean perch sampled from the Bering Sea (Atka-Bering region) to Vancouver Island, British Columbia. Slopes in italics are significantly different from one (P < 0.05). ANCOVA indicates significance of tests for geographic variation over the six zones by analysis of covariance (if slopes tested significantly different, intercepts were not tested). Males Females Region N Intercept Slope R2 N Intercept Slope R2 NAPE Atka-Bering 168 -0.57942 1.04767 0.986 215 -0.54925 1.03434 0.988 Atka-Gulf 196 -0.49916 1.01264 0.946 197 -0.59329 1.05137 0.964 Kodiak 150 -0.52788 1.02511 0.984 144 -0.50253 1.01401 0.989 Yakutat 66 -0.52515 1.02721 0.965 102 -0.37849 0.96596 0 960 Southeastern 103 -0.52149 1.02716 0.984 65 -0.31086 0.92727 0.983 Vancouver 45 -0.60283 1 .06479 0.957 24 -0.55937 1 .04431 0947 ANCOVA — *** — *** SPINOUS DORSAL-FIN LENGTH Atka-Bering 168 -0.46575 1 .00808 0.970 215 -0.51738 1.03007 0.984 Atka-Gulf 196 -0.48033 1.01155 0.938 197 -0.53388 1.03434 0.950 Kodiak 151 -0.40715 0.98179 0.968 144 -0.47812 1.01100 0.937 Yakutat 65 -0.41767 0.98430 0.945 103 -0.60298 1.06233 0,939 Southeastern 104 -0.65332 1.07876 0.965 66 -0.66303 1.08146 0.975 Vancouver 45 -0.18623 0.88693 0.887 24 -0.46602 1.00011 0.895 ANCOVA — • *• *** NS HIND-TRUNK DORSAL Atka-Bering 168 -0.34928 0.96296 0.977 213 -0.31407 0.94808 0.981 Atka-Gulf 196 -0.43921 1.00161 0.932 197 -0.34965 0 96475 0.958 Kodiak 151 -0.52480 1.03743 0.977 144 -0.49997 1 .02683 0.973 Yakutat 65 -0.48155 1.01525 0.965 103 -0.48653 1.01744 0.955 Southeastern 104 -0.31689 0.94900 0.976 66 -0.55759 1.05007 0.978 Vancouver 45 -0.43589 0.99412 0.900 24 -0.52323 1.02901 0.930 ANCOVA — *"• — *** HIND-TRUNK VENTRAL Atka-Bering 168 -0.36545 0.95394 0.976 215 -0.17229 0.86822 0.970 Atka-Gulf 196 -0.36803 0.95720 0.908 197 -0.15326 0.86477 0.908 Kodiak 151 -0.30535 0.93808 0.928 144 -0.14297 0.86656 0.942 Yakutat 65 -0.21169 0.89828 0.900 103 -0.10019 0.84914 0.872 Southeastern 104 -0.36519 0.96875 0.978 66 -0.30279 0.94197 0.980 Vancouver 45 -0.15815 0.88163 0.942 24 -0.06116 0.79192 0.804 ANCOVA **♦ NS •** NS BELLY Atka-Benng 168 -0.37364 0.94635 0.956 215 -0.70418 1.09162 0.965 Atka-Gulf 196 -0.33800 0.92788 0.808 197 -0.62569 1.05150 0.906 Kodiak 151 -0.60457 1.02908 0.832 144 -0.85483 1.14125 0.909 Yakutat 63 -0.97655 1.17854 0.880 97 - 1 .03204 1.20777 0.872 Southeastern 102 -0.61599 1.01693 0.903 66 -0.83303 1.11490 0.919 Vancouver 45 -0.81199 1.10045 0.802 24 -1.13734 1.24044 0.862 ANCOVA — *•* — * PELVIC INSERTION 1 Atka-Bering 168 -0.55435 1.06386 0.988 215 -0.46702 1.02686 0.988 Atka-Gulf 196 -0.51060 1.04589 0.958 197 -0.55050 1.06143 0.976 Kodiak 151 -0.43680 1.01440 0.973 144 -0.44339 1.01430 0.982 Yakutat 61 -0.40669 1.00918 0.961 97 -0.29310 0.95993 0.959 Southeastern 104 -0.40062 1.00887 0.970 66 -0.24953 0.94551 0.968 Vancouver 45 -0.53731 1.06759 0.944 24 -0.54122 1 .06600 0.926 ANCOVA — * — • *• HEAD Atka-Bering 167 -0.66929 1.08599 0.985 211 -0.54938 1.03404 0.986 Atka-Gulf 193 -0.62867 1.06805 0.956 95 -0.66243 1.08088 0.974 Kodiak 149 -0.58928 1.04964 0.974 141 -0.46700 0.99665 0.987 Yakutat 65 -0.56204 1 .04073 0.961 103 -0.38096 0.96489 0.962 Southeastern 104 -0.52223 1 .02793 0.979 65 -0.25528 0.91393 0.979 Vancouver 45 -0.45179 1 .00443 0.935 24 -0.41197 0.98543 0.935 ANCOVA — * — *'* NS = P > 0.05; "P ^ 0.05; "P ^ 0.01. '"P s 0.005. 666 QUAST: MORPHOMETRir VARIATION ON PACIFIC OCEAN PERCH Table 2.— Continued. Males Females Region N Intercept Slope R2 N Intercept Slope R2 BODY-DEPTH PELVIC Atka-Bering 168 -0.51033 1.00257 0.976 215 -0.54783 1.01989 0.980 Atka-Gulf 196 -0.53594 1.00921 0.928 197 -0.69474 1.07506 0.951 Kodiak 151 -0.52962 1 .00452 0.967 144 -0.62554 1.04313 0.982 Yakutat 65 -0.61821 1 .04588 0.961 103 -0.55126 1.01765 0.946 Southeastern 104 -0.71748 1.08660 0.978 66 -0.64909 1.05958 0.982 Vancouver 45 -0.79012 1.12032 0.951 24 - 1 .02734 1.21485 0.973 ANCOVA — *** — *** BODY-DEPTH ANAL Atka-Benng 168 -0.31899 0.88824 0.970 215 -0.23926 0.85418 0.959 Atka-Gulf 196 -0.35517 0.89788 0.903 197 -0.31260 0.87925 0.930 Kodiak 150 -0.48843 0.94820 0.949 144 -0.48666 0.94522 0.972 Yakutat 66 -0.44835 0.93799 0.930 103 -0.46838 0.94474 0.933 Southeastern 104 -0.60709 1 .00067 0.958 66 -0.69547 1 .03764 0.960 Vancouver 45 -0.45563 0.94021 0.848 24 -0.47752 0.94904 0.769 ANCOVA — •• * — **• CAUDAL PEDUNCLE Atka-Bering 168 -0.89038 0.92634 0.970 215 -0.75477 0.86857 0.964 Atka-Gulf 196 -0.96731 0.95528 0.932 197 -0.83252 0.89772 0.928 Kodiak 151 -1.03039 0.97784 0.954 144 -1.04643 0.98133 0.972 Yakutat 66 -0.90622 0.93157 0.947 103 -0.86913 0.91426 0.944 Southeastern 102 -0.98737 096392 0.958 66 -0.99718 0.97023 0.964 Vancouver 45 -0.86566 0.91611 0.818 24 -0.88816 0.92524 0.838 ANCOVA — *** — **" ORBIT Atka-Bering 168 -1.12729 1.03674 0.943 215 -0.97536 0.97242 0.949 Atka-Gulf 196 -1.02260 0.99192 0.793 197 - 1 .05490 1.00415 0.861 Kodiak 151 -1.07754 1 .00791 0.802 144 -0.83433 0.90505 0.786 Yakutat 65 -0.89752 0.92381 0.866 102 -0.71441 0.84725 0.830 Southeastern 103 -1.16845 1.03324 0.899 66 -0.83621 0.89158 0.903 Vancouver 45 -0.14594 0 62383 0.862 24 -0.22779 0.65647 0.667 ANCOVA — '** — *• LONGEST PECTORAL-FIN RAY Atka-Benng 168 -0.64359 1 .02265 0.970 215 -0.53922 0.97618 0.978 Atka-Gulf 196 -0.63481 1.01701 0.937 197 -0.51453 0.96360 0.948 Kodiak 151 -0.55655 0.98090 0.947 144 -0.44668 0.93236 0.961 Yakutat 66 -0.72067 1 04225 0.923 102 -0.61357 0.99762 0.942 Southeastern 101 -088825 1.12077 0.968 66 -0.78218 1.07402 0.956 Vancouver 45 -0.50319 0.96330 0.901 23 -0.73958 1.05622 0.934 ANCOVA — ** — «* UPPER-JAW LENGTH Atka-Bering 168 -1.07654 1.10238 0.947 215 -0.92793 1.03735 0.955 Atka-Gulf 196 -1.11966 1.12050 0854 195 -1.04190 1.08558 0.910 Kodiak 151 -0.84319 1 .00090 0.898 144 -0.65487 0.91998 0.932 Yakutat 64 -0.64407 0.91575 0.788 101 -0.57472 0.88516 0.770 Southeastern 104 -099592 1.06287 0.838 66 -0.81778 0.98171 0.819 Vancouver 45 -052744 0.86802 0.711 24 -0.16892 0.72101 0.651 ANCOVA — •'* — *** UPPER-JAW WIDTH Atka-Bering 168 -1.37815 1.03050 0.891 215 -1.38585 1 .02967 0.915 Atka-Gulf 196 -1 49225 1 07165 0.804 197 -1.41203 1 .03808 0.838 Kodiak 151 - 1 .23784 0.96166 0.776 143 -1.20719 0.94924 0.880 Yakutat 66 -1.31512 1.00107 0.844 103 -1.20986 0.95659 0.839 Southeastern 104 -1.56596 1.11473 0.880 66 -1.00614 0.87986 0.881 Vancouver 45 -1.91421 1 25943 0.775 23 -1.81392 1.21350 0.748 ANCOVA — ' — *** NS = P > 0 05: 'P s 0.05; "P ^ 0.01; "•P ^ 0.005. 667 FISHERY BULLETIN; VOL. 85, NO. 4 Table 2. — Continued. Males Females Region N Intercept Slope R2 N Intercept Slope R2 SYMPHYSEAL KNOB Atka-Bering 166 -3.09541 1.56992 0.769 215 -2.81549 1.44987 0.814 Atka-Gulf 196 -2.80011 1.43432 0.540 197 -2.87970 1.46848 0.576 Kodiak 151 -3.82404 1.87456 0.876 144 -3.23909 1.63193 0.857 Yakutat 66 -3.01114 1.52401 0.785 103 -2.54090 1.32784 0.714 Southeastern 104 -2.53152 1.32159 0.808 65 -3.23953 1.60663 0.817 Vancouver 45 -2.90945 1.47650 0.638 24 -2.42106 1.28184 0.570 ANCOVA — *** *** NS 6TH SPINOUS RAY IN DORSAL FIN Atka-Bering 167 -0.94686 1 .00886 0.882 209 -0.79653 0.94481 0.500 Atka-Gulf 194 -0.99815 1.03241 0.735 191 -1.07166 1 .05966 0.783 Kodiak 148 -0.82023 0.95342 0.836 136 -0.62671 0.86982 0.845 Yakutat 64 -1.19498 1.10184 0.818 97 -0.91257 0.98441 0.800 Southeastern 90 -1.17989 1.10021 0.894 59 - 1 .00276 1.02520 0.877 Vancouver 44 -0.64705 0.88566 0.616 19 -0.81268 0.94818 0.649 ANCOVA *•• NS — "* 13TH SPINOUS RAY IN DORSAL FIN 1 Atka-Bering 165 -0.87157 0.91366 0.804 208 -0.72181 0.85087 0.834 Atka-Gulf 192 -0.98173 0.96415 0.621 183 -1.02884 0.98308 0.730 Kodiak 141 -0.85834 0.91216 0.766 133 -0.90253 0.92803 0.780 Yakutat 63 -0.94187 0.94260 0.771 87 -0.85833 0.90897 0.744 Southeastern 95 -0.79462 0.88878 0.816 57 -0.95486 0.95753 0.783 Vancouver 42 -0.39052 0.73692 0.394 20 -1.10477 1.01967 0.638 ANCOVA **• NS • ** NS 3D ANAL-FIN SPINE Atka-Bering 147 -0.50863 0.81717 0.849 190 -0.39317 0. 76948 0.865 Atka-Gulf 166 -0.24648 0.71104 0.682 172 -0.18951 0.68697 0.696 Kodiak 131 -0.36994 0.75870 0.785 126 -0.32305 0 73849 0.818 Yakutat 62 -0.14422 0.66714 0.690 93 -0.45754 0. 79909 0.714 Southeastern 66 -0.15004 0.67029 0.061 44 -0.54744 0.82184 0.742 Vancouver 40 -H0.33516 0.46515 0.374 19 -^0.281 73 0.48651 0.451 ANCOVA — ••• •*• NS NS = P > 0.05; "P s 0 05; "P s 0 01; "•P < 0.005. agree with visual impressions of negligible pro- portional changes with growth, gained from working with the species in the field. To compare measurement variability for sexual and geographic factors in untransformed data, in the section on character discrimination, I ob- tained symmetrical estimates of variation about mean values by taking limits formed by one standard deviation (about each regression by sex within region) on each side of the transformed means. I then back-transformed the limits and halved the difference between them. This sym- metrical substitute for the asymetrical standard deviation about regression is called the "alternate standard deviation". Westrheim (1975) studied sexual maturity in Pacific ocean perch from the Gulf of Alaska and concluded that onset occurred between 195 and 260 mm SL (fork length (FL) converted to SL by relationships established from the present mor- phometric data: SL= -3.272 + 0.879 FL, where N- 1,528 and R^=0.997]. In the present study, the boundary between juveniles and adults is taken as 230 mm SL. Body Form Varied Geographically and by Sex Although later, more detailed, analyses dis- closed that the constant slopes required by the ANCOVA program BMDP P2V (Dixon et al. 1977) for each character (transformed data) were technically unmet, a preliminary analysis with this program indicated likely significant geo- graphic and sexual variation in all characters (subsequently confirmed in detailed analyses with a more general ANCOVA model. Table 2), and a general lack of interaction between these factors. Also, I determined that only negligible bias has been induced in the regressions by loga- rithmic transformation of data by computing mean response variables, at 260 mm SL, for un- transformed measurements fit by nonlinear least squares (BMDP PAR, Dixon et al. 1977). For six 668 QUAST: MORPHOMETRIC VARIATION ON PACIFIC OCEAN PERCH measurements, by sex and region, over which the two methods were compared, the prediction of character sizes in fish of 260 mm SL by nonUnear least squares was close to the mean and within the 957c confidence interval for prediction from transformed data (Fig. 2) and did not change the results. Graphical comparison of character measure- ments, related to the standardized fish of 260 mm SL, also disclosed that geographic and sexual variation were frequent: Nape (Fig. 2A). — Geographic variation in dis- tance between tip of snout and the dorsal fin was significant. Napes averaged shortest in the Atka-Gulf and Kodiak regions and lengthened in a cline to the Vancouver region. Females averaged smaller napes than males in all re- gions. Spinous dorsal-fin length (Fig. 2B). — Geographic variation in length of spinous dorsal fin was significant, and the fin shortened in a cline from the Bering Sea to the Southeastern and Vancouver regions. Because geographic varia- tion in length of the spinous dorsal fin was nearly opposite to geographic variation in the nape, changes in position of the anterior fin insertion probably caused the reciprocal clines. Sexual dimorphism was not important. Hind-trunk dorsal (Fig. 2C). — Distance between the spinous dorsal fin and the tail changed sig- nificantly between regions. It averaged largest in specimens from the middle regions (Atka- Gulf and Kodiak) and smallest in the Vancou- ver region. Sexual dimorphism was not impor- tant. Hind-trunk ventral (Fig. 2D). — Distance between the second spine of the anal fin and the tail changed significantly between regions. It aver- aged largest in the Southeastern and Vancou- ver regions and shortest in the Atka-Bering re- gion. The data formed a geographic cline opposite to that of the belly measurement. From the Yakutat region westward, females averaged significantly shorter in this measure- ment than males. Belly (Fig. 2E). — Distance between the pelvic fins and anal-fin spines varied significantly geo- graphically and decreased in a cline from northwest to southeast (Atka-Bering region to the Southeastern and Vancouver regions). Belly measurements averaged about 1.6 cm smaller in the southeastern extreme of the sampling range than in the northwestern. The cline apparently is caused by opposing rela- tional movements of pelvic girdle and anal-fin spines along the body axis because the pelvic insertion and hind-trunk ventral measure- ments decreased from southeast to northwest. Sexual dimorphism was significant, with bel- lies of males averaging about 4 mm smaller than those of females. Pelvic insertion (Fig. 2F). — Geographic variation in distance between snout and pelvic fins was significant, and the distance increased from northwest to southeast, from the Atka-Bering region to the Vancouver region. The measure- ment averaged shorter in females than males in all regions, evidence for significant sexual dimorphism. Head (Fig. 2G). — Geographic variation in head length was significant, and heads averaged shortest in the Kodiak region and longest in the Vancouver region. Sexual dimorphism was usually significant, and females averaged smaller heads than males. Trends in geo- graphic and sexual variation between the head and pelvic insertion were similar, probably be- cause both measurements include similar re- gions of the head. Body -depth pelvic (Fig. 2H). — Geographic varia- tion in body depth at the pelvic fins was signif- icant and formed a broken cline. Deepest bodies occurred at the extremes of the sampling range (Atka-Bering and Vancouver regions), and were shallowest in the Kodiak region. Sexual dimorphism was inconsistent. Body-depth anal (Fig. 21). — Geographic variation in body depth at the anal spines was signifi- cant, and depth was shallowest in Kodiak spec- imens and deepest in Atka-Bering specimens. Body depths averaged smaller in females than in males from the Yakutat region to the Atka- Bering region, but the differences may not be significant. Caudal peduncle (Fig. 2J). — Geographic varia- tion in depth of caudal peduncle was signifi- cant, but important geographic differences were limited to regions west of Yakutat. Speci- mens from the Kodiak region averaged narrow- est peduncles, and specimens from the Atka- Bering, Southeastern, and Vancouver regions averaged widest peduncles. Caudal peduncles averaged significantly narrower in females than in males except in the Southeastern and Vancouver regions. Orbit (Fig. 2K). — Geographic variation in orbit diameter was significant. Diameters were 669 FISHERY BULLETIN: VOL. 85, NO. 4 REGION 94- 93- 92 91 - 90 88 E89 E CC LU H O < IT < X o o 111 t^96 CO ATKA-B|ATKA-G| KOD , YAK . SE . VAN ATK A-B , ATK A-G, KOD , YAK , SE , VAN ♦ < O-O 0. >4.NAPE o LU F, 95- LLI DC a 94 93 92 91 90 -s- ■-»- •-- ■o o C. HIND-TRUNK DORSAL •--4 o o 88.97 94 93 92 91 90 89 88 96- 95- 94- 93- 92- 91 - 90- 89- 88- 87- 86- 85- 84- 83- ■O ■9- ■o -o-i -o B. SPINOUS DORSAL -FIN LENGTH [sS-SO O -G ■O O -o -o D. HIND-TRUNK VENTRAL Figure 2. — Geographic and sexual variation in measurements of Pacific ocean perch in the northeastern Pacific Ocean and eastern Bering Sea as represented by mean responses to regression functions (Table 2) (circles) and 95% confidence intervals (brackets) for these responses as related to a standard-sized fish of 260 mm SL. Solid symbols represent males and open symbols females. Triangles represent mean responses by nonlinear least squares analysis (Dixon et al. 1977) for 260 mm SL fish. 670 QUAST: MORPHOMETRIC VARIATION ON PACIFIC OCEAN PERCH REGION ATKA-B ,ATKA-G, KOD YAK SE VAN 86- 85- 84- 83- 82- 81- 80- 79- 78- n - 76- £74k ^73h ir ^72h O LLI N 70|- W 96 95 94 93 92- 91 90 89 88 87 .c] OO » 4 O o ■o 0 o t> E. BELLY -o. -o-" [o [. -O" G. HEAD -o o Figure 2.— Continued. largest in the Atka-Bering and Atka-Gulf re- gions and smallest in the Southeastern region. Diameters decreased continuously from the Atka-Bering region to the Southeastern region, but the trend was broken by Vancouver sam- ples. Although not significant within regions, ATKA-B ,ATKA-G, KOD YAK SE , VAN 114 1 13 112 1 1 1 110 109 108 107 106 105 104 103 102 101 100 99 98 97 83 82 81 - 80 79 78 O o o -[• F. PELVIC INSERTION t>-o •^ * !>-§..< c>-o o •^ < [> H. BODY- CH DEPTH PELVIC o -o-* sexual differences in orbit diameter probably were significant overall because females aver- aged smaller orbits than males in all regions. Longest pectoral-fin ray (Fig. 2L). — Geographic variation in length of pectoral fins was signifi- cant. The fins averaged longest at the eastern 671 FISHERY BULLETIN: VOL. 85, NO. 4 REGION ATKA-B|ATKA-G| KOD i YAK i SE i VAN ATKA-B |ATKA-G| KOD , YAK , SE , VAN 67 66 65 64 'i s cc HJ 63 I- o < cc < X o 68.84 O-l •o- o I. BODY-DEPTH r ANAL -O 61.79-11 111 N W Q LJJ I- O Q LU CC a. 25 24 23- 22- 21 20 T^ i^} l4 ^■'?, •--* K. ORBIT 24- 23 22 - [o- 21- 20 67 66 65 64 63 62 E«" ♦3 l^ ■-•-1 o- J. CAUDAL PEDUNCLE o o o o L. LONGEST PECTORAL- FIN RAY -o Figure 2.— Continued. and western extremes of the sampling range and shortest in the Yakutat region. In all re- gions but Yakutat, sexual dimorphism was high and significant, and fins in females aver- aged >1 mm shorter than in males. Upper -jaw length (Fig. 2M). — Geographic varia- tion was significant, and jaws averaged longest in eastern Aleutian samples (Atka-Bering and Atka-Gulf regions). Sexual dimorphism was significant in three regions (Atka-Bering, Atka-Gulf, and Southeastern), with females av- eraging shorter upper jaws than males in each. Upper -jaw width (Fig. 2N). — Geographic varia- tion was significant, with upper jaws averaging narrowest in the Kodiak region. Sexual dimor- phism was important in only the Atka-Bering and Southeastern regions, where females aver- aged narrower upper jaws than males. Symphyseal knob (Fig. 20). — Geographic varia- tion was significant but erratic. On average, specimens from the Atka-Gulf region probably have the largest symphyseal knobs. Sexual di- morphism seems unimportant. 6th spinous ray in dorsal fin (Fig. 2P). — Geo- 672 QUAST: MORPHOMETRIC VARIATION ON PACIFIC OCEAN PERCH ATKA-B , ATKA-G| KOD YAK SE REGION VAN ATKA-B, ATKA-G, KOD , YAK 39 38 37 36 'i B DC ^ 35 O < cr < I 0 33 LU N W 32 Q LU O Q m 31 cr Q. O -O o- A7. UPPER-JAW LENGTH O o 30 29 28 O -O o o ^■9- t> -o [> o O. SYMPHYSEAL KNOB SE VAN 13 12 1 1 5.3 5.2 5.1 5.0 4.9 4.8 4.7 4.5 4.4 4.3 4.2 4.1 f* w ,, E*3 N. UPPER-JAW WIDTH P -O o o --0 p. 6**^ SPINOUS RAY IN DORSAL FIN •o Figure 2.— Continued. graphic variation was significant but erratic in this index to height of the spinous dorsal fin. Kodiak specimens had the highest spinous dor- sal fins, on average. Sexual variation was in- consistent and is probably unimportant. 13th spinous ray in dorsal fin (Fig. 2Q). — Geo- graphic variation was significant but erratic in this index to height of the notch between spinous and soft dorsal fins. Sexual dimorphism seems generally unimportant. 3d anal-fin spine (Fig. 2R). — Geographic and sex- ual variation in length of the spinous ray was minor, except that the fin spine was unusual- ly short in females from the Southeastern region. The measurements (as related to the standard- ized fish of 260 mm SL) usually varied geograph- ically either in generally monotonic clines over the study area (Atka-Bering to Vancouver re- gions) or V-shaped clines that were broken in the Yakutat or Kodiak region. Only two sets of char- acters varied almost monotonically, and varia- tion within each can be ascribed to a progressive shift in boundary features for body regions: Length of nape generally decreased and length of the spinous-dorsal fin increased from southeast to northwest (Fig. 2A, B), probably because of a rela- tional shift in the dorsal-fin insertion. Belly size increased from southeast to northwest (Fig. 2E), probably because the pectoral girdle and associ- 673 FISHERY BULLETIN: VOL. 85, NO. 4 ATKA-B, ATKA-G, KOD , YAK , SE REGION VAN ATKA-B, ATKA-G, KOD , YAK , SE , VAN Figure 2.— Continued. ated pelvic fins and the complex of anal-fin spines moved relationally apart (Fig. 2F, D). In contrast, depth of the entire body of Pacific ocean perch varied similarly geographically because the three measures of body depth — body-depth pelvic, body-depth anal, and caudal peduncle — varied concordantly (Fig. 3). Sexual dimorphism was significant in most measurements, and except for belly size, mea- surements usually averaged larger in males than in females. The combined effects of geographic and sexual variation meant that belly measure- ments averaged about 16 mm larger in standard- sized females from the Atka-Bering region than in males from the Southeastern or Vancouver re- gion (Fig. 2E). Slopes in over one-half of the measurement re- gressions differed significantly from unity (Table 2), indicating growth allometry, particularly when differences from unity were consistent. The symphyseal knob was the only character with strong positive allometry (Table 2). Only two characters were strongly or consistently nega- tively allometric (Table 2): hind-trunk ventral, significant in 11 of 12 sex/region cells (slopes av- eraging 0.94 in males and 0.87 in females); and length of 3d anal-fin spine, significant in all 12 sex/region cells (slopes averaging 0.72 in males and 0.74 in females). With growth, the trunk pos- terior to the anal-fin spines becomes proportion- ally smaller relative to the rest of the body be- cause of negative allometry in the body-depth + 80- 60- 40- 20 O z < I o I- 0 z m a. 20 m Q. - 40 60- 80- O I < CD ^ i< H O < (- o I < < Q o o < Q o < > < o > I- T z < > uj o « I- T BODY-DEPTH PELVIC BODY-DEPTH ANAL CAUDAL PEDUNCLE J_ Figure 3. — Geographic variation in depth measurements of three body characters in Pacific ocean perch. Character mea- surements were related to a standard-sized fish of 260 mm SL from regressions in Table 2, then back transformed. Because measurements for males and females were usually different (Fig. 2), midvalues between sexes were used. Percentage change is the change in a measurement between neighboring regions as a percentage of the measurement's range over the geographic range. anal, hind-trunk ventral, and caudal peduncle measurements. 674 QUAST: MORPHOMETRIC VARIATION ON PACIFIC OCEAN PERCH Inflections and Sexual Crossover in the Regressions I examined 216 computer-drawn scattergrams for sexes within regions that represented the measurement regressions (Table 2), by transpar- ent overlay with incised straight line. None shows obvious curvature, and only three show possible inflections: In females, the regression ap- pears to bend upward about 8.3° at about 262 mm SL for belly measurements in the Atka-Bering region; downward about 7.0° at about 242 mm SL for hind-trunk ventral in the same region; and possibly upward about 13.5° at about 260 mm SL for the belly in the Southeastern region. Whether the apparent breaks in the three regressions are artifacts or real is moot. Evidence for reality in- cludes their visibility in scattergrams; that all occur in a single sex (females) and in a measure- ment possibly influenced by sexual development with growth; and that the possible break in the belly regression in the Atka-Bering region has a near complement in the hind-trunk ventral mea- surements, which extend from the belly to the tail. Also, as discussed in a succeeding paragraph, mild inflections probably are hidden in variabil- ity of the data. Evidence against reality includes the extreme rarity of visible indications of possi- ble breaks (none in 16 of 18 measurements and only 3 in the remaining 24 regressions), and the extreme goodness of fit (i?^) to a straight line shown by the regressions with possible inflections (Table 2). The measures for goodness of fit for belly are the best among regions in the two re- gressions with possible inflections, and both mea- sures are higher than those for corresponding male regressions, which show no indications of inflection. The weight of evidence seems to side with the visible breaks being artifacts. Yet, even if the breaks represent real inflections, the mor- phometric data seem to fit linear criteria well as far as conventional measures are concerned. The regressions for sexes, within measure- ments and regions (Table 2), diverge with in- creasing SL, indicating that sexual dimorphism increases with growth. The increases are slight, however, as evidenced by the similarity in regres- sion parameters between sexes, and on average, characters will differ in size between sexes by only a few millimeters in the standard 260 mm SL fish (Fig. 2). Given normal variation, the differ- ences should not be obvious or reliable in differen- tiating sexes by gross examination of even the largest Pacific ocean perch. However, divergent sexual-regression pairs pose an apparent contradiction when they inter- sect within their data domain. Such intersections infer 1) sexual differences in fish on the juvenile side of the intersection, 2) differences on the juve- nile side the reverse of those on the adult side, and 3) differences between sexes in juveniles that increase as SL's become smaller. In the morpho- metric data, the regressions for sexes do intersect, and the intersections form a symmetrical unimo- dal distribution with a strong peak near 230 mm SL. By itself, crossover need not be a problem; e.g., the symmetrical confidence limits about re- gression (Sokal and Rohlf 1969) are evidence that crossover within these limits is normal in sam- ples from a single population. Yet, differences in measurements for sexes on the juvenile side of the modal point for sexual crossovers apparently are greater than can be accommodated by confidence limits based on the regressions. In 108 compari- sons of mean estimates for measurements and their confidence limits at 170 mm SL, 18 have sig- nificant differences, when only about 5 are expected with 95% confidence limits. Apparently, sexually associated crossover is slightly too severe in the morphometric data to be adequately contained by confidence limits for single populations. The evidence is strong that most of the mea- surement regressions do not fit their data per- fectly and, because of the nature of the error, that slight growth inflections likely are concealed by variation. The three apparent inflections men- tioned previously might indicate the size at in- flection, but the evidence is weak. It is apparent, however, that significant sexual differences in measurements at SL's below the crossover mode should not be taken literally. (Assuming that measurement regressions may incorporate hid- den inflections, most reliable estimates of juve- nile measurements among sexual pairs within their data domains will be from the regression whose slope is nearest unity, particularly if the sample number is large and the slope not signifi- cantly different from unity.) Because the bulk of specimens were larger than 230 mm SL, and the regressions fit their data closely, conclusions re- garding adult relationships in the present data, particularly trends, should be reliable. Strength of Geographic and Sexual Variation Shown in the Characters Analysis of morphological diversity is most use- 675 FISHERY BULLETIN: VOL. 85, NO. 4 ful where assigned components of variation can be maximized relative to unassigned components. I compared indices of variation assignable to mea- surement size (standard specimens), geographic, and sexual causes in the 18 characters by means of the alternate standard deviation (see Methods). First, I investigated relationships between varia- tion and the size of a character. When mean re- sponses for measurements in standard-sized fish from the Kodiak region were used as a basis of comparison, variability was positively related to size of the character, with the two variates signif- icantly correlated (Fig. 4). Size accounted for about 62% iR'^) of the varia- tion between characters. To remove its effect in further comparisons, I used an alternate version of the coefficient of variation; i.e., the alternate standard deviation for each character divided by a size index (Kodiak) for that character xlOO. Although the alternate coefficient of variation re- duced the unassigned variability, some sizable differences remained between characters (Table 3) — major dimensions of the head and dorsum varied least, fin spines and small features of the head varied most, and the symphyseal knob varied considerably more than any other charac- ter. With the exception of belly, the major trunk Table 3. — Evidence for unassigned variation in measurements in Pacific ocean percfi after correction for measurement size. Mea- surements (Y) are related to a standard-sized fish of 260 mm SL from the Kodiak region by regressions in Table 2. Relative variation for each measurement is indexed by the alternate coefficient of variation (ACV) (see text) pooled over all regions and expressed as a percentage. Measurements arranged by increasing alternate co- efficient of variation in the last column. Males Females Both s Y, mm exes Measurement Y, mm ACV Y, mm ACV ACV Nape 88.66 2.50 88.36 2.29 88.52 2.40 Head 88.22 3.32 87.07 2.53 87.66 2.94 Pelvic insertion 103.03 3.20 101.42 3.01 102.24 3.11 Body-depth pelvic 78.75 3.53 78.27 3.04 78.52 3.29 Hind-trunk dorsal 95.62 3.01 95.46 3.72 95.54 3.36 Spinous dorsal-fin length 92.01 3.43 91.92 3.63 91.97 3.53 Body-depth anal 63.31 4.19 62.52 3.51 62.92 3.86 Caudal peduncle 21.43 4.08 21.06 3.65 21.25 3.87 Longest pectoral-fin ray 64.91 4.43 63.82 4.10 64.38 4.27 Hind-trunk ventral 91.22 4.98 89.07 4.69 90.07 4.84 Upper-jaw length 36.60 6.43 36.89 5.43 36.74 5.94 3d anal-fin spine 28.99 6.68 28.87 6.61 28.93 6.64 6th spinous ray in dorsal fin 30.36 7.92 29.78 7.58 30.08 7.76 Belly 75.97 8.82 79.66 7.87 77.77 8.36 Upper-jaw width 12.15 9.86 12.17 7.54 12.16 8.73 13th spinous ray in dorsal fin 22.11 9.45 21.81 10.05 21.96 9.74 Orbit 22.73 9.57 22.46 10.32 22.60 9.94 Symphyseal knob 5.05 13.51 5.04 14.59 5.04 14.04 5.0 4.0 Y=1 . 1126 + 0. 0246X N = 18, r = 0.787* • > 111 Q Q ir < Q 3.0 2.0 - • • w 1.0- _L _L J_ P<0.01 J_ _L _L 10 20 30 40 50 60 70 80 90 100 110 120 AVERAGE SIZE OF MEASUREMENT Figure 4. — Relationship between absolute size and variability of measurements in Pacific ocean perch. Character sizes are mean responses for hypothetical 260 mm SL fish from the Kodiak region as estimated from the regressions of Table 2; variability is estimated by alternate standard deviations for each character (see Methods) pooled over all regions. 676 QUAST: MORPHOMETRIC VARIATION ON PACIFIC OCEAN PERCH dimensions had the lowest unassigned variabiHty when corrected for size. As would be expected from the plots of charac- ter measurements (Fig. 2), sexual and geographic variation were important components after the size effects were removed (Table 4). For sexual variation, the hind-trunk dorsal, spinous-dorsal fin, and body-depth pelvic measurements were the poorest discriminators, and belly, hind-trunk ventral, and longest pectoral-fin ray the best. For geographic variation, 3d anal-fin spine, symphy- seal knob, and 6th spinous ray in dorsal fin were the poorest discriminators, and belly, hind-trunk ventral, and head measurements the best. Present Nominal Subspecies are Questionable Barsukov (1964) synonymized the nominal spe- cies Sebastodes alutus (Gilbert) and S. pau- cispinosus (Matsubara) but suggested that the eastern and western Pacific representatives may be separate subspecies: S. a. paucispinosus ranging from Honshu to Olyutorskii Bay and along the northern Bering Sea slope, perhaps to Bristol Bay; and S. a. alutus ranging from Cali- fornia to the Gulf of Alaska and along the Aleu- tian Arc to, and including, the Commander Is- lands. Barsukov morphologically distinguished the subspecies by "Alaskan Seb. alutus longer than 23 cm are quite noticeably distinguished from Seb. alutus from other parts of the range by body depth", and in a key gives the principal sub- species discriminator as whether the ratio SL/ body depth of 170-360 mm SL fish is greater than 3.2 iS. a. alutus) or less (S. a. paucispinosus). The question of eastern and western subspecies in S. alutus seems more complex than Barsukov (1964) suggested. The weight of present evidence, although preliminary, does not seem to justify the nominal subspecies. First, there is the problem of how populations sympatric over a distance as great as the Aleutian Arc could maintain repro- ductive isolation adequate to insure genetic dis- tinctiveness. There is no evidence for isolating mechanisms in the species — on the contrary, the larvae are pelagic (Hart 1973), which should pro- mote rapid genetic exchange over major distances Table 4. — Relative degree that measurements in standard-sized Pacific ocean perch of 260 mm SL reveal geographic and sexual variation. In the variation sections, variation not related to character size was indexed by alternate coefficients of variation that were pooled over all samples, sexual variation was indexed by differences between sexes as a percentage of their mid-size in each region (positive if males averaged larger than females and negative if males were smaller, but only absolute values were used in calculations), and geographic variation was indexed by the maximum difference between regions as a percentage of mean-estimate size of measurements from the Kodiak region (Table 3). In the discnmination section, indices indicate relative magnitudes of sexual and geographic variation relative to variation not related to character size. Low values or ranks (in parentheses) indicate that variation not related to character size was high relative to sexual (data column 4) or geographic (data column 5) variation; hence, the measurement is a poor indicator of sexual or geographic variation, and vice versa. Kendall Coefficient of Concordance (Siegel 1956) between ranks for sexual and geographic vanation was not significant. Variation not related Relative > i/ariation Discrimination Geo- (Sexual ^ Geographic h- Character to character size Sexual graphic col. 1) X 10 column 1 Nape 2.91 (1) 0.493 (4) 4.60 (7) 1.69 (10) 1.58 (13) Spinous dorsal-fin length 3.73 (6) (-)0.088 (2) 4.39 (5) 0.24 (2) 1.18 (9) Hind-trunk dorsal 3.35 (4) (-)0.019 (1) 3.80 (2) 0.06 (1) 1.13 (7.5) Hind-trunk ventral 4.05 (9) 2.038 (17) 9.57 (15) 5.03 (17) 2.36 (17) Belly 6.45 (11) (-)3.873 (18) 16.47 (18) 6.00 (18) 2.55 (18) Pelvic insertion 3.08 (3) 0.875 (11) 6.53 (12) 2.84 (14) 2.12 (15) Head 3.07 (2) 1.104 (13) 6.69 (13) 3.60 (15) 2.18 (16) Body-depth pelvic 3.63 (5) (-)0.104 (3) 4.32 (4) 0.29 (3) 1.19 (10) Body-depth anal 4.21 (10) 0.628 (8) 6.19 (11) 1.49 (9) 1.47 (11) Caudal peduncle 4.02 (7.5) 0.828 (9) 4.53 (6) 2.06 (13) 1.13 (7.5) Orbit 7.01 (13) 0.503 (5) 13.39 (17) 0.72 (6) 1.91 (14) Longest pectoral-fin ray 4.02 (7.5) 1.663 (16) 6.03 (10) 4.14 (16) 1.50 (12) Upper-jaw length 6.68 (12) 1.305 (14) 4.93 (8) 1.95 (12) 0.74 (4) Upper-jaw width 8.09 (15) 0.970 (12) 9.03 (14) 1.20 (8) 1.12 (6) Symphyseal knob 15.49 (18) 0.873 (10) 11.29 (16) 0.56 (4) 0.73 (2.5) 6th spinous ray in dorsal fin 8.21 (16) 1.410 (15) 5.97 (9) 1.72 (11) 0.73 (2.5) 13th spinous ray in dorsal fin 9.54 (17) 0.589 (6) 0.94 (1) 0.62 (5) 0.99 (5) 3d anal-fin spine 7.31 (14) 0.602 (7) 3.87 (3) 0.82 (7) 0.53 (1) 677 FISHERY BULLETIN: VOL. 85, NO. 4 and erase local differentiation. Further, the spe- cies appears to be genetically nearly homoge- neous over as great, and environmentally vari- able, a distance in the eastern part of its range — Seeb and Gunderson (in press) demon- strated "very high similarity" among populations from Washington State to the Bering Sea and found no evidence for a barrier at the Aleutian Chain. Second, as already mentioned, morpholog- ical (including morphometric) differences can be environmentally induced through modification of growth rate. The growth differences Quast (fn. 2) found, which appeared to conform more closely to latitude than ocean temperatures, resemble geo- graphic trends in some measurements examined in the present study, including body-depth pelvic (Fig. 2H). Last, Barsukov's criteria for the sub- species can be called into question. When referred to fish of two standard sizes, 260 and 300 mm SL, my data on body-depth pelvic (Barsukov's "body depth") indicate that only representatives from the Kodiak region have 95% confidence limits for mean population values that lie consistently on the S. a. alutus (the nominal eastern subspecies) side of Barsukov's criterion-confidence limits for most measurement means for other regions indicate ambiguous or improper identification (Table 5), and that a majority of specimens will be improperly identified. Further, neighboring populations in the eastern subspecies' range frequently differ significantly in one or more measurements (Fig. 2, Table 2). If significant geographic variation were a sole criterion for subspecies then a number might need be named. Barsukov (1964) gave further criteria for sepa- rating the nominal eastern and western subspe- cies, but the criteria seem subjective and imprac- tical: Prominence and apparent squamation of occipital crests do not seem of value; as I have observed, development of crests may be highly variable within regions, and evidence for squa- mation can be altered in specimens collected by bottom trawl. His analysis of variation in the oc- cipital crests is too short and subjective to be use- ful. Size of symphyseal knob ("larger at similar body lengths in the eastern subspecies") is not reliable because the character has considerable Table 5. — Mean, upper, and lower 95% confidence limits for the mean of body-deptfi pelvic (BDP) measurements and their derived proportions of standard length (SL) for Pacific ocean perch of 260 and 300 mm SL. Ratios of SL divided by body depth are categorized according to Barsukov's (1964) criterion of 3.2 for the ratio as follows: Ratios rounding to greater than the interval 3.15-3.24 (3.2 expanded to its inclusive values with two decimal points) are followed by a blank, those within the interval are followed by an "A", and those lower are followed by an 1". Ratios followed by a blank would identify the eastern nominal subspecies (S. a. alutus) by the 3.2 criterion, those followed by an "A" would give an ambiguous identification (S. a. alutus or S. a. paucispinosus), and those followed by an '!" would identify the western nominal subspecies (S. a. paucispinosus). 260 mm SL 300 mm SL 1 \^ales Females Males Females Region (limit) BDP SL7BDP BDP SLyBDP BDP SL/BDP BDP SL/BDP Atka-Bering Lower 81.909 3.17 A 82.665 3.15 A 94.715 3.17 A 95.717 3.13 1 Mean 81.422 3.19 A 82.258 3.16 A 94.006 3.19 A 95.183 3.15 A Upper 80.957 3.21 A 81.853 3.18 A 93.302 3.22 A 94.652 3.17 A Atka-Gulf Lower 80.115 3.25 80.145 3.24 A 92.859 3.23 A 93.687 3.20 A Mean 79.666 3.26 79.708 3.26 92.044 3.26 92.964 3.23 A Upper 79.219 3.28 79.274 3.28 91 .236 3.29 92.247 3.25 Kodiak Lower 79.214 3.28 78.662 3.31 91.513 3.28 91.554 3.28 Mean 78.753 3.30 78.269 3.32 90.928 3.30 90.870 3.30 Upper 78.295 3.32 77.878 3.34 90.346 3.32 90.389 3.32 Yakutat Lower 81.560 3.19 A 81.271 3.20 A 94.997 3.16 A 94.278 3.18 A Mean 80.828 3.22 A 80.600 3.23 A 93.877 3.20 A 93.236 3.22 A Upper 80.102 3.25 79.934 3.25 92.770 3.23 A 92.186 3.25 Southeastern Lower 81.125 3.20 A 81.878 3.18 A 94.906 3.16 A 95.531 3.14 1 Mean 80.655 3.22 A 81.239 3.20 A 94.224 3.18 A 94.540 3.17 A Upper 80.187 3.24 A 80.605 3.23 A 93.547 3.21 A 93.559 3.21 A Vancouver Lower 83.391 3.12 1 82.356 3.16 A 97.376 3.08 1 97.027 3.09 1 Mean 82.304 3.16 A 80.629 3.22 A 96.616 3.11 1 95.938 3.13 1 Upper 81.231 3.20 A 78.958 3.29 95.861 3.13 1 94.861 3.16 A 678 QUAST: MORPHOMETRIC VARIATION ON PACIFIC OCEAN PERCH unassigned variation for its size (Table 3), as well as high positive allometry (Table 2). The possibility exists, since Barsukov (1964) measured body depth directly on preserved speci- mens, a method different from that used in the present paper, that the two methods give biased measurements relative to the other. The question cannot be fully resolved; Barsukov gave sparse collection information (e.g., his conclusions on Bristol Bay representatives were based on eight or fewer specimens between 30 and 340 mm SL), and he gave no data on statistical parameters or data peculiarities. Although the body depth mea- surement at pelvic fins is simple to perform, high accuracy and undistorted material are necessary because geographic variation is slight but signifi- cant (e.g., maximum geographic difference be- tween means for body-depth pelvic at 260 mm SL is around 4 mm in Figure 2H). Indirect evidence indicates that the combina- tion of photogrammetry and fresh specimens used in the present study probably gave more precise measurements than the hand methods and mu- seum specimens used by Barsukov (1964), but likely that bias between methods was unimpor- tant relative to other factors. Barsukov stated that body depth in specimens attributed by him to S. a. paucispinosus, presumably including those from Bristol Bay, averages 3.05 into SL, and that his specimens of iS. a. alutus average 3.42. In con- trast, in the present study, the extreme regional confidence limits for means lie between 3.08 and 3.34 (Table 5), well within the span of Barsukov's means (my data for Bristol Bay are nearly central between his values, with confidence limits of 3.12 and 3.22). Rather than methods bias, the wide range of mean values for body depth given by Barsukov (1964) relative to those in Table 5 may have been caused in part by chance overweighting of ex- treme data values because of his relatively small sample sizes. His 124 specimens were relatively few for a considerable geographic range — 82 from Bristol Bay to Washington and 42 from Olyu- torskii Bay and the Commander Islands. Perhaps, body-depth variation was falsely indicated as bi- modal in Aleutian Arc representatives, leading to the interpretation that the data represented shallow- and deep-bodied populations. Finally, Barsukov may have been misled by variable distortion and shrinkage of his speci- mens owing to conditions of preservation and storage. Although he stated that his specimens shortened 0.3-4.0% after "several" months of preservation in alcohol, and that 200 mm SL fish lost 1% and 300 mm fish lost 2% on average, he apparently did not try to compensate for this loss in length and apparently did not measure corre- sponding changes in body depth at the pelvic fins. Some of his material had been preserved much longer than several months and may have been even less representative of fresh material — the Olyutorskii Bay and Commander Island speci- mens were collected by A. P. Andriyashev in 1932 and 1950-52, indicating probable 9-30 yr storage in alcohol before measurement. CONCLUSIONS Because geographic variation was expressed in all parts of the morphology of Pacific ocean perch that I investigated, I conclude that the variation pervades body growth in the species. Over the eastern Bering Sea and eastern Pacific Ocean, adult measurements usually vary as V-shaped clines. Here, representatives of the same SL from the extremes of the sampling range (Vancouver Island and the eastern Bering Sea) resemble each other more than they resemble fish from near the midrange (Kodiak and Yakutat regions), where measurements often are smallest. Only measurements of belly size and neighbor- ing parts of the body have single, monotonic clines over the regions. Belly size increases dra- matically from Vancouver Island to the eastern Bering Sea accompanied by corresponding size decreases in neighboring body measurements. The anterior and posterior boundaries of the belly, pelvic girdle (given by pelvic insertion), and anal-fin spines (given by body-depth anal) move relationally farther apart to give progressively larger bellies in populations farther from Vancou- ver Island and closer to the Bering Sea. Length of the spinous dorsal fin generally increases from southeast to northwest and length of the nape decreases, both apparently because of a relation- ally forward shift in the anterior insertion of the dorsal fin. Nearly all morphometric characters apparently grow allometrically in Pacific ocean perch, but average body form does not change markedly with growth because allometric coefficients of most characters are near unity. Often, measure- ments vary between apparent slight but signifi- cant positive and negative allometry, depending on the sampling region. The symphyseal knob and 3d anal-fin spine (allometric coefficients were 1.52 and 0.73, respectively, including both sexes) 679 FISHERY BULLETIN: VOL. 85, NO. 4 may be sufficiently allometric for the symphyseal knob to seem exceptionally prominent and the 3d anal-fin spine exceptionally small in some large specimens. Most body dimensions average smaller in fe- males than in males, but bellies are larger in females. The same mechanism, opposing rela- tional movements of the pelvic girdle and the anal-fin spines, is responsible for sexual and geo- graphic variation in belly size. Although sexual dimorphism increases with growth, sexual differ- ences are not prominent and have broadly over- lapping distributions. As a result, sexes probably cannot be reliably identified by gross examina- tion of any of the 18 characters. Neither geographic nor sexual variation is quantitatively similar between measurements in Pacific ocean perch — characters tend to vary with their absolute size. Overall, geographic variation is most poorly discriminated by the measure- ments of the 3d anal-fin spine, symphyseal knob, and 6th spinous ray in dorsal fin, and best dis- criminated by the belly, hind-trunk ventral, and head. Sexual dimorphism is most poorly discrimi- nated by the measurements of the hind-trunk dorsal, spinous dorsal-fin length, and body-depth pelvic, and best discriminated by the belly, hind- trunk ventral, and pelvic insertion. Measure- ments that have been used for taxonomy of Pacific ocean perch in the past are relatively poor discriminators of geographic variation or possible genetic stocks or subspecies. Although body form changes significantly with geographic region, sex, and growth, differences are too small and unexplained variation too large for differences to be of value for distinguishing single specimens geographically. Because of questions concerning validity and importance of published morphological informa- tion supporting supposed subspecies of Pacific ocean perch, it seems prudent that further claims for subspecies based on morphology be postponed until variation is reliably assessed over the entire species' range and definitive characteristics are known to be genetically based. ACKNOWLEDGMENTS My special thanks to Elizabeth Lu Hall for her exacting work in converting photographic images into data and to Richard E. Haight and Richard H. Carlson, also of the Auke Bay Laboratory, for assistance in photographing the specimens. Many of the photographs were taken aboard the Univer- sity of Hokkaido Faculty of Fisheries' RV Oshoru Maru and through the kindness of her captain, T. Fugii. My thanks also to Jergen Westrheim of the Nanaimo Laboratory and Fisheries and Oceans Canada, for permission to photograph specimens aboard the RV G. B. Reed. Evan B. Haynes and Bruce L. Wing of the Auke Bay Laboratory pro- vided helpful reviews of the manuscript. LITERATURE CITED Barsukov, V V. 1964. Intraspecific variability in the Pacific rockfish (Se- bastodes alutus) Gilbert. In P. A. Moiseev (editor), So- viet fisheries investigations in the northeast Pacific, vol. 49, p. 241-267. Tr. Vses. Naucho-Issled. Morsk. Rybn. Khoz. OkeanogT., Part II. (Translated by Israel Pro- gram for Scientific Translations, 1968.) Chen, L -C 1971. Systematics, variation, distribution, and biology of rockfishes of the subgenus Sebastomus (Pices, Scor- paenidae, Sebastes). Bull. Scripps Inst. Oceanogr., Univ. Calif 18, 115 p. Dixon, W J (ed.), M B Brown (ed.), L Engelman, J W. Frane, AND R I Jennrich 1977. BMDP-77 Biomedical Computer Programs, P- Series. Univ. of Calif Press, Berkeley, 880 p. Hart, J. L 1973. Pacific fishes of Canada. Bull. Fish. Res. Board Can. 180, 740 p. HUBBS, C L , AND K F Lagler 1949. Fishes of the Great Lakes region. Cranbook Inst. Sci. Bull. 26, 186 p. Matsubara. K 1943. Studies on the scorpaenoid fishes of Japan: Anatomy, phylogeny and taxonomy (II). Trans. Sigenk- agaku Kenkyusyo 2:172-486. SEEB, L W , AND D R GUNDERSON In press. Population structure and geographic variation in Pacific ocean perch (Sebastes alutus ). Can. J. Fish. Aquat. Sci. SlEGEL, S 1956. Nonparametric statistics for the behavioral sci- ences. McGraw-Hill, N.Y., 312 p. Simpson, G. G . A Roe, And R C Lewontin. 1960. Quantitative zoology. Rev. ed. Harcourt Brace and Co., N.Y., 440 p. SOKAL, R R , AND F J ROHLF 1969. Biometry. W. H. Freeman and Co., San Franc, 776 p. Westrheim, S J. 1973. Age determination and growth of Pacific ocean perch {Sebastes alutus ) in the northeast Pacific Ocean. J. Fish Res. Board Can. 30:235-247. 1975. Reproduction, maturation, and identification of lar- vae of some Sebastes (Scorpaenidae) species in the north- east Pacific Ocean. J. Fish. Res. Board Can. 32:2399- 2411. 680 GENETIC VARIATION IN CHINOOK, ONCORHYNCHUS TSHAWYTSCHA, AND COHO, O. KISUTCH, SALMON FROM THE NORTH COAST OF WASHINGTON R. R Reisenbichler' and S R Phelps^ ABSTRACT We used starch-gel electrophoresis to genetically characterize the populations of chinook salmon, Oncorhynchus tshawytscha . and coho salmon, O. kisutch, in the major drainages of the north coast of Washington (the Quillayute, Hoh, Queets, and Quinault Rivers). Of 55 loci examined for elec- trophoretically detectable variation, 6 were polymorphic (frequency of the common allele was less than 0.95) in chinook salmon and 3 in coho salmon. Statistical tests of interdrainage and intra- drainage variation for coho salmon were tenuous because most of the fish examined were from a single year class so that we could not account for variation among year classes. Nevertheless, these tests suggested that distinct stocks of coho salmon exist within drainages, and that variation was not significantly greater am6ng drainages than within drainages. Interdrainage variation for wild chi- nook salmon was not significant. The data suggested that summer chinook salmon were elec- trophoretically different from fall chinook salmon, and the hatchery populations of chinook salmon were distinct from wild fish. A hatchery population developed primarily from north coast fish was electrophoretically more similar to wild chinook salmon than were the others. Effective conservation and management of natu- ral organisms require protection of the genetic resources (genes, gene combinations, gene pools) of these organisms (Altukhov 1981; Frankel 1983). Conservation of anadromous salmonids from the north coast of Washington (the area from the Quinault River to the Strait of Juan de Fuca) is receiving national attention because many of these fish spawn or rear in Olympic Na- tional Park, and the United States Congress has directed that the natural resources of National Parks be conserved. Olympic National Park is the only natural area administered by the National Park Service outside Alaska with substantial numbers of native anadromous salmonids. There is also international concern for conservation of natural (including genetic) resources in Olympic National Park, as indicated by inclusion of the park in the International Biosphere Reserve Pro- gram (Franklin 1977). The present study was initiated to genetically characterize the populations of chinook salmon, Oncorhynchus tshawytscha , and coho salmon, O. kisutch , from the major drainages of the north lU.S. Fish and Wildlife Service, Seattle National Fishery Re- search Center, Building 204, Naval Station, Seattle, WA 98115. '■^U.S. Fish and Wildlife Service, Seattle National Fishery Re- search Center, Building 204, Naval Station, Seattle, WA 98115; present address: Washington Department of Fisheries, Room 115, General Administration Building, Olympia, WA 98504. Manu.scnpt accepted July 1987. FISHERY BULLETIN VOL 85, NO. 4. 1987. coast: the Quillayute, Hoh, Queets, and Quinault Rivers (Fig. 1). Coho salmon from two other streams in northwestern Washington (the Sno- homish River and Snow Creek) and chinook salmon from Elwha Hatchery and the Wynoochee River were also sampled to enhance our perspec- tive for examining north coast fish. Chinook and coho salmon are native to the west coast of North America from California to Alaska (Scott and Crossman 1973) and are the only species of Pacific salmon that are abundant in each of the major north coast drainages. Starch-gel elec- trophoresis was used to genetically characterize the fish. Our objectives were 1) to develop a baseline set of allele frequency data; 2) to determine whether allele frequencies varied among major drainages; 3) to determine the degree of genetic structuring in coho salmon within major drainages; 4) to de- termine whether summer chinook salmon are electrophoretically distinct from fall chinook salmon; and 5) to determine whether hatchery populations of chinook salmon are electrophoreti- cally distinct from wild (i.e., naturally spawned) fish. We could not examine genetic structuring in chinook salmon within major drainages because wild adults were sampled in the lower portions of the rivers and thus their destinations within the major drainages were unknown, and samples of 681 FISHERY BULLETIN; VOL. 85, NO. 4 Figure 1. — Study area in northwestern Washington. This study focused on the four major stream systems of the north coast: the Quillayute (1), Hoh (7), Queets (9), and Quinault (13) drainages. Numbers identify sampHng areas ("nets" indicates that adults were taken in the Indian gill net fisheries): (1) Quillayute River (nets); (2) Dickey River; (3) Soleduck River; (4) Soleduck Hatchery; (5) Calawah River; (6) Bogachiel River; (7) Hoh River (nets); (8) Hoh River; (9) Queets River (nets); (10) Clearwater River; (11) Upper Queets River, i.e., above the Salmon River; (12) Salmon River; (13) Quinault River (nets); (141 Lower Quinault River, i.e., below Lake Quinault; (15) Quinault National Fish Hatchery; (16) Quinault pens; (17) Upper Quinault River, i.e., above Lake Quinault; (18) Wynoochee River; (19) Snohomish River; (20) Snow Creek; (21) Elwha Hatchery. wild juveniles contained unknown proportions of fish from genetically distinct runs. MATERIALS AND METHODS Three "runs" of chinook salmon and two runs of coho salmon occur in the study area. The runs are primarily distinguished by the time of year when the fish return to fresh water as adults. In gen- eral, spring chinook salmon return to fresh water from March to early June, summer chinook salmon from late June to August, and fall chinook salmon from mid-September to November. Simi- larly, summer coho salmon return to fresh water during August and early September, and fall coho salmon return from mid-October through Novem- ber. Spring chinook salmon and summer coho salmon were not included in this study because returns to fresh water were low and few of these fish were available during our study. Adult salmon spawn in the autumn, and juveniles emerge from the gravel during the following win- ter or spring. Juvenile chinook salmon typically remain in the streams for several weeks to sev- eral months after emerging from the gravel, and enter the ocean during the summer or autumn; juvenile coho salmon remain in the streams for a year and enter the ocean during the spring. Almost all summer coho salmon in the study area spawn in the Soleduck River (Quillayute River system) above Salmon Cascades (Houston lOSS"^). Our samples of fall-run juvenile coho salmon for the Soleduck River were taken from tributaries below Salmon Cascades to reduce the chance of including summer-run fish. In addition to the fish rearing in streams. ^Houston, D. B. 1983. Andromous fish in Olympic Na- tional Park: a status report. Unpubl. rep. U.S. National Park Service, Port Angeles, WA. 682 REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON salmon are raised in one federal, one state, and two tribal hatcheries along the north coast. Sam- ples were taken from six hatchery populations (Table 1). mouths of the rivers. At the hatcheries, samples of tissue were taken within 3 hours after the fish were killed for spawning. Adults from the fish- eries were not available to us until they had been Table 1 . — Run times and stock origins for hatchery populations used in genetic characterization. Species of salmon Run Hatchery Stock origin 1 Chinook Fall Quinault National Fish Quinault River and transfers from Hatchery (Quinault Hoh and Queets Rivers, and NFH) University of Washington, Willapa, Nemah, Finch Creek, Deschutes, Green River, and Samish Hatcheries. Chinook Fall Quinault Tribal Penned Queets River and transfers from Rearing Facility (Quin- Quinault, Green River, Samish, ault Pens) and Deschutes Hatcheries. Chinook Fall Washington Department Primarily Soleduck River; some of Fisheries transfers from Dungeness Hatch- Soleduck Hatchery ery. Chinook Spring- Washington Department Soleduck River and transfers summer of Fisheries from Dungeness, Cowlitz, and Soleduck Hatchery Umpqua Hatcheries. Coho Fall Quinault NFH Transfers from Quilcene, Purdy Creek, Moclips, Willapa, Sole- duck, Simpson, Skagit, Green River, Hood Canal, and Cowlitz Hatcheries. Coho Fall Washington Department Primarily Soleduck River; some of Fisheries transfers from Dungeness Hatch- Soleduck Hatchery ery. 'From Houston (see text footnote 3). Sample Collection Fish were collected during 1983 from the 21 areas identified in Figure 1 (some juvenile Chi- nook salmon were also available from collections made in 1982). Juvenile fish at hatcheries were collected with dip nets at several locations along each raceway containing the fish to be studied. Juveniles in streams were collected by trapping, electrofishing, and seining. A few juvenile coho salmon (usually <15 in each age group) were taken from each of several sites throughout each drainage. Juvenile chinook salmon were taken from several sites in the lower portions of the rivers. Juveniles of both species were collected from areas where no hatchery fish were released or before hatchery fish were released; they were either kept alive or held on ice for up to 8 hours and then frozen at -10°C or -70°C until thawed for electrophoretic analysis. Samples of tissue from eye, liver, white muscle, and heart were taken from adult fish spawned at hatcheries or caught in gill net fisheries at the delivered to wholesale fish buyers. Some fish were delivered more than a day after the fish were killed; although most were kept on ice or refrigerated during this interval, some isozyme activity was lost. Tissue samples from all adults were placed on ice within 30 minutes after exci- sion and were frozen at -10°C or -70°C within 6h. Electrophoresis We used horizontal starch-gel electrophoresis (Utter et al. 1974; May et. al. 1979) to assay fish tissues. Eye, heart, liver, and muscle tissues were removed from partly thawed juveniles just before electrophoretic analysis. We identified alleles at loci encoding specific enzymes, using the staining methods of Harris and Hopkinson (1976) and Allendorf et al. (1977). The nomenclature used to describe the gene loci and the allele variants fol- lowed Allendorf and Utter (1979). Of the 40 enzymes examined, 30 had sufficient activity and resolution to be used in this study 683 FISHERY BULLETIN: VOL. 85, NO. 4 (Table 2). Initially all 30 enzymes were examined in all fish; in later samples, however, we omitted the loci in chinook salmon that had been deter- mined to be monomorphic in previous studies or in our initial screening. Table 2. — Enzymes and loci examined in chinook and coho salmon. Enzyme commission numbers are in parentheses. Tissue E refers to eye, H to heart, L to liver, and M to white muscle. Buffer system 1 was described by Ridgway et al. (1 970), 2 by Clayton and Tretiak (1972), and 3 by Markert and Faulhaber (1965) and Kobayashi et al. (1984). Enzyme p-N-Acetyl-galactosaminidase (3.2.1 .23) N-Acetyl-B-glucosaminidase (3.2. 1 .30) Acid phosphatase (3.1.3.2) Aconitate hydratase (4.2.1.3) Adenosine deaminase (3.5.4.4) Adenylate kinase (2.7.4.3) Alanine aminotransferase (2.6.1.2) Alcohol dehydrogenase (1.1.1.1) Aspartate aminotransferase (2.6.1.1) Creatine kinase (2.7.3.2) Diaphorase-NADH (1.6.*.*) Diaphorase-NADPH (1.6.*.*) Fructose bisphosphate aldolase (4.1.2.3) Fumarate hydratase (4.2.1.2) Glucose-6-phosphate isomerase (5.3.1.9) p-Glucuronidase (3.2.1.31) Glyceraldehyde-3-phosphate dehydrogenase (1.2.1.12) Glycerol-3-phosphate dehydrogenase (1.1.1.8) L-lditol dehydrogenase (1.1.1.14) Isocitrate dehydrogenase (1.1.1.42) L-Lactate dehydrogenase (1.1.1.27) Lactoylglutathione lyase (4.4.1.5) fylalate dehydrogenase (1.1.1.37) Malate dehydrogenase (NADP + ) (1.1.1.40) IVIannose-6-phosphate isomerase (5.3.1.8) a-Mannosidase (3.2.1.24) Phosphoglucomutase (5.4.2.2) Phosphogluconate dehydrogenase (1.1.1.44) Phosphoglycerate kinase (2.7.2.3) Superdioxide dismutase (1.15.1.1) Chinook salmon Coho salmon Loci Tissue Buffer Loci Tissue Buffer ^bGala-2 L 2 ^bGala-1 L 2 bGala-2 L 2 WGa-1 L 1 'bGa-1 L 1 ^Acp-1 L 2 ^Acp-1 L 2 ^Acp-2 L 2 'Acp-2 L 2 ^Ah-1 H 2 'Ah-1 H 2 ^Ah-2 H 2 'Ah-2 H 2 Ah-3 L 2 Ah-3 L 2 — — — ^Ada-1 M 1,3 Ada-2 M,E 1,3 'Ak-1 M 2 ' 'Ak-1 M 2 ^AIat-1 M 1 'Alat-2 M 1 Adh-1 L 2 'Adh-1 L 2 ^Aat-1 L 2 'Aat-1.2 L 2 'Aat-3.4 M 2 Aat-3.4 M 2 'Aat-5 E 2 'Aat-5 E 2 'Ck-1 M 1 Ck-1 f^ 1 ^Ck-2 M 1 'Ck-2 M 1 ^Ck-3 M 1 'Ck-3 M 1 Wia-1 L 1 'Dia-1 L 1 WiaP-1 L 1 'DiaP-1 L 1 ^Fbald-1 E 2 'Fbald-1 E 2 ^Fbald-2 E 2 'Fbald-2 E 2 'Fh-1 M 1 'Fh-1 M 1 ^ Gpi-1 M 1 Gpi-1 M 1 Gpi-2 M 1 Gpi-2 M 1 Gpi-3 M 1 Gpi-3 M 1 ^bGus-1 L 1 'bGus-1 L 1 ^Gapdh-3 E 2 'Gapdh-3 E 2 ^Gapdh-4 E 2 'Gapdh-4 E 2 ^G3pdh-1 M 2 G3pdh-1 1^ 2 ^G3pdh-2 M 2 'G3pdh-2 M 2 ^G3pdh-3.4 H 2 lddh-1,2 L 1 — — — Udh-1 M 2 'ldh-1 M,H 2 ldh-2 M 2 'ldh-2 M,H 2 ldh-3.4 fVI.L 2 ldh-3.4 L,H 2 ^Ldh-1 M 'Ldh-1 M ^ Ldh-2 M 'Ldh-2 M ^Ldh-3 E Ldh-3 E ^Ldh-4 L Ldh-4 L 'Ldh-5 E 'Ldh-5 E ^Lgl-1 E,M Lgl-1 E,tVI Mdh-1,2 L 2 Mdh-1.2 L 2 Mdh-3,4 M 2 Mdh-3.4 M 2 mdhP-1 M 2 MdhP-1 M 2 mdhP-2 M 2 'MdhP-2 M 2 iMdhP-3 L 2 'MdhP-3 L 2 Mpi-1 E 2 Mpi- 1 E 2 ^aMan-1 L 1 'aMan-1 L 1 Pgm-1 t^ 1 'Pgm-1 M 2 Pgm-2 M 2 Pgdh-1 M 2 Pgdh-1 M 2 'Pgk-1 M 2 'Pgk-1 M 2 Pgk-2 M 2 'Pgk-2 M 2 Sod-1 L,H 1 'Sod-1 L 1,2 TSod-2 L 2 ^ No Isozyme variation observed. 684 REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON Data Analysis Goodness-of-Fit Tests We used the chi-square test to examine geno- type frequencies for deviation from the (Hardy- Weinberg) proportions expected with random mating. Cells with an expected number <5 were combined with the next larger cell. The signifi- cance level for each test was modified to account for the increase in type I error when multiple tests of the same hypothesis are made (Cooper 1968). Tests were considered significant if the chi- square statistic exceeded the critical value for chi-square associated with a probability of 0.05/n , where n was the number of loci tested within a sample. In this way the overall probability of re- jecting Hq by chance alone was approximately 1 - (1 - 0.05/;?)" = 0.05 for each sample. Geno- types for Idh-3,4, Mdh-1^, and Mdh-3,4 were not tested because these systems consisted of pairs of loci with identical electrophoretic mobility, and genotypes at each locus could not be determined. The likelihood ratio test (G-test; Sokal and Rohlf 1981) was used to test equality in allele frequencies between year classes. Here also, cells with an expected number <5 were combined with the next largest cell. The G -statistics, summed over all loci, were considered significant if they exceeded the critical value for chi-square associ- ated with a probability of 0.05/s, where s was the number of samples tested. Samples from streams and samples from hatcheries were tested as sepa- rate groups. The correction for multiple compari- sons was made because each of the three Hq — no interbrood variation by drainage, by streams within drainages, or by hatchery — was independ- ently tested for several drainages, streams, or hatcheries, respectively. Analysis of Variance We used analysis of variance (ANOVA) to test interdrainage differences, differences between hatchery and wild chinook salmon, and differ- ences between summer and fall runs of chinook salmon. Data for coho salmon were not tested by ANOVA because data were available for only one year class from most locations, and estimates of interbrood variation in allele frequencies would have come from only two sample locations. The data used were from the loci scored for fish from each major north coast drainage and with fre- quencies <0.95 for the common (100) allele. The values used in the analysis were the arcsin of the square root of the frequency of the common allele at each locus. Differences were tested by contrasts (Table 3) or by partitioning the sum of squares within a one-way ANOVA for each locus (Snedecor and Cochran 1967; SPSS, Inc. 1983). Groups included in this analysis were as follows (adults would have spawned in 1983): Cell Group Run Replicate 1 Quillayute River Mixed 1981 brood 1982 brood 2 Hob River Mixed 1981 brood 1982 brood 3 Queets River Mixed 1981 brood 1982 brood 4 Quinault River Mixed 1981 brood 1982 brood 5 Wynoochee River Mixed 1982 brood 6 Quinault Pens Fall 1982 brood 7 Quinault NFH Fall 1981 brood 1982 brood 8 Soleduck Hatchery Fall 1981 brood 1982 brood 9 Hoh River Fall Adults 10 Queets River Fall Adults 11 Soleduck Hatchery Spring- (data from Milner summer et al. 19831) 1982 brood Adults 12 Hoh River Summer Adults 13 Queets River Summer Adults 'Milner, G. B., D J. Teel, and F M Utter. 1983. Genetic stock iden- tification study; final report of research Unpubl. Rep. Natl. Mar. Fish. Serv., NOAA, Seattle, WA. Juveniles from the different runs of chinook salmon were morphologically indistinguishable and our estimates of error variance were probably inflated because they were based on samples (of juveniles) that vary from year to year in the pro- portion of fish from each race. As a result, the (discriminatory) power for detecting differences between groups was impaired. In view of this re- duced discriminatory power, differences with 0.05 < P < 0.1 were noted in the text; statistical sig- nificance, however, was reserved for differences with P < 0.05. Adult fall and summer chinook salmon from the Quillayute River and adult fall chinook salmon from the Quinault River were not in- cluded in the ANOVA because adults returning to these streams include large numbers of hatch- ery fish (Houston fn. 3). Adult summer chinook 685 FISHERY BULLETIN: VOL. 85, NO. 4 Table 3.— Chinook salmon — coefficients for contrasts (Snedecor and Cochran 1967) within the analysis of variance. Cell numbers refer to groups identified in text. Within each contrast, the mean allele frequencies for groups with positive coefficients were compared with the mean frequencies for groups with negative coefficients. Cell Contrast 1 2 3 4 5 6 7 8 9 10 11 12 13 Interdrainage variation 1 Fall-run adults 0 0 0 0 0 0 0 0 -1 1 0 0 0 2 Summer-run adults 0 0 0 0 0 0 0 0 0 0 0 -1 1 Hatchery vs. wild 3 Summer run 0 0 0 0 0 0 0 0 0 0 2 1 1 Fall run: 4 Quinault Pens 0 0 0 0 0 -2 0 0 1 1 0 0 0 5 Quinault NFH 0 0 0 0 0 0 2 0 1 1 0 0 0 6 Soleduck Hatchery 0 0 0 0 0 0 0 2 1 1 0 0 0 Summer vs. fall 7 Adults 0 0 0 0 0 0 0 0 -1 -1 0 1 1 salmon from the Quinault River were not in- cluded because many hatchery fall chinook salmon return to the Quinault River with the summer-run salmon (during August, when most of our sampling was done) and our samples prob- ably included a high proportion of fall-run hatch- ery fish (Larry Gilbertson'*). Gene Diversity Analysis We used a modification of Chakraborty's (1980) gene diversity analysis to examine the hierarchi- 4 Larry Gilbertson, Quinault Tribal biologist, Quinault In- dian Nation, P.O. Box 189, Taholah, WA 98587, pers. commun. August 1983. cal structure of genie diversity among the sam- ples of wild coho salmon from the north coast. This analysis partitions total gene diversity (//,, heterozygosity of allele frequencies over loca- tions) into interdrainage and intradrainage com- ponents (Nei 1973). We considered three levels of population subdivision (Fig. 2) — broods (6), streams within drainages (w), and drainages id) — so that Hi = Hg + Di,w + 0^,^ + D^t, where Hg is the average heterozygosity within samples, Z)^y, is the gene diversity between broods, 0^,^ is the diversity within drainages, and Z)^/ is the di- versity among drainages. Relative gene diversi- ties {G,j ) are the proportions of //^ associated with a particular hierarchical level; for example, Gwd = Du,d/Ht. Drainage Quillayute River North Coast Coho Salmon Hoh River Queets River Stream Dickey River Soleduck River Quinault River .Bogachiel River vCalawah River Clearwater River Brood 1981 < 1982 1982 1982 1981 < 1982 ■'^Upper Queets River 1982 1982 Figure 2. — Coho salmon — hierarchical subdivisions used in the gene diversity analysis for wild fish from the north coast of Washington (see test). Where brood is not identified, fish were from both the 1981 and 1982 broods. 686 REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON The modification to Chakraborty's (1980) anal- ysis consisted of giving equal weight to subgroups within a cell, rather than weighting them accord- ing to the number of samples within each sub- group. Our sampling design did not include all possible or desirable subgroups; the design was a compromise that allowed us to evaluate the dif- ferent levels of subdivision and still remain within our budget. We felt that equal weighting was necessary because the number of subgroups within a cell usually did not reflect the "true" number of subgroups that may have existed for that cell. Donald Campton (University of Florida, Gainesville) provided a computer program, coded in Fortran 77, that included the required modifi- cation to Chakraborty's equations. Cluster Analysis The unweighted pair group method of cluster analysis (UPGM analysis; Sneath and Sokal 1973) and (nonmetric) multidimensional scaling (Gordon 1981; Kruskal and Wish 1977) were used to illustrate genetic similarities among samples. These two cluster analyses were applied to values of Nei's (1972) genetic distance calculated for each pair of samples. Data from the separate broods were pooled with equal weight for these analyses. RESULTS Chinook Salmon Although fish from two locations showed signif- icant deviation from Hardy-Weinberg propor- tions (P < 0.05/n, where n, was the number of loci tested for location i) — juveniles of the 1982 brood from the Bogachiel River were deficient in het- erozygotes at the Pgk-2 locus and juveniles of the 1982 brood from the Hoh River had an excess of heterozygotes at the Gpi-2 locus — these devia- tions are probably spurious, given the large num- ber (20) of samples tested. Interbrood variation in allele frequencies was significant (P < 0.01) for wild fish and for hatch- ery fish (Table 4). Six loci, or pairs of loci, showed sufficient variation and were scored for enough fish (« > 25) to be used in the ANOVA (Fig. 3, App. Table 1). Variation between drainages was not significant, although summer-run fish may differ between drainages iP = 0.07, Table 4). Hatchery fish were different from wild fish (con- trasts 3 to 6 in Table 5). The UPGM cluster analysis showed that the hatchery populations were distinct from wild ju- veniles and from all but one (Quinault River) sample of adults (Fig. 4). Of the hatchery popula- tions, fall-run fish from Soleduck Hatchery were Table 4. — Chinook salmon — likelihood ratio analysis of interbrood variation at 10 codominant loci. Significant levels were evaluated for totals only. G = likelihood ratio statistic. Ah-3 G, pi-2 ldh-3,4 G df G Mdh-3.4 Mpi-1 df G Pgm-1 df G Pgk-2 df G Sod-1 Total df G df df G df G df G Interbrood variation for drainages Quillayute River 1 1.36 — — 1 15.61 1 6.26 1 0.25 1 0.00 1 3.90 1 3.63 7 31.01** Hoh River 2 2.66 — — 1 10.21 — — 1 1.78 1 3.22 1 0.09 1 0.21 7 18.16* Queets River 1 0.00 — — 1 0.05 — — 1 2.73 1 0.19 1 1.12 1 0.32 6 4.41 Group total 20 53.58t Interbrood variation for streams (withi n drainages) Soleduck River 1 4.77 — — — — — — 1 1.83 1 0.65 1 4.79 1 3.85 5 15.89* Bogachiel River 1 0.20 — — — — 1 5.81 1 0.33 1 0.44 1 0.55 1 0.81 6 8.14 Hoh River 2 266 — — 1 10.20 — — 1 1.78 1 3.22 1 0.08 1 0.21 7 18.16* Queets River above Salmon River 1 0.42 — — 1 2.58 — — 1 3.64 1 16.55 1 2.22 1 0.34 6 25.76** Clearwater River 1 0.43 — — 1 2.73 — — 1 0.57 1 6.57 1 0.07 1 0.02 6 10.39 Group total 30 78.34t Interbrood variation for hatcheries Soleduck Hatchery Spring/summer 2 1.98 1 029 2 6.46 — — 2 9.56 — — 2 4.45 2 12.58 11 35.31" Fall 1 2.26 — — — — — — 1 0.89 — — — — 1 0.31 3 3.46 Quinault NFH (Fall) 1 1.63 — — 1 5.99 — — 1 9.53 1 2.21 1 11.41 1 1.29 6 32.06** Elwha Hatchery — — 1 9.58 — — — — 1 2.28 — — 1 7.15 1 0.30 4 19.34** Group total 24 90.17t •P < 0.05 n "P < 0 01 n tP<0.01. where n = 3 for interbrood variation within drainages, n = 5 for variation within streams, and n eries These are corrections for multiple comparisons (Cooper 1968). 4 for variation within hatch- 687 FISHERY BULLETIN: VOL. 85, NO. 4 Elwha H Soleduck H. IF) Qumault Pens Ouinaull NFH Ouillayule R Ad (F) — Hoh R, Ad Queets R Ad. Soleduck R. Bogachiel R. Hoh R. Clearwater R. Salmon R Upper Queets H Qumault R. Wynoochee R- Soleduck H ISP/SU) Ouillayute R. Ad (SU) — Hoh R Ad (SU) Queets R Ad. (SU) Qumault R. Ad (SU) — Ah-3 0.0 0.2 0.4 0.6 0.8 Frequency of 100 allele — I 1.0 Gpi-2 ~ • 1 I 1 • — » 1 1 1 1 I I 1 0.0 0.2 0.4 0.6 0.8 Frequency of 100 allele 1.0 Elwha H. Soleduck H. (F) Qumault Pens Qumault NFH Qurllayute R Ad. (F) Hoh R. Ad. Queets R. Ad. Soleduck R Bogachiel R. Hoh R. Clearwater R. Salmon R Upper Queets R Quinault R. Wynoochee R. Soleduck H. (SP/SU) Quillayute R. Ad (SU)— Hoh R. Ad. (SU) Queets R Ad (SU) Qumault R Ad. (SU) — ldh-3,4 T" T" 0.0 1 1 1 — 0.2 0.4 0.6 0.8 Frequency of 100 allele — 1 1.0 0.0 Mpi-1 — 1 1 1 1 1 1 1 1 1 0.2 0.4 0.6 0.8 Frequency of 100 allele 1.0 Figure 3a. — Chinook salmon — common-allele frequencies iq ) for four protein-coding loci, or pairs of loci. Each horizontal bar is 4\'q(\ - q )/2n in length and approximates the 95Vc confidence interval; n = number offish scored. Frequencies for fewer than 25 fish are not presented and were not used in the analyses. Data for Gpi-2 were not included in the ANOVA because of missing data (see Appendix Table Al). H. = hatchery; Ad. = adults; F = fall run; SP/SU = mixed spring/summer run; SU = summer run. Adults were from the fall run unless specified otherwise. Table 5. — Chinook salmon — results from multivariate (MANOVA) and univariate analyses of the variance among frequencies (q) of the 100 allele at each of six loci or pairs of loci. Actual values in the analyses were transformed frequencies: arcsin Vq. Hypothesis numbers correspond to those In the text table for contrasts under Ivlaterials and Methods. F = F statistics, df = degrees of freedom for the F statistics. P value from Tests at individual loci Hypothesis MANOVA Ah-3 ldh-3,4 Mpi-1 Pgm-1 Pgk-2 Sod-1 Interdrainage variation 1 Fall run adults 0.54 F 0.15 0.00 0.14 0.52 0.27 0.18 df 1,7 1,7 1,7 1,7 1,7 1,7 2 Summer run adults 0.07 F 5.39 0.00 0.51 0.00 1.00 0.14 df 1,7 1,7 1,7 1,7 1,7 1,7 Juveniles — F 0.43 1.45 0.54 0.38 0.85 0.31 df 3,6 3,7 3,7 3,7 3,7 3,6 Hatchery vs. wild 3 Summer run 0.34 F 11.46 0.54 0.44 22.15* 0.28 5.67 df 1,7 1,7 1,7 1,7 1,7 1,7 Fall run 4 Quinault Pens 0.03* F 4.93 1.50 0.00 0.06 4.40 0.18 df 1,7 1,7 1,7 1,7 1,7 1,7 5 Quinault NFH 0.06 F 0.18 0.33 0.09 0.53 8.91 7.45 df 1,7 1,7 1,7 1,7 1,7 1,7 6 Soleduck Hatchery 0.03- F 6.06 4.37 2.79 11.44 0.00 2.84 df 1,7 1,7 1,7 1,7 1,7 1,7 Summer vs. fall 7 Adults 0.06 F 0.00 0.05 0.10 1.32 1.71 0.67 df 1,7 1,7 1,7 1,7 1,7 1,7 •P < 0.05 for MANOVA. or P < 0.05/6 for univariate tests. 688 REISENBICHLER and PHELPS; GENETIC VARIATION IN CHINOOK AND COHO SALMON Elwha H Soleduck H (F) Quinault Pens Oulnaull NFH Ouillayute R Ad (F) — Hoh R Ad Oueets R Ad Soleduck R, Bogachiel R Hoh R Clearwater R Salmon R Upper Oueets R Quinault R Wynooctiee R Soleduck H (SP SUI — Ouillayute R Ad (SU)— Hoh R Ad (SUI Oueets R Ad ISUI Outnaull R Ad (SUI — Pgk-2 T T" T" "T "T" 0.0 n — 0.2 0.4 0.6 0.8 Frequency of 100 allele Pgm-1 — I 1.0 0.0 T" T" T" T 0.2 0.4 0.6 Frequency of 100 ; 0.8 lele 1.0 Elwha H Soleduck H (F| Quinault Pens Quinault NFH Ouillayute R Ad (F) — Hoh R Ad Oueets R Ad Soleduck R. Bogachiel R. Hoh R Clearwater R Salmon R Upper Oueets R Quinault R Wynoochee R Soleduck H (SP/SU) — Ouillayute R Ad (SU) — Hoh R Ad (SUI Oueets R. Ad (SU) — Oulnaull R Ad (SU) — Sod-1 0.0 "T "T ~r "T "T" "T" — r 0.2 0.4 0.6 0.8 Frequency of 100 allele 1.0 Figure 3b.— Chinook salmon- common-allele frequencies (q) for three protein-coding loci, or pairs of loci. Each horizontal bar is 4\'q(l - q )l2n in length and approximates the 95% confidence interval; n = number of fish scored. Frequencies for fewer than 25 fish are not presented and were not used in the analyses. Data for Gpi-2 were not included in the ANOVA because of missing data (see Appendix Table Al). H. = hatchery; Ad. = adults; F = fall run; SP/SU = mixed spring/summer run; SU = summer run. Adults were from the fall run unless specified otherwise. Figure 4. — Chinook salmon — dendrogram showing results of analysis, by the unweighted pair group method, of genetic dis- tance between samples. Distances (Nei 19721 were based on 11 protein-coding loci or pairs of loci. The letters following the names of samples correspond to the points in Figure 5. Data were from juvenile fish unless adults are indicated. FR = fall run; SR = spring or summer run. SOLEDUCK RIVER A OUILLAYUTE SR ADULTS B UPPER QUEETS RIVER . . C HOH FR ADULTS D CLEAR WAT ER RIVER .. . E HOH SR ADULTS F QUEETS FRADULTS . . . G BOGACHIEL RIVER . . . . H HOHRIVER I OUILLAYUTE FR ADULTS J QUINAULT RIVER K OUEETS SR ADULTS . . . L WYNOOCHEE RIVER.... M SALMON RIVER N SOLEDUCK HATCHERY FR 0 SOLEDUCK HATCHERY SR P QUINAULT ADULTS 0 QUINAULT NFH R ELWHA HATCHERY . . . . S QUINAULT PENS T J_ 0.010 0.005 Genetic distance 689 FISHERY BULLETIN: VOL. 85, NO. 4 most similar to wild fish. Summer-run adults and fall-run adults from the Quillayute River both clustered with the wild fish, suggesting that a large proportion of the fish in these samples were wild fish. Multidimensional scaling gave similar results and more clearly illustrated that hatchery populations were distinct not only from the wild fish but also from each other (Fig. 5). Coho Salmon Coho salmon showed genie variability at 21 loci or pairs of loci; however, the frequency of the common allele was <0.95 for most samples at only 2 loci: bGala-2 and Idh-3,4 (Fig. 6, App. Table 2). Allendorf and Utter (1979) found a sim- ilar lack of variation, reporting that coho salmon display the least amount of electrophoretic varia- tion of the five Pacific salmon species in North America. Hierarchical analysis of genie diversity (het- erozygosity) showed that the interbrood level ac- counted for 2% (= 0.09/(0.09 + 0.85 + 3.97; Table 6) of the genie diversity observed among samples of coho salmon; the within-drainage level ac- counted for 17% and the interdrainage level for 81%. Variation at Pnp-1 had a substantial influ- ence on the average locus values. Unfortunately, data for Pnp-1 were missing for several of the samples because the methodology for this enzyme was not stabilized until we were well into our study. With Pnp-1 excluded from the analysis, the interbrood level accounted for 5% of the genie diversity observed among samples, the within- drainage level accounted for 39%, and the in- terdrainage level accounted for 56%. Variation in allele frequencies among streams within the Quillayute and Queets drainages was statistically significant (tested at bGala-2, Idh- 3,4, and Pnp-1 ; G = 11.27 with 5 degrees of free- dom; P < 0.05); however, interpretation of this result is complicated because data were not avail- able to adequately account for variation among year classes. Variation among drainages was not significantly greater than variation within drainages (P > 0.10, hierarchical likelihood ratio S OLEOUCK H.--FR o SOLEDUCK H.-- SR QUINAULT PENS Q QUINAULT ADULTS ELWHA HATCHERY OUINALT NFH Figure 5. — Chinook salmon — two-dimensional representation (from multidimensional scaling) of ge- netic distances among samples collected for this study. The letters correspond to the groups identified in Figure 4. The polygon encloses the samples of wild fish (A through N). The aim of multidimensional scaling is to represent each group by a point in two-dimensional space so that the relative distances among points represent the relative (genetic) distances between groups. 690 REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON Figure 6. — Coho salmon — common-allele frequencies {q ) for several protein-coding loci. Each horizontal bar is 4V'g(l - q )l2n in length and approximates the 95% confidence interval; n = number of fish scored. Fre- quencies for fewer than 25 fish are not shown and were not used in analysis. Bogachiel R. Calawah R. Clearwater R. Dickey R. Hoh R. Oueets R. Quinault R. Soleduck R. Snohomish R. Snow Creek Quinault NFH Soleduck H. ^ bGala-3 0.0 0.2 0.4 0.6 0.8 Bogachiel R. Calawah R. Clearwater R. Dickey R. Hoh R. Oueeis R. Quinault R. Soleduck R. Snohomish R. Snow Creek Quinault NFH Soleduck H. ldh-3,4 1.0 0.0 0.2 0.4 0.6 0.8 1.0 Bogachiel R. Calawah R. Clearwater R. Dickey R. Hoh R. Oueets R. Quinault R. Soleduck R. Snohomish R. Snow Creek Quinault NFH Soleduck H. Np-1 \ \ I I \ \ \ r- 0.0 0.2 0.4 0.6 0.8 Frequency of 100 allele 1.0 Table 6. — Coho salmon — hierarchical analysis of electrophoretl- cally detectable gene diversity for coho salmon from the Quillayute, Hoh, Queets, Quinault and Wynoochee Rivers. Analysis was based on 58 loci, including 36 that were monomorphic. The hierar- chical design is shown in Figure 3. Relative gene diversity (%) Locus Total gene diversity (Ht) Within samples Among broods Average Average excluding Pnp-1 0.021 0.016 95.09 0.09 97.64 0.12 Within drain- ages 0.85 0.93 Among drain- ages 3.97 1.31 analysis; Grant et al. 1980; Smouse and Ward 1978). Samples without data for bGala-1 or Idh-3,4, the most variable loci, were omitted from the UPGM cluster analysis (Fig. 7) and multidimen- sional scaling (Fig. 8). Both analyses showed that fish from Quinault NFH were distinct from wild fish; much of this distinctiveness occurred at the bGala-2 locus (Fig. 6). Fish from Snow Creek and the Snohomish River clustered among the wild fish from the north coast. The results were simi- lar when Pnp-1 was excluded from the analysis, except that fish from the upper Queets River were no longer distinct from the other wild fish. DISCUSSION Wild Populations Variation in allele frequencies among drain- ages for chinook salmon was not statistically sig- nificant. The inability to detect differences among drainages could have resulted from 1) low statis- tical power (probability of rejecting Hq if it is false) because we had too few broods or because variation in racial composition of juveniles in dif- ferent years inflated the estimates of error vari- ance, 2) our exclusive reliance on data for genes that can be sampled by electrophoresis, or 3) a lack of true genetic difference among groups. We 691 FISHERY BULLETIN: VOL. 85, NO. 4 £ H SOLEDUCK RIVER HOH RIVER CLEARWATER RIVER BOGACHIEL RIVER SNOW CREEK CALAWAH RIVER SNOHOMISH RIVER UPPER OUEETS RIVER QUINAULT NFH 0.004 0.003 0.002 0.001 0 Genetic distance Figure 7. — Coho salmon — dendrogram showing results of analysis, by the unweighted pair group method, of genetic distance between samples. Distances were based on 24 protein-coding loci or pairs of loci. QUEETS R. • CLEARWATER R. HOH R. • • QUINAULT NFH SOLEDUCK R. • • SNOHOMISH R. BOGACHIEL R. • • SNOW CR. CALAWAH R. • • Figure 8. — Coho salmon — two-dimensional representation (from nonmetric multidimensional scaling) of genetic distances among samples collected for this study. Only samples scored for bGala-2 and Idh-3,4 were included in the analysis. emphasize that the lack of differentiation in fre- quencies of electrophoretically detectable alleles does not preclude the existence of important ge- netic differences or status as separate stocks (ge- netic populations). The high degree of "homing" by both chinook salmon (see, e.g., Rich and Holmes 1928) and coho salmon (see, e.g., Shap- avalov and Taft 1954) to the streams from which they originate suggests that salmon from differ- ent drainages should be considered as separate 692 REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON stocks unless strong evidence exists to the con- trary. Our data suggested that summer chinook salmon were distinct from fall chinook salmon (P = 0.06, Table 5). Electrophoretic differences between distinct runs or life history types of chi- nook salmon were also found within the Nanaimo River system (Carl and Healey 1984) and within the Columbia River system (Kristiansson and Mclntyre 1976). Summer-run fish from different streams along the north coast were not suffi- ciently similar to form a cluster separate from the fall-run fish (Figs. 4, 5), and the differences among populations of summer-run fish may be as great as the differences between summer- and fall-run fish. Unfortunately the small number of populations precluded rigorous comparison of these differences. The (significant) variation in allele frequencies between year classes of juvenile chinook salmon may have been exaggerated by variation between years in the proportion of fish from the three dif- ferent runs. This possibility illustrates the need for sampling adult chinook salmon (only adults can be distinguished according to run) in river systems where juveniles from different runs occur together. Of course, the utility of sampling adults to genetically describe wild populations is com- promised if adult hatchery and wild fish occur together and cannot be reliably separated. The gene diversity analysis for coho salmon showed that diversity within drainages was eight to nine times the diversity among broods, with or without Pnp-1 included in the analysis, and sug- gested that separate breeding units exist within drainages as well as between drainages. Separate breeding units within drainages were also sug- gested by the likelihood ratio analysis. Hatchery Fish Versus Wild Fish Analysis of variance for hatchery and wild chi- nook salmon, and the cluster analyses for both chinook and coho salmon showed that the hatch- ery populations of the north coast were geneti- cally distinct from the populations of wild fish. Indeed, coho salmon from Snow Creek or from the Snohomish River were more similar to wild coho salmon from the north coast than were coho salmon from Quinault National Fish Hatchery (Fig. 7). The differences between hatchery and wild fish were to be expected because the hatchery popula- tions were developed with fish from locations in addition to the local stream or exclusive of the local stream. Among chinook salmon, fall-run fish at Soleduck Hatchery were the most similar to wild fish (Fig. 5), probably because the Sole- duck Hatchery population was the only hatchery population developed primarily with local fish (Houston fn. 3). Fall coho salmon at Soleduck Hatchery were also primarily developed with local fish but were not included in the analysis because of missing data. We would expect these coho salmon to be more similar to wild fish than were the coho salmon from Quinault National Fish Hatchery — and that expectation held for al- lele frequencies at Ada-2 and Ldh-4 , and was not countered by evidence from any other loci (App. Table A2). It is reasonable to assume that interbreeding with fall chinook salmon (or fall coho salmon) from Soleduck Hatchery will cause less reduction of fitness and less genetic change for wild fish than will interbreeding with the other (less simi- lar) hatchery fish (Helle 1981; Reisenbichler 1984). The observed differences between fall chi- nook salmon at Soleduck Hatchery and wild fish probably exist because few wild fish are included in the hatchery brood stock. Data for steelhead, Salmo gairdneri, (Reisenbichler and Phelps 1985^) illustrate that the continued use of wild fish in the hatchery brood stock and avoidance of selective breeding are necessary to maintain a hatchery population that is genetically similar to wild fish. Where hatchery populations can be managed separately from wild populations and where few hatchery fish stray onto natural spawning areas, perhaps there is little reason to ensure that hatchery fish are genetically similar to wild fish. However, where substantial numbers of hatchery fish successfully spawn in streams and where genetic resources are to be conserved, hatchery fish should be as genetically similar as possible to the wild fish (e.g., Helle 1981). ACKNOWLEDGMENTS We are grateful to the many persons who pro- vided advice, information, samples, or other assis- tance. The efforts of Scott Corley and Kurt Nelson are especially appreciated. Dave Agee, Susan Glenn, Denny Offutt, and Dave Teel assisted with portions of the electrophoretic analysis. Carl SReisenbichler, R. R., and S. R. Phelps. 1985. Genetic structure of steelhead, Salmo gairdneri , from the north coast of Washington State. Unpubl. rep. National Fishery Research Center, Seattle, WA. 693 FISHERY BULLETIN: VOL. 85, NO. 4 Burger, Steve Landino, Jim Shaklee, and Dick Wilmot reviewed the manuscript and made sug- gestions. J. Aho, R. Contor, A. Fox, and D. Hous- ton provided valuable support for the project. This study was supported by funds from the National Park Service and the U.S. Fish and Wildlife Ser- vice. LITERATURE CITED Allendorf. F W . N Mitchell, N Ryman, and G Stahl 1977. Isozyme loci in brown trout (Salmo trutta L.): detec- tion and interpretation from population data. Hereditas 86:179-190. Allendorf, F W., and F M Utter. 1979. Population genetics. In W. S. Hoar, D. S. Randall, and J. R. Brett (editors), Fish physiology. Vol. 8, p. 407- 454. Acad. Press, N.Y. Altukhov, Yu. p. 1981. The stock concept from the viewpoint of population genetics. Can. J. Fish. Aquat. Sci. 38:1523-1538. Carl, L. M , and M C Healey 1984. Differences in enzyme frequency and body morphol- ogy among three juvenile life history types of chinook salmon iOncorhynchus tshawytscha) in the Nanaimo River, British Columbia. Can. J. Fish. Aquat. Sci. 41:1070-1077. Chakraborty, R. 1980. Gene-diversity analysis in nested subdivided popu- lations. Genetics 96:721-726. Clayton, J W , and D. N Tretiak 1972. Amine-citrate buffers for pH control in starch gel electrophoresis. J. Fish. Res. Board Can. 29:1169-1172. Cooper, D W 1968. The significance level in multiple tests made simul- taneously. Heredity 23:614-617. Frankel, O H 1983. Foreword. In C. M. Schonewald-Cox, S. M. Cham- bers, B. MacBryde, and W. L. Thomas (editors). Genetics and con.servation; a reference for managing wild animal and plant populations, p. xiii-xv. Benjamin/Cummings Publ. Co., Inc., Lond. Franklin, J. F 1977. The biosphere reserve program in the United States. Science 195:262-267. Gordon, A D 1981. Classification. Chapman and Hall, N.Y., 193 p. Grant. W. S., G. B. Milner, P. Krasnowski, and F M Utter 1980. Use of biochemical genetic variants for identifica- tion of sockeye salmon iOncorhynchus nerka) stocks in Cook Inlet, Alaska. Can. J. Fish. Aquat. Sci. 37:1236- 1247. Harris. H , and D A Hopkinson 1976. Handbook of enzyme electrophoresis in human ge- netics. Am. Elsevier Publ. Co., Inc., N.Y. Helle, J H 1981. Significance of the stock concept in artificial propa- gation of salmonids in Alaska. Can. J. Fish. Aquat. Sci. 38:1665-1671. Kobayashi, T . G B Milner, D J Teel, and F M Utter 1984. Genetic basis for electrophoretic variation of adeno- sine deaminase in chinook salmon. Trans. Am. Fish. Soc. 113:86-89. KRISTIANSSON, A C , AND J D MClNTRYE. 1976. Genetic variation in chinook salmon (Oncorhyn- chus tshawytscha) from the Columbia River and three Oregon coastal rivers. Trans. Am. Fish. Soc. 105:620- 623. Kruskal, J B , AND M Wish 1977. Multidimensional scaling. Sage Publications, Beverly Hills, CA. MARKERT, C L , AND I FAULHABER. 1965. Lactate dehydrogenase isozyme patterns offish. J. Exp. Zool. 159:319-332. May, B , J E Wright, and M Stoneking. 1979. Joint segregation of biochemical loci in Salmonidae; results from experiments with Salvelinus and review of the literature on other species. J. Fish. Res. Board Can. 36:1114-1128. Nei, M 1972. Genetic distance between populations. Am. Nat. 106:283-292. 1973. Analysis of gene diversity in subdivided popula- tions. Proc. Natl. Acad. Sci. (USA) 70:3321-3323. Reisenbichler, R R 1984. Outplanting: potential for harmful genetic change in naturally spawning salmonids. In J. M. Walton and D. B. Houston (editors). Proceedings of the Olympic Wild Fish Conference, Peninsula College, p. 33-39. Fisheries Technology Program, Port Angeles, WA. Rich, W H , and H B Holmes 1928. Experiments in marking young chinook salmon on the Columbia River, 1916 to 1927. U.S. Bur. Fish., Bull. 44:215-264. Ridgway. G J , S W Sherburne, and R D Lewis 1970. Polymorphisms in the esterases of Atlantic her- ring. Trans. Am. Fish. Soc. 99:147-151. ScoTT, W B , AND E J Grossman 1973. Freshwater fishes of Canada. Fish. Res. Board Can. Bull. 184, Shapovalov, L , and a Taft 1954. The life histories of the steelhead rainbow trout iSalmo gairdneri gairdneri ) and silver salmon iOncor- hynchus kisutch) with special reference to Waddell Creek, California, and recommendations regarding their management. Calif Dep. Fish Game, Fish Bull. 98, 375 p. Smouse, P E , and R H Ward. 1978. A comparison of the genetic infrastructure of the Ye'cuana and the Yanomama: a likelihood analysis of genotypic variation among populations. Genetics 88:611-631. SNEATH, P H a , AND R R SOKAL 1973. Numerical taxonomy. Freeman and Co., San Franc, 573 p. Snedecor, G W , AND W G Cochran 1967. Statistical methods. 6th ed. Iowa State Univer- sity Press, Ames, lA, 594 p. SOKAL, R R , AND F J ROHLF 1981. Biometry. 2d ed. Freeman, San Franc, 859 p. SPSS, iNC 1983. A complete guide to SPSS language and opera- tions— user's guide. McGraw-Hill, N.Y., 806 p. Utter, F M , H O Hodgins, and F W Allendorf 1974. Biochemical genetic studies of fishes: potentialities and limitations. In D. C. Mallins and J. R. Sargent (ed- itors). Biochemical and biophysical perspectives in marine biology, p. 213-238. Acad. Press, N.Y. 694 REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON Appendix Table 1 —Allele frequencies for chinook salmon from Washington. Each allele is designated by its mobility (relative to the common allele) times 100. N is the number of fish scored for most loci; however, fewer fish may have been scored at some loci. Frequencies from fewer than N/2 fish are identified with an asterisk, and frequencies from fewer than 25 fish are not shown and were not used in our analyses. Numbers preceding sample names correspond to locations shown in Figure 1. Brood N Ah-3 Adh-1 Gpi-2 Location and sample 100 85 118 108 -100 -50 100 67 -15 150 Quillayute River 1 Fall-run adults 1(1983) 99 0.892 0.097 0.011 — 0.990 0.010 0.714* 0.276- 0.010 — . 1 Summer-run adults '(1983) 120 0.906 0.094 — — 0.996 0.004 — — — — 3 Soleduck River 1981 70 0.884 0.101 0.014 — 0.971 0.029 — — — — 1982 40 0.971 0.029 — — — — 0.462 0.488 0.010 — 4 Soleduck Hatchery Fall run 2pre-1982 1982 40 0.840 0.130 I 0.030 0.990 0.010 0.662 0.288 0.012 0.038 Spring/summery run 2pre-1982 0.850 0.150 — — 0.980 0.020 — — — — 1982 50 0.830 0.170 — — 1.000 — 0.700 0.300 — — 1(1983) 77 0.889 0.111 — — 1.000 — . 0.761 0.239 — — 6 Bogachiel River 1981 70 0.926 0.066 0.008 — 0.985 0.015 — — — — 1982 40 0.894 0.091 — 0.015 - — — 0.650 0.338 — 0.012 Hoh River 7 Fall-run adults 1(1983) 37 0.957 0.043 — — 0.973 0.027 0.650 0.350 — — 7 Summer-run adults 1(1983) 86 0.983 0.017 — — 0.960 0.040 0.574 0.426 — — 8 Juveniles 1981 70 0.900 0.064 0.036 — 0.991 0.009 — — — — 1982 76 0.950 0.029 0.021 — — — 0.592 0.388 0.020 — Queets River 9 Fall-run adults 1(1983) 94 0.944 0.044 0.012 — 0.978 0.022 0.595 0.399 0.006 — 9 Summer-run adults 1(1983) 60 0.907 0.074 0.019 — 0.969 0.031 0.652 0.348 — — 10 Clearwater River 1981 70 0.891 0.094 0.014 — 0.957 0.043 — — — — 1982 48 0.917 0.052 0.031 — 0.980 0.020 0.650 0.350 — — 12 Salmon River 1982 48 0.880 0.109 0.011 — 0.943 0.057 0.531 0.469 — — 1 1 Upper Queets River 1981 70 0.906 0.087 0.007 — 0.957 0.043 — — — — 1982 54 0.880 0.070 0.050 — — — 0.491 0.500 0.009 — Quinault River 13 Adults 1(1983) 64 0.927 0.073 — — 0.976 0.024 — — — — 14 Lower Quinault River 1982 55 — — — — — — 0.750 0.236 0.014 — 17 Upper Quinault River 1982 53 0.904 0.096 — — — — — — — — 15 Quinault NFH 2pre-1982 99 0.920 0.080 — — 0.980 0.020 — — — — 1982 50 0.958 0.042 — — 1.000 — 0.411 0.589 — — 16 Quinault Pens 1982 50 0.870 0.054 0.076 — 1.000 — 0.500 0.500 — — Others 18 Wynoochee River 1982 66 — — — — — — 0.635 0.365 — — 21 Elwha Spawning Channel 1981 39 — — — — — — 0.500 0.500 — — 1982 40 0.962 0.038 jid have belonged to the 1983 year class. 1.000 0.237 0.745 — 'Offspring from these adults woi 2Milner. G B , D. J Teal, and F M. Utter. 1983. Genetic stock identification study; final report of research. Unpubl. rep. Natl. Mar. Fish. Sen/., . NCAA. 695 FISHERY BULLETIN: VOL. 85, NO. 4 Appendix Table 1 . — Continued. Brood lddh-1.2 100 36 ldh-3,4 Mdh-1,2 Mdh-3.4 Location and sample 100 120 87 60 100 120 100 115 67 Quillayute River 1 Fall-run adults 1(1983) 1 Summer-run adults 1(1983) 3 Soleduck River 1981 1982 4 Soleduck Hatchery Fall run 2pre-1982 1982 Spring/summer run 2pre-1982 1982 1(1983) 6 Bogachiel River 1981 1982 Hoh River 7 Fall-run adults 1(1983) 7 Summer-run adults 1(1983) 8 Juveniles 1981 1982 Queets River 9 Fall-run adults 1(1983) 9 Summer-run adults 1(1383) 10 Clearwater River 1981 1982 12 Salmon River 1982 1 1 Upper Queets River 1981 1982 Quinault River 13 Adults 1(1983) 14 Low/er Quinault River 1982 17 Upper Quinault River 1982 15 Quinault NFH 2pre-1982 1982 16 Quinault Pens 1982 Others 18 Wynoochee River 1982 21 Elwha Spawning Channel 1981 1982 0.976 0.024 0.928 0.072 0.911 0.057 1 .000 — 0.990 1.000 0.955 0.985 0.915 0.946 1.000 0.035 0.010 0.078 0.054 0.910 0.090 0.922 0.078 0.929 0.071 0.996 0.004 0.915 0.085 0.897 0.103 0.928 0.903 0.825 0.938 0.873 0.072 0.060 0.175 0.062 0.127 0.943 0.931 0.900 0.974 0.978 0.057 0.069 0.090 0.026 0.022 0.032 — 0.010 — 0.010 0.005 0.003 0.004 0.037 — — — 0.936 0.064 — — — 0.952 0.048 — 0.010 1.000 1.000 0.987 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 0.990 1.000 1.000 1.000 000 000 000 000 000 — — 0.980 0.020 — — — — 0.894 0.106 — — 0.938 0.062 0.950 0.050 — — 'Offspring from these adults would have belonged to the 1983 year class ^Milner, G. B., D J Teel, and F M Utter. 1983 Genetic stock identification study; final report of research, Seattle, WA 1.000 1.000 1.000 0.013 0.980 0.971 0.993 0.975 0.990 0.938 0.975 0.985 0.987 0.975 0.888 0.010 0.992 0.961 0.970 0.974 0.996 0.991 0.990 0.990 1.000 0.995 0.020 — 0.025 0.004 0.007 — 0.025 — 0.010 0.062 0.015 0.015 0.006 0.025 0.112 0.959 0.041 0.973 0.027 0.996 0.004 0.990 0.010 1.000 — 0.979 0.021 0.988 — 0.009 0.010 0.010 0.005 0.010 0.007 0.008 — 0.039 — 0.025 0.005 0.026 — 0.004 — 0.995 0.005 — 0.012 0.996 0.004 — 0.929 — 0.071 1 .000 — — Unpubl, rep- Natl. Mar. Fish. Serv,. NOAA, 696 REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON Appendix Table 1. — Continued. Location and sample Mpi-1 Pgm-1 Pgdh-1 Brood 100 116 90 100 129 150 100 90 Pgk-2 100 81 Quillayute River 1 Fall-run adults 1 Summer-run adults 3 Soleduck River 4 Soleduck Hatchery Fall run Spring/summer run 6 Bogachiel River Hoh River 7 Fall-run adults 7 Summer-run adults 8 Juveniles Queets River 9 Fall-run adults 9 Summer-run adults 10 Clearwater River 12 Salmon River 11 Upper Queets River Quinault River 13 Adults 14 Lower Quinault River 17 Upper Quinault River 15 Quinault NFH 16 Quinault Pens Others 18 Wynoochee River 21 Elwha Spawning Channel 1(1983) 1(1983) 1981 1982 2pre-1982 1982 2pre-1982 1982 1(1983) 1981 1982 1(1983) 1(1983) 1981 1982 1(1983) 1(1983) 1981 1982 1982 1981 1982 1(1983) 1982 1982 2pre-1982 1982 1982 1982 1981 1982 0.672 0.688 0.743 0.650 0.810 0.862 0.620 0.580 0.753 0.621 0.663 0.743 0.738 0.593 0.669 0.704 0.661 0.732 0.775 0.638 0.636 0.750 0.746 0.632 0.654 0.610 0.786 0.730 0.723 0.500 0.632 0.328 0.312 0.257 0.350 0.190 0.138 0.370 0.410 0.247 0.379 0.337 0.257 0.262 0.407 0.331 0.296 0.339 0.268 0.225 0.362 0.364 0.250 0.254 0.368 0.346 0.390 0.214 0.270 0.269 0.482 0.368 0.010 0.010 0.008 0.018 0.909 0.951 0.936 0.962 1.000 0.988 1.000 0.990 0.980 0.949 0.925 0.946 0.886 0.900 0.954 0.914 0.882 0.943 0.843 0.906 0.864 0.991 0.984 0.864 0.896 0.930 0.970 0.940 0.917 0.987 1.000 0.091 0.049 0.064 0.038 0.012 0.010 0.020 0.022 0.075 0.054 0.114 0.086 0.039 0.086 0.118 0.050 0.147 0.052 0.079 0.009 0.016 0.136 0.104 0.050 0.030 0.040 0.083 0.029 0.014 0.007 0.007 0.010 0.042 0.057 0.020 0.020 — 0.013 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 .000 .000 .000 .000 .000 .000 .000 0.992 1.000 1.000 0.980 1.000 1.000 1.000 1.000 0.008 0.020 0.512 0.488 0.475 0.525 0.486 0.514 0.325 0.675 0.370 0.630 0.490 0.610 0.617 0.543 0.487 0.467 0.591 0.421 0.438 0.562 0.369 0.463 0.597 0.539 0.500 0.580 0.776 0.235 0.632 0.468 0.250 0.510 0.390 0.383 0.457 0.513 0.405 0.595 0.473 0.527 0.470 0.530 0.454 0.546 0.533 0.409 0.579 0.562 0.438 0.631 0.537 0.403 0.461 0.500 0.420 0.224 0.765 0.368 0.532 0.750 1 0ffspring from these adults would have belonged to the 1983 year class. 2Milner, G B , D J Teel. and F f^ Utter 1983 Genetic stock identification study: final report of research. Seattle. WA. Unpubl rep Natl. Mar. Fish. Serv., NOAA, 697 FISHERY BULLETIN; VOL. 85, NO. 4 Appendix Table ^ —Continued. Brood Sod-1 Location and sample -100 -225 400 Quillayute River 1 Fall-run adults 1(1983) 0.904 0.096 — 1 Summer-run adults 1(1983) 0.860 0.140 — 3 Soleduck River 1981 0.903 0.097 — 1982 0.975 0.025 — 4 Soleduck Hatchery Fall run 2pre-1982 0.800 0.200 — 1982 0.833 0.167 — Spring/summer run 2pre-1982 0.620 0.380 — 1982 0.840 0.160 — 1(1983) 0.724 0.276 — 6 Bogachiel River 1981 0.885 0.115 — 1982 0.926 0.074 — Hoh River 7 Fall-run adults 1(1983) 0.892 0.108 — 7 Summer-run adults 1(1983) 0.879 0.121 — 8 Juveniles 1981 0.886 0.114 — 1982 0.868 0.132 — Queets River 9 Fall-run adults 1(1983) 0.919 0.081 — 9 Summer-run adults 1(1983) 0.852 0.148 — 10 Clearwater River 1981 0.913 0.080 0.007 1982 0.907 0.093 — 12 Salmon River 1982 0.927 0.073 — 1 1 Upper Queets River 1981 0.886 0.107 0.007 1982 0.861 0.139 — Quinault River 13 Adults 1(1983) 0.703 0.297 — 14 Lower Quinault River 1982 0.949 0.051 — 17 Upper Quinault River 1982 0.824 0.176 — 15 Quinault NFH 2pre-1982 0.780 0.210 0.010 1982 0.720 0.200 0.080 16 Quinault Pens 1982 0.929 0.071 — Others 18 Wynoochee River 1982 0.821 0.179 — 21 Elwha Spawning Channel 1981 0.741 0.259 — 1982 0.697 0.303 — lOffspring from these adults would have belonged to the 1983 year class 2Milner, G. B , D J Teel. and F M Utter 1983 Genetic stock identification study; final report of research. Unpubl. rep. Natl. Mar. Fish. Serv., NCAA, Seattle, WA. 698 REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON Appendix Table 2. — Allele frequencies for coho salmon from Washington. Each allele is designated by its mobility (relative to the common allele) times 100. N is the number of fish scored for most loci; however, fewer fish may have been scored at some loci. Frequencies from fewer than W 2 fish are identified with an asterisk, and frequencies from fewer than 25 fish are not shown and were not used in our analyses. Numbers preceding sample name correspond to locations shown in Figure 1. Brood bGala-2 Ah-3 Ada-2 Aat-3.4 N 100 128 100 91 115 130 100 110 91 100 87 Quillayute River 2 Dickey River 1981 52 — — 0.990 — — 0.010 0.990 0.010 — 1.000 — 3 Soleduck River 1981 37 0.660 0.340 0.973 0.027 — — 0.933 0.967 — 1.000 — 1982 48 0.583 0.417 1.000 — — — 1.000 — — 1.000 — 4 Soleduck Hatchery 1981 40 — — 0.988 0.012 — — 1.000 — — 1.000 — 5 Calawah River 1982 40 0.551 0.449 0.925 0.050 0.025 — 0.988 0.013 — 1.000 — 6 Bogachiel River 1982 74 0.546 0.454 1 .000 — — — 1.000 — — 1.000 — Hoh River 8 Winfield, Nolan, 1981 48 0.061 0.399 1 .000 — — — 0.990 0.010 — 1.000 — Pin Creeks 1982 44 — — 0.989 0.011 — — 1.000 — — 1.000 — 8 Other tributaries 1982 45 0.616 0.384 1.000 — — — 1.000 — — 1.000 — Queets River 10 Clearwater River 1981 210 0.637* 0.363 0.979 0.021 — — 0.995 0.005 — 1.000 — 1 1 Upper Queets River 1981 76 0.674 0.326 1.000 — — — 0.993 0.007 — 1.000 — Quinalt River 13 Lower Quinault River 1982 60 — — — — — — 1.000 — — 0.923 0.077 15 Quinault NFH 1981 40 — — 1 .000 — — — 0.961 0.039 — 1.000 — 1982 40 0.814 0.186 0.988 0.012 — — 0.988 — 0.012 1.000 — Others 19 Snohomish River 1981, 1982 106 0.595 0.405 0.986 0.009 0.005 — 1.000 — — 1.000 — 20 Snow Creek 1981 60 0.542 0.458 1.000 — — — — — — 1.000 — Brood Ck-1 Gpi-1 Gpi-2 Gpi- 3 G3pi. -100 1h-1 100 127 100 250 100 157 67 100 90 -15 Quillayute River 2 Dickey River 1981 1 .000 — 1 .000 0.990 0.010 — 1.000 — 0.990 0.010 3 Soleduck River 1981 1 .000 — 1 .000 0.986 0.014 — 1.000 — 1.000 — 1982 1 .000 — 1 .000 1 .000 — — 1.000 — 1.000 — 4 Soleduck Hatchery 1981 1 .000 — 1 .000 1 .000 — — 1.000 — 0.969 0.031 5 Calawah River 1982 1 .000 — 1 .000 1 .000 — — 1.000 — 0.987 0.013 6 Bogachiel River 1982 1.000 — 1 .000 1 .000 — — 1.000 — 1.000 — Hoh River 8 Winfield, Nolan, 1981 1 .000 — 1 .000 1 .000 — — 1.000 — 1.000 — Pin Creeks 1982 1 .000 — 1 .000 1 .000 — — 1.000 — 1.000 — 8 Other tributaries 1982 1 .000 — 1.000 0.989 0.011 — 1.000 — 1.000 — Queets River 10 Cleanwater River 1981 1 .000 — 0.998 0.002 0.993 0.007 — 1.000 — 1.000 — 1 1 Upper Queets River 1981 1.000 — 0.993 0.007 0.993 — 0.007 1.000 — 0.987 0.013 Quinalt River 13 Lower Quinault River 1982 1 .000 — 1 .000 0.991 0.009 — 1.000 — — — 15 Quinault NFH 1981 1 .000 — 1 .000 1 .000 — — 1.000 — 0.959 0.041 1982 1 .000 — 1 .000 0.950 0.050 — 0.975 0.025 1.000 — Others 19 Snohomish River 1981, 1982 0.995 0.005 1.000 0.981 — 0.019 1.000 — 1.000 — 20 Snow Creek 1981 1 .000 — 1 .000 1 .000 — — 1.000 — 0.992 0.008 699 FISHERY BULLETIN: VOL. 85, NO 4 Appendix Table 2— Continued. Brood ldh-3.4 Ldh-3 Ldh-4 LQ' 1-1 100 130 70 123 157 100 45 140 100 110 100 80 Quillayute River 2 Dickey River 1981 0.825 0.169 0.006 — — 0.971 0.029 — 1.000 — 1.000 3 Soleduck River 1981 0.973 — 0.007 0.020 — 0.986 . — 0.014 1.000 — 1.000 — 1982 — — — — — 1.000 — — 1.000 — 1.000 — 4 Soleduck Hatchery 1981 0.964 0.036 — — ^ 1.000 — — 1.000 — 1.000 — 5 Calawah River 1982 1.000 — — — — 1.000 — — 1.000 — 0.963 0.037 6 Bogachiel River 1982 0.996 — 0.004 — — 0.993 — 0.007 1.000 — 1.000 — Hoh River 8 Winfieid, Nolan, 1981 0.995 — 0.005 — — 1.000 — — 1.000 — 1.000 — Pin Creeks 1982 0.978 — 0.022 — — 0.988 0.012 — 1.000 — 1.000 — 8 Other tributaries 1982 1.000 — — — — 0.989 0.011 — 1.000 — 1.000 — Queets River 10 Clearwater River 1981 0.924 0.073 0.001 0.002 — 0.993 0.007 — 1.000 — 1.000 — 1 1 Upper Queets River 1981 0.858* 0.132* — 0.006- ■ 0.004' 1.000 — — 0.994 0.006 1.000 Quinalt River 13 Lower Quinault River 1982 0.905 0.095 — — — 1.000 — — 1.000 — 1.000 — 15 Quinault NFH 1981 0.917 0.077 — — 0.006 1.000 — — 1.000 . — 1.000 — 1982 1.000 — — — — 0.988 0.012 — 0.975 0.025 1.000 Others 19 Snohomish River 1981, 1982 0.920 0.070 0.007 — 0.003 1.000 — — 0.972 0.028 1.000 — 20 Snow Creek 1981 0.985 0.015 — — — 1.000 — — 1.000 — 1.000 — Brood Mdh-1.2 Mdh-3,4 Mdh-5 MdhP-1 100 37 210 100 123 110 89 140 100 107 100 130 Quillayute River 2 Dickey River 1981 1.000 — — 0.991 0.009 — — — 0.971 0.029 1.000 — 3 Soleduck River 1981 1982 1.000 1.000 — — 0.967 1.000 0.020 — — 0.013 — — 1.000 — 4 Soleduck Hatchery 1981 1.000 — — 1.000 — — — — — — 1.000 — 5 Calawah River 1982 1.000 — — 0.988 0.012 — — — 1.000 — 1.000 — 6 Bogachiel River 1982 1.000 — — 0.993 0.007 — — — — — 1.000 Hoh River 8 Winfieid, Nolan, 1981 1.000 — — 0.985 0.005 0.010 — — 1.000 — 1.000 Pin Creeks 1982 1.000 — — 0.955 0.040 — 0.006 — 1.000 — 1.000 8 Other tributaries 1982 1.000 — — 0.956 0.044 — — — — — 1.000 — Queets River • 10 Cleanwater River 1981 1.000 — — 0.985 0.010 — 0.005 — 0.960 0.040 0.990 0.010 1 1 Upper Queets River 1981 0.994 0.003 0.003 0.997 0.003 — — — 1 .000* 1.000 — Quinalt River 13 Lower Quinault River 1982 1.000 — — 1.000 — — — — 1.000 1.000 15 Quinault NFH 1981 1.000 — — 1.000 — — — — 0.950 0.050 1.000 — 1982 1.000 — — 0994 0.006 — — — 1.000 Others 19 Snohomish River 1981, 1982 0.997 0.003 — 0.998 0.002 — — — 1.000 1.000 20 Snow Creek 1981 1.000 — — 1.000 — — — — — 1.000 — 700 REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON Appendix Table 2. — Continued. Mpi-1 Pgm-2 Pgdh-1 Pnp-1 Brood 100 123 -100 -55 Quillayute River 2 Dickey River 1981 1.000 — 0.990 0.010 3 Soleduck River 1981 1.000 — 1.000 — 1982 1.000 — 1.000 — 4 Soleduck Hatchery 1981 1.000 — 1.000 — 5 Calawah River 1982 1.000 — 1.000 — 6 Bogachiei River 1982 1.000 — 1.000 — Hoh River 8 Wintield, Nolan, 1981 1.000 — 1.000 — Pin Creeks 1982 1.000 — 1.000 — 8 Other tributaries 1982 1.000 — 1.000 — Queets River 10 Clearwater River 1981 0.995 0.005 1.000 — 1 1 Upper Queets River 1981 1.000 — 1.000 — Quinault River 13 Lower Quinault River 1982 1.000 — 1.000 — 15 Quinault NFH 1981 1.000 — 1.000 — 1982 1.000 — 1.000 — Others 19 Snohomish River 1981, 1982 1.000 — 1.000 — 20 Snow Creek 1981 1.000 — 1.000 — 100 92 100 155 1.000 _ _ _ 1.000 _ _ _ 1.000 _ _ _ 1.000 _ _ _ 1.000 — 1.000 — 1.000 _ _ _ 1.000 _ _ _ 1.000 — 0.673 0.327 1.000 _ _ _ 1.000 — 0.877* 0.123* 1.000 — 0.780* 0.220* 0.974 0.026 — — 1.000 _ _ _ 1.000 — 1.000 — 1.000 — 0.995 0.005 1.000 — 1.000 — 701 ASSESSMENT OF INTERACTION BETWEEN NORTH PACIFIC ALBACORE, THUNNUS ALALUNGA , FISHERIES BY USE OF A SIMULATION MODEL p. Kleiber and B. BakerI ABSTRACT Using a simulation model of a typical year in the North Pacific albacore fisheries in the 1970s, we tested for the degree to which the activity of fleets affects the performance of other fleets. The results show that rather drastic (factor of two) changes in the activity of any of the three principal albacore fleets have only a mild effect on the catch of the other fleets. With the overall exploitation rate in the model close to the exploitation rate determined from tagging results (6%), the maximum degree of interaction was a 7.5'7f drop in longline catch resulting from doubling the baitboat effort. The mild degree of interaction was insensitive to exploitation rate up to approximately 10% exploitation, although interaction became more severe at higher levels of exploitation. Fishery interaction, the effect of one fishing fleet on another, is a phenomenon of growing concern to those involved in the management and devel- opment of pelagic fisheries. This concern has arisen from the growing awareness that oceanic fishery resources are not unlimited and from the evolution of exclusive economic zones to protect local interests against large international fishing fleets. Assessing the potential for interaction be- tween tuna fisheries in different island countries was one of the principal reasons that the South Pacific Commission conducted the Skipjack Sur- vey and Assessment Programme (Kearney 1983). Workshops on this topic have been held during international tuna fishery meetings, and a Tuna Fisheries Interaction Programme has been pro- posed within the Indo-Pacific Tuna Development and Management Programme. Because there is a multiplicity of fleets and na- tions involved in harvesting albacore, Thunnus alalunga , a tuna, in the North Pacific, there is a potential concern about interaction between these fleets. A history of North Pacific alba- core fishing since the 1950s is summarized by Laurs (1983). Three principal fleets have been responsible for the catch: the Japanese baitboat, the Japanese longline, and the United States jig- boat fleets (Fig. 1). In the 1970s these accounted for more than gO'/f (60%, 15%, and 18%, respec- tively) of the total catch. In recent years, Japanese gill net gear has become important, ac- 'Southwest Fisheries Center La JoUa Laboratory, National Marine Fisheries Service, NOAA, P.O. Box 271, La JoUa, CA 92038. Manuscript accepted July 1987 FISHERY BULLETIN: VOL 85. NO 4, 1987 counting for approximately 20% of the total catch from 1981 through 1983 (Fig. 1). Detailed statis- tics on this emerging fishery are not currently available. Among the three principal fleets, the U.S. fleet tends to take the smallest fish, and the longline the largest, but the size distributions in the catch overlap to a large extent (Fig. 2). The geographic distribution of the fleets is indicated in Figure 3, but the overlap is overemphasized because there is seasonal separation in many cases. Neverthe- less, the migratory nature of albacore makes for potentially significant interaction between fleets that are separated in time and space. Because there have been no clear trends in catch or in catch per effort (Laurs 1983), it has been assumed that the albacore stocks have not been adversely affected by the fisheries, and such woes as the fishermen have had have not been blamed on poor status of stocks. Therefore there has been little reason for fleets to accuse one another of depleting the stocks and thus little concern about fishery interaction. To verify that sanguine view, we have estimated the degree of interaction between the three principal albacore fisheries in the North Pacific. We defined interac- tion to be the degree to which changes in the activity (effort) of one fleet affect the performance (catch) of another fleet. The magnitude of this kind of interaction cannot be calculated di- rectly from fishery data, nor can controlled, real- life experiments be conducted on the grand scale necessary to address this topic. However, experi- ments conducted on simulation model are feasi- ble. The results of such experiments with an 703 FISHERY BULLETIN: VOL 85, NO. 4 1970 1982 Figure 1. — Annual catch of albacore by gear type. [Data from Majors and Miller. 1985. Summary of the 1984 North Pacific albacore fishery data. U.S. Natl. Mar. Fish. Serv., Southwest Fish. Cent. Admin. Rep. LJ-85-14. 45 p.] albacore simulation model are the subject of this report. THE MODEL We used a model that incorporates recruitment, growth, migration, natural mortality, and har- vest of albacore by the Japanese baitboat fleet, the Japanese longline fleet, and the United States surface fleet (primarily jig gear). Our approach was to manipulate the effort of one fleet at a time and note the effect on the catch of the other fleets. A full technical description of the model is given by Kleiber and Baker^. We discretized fish size into 5 cm length classes and the North Pacific range of albacore into nine geographic zones (Fig. 3). The basic dynamics within a size class and zone are described by the following differential equation: dP.Jt) dt - Gs-jPg-i^it. + ^i^i P.^it) M Gs + ^ M-^ Ps^it) -2 'S.Z.g 2Kleiber, P., and B. Baker. 1987. The North Pacific alba- core simulation model. U.S. Natl. Mar. Fish. Serv., Southwest Fish. Cent., Admin. Rep. LJ-87-2, 38 p. where Cg^^ = Qs^ fzji^ Ps/t^ is the catch rate (number per unit time) by size, zone, and gear. The symbols are defined as follows: s — index for size class z — index for geographic zone z — index for zone adjacent to z g — index for gear type Ps,z^t ' — population (numbers) by size and zone at time t the following being input parameters: Po,z^t^ — recruitment rate by zone at time t Gg — proportion growing out of size s per unit time Go —always = 1 (so thatPQ ^(t) is recruitment rate) M-zi—eo — coefficient of migration from zone Zj to zone 22 M — natural mortality Qs,g — catchability by size and gear fz^it) — effort by zone and gear at time t. INPUT PARAMETER VALUES Full details of how input parameters were esti- mated are given by Kleiber and Baker (fn. 2). The following is a summary. The most complete catch and effort data sets that were available to us and that cover the three 704 KLEIBER and BAKER: INTERACTION BETWEEN NORTH PACIFIC ALBACORE FISHERIES Baitbaat Longllne l.Ot 65 85 Length (cm) 125 Figure 2. — Average albacore annual catch-at-size in numbers (solid lines), and annual catch-at-size predicted by model under nominal conditions (dashed lines), 1970-80. l.Ot 65 85 Length (cm) U.S. I.Ot 45 65 85 Length (cm) 105 105 125 1: 0.4 b: 0.6 1: 0.2 j:0.3 b: <.05 1: 1.7 j: <.05 b: Baltboat 1: Longline j: U.S. Jig 100 120 140 160 180 160 140 120 100 80 longitu(de Figure 3. — Model zones with average annual albacore catch (numbers x 105) by gear type, 1970-80. 705 FISHERY BULLETIN: VOL. 85, NO. 4 major fleets span the years 1970 through 1980. We processed the data into an "average year", that is, the average (over the years 1970-80) of catch by size, zone, month, and gear and the aver- age of effort by zone, month, and gear (Kleiber and Baker"^). The effort values were used directly in the model and the catch and effort used to esti- mate other input parameters. We estimated the 1970-80 average recruitment and preliminary catchability values by size and gear by use of a size-structured cohort analysis (Jones 1981). The catch-at-size vector necessary for this cohort analysis was obtained from the average year by aggregating over zones, averag- ing over months, and smoothing over size classes. To conduct the cohort analysis, we needed to specify an average final cohort size, which was unknown to us. We tried a series of values and chose results for which the overall exploitation rate (catch divided by recruitment estimate) was close to the overall exploitation rate estimated from tagging. Tagged albacore have been re- leased in the U.S. fishery at an average size of approximately 65 cm, and approximately 6% of the tags have been recovered (Laurs"^). Nonre- porting losses are small for the major fisheries that recovered the tags (Laurs^). Assuming a value of 10% for nonreporting and Type I and II tag losses of 12% and 0.098 year'^ respectively (Laurs et al. 1976), the exploitation rate of re- cruits to 65 cm should be approximately twice the raw recovery percentage. But the exploitation rate in the cohort analysis and the simulation model is based on recruits to 25 cm which should be approximately twice as numerous as recruits to 65 cm (based on growth and natural mortality rates used in the model). Therefore, the exploita- tion rate of recruits to 25 cm should be approxi- mately equal to the raw tag recovery percentage (6%). We chose a cohort analysis with an exploita- tion rate of 6.3% as the basis for the results pre- sented below except where we discuss sensitivity to exploitation rate for which we repeated the analysis several times starting at this point with a series of cohort analyses at a series of higher exploitation rates. 3Kleiber, P., and B. Baker. 1986. Development of catch and effort data base for the North Pacific albacore simulation model. U.S. Natl. Mar. Fish. Serv., Southwest Fish. Cent., Ad- min. Rep. LJ-86-26, 21 p. 4Laurs, R. M. 1979. Results from North Pacific albacore tagging studies. U.S. Natl. Mar. Fish. Serv., Southwest Fish. Cent., Admin. Rep. LJ-79-17, 10 p. 5R. M. Laurs, Southwest Fisheries Center La JoUa Labora- tory, National Marine Fisheries Service, NOAA, P.O. Box 271, La Jolla, CA 92038, pers. commun. March 1987. The cohort analysis yielded a Pacific-wide re- cruitment estimate. We apportioned one third of this recruitment to each of the three southern zones, where albacore larvae are predominantly found (Nishikawa et al. 1984). Size-specific overall fishing mortalities ob- tained from the cohort analysis were apportioned to gear type by the proportion of the total catch at each size that is taken by each gear type in the average year. We then converted the fishing mor- talities into catchabilities by dividing by the over- all average effort for each gear type. Size-specific growth coefficients, Gg, were the growth rates in length (dl/dt) at the upper end of each size class divided by the length of the size classes (5 cm). We estimated the growth rates from the derivative form of the von Bertalanffy growth equation ^J-^(Zoo-l) where U is 135.6 cm and k is 0.014 month" ^ (Clemens 1961). We used these same values in the size-structured cohort analysis, which re- quired input of growth information. A value for natural mortality was also needed both in the cohort analyses and in the model. We used a value of 0.017 month"^ (0.2 year"^) (Suda 1966). The tag data would be a good source of informa- tion to estimate migration coefficients except that the tag recovery effort is not uniformly dis- tributed and analytical techniques to deal with that situation are not well developed. To get a reasonable set of migration coefficients, we quan- tified the experience of three experts, scientists knowledgeable about the North Pacific albacore fisheries and the available tag data. We first asked the experts to identify the significant paths (movement from one zone to an adjacent zone) for each of a series of broad (10 cm) size classes. For each path we then asked how the intensity of migration via that path is distributed over the months of the year and we evaluated the average intensity by asking the experts the following question: On average, during the season of this migration, if 100 fish of the given size class are now in the origin zone, how many of these (irre- spective of mortality) would be expected to be in the destination zone one month from now? We calculated the average migration coefficient for the particular path by 706 KLEIBER and BAKER: INTERACTION BETWEEN NORTH PACIFIC ALBACORE FISHERIES fJL = In 100 100 -X where X is the answer to the above question. The monthly migration coefficients were then ob- tained by scahng the distribution of intensity over months so that the average was equal to |jl. The migration coefficients for the 10 cm size classes were then assigned to the smaller (5 cm) size classes used in the model, and the coefficients smoothed over size to soften discontinuities. The pattern of movement represented by the migration coefficients can be summarized as fol- lows: For immature fish (<85 cm) in the zones north of 25° north the pattern is vigorous seasonal movement toward the east in the summer and toward the west at other times. New recruits, which appear in the southern zones, migrate mainly northward throughout the year and are entrained in the east-west excursions of the northern zones. Mature fish (>85 cm) accumulate in the southern zones with brief movements northward in April and May. RESULTS When we ran the model with input values esti- mated as described above and with the effort of all fleets set at nominal levels, that is, the average seasonal and geographic pattern and magnitude of effort for the 1970s with the pattern repeated year after year, we found that after 10 years of simulation the seasonal and geographic pattern and magnitude of catch closely repeated itself year after year. Therefore in making comparisons of model results under different conditions, we allowed the model to run at least 10 years under a given repetitive annual regime before recording the catch results during 1 year of simulation. In using the preliminary catchability values in the model, we found that the predicted catches were too low and the exploitation rate achieved (2.6%) was less than half the exploitation rate in the cohort analysis, which estimated those catch- ability values (6.3%). This is because the cohort analysis could not deal with geographic and sea- sonal variability. The fleets were presumed to be harvesting the ocean-wide population rather than the fish in a localized area and time as in the simulation model. We therefore scaled the catcha- bilities of each fleet upward to make the annual catches in number in the model (after 10 years of simulation) close to the real average annual catches (Kleiber and Baker fn. 2). With the cor- rected catchabilities, an exploitation rate of 5.1% was achieved and we took the results in this case to be our nominal (control) results (Fig. 2, Table 1). We then made runs in which the original sea- sonal and geographic pattern of effort was main- tained but the magnitude of effort of one of the fleets was either doubled or halved. We could then compare the annual catch of each fleet under the changed (experimental) conditions with the annual catch under nominal conditions. Table 1, — Average albacore annual (1970-80) catch in numbers and metric kilotons (kt) by baitboat, longline, and U.S. fleets plus annual catch from model after at least 10 years of simulation under nominal conditions and under various conditions of altered fishing effort. baitboat longi ne U.S. number number number catch (106) kt (106) kt (106) kt average 6.29 57.52 0.68 13.30 2.65 19.56 nominal effort 6.87 55.12 0.66 9.47 2.67 18.32 baitboat X 2 12.89 102.00 0.62 8.76 2.63 18.05 effort ^ 2 3.55 28.69 0.69 9.86 2.69 18.46 longline X 2 6.85 54.91 1.32 18.79 2.67 18.30 effort -^ 2 6.88 55.23 0.33 4.75 2.67 18.33 U.S. X 2 6.80 54.43 0.65 9.23 5.19 35.41 effort -^ 2 6.90 55.49 0.67 9.59 1.35 9.32 The catch-at-size for the three fleets under nominal and experimental conditions is plotted in Figures 4 to 6. Changes in effort in one fleet ap- pear to have little effect on the size distribution in the catch of any of the fleets. The effect on amount caught is another matter. Total catch of all sizes, both in numbers and in weight, is given in Table 1. We obtained catch in weight by converting the number caught in each length category to weight using the length- weight relationship of Clemens (1961) and then by summing over length categories. The effects are summarized in Tables 2 and 3 where the change from nominal catch for each fleet is given for each experimental treatment. By far the largest effect of a change in effort of any fleet is the eflect on its own catch. A doubling of the bait- boat eflbrt causes the largest between-fleet effect, which is a 7.5% depression of the longline catch in weight, a loss of approximately 700 t (Table 3). A similar loss to the baitboat fleet, due to doubling of U.S. eff'ort, is only a 1.3% decrease in the bait- boat catch (Table 3). We tested the sensitivity of our results to the 707 FISHERY BULLETIN: VOL. 85, NO. 4 Ba i tboa t Catch Effortx2 : . . Baitboat Longl ine U.S. Effort/2 : Ba itboat Longl ine U.S. 115 125 Length (cm) Figure 4. — Annual albacore baitboat catch-at-size in numbers predicted by the model under nominal and experimental conditions. Table 2. — Interaction matrix for annual albacore catch in numbers. The values given are the differences between the catch under altered effort and the nominal catch (percent of nominal catch in parentheses). effect 4> A catch: number y 106 (%) cause 4> baitboat longline U.S. baitboat effort X 2 - 2 6.02 (87.6) -3.32 (48.3) -0.04 (6.1) 0.03 (4.5) -0.04 (1.5) 0.02 (0.7) longline effort X 2 - 2 -0.02 (0.3) 0.01 (0.1) 0.66 (100.0) -0.33 (50.0) 0.00 (0.0) 0.00 (0.0) US effort X 2 - 2 -0.07 (1.0) 0.03 (0.4) -0.01 (1.5) 0.01 (1.5) 2.52 (94.4) -1.32(49.4) Table 3. — Interaction matrix for annual albacore catch in weight. The values given are the differences between the catch under altered effort and the nominal catch (percent of nominal catch in parentheses). effect => A catch: kt (%) cause X 2 - 2 baitboat longline U.S. baitboat effort 46.88 (85.1) -26.43 (47.9) -0.71 (7.5) 0.39 (4.1) -0.27 (1.5) 0.14 (0.8) longline effort X 2 - 2 -0.21 (0.4) 0.11 (0.2) 9.32 (98.4) -4.72 (49.8) -0.02 (0.1) 0.01 (0.1) U.S. effort X 2 - 2 -0.69 (1.3) 0.37 (0.7) -0.24 (2.5) 0.12 (1.3) 17.09 (93.3) -9.00 (49.1) overall exploitation rate by repeating the whole analysis at higher exploitation rates, starting with the cohort analysis, correcting catchabilities to give a new set of nominal results, and finally measuring the most sensitive interaction, the ef- fect of doubled baitboat effort on the longline catch (Fig. 7). The degree of interaction is not affected very much when the exploitation rate is below 10%, but it rises quickly at higher exploita- tion rates. DISCUSSION Our results support the notion that fishery in- teraction is not of great consequence, at least for the North Pacific albacore fisheries typical of the 708 KLEIBER and BAKER: INTERACTION BETWEEN NORTH PACIFIC ALBACORE FISHERIES 20t Longline Catch 15-. sz. u to u 10-. 5 ■■ Effortx2 : Baitboat Longl ine U.S. Effort/2 : Baitboat Longl ine U.S 25 I 35 t - 45 105 115 125 Length (cm) Figure 5. — Annual albacore longline catch-at-size in numbers predicted by the model under nominal and experimental conditions. 1970s. The reliability of this conclusion, of course, depends on the reliability of our simulation model, but in evaluating the behavior of the model, we should remember that what is impor- tant is the response of the model to experimental manipulation not the exactitude of the nominal behavior in comparison to real data. Of course, if the nominal behavior is outlandish, the responses to manipulation will be suspect. Therefore we used the average year as a signpost to tune the nominal results of the model into the range of plausible behavior, but we did not insist on exact duplication of the average year (itself an abstrac- tion that never happened in reality). A case in point is the longline catch, which under nominal conditions in the model is less (in weight) than any of the real annual longline catches for the years 1970-80. The average over those years is 13.3 metric kilotons (kt) per year whereas the nominal longline catch in the model is 9.47 kt/year (Table 1). The discrepancy is ex- plained by the fact that the average size offish in the model longline catch is less than the average size in the real longline catch, because large fish in the model migrate out of reach of the longline fleet more than they should. We have not cor- rected this problem because we are waiting for further information from tagging studies to get better estimates of migration coefficients. We ex- pect the corrections to be quantitative refine- ments of the existing values and not a qualitative change in the current migration pattern in the model. What is important in the current context is that bias in the nominal results is bound to show up in the experimental results as well. The migration coefficients were the same in both control and experimental situations in the model. Therefore, refinements to the coefficients are not likely to make much difference in the relative values in Tables 2 and 3, particularly in the percentages. It is pertinent that our conclusion of low interaction 709 FISHERY BULLETIN: VOL. 85, NO. 4 i.ar U.S. Catch 1.0- ID £1 a (0 CJ Effort«2 : . . Baitboat Longl ine _... U.S. Effort/2 : . Baitboat Longl ine . . U.S. Nominal : 105 115 125 Length (cm) Figure 6. — Annual albacore U.S. catch-at-size in numbers predicted by the model under nominal and experimental conditions. (expressed as percent) persisted through a series of updates of the model (such as changes in config- uration of geographic strata) and inevitable up- dates and corrections of the fishery data base. Though the nominal behavior of the model need not conform precisely to the mean behavior of albacore fisheries in the 1970s (however that might be defined), the behavior should, nonethe- less, be a plausible representation of the albacore fisheries in that period. We have seen, for exam- ple, that our conclusion would be suspect if the actual exploitation rate were considerably higher than the 6% value that we assumed (Fig. 7), but such high exploitation levels would be contrary to the tag return results. We have only tested the effects of changes in the magnitude of effort, not changes in seasonal and geographic pattern of effort, which might cause the fleets to overlap much more than they do. However, our experimental treatment of dou- bling the effort of a fleet is tantamount to adding u U c o ■D cr 5 10 15 Exploitation rate (%) Figure 7. — Sensitivity of interaction to overall exploitation rate. The ordinate is the percent reduction from nominal levels in annual albacore longline catch in weight as a result of dou- bling the baitboat effort. 710 KLEIBER and BAKER: INTERACTION BETWEEN NORTH PACIFIC ALBACORE FISHERIES a completely overlapping, competing fleet. The response of each fleet to doubling of its own effort was close to a 1009f increase in catch (Tables 2, 3), indicating that the degree of competition was low. Therefore, it would be difficult to design a realis- tic experimental treatment that 1) would be sim- ply a shift in the geographic and seasonal pattern of effort in one fleet (not a change in magnitude), and 2) would have a strong impact on another fleet. A legitimate question is whether our conclu- sions, which are based on 1970s data, can be ex- trapolated to the current conditions. The most striking change in recent years is the emergence of the gill net fishery for albacore, which now takes approximately 207f of the total catch. How- ever, because the total catch has not increased, the exploitation rate must still be mild, and we would expect that interaction between fleets would also still be mild. We cannot use our model to estimate interaction quantitatively in this sit- uation because we lack detailed data on the gill net fishery. CONCLUSIONS The implication of our results is that fleet inter- action is not likely to be significant if the pattern and magnitude of effort in the 1970s are main- tained. This assessment could change if the over- all exploitation rate increases considerably. The recent emergence of the gill net fishery could be of significance in this regard. The levels of annual catch that have been reported by this fishery are not likely to be of concern, but the significance cannot be confidently evaluated unless detailed catch, effort, and size distribution data are made available. LITERATURE CITED Clemens, H. B. 1961. The migration, age, and growth of Pacific albacore (Thunnus germo), 1951-1958. Calif. Dep. Fish Game, Fish. Bull. 115, 128 p. Kearney, R E 1983. Assessment of the skipjack and baitfish resources in the central and western tropical Pacific Ocean: Summary of the Skipjack Survey and Assessment Pro- gramme. South Pacific Commission, Noumea, New Caledonia, 37 p. Jones. R 1981. The use of length composition data in the fish stock assessment (with notes on VPA and cohort analy- sis). FAO Fish Circ. 734, 55 p. Laurs, R M 1983. The North Pacific albacore — An important visitor to California Current waters. Calif. Coop. Oceanic Fish. Invest. Rep. 24:99-106. Laurs, R M., W H Lenarz, and R. N. Nishimoto. 1976. Estimates of rates of tag shedding by North Pacific albacore, Thunnus alalunga. Fish. Bull., U.S. 74:675- 678. Nishikawa, Y . M Honma. S Ueyanagi, and S Kikawa 1984. Average distribution of larvae of oceanic species of scombroid fishes, 1956-1981. Far Seas Fish. Res. Lab., Jpn., Contrib. No. 236, 99 p. SUDA. A. 1966. Catch variations in the north Pacific albacore- VI. The speculations about the influences of fisheries on the catch and abundance of the albacore in the north Pacific by use of some simplified mathematical models. Rep. Nankai Reg. Fish. Res. Lab. 24:1-14. 711 AGE DETERMINATION OF PACIFIC COD, GADUS MACROCEPHALUS , USING FIVE AGEING METHODS^ Han-Lin Lai,^ Donald R. Gunderson,' and Loh Lee Low" ABSTRACT A comparative study of age determination methods for Pacific cod, Gadus macrocephalus , was carried out using dorsal and pectoral fin rays, scales, otoliths, and coracoids. A preliminary validation using the modal length of a strong year class confirmed that sections of dorsal fin rays were the most reliable ageing method. A Monte Carlo method was developed for converting scale ages to dorsal fin-ray ages. An analysis by log-linear model was developed for testing the effects of ageing method and age class on repeatability of age reading. Scales have been widely used for ageing Pacific cod, Gadus macrocephalus , in the North Pacific since Kennedy (1970) developed the method for fish in Hecate Strait, British Columbia. However, Bakkala (1981)^ found that the scale method may not be an appropriate ageing method since the estimated ages from scales do not appropriately reflect the progress of known year classes in the eastern Bering Sea. Beamish et al. (1978) also found that Kennedy's criteria might not be satis- factory for scales from juvenile Pacific cod in Canadian waters. Beamish (1981) reported that thin sections of fin rays can be reliably aged and might be more accurate than scale ages when ageing older fish. However, Westrheim and Shaw (1982) encoun- tered difficulties with fin-ray cross sections and reported that fin-ray ages were younger than scale ages. They also discovered false checks on the scales during the first year of life, which fitted the annulus criteria of Kennedy (1970), and vali- dated the scale ageing method for age-1 and -2 Pacific cod. Chilton and Beamish (1982) noted problems associated with scales and fin rays, and ^Contribution No. 726, School of Fisheries, University of Washington, Seattle, WA. ^School of Fisheries, WH-10, University of Washington, Seat- tle, WA 98195; present address: Center for Quantitative Sci- ences, HR-20, University of Washington, Seattle, WA 98195. ^School of Fisheries, WH-10, University of Washington, Seat- tle, WA 98195. ''Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., BIN C15700, Seattle, WA 98115. ^Bakkala, R. G. 1981. Pacific cod of the eastern Bering Sea. Document submitted to the annual meeting of the Inter- national North Pacific Fisheries Commission, Vancouver, BC, Canada, October 1981, 32 p. Available at the Northwest and Alaska Fisheries Center, NMFS, NOAA, Seattle, WA 98115. Manuscript accepted August 1987. FISHERY BULLETIN: VOL. 85, NO. 4, 1987. recommended a judicious mixture of scale ages, fin-ray ages and length-frequency analysis for ageing Pacific cod in Canadian waters. In earlier studies, Mosher (1954), Moiseev (1953), and Ketchen (1970) reported that the otolith surface ageing method was not satis- factory for Pacific cod. Ketchen (1970) also had no success with ageing of vertebrae, or opercula. Lai (1985) reported that age determination from the bony tissues in the gillcover, scapula, and cleithrum was not feasible. This paper reports the results of a comparative study and preliminary validation of age determi- nation methods for Pacific cod in the eastern Bering Sea using dorsal and pectoral fin rays, scales, otoliths and coracoids. In addition, we de- velop a method to convert scale ages to dorsal fin-ray ages. This age conversion method makes it possible to use existing ages estimated from scales. MATERIALS AND METHODS Dorsal and pectoral fin rays, otoliths, and scales were taken from 230 Pacific cod collected from foreign fishing vessels operating in the east- ern Bering Sea during September 1983 to March 1984. Among these samples, coracoids were also taken from 101 fish. Dorsal and pectoral fin-ray sections were pre- pared using an Isomet low-speed saw (Lai 1985), and the annuli were identified by the criteria of Beamish (1981) and Chilton and Beamish (1982). Scale images were made by acetate plate (Koo 1962) and then read by a microfisch reader. An- nuli on scales were identified by the methods de- 713 FISHERY BULLETIN: VOL. 85, NO. 4 scribed in Westrheim and Shaw (1982). Otoliths must be sectioned with a low-speed electric saw and burnt with care, as they are quite brittle. The interpretation of annuli on break-and-burn sur- faces was adapted from that of walleye pollock (Lai and Yeh 1986). The coracoid is a dermal bone connected to the scapula and radii in the pectoral girdle. A trian- gular component within the coracoid (Fig. 1) shows alternate opaque (dark) and translucent (light) zones under a binocular microscope and transmitted light, by which annuli can be identi- fied. The center of this structure occasionally con- tains some checks that are parallel to the an- nulus; however, their spacing is narrower than that of regular annuli. The age samples were read Figure 1. — Upper: The pectoral girdle. A, coracoid; B, scalpula; C, radii; D, the portion used for age determination. Lower: Enlarged Part D. •, translucent zone, counted as an annulus. 714 LAI ET AL.: AGE DETERMINATION OF PACIFIC COD twice by the senior author, who had 3 years expe- rience in age determination using all of the age- ing methods. A loglinear model (Fienberg 1981) was used to test the independence of repeatability (R ,k = 1,2 corresponding to agreement and disagreement) of age readings corresponding to age-class (A , J = 1,...,8 corresponding to age classes given in Table 1) and ageing method (M, i = 1,...,5 corresponding to ageing methods in the sequence given in Table 1). See Appendix A for the details Table 1 . — Observed frequency of agreement between two age readings using different ageing methods. Percent agreement is shown in parentheses. Method Age Repeatability (R) {M) (A) Agree Disagree Total Coracoid under 2 9 100) 0 9 3 1 (50) 1 2 4 10 (77) 3 13 5 12 60) 8 20 6 18 67) 9 27 7 5 45) 6 11 8 3 43) 4 7 over 9 4 40) 6 10 Total 62 63) 37 99 Break-and- under 2 9 90) 1 10 burn 3 7 64) 4 11 4 8 73) 3 11 5 41 67) 20 61 6 35 58) 25 60 7 14 39) 22 36 8 6 35) 11 17 over 9 9 41) 13 22 Total 129 57) 99 228 Dorsal under 2 9 100) 0 9 fin ray 3 2 67) 1 3 4 13 81) 3 16 5 36 77) 11 47 6 40 82) 9 49 7 24 69) 11 35 8 9 39) 14 23 over 9 7 29) 17 24 Total 140 68) 66 206 Pectoral under 2 8 100) 0 8 fin ray 3 6 (75) 2 8 4 18 64) 10 28 5 36 68) 17 53 6 24 45) 29 53 7 12 43) 16 28 8 10 45) 12 22 over 9 3 (25) 9 12 Total 117 (55) 95 212 Scale under 2 9 (82) 2 11 3 11 (61) 7 18 4 60 (74) 21 81 5 48 (64) 27 75 6 11 (38) 18 29 7 1 (11) 8 9 8 0 (0) 1 1 over 9 0 (0) 1 1 Total 140 (62) 85 225 of this statistical method. The computer program P4F in BMDP (Dixon 1983) was used for compu- tation and analysis. An analysis of variance (ANOVA) model with repeated measures (Winer 1971) was used to ex- amine variability in age readings due to the methods. The ANOVA model was Xy„ = M- + TT^j + M, + (M7T)j„ + Rj + (i?1T)j„ + {MRTT)ijn +ey„ where, Xy„ is the observed age of the nth fish by the ith ageing method and jth reading, i = 1, 2, 3, 4, 5 indicates ageing methods in the sequence shown in Table 1, J = 1, 2 indicates the first and sec- ond age readings, n = 1,...JSI indicates number of fish, |JL is the grand mean, ■n is the effect between subjects (indi- vidual fish), M is the effect of ageing method, R is the effect of reading, (Mir), {R tt), and (MR -it) are the two- and three-factor interactions of effects M , R , and tt, and e is the random error. Readings from coracoids were not included in the analysis because of the small sample size. Speci- mens with missing values were also excluded from this analysis. The Q -statistic (Snedecor and Cochran 1967) was used to test the differences between the mean ages between readings and be- tween ageing methods. Validation of age determination was carried out by comparing the mean lengths at age esti- mated from the five ageing methods to the modal lengths of the 1977 year class. The progression of this year class could be traced by examining the length-frequency distributions for 1978 to 1983. Because existing age data files for Pacific cod at the Northwest and Alaska Fisheries Center (NWAFC) were derived from scales, it is impor- tant to explore whether or not the scale age data can be used for stock assessment. A simulation study was carried out to convert a scale age- length key to a dorsal fin-ray age-length key. 715 FISHERY BULLETIN: VOL. 85. NO. 4 since the dorsal fin-ray method provides more ac- curate ages. Age-length keys, derived from scales collected in 1979 and 1980 NWAFC demersal trawl sur- veys in the eastern Bering Sea, and length- frequency distributions obtained from these sur- veys were used as basic data to be converted. The classification probabilities for fin rays vs. scales were constructed from 966 age readings on dorsal fin rays and scales from the same fish collected in 1983-84 (Table 2). Each fish in the 1979 and 1980 scale age-length key was assigned a "pseudo" dor- sal fin-ray age, which was a random variate gen- erated from the subprogram GGDA in IMSL (International Mathematical and Statistical Library). The GGDA subroutine used the prior probability density in Table 2 corresponding to the scale age and length group to which the fish belongs. This was done 30 times to create 30 con- verted age-length keys each for 1979 and 1980. Then, mean lengths at age were determined from each converted age-length key, and age composi- tions were estimated by the method described in Lai (1987). These simulated results were com- pared with those from length-frequency analysis (Lai 1985). RESULTS Table 1 shows that the age readings from dorsal fin rays had the highest percent agreement (68%), followed by coracoids (63%), scales (62%), otoliths (57%), and pectoral fin rays (55%). Furthermore, the percent agreement for the major age classes in the fishery (ages 4-7) were much higher than for the other methods. To confirm this result, a log-linear model was fitted to the data in Table 1. Table 3 shows that the best log-linear model (Appendix A) was e,.^=log(m,,^) = |x + Xf + Xf +\f ik jk y (2) Table 2. — Observed classification probability for age readings from scales and dorsal fin-rays. Scale Length age (cm) Dorsal fin-ray age 8 10 11 12 13 14 Total 1 20-29 1 .0 1 30-39 1.0 2 30-39 1.0 2 40-49 0.750 0.250 3 30-39 1.0 3 40-49 0.333 0.667 3 50-59 0.188 0.375 0.375 0.063 3 60-69 0.087 0.565 0.174 0.130 0.043 4 40-49 0.133 0.467 0.400 4 50-59 0.604 0.264 0.094 0.038 4 60-69 0.066 0.377 0.221 0.197 0.098 0.041 4 70-79 0.153 0.350 0.190 0065 0.058 0.066 0.051 0.036 4 80-89 0.348 0.217 0.174 0.130 0.043 0.087 5 50-59 0.143 0.857 5 60-69 0.043 0.391 0.239 0.196 0.109 0.022 5 70-79 0.005 0.095 0.264 0.206 0.180 0.116 0.069 0.048 0.016 5 80-89 0 154 0.128 0.205 0.256 0.154 0.077 0.026 5 90-99 0.500 0.500 6 60-69 0.611 0.222 0.167 6 70-79 0.017 0.083 0.333 0.233 0,167 0.117 0.033 0.017 6 80-89 0.188 0.344 0.156 0.125 0.125 0.063 6 90-99 0.059 0.412 0.176 0.118 0.118 0.118 6 100-109 0.500 0.500 7 70-79 0.333 0.111 0.333 0.111 7 80-89 0.182 0.545 0.182 7 90-99 0.063 0.187 0.250 0.187 0.063 0.250 8 80-89 0400 0.200 0.400 8 90-99 0.333 0.167 0.333 0.029 0.111 0.090 0.167 20 6 8 8 2 3 32 23 15 53 122 137 23 14 92 189 39 2 18 60 32 17 4 9 11 16 5 6 716 LAI ET AL.: AGE DETERMINATION OF PACIFIC COD Table 3. — Test for the independence of percent agreement {R) of age readings correlated to age-class {A) and ageing method (M). Model Proba- df G2 bility M. + Xf* + \;^ + \« + \MR + ),AR + xMA + \MAR Q 0.00 1 .0000 ' I k ik ik II Ilk (jL + \W + \>< + XR + \Mfl + )^AR + xMA I I k ik jk ij H + \'^< + \^ + \f + X'Wf? + ),AR I I k IK jk u. + \M + X^* + \fl + \AR + XMA I ] k jk I] u. + XW + x>< + xR + xMR + },MA I j k ik ij 28 28.08 0.4603 32 48.87 0.0285 56 241.75 0.0000 35 125.03 0.0000 which impHed that percent agreement was corre- lated to age-class and ageing method, and the estimated age frequencies differed by ageing method. However, these pairwise relationships between any two factors are unrelated to the third. Because we are interested in the effects of age- ing method and age class on repeatability, it is reasonable to look at the ratio between agree- ment (^ = 1) and disagreement (k = 2) for each combination of ageing method and age class, i.e., myi/my2 for ^ll i andj. The logarithm of this ratio is known as the logit model (Fienberg 1981, chap- ter 6). The logit model for Equation (2) can be derived as Table 4. — Estimated logit effects corresponding to the loglin- ear model In Equation (2). Factor Logit effect constant (w) 0.540 Age (^A) <2 2.204 1 age 3 0.240 age 4 0.676 age 5 0.278 age 6 -0.196 age 7 -0.738 age 8 -1.064 over 9 -1.400 Age method (w^) Dorsal fin ray 0.568 Coracoid 0.096 Otolith -0.058 Pectoral fin ray -0.208 Scale -0.398 logitiij) = log (miji/mij2) = w -I- co M + Ui^ (3) O) (Xf Xf) X^„^). The ef- where, the logit effects, wf = (X^^ - Xj^«), and cof = {\ff ,. 2 '• ^"^ ^^ fects (X's) without factor R in Equation (2) are cancelled out by subtraction in Equation (3). The values of X's can be obtained from the BMDP pro- gram and substituted into Equation (3). The re- sults show that there was a significant declining effect on agreement as age increased (Table 4). This indicated that percent agreement decreased with increasing age. Age determination using coracoids and dorsal fin rays had a positive effect on agreement, which indicated that agreement of these methods was higher than the average of the five methods, but the other ageing methods had a negative effect, i.e., agreement was lower than average. Table 5 shows that the effect of ageing method was significant for all fish older than age 3. There was no significant difference (5% level) between readings except for ages 5-6. This difference prob- ably resulted from differences between readings for otoliths and pectoral fin rays (Table 6). Mean square (MS) for the ageing method eff'ect in- creased with age and was the predominant com- ponent in the within-subject variation for all age categories. Therefore, variability in age determi- nation was mainly due to ageing method rather than inconsistent annulus interpretation by the reader. Using the Q -statistic, the mean ages of the two readings were not significantly different except for age group 5-6 using otolith and pectoral fin- ray ageing methods (Table 6). Significant differ- ences between ageing methods were found in all age categories except the youngest. Age readings from dorsal fin rays and pectoral fin rays were not significantly different for fish younger than age 6. Age readings from otoliths and pectoral fin rays were not significantly different for fish older than age 7. Otolith readings were older than other methods for fish younger than age 6 but were younger than dorsal fin-ray readings for fish older than age 7. Scale readings gave consistently younger ages than the other methods. Bakkala and Wespestad (1984) reported that recruitment of the 1977 year class was uniquely strong when compared to its neighboring year classes. The modal length of this year class can be 717 FISHERY BULLETIN: VOL. 85, NO. 4 Table 5. — Comparison of ageing variability of Pacific cod by ANOVA. Dorsal fin-ray age N Between subject (it) Within subject Method (M) /Wtt Reading (fl) R-n MR MRt^ 1-2 8 SS df F 14.359 7 2.051 0.297 2.578 3 21 0.099 0.123 0.81 0.016 0.109 1 7 0.016 0.016 1.00 0,047 1,328 3 21 0.016 0.063 0.25 3-4 19 SS df MS F 52.395 18 2.911 8.967 38.658 3 54 2.989 0.716 4.18** 0.164 3.711 1 18 0.164 0,206 0.80 0.072 7.553 3 54 0.024 0.140 0.17 5-6 93 SS df MS F 270.495 92 2.940 166.154 273.721 3 276 55.385 0.992 55.85" 1 .840 26,285 1 92 1.840 0.286 6.44** 1,122 74,253 3 276 0,374 0,269 1.39 7-8 57 SS df MS F 217.244 56 3.879 344.018 267.232 3 168 114.673 1.591 72.09** 0.219 26.031 1 56 0.219 0.465 0.47 0.921 56.829 3 168 0.307 0.338 0.91 9-1- 23 SS df MS F 339.457 22 15.430 498.283 198.717 3 66 166.090 3.011 55.16** 0.035 21,652 1 22 0.348 0.984 0,35 1.869 53.130 3 66 0,623 0,805 0,77 significant at 1% level. Table 6 — Tests for differences between mean ages of Pacific cod using various ageing methods. Bracket and underline: not significantly different at 5% level. Dorsal fin-ray age N Reading Dorsal fin ray Otolith Pectoral fin ray Scale SD df 1-2 8 1 1,375 1.3751 1.3751 1 .250 1 0,063 7 2 1,375 1.475 1.375 1.250 Method mean 1,375 1.425 1.375 1.250 0.124 21 3-4 19 1 3.842 4.211 1 3.842 3.895 3.579 0.147 18 2 3.847 4.315 J 3.579 Method mean 3.895 4.263 3.869 3.579 0.194 54 5-6 93 1 5.5051 5.774 5.430 4.581 1 0.078 92 2 5.588 5.979 5.591 4,570 Method mean 5.547 5.877 5.511 4,576 0,103 276 7-8 57 1 7.386 6.561 1 6.614] 5.000 1 0,128 56 2 7.316 6.702 J 6.737 4.982 Method mean 7.351 6.632 6.676 4.991 0.167 168 9-1- 23 1 10.1301 8.000 8.304 ] 5.522 0.293 22 2 9.913 8.261 J 8.130 5.304 Method mean 10.022 8.131 8.217 5.413 0.362 66 traced from the length-frequency distributions from 1978 to 1983 (Fig. 2). Using the method of Macdonald and Pitcher (1979), the mean lengths for ages 1-6 were 22, 35, 43, 52, 61, and 64 cm respectively (Lai 1985) and were very close to the modal length of length-frequency distributions. Figure 2 shows the mean lengths at age esti- mated from the 1983-84 samples by the five age- ing methods and the comparison to the modal lengths. It is apparent that the mean lengths at age estimated from dorsal fin rays were closest to the modal lengths. Also, the variability around mean length at age estimated from dorsal fin rays was generally smaller than for the other methods. We also used the index of variation^ (IV, Lai 61 V = 100% • Vscy, -X,m{n - l)(X + Y)/2], where X, is the first age reading, Y, js the second age reading, n is the sample size, and X and Y are mean of the X, and y, (Lai 1985). This indicates the degree of variation between the two 718 LAI ET AL.: AGE DETERMINATION OF PACIFIC COD 12 ^ 10 - 8 - 6 - 4 2 0 J 1978 10 n 6 4 2 1981 y ,--^i--> T 1 1 1 0 T 1 1 1982 Figure 2. — Length-frequency distributions collected from trawl surveys in 1978-83 (after Bakkala and Wespestad 1984). The mode in 1978 represents age-1 Pacific cod of the 1977 year class, and progresses from age 2 to age 6 in the subsequent years. The estimated mean lengths at ages 1-6 estimated from the five ageing methods (1, coracoids; 2, otoliths; 3, dorsal fin rays; 4, pectoral fin rays; and 5, scales) correspond to the age of 1977 year class. Horizontal lines indicate 95% confidence interval around means. Vertical lines indicates the mean length from length-frequency analysis. 1985) to examine the degree of precision. Among the five ageing methods, the IV for dorsal fin rays was the lowest ( ISVc ) when compared with other methods (149c, 14%, 15%, and 16% respectively for otoliths, coracoids, scales, and pectoral fin rays). The accuracy of converting scale ages to dorsal fin-ray ages was also evaluated. Mean length at age and age composition (obtained by using con- verted dorsal fin-ray ages) were compared with corresponding results from length-frequency analysis (Macdonald and Pitcher 1979). The 95% ages being compared taking the age distribution of the sample into account. confidence interval (Fig. 3) for each converted mean length at age in 1979 and 1980 included the corresponding value estimated from the length- frequency analysis. However, the mean lengths at age derived from scales were significantly dif- ferent from the other two. The age composition estimated by the scale method showed a monotonic decrease with age in 1979, and the strong 1977 year class (age 2) was not evident (Fig. 4). Nevertheless, length- frequency data from surveys indicated that age-2 fish were dominant in 1979 (Bakkala and We- spestad 1984). The age composition based on con- verted dorsal fin-ray ages was dominated by age- 2 fish and was similar to that estimated from the 719 FISHERY BULLETIN: VOL. 85, NO. 4 80- 2 u X (- o u 60- 40- 20- 0 I <^- 1979 I I I I I I I I I I I I — r 1 5 10 1980 I I I I I I I r I — I I I I 15 10 AGE (YRS) Figure 3. — Comparison of mean length at age estimated from scales ( — A — ), converted dorsal fin-ray ages {^), and length-frequency analysis ( — # — ) for 1979 and 1980. Vertical line indicates the 95% confidence interval from the 30 simulation runs. -^ 60 g40H q: O Q_ O I 20 0 1979 ^^'♦^♦♦♦^ 10 AGE (YRS) 1980 10 Figure 4. — Comparison of age composition estimated from scales ( — A — ), converted dor- sal fin-ray ages (o<), and length-frequency analysis ( — 9 — ). Vertical line indicates 95% confidence interval of the 30 simulation runs. length-frequency analysis. In 1980, the propor- tions estimated by using the scale method were somewhat higher than for the other two methods at ages 2 and 3 (Fig. 4), but were lower at ages 4 and 5. Since the classification probabilities and scale age-length keys used in this study were derived independently from different years, the method of age conversion appears to be relatively insensitive to interannual differences in the clas- sification probabilities. DISCUSSION Although Westrheim and Shaw (1982) vali- dated the interpretation of the annuli on scales for age groups 1 and 2, this validation was not 720 LAI ET AL.: AGE DETERMINATION OF PACIFIC COD considered to be sufficient for all age groups. Beamish et al. (1978, figs. 12 and 13) showed that the difference between readers was substantial even in age groups 1 and 2. In our study, we found that age readings from the scale method had low precision, and that scale ages were much younger than those obtained by any other method. In this study, validation for age groups 1-6 showed that dorsal fin rays gave the most reliable ages for Pacific cod and thus should be used in the future. This method provided estimates of mean length at age that agreed most closely to observed growth of the 1977 year class, and the precision of this method was the highest attained in this study. Another major advantage of this ageing method is that additional fin rays can be taken from fish with a previous history of fin-ray re- moval to verify the accuracy of age determina- tions between time of release and recapture. We used the Monte Carlo method to convert scale ages to dorsal fin-ray ages. The results indi- cated that the previously collected scale age data can be used in age-dependent methods of stock assessment. Since the 1983-84 classification prob- abilities used in this study were completely inde- pendent of the 1979-80 scale age data, the method appears to be robust with respect to interannual variability in the classification probabilities for Pacific cod. However, application of this method to other species will require caution when the classification probabilities are applied to the data from different years, since interannual variabil- ity could be a source of error. Still, any errors that arise will probably be smaller than those pro- duced from an inappropriate ageing method. Analytical methods, such as those of Pella and Robertson (1979) and Cook (1982), could also be used for converting scale age-length keys to dor- sal fin-ray age-length keys. However, these meth- ods are mathematically more complicated and occasionally yield negative values in some of the converted age-length keys (Cook 1982). The method of Hoenig and Heisey (1987)^ is of partic- ular interest because it avoids negative values by applying an incomplete E-M (estimation and maximization) algorithm to fit a log-linear model to the classification matrix. Nevertheless, this method may not be valid if there is a substantial systematic ageing error (as in our case, dorsal ^Hoenig, J. M., and D. M. Heisey, 1987. Using log-linear models with the EM algorithm to correct estimates of stock composition and convert to age. Manuscr. in prep. Rosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, FL. fin-ray ages vs. scale ages) because too many empty cells are in the classification matrix. Alter- natively, the method of Barlow (1984) can also be used for this purpose, although the result will be very similar to our Monte Carlo method, as Bar- low's method is a deterministic version of our own. ACKNOWLEDGMENTS This study was funded by Northwest and Alaska Fisheries Center, National Marine Fish- eries Service, NOAA, Seattle, WA. We thank Richard G. Bakkala for providing length frequen- cies and the scale age-length keys. Particular thanks are due to Richard J. Beamish and an anonymous reviewer for their constructive com- ments. LITERATURE CITED Bakkala, R G., and V G. Wespestad. 1984. Pacific cod. In R. G. Bakkala and L. L. Low (edi- tors). Condition of groundfish resources of the eastern Bering Sea and Aleutian Islands region in 1983, p. 21- 36. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-53. Barlow, J. 1984. Mortality estimation: biased results from unbiased ages. Can. J. Fish. Aquat. Sci. 41:1843-1847. Beamish. R J 1981. Use of sections of fin-rays to age walleye pollock, Pacific cod, albacore, and the importance of this method. Trans. Am. Fish. Soc. 110:287-299. Beamish, R J . K R Weir. J R Scarsbrook, and M S Smith. 1978. Growth of young Pacific hake, walleye pollock. Pacific cod and lingcod in Stuart Channel in 1976. Fish Mar. Serv. MS. Rep. 1518, 111 p. Chilton, D E . and R J Beamish 1982. Age determination methods for fishes studied by the groundfish program at the Pacific Biological Sta- tion. Can. Spec. Publ. Fish. Aquat. Sci. 60, 102 p. Cook, R C 1982. Estimating the mixing proportion of salmonids with scale pattern recognition applied to sockeye salmon {Oncorhynchus nerka ) in and around the Japanese land- based drifnet fishery. Ph.D. Thesis, Univ. Washington, Seattle, 214 p. Dixon, W J (editor) 1983. BMDP statistical software. Univ. Calif. Press, Berkeley, CA, 733 p. FlENBERG, S E 1981. The analysis of cross-classified categorical data. 2d ed. MIT Press, Cambridge, 198 p. Kennedy, W A 1970. Reading scales to age Pacific cod (Gadus macro- cephalus) from Hecate Strait. J. Fish. Res. Board Can. 27:915-922. Ketchen, K S 1970. An examination of criteria for determining the age of Pacific cod (Gadus macrocephalus) from otoliths. 721 FISHERY BULLETIN: VOL. 85, NO. 4 Fish. Res. Board Can. Tech. Rep. 171, 42 p. KOO, T S. Y. 1962. Age and growth studies of red salmon scales by geo- graphical means. In W. F. Thompson (editor). Studies of Alaska red salmon, p. 1-164. Univ. Washington Press, Seattle. Lai. H L 1985. Evaluation and validation of age determination for sablefish, pollock, Pacific cod and yello\Affm sole; opti- mum sampling design using age-length key; and implica- tions of ageing variability in pollock. Ph.D. Thesis, Univ. Washington, Seattle, 426 p. 1987. Optimum allocation for estimating age composition using age-length key. Fish. Bull., U.S. 85:179-185. Lai. H L . and S Y Yeh 1986. Age determination of walleye pollock (Theragra chalcogramma ) using four age structures. Bull. Int. N. Pac. Fish. Comm. 45:66-78. Macdonald, P D M , AND T J Pitcher. 1979. Age groups from size-frequency data: a versatile and efTicient method of analyzing distribution mix- tures. J. Fish. Res. Board Can. 36:987-1001. MOISEEV. P A 1953. Cod and flounders of far eastern seas. Izv. Tikhookean. Nanch No-Issled. Inst. Rybn. Khoz. Okeanogr. 40, 27 p. (Transl. Fish. Res. Board Can., Transl. Ser. 119, 596 p.) MOSHER, K. H 1954. Use of otoliths for determining the age of several fishes from the Bering Sea. J. Cons. Perm. Int. Expl. Mer 19:337-344. Pella, J. J , AND T L Robertson. 1979. Assessment of composition of stock mixtures. Fish. Bull., U.S. 77:387-398. Snedecor, G W , AND W A Cochran. 1967. Statistical methods. 6th ed. Iowa Univ. Press, Ames, 593 p. Westrheim. S J , and W. Shaw. 1982. Progress report on validating age determination methods for Pacific cod (Gadus macrocephalus). Can. MS Rep. Fish. Aquat. Sci. 1670, 41 p. Winer, B. J 1971. Statistical principles in experimental de- sign. McGraw-Hill Co., N.Y., 907 p. 722 LAI ET AL.: AGE DETERMINATION OF PACIFIC COD APPENDIX A.— ANALYSIS BY LOG-LINEAR MODEL In our study, there are three factors, ageing method (M), age class (A), and re- peatability (R ). Our sampling model is product-multinomial, using the terminology of Fienberg (1981, Sec. 3-2), since the number offish being aged is fixed for each ageing method after deleting unreadable or damaged age-structures. The aged fish were cross- classified into corresponding cells denoted by factors A and R (Table 1). Let y,jfi be the observed cell frequency for the j'th ageing method, thejth row of age class, and the /jth column of repeatability, and let m^^ be the expected value ofyijk ■ The general log-linear model (called the saturated model because it includes the highest three-factor interaction) for our three-way (5x8x2) contingency table is Qyk = \0g{m,jk) = JJL + \f^ + \/ + \f + Xj"" + ^fk"" + ^r + ^yk"" (A.l) where, as in the usual analysis of variance model, all effects sum up to zero over any subscript. Let 0 be the marginal mean of 8^;^ over the subscript which is replaced by " + " to indicate averaging, then the parameters in (Equation A.l) can be written as M- = 6 + + + ^ij = 6y+ ~ 9i++ ~ 9+7+ + 6 + + + kf - e,++ - e+++ x^^ - 8,+^ - e,++ - 0++^ + 8+++ kf = 0+7+ - 8+ + + kf^ ^ d+jk - Q+J+ -Q++k + Q+++ ( A.2) Xf = 0++k - 0+ + + Xjf/« = Q,jk - 8,,+ - 8,+^ + 8,+ + ~^+jk + 0+J+ + 0 + +* ~ 0 + + + - Log-linear models are "hierachical", i.e., higher-order interaction terms can be included only if related lower-order terms are included. For example, X^"*^ is not included unless X^, X^^, and X^'^ are all included. Once all expected cell frequencies (m^^) are estimated, the goodness-of-fit for the selected model can be tested using the likelihood ratio test statistic G^-2211y..^o,i^) (A.3) which has approximately a x^ distribution with degrees of freedom (df) = number of cells - number of parameters (Fienberg 1981, sec. 3-3 and 3-4). Using the partition property of G^, we can decide whether an effect or an interaction should be included. In Table 3, for example, Hq: k^^^ = 0 can be tested by examining G2 = 28.08 - 0.00 - 28.08. This is not significant at the 1% level (referred to a x^ distribution with df = 28). Similarly, G^ = 20.79 for Hq: k^^ = 0, which exceeds the upper 1% tail value of a x^ distribution with df = 4, and is rejected. This means that ^MAR ^j]j j^Q^ Yte included in the model but X^"* will. Hence, our best log-linear model is Equation (2). 723 SIDESCAN SONAR AS A TOOL FOR DETECTION OF DEMERSAL FISH HABITATS K W Able.i D C. Twichell,2 C B. Grimes,^ and R. S Jones'* ABSTRACT Sidescan sonar can be an efTective tool for the determination of the habitat distribution of commer- cially important species. This technique has the advantage of rapidly mapping large areas of the seafloor. Sidescan images (sonographs) may also help to identify appropriate fishing gears for differ- ent types of seafloor or areas to be avoided with certain types of gears. During the early stages of exploration, verification of sidescan sonar sonographs is critical to successful identification of impor- tant habitat types. Tilefishes (Lopholatilus and Caulolatilus) are especially good target species be- cause they construct large burrows in the seafloor or live around boulders, both of which are easily detectable on sonographs. In some special circumstances the estimates of tilefish burrow densities from sonographs can be used to estimate standing stock. In many localities the burrow and boulder habitats of tilefish are shared with other commercially important species such as American lobsters, Homarus americanus: cusk, Brosme brosme; and ocean pout, Macrozoarces americanus. Acoustic techniques have become important tools in fishery research in the last 20 years. Of these, sonar has proven useful in a number of related efforts for pelagic fisheries (Forbes and Nakken 1972) including the detection of fishes in the water column (Harden-Jones and McCartney 1962; Anderson and Zahuranec 1977) and estima- tion of fish numbers and biomass (Smith 1970; Hewitt et al. 1976; Suomala and Lozow 1980; Barans and Holliday 1983; Nakken and Venema 1983). More recent studies have demonstrated how sidescan sonar, in combination with acousti- cally tagged fish, can be used to evaluate trawling gear (Harden-Jones 1980). Sidescan sonar has been used infrequently to assess critical habitat for demersal fishery resources with the exception of an early attempt to map a herring (Clupea harengus) spawning area (Stubbs and Lawrie 1962). Our research has focused on detection of tilefish burrows (Twichell et al. 1985; Grimes et al. 1986; Able et al. 1987), but an outgrowth has been the identification of the habitats of other species. Here we describe the use of sidescan ^Center for Coastal and Environmental Studies and Depart- ment of Biological Sciences, Rutgers University, New Brunswick, NJ 08903; present address: Marine Field Station, Rutgers University, Tuckerton, NJ 08087. 2U.S. Geological Survey, Woods Hole, MA 02543. ^Southeast Fisheries Center Panama City Laboratory, Na- tional Marine Fisheries Service, NOAA, Panama City, FL 32407. ■^University of Texas Marine Science Institute, Port Aransas, TX 78373. sonar to map the extent and distribution of differ- ent habitat types and, in the case of tilefish, derive an estimate of standing stock and potential yield. TECHNIQUE Sidescan sonar is similar to low-angle, oblique, aerial photography except that the images (sono- graphs) are based on differences in the intensity of the reflected acoustic signal rather than the intensity of the reflected light (Belderson et al. 1972). The system consists of a towed vehicle (Fig. 1) in which is housed two sets of transducers that scan to each side, a conducting tow cable, a winch, and a dual-channel recorder for displaying the signals. The transducers are constructed so that their beams form a very narrow arc (1-2°) in the direction perpendicular to the ship's track, but a broad arc in the direction parallel to the ship's track (Fig. 1). As the ship moves, successive bands of seafloor are insonified, and in this way an acoustic areal map is recorded of the scanned area. We used a 100 kHz Klein^ sidescan sonar sys- tem. This system can resolve features as small as 0.5 m diameter (see Results) at a scanning range of 100 m to each side of the towed vehicle. The ^Use of trade names in this report does not constitute endorse- ment by the U.S. Geological Survey or the National Marine Fisheries Service. Manuscript accepted June 1987. FISHERY BULLETIN: VOL. 85, NO 4. 1987 725 FISHERY BULLETIN: VOL. 85, NO. 4 Figure 1. — Schematic diagram of a sidescan sonar vehicle being towed over the seafloor (upper) and resulting sonograph (lower) with images of trawl door tracks, tilefish burrow, gravel, and a boulder. sidescan vehicle was towed 10-15 m above the seafloor at speeds of 3-7 km/hour and was set to scan 100 m or 150 m to each side of the towed vehicle. At these speeds and scanning range, 0.6-1.4 km^ of seafloor could be mapped per hour. The sidescan sonograph signatures that char- acterize different habitat types are largely deter- mined by two conditions, topography and fine- scale roughness (in particular, differences in sediment texture). The signals received from tilefish burrows (Fig. 2) and boulders (Fig. 3) pro- vide good examples of differences in strength of the recorded signal due to topographic effects. A strong signal (dark) is received from the side of the feature facing the transducer while a weak signal or shadow (light) is received from the side sloping away from the transducer. Thus, boulders have the strong return nearest the transducer fol- lowed by a shadow (Fig. 3), while burrows appear as a shadow preceding the strong return (Fig. 2). Gravel gives a much stronger signal than silt be- cause of the many small facets facing the trans- ducers. Textural differences usually can be distin- guished from topographic differences because there is no shadow associated with them (Fig. 4). Although the sonograph is a map view of the seafloor, there are two distortions that must be compensated for when interpreting these images. The first is the across-track, slant-range distor- tion which results from distances being measured from the sidescan vehicle that is positioned above the bottom and not the zero line on the seafloor below the towed vehicle (Fig. 1). For this reason, the point on the seafloor directly below the fish is plotted away from the actual zero line by the dis- tance the fish is off the bottom (distance h on Figure 1). The second geometric distortion is in 726 ABLE ET AL.: SIDESCAN SONAR TO DETECT DEMERSAL FISH HABITATS \ o o CD O O LO CD O O ^ CO o CD CO o o o o CD CO X ea u o o ■n 1 es T3 C & c o c 3 o s .a -2 8 o c -c ■- a. t— o^ o >-J c 4-B a o i- 3 O C o o o C a. ca & o B o o c m ? o > CO m C o fcUD C '-3 CO -a c 3 o -a c 3 o >> C O O C cS 0) a. CO o C o CO C « u 10 -a in O o o o (lAI)39NVd 1NV1S 728 ABLE ET AL : SIDESCAN SONAR TO DETECT DEMERSAL FISH HABITATS o On U 0) o CO m jz _o 3 01 a o 0) ■*-> 01 -o ^ c 0) o CUD C E o u c o> .2 — -a o aj b. c u a> a-^ -c £ .ti t — >> > 0; c K o =^ ir e o> « 1^ O .c 0) ^ .S c is « 3 03 o C o -a >. J o J= o. CO o c I o o CO r- (N)39NVa iNvns I o o ro c US u CO 0) en a: O 729 FISHERY BULLETIN: VOL. 85, NO 4 the along-track direction and is due to the ship's speed. Normally the records are compressed in the direction the ship travelled relative to the true geometry. A circular feature will look ellipti- cal on the sonograph, with the long axis of the ellipse perpendicular to the ship's track, and a linear feature will look more perpendicular to the track than it really is. RESULTS AND DISCUSSION The Tilefish Example Our initial discovery (Able et al. 1982) that tile- fish, Lopholatilus chamaeleonticeps , construct large (up to 4-5 m diameter and 2-3 m deep), ver- tical burrows in the substrate (Fig. 5A) suggested that we might be able to detect these burrows with sidescan sonar. Since then we have success- fully determined tilefish occurrence, distribution patterns, and relative abundance based on side- scan sonar observations at the edge of the conti- nental shelf in the Mid-Atlantic Bight (Twichell et al. 1985; Grimes et al. 1986) and the upper slope off Florida (Able et al. 1987) (Table 1). Ver- ification of sonograph images as tilefish burrows (Figs. 2, 5A) was accomplished by in situ observa- tions fi-om the Johnson-Sea-Link submersibles (Askew 1985) (Table 1). As a result of these studies, we have demon- strated that sidescan sonar can consistently de- tect tilefish burrows both where the substrate consists of semilithified clay (Mid-Atlantic Bight: Twichell et al. 1985; Grimes et al. 1986) and softer carbonate muds (off east coast of Florida; Able 1987). During sidescan and submersible op- erations near Veatch Submarine Canyon, we de- termined that sidescan sonar could detect tilefish burrows as small as 0.5 m in diameter. Detection of small burrows on sidescan was confirmed when burrows in the area were measured in situ and found to be 35-65 cm in diameter (mean = 48 cm, sample number = 8). Under certain situations, tilefish abundance could be estimated directly from sonographs. Usually one tilefish is associated with each bur- row (Able et al. 1982), therefore sidescan sono- grams providing burrow counts could be used to estimate standing stocks in areas surveyed, with a modification of the area-density method (Ever- hardt and Youngs 1981). Frequency distributions of burrow density per unit area were log-normal, and there were considerable numbers of zero ob- servations (i.e., about 14-24% zero observations). Therefore we log,, transformed the burrow density data, and calculated the sample mean and vari- ance of the delta distribution according to Pen- nington (1983). We present sample estimates from our data for two different locations: Case 1: Middle Atlantic Bight in the vicinity of Hudson Submarine Canyon (number of obser- vations = 407, number of nonzero observations = 316, data from Twichell et al. 1985) based on the formula a where A^ = total number offish (burrows) in sur- delta-distribution mean number veyed area, — of burrows observed per unit area surveyed, A = total area surveyed, SD = standard devia- tion, and C.I. (confidence interval) = 95% (1.96 X SD) calculated from the delta-distribution vari- ance, thus Table 1.— Sidescan sonar observations of tilefish, Lopholatilus and Caulolatilus, burrows on the seafloor off the east coast of the United States. Sidescan trackline Depth No. of distance range verification Location Date (km) (m) dives Vicinity of Hudson July 1982; 100 90-200 6 Submarine Canyon, August 1982 1 Mid-Atlantic Bight Between Block and July-August 129 100-350 2 Veatch Submarine 1984 Canyons, Mid- Atlantic Bight Off Cape Canav- May 1984 36 100-250 2 eral, FL N = ^^^ • 0.407 km2 with SD = 123 and 95% km2 C.I. = 241 N = 1,041 ± 98 tilefish [in the surveyed area]. Case 2: South Atlantic Bight off Ft. Pierce, FL (number of observations = 46, number of nonzero observations = 40). The data was obtained with 167 kHz sidescan sonar from the research sub- mersible NR-1 (Able et al. unpubl. data). In this instance 730 ABLE ET AL.: SIDESCAN SONAR TO DETECT DEMERSAL FISH HABITATS Figure 5. — Photographs of A) tilefish, Lopholatilus chamaeleonticeps , and American lobster, Homarus amer- icanus. in a vertical burrow, and B) tilefish in boulder habitat. These are the same kind of habitats shown as sidescan sonographs in Figures 2 and 3 respectively. 731 FISHERY BULLETIN. VOL. 85, NO. 4 N 369 • 0.42 km2 with SD = 64 and 95% C.I. = 125 N = 154 ± 53 tilefish [in the surveyed area]. The estimates of A'^ in cases 1 and 2 could be extrapolated to the entire fishing grounds using an estimate of the area of the entire grounds to provide an estimate of standing stock. However, we believe that extrapolation to areas where no density data is available is imprudent for several reasons. First, the density of burrows in different locations is quite variable as shown in the two above examples, and density on the Middle Atlantic-Southern New England ground (case 1) varies over the grounds at least tenfold (Grimes et al. 1986). Second, some burrows, at least in the Middle Atlantic-Southern New England area, may not be occupied during all seasons of the year (Grimes et al. 1986). Although we do not have as much background knowledge for case 2, we know that burrow density at different sites off the Flor- ida east coast varied at least fivefold (Able et al. unpubl. data). Another possible source of error in using bur- row density to estimate tilefish stock size is that some burrows may be unoccupied. This should be of particular concern in exploited fishing areas. However, Twichell et al. (1985) and Able et al. (unpubl. data) have shown that abandoned bur- rows are filled by sedimentation relatively rapidly, i.e., less than one year, somewhat ameliorating the problem, at least over longer time periods. Perhaps the most constructive aspect of cases 1 and 2 is the opportunity to examine the error associated with sidescan sonar estimates of A^. These results show that the standard deviation varied from about 5 to 20% of the mean. Hen- nemuth (1976) found that the standard deviation in the numbers of different demersal species caught per tow within a stratum during stratified bottom trawl surveys approximately equalled the mean. Thus, this comparison suggests that area density estimates of abundance (calculated using the delta distribution) from sidescan sonar sur- veys will provide abundance estimates of much greater precision than trawl surveys. Reduced manpower needs and rapid application are addi- tional factors that favor the sidescan sonar methodology. However, because the sidescan methodology is only useful for certain three di- mensional habitats (e.g., reefs, rocks, and bur- rows) that would damage or make a trawl useless, application of the two techniques may usually be mutually exclusive. Tilefish are known to occur in other habitats (Grimes et al. 1986) such as boulder fields, which can be detected on sidescan sonographs (Figs. 2, 5B). Another habitat type (pueblo habitats, Warme et al. 1977; Grimes et al. 1986) occurs in the clay outcrops along the walls of submarine canyons (Figs. 4, 6A). Neither of these habitat types lend themselves to quantification of fish abundance. Recently, we have been able to con- firm that the burrows of other tilefish {Caulo- latilus spp.) are also detectable with sidescan sonar (Able et al. 1987; Figs. 6B, 7). Subsequent observations from a submersible confirmed that these burrows were occupied by C. microps and C. cyanops with frequent multiple occupancy. Given that it has now been demonstrated that represen- tatives of four of the five genera of tilefishes con- struct burrows (see Able et al. 1987), it is reason- able to suspect that all tilefish construct burrows. Thus, those larger species of commercial interest, such as red tilefish, Branchiostegus japonicus japonicus (Lim and Misu 1974), also may have burrows that are detectable by sidescan sonar. Other Examples and Possibilities As an outgrowth of our studies of Lopholatilus we have observed other species-specific habitats that can be detected with sidescan sonar. Ameri- can lobster, Homarus americanus, typically oc- cupy scour basins around large boulders (Cooper and Uzmann 1977; Valentine et al. 1980) as do cusk, Brosme brosme, and ocean pout, Macro- zoarces americanus (Valentine et al. 1980; Grimes et al. 1986), and these habitats also are detectable with sidescan sonar. American lobster (Fig. 5) and conger eels. Conger oceanicus, (Able et al. 1982; Grimes et al. 1986) have been ob- served in tilefish vertical burrows as well. Simi- larly, it would not be surprising if the habitats of other clawed lobsters are detectable with sidescan sonar. For example, H. gammarus from the east- ern North Atlantic is similar to H. americanus in that it is shelter seeking and occurs around boul- ders (Dybern 1973). In addition, recent in situ observations in the Gulf of Mexico have discov- ered that yellowedge grouper [Epinephelus flavolimbatus ) also occupy burrows and elongate trenches (R. S. Jones, E. Gutherz, and W. R. Nel- son, pers. obs.) that could easily be detected by sidescan sonar. 732 ABLE ET AL.: SIDESCAN SONAR TO DETECT DEMERSAL FISH HABITATS Figure 6. — Photograph of tilefish, Lopholatilus chamaeleonticeps , in A) pueblo habitat that is part of a clay outcrop and B) Caulolatilus tilefish in a burrow. These are the same kind of habitats as shown in Figures 4 and 7 respectively. 733 FISHERY BULLETIN: VOL. 85, NO. 4 -o CO 'm u C CS u to (U in Ol J= >-> OJ u t. c ^ CO CO a; o Q. C =0 ii o > CO O) X ,2 T3 ~ a* CO O a ~~ "— I. ■C o a o O {lAI)39NVd INVnS o o c CO CO CO o c o 01 t: D o 734 ABLE ET AL.: SIDESCAN SONAR TO DETECT DEMERSAL FISH HABITATS With sidescan sonar, detection and mapping of general habitat types such as rock outcroppings and wrecks, which support populations of com- mercially important species (e.g.. Grimes et al. 1982; Sedberry and Van Dolah 1984), could be done efficiently. Sidescan sonar also could prove very effective (see Wong et al. 1970) in mapping the distribution and relief of a coral reef, and other outcroppings which are often the habitats of groupers, snappers, porgies, and grunts. In addition to these specific examples, general characteristics of sidescan sonar are advanta- geous in detecting demersal fish habitats. The system has a wide effective search image (up to 150 m to each side for the 100 kHz Klein sidescan unit) that enables it to map large areas of the bottom during a single transect. With multiple transects a complete picture of the bottom can be obtained. Also, a sonograph could determine po- tentially appropriate habitats for several species simultaneously. For example, in our studies we have been able to detect boulders (potential lob- ster, tilefish, and cusk habitat) and vertical bur- rows (potential tilefish, lobster, and conger eel habitat) in the same transect of the sidescan sonar. Verification of the various images that appear on the sonograph is critical to successful opera- tion of sidescan sonar for fish habitat detection. We have been able to do this using observations from the Johnson-Sea-Link submersibles (Twichell et al. 1985; Grimes et al. 1986; Able et al. 1987). However, this is an expensive option and not generally available. Others have been able to verify sonograph targets from underwater photographs (Bouma and Rappeport 1984) or underwater television (Powles and Barans 1980). The simplest technique, and one that would offer the most information to a fishermen, is directed fishing at the location of sonograph targets of par- ticular interest. Even with these advantages, sidescan sonar op- erations are still expensive. However, a consider- able body of sonograph data already exists that has not been utilized by fishermen or fishery biol- ogists. A large number of sidescan sonar surveys have been conducted in North American waters in recent years, largely as a result of exploration for oil and related impact studies (Carpenter and Roberts 1979; Neurauter 1979; Carpenter et al. 1982). We have taken advantage of one of these surveys to identify possible tilefish burrows off the west coast of Florida, an area in which we had no prior experience. Individual burrows were clearly visible on sonographs (Neurauter 1979; target type No. 3, fig. 39) originally made to iden- tify geologic bedforms. ACKNOWLEDGMENTS A number of individuals and institutions as- sisted in this effort. Greg Kennedy (Harbor Branch Foundation, Inc.) and Greg Miller (U.S. Geological Survey) helped to operate the sidescan sonar equipment and kept it functioning prop- erly. The submersible and ships crews of the Johnson-Sea-Link submersibles and the RV Johnson and RV Sea Diver provided their usual professional operations. G. Scott (National Marine Fisheries Service, Miami) provided the delta distribution computer program. Helpful comments on an earlier draft of the manuscript were provided by Elizabeth Winget, Sally Needell, and Ron Circe (U.S. Geological Survey) and an anonymous reviewer. Denise Rucci and Patricia Eager prepared the manuscript. This work was supported by grants from the NOAA Office of Undersea Research, New Jersey Sea Grant, and Florida Sea Grant (No. NA80AA-D- 00038) and logistical support was provided by Harbor Branch Foundation, U.S. Geological Sur- vey and the Center for Coastal and Environmen- tal Studies, Rutgers University. 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. Twichell, R. S. Jones, and C. B Grimes 1987. Tilefishes of the genus Caulolatilus construct bur- rows in the sea floor. Bull. Mar. Sci. 40:1-10. Anderson. N R , and B. J. Zahuranec. 1977. Oceanic sound scattering prediction. Plenum Press, N.Y. Askew. T. M 1985. J ohnson -Sea-Link user's majwiaX. Harbor Branch Foundation Misc. Rep. No. 17, 29 p. Barans, C. A., and D. V. Holliday. 1983. A practical technique for assessing some snapper/ grouper stocks. Bull. Mar. Sci. 33:176-181. Belderson, R. H., N. H. Kenyon, A H. Stride, and A. R. Stubbs. 1972. Sonographs of the seafloor. Elsevier, N.Y., 185 p. Bouma, N. H , and M L. Rappeport. 1984. Verification of sidescan sonar acoustic imagery by underwater photography. In P. F. Smith (editor), Underwater photography. Scientific and engineering ap- plications, p. 279-294. Van Nostrand Reinhold Co. Carpenter, G. B., A P. Cardinell, D. K. Francois, L. K Good, R L Lois, and N T Stiles 1982. Potential geologic hazards and constraints for 735 FISHERY BULLETIN: VOL. 85, NO. 4 blocks in proposed North Atlantic OCS oil and gas Lease Sale 52. U.S. Geol. Surv. Open File Rep. 82-36, 51 p. Carpenter. G. B , and J W Roberts. 1979. Potential geologic hazards and constraints for blocks in Mid-Atlantic OCS oil and gas Lease Sale 40. U.S. Geol. Surv. Open File Rep. 79-1677, 191 p. Cooper. R. A., and J. R. Uzmann 1977. Ecology of juvenile and adult clawed lobsters, Homarus americanus, Homarus gammarus, and Nephrops norvegicus . In B. K. Phillips and J. S. Cobb (editors). Workshop on lobster and rock lobster ecology and physiology, p. 187-208. Commonwealth Scientific and Industrial Research Organization, Division of Fish- eries and Oceanography No. 7. Dybern. B I 1973. Lobster burrows in Swedish waters. Helgol. wiss. Meeresunters. 24:401-414. Everhardt. H W . AND W. D. Youngs. 1981. Estimating population size. In Principles of fishery science, p. 88-1122. Cornell Univ. Press. Ithaca, NY. Forbes. S. T.. and 0. Nakken (editors). 1972. Manual of methods for fisheries resource survey and appraisal. Part 2. The use of acoustic instruments for fish detection and abundance estimation. FAO Manu- als Fish. Sci. No. 5, Rome. Grimes, C. B., K. W. Able, and R S. Jones (1986). Tilefish (Lopholatilus chamaeleonticeps) habitat, behavior and community structure in Middle Atlantic and southern New England waters. Environ. Biol. Fishes 15(4):273-292. Grimes, C B , C S Manooch. and G R Huntsman 1982. Reef and rock outcropping fishes of the South At- lantic Bight and notes on the ecology of vermillion snap- per, Rhomboplites aurorubens , and red porgy, Pagriis pa- grus. Bull. Mar. Sci. 32(l):277-289. Harden-Jones, F. R. 1980. Acoustics and the fisheries: Recent work with sector-scanning sonar at the Lowestoft Laboratory. In F. P. Diemer, F. J. Vernberg, and D. Z. Mirkes (editors), Advanced concepts in ocean measurements for marine biology, p. 409-421. The Belle W. Baruch Library in Marine Science No. 10, Univ. South Carolina Press. Harden Jones. F. R , and B S McCartney 1962. The use of electronic sector-scanning sonar for fol- lowing the movements of fish shoals, sea trials of R.R.S. "Discovery II." J. Cons. Perm. Int. Explor. Mer 27:141- 149. Hennemuth, R. C. 1976. Variability in Albatross IV catch per tow. Res. Doc. ICNAF 761VI'104, 18 p. Hewitt, R P., P E. Smith, and J. C. Brown. 1976. Development and use of sonar mapping for pelagic stock assessment in the California current area. Fish. Bull., U.S. 74:281-300. LiM, P -Y , AND H Mlsu 1974. On the age determination of the Aka-amadai, Bran- chiostegus japonicus japonicus (Houtluyn) in the adja- cent waters of Tsushima Islands. Bull. Seikai Reg. Fish. Res. Lab. 46:41-51. Nakken, 0., and S. C Venema (editors). 1983. Symposium on fisheries acoustics. Selected papers of the ICES/FAO Symposium on fisheries acoustics. Bergen, Norway, 21-24 June 1982. FAO Fish. Rep., (300):331 p. Neurauter, T. W. 1979. Bed forms on the west Florida Shelf as detected with sidescan sonar. M.S. Thesis, Univ. South Florida, Tampa, 125 p. Pennington, M 1983. Efficient estimators of abundance, for fish and plankton surveys. Biometrics 39:281-286. Powles, H., and C a Barans 1980. Groundfish monitoring in sponge-coral areas off the southeastern United States. Mar. Fish Rev. 42(5):21- 35. Sedberry, G R , and R F Van Dolah 1984. Demersal fish assemblages associated with hard bottom habitat in the South Atlantic Bight of the U.S.A. Environ. Biol. Fishes 11:241-258. Smith. P E 1970. The horizontal dimensions and abundance of fish schools in the upper mixed layer as measured by sonar. In G. B. Forquar (editor). Proceedings of an International Symposium on Biological Sound Scattering in the Ocean, p. 563-600. Maury Center for Ocean Science, Wash., DC. Stubbs, a R , and R G G Lawrie 1962. Asdic as an aid to spawning ground investiga- tions. J. Cons. Int. Explor. Mer 27(3):248-260. SUOMALA, J R , JR , AND J B LOZOW 1980. Hydroacoustics and fisheries biomass estima- tions. In F. P. Diemer, F. S. Vernberg, and D. Z. Mirkes (editors). Advanced concepts in ocean measurements for marine biology, p. 353-365. Belle W. Baruch Library in Marine Science No. 10, Univ. South Carolina Press. Twichell, D C , C B Grimes, R S. Jones, and K W Able 1985. The role of bioerosion in shaping topography around Hudson Submarine Canyon. J. Sediment. Petrol. 55:712-719. Valentine, P C , J R Uzmann, and R A Cooper. 1980. Geology and biology of Oceanographer Submarine Canyon. Mar. Geol. 38:283-312. Warme, J E , R A Slater, and R A Cooper 1977. Bioerosion in submarine canyons. In D. J. Stanley and G. Kelling (editors). Submarine Canyon, fan and trench sedimentation, p. 65-70. Hutchinson and Ross Publ., Dowden. Wong, H K , W D Chesterman, and J D Bromhall 1970. Comparative sidescan sonar and photographic sur- vey of a coral bank. Int. Hydrogr. Rev. 47(2):ll-23. 736 THE EFFECTS OF BOTTOM TRAWLING ON AMERICAN LOBSTERS, HOMARUS AMERICANUS, IN LONG ISLAND SOUND Eric M. Smith and Penelope T. Howell' ABSTRACT American lobsters taken in the commercial trawl fishery in Long Island Sound, U.S.A., were in- spected for incidence of damage and immediate mortality associated with bottom trawling. Similar sampling was conducted in the pot fishery. American lobsters from trawl and pot catches were held in controlled conditions for 14 days to determine the level of delayed mortality associated with the two fisheries. Trawl-caught lobsters were exposed to subfreezing (-9.5°C) temperatures for periods from 30 to 120 minutes and then returned to seawater to determine the rate of freeze-induced mortality. Major damage rates due to trawling ranged from 12.6-14.0% during molting periods to 0-5.6% during intermolt periods. Delayed mortality ranged from 19.2% during the July molt to 1% during August and appeared to be related to the incidence of damage, molt condition, and temperature. Mortality of American lobsters held in subfreezing temperatures occurred after 30-minute exposure and reached 100% at 120-minute exposure. The American lobster, Homarus americanus, supports one of the most valuable commercial fisheries in the northwest Atlantic Ocean with landings of approximately 20,900 t per year val- ued at $115 million (Anonymous 1985). The fish- ery is conducted predominantly with traps or pots and, secondarily, with bottom trawl nets. Long Island Sound is a 2,908 km^ embayment of the Atlantic Ocean in southern New England, lying between Connecticut and New York at ap- proximately lat. 41°N. The Sound supports a spawning stock of American lobsters and a valu- able commercial lobster fishery, which, in 1985, generated landings of 1,134 t valued at $7.0 mil- lion for some 900 commercial fishermen. In Con- necticut, over 907^ of the commercial landings are taken by the pot fishery, and over 90^^ of Connect- icut's commercial fishermen are lobster pot fish- ermen. Connecticut trawlers, who catch Ameri- can lobsters in a mixed species bottom fishery, take <109c of Connecticut commercial landings and constitute about 10% of all commercial fish- ermen taking lobsters. Recreational lobstermen, both potters and scuba divers, totaled 2,440 in 1985 but only accounted for about 5% of total landings (CT DEP2). The resource is heavily exploited with annual exploitation rates ranging from 85 to 959c (Briggs 1985; Blake 1986). Principal management mea- 1 Bureau of Fisheries, Connecticut Department of Environ- mental Protection, P.O. Box 248, Waterford, CT 06385. 2Connecticut Department of Environmental Protection (CT DEP) unpublished fishery statistics. sures include prohibitions on the taking of fe- males bearing external eggs and the retention of any American lobster <81 mm carapace length (CD. These measures are intended to protect American lobsters from exploitation until they have reproduced at least once. During late 1982, commercial catch per unit effort rates doubled from those of the previous 5 years (CT DEP fn. 2). This increase stimulated a shift in directed fishing by some trawlers from mixed finfish to lobsters. The redirection of effort also generated competition for lobsters and fish- ing space, and an extremely emotional contro- versy arose between potters and trawlers. This paper addresses the resource considerations of the controversy, that is, the impacts of mobile trawl gear on lobsters. This study was designed to measure 1) the physical injury and immediate mortality in- curred by American lobsters in the trawl fishery; 2) the potential level of trawl-induced delayed mortality of American lobsters less than the min- imum length upon return to the water; and 3) the rate of mortality of American lobsters due to ex- posure to subfreezing air temperatures during winter fishing. METHODS Incidence of Damage Biologists made 63 trips aboard commercial stern trawlers from 12 to 26 m and 12 trips on pot vessels from 12 to 14 m. Except during January Manuscript accepted June 1987 FISHERY BULLETIN: VOL. 85, NO. 4. 1987. 737 FISHERY BULLETIN: VOL. 85, NO. 4 1984, trawl samples were obtained each month from July 1983 through January 1985 during tows of 1-3 h duration. Similar observations of trawl catches were obtained from cruises made by the CT DEP's research vessel (a 13 m stern-rigged trawler) from May through November in 1983 and 1984 during tows of 0.5-2 h duration. Com- mercial pot fishery sampling was conducted inter- mittently during summer and fall in both years. All samples were obtained in Long Island Sound west of long. 72°52'W. American lobsters were examined for evidence of physical damage, denoted as "new" (resulting from the current net tow or pot haul) or "old" (previously sustained). Old damage was inferred by the presence of discolored and healed tissue and by the absence of bleeding. New damage was further categorized as "major" and "minor". Major damage was defined as death, broken or crushed body parts or claw(s), and multiple in- juries, even when each individual wound was minor. Minor injury included walking leg loss or damage, antenna damage, minor breakage to a claw or rostrum tip, and recently autotomized claw(s). Recent autotomy (reflex amputation) was defined as a fresh partition without discoloration of the covering membrane at the "breaking plane" (Herrick 1909). Additional data were recorded on length, sex, shell hardness, presence of eggs, and absence of claws. Soft, newshell, and hardshell condition generally followed stages A, B, and C, respectively, as described by Passano (1960). In this paper, the term "newshell" in- cludes both soft and newshell lobsters (stages A and B). In comparing damage rates between commer- cial and research vessels, we found that hardshell American lobsters <81 mm CL were damaged more frequently in research than in commercial samples (x^ = 6.143). Since the research vessel used a smaller mesh (51 mm cod end) than com- mercial vessels (75-114 mm cod end), it caught a proportionately greater number of American lob- sters <81 mm CL (86% research vs. 72% commer- cial). Therefore, only lobsters >81 mm CL were analyzed because the greater proportion of sub- legal American lobsters and other biota in the small mesh sample might have resulted in an overestimate of damage relative to the commer- cial fishery, due both to intraspecific, agonistic behavior and to compacting of net contents. The incidence of major damage, compared be- tween pot and trawl gears, large and small trawl vessels, and commercial trawlers and research trawler were analyzed by chi-square goodness of fit (Sokal and Rohlf 1969). Delayed Mortality Damaged and undamaged American lobsters less than the minimum length were taken from the same area of Long Island Sound, on the same date, from commercial pot vessels and from either commercial trawl vessels or research vessel. Sam- ples were obtained in November 1983 and May, July, and August 1984. They were transported in tanks with circulating seawater to a laboratory equipped with an open circulating seawater sys- tem and were observed twice daily for a period of 14 days. Lobsters were fed daily with assorted fish species. Dead lobsters were removed upon observation. Lobsters were selected nonrandomly from catches to ensure that damaged lobsters were adequately represented in the tests. Since damaged lobsters often made up only a small per- centage of the total, our sample consisted of pro- portionally more damaged lobsters than their ac- tual proportion in the catch. American lobsters with minor damage were combined with undamaged ones for analysis be- cause undamaged lobsters rarely experienced de- layed mortality (2/309) and those with minor damage never did. To estimate total delayed mor- tality from each gear, the observed delayed mor- tality rate in each test category (undamaged hardshell, damaged hardshell, and newshell) was extrapolated to the corresponding trawl catches of undamaged, damaged, and newshell lobsters ob- served during the same months in 1984. The lab- oratory mortality rate for newshell lobsters was applied to all newshell lobsters in commercial catches rather than only damaged newshell lob- sters since internal (unobserved) damage may have occurred. Field data from 1984 commercial catches were used to estimate total trawl-induced mortality because the proportion of newshell lob- sters in commercial catches was highest during 1984, and we believe that that year's data most accurately represent the proportion of newshell lobsters taken in the fishery. The chi-square goodness of fit (Sokal and Rohlf 1969) with Yates' correction for continuity (Zar 1974) was utilized to determine whether lobster mortality rates differed between large and small trawl vessels and between commercial and re- search trawl vessels. The effect of season and molt condition was examined using a log-likelihood ratio test (G-test; Sokal and Rohlf 1969). Seasonal 738 SMITH and HOWELL: BOTTOM TRAWLING EFFECTS ON AMERICAN LOBSTERS changes were defined by water temperature vari- ation whereas molt condition (newshell vs. hard- shell) was based on the bimodal distribution of lobster molting (June-July and October- November) observed in Long Island Sound (Lund et al. 1973). The relationship between tempera- ture and shell condition was evaluated by com- paring data derived from postmolt periods follow- ing the summer (warming seawater temperature) vs. fall (cooling seawater temperature) molts. Postmolt lobsters recovered from the capture process in flowing seawater of 20°C in the July test and 15°-16°C in the November test. Intermolt lobsters in May recovered in 12°C seawater and, in August-September, at 22°C. In all cases, sea- water temperatures reported here were elevated approximately 1°-2°C from passage through the circulation system. The trawl-caught sample included 74 lobsters taken from research vessel catches. Comparison of delayed mortality rates between research and commercial samples revealed no significant dif- ference (x^ = 0.305) so the samples were pooled. Delayed mortality rates from samples taken from small (12 m) and large (20 m) trawl vessels in November 1983 and August 1984 were not signif- icantly different (x^ = 0.05), therefore, large and small vessel data were also pooled. All damage observations were recorded in the same manner as reported above. Mortality Due to Freezing Seventy hardshell, undamaged, trawl-caught American lobsters <81 mm CL were held in an open circulating seawater system at 8°C. On 15 January 1985, three groups of 20 lobsters were exposed in a wooden box (2 m x 1 m x 0.3 m) to ambient air temperature of -9.5°C for periods of 30, 60, and 120 minutes. The remaining 10 lob- sters were held without exposure as a control. After the prescribed time had elapsed, the test lobsters were returned to the holding system and observed several times during the following 48-h period for incidence of mortality. RESULTS Incidence of Damage Due to Trawling Damage to American lobsters of all sizes caused by commercial trawling throughout the 19-mo sampling period is summarized in Table 1. The monthly incidence of major damage (including immediate mortality) varied seasonally, from 0 to 14.0% in the trawl fishery and from 0 to 3.5% in the pot fishery. Minor damage ranged from 0.9 to 8.1% in the trawl fishery and from 0 to 10.7% in the pot fishery. Differences between pot and trawl damage rates were compared only for months in which both gears were sampled. The incidence of major damage was significantly greater for trawl samples in July and October-November, but not in August or September (Table 2). The incidence of damage due to the two gears Table 1. — Incidence of damage to American lobsters taken by commercial otter trawl from July 1983 through January 1985.1 Major damage No major damage Minor Un- damage damaged Month Broken parts Multiple injuries Dead Total (n) Jan. (n) (%) 1 1.0 0 0 2 2.0 6 6.0 91 91.0 100 Feb. (n) (%) 0 0 0 0 0 0 1 6.7 14 93.3 15 Mar. (n) (%) 6 1.0 3 0.5 0 0 9 1.6 557 96.9 575 Apr. in) (%) 19 1.3 8 0.5 4 0.3 25 1.6 1,461 96.3 1,517 May (n) (%) 22 1.9 7 0.6 0 0 21 1.9 1,087 95.6 1,137 June (n) (%) 1 0.5 1 0.5 4 1.8 2 0.9 216 96.3 224 July (n) %) 72 6.9 44 4.2 16 1.5 57 5.5 854 81.9 1,043 Aug. in) (%) 18 1.6 4 0.4 1 0.1 36 3.2 1,056 94.7 1,115 Sept. in) %) 49 3.6 4 0.3 23 1.7 36 2.7 1,242 91.7 1,354 Oct. n) %) 94 5.5 47 2.7 25 1.5 116 6.8 1,426 83.5 1,708 Nov. n) %) 133 6.0 128 5.8 50 2.2 181 8.1 1,738 77.9 2,230 Dec. n) %) 77 6.5 20 1.7 13 1.1 68 5.8 998 84.9 1,176 'Data from months in successive years are pooled. Table 2. — Incidence of damage to American lobsters taken by commercial pot and otter trav^^l gears, 1983-84. Pot fishery Trawl Major damage fishery N Month Major damage N X2 July 0.9% 2,165 12.7% 1,043 212.2" August 1 .3% 1,512 2.1% 1,115 1.8 September! 3.5% 424 5.9% 1,344 2.4 Oct.-Nov.i 0.6% 661 14.4% 2,672 96.8" ■"significantly different (P < 0.001). '1983 data only (no pot fishery samples taken in 1984). 739 FISHERY BULLETIN: VOL. 85, NO. 4 was analyzed with respect to shell hardness and carapace length (<81 mm vs. >81 mm). Size- specific damage to newshell American lobsters taken in the trawl fishery was significant with smaller lobsters incurring more damage (43% vs. 30%, x^ = 6.64, P =0.01). There were no size- specific differences in damage to newshell lob- sters taken in the pot fishery or to hard-shelled ones in either fishery. Trawled egg-bearing female American lobsters (always hard shelled) incurred 1.9% major dam- age, no immediate mortality, and 2.1% minor damage throughout the sampling period in = 909). Eggbearers >81 mm CL (n = 585) ex- hibited 2.2% major damage and 2.4% minor dam- age while those <81 mm CL (n = 306, 18 size unspecified) incurred 1.3% major damage and 1.6% minor damage. Pot-caught eggbearers sus- tained 0.9% major, and 0.8% minor damage throughout the sampling period in = 1,926). One of 1,926 pot-caught egg-bearers (0.05%) died on deck. These data suggest that the damage and immediate mortality to gravid American lobsters associated with both fisheries is minimal. From September 1983 through December 1984, trips were made with both large (15-26 m) and small (12-14 m) trawl vessels during eight differ- ent months. In two of those months, damage rates were equal, in two they differed by <1%, and in four they ranged from 7.3% higher for small ves- sels to 13.7% higher for large vessels. The differ- ence exhibited in the four months for which devi- ations of more than 1% were observed was not significant (x^ = 0.019, P < 0.5) indicating that damage to trawled American lobsters is inde- pendent of vessel size. Trawl-induced damage occurred at similar rates in cold-water vs. warm-water intermolt pe- riods (2.2% January-June vs. 3.1% August- September) and between cooling and warming postmolt periods (11.5% October-December vs. 12.6% July; Table 1). This suggests that damage due to trawling is more a function of shell condi- tion than water temperature. Delayed Mortality From November 1983 to August 1984, 526 Table 3. — Estimated mortality to American lobsters <81 mm CL (delayed and imme- diate) due to otter trawling in 1984. Actual delayed mortality is computed by multiplying the percent occurrence of each damage category in trawl catches by the laboratory delayed mortality rate for that category. Total mortality rate is the sum of the actual delayed mortality plus immediate mortality observed on deck. Laboratory Actual Observed Total Trawl delayed delayed immediate mortality catches mortality mortality mortality rate IVIonth % rate % % % % fVlay {N = 608, trawler catches; N Hardshell 41 , laboratory samples) undamaged 96.5 0 0 0 0 damaged 2.0 85.7 1.7 0 1.7 Newshell 1 .5 33.3 0.5 0 0.5 2.2 0 2.2 July (N = 533, trawler catches; N = 40, laboratory samples) Hardshell undamaged 81.8 10.5 8.6 0 8.6 damaged 9.9 85.7 8.5 2.1 10.6 Newshell 6.2 33.3 2.1 0 2.1 19.2 2.1 21.3 August (N = 456, trawler catches; N = ' 146, laboratory samples) Hardshell undamaged 98.2 0 0 0 0 damaged 0.7 85.7 0.6 0 0.6 Newshell 1.1 33.3 0.4 0 0.4 1.0 0 1.0 November {N = 408, trawler catches; N = 147, laboratory samples) Hardshell undamaged 86.3 0 0 0 0 damaged 3.9 42.4 1.7 1.2 2.9 Newshell 7.8 33.3 2.6 0.8 3.4 4.3 2.0 6.3 740 SMITH and HOWELL: BOTTOM TRAWLING EFFECTS ON AMERICAN LOBSTERS American lobsters <81 mm CL, taken from pot and trawl vessels, were held to estimate trawl- induced delayed mortality (Table 3). Of 374 trawl-caught lobsters held in the laboratory, 47 (12.6%) had sustained major damage. Eighteen were newshell and were treated as one category regardless of damage sustained. Two of 309 (0.6%) undamaged, hardshell trawl-caught lob- sters died whereas 55.3% of damaged ones died within the 14-d period, most within the first 7 days. Six of 18 newshell lobsters died (33.3%). Of 153 pot-caught lobsters, 8 (5.2%) had major dam- age; none were newshell. No pot-caught lobsters experienced delayed mortality. Laboratory delayed mortality rates for dam- aged hardshell American lobsters were pooled in May, July, and August because of the small sam- ple sizes (n = 14). Delayed mortality rates ranged from 50 to 100% in the three months. The mean (85.7%) was applied to the proportion of damaged hardshell lobsters in the commercial trawl fish- ery in each of those months. In November, a 42.4% delayed mortality rate (n = 33) was ap- plied to the damaged lobster category in Novem- ber commercial samples. Although only two un- damaged hardshell lobsters died, both occurred in July (2 of 19). This proportion (10.5%) was applied to the proportion of undamaged hardshell lobsters in the July commercial catches. Since no delayed mortality occurred to undamaged hardshell lob- sters in May, August, or November, no delayed mortality rate was applied to that category for those months. The delayed mortality rate for newshell lobsters (33.3%, n = 18) was applied to the proportion of newshell lobsters in commercial catches (1-8%). Immediate (on-deck) and delayed mortality rates for lobsters <81 mm CL and for each shell condition (undamaged hard, damaged hard, and newshell) were summed, resulting in an estimate of total mortality for each season (Table 3). Estimated trawl-induced mortality rates for the four sample periods were tested to determine the relative importance of high seasonal water temperatures (July and August) and molting (November and July). November and July had significantly higher rates of mortality than May and August (Tables 3, 4), indicating that molt condition (July and November) is more important in determining the extent of mortality than water temperature. There was no significant difference between May and August delayed and immediate mortality estimates (Table 4) despite a 10°C dif- ference in water temperature. However, warm temperatures appeared to increase the incidence of mortality after molting since the July rate was significantly higher than the November rate (21.3% vs. 6.3%, Table 4). Table 4. — Log-likelihood ratio test of expected trawl- induced mortality to American lobsters <81 mm CL, by season. Underlining denotes no significant difference at P = 0.05. x2 = 178.4* overall, 44.8* (July- November), August). 11.8* (November-May), and 1.8 (May- May August November July Water temperature cold warm cold warm Molt condition nonmolting molting Estimated mortality 2.2% 1 .0% 6.3% 21 .3% *= P< 0.005. Mortality Due to Freezing Exposing undamaged American lobsters to am- bient air temperatures of -9.5°C produced no mortality at 30-min exposure, and 70% and 100% mortality at 60 minutes and 120 minutes, respec- tively. Damaged lobsters were not tested since they occur so infrequently during cold-water peri- ods of the year (Table 1). Freezing temperatures may also induce reflex amputation (autotomy) of claws. One lobster of 20 in the 30-min sample and four American lobsters of 20 in the 60-min sample autotomized one claw. DISCUSSION Jamieson and Campbell (1985) found that sea scallop dragging in eastern Canada could damage American lobsters, but since the sea scallop and lobster fisheries generally were not simultaneous (lobsters tended to emigrate from the scalloping areas each season prior to the advent of the drag fishery), the use of scallop gear over beds that hold lobsters at other times posed no significant impact to the resource. Scarratt (1972) observed a similar situation in the eastern Canadian Irish moss rake fishery. Although American lobsters did suffer damage from the gear, most lobsters emigrated before the moss harvest season, so the damage associated with the gear was minimal. In western Long Island Sound, the lobster pot and mixed species trawl fisheries both operate throughout the year. Since the seabed in this area is relatively uniform and generally free of 741 FISHERY BULLETIN: VOL. 85, NO. 4 obstructions to trawling, lobsters can be taken simultaneously with both gears, raising concerns that trawl gear is detrimental to the resource. In Narragansett Bay, RI, using a 13 m stern trawl-rigged research vessel, Ganz (1980) found low immediate mortality to trawl-caught Ameri- can lobsters and low damage rates during inter- molt periods, and higher damage rates immedi- ately following molting. He observed that hardshell lobsters were not likely to sustain criti- cal injuries but postulated that commercial-scale operations might produce higher damage rates than the research vessel because of net com- paction. We found that major damage (including immediate mortality) was greatest during molt- ing periods, ranging from 12.6% in July to 14.0% in November. Hardshell (intermolt) lobsters suf- fered little damage by commercial trawling, with the monthly incidence of major damage and im- mediate mortality <3% from January through June and in August. Minor damage, including autotomy of claws, ranged from 0.9 to 8.1% and was greatest in October-November. Minor dam- age was <2% from March through June. The incidence of immediate mortality by month never exceeded 0.5% in the pot fishery or 2.2% in the trawl fishery. Moreover, we found that dam- age rates were independent of vessel size, whether in the commercial fishery or between 12 m commercial and 13 m research vessels (see below). Newly molted American lobsters were damaged by both trawl and pot gears but trawl- ing caused greater damage. Spurr (1978) found experimental otter trawl- induced injury in summer to be greater in July than in September; however, since he aggregated minor damage with major damage and did not provide sample sizes, quantitative comparison of major damage between the two studies is not pos- sible. Spurr concluded that lobster damage due to trawling in winter would be minor, a conclusion supported by our data. Ganz (1980) suggested that damaged, sublegal American lobsters might suffer mortality upon release, although he did not investigate this pos- sibility. We found that delayed mortality ap- peared to be influenced by the condition (inci- dence of damage) of American lobsters. Only 2 of 309 undamaged, hardshell lobsters experienced delayed mortality after trawling, compared to 26 of 47 (55.3%) damaged, hardshell, trawl-caught lobsters. Delayed mortality never occurred to lob- sters with minor damage or autotomized claws. Although the sample size was small in = 18), only one-third of all trawled newshell American lobsters experienced delayed mortality. While both trawl and pot gear damage lobsters, trawl-caught lobsters alone sustained delayed mortality. One of the initial concerns about trawl- ing was that visibly undamaged American lob- sters less than the minimum length, returned to the water after trawling, would suffer a high level of unobserved mortality. Our results indicate that such mortality rarely occurs to undamaged Amer- ican lobsters; consequently, potential delayed mortality may be ascertained simply by inspec- tion of the incidence of major damage in the catch. Of the two molting periods in Long Island Sound, a higher rate of mortality was observed in July than in November, possibly related to the warmer postmolt water temperatures which occur in summer compared with late fall. Damage rates during postmolt periods were similar (12.6% July vs. 14.0% November) notwithstanding the difference in water temperature. These results suggest that the resistance of trawled postmolt American lobsters may be lowered by warm sea- water temperatures and that such temperatures may increase the occurrence of delayed mortality independent of the incidence of damage. Mean values of immediate and delayed mortality for all samples taken during intermolt periods were neg- ligible (<1.0% and <2.0%, respectively). There appeared to be some variability in mor- tality depending on the type of damage sustained. Damage to the abdomen and carapace was almost always lethal (100% and 92%, respectively), while broken parts such as claws or rostrum resulted in 25% and 50%' mortality, respectively. However, small sample sizes precluded definitive analysis. Ganz (1980) speculated that the most signifi- cant impact of trawling might be related to the cumulative effect of trawling and damaging sub- legal American lobsters, and subsequently re- leasing them to be taken again. In discussing this possibility, he reported an immediate claw loss (cull) rate of 3.5% and a prior cull rate of 8.8%. This is a valid concern, and one which should be considered in both trawl and pot fisheries. For example, Smith (1977) reported a cull rate of 26% in an area of Long Island Sound which had not been trawled during recent years and a rate of 23% in an area lightly fished with trawls, sug- gesting that a high cull rate can occur in the ab- sence of trawling (both areas were heavily fished with pots). These observations, while higher than Ganz's, included both new and old claw loss as well as lobsters with regenerated claws. In the 742 SMITH and HOWELL: BOTTOM TRAWLING EFFECTS ON AMERICAN LOBSTERS present study, minor damage (which included im- mediate claw loss) ranged from 0.9 to 8.1% per month in the trawl fishery and from 0 to 10.7% in the pot fishery. While the mortality of trawled newshell Amer- ican lobsters was high during the two molting periods, they represented such a small percentage of the total trawl catch (1-8%) that the estimated total mortality to the entire catch was little changed by their presence. Similarly, while dam- aged lobsters sustain a high rate of mortality, it is their proportion in the catch which determines the rate of additional mortality experienced by the population. No truly soft American lobsters (stage A, after Passano 1960) were observed in the pot catches sampled for delayed mortality estimates, and they occurred rarely in other commercial pot sam- ples (up to 3% in July 1984). They were observed infrequently in trawl catches as well (up to 7.8% in November 1984). The low incidence of soft lob- sters in commercial catches is probably due to reduced mobility during the shell hardening proc- ess (Herrick 1909) when American lobsters are most vulnerable to damage and predation. No significant difference was observed in dam- age rates or delayed mortality based on sampling of vessels <15 m or >15 m. There was no differ- ence in delayed mortality rate between 13 m re- search vessel samples and either 12 m commer- cial or 15-26 m commercial trawlers. Finally, there was no difference between damage rates to American lobsters >81 mm CL taken by research and commercial vessels in October-November 1984, the only period for which comparable data (tows of 2-h duration) were available for both ves- sel categories. The former results suggest that trawl-induced damage and mortality is indepen- dent of vessel size; the latter suggests that obser- vations of fishery-induced damage made by biolo- gists were representative of actual fishing conditions. While American lobsters may succumb to sub- freezing air temperatures, consideration of this source of mortality is a function of both tempera- ture and exposure time and must be judged on the behavior of the fishery in question. Edwards and Bennett (1980) reported a survival rate of 42-85% for Nephrops norvegicus after 1-h exposure to air, noting that favorable weather conditions and the type of vessel used were factors contributing to higher survival. In the present study, American lobsters exposed to -9.5°C air temperatures for 30 minutes all survived. Those exposed for 120 minutes all succumbed. Intermediate exposure (60 minutes) produced intermediate results (30% survival). In Long Island Sound, field observa- tions during a 19-mo period suggested that opera- tors in the trawl fishery commonly sorted the catch within 15-45 minutes. Our design tended to maximize the debilitating effect of freezing temperatures. Trawl net con- tents are commonly released onto the vessel deck in a pile. Consequently, organisms on the inside of the pile are protected from cold air tempera- tures. In our test, all American lobsters were placed on a flat table with 0.3 m high sides; thus, all lobsters were equally exposed to freezing tem- peratures and none were able to benefit from the "piling" of a normal catch. As a consequence, this experiment very likely overstated the impact of subfreezing air temperatures on commercially taken American lobsters. An additional consideration beyond the di- rected fishery is the mortality to "sublegal" American lobsters (those <81 mm CL) and the damage to lobsters during molting periods which may result from a mixed species trawl fishery which includes a so-called incidental catch or "bycatch" of lobsters. As with freeze-induced mor- tality, this factor must be judged based on the performance of the fishery in question. In Connecticut, trawlers reported <10% of the commercial American lobster landings reported by all fishermen during the period 1982-86 (CT DEP fn. 2). However, given the controversy sur- rounding trawling during this investigation and the weaknesses inherent in catch reporting sys- tems during periods of controversy (Matlock 1986; Ferguson 1986), a number of independent methods of observation were used to determine the actual impact to the resource associated with trawling. Unannounced boardings by Connecti- cut Conservation Officers were utilized to esti- mate the true magnitude of lobster catches. Biol- ogists made sampling trips on the vessels of commercial trawl fishermen fishing in Long Is- land Sound to document the fishery-induced inci- dence of damage. Research vessel sampling was used to estimate a fishery-independent rate of trawl-induced damage. The results of these obser- vations suggested that the magnitude of Ameri- can lobster catches per trip was about the same as those reported in mandatory logbooks and, except during molting periods, trawl-induced damage was minimal. Observations in controlled laboratory condi- tions were utilized to estimate the delayed 743 mortality which might be expected to occur to sublegal American lobsters returned to the water after trawling. These observations indicated that this source of mortality should only be a concern during molting periods. Since delayed mortality to sublegal American lobsters occurs to a signifi- cant degree only during molting periods, an inci- dental limit of some number of lobsters per day for trawlers during those periods represents an effective means to deter the directed fishery and protect sublegal American lobsters while allow- ing the finfish fishery to continue. There has been considerable controversy in New England regarding the effects of trawling on the American lobster resource. This study pro- vides three results of assistance to fishery man- agers dealing with this question. First, both pot and trawl gear damaged American lobsters, but trawl-induced damage occurred more frequently, and particularly during molting periods. How- ever, damage was not always lethal and visibly undamaged lobsters virtually never sustained de- layed mortality. Second, during molting periods, mortality caused by trawling reached 6-21%, de- pending on season. Delayed mortality was influ- enced most by the degree of damage sustained by the lobster. Therefore, while delayed mortality may be of considerable consequence to the re- source during molting periods, it can be estimated by inspection of the condition of lobsters in the catch. Third, during intermolt periods, both im- mediate and delayed mortality due to trawling occurred infrequently { \- 4 Li— ^ Figure 3. — Selected Pacific whiting fillet areas examined for cooked texture and pseudocyst intensity. 747 FISHERY BULLETIN: VOL. 85, NO. 4 thickness, cooked in an oven at 375°F for 20 to 50 minutes. The texture of the fish was then deter- mined organoleptically. Mechanical texture was determined on 180 randomly selected fish, in which duplicate shear press readings were taken on 15 g of muscle tissue removed from the dorsal portions of the fish samples to correlate with the portions taken for sensory evaluations. The successful use of me- chanical texture analysis to correlate with the sensory texture of Pacific whiting was previously reported by Nelson et al. (1985). The tissues were teased into flakes which were leveled to a height of 8 mm in a Kramer shear-compression cell which was reduced in size to 29 mm wide, 71 mm long, and 64 mm deep to accommodate the sam- ple. The assembly, consisting of four blades, was similar to the one described by Bilinski et al. (1977) and used by Tsuyuki et al. (1982) to evalu- ate objectively the texture of Pacific whiting. The cell operated in conjunction with the Food Tech- nology Corporation FTA 3000 transducer^, TP-4 Texturepress, and the TR-5 Texturecorder. A plot was made of the force required to drive the blades through the sample at a ram speed of 1 cm per minute and at a set recorder range. The peak force in pounds per 15 g tissue was calculated from the plot (Bourne 1982). Culling was performed on partially thawed halves offish. Pseudocysts are visible as white or black threads of varying intensities imbedded longitudinally along the muscle fibers. The culling categories were modeled after the scheme of Patashnik et al. (1982) for both white and black pseudocysts as none, light (<20%), moderate (20 to 30%), and heavy (>30%) as determined visually, based on the percent area of fillet affected. Only fish over 27 cm were used for all analyses, since in commercial operations fish smaller than 27 cm would not likely be taken because domestic and foreign fishermen use 50 mm (2 in) to 100 mm (4 in) cod end mesh size as regulated by the Pacific Coast Groundfish Fishery Management Plan.5 Data representing sensory textures and Kudoa pseudocyst counts (made on a total of 562 fish exclusive of fish under 27 cm) were analyzed on ■^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. y. Wall, REFM Division, Northwest and Alaska Fisheries Center, National Marine Fisheries Service, NOAA, 7600 Sand Point Way N.E., BIN C15700, Seattle, WA 98115, pars, com- mun. March 1985. the NWAFC Burroughs 7800 computer system using the SPSS software package described by Nie et al. (1970) and SPSS Update 7-9 (Hull and Nie 1981). Fitted regression curves between cooked tex- ture values and white and black pseudocyst counts were drawn using robust locally weighted regression analyses described by Cleveland (1979). The method was used to smooth scatter- plots by calculating a polynomial fit to data using weighted least squares. RESULTS AND DISCUSSION Relationship Between Sensory Texture Ratings and Shear Press Values Since any nonsensory evaluations of a fishery product must ultimately relate to the products' intrinsic organoleptic properties, emphasis in this study was placed upon taste tests despite the inherent fatigue factor associated with testing large numbers of samples. No consumer-type panel was carried out to test the accuracy of the texture evaluation, because fish with abnormal o Q. 7 - > o • • t* • • •••9 • ••• • •••••• Al* t • • •• • •••»* I I • *•«•« t •• •• m\ • • t« ••• •• I • • I • I t . • • • ••• I r = 0.90 20 40 60 80 100 Shear press in pounds / 15 grams tissue 120 Figure 4. — Scatterplot and Spearman correlation coefficient (r) between sensory texture rating and shear press force of cooked flesh of Pacific whiting fillets. 748 KUDO ET AL.: TEXTURE QUALITY OF PACIFIC WHITING texture were purposely included in the present study to determine the extent of the problem. On randomly selected samples, shear press measure- ments were run concurrently as an objective sup- port for the sensory data. A scatterplot and the Spearman correlation co- efficient (nonparametric) determined for the sen- sory texture ratings and shear press values shown in Figure 4 was produced by the SPSS soft- ware. The high degree of correlation (r = 0.90 at P = 0.001) was found to support the credibility of sensory texture evaluation in this study. Effect of White and Black Pseudocysts and Kudoa Species Upon Sensory Texture Fitted regression curves for sensory texture values and white pseudocyst counts are shown in Figure 5A. Mixed infections include both K. pan- iformis and K. thyrsitis pseudocysts in the same fish. The curve representing the white mixed in- fection shown superimposed on that of the white K. paniformis curve suggests that most of the white mixed pseudocysts consisted of K. pani- S E N S 0 R Y T E X T U R E B S E N S 0 R Y ■j E X T U R E 19 e- 7 5- 5 8 2 5 0 0 WHITE PSEUDOCYSTS OPANIFORMIS X ^XTHYRSITIS A- A MIXED 0 200 400 600 800 KUDOA PSEUDOCYST COUNTS PER GRAM TISSUE 10 0—1 7 5 5 0 2 5 0 0 BLACK PSEUDOCYSTS G- -OPANIFORMIS X XTHYRSITIS A- AMIXED "T T 0 200 400 600 800 KUDOA PSEUDOCYST COUNTS PER GRAM TISSUE Figure 5. — Fitted regression curves between cooked texture of Pacific whiting and intensity of (A) white Kudoa pseudocysts and (B) black pseudocysts. 749 FISHERY BULLETIN: VOL. 85, NO. 4 formis. Correlation values for these two curves were r = -0.79 and -0.80, respectively (Table 2). Pseudocyst counts of white K . thyrsitis were <10 and did not appear to adversely affect the texture of the Pacific whiting (Table 3A). The low count is thought by Kabata and Whitaker (1981) to be the result of an evolutionary development of defense mechanisms in the host fish. The correlation coef- ficient representing sensory textures and white K. thyrsitis infection was 0.12. This suggests that K. thyrsitis plays an insignificant role in the tex- ture quality of Pacific whiting, a view also held by Kabata and Whitaker (1981). Fitted regression curves between sensory tex- tures and black pseudocyst counts are shown in Figure 5B. Mixed black pseudocyst counts did not Table 2. — Spearman correlation coefficient between sensory tex- ture rating and Kudoa pseudocysts in Pacific wtiiting caughit during the summer of 1983. Sample Correlation Signifi- Sensory texture witfi: size coefficient cance White pseudocysts total population 562 -0.74 0.001 Black pseudocysts total population 562 -0.38 0.001 K. paniformis white pseudocysts 201 -0.80 0.001 black pseudocysts 201 -0.66 0.001 K. thyrsitis white pseudocysts 191 -0.12 0.05 black pseudocysts 191 0.07 0.15 Mixed infection white pseudocysts 259 -0.79 0.001 black pseudocysts 259 -0.49 0.001 Table 3. — Percentage of sensory texture ratings for (A) each Kudoa species, (B) fork length, and (C) sex composition for Pacific whiting caught dunng the summer of 1983. Sample Sensory texture rating ( ;%) Description size Normal Soft Abnormal Total (A) Kudoa species K. paniformis 144 17 5 3 26 K. thyrsitis 125 22 0 0 22 fvlixed infection 190 23 7 4 34 None 103 18 0 0 18 Total 562 80 12 7 100 (B) Fork length composition of whiting Small (27-39 cm) 144 93 5 2 100 Medium (40-49 cm) 172 79 17 4 100 Medium large (50-59 cm) 170 74 14 12 100 Large (60-80 cm) 76 72 15 13 100 All lengths 562 80 13 7 100 (C) Sex composition of whiting Female 332 81 12 6 100 Male 227 79 13 8 100 Total 559 80 13 7 100 follow the pattern for black K . paniformis counts, as observed with the white mixed and K . pani- formis infections. This suggests that the black pseudocyst counts cannot be used as a reliable predictor of sensory texture. Magnitudes of black K. paniformis counts and black mixed counts were <320, while counts were <16 for black K. thyrsitis. Like those infected with white K. thyr- sitis, fish infected with black K. thyrsitis para- sites had normal, firm cooked textures (Table 3A). The absolute value of the Spearman correla- tion coefficient (Table 2) between texture quality and black pseudocyst intensity was lower (-0.38) than that of the coefficient for texture quality and white pseudocyst intensity ( -0.74), both at signif- icance level of 0.001. From the correlation coefficients for the white and black pseudocyst counts of both species, the square of the coefficients (coefficient of determi- nation) was calculated to compare their relative importance in terms of differences in their magni- tude with respect to one another. Fifty-five per- cent of the observed variability in all sensory tex- ture ratings can be accounted for (predicted) by the observed variability in the white pseudocyst counts, while only 14% can be accounted for by the observed variability in the black pseudocysts counts. However, infections do not occur in Pacific whiting as only white or black pseudocysts; they also occur as a mixture of the two. Thus, when the quantitative effects of the white or black pseudo- cyst counts on sensory texture were evaluated by multiple regression analysis, we found that only 1.5% of the variability in sensory texture rating was accounted for by black pseudocyst counts, and 45% accounted for by white pseudocyst counts. These figures represent partial coeffi- cients of determination which indicate the rela- tionship between two variables while controlling the effects of one or more other variables and are consistent with the findings of Patashnik et al. (1982) and Tsuyuki et al. (1982). Intensity of White and Black Pseudocyst Infection in Relation to Sensory Texture Ratings The magnitude of white and black pseudocyst counts in relation to their corresponding sensory texture scores are shown in Tables 4 and 5. A total of 214 fish, 38% of the fish examined, did not have white pseudocysts. Table 4 shows that texture scores for this group of fish range from 4 to 9 with 97% of the scores in the range of 750 KUDO ET AL.: TEXTURE QUALITY OF PACIFIC WHITING 6 to 8. Fish infected with white pseudocysts to the degree of 26 to 50 counts (average for the six areas in each fish tested) resulted in 1 out of 50 fish tested with a sensory texture score of 2. Thus, 2% of the fish in this category must be regarded as too soft and are organoleptically unacceptable (Table 1). In the sample of 562 fish, 25 specimens were found with white pseudocyst counts ranging be- tween 51 and 75, none of which were judged to have a texture score of 2 or lower. However, note that 11 of 25 (449c) fish samples were soft tex- tured. When the white pseudocyst counts were 76 or higher, nearly all of the fish were soft or abnor- mally soft textured. The frequency distribution of sensory texture scores vs. black pseudocyst counts (Table 5) does not show the same relationship that intensity of white pseudocyst infection had on sensory texture (Table 4). This becomes evident after examining sensory texture scores for fish with black pseudo- cyst counts of 1 to 25 (Table 5) when the entire range of texture scores is represented. Note, however, that the observed distributions of white and black pseudocysts differ. For exam- ple, of the 98 fish with pseudocyst counts exceed- ing 50, 81 had white pseudocysts but only 17 had black pseudocysts. Also, for Pacific whiting with counts >100, 37 of 45 (82%) with white pseudo- Table 4. — Frequency distribution of fish in each sensory texture score category and intensity of white pseudocyst infection in Pacific whiting over 27 cm. n = 562. White pseudocyst count (ave. fillet) Number of fish in each sensory texture score category Row total Percent of n 9 8 7 6 5 4 3 2 1 0 5 131 69 7 1 1 214 38.1 1-10 2 49 59 24 14 2 1 151 26.9 11-25 1 5 15 23 11 9 2 66 11.7 26-50 3 5 18 14 5 4 1 50 8.9 51-75 2 3 9 10 1 0 25 4.4 76-100 0 1 2 3 3 0 2 11 2.0 101-125 1 0 1 1 1 4 5 1 14 2.5 126-150 1 1 1 0 3 0.5 151-200 2 3 9 2 16 2.8 201-300 1 1 4 6 1.1 301-400 1 0 0 0 1 0.2 401-800 1 4 5 0.9 Total of infected fish 3 58 81 70 51 34 20 18 13 348 Percent of n 0.5 10.3 14.4 12.5 9.1 6.0 3.6 3.2 2.3 61.9 100.0 Table 5. — Frequency distribution of fish in each sensory texture score category and intensity of black pseudocyst infection in Pacific whiting over 27 cm. n = 562. Black pseudocyst Number of fish in each sensory texture score category Row total Percent of n (ave. fillet) 9 8 7 6 5 4 3 2 1 0 2 97 72 15 9 4 3 2 0 204 36.3 1-10 11-25 26-50 51-75 76-100 101-125 126-150 151-200 201-300 301-400 401-800 74 62 42 31 14 4 10 3 1 2 9 7 2 0 8 4 0 0 0 0 0 0 21 5 4 0 0 1 0 0 0 6 5 4 0 0 0 0 0 1 4 2 3 0 1 1 0 1 0 1 253 58 30 5 4 2 0 2 3 1 0 45.0 10.3 5.3 0.9 0.7 0.4 0 0.4 0.5 0.2 0 Total of infected fish 92 78 62 43 31 17 16 13 358 Percent of n 1.1 16.4 13.9 11.0 7.7 5.5 3.0 2.8 2.3 63.7 100.0 751 FISHERY BULLETIN: VOL. 85, NO. 4 cyst counts and 5 of 8 (63%) with black pseudocyst counts fell in the 1 to 3 texture category. Al- though the data are limited, this may suggest that for lower counts black pseudocysts are not related to poor texture, whereas for higher counts, they are (r = -0.38 at P = 0.001). In this study, 459 out of 562 fish were infected with the myxosporean parasite Kudoa . However, knowing that only white pseudocysts contain the parasites that produce the proteolytic enzymes that adversely affect texture (Tsuyuki et al. 1982), the assessment of the effect of pseudocyst infections on texture was necessarily confined to white pseudocyst counts only. Therefore, in order to determine the most likely white pseudocyst count which, when exceeded, would produce an abnormal texture in the cooked fish, the following analysis was made. On the scale of firmness in Table 1, the minimum acceptable texture value was defined as 3. Only 20 fish were rated 3. Their mean intensity of infection was 94.9 pseudocyst counts and median intensity was 88. Conse- quently, fish with median intensity >88 pseudo- cyst counts were hypothesized to be abnormally textured. On a qualitative scale this level of infec- tion was considered heavy. To test this hypothe- sis, all fish having infection intensities >88 white pseudocyst counts were computer selected. There were 50 such fish. These included 90.3% of all the fish with abnormal textures (sensory rating of 1 to 2), but also included 3.9% of all the normal or soft-textured fish (sensory rating of 3 to 9). Next, fish with average sensory textures (5) were se- lected to determine the white pseudocyst counts below which sensory textures would most likely be normal. The median pseudocyst count for these fish was 23. Counts below 23 were considered in- dications of a light infection. The level of infection between these two counts (23 to 88) was consid- ered to be moderate. Because black pseudocyst counts correlated poorly with sensory texture in this study, the de- gree of black pseudocyst infection had to be ar- rived at from cullability figures using the culling techniques described previously. The cullability categories were determined visually and, like the white pseudocysts, described as light, moderate, and heavy. Confidence limits for black pseudocyst counts for each cullability category were statisti- cally determined and the midpoint between the high end of one confidence limit and the low end of the adjoining confidence limits was taken as the dividing point. The following range of black pseudocyst counts was arrived at for each cate- gory: none (0); light (1 to 28); moderate (29 to 79); heavy (80 or more). Effect of Geographical Areas Data in Figure 6A shows the percentage of the catch from the various survey areas sampled in this study and their related sensory textures. Based on results of our sensory evaluations, 13% of the Pacific whiting harvested from the Monterey-Eureka, CA, sampling areas had ab- normal textures, whereas only 1 to 3% of the Pacific whiting caught between the Columbia River and Vancouver Island had abnormal tex- tures. The correlation coefficient between survey area and sensory texture rating was 0.20 at P =0.001. Similarly, the incidence of heavy white pseudo- cyst infection (>88 pseudocyst counts) was about threefold greater in the Pacific whiting from the Monterey-Eureka, CA, corridor (Fig. 6B) than in whiting caught between the Columbia River and Vancouver Island, i.e., 16 to 11% vs. 5 to 4%. The percentage of no white pseudocyst found in the fish samples from all survey areas ranged from 25 to 34%. The trend in black pseudocyst counts (Fig. 6C) was similar to the trend in white pseudocyst counts in that more heavy (>80) black pseudocyst counts were observed in Pacific whiting caught between Monterey-Eureka, CA, and the Colum- bia River than in Pacific whiting from the Van- couver Island area. Two to three percent of the Pacific whiting sampled from the area between Monterey, CA, and the Columbia River were heavily infected with black pseudocysts, whereas no heavy black pseudocyst infections were found in the Pacific whiting caught from Vancouver Is- land area. Infections of black pseudocysts in Pacific whiting from the Vancouver Island area were primarily light (1 to 29), the category into which 84% of the infected fish fell, and moderate infections (29 to 79), the category in which 4% of the infected fish fell. Table 6A shows the prevalence of the Kudoa species found in Pacific whiting caught between Monterey, CA, and Vancouver Island, Canada. Kudoa paniformis was the predominant species (average 36 to 37%) found in the Pacific whiting taken in the Monterey-Eureka area, whereas K. thyrsitis was the predominant species (29 to 40%) in the fish caught north of the Columbia River. Mixed infections averaged 34% for the combined survey areas. 752 KUDO ET AL.: TEXTURE QUALITY OF PACIFIC WHITING A ''' p E R I 7SH N T A G E se — 0 F C A T C H 25- B 188—1 P E R C E N T A G E 0 F C A T C H 75- 58 — 25 — 188—1 P E R C E N T A G E 0 F C A T C H 75 — 58- 25 — Z gg^z: SENSORY TEXTURES NORMAL SOFT ABNORMAL MONTEREY EUREKA COLUMBIA VANCOUVER WHITE PSEUDOCYSTS NONE LIGHT MODERATE HEAVY MONTEREY EUREKA COLUMBIA VANCOUVER m 7ZZ PCXXl ixx?n I7Z71 BLACK PSEUDOCYSTS NONE LIGHT MODERATE HEAVY MONTEREY EUREKA COLUMBIA VANCOUVER Figure 6. — Occurrence (%) of (A) sensory texture categories, (B) degree of white pseudocyst infection, and (C) degree of black pseudocyst infection of Pacific whiting, by INFPC areas. 753 FISHERY BULLETIN: VOL. 85, NO. 4 Table 6.— (A) Prevalence of Kudoa species, (B) fork length compo- sition (%), and (C) sex ratio of Pacific whiting samples in the INPFC catch survey area. INPFC survey area Mon- Colum- Van- All Description terey Eureka bia couver areas Sample size (fishes) 136 117 234 75 562 (A) Kudoa species (%) K. paniformis 36 37 18 15 26 K. thyrsitis 7 15 29 40 22 Mixed infection 35 27 35 39 34 None 22 21 19 7 18 (B) Fork length composition (%) Small (27-39 cm) 30 33 26 7 26 Medium (40-49 cm) 31 29 31 32 31 Medium large (50-59 cm) 27 28 30 40 30 Large (60-80 cm) 13 10 13 21 14 (C) Ratio of female/ male 1.2 1.3 1.6 2.2 1.5 Effect of Biological Factors on Sensory Texture Biological data showing sex composition, the ratio of females to males, and representative fork lengths (FL) and their relationship to correspond- ing sensory textures and survey areas are given in Tables 3B, 3C, 6B, and 6C, respectively. The ratio of females to males nearly doubled as fish- ing activities in this study moved from south to north along the Pacific coast. Concurrently, the fork length of the fish increased as well. About 60% of the largest fish (60 to 80 cm FL) were caught between the Columbia River and Vancou- ver Island. More abnormal textures were oberved in the larger fish than in the smaller fish (Table 3B). Based on the total number of specimens ex- amined, however, the correlation coefficient for the relationship between fork length and sensory texture was low (r = -0.21 at P = 0.001). Sensory texture was not found to be related to the sex of the fish (r = 0.04 at P = 0.181, Table 3C). The percentage of abnormal textures in the female and male fish was about the same, i.e., 6 and 8%, respectively, confirming the reports of Kabata and Whitaker (1981). Similarly, males and females were evenly distributed (approxi- mately 12%) in the soft-texture category. The Relationship of Pseudocyst Counts to Location of hifection in a Fillet Area to Texture Results of analyses to determine the relation- ship of intensity of infection in a fillet area (Fig. 3) to texture quality are shown in Table 7. Pseu- docyst infections were found throughout the fillet areas. The highest percentage (11 to 12%) of heavy infections of white pseudocysts were located in the nape area, whereas the lowest inci- dence (8%) was located in the tail (r = 0.89 at Table 7. — Percentage of degree of Kudoa pseudocyst infection and corresponding sensory textures found in preselected examination areas of Pacific whiting fillets. Fillet area examined r •Jape Middle Tail Dorsal Ventral Dorsal Ventral Dorsal Ventral Degree of white pseudocyst infection (%)/examination area None 38 41 41 43 46 46 Light (1-22)1 32 34 33 37 32 34 Moderate (23-87)i 18 15 17 13 15 13 Heavy (88 and over)' 12 11 9 8 8 8 Degree of black pseudocyst infection (%)/examination area None 41 34 43 40 44 41 Light (1-28)1 51 55 51 52 50 54 Moderate (29-79)' 6 8 4 6 5 3 Heavy (80 and over)' 2 3 2 2 1 2 Sensory texture(s) (%)/ examination area Normal 77 82 81 85 85 87 Soft 12 10 11 10 10 10 Abnormal 11 8 8 5 5 4 Sample size 502 548 562 556 560 551 'Pseudocyst counts. 754 KUDO ET AL.: TEXTURE QUALITY OF PACIFIC WHITING P = 0.008). Incidences of heavy infections of black pseudocysts were 2 to 3% at the nape and about 1 to27f atthetaiUr = 0.71 at P = 0.06). In general, there tend to be more black and/or white pseudo- cysts in the nape than in the tail area, although the correlation of white pseudocysts in the dorsal to ventral direction was rather low (r = -0.21 at P = 0.035). Sensory texture profiles shown in Table 7 indi- cated that more abnormal textures were found in the nape area (11 to 8%) than in the tail area (5 to 4%) ofthe fish examined (r = 0.48 at P = 0.16). Effect of Culling At the present time, there is no accepted method or methods for efficiently detecting and culling Pacific whiting infected with white or black pseudocysts. Ultraviolet light and back lighting with white light have been tried on occa- sion, but these techniques have not been devel- oped enough to be effective for use in Pacific whit- ing production. This leaves visual detection as the only on-site method for detecting and removing suspect Pacific whiting from the production line. However, as there are no reliable data available concerning the effectiveness of visual culling, an attempt was made in this study to estimate its potential usefulness. Criteria for culling (Patashnik et al. 1982) in this study was described in a previous section. Of the 562 fish examined, 34 were visually culled on the basis of a moderate to heavy degree of white and black pseudocyst infection. Of these, 10 fillets were moderately to heavily infected with white pseudocysts, 7 with both black and white pseudo- cysts, and 17 with only black pseudocysts. Based on the criteria developed by Patashnik et al. (1982), these results suggest that culling fillets that are moderately or heavily infected with pseudocysts appears possible. However, since culling has not been successfully demonstrated in a commercial setting, the technique may prove to be too difficult and time consuming to be practi- cal. CONCLUSIONS Overall, 18^^ ofthe Pacific whiting samples col- lected for this study were uninfected with Kudoa . Furthermore 65^7? had counts of <10 white pseu- docysts, and only 10% had counts over 100 white pseudocysts. By comparison, 81% ofthe fish sam- ples had <10 black pseudocysts counted and only 1% were infected with over 100 black pseudocysts counted. When both the white and black pseudocyst counts were considered collectively, the varia- tion in white pseudocysts explained 55% of the variation in sensory texture, whereas black pseudocysts accounted for 14%. However, when the effect of the white pseudocysts was mathe- matically removed from the fish samples having both, the black pseudocysts were found to explain only 1.5% of the variation in sensory texture. Infections of white K. panifor/nis and white mixed infections correlated (r = 0.80 and 0.79, re- spectively) with the variations in sensory texture better than the black K. paniformis or black mixed infections (0.66 and 0.49, respectively). Neither the white nor the black /if . thyrsitis pseu- docyst counts correlated well with sensory tex- tures. Although common in the Pacific whiting samples examined in this study, K. thyrsitis con- sistently were found in low numbers. On the other hand, K. paniformis were identified in 26% ofthe sample, and mixed infections were observed in 34% of the fish examined. The heaviest infections of white and black Kudoa sp. pseudocysts were found in the Pacific whiting caught off the coast of California. The highest percentage of abnormal sensory textures were also observed in the fish harvested off the California coast. Generally, in this study, we found that the larger the fish the greater the incidence of abnor- mal textures. Sex of the fish had no apparent ef- fect on the quality of sensory texture. Anatomically, the nape and dorsal areas of the Pacific whiting samples examined tended to have higher counts of white pseudocysts, and therefore more abnormal textures, than the other areas of the fish examined. The occurrence of heavy white pseudocyst infections in the nape, middle, and tail sections ofthe fish samples averaged 11.5%, 8.5%, and 8%, respectively. Heavy black pseudocyst in- fections were 2.5%, 2%, and 1.5% for nape, mid- dle, and tail sections. Overall, abnormal textures were found 9.5% ofthe time in the nape, 6.5% in the middle, and 4.5% in the tail. Differences in the number of white pseudocyst counts found be- tween the dorsal and ventral sides ofthe fish were small. The occurrence of abnormal texture was 30% greater for the dorsal side ofthe fish than the ventral side. Results of visual culling in this study suggest that the method may have some potential, but 755 FISHERY BULLETIN: VOL. 85. NO. 4 that the technique has yet to be successfully demonstrated under commercial conditions. ACKNOWLEDGMENTS The authors express their appreciation to Thomas Dark, Mark Wilkins, and Kenneth Wein- berg of the Resource Assessment and Conserva- tion Engineering Division, Northwest and Alaska Fisheries Center, NMFS, for collecting fish and their biological data during their bottom trawl survey; to Z. Kabata and D. J. Whitaker of the Department of Fisheries and Oceans, Fish- eries Research Branch, Pacific Biological Station, Nanaimo, B.C., Canada, for undertaking the enormous work of enumerating the white and black pseudocysts and identifying each Kudoa paniformis and K. thyrsitis for this study; to Russ Kapperman, Mathematical Statistician, Fish- eries Data and Management System Division, Northwest and Alaska Fisheries Center, NMFS, for setting up and running the valuable computer program for the robust locally weighed regression curves; and to Carla Stehr, Fishery Research Bi- ologist, Environmental Conservation Division, Northwest and Alaska Fisheries Center, NMFS, for providing the photomicrographs of both the /T. paniformis and K . thyrsitis . LITERATURE CITED BiLiNSKi, E., Y. C Lau. and R. E E. Jonas 1977. Objective measurement and control of the firmness of canned herring. Tech. Rep. Mar. Fish. Serv. 727, 22 p. Ind., Technol. Insp. Dir., Dep. Fish. Environ., Van- couver, BC, Can. Bourne, M. 1982. Food texture and viscosity: Concept and measure- ment (food science and technology). Acad. Press, Inc., N.Y., 312p. Cleveland, W S. 1979. Robust locally weighed regression and smoothing scatterplots. J. Am. Stat. Assoc, Theory Methods Sect. 74(368):829-836. Dark, T A 1985. Pacific whiting, Merluccius productus: The re- source, the industry, and a management history — Intro- duction. Mar. Fish. Rev. 47(2):1. Hull, C. H, AND N H NiE. 1981. SPSS update 7-9 new procedures and facilities for releases 7-9. McGraw-Hill Book Co., N.Y., 402 p. Kabata. Z , and D. J. Whitaker. 1981. Two species of Kudoa (Myxosporea: Multivalvulida) parasite in the flesh of Merluccius productus (Ayres, 1855) (Pisces: Teleostei) in the Canadian Pacific. Can. J. Zool. 59:2085-2091. 1986. Distribution of two species of Kudoa (Myxo- zoa:Multivalvulida) in the offshore population of the Pacific hake, Merluccius productus (Ayres, 1988). Can. J. Zool. 64:2103-2110. Konagaya, S. 1982. A review of the abnormal conditions of fish meat: Jellied meat and yake-niku, spontaneously done meat. (In Jpn.) Collected Reprints 1982-1983 from the Tokai Regional Fisheries Research Laboratory B713, p. 379-388. Tokyo, Japan. Nagahisa, E.. S Nishimura, and T. Fujita 1983. Kamaboko-forming ability of the jellied meat of Pacific hake. (In Jpn.) Bull. Jpn. Soc. Sci. Fish. 49(6):901-906. Translated by George Kudo, Utilization Res. Div., Northwest Alaska Fish. Cent., Natl. Mar. Fish. Serv., NOAA, 2725 Montlake Boulevard East, Seattle, WA 98112. Nelson, R. W , H J Barnett, and G Kudo. 1985. Preservation and processing characteristics of Pacific whiting, Merluccius productus. Mar. Fish. Rev. 47(2):60-74. NiE, N. H., C H. Hull, J. G Jenkins, K, Steinbrenner, and D. H. Bent 1970. Statistical package for the social sciences. 2d ed. McGraw-Hill, Inc., N.Y., 657 p. Oppenheimer, C H. 1962. Fish as food. In George Borgstrom (editor). On marine fish diseases. Vol. 2, p. 541-566. Acad. Press, N.Y. Patashnik, M. H , S. Groninger, Jr . H. Barnett, G Kudo, and B KOURY. 1982. Pacific whiting, Merluccius productus: 1. Abnor- mal muscle texture caused by myxosporidian-induced proteolysis. Mar. Fish. Rev. 44(5): 1-12. Robins, C R , R M Baily, C E Bond, J R Brooker, E. A. Lach- ner. R N Lea. and W. B Scott. 1980. A list of common and scientific names of fishes from the United States and Canada. Am. Fish. Soc. Spec. Publ. No. 12, p. 76. TSUYUKI. H., S. H. WiLLISCROFT, Z.KaBATA. AND D. J. WHITAKER. 1982. The relationship between acid and neutral protease activities and the incidence of soft cooked texture in the muscle tissue of Pacific hake (Merluccius productus ) in- fected with Kudoa paniformis and/or K. thyrsitis, and held for varying times under different pre-freeze chilled storage conditions. Can. Tech. Rep. Fish. Aquat. Sci. No. 1130, 39 p. Weinberg, K. L , M E. Wilkins, and T A Dark. 1984. The 1983 Pacific West Coast bottom trawl survey of groundfish resources: Estimates of distribution, abun- dance, age, and length composition. U.S. Dep. Commer. NOAA Tech. Memo. NMFS-F/NWC-70, 376 p. 756 DISTRIBUTION AND ABUNDANCE OF BILLFISH LARVAE (PISCES: ISTIOPHORIDAE) IN THE GREAT BARRIER REEF LAGOON AND CORAL SEA NEAR LIZARD ISLAND, AUSTRALIA Jeffrey M, Leis,' Barry Goldman,^ and Shoji Ueyanagi^ ABSTRACT Istiophorid larvae were relatively common in plankton samples from the Lizard Island region in November to early March 1980-85. Black marlin, blue marlin, and sailfish larvae were captured. Larvae of all three taxa were most concentrated and abundant in the Coral Sea immediately seaward (= windward) of the outer ribbon reefs. Concentration and abundance within the Great Barrier Reef Lagoon were not usually different from those more than 0.25 nautical miles offshore in the Coral Sea. Size-frequency data combined with the distributional information suggest that spawning or at least hatching of eggs was concentrated in the area within 0.25 nautical mile seaward of the reef crest. Preflexion larvae of blue marlin and sailfish were essentially confined to the upper 6 m of the water column (and perhaps the upper half of that), but not the neuston. Preflexion larvae of all three species dominated the oblique bongo net tows (98%), while postflexion larvae dominated the neuston samples (76%). This suggests an upward ontogenetic movement. The horizontal distribution of istiophorid larvae is probably the result of spawning close to the reef front, an area of supposed downwelling, combined with the proclivity of the larvae to occupy surface waters. This should lead to retention of larvae in the forereef area. Some caveats about accepting this hypothesis as a complete explanation for the horizontal distribution of istiophorid larvae are discussed. Near-reef areas appear to be important in the early life history of istiophorids at least in the Coral Sea and for the three taxa studied. The billfishes of the family Istiophoridae are large, high trophic level, pelagic fishes of consid- erable sport and commercial importance through- out tropical and subtropical oceans (Nakamura 1985). Information on their early life history is limited and investigations have been hampered by the relative rarity of the larvae. Studies on the distribution of istiophorid larvae in the Indo- Pacific have dealt with distributions over very broad areas and have not examined distributions on a small scale, particularly those very close to reefs. (The considerable Japanese work was sum- marized by Nishikawa et al. 1985 and the Rus- sian work by Gorbunova 1976.) Size of larvae in relation to horizontal distribution has only rarely been considered. Aside from reports that istio- phorid larvae had been captured in neuston tows iDivision of Vertebrate Zoology, The Australian Museum, P.O. Box A285, Sydney South, NSW 2000, Australia. 2Lizard Island Research Station, PMB 37, Cairns, Queens- land 4870, Australia; present address: Research Associate, The Australian Museum, P.O. Box A285, Sydney South, NSW 2000, Australia. ^Faculty of Marine Science and Technology, Tokai Univer- sity, 3-20-1 Orido, Shimizu, Shizuoka 424, Japan. Manuscript accepted July 1987. FISHERY BULLETIN: VOL 85, NO. 4. 1987. (e.g., Bartlett and Haedrich 1968; Gorbunova 1976) the only published information on vertical distribution of istiophorid larvae was provided by Ueyanagi (1964), who concluded billfish larvae were largely confined to surface waters during the day and dispersed through the upper 50 m at night. During studies on the distributional ecology of the larvae of reef fishes in the vicinity of Lizard Island in the northern region of the Great Barrier Reef, Australia, two of us (Leis and Goldman) have sampled extensively in the Great Barrier Reef Lagoon and the near-reef waters of the Coral Sea. In our samples, we captured a relatively large number of istiophorid larvae. This has pro- vided information which sheds light on little known aspects of the early life history of istio- phorids and in view of the widespread interest in istiophorid biology, we have prepared this sum- mary on the horizontal and vertical distribution of istiophorid larvae over relatively small scales and how these relate to development of the lar- vae. Because istiophorid larvae are difficult to identify, we have collaborated to insure accuracy in identification of the larvae. 757 FISHERY BULLETIN: VOL. 85, NO. 4 MATERIALS AND METHODS All samples considered here were taken in an area between Lizard Island, approximately mid- way across the Great Barrier Reef Lagoon, to 10 nmi seaward of the outer ribbon reefs of the Great Barrier Reef in the Coral Sea (Fig. 1). The Great Barrier Reef Lagoon is modally 30 m deep in this region (range 25-40 m). The outer ribbon reefs are located along the shelf break, and the bottom falls off sharply with distance into the Coral Sea reaching depths >2,000 m within 6 nmi (Fig. 1). The samples immediately to windward of Lizard Island were taken from a 7 m boat with a net of 0.4 m^ mouth area (see Leis 1986 for further details of sampling). Other samples were taken from RV Sunbird, a 14 m catamaran. The neu- ston net, with mouth dimensions of 1 x 0.3 m and 0.5 mm mesh, was towed between the bows of the catamaran, and normally fished to a depth of 0.1 m. Bongo nets of 0.85 m diameter and fitted with a depressor were towed in a double-oblique pat- tern to study horizontal distribution or in opening-closing mode to study vertical distribu- tion. Nets were towed at approximately 1 m/sec- ond, were of 0.5 mm mesh, and were equipped with a flowmeter. During the vertical distribu- tion study, a depth sensor with a deck display was fitted to the net. At other times, a mechanical depth-distance recorder was used. The neuston net, when used, fished during the bongo net tow. The oblique bongo net tows were done to the greatest calculated depth that was Figure 1. — Map of the study area. One sampling block is the Great Barrier Reef Lagoon beween Lizard Island and the ribbon reefs, this is delineated by the broken lines. In the Coral Sea, there are five sampling blocks A-E as defined in the t«xt. Depths in the Lagoon range from 20 to 40 m. Depths to 400 m are encountered in Coral Sea block A, to 1,000 m in block B and > 1,500 m in C-E. Outer ribbon reefs are 1) Day, 2) Carter, 3) Yonge, 4) No Name, and 5) Number 10 Ribbon. None of the outer reefs are emergent. 758 LEIS ET AL.: DISTRIBUTION AND ABUNDANCE OF BILLFISH LARVE considered safe in the lagoon and close to the windward face of the reef. Because of the great variations in bottom topography in the latter area, the net actually hit bottom upon occa- sion. Further offshore, bongo net tows were done with a standard amount of wire out which en- sured a maximum sampling depth in excess of 100 m. Position fixing in the Coral Sea was by radar reflection of the waves breaking on the reef crest when close to the reef. This meant that actual distance off the reef varied somewhat ( — 100 m) depending on sea state and tide. Four cruises were made to investigate horizontal distribution: 1) 2-5 November 1984, 2) 17 and 20-22 November 1984, 3) 30 January-2 February 1985, and 4) 9-13 February 1985. On each cruise, six samples were taken between Lizard Island and the outer reef on one day (Fig. 1), and three days were spent in the Coral Sea running a transect each day and start- ing at opposite ends of the transect on alternate days. On each transect, two randomly located samples were taken in each of five offshore blocks defined by distance (nmi) from the outer reef crest (Fig. 1): A, 0-0.25 nmi; B, 0.25-1.0 nmi; C, 1.0-3.0 nmi; D, 3.0-6.0 nmi; E, 6.0-10.0 nmi. Therefore, six samples were taken in each block on each cruise. The three transects on a cruise were each centered off a different reef (i.e., either of Day, Carter, Yonge, No Name, or Number 10 Ribbon Reefs). Bad weather and high volumes of floating pumice precluded the routine use of the neuston net. There were some variations in this plan owing to weather or equipment problems, the most serious of which was missing 4 of 6 samples in block A on the second cruise. Larvae from other samples taken with similar methods in November 1983 were also included where appropriate. Funding limitations prevented processing of sam- ples from block D. The vertical distribution samples were taken in the lee of Carter Reef primarily in the Great Bar- rier Reef Lagoon, but partially in the pass to the north of Carter Reef (Fig. 1). Samples were taken in sets; each set consisted of a neuston tow and 3 bongo net (0-6 m, 6-13 m, and 13-20 m) tows. The 0-6 m stratum was sampled in the undisturbed water flowing between the hulls of the catama- ran. In February-March 1983, 22 such sets were taken, 8 each in morning and afternoon and 6 at night. Additional samples from within the Great Bar- rier Reef Lagoon as reported by Leis and Gold- man (1984, 1987) and Leis (1986) were used for seasonality information. Samples in the Lagoon were taken in all months but May, June, August, September, and December. Samples were taken in the Coral Sea in October, November, January, and February. Oblique bongo net tows typically filtered 1,000- 1,500 m^ and horizontally stratified tows filtered 400 m'^. Neuston tows typically travelled 1,200- 2,000 m. All nets were carefully washed after each tow and the sample preserved in 5-10% seawater-formalin. In the laboratory, samples were sorted using a dissection microscope (~10x) and all larvae re- moved. Samples from both sides of the bongo net were fully sorted except for the Great Barrier Reef Lagoon samples from February-March 1983 and November 1984 when only side was sorted because of high plankton volume. Larvae were placed in 70% ethanol prior to measurement. Identification of larvae followed Ueyangi (1963, 1974a, b). Larvae were measured using an eye- piece micrometer of a dissection microscope to the nearest 0.1 mm. Notochord length and standard length were measured for preflexion and postflex- ion larvae, respectively (Leis and Rennis 1983). Larvae from these samples are deposited in the Australian Museum, Sydney. Numbers of larvae per sample were converted to numbers per volume (concentration) and num- bers per area (abundance) using standard meth- ods (Leis 1986). In analysis of vertical distribu- tion data, only positive sets (i.e., those in which at least one larva was captured) were considered. Statistical methods followed Conover (1971) and Zar (1974). References to the Student-Newman- Keuls (SNK) test refer to the version based on ranks (Zar 1974). RESULTS Identification We captured larvae of black marlin, Makaira indica; blue marlin, Makaira mazara; striped marlin, Tetrapturus audax; and Indo-Pacific sail- fish, Istiophorus platypterus . The larvae here identified as black marlin correspond to the "non- pigmented" sailfish of Ueyanagi (1974a, b). This type of istiophorid larva has been captured only in the seas off northern Australia and the south- ern portion of the Indonesia-New Guinea archipelago (Ueyanagi 1974a, b). In our Coral Sea samples these larvae were found almost exclu- sively during November, the time when large 759 numbers of gravid female black marlin occur in the area (B. Goldman pers. obs.; J. Pepperell^). A major sport fishery is based on this apparent spawning migration and the catches are made primarily just off the windward reef faces in the northern Coral Sea. When sampling in November 1984, our research vessel was frequently operat- ing in the midst of the sport fishing fleet. This circumstantial evidence suggested the possibility that the "non-pigmented" sailfish larva was in fact the larva of the black marlin, and led us to recheck these "non-pigmented" sailfish larvae. Two specimens (5.6 and 9.7 mm) were cleared and stained for bone and cartilage (Potthoff 1984); both specimens had vertebral formulae of 11-1-13 = 24. This confirms that they are of the genus Makaira (Nakamura 1985). These larvae can be distinguished from those of the only other Indo-Pacific member of the genus, the blue mar- lin, by head profile and depth and minor pigment differences. Therefore, we concluded that the "non-pigmented" sailfish larva captured in the present study were black marlin. A more detailed treatment of the identity of "non-pigmented" sail- fish larvae will be given separately (Ueyanagi and Leis in prep.). The larvae identified here as sailfish are nor- mally pigmented sailfish larvae which had not previously been reported from the Coral Sea (Ueyanagi 1974a, b). Only a few striped marlin larvae were captured, and because nearly all were small and only tentatively identified, they are not considered further. Seasonal Occurrence Sailfish larvae were taken only in January, February, and March. Blue marlin larvae were taken in mid-November, January, February, March, and April, although only one larva was taken in April. Black marlin larvae were taken throughout November, and three were taken in January-February. A sequence of occurrence of larvae and pre- sumably of spawning in the area begins with the appearance of black marlin larvae in late spring- early summer, followed by blue marlin in summer-autumn, and finally sailfish in late summer-early autumn. •*J. Pepperell, Fisheries Research Institute, N.S.W. Depart- ment of Agriculture, Cronulla, N.S.W., Australia, pers. com- mun. 1986. FISHERY BULLETIN: VOL. 85, NO. 4 Horizontal Distribution Black marlin larvae were most concentrated in block A adjacent to the seaward side of the reef on all cruises (Table 1). Concentrations elsewhere were low, with median values usually of zero. However, data from only one cruise could be tested statistically. The distribution of abun- dance was similar to that of concentration, with the exception that abundance in the two near-reef blocks could not be shown to be significantly dif- ferent during the first cruise. During the first two (November) cruises, black marlin larvae were taken in 7 of 8 samples from the near-reef area (block A). Only three black marlin larvae were taken on cruises three and four (January- February), all in block A. Black marlin larvae were present in only 13 of 96 samples taken else- where, and of these areas, block B (0.25-1.0 nmi offshore) had the highest frequency of occurrence, 5 of 24 samples. Clearly, black marlin larvae were consistently found in greatest numbers closest to the seaward side of the reef. The offshore extent of this high density zone of black marlin larvae was very lim- ited, extending at most to 1 nmi seaward (block B) of the reef crest, but more likely to only 0.25 nmi. Blue marlin larvae were less abundant than black marlin in our samples but had a similar distributional pattern. Again, data from only one cruise (the third) could be tested statistically. Ex- cept for the second cruise, blue marlin larvae were both most concentrated and abundant in block A, the area closest to the seaward face of the reef (Table 1). Further, 8 of the 13 occurrences were in this block. During the second cruise, blue marlin larvae seemed most concentrated and abundant at block B (0.25-1.0 nmi off), but only six larvae were captured on this cruise and only two samples were taken in block A so the signifi- cance of these results is questionable. Blue mar- lin were, with the possible exception of the second cruise, consistently found in greatest numbers closest to the seaward side of the reef. This is similar to the pattern for black marlin. However, small numbers of blue marlin larvae were cap- tured in block E, the most offshore segment of the transect, and this offshore area had the second highest frequency of occurrence of blue marlin larvae (Table 1). Only 13 sailfish larvae were taken, and the data are too sparse to indicate much more than all but 1 of the 7 occurrences were in the two blocks nearest the reef front (A and B). Sailfish larvae 760 LEIS ET AL.: DISTRIBUTION AND ABUNDANCE OF BILLFISH LARVE Table 1. — Distribution of istiophorid larvae based on transects from the Great Barrier Reef Lagoon into tfie Coral Sea. Co, concentration (larvae'1,000 m3); Ab, abundance (larvae/100 m^); f, frequency (i.e., nunnber of positive hauls). Values for Co and Ab are medians, and parenthetically, ranges. P is for Kruskal-Wallis test. For tested data sets, values with the same superscript symbol (# or t) are not significantly different (P > o.05, SNK Test). NT, not tested statistically: T, because only 2 samples were taken in block A; F, because too few larvae were taken. Normally, 6 samples were taken in each block on each cruise. No larvae were taken on the cruises not listed. Great Barrier Coral Sea blocks A B C E p Reef Lagoon (0-0.25 nmi) (0.25-1.0 nmi) (1.0-3.0 nmi) (6.0-10.0 nmi) Black marlin 1st cruise Co 0 (0-4.4)# 3.8 (0-11.3) 0.8(0-1.9)* 0 (0-1.9)* 0 (0-0.5)* 0.04 Ab 0 (0-10.9)* 15.0(0-56.5)t 10.5 (0-33.0)*t 0 (0-47.1)* 0 (0-13.0)* 0.06 f 1 5 4 2 2 2d cruise Co 0 (0-1.4) 2.8(1.1-4.5) 0 (0-0.5) 0 (0-1.2) 0 NT, T Ab 0 (0-3.4) 12.8(5.1-20.4) 0 (0-6.9) 0(0-14.9) 0 f 2 2 (of 2) 1 1 0 3d cruise Co 0 0 (0-0.7) 0 0 0 NT, F Ab 0 0 (0-2.0) 0 0 0 f 0 2 0 0 0 4th cruise Co 0 0 (0-1.2) 0 0 0 NT, F Ab 0 0 (0-1 .8) 0 0 0 f 0 1 0 0 0 Blue marlin 2d cruise Go 0 0 0 (0-2.2) 0 (0-0.6) 0 NT, T, F Ab 0 0 0 (0-23.0) 0 (0-7.2) 0 f 0 0 (of 2) 2 1 0 3d cruise Co 0# 1.5(0-8.4) 0 (0-0.5)* 0# 0 (0-0.6)* 0.02 Ab 0# 12.6(0-25.2) 0 (0-5.2)* 0# 0 (0-7.0)* 0.02 f 0 5 1 0 2 4th cruise Co 0 0.33 (0-2.4) 0 0 0 (0-1.8) NT, F Ab 0 1.0(0-7.7) 0 0 0 (0-21.1) f 0 3 0 0 2 Sailfish 3d cruise Co 0 0 (0-4.2) 0 (0-0.6) 0 0 NT, F Ab 0 0 (0-12.6) 0 (0-6.1) 0 0 f 0 1 2 0 0 4th cruise Co 0 0 (0-2.6) 0 (0-0.8) 0 (0-0.7) 0 NT, F Ab 0 0(0-10.4) 0 (0-9.3) 0 (0-7.8) 0 f 0 2 1 1 0 may have a distribution similar to that of blue marlin and black marlin larvae (Table 1). Sizes of Larvae From Bongo Net Tows Black marlin larvae ranged from 2.5 to 6.8 mm with a strong mode at 2.8-2.9 mm (Table 2a). Statistical comparison of the size-frequency data between areas could only be undertaken for the first cruise. Data from block A were compared with data from all other areas pooled. The size- frequency distributions were possibly different (Kolmogorov-Smirnov test, P = 0.07): a greater proportion of the larvae were of the smaller size classes (<4 mm) in block A than in the other blocks. Inspection of the limited size-frequency data from the other cruises indicates a similar situation. More than one cohort of larvae was present because larvae on the second cruise were not larger than those on the first. Blue marlin larvae ranged from 2.5 to 8.3 mm with a weak mode at 3.1 mm (Table 2b). Too few blue marlin larvae were captured to allow rigor- ous analysis of the size-frequency data, but there did not appear to be any difference in the size 761 FISHERY BULLETIN: VOL. 85, NO. 4 Table 2. — Size frequency of a) black and b) blue marlin and c) sailflsh larvae. If a block or cruise is not listed, no larvae were taken there. X indicates a hiatus in the size sequence. A few larvae too badly damaged to be measured were omitted. Blanks indicate zero. a. Black marlin Size class (mm) Block Cruise 2.5 .... 3.0 .... 3.5 .... 4.0 .... 4.5 .... 5.0 X 5.6 5.7 5.8 5.9 6.0 X 6.8 E 1st 11 C 1st 111 1 2d 1 1 8 1st 1111 11 11 2d 1 1 A 1st 328643213431 1 1 1 1 2d 13 12 3 3d 1 4th 1 1 Lagoon 1st 11 1 2d 1 1 b. Blue marlin Size class (mm) Block Cruise 2.5 .... 3.0 .... 3.5 .... 4.0 .... 4.5 .... 5.0 . . . . 5.5 X 8.3 E 3d 2 4th 1 111 1 C 2d 1 B 2d 1 1 1 1 3d 1 A 3d 211 4212 11 11 11 4th 1 112 c. Sailfish Size class (mm) Block Cruise 2.5 .... 3.0 .... 3.5 .... 4.0 C 4th 1 B 3d 11 4th 1 A 3d 2 3 1 4th 111 composition of the larvae between areas. Larvae on the fourth cruise were apparently larger than those on the third, however it is doubtful that only one cohort was involved because of the small size difference between the two cruises which were about 10 days apart (Table 2b). Sailfish larvae ranged from 2.5 to 3.8 mm, with a mode at 2.6 mm (Table 2c). Only 12 larvae were captured, but there is a suggestion that smaller larvae were taken nearest the windward reef face, and that the size of larvae increased with distance into the Coral Sea. Vertical Distribution Our information on vertical distribution comes primarily from samples taken within the Great Barrier Reef Lagoon. It is limited, but it is consis- tent. In the sampling with opening-closing bongo net and neuston net, larvae of two species, sailfish and blue marlin, were captured. No istiophorid larvae were present in the neuston tows of the vertical distribution sets. Sailfish larvae were captured in 7 of 16 day-time vertical sets and none of the 6 night-time sets. All the sailfish lar- vae were captured in the 0-6 m stratum with the exception of two larvae, one from each of the two deeper strata, which came from two sets taken in one of the turbulent interreef channels during a falling tide (Table 3). Even with the inclusion of the data from the interreef channel, sailfish lar- vae were most concentrated in the 0-6 m stratum while concentrations in the other strata did not differ (Friedman test, SNK test, P < 0.05). Blue marlin larvae were captured in only three of the day-time vertical sets. All the blue marlin larvae were captured in the 0-6 m stratum, with the 762 LEIS ET AL.: DISTRIBUTION AND ABUNDANCE OF BILLFISH LARVE Table 3. — Day-time vertical distribution of sailfish and blue marlin larvae in the vicinity of Carter Reef in Febru- ary and March 1983. N refers to number of vertical sets (i.e., a tow in each stratum) that contained at least one larva of that species. Sailfish (A/ = 7) Depth stratum Concentration (larvae/400 m3) Median Number of positive Range hauls (of 7) Neuston Bongo net 0-6 m 6-13 m 13-20 m 0 1.7 0 0 0-0 0-7.8 0-1.6 0-1.0 0 6 1 1 Blue marlin {N = 3) Depth stratum Concentration (lan/ae/400 m3) Median Number of positive Range hauls (of 3) Neuston Bongo net 0-6 m 6-13 m 13-20 m 0 1.8 0 0 0-0 1.8-4.3 0-0 0-1.0 exception of a single larva from 13 to 20 m from one of the interreef channel sets. In the three positive sets, blue marlin larvae were always most concentrated in the 0-6 m stratum, but there were too few data for rigorous testing. Blue marlin and sailfish larvae occurred in one vertical set taken on the windward side of Lizard Island in January 1980 (see Leis 1986). One larva of each species was taken in each of the 0-1 m and the 3-4 m tows, while none were taken in the 6-7 m tow. Istiophorid larvae from our neuston samples were developmentally more advanced (older) than those from bongo net samples. In all our samples, the bongo net captured 160 istiophorid larvae (black marlin, blue marlin, sailfish), three of which were postflexion stage, while the neu- ston net captured 17 istiophorid larvae (black marlin and blue marlin), 13 of which were post- flexion stage (chi square, P < 0.001). During the day preflexion blue marlin and sail- fish larvae inhabit the upper 6 m, and possibly the upper half of that, but not the neuston. It appears that once the caudal fm is formed, istio- phorid larvae move upward even more and enter the neuston. DISCUSSION Distribution of istiophorid larvae over such a small scale has not been studied previously, nor have such high concentrations of larvae been re- ported. Our results were surprising. Highest con- centrations and abundances of istiophorid larvae in our study area were consistently found in the Coral Sea very close to the windward side of the ribbon reefs at the outer edge of the Great Barrier Reef The size-frequency data (see below) suggest that this near-reef environment was a spawning area or just down wind of one for the three types of billfishes considered here. Concentration and abundance of istiophorid larvae in the Great Barrier Reef Lagoon (here- after referred to as the Lagoon) were always lower than in block A when both areas were sam- pled, but lagoonal numbers were generally not different from those further offshore in the Coral Sea. We cannot exclude the possibility that some istiophorid spawning takes place within the Lagoon, but believe it is more likely that the lar- vae were advected into the Lagoon through the interreef channels, as are larvae of many other oceanic fishes (Leis 1986; Leis and Goldman 1987). Still, concentrations of istiophorid larvae were high at times in the Lagoon (e.g., February- March 1983), and the relative survival of the lar- vae in the Lagoon vs. the Coral Sea is an open question. The marginally significant difference between areas in size frequency of black marlin larvae suggests that hatching of the eggs takes place very near the windward face of the reefs. This also suggests that black marlin larvae found else- where were largely the result of dispersal away from the near-reef area, and these dispersed lar- vae had grown somewhat during their dispersal. Spawning could either be concentrated in the near-reef area or more widely spread, in which case the eggs would have become concentrated in the near-reef area through wind-induced surface drift and forereef down welling (see below). Alter- natively, larval growth rates could be higher or mortality lower in the areas further from the reef. Our data do not allow us to distinguish between these alternatives, but we believe the first is the most likely. The data on blue marlin larvae gave no indica- tion of differences in size frequency between areas. The lack of difference in size-frequency dis- tribution could indicate that spawning in blue marlin was more evenly spread than in black marlin. If so, the increase in numbers nearest the windward side of the reef would be attributable to concentration and retention of larvae there. We cannot differentiate between this possibility and 763 FISHERY BULLETIN: VOL. 85, NO 4 the alternative that spawning is most intense near the reef. Too few sailfish larvae were taken to make any firm statements on distribution of larvae of differ- ent sizes. However, they appeared to have a pat- tern of size distribution with location similar to that of black marlin larvae. The vertical distribution data show that, at least during the day, preflexion larvae of blue marlin and sailfish concentrate in the upper few meters (perhaps upper 3 m) of the water column, but not in the neuston. However, postflexion lar- vae of blue marlin and black marlin are neu- stonic. This ontogenetic vertical migration has not been noted previously. The somewhat differ- ent results from the limited interreef channel samples could have been caused by turbulence due to strong tidal currents in these narrow passes. Using nonclosing nets, Ueyanagi (1964) stud- ied vertical distribution of istiophorid larvae (all taxa combined) over the upper 50 m and con- cluded that during the day larvae were most often caught at the surface and frequency of capture decreased with depth. At night catches of larvae were approximately evenly distributed over the upper 50 m. More recent data (Ueyanagi unpubl. data) confirmed this pattern for blue marlin, striped marlin, spearfish, and sailfish larvae. It is possible that the observed horizontal dis- tribution of istiophorid larvae in the Lizard Is- land area results solely from a concentration of spawning or at least hatching of eggs close to the windward side of the reefs. However, it is likely that additional factors are involved. The south- east trade winds push surface water against the windward sides of these reefs and although some of the water flows across the reefs into the La- goon, downwelling (anstau conditions) should occur seaward of the reef. An organism which maintains a position near the top of the water column, as do the istiophorid larvae (or positively bouyant fish eggs), would accumulate in such a downwelling zone. A similar situation has been described off the windward reef at Lizard Island where larvae of a number of reef fishes with shallow-living larvae were apparently retained (Leis 1986). However, the istiophorid larvae ap- parently disperse away from the surface at night (Ueyanagi 1964) whereas the larvae retained off windward Lizard Island tended to maintain their day-time vertical distribution at night (Leis 1986). If they did leave the surface, the istio- phorids might be advected away from the reef front. A further caveat against accepting the "anstau hypothesis" as a full and simple explana- tion for the distribution of istiophorid larvae in the area involves the trade winds. During the time the near-reef peak in istiophorid larvae was best developed (2-5 November), the winds varied from 0 to 10 kt and from northeast to southeast while on the other cruises, the wind was stronger and varied from 10 to 30 kt and from east to southeast. Finally, preliminary analysis of data from the samples in which the istiophorids were captured revealed that high abundances of a number of reef fish larvae also occur off the wind- ward reef face. Many of these were not near- surface dwelling larvae. Further study of larval fish distributions and their causes in this area is clearly required. Whatever the causes for the distributions of the istiophorid larvae very near the windward reef face, it is somewhat surprising that the larvae of epipelagic, oceanic fishes should be so abundant in such a narrow band along the reefs. Sailfish are known to spawn relatively close to land masses rather than in the open ocean (Ueyanagi 1974c) and black marlin are often found nearshore (Nakamura 1985); blue marlin are truly oceanic fishes (Nakamura 1985; Nishikawa et al. 1985). Yet larvae of all three were concentrated in a narrow band only 0.25 nmi (possibly to 1 nmi) off the reef crest. If pelagic fishes such as istiophorids concentrate their spawning very close to reefs or if the larvae are retained there, it will be essen- tial for such areas to be included in studies of the larval biology of these fishes. The assumption that open oceanic areas are the important nurs- ery areas for epipelagic fishes seems at best ques- tionable for istiophorids in the Coral Sea and sim- ilar factors may apply to other taxa in this and other areas. For example. Miller (1979) reported much higher concentrations of yellowfin tuna lar- vae, Thunnus albacares, in areas 200 m off the Oahu shoreline than had been reported else- where. Nearshore or near-reef areas may provide more favourable habitats for fish larvae, including those of many pelagic species, than do oceanic areas. The larvae of jack mackerel, Trachurus symmetricus , an epipelagic (albeit, neritic) fish, are spread widely over oceanic and coastal areas off California, yet larval mortality due to starva- tion in oceanic areas can be much higher than in coastal areas, presumably because of insufficient concentrations of food offshore (Theilacker 1986). This may apply to other pelagic fishes as well and 764 LEIS ET AL.: DISTRIBUTION AND ABUNDANCE OF BILLFISH LARVE is a further indication that very nearshore (and near-reeD areas must not be excluded from stud- ies of the larvae of epipelagic fishes. In summary, we found the highest concentra- tions and abundances of istiophorid larvae of three taxa very close to the windward face of the Great Barrier Reef in the Coral Sea in late spring and summer. Size-frequency analysis suggested that these high concentrations of larvae were due to spawning or at least hatching of eggs very close to the reef. The larvae might be retained in this forereef area of supposed down welling because, at least during the day, they concentrate in the upper few meters of the water column as preflex- ion larvae and in the neuston as postflexion lar- vae. These results have potentially important im- plications for the study of the larval biology of epipelagic fishes. ACKNOWLEDGMENTS We thank the many people who helped take and process the plankton samples reported upon here, particularly E. Moodie, S. Reader, and S. Thompson. T. Gob typed the manuscript and S. Bullock provided editorial assistance. T. Trnski and W. J. Richards commented on the manuscript; Richards comments were particu- larly valuable in that they led us to the proper identification of black marlin larvae. This work was supported primarily by Australian Marine Science and Technology Grants 80/2016 and 83/ 1357 to Goldman and Leis, respectively, but also by a Queen's Fellowship in Marine Science, Great Barrier Reef Marine Park Authority grants and Australian Museum grants to Leis. LITERATURE CITED Bartlett, M R , AND R L Haedrich 1968. Neuston nets and South Atlantic larval blue marlin {Makaira nigricans). Copeia 1968:469-474. CONOVER, W J 1971. Practical nonparametric statistics. J. Wiley, N.Y., 462 p. GORBUNOVA, N N 1976. The classification and distribution of the larvae of Indo-Pacific species of billfishes from the Family Istio- phoridae. J. Ichthyol. 16:437-452. Leis, J M 1986. Vertical and horizontal distribution of fish larvae near coral reefs at Lizard Island, Great Barrier Reef. Mar. Biol. 90:505-516. Leis, J. M., and B Goldman. 1984. A preliminary distributional study of fish larvae near a ribbon coral reef in the Great Barrier Reef. Coral Reefs 2: 197-203. 1987. Composition and distribution of larval fish assem- blages in the Great Barrier Reef Lagoon, near Lizard Island, Australia. Aust. J. Mar. Freshwater Res. 38: 211-223. Leis, J M , and D. S. Rennis. 1983. The larvae of Indo-Pacific coral reef fishes. New South Wales University Press, Sydney and University of Hawaii Press, Honolulu, 269 p. Miller, J M. 1979. Nearshore abundance of tuna (Pisces: Scombridae) larvae in the Hawaiian Islands. Bull. Mar. Sci. 29:19- 26. Nakamura, I. 1985. Billfishes of the world. FAO species catalogue, Vol. 5, 65 p. FAO Fish. Synop. 125, FAO, Rome. NiSHIKAWA, Y., M. HONMA, S. UEYANAGI, AND S. KiKAWA. 1985. Average distribution of larvae of oceanic species of scombroid fishes, 1956-1981. Far Seas Fish. Res. Lab. Shimizu, Jpn., 99 p. POTTHOFF, T. 1984. Clearing and staining techniques. In H. G. Moser, W. J. Richards, D. M. Cohen, M. P. Fahay, A. W. Kendall, Jr., and S. L. Richardson (editors). Ontogeny and system- atics of fishes, p. 35-37. Am. Soc. Ichthyol. Herpetol., Spec. Publ. 1. Theilacker, G. H 1986. Starvation-induced mortality of young sea-caught jack mackerel, Trachurus symmetricus , determined with histiological and morphological methods. Fish. Bull., U.S. 84:1-18. Ueyanagi, S 1963. Methods for identification and discrimination of the larvae of five Istiophorid species distributing in the Indo- Pacific. [In Jpn., Engl, synop.] Rep. Nankai Reg. Fish. Res. Lab. 17:137-150. 1964. Description and distribution of larvae of five istiophorid species in the Indo-Pacific. In Proceedings of the symposium on Scombroid fishes, pt. 1, p. 499-528. Mar. Biol. Assoc. India, Symp. Ser. 1. 1974a. On an additional diagnostic character for the identification of billfish larvae with some notes on the variations in pigmentation. In R. S. Shomura and F. Williams (editors), Proceedings of the International Billfish Symposium, Kailua-Kona, Hawaii, 9-12 August 1972, part 2, p. 73-78. NOAA Tech. Rep. NMFS SSRF- 675. 1974b. Present state of billfish larval taxonomy. InJ.H. S. Blaxter (editor), The early life history of fish, p. 649- 658. Springer- Verlag, N.Y. 1974c. Some considerations on the early life stage of the sailfish, Istiophorus platypterus, particularly regarding the transport of larvae by surface currents. [In Jpn., Engl, synop.] Bull. Far Seas Fish. Res. Lab., 10:189- 191. Zar.J H 1974. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, NJ, 620 p. 765 PREVALENCE AND EFFECTS OF INFECTION OF THE DORSAL AORTA IN YELLOWFIN TUNA, THUNNUS ALBACARES, BY THE LARVAL CESTODE, DASYRHYNCHUS TALISMANI Richard W. Brill,' Robert Bourke,' James A. Brock,^ and Murray D. Dailey^ ABSTRACT Approximately 60% of small (<3 kg) yellowfin tuna, Thunnus albacares, caught near the Hawaiian Islands carry the plerocercoid (larval) stage of the cestode (tapeworm), Dasyrhynchus talismani, in their anterior dorsal aortas. Because the worms and the resultant host inflammation appear to occlude the vessel almost totally, we assumed that the parasite could increase natural mortality rates. Tuna could be limited in their ability to capture prey and therefore should show evidence of long- or short-term food deprivation. We measured body weight, fork length, liver weight, heart weight and, in fish captured from one school, RNAVDNA ratios (a measure of short-term growth rate), and otolith weight (a measure of long-term growth rate) from parasitized and unparasitized fish. We found no significant differences between infected and uninfected fish nor any evidence of starvation in infected fish. How small yellowfin tuna remain apparently unaffected by the parasitic occlusion of their dorsal aorta remains to be demonstrated. We also examined changes in incidence of infection in small yellowfin tuna caught between Febru- ary 1985 and March 1986 as well as the prevalence in large (>45 kg) fish. Large yellowfin tuna were rarely parasitized (5.2'7f ) in the dorsal aorta, but showed a high rate (>80%) of infection within other major arteries. The prevalence in small fish varied dramatically with season, dropping suddenly from 66% in June-July 1985 to 11% in August-September 1985. Unparasitized fish caught during August- September 1985 showed significantly higher condition factors, relative heart weights, and relative liver weights than did unparasitized fish caught at other times of the year. We hypothesize that the sudden decrease in prevalence was due to influx of a separate group of small yellowfin tuna into the Hawaiian fishery. We believe that this parasite may therefore serve as a marker for tracing the movements of small yellowfin tuna into and out of specific fisheries or areas. During a series of experiments that involved catheterization of the anterior dorsal aorta of small ( 1-3 kg) yellowfin tuna, Thunnus albacares, we discovered that approximately 60% of the ex- perimental fish had this blood vessel infected with parasites. The parasites were white, round (2-4 mm in diameter), often more than 4 cm long, and usually folded repeatedly. As a result of the parasites and the tissue inflammation that devel- ops as part of host immune response, the lumen of the infected aortas appeared almost totally oc- cluded. Because all the blood to the internal or- gans and swimming muscles must flow past this occlusion, we assumed the parasite could be a major factor contributing to the natural mortality of small yellowfin tuna. The first demonstration of a dorsal aorta para- 1 Southwest Fisheries Center Honolulu Laboratory, National Marine Fisheries Service, NOAA, Honolulu, HI 96822-2396. 2Aquaculture Development Program, Department of Land and Natural Resources, State of Hawaii, Honolulu, HI 96819. 3Southern California Ocean Studies Consortium, California State University, Long Beach, Long Beach, CA 90840. Manuscript accepted July 1987 FISHERY BULLETIN: VOL 8.5. NO 4. 1987. site in yellowfin tuna was by Kishinouye (1923), who stated, "Often a species of nematod [sic] is found in the dorsal aorta of Neothunnus macrop- terus [now Thunnus albacares]; the parasite causes the tissues of the canal to become thick and tough, giving it at the same time a yellowish tint." Other investigators described intravascular parasites from the branchial vessels and arteries serving the stomach, liver, spleen, pyloric cae- cum, and gall bladder of this species (Baudin Lau- rencin 1971). The parasites have been described simply as the plerocercoids (larval cestodes) (Chen and Yang 1973), identified to the family Dasyrhynchidae (Ward 1962), or as the species Dasyrhynchus talismani (Baudin Laurencin 1971). Intravascular infection by plerocercoids has been reported from yellowfin tuna caught in the western Pacific (Chen and Yang 1973), eastern Atlantic (Baudin Laurencin 1971), Gulf of Mexico (Ward 1962), and now central Pacific (this re- port). Infection rates have been reported to be as high as 100% (Baudin Laurencin 1971). If all re- 767 FISHERY BULLETIN: VOL. 85, NO. 4 ports involve the single parasite species, D. talis- mani, this parasite is both common and ubiqui- tous. Because of the apparent vessel blockage by par- asites and tissue inflammation, we hj^jothesized that infected fish would be severely activity lim- ited and not function well as predators. If so, in- fected fish should show evidence of short- and/or long-term food deprivation, including lower rela- tive condition factors, smaller livers, and slower long- and short-term growth rates (Bulow 1970; Pollard 1971, 1973; Bulow et al. 1981). We also expected blockage of the dorsal aorta to cause in- creased blood pressures, increased cardiac work, and therefore cardiac hypertrophy (Poupa and Ostadal 1969). To test our hypotheses, we measured fork length, body weight, liver weight, and ventricle weight from parasitized and unparasitized fish. We also determined RNA/DNA ratios as a mea- sure of the parasite's effects on short-term growth rates (Bulow 1970; Bulow et al. 1981), and rela- tive otolith weights as a measure of the parasite's effects on long-term growth rates. Since otolith weights are linearly related to a fish's chronolog- ical age (Boehlert 1985), fish that are older at a given body size should have relatively heavier otoliths. We also recorded the prevalence of infec- tion in small (<3 kg) and large 045 kg) yellowfin tuna, and in skipjack tuna, Katsuwonus pelamis, kawakawa, Euthynnus affinis, and bigeye tuna, T. obesus. Additionally, monthly prevalence in small yellowfin tuna was noted. To measure directly the effect of the occlusion caused by the parasites and host inflammation, we used an in vitro perfusion test. We measured the pressure required to push various saline flow rates down the dorsal aorta in freshly dead in- fected and uninfected fish. MATERIALS AND METHODS Sampling Procedures The small tunas (0.2-3 kg body weight) used in this study were captured at sea near Oahu, HI, returned alive to be held in shoreside tanks at the Kewalo Research Facility (National Marine Fish- eries Service, Southwest Fisheries Center Hon- olulu Laboratory), or sacrificed at sea and held on ice. During necropsy, fork length and body weight were measured, ventricles and livers were re- moved, blotted dry, and weighed to the nearest 10 mg. Only those fish in captivity for 3 days or less were included in the data for condition factor, and relative heart and liver weights. Large yellowfin tuna (>45 kg body weight) were examined at the Honolulu Fish Auction for the presence of dorsal aorta parasites while the fish were being pre- pared for sale. All fish were examined within 24 hours of death. Thirty-five live yellowfin tuna, from one school, were caught 15 January 1986 and transported alive to the Kewalo Research Facility. Im- mediately upon arrival, the animals were sacri- ficed, weighed, and measured. Lateral white muscle samples were taken within 4 minutes of death and immediately frozen on dry ice. These samples were subsequently used for measure- ment of RNA/DNA ratios using the Schmidt- Thannhauser procedure as described in Murno and Fleck (1966). The ventricles and livers of the fish were also removed, blotted dry, and weighed to the nearest 10 mg. Sagittal otoliths were re- moved, cleaned, dried, and weighed to the nearest microgram. Parasites for species identification were ob- tained most often from the major artery within the spleen of large (>45 kg) yellowfin tuna. After being removed from a surrounding capsule, para- sites were placed in tap water until the holdfasts everted. For histological examination, sections of dorsal aorta were fixed in 10% buffered formalin and processed by routine laboratory procedures. Tis- sue sections were stained with hematoxylin and eosin. Direct Measurement of Pressure-Flow Relationships in the Anterior Dorsal Aorta To quantify blockage, we choose to measure the pressures required to push various flow rates of saline through the anterior dorsal aorta of in- fected and uninfected fish. The dorsal aorta of freshly killed fish was exposed from the conflu- ence of the efferent arteries of the first and second gill arches to the point where it enters the first hemal arch. All efferent and afferent vessels were tied off except for the confluence of the efferent arteries from either the left or right first and sec- ond gill arches. This portion of the vessel was cut and a short length of flared polyethylene tubing (PE160, 2.4 mm OD) inserted. The dorsal aorta was also transected at the point where it entered the first hemal arch to allow the saline perfusate to flow out. Parasites were never found poste- 768 BRILL ET AL.: INFECTION OF DORSAL AORTA IN YELLOWFIN TUNA rior to this point. Saline perfusion pressures, at various constant flow rates provided by an infu- sion pump, were recorded via a Uonix"* pressure transducer. Calculation of Relative Condition Factor and Relative Organ Weights Use of relative condition factor, and relative organ and otolith weights allow groups of fish containing individuals of a range of body sizes to be directly compared (Pollard 1972). Using data from unparasitized fish, regressions of body weight (g) on fork length (cm), liver weight (g) on body weight (g), and heart weight (g) on body weight, were fitted by a least squares technique to the exponential equation: Y = a- Xf" using a log-log transformation of the data. Rela- tive condition factor and relative organ weights for individual fish were calculated using the re- gression parameters (a, 6 ) with the equation: K = W/a- XK For relative condition factor, W = body weight and X = fork length. For relative organ weights, W = liver or heart weight, and X = body weight. The relationship of otolith weight (mg) to body weight was found best fit with the simple linear regression: Y = a + ib ■ X). Relative otolith weights were therefore calcu- lated using the equation: K = WKa + b- X) where W = otolith weight and X = body weight. A relative condition factor <1 indicates that an individual is lighter for its fork length than pre- dicted based on data from unparasitized fish. Sim- ilarly, a relative liver or heart weight <1 indi- cates a smaller liver or heart for a given body size than that found for unparasitized fish. A relative otolith weight >1 means that an individual expe- rienced a relatively slower long-term growth rate ^Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. (i.e., is relatively older for a given body size and therefore has a larger otolith). RESULTS Identification of the Parasite Tapeworms (class: Cestoda, order: Try- panorhyncha) can be identified to species based on scolex morphology and tentacular hooks (onchotaxy), mature segments are not required. The larval cestodes recovered from the yellow- fin tuna during this study showed proboscis chainettes flanked by a single row of inter- calary hooks, a characteristic that distinguishes Dasyrhynchus talismani from its congeners. This parasite was originally described from five mature worms removed from the spiral valve of Galeus glaucus (= Prionace glauca, the blue shark) off" Cape Verde, West Africa (Dollfus 1935). Dasyrhynchus talismani has also been re- ported in the Pacific from Carcharinus longi- manus (Heinz and Dailey 1974). All other reports describe plerocercoids from the vascular systems of teleost fishes (Bussieras and Aldrin 1965; Baudin Laurencin 1971; Chen and Yang 1973). Prevalence of hifection by Host Species, Fish Size, and Season A total of 53 skipjack tuna, 27 kawakawa, 10 bigeye tuna, and 470 yellowfin tuna were exam- ined for the presence of parasites. We found only yellowfin tuna to be infected. Infection in yellowfin tuna varied with size class. We found a significantly lower incidence of dorsal aorta infection in large fish. Of 220 indi- viduals weighing 0.21 to 2.7 kg, 48% were in- fected, while of 250 fish weighing more than 45 kg, only 5.2% carried the parasite in their dorsal aortas. Viscera of a small subsample (N = 8) of the larger fish were examined and indicated that in larger fish the parasite infects (in the order of prevalence) the major arteries of the spleen, in- testinal caeca, liver, mesenteries, and lateral blood vessels. Fish in intermediate size classes were not available for this study. Fish >3 kg do not survive the trip from the fishing grounds to the Kewalo Research Facility and are therefore not normally captured by commercial fishermen for return to the laboratory. Fish <45 kg are not common at the Honolulu Fish Auction where they could be examined during normal processing. The 769 FISHERY BULLETIN: VOL. 85, NO. 4 purchase of intermediate-sized yellowfin tuna specifically for this study (yellowfin tuna in- tended for market cannot be necropsied and then sold) was prohibitively expensive. Infection of yellowfin tuna appears to vary sea- sonally. Figure 1 shows changes in prevalence of dorsal aorta infection in small (<3 kg) yellowfin tuna captured between February 1985 through March 1986. Prevalence remained stable for ap- proximately 6 months during the winter through early summer. Then in late summer of 1985, prevalence dropped dramatically from 66 to 11%. Beginning in October 1985, prevalence increased steadily, reaching 39% in February-March 1986, the last months for which data are available. NFESTATION OF YELLOWFIN TUNA - BIMONTHLY INTERVALS 70 53 56 49 42 35 28 21 14 07 0 . ^ . .A in A v . n=10 \ . \ mean=i8% • \ / /^8 "3 . \ / \ / /" ' v ■ \ — 1 45 — 1 — 1 1 FEB-MAR JUN-JUL OCT-NOV FEB-MAR APR-MAY AUG-SEP DEC-JAN Figure 1. — Prevalence (percent infection) of the dorsal aorta by the plerocercoid stage of the cestode Dasyrhynchus talismani in small (0.3-3 kg) yellowfin tuna caught near the Hawaiian Is- lands from February 1985 to March 1986. Pathology In infected fish, the anterior dorsal aorta was partially to nearly completely occluded by a para- sitic embolus which contained one to several larval cestodes. Figure 2a shows this vessel in a moderately infected yellowfin tuna. Parasites and a small amount of host inflammation are ev- ident in the anterior end (to the right). A normal portion of vessel, with a smooth wall, is seen to the left. Figure 2b shows the anterior dorsal aorta from a very heavily infected fish. Histological examination revealed that para- sitic emboli were primarily composed of larval cestodes, mononuclear cells with eosinophilic granules (presumed to be eosinophils), epitheloid cells (histocytes), fibroblasts, and collagen fibers (Fig. 3). Larval cestodes within the dorsal aorta were associated with a chronic severe endarteritis and, to a lesser extent, mesoarteritis. In heavily infected fish, collapsed channels were commonly found within emboli. Undoubtedly these channels expanded in life with increases in intraluminal blood pressure to allow blood to flow through the vessel. Necrotic worms were also seen, suggesting that the host's defense system was at least par- tially capable of killing the larvae located in the dorsal aorta. Infection of other arteries was usually by one, or at most two, larger parasites which were never seen to be folded. These parasites had a bulbous anterior end which always pointed downstream. No evidence of host immune response was ob- served when the parasites were in vessels other than the dorsal aorta. Effect of Infection on Measures of Physiological Fitness Mean (±1 SD) relative condition factor and mean {±1 SD) relative organ weight data are given in Table 1 for the fish sampled between May 1985 and September 1986. Table 2 lists the same parameters plus mean relative otolith weight, and mean RNA/DNA ratio for fish sam- pled from the single school captured on 15 Janu- ary 1986. Means were compared with Student's two-tailed T test, with P = 0.05 taken as the min- imum level for statistical significance. Examination of data from all the fish caught May 1985-September 1986 reveals statistically significant differences in relative liver and heart weights for parasitized and unparasitized fish. On the average, parasitized fish, have 15% smaller livers, and 9% smaller hearts than unparasitized fish. Relative condition factors were not signifi- cantly different. When data from fish coming from the single school are examined, there are no statistically significant differences in these three parameters, relative otolith weight, or in RNA/ Table 1 . — Relative condition factor, relative liver weight, and rela- tive heart weight, of yellowfin tuna sampled between May 1985 and September 1986. Fish Mean (±SD) relative condition factor Mean (±SD) relative liver weight Mean (±SD) relative heart weight Uninfected 1.00 (±0.10) N = 109 1.02 (±0.22) W = 82 1.01 (±0.18) W = 88 Infected 0.976 (±0.057) 0.867 (±0.098)* 0.919 (±0.085)* A/ = 35 N = 22 A/ = 32 "Uninfected and infected groups different at P = 0.05 level. 770 BRILL ET AL.: INFECTION OF DORSAL AORTA IN YELLOWFIN TUNA Figure 2b. -The anterior dorsal aorta of a moderately infected yellowfin tuna. In the anterior end of the vessel (to the right) are parasites and host inflammation. -The anterior dorsal aorta of a heavily infected yellowfin tuna. The vessel's lumen is all but occluded by parasites and host inflammation. Ruler divisions are in mm. Table 2. — Relative condition factor, relative liver weight, relative heart weight, relative otolith weight, and RNA/DNA ratio of yellov\4in tuna sampled 15 January 1986. Fish Mean (±SD) relative condition factor Mean (±SD) relative liver weight Mean (±SD) relative heart weight Mean (±SD) relative otolith weight Mean (±SD) RNA/DNA ratios Uninfected Infected 1.00 (±0.04) N = 23 0.995 (±0.035) N = 12 1.00 I N t0.08) 23 1.00 (±0.07) N = ^2 1.00 (±0.06) A/ = 23 0.975 (±0.074) A/= 12 1.00 (±0.04) A/ = 12 0.980 (±0.058) N = 8 28.9 (±8.9) N = 2^ 33.1 (±11.1) N= 10 771 FISHERY BULLETIN: VOL. 85, NO. 4 Figure 3. — Photomicrograph of a cross section of the dorsal aorta containing parasites and host inflammation of a heavily infected yellowfin tuna. The host inflammation (I) and cross sections of the parasites (W) can be easily differentiated. Hematoxylin and eosin stain. DNA ratios. RNA/DNA ratios were compared di- rectly (rather than by calculating relative RNA/ DNA ratios) because they were found not to be correlated with body weight (correlation coeffi- cient - 0.04). Data from unparasitized fish collected in Au- gust and September 1985 were analyzed sepa- rately from all remaining unparasitized fish. This was done because we found dramatically lower rates of infection (11%) during those 2 months than during the preceding 2 months (67% in- fected) and assumed this to be due to an influx of a new group of small yellowfin tuna into the Hawaiian fishery. New regression parameters for body weight on fork length, heart weight on body weight, and liver weight on body weight were calculated using data from unparasitized fish ex- cluding those caught during August-September 1985. Mean relative condition factors and relative organ weights were then recalculated. Table 3 shows mean relative condition factors, mean relative liver weights, and mean relative heart weights for unparasitized fish captured dur- ing August and September 1985, unparasitized Table 3. — Relative condition factor, relative liver weight, and relative heart weight from uninfected yellowfin tuna sampled during August and September 1985, from all other uninfected yellowfin tuna, and all infected fish. Group Mean (±SD) relative condition factor t^ean (±SD) relative liver weight Mean (±SD) relative heart weight August-September 1985 uninfected fish All other uninfected fish Infected fish 1.11 (±0.08)* W = 36 1.00 (±0.09) N = 73 1.01 (i N 0.06) = 35 1.24 (±0.28)* N = 35 1.01 (±0.17) N = 47 0.937 (±0.102) N = 22 'Uninfected groups different at P = 0.01 level. 1.22 (: N = 1.00 (: N = :0.18)* 36 :0.09) 52 1.00 (±0.09) A/ = 32 772 BRILL ET AL.: INFECTION OF DORSAL AORTA IN YELLOWFIN TUNA fish captured at all other times of the year, and all parasitized fish, based on the new regression parameters. The two groups of unparasitized fish show statistically significant differences in all three. Fish caught during August and September were on the average \1% heavier at a given body size (i.e., relative condition factor = 1.11), had liv- ers an average of 23% heavier, and had hearts an average of 22% heavier than those unparasitized fish captured at other times of the year. When unparasitized fish, excluding those caught during August and September, are compared with para- sitized fish, there are now no statistically signifi- cant differences in mean relative condition fac- tors, mean relative liver weights, or mean relative heart weights. The data from unpara- sitized fish captured in August and September, when included in the complete data set, are there- fore responsible for the observed differences in relative heart and liver weights between infected and uninfected fish seen in Table 1. In Vitro Perfusion of the Dorsal Aorta Three parasitized and five unparasitized fish, ranging in weight from 0.915 to 2.666 kg, were used in this series of experiments. The intensity of infection was subjectively classified as slight, moderate, or heavy. Perfusion pressures were normalized to a 1 kg fish weight by dividing the observed pressures by the reciprocal of body 300 290 ■• en X E e 180 -■ I50-- 90-- 60 30-- HEAVY MODERATE -^ SLIGHT -t- ■+■ -+- -t- 20 30 iO 50 60 Perfusion Flow ml/mm 70 80 Figure 4. — In vitro pressure required to pump various rates of saline down the dorsal aorta of parasitized and unparasitized yellowfin tuna. The fish ranged in size from 0.915 to 2.666 kg. Data have been normalized to a body weight of 1 kg. weight, in kilograms. Moderately and heavily in- fected fish showed higher perfusion pressures at a given flow rate than did unparasitized animals (Fig. 4). While no data are available on the nor- mal cardiac outputs and blood pressures in swim- ming yellowfin tuna, restrained and lightly anes- thetized yellowfin tuna have cardiac outputs of approximately 40-60 mL/kg and dorsal aorta blood pressures of 50-70 mm Hg (D. R. Jones and R. W. Brill, unpubl. obs.). Figure 4 shows that at normal cardiac outputs, the apparent occlusion of the dorsal aorta caused by the parasites and host inflammation is indeed real and should cause moderately and heavily infected fish to have ex- cessively high blood pressures, high cardiac en- ergy demands and presumably reduced fitness. DISCUSSION Prevalence of Infection by Species, Size, and Season Our data indicate that in Hawaiian waters D. talismani is limited to yellowfin tuna. However, D. talismani has been reported to occur in At- lantic bigeye tuna (Bussieras and Aldrin 1965). Since we examined relatively few individuals of this species, we cannot rule out the occurrence of this parasite in bigeye tuna in the central Pacific. Skipjack tuna, bigeye tuna, kawakawa, and yellowfin tuna often occur simultaneously in the same areas and show a great overlap in prey spe- cies (King and Ikehara 1956; Waldron and King 1963). It is therefore unlikely that host specificity is attributable to only yellowfin tuna ingesting the intermediate host, which is not known but is most likely a small crustacean (Deardorff et al. 1984). Host specificity could arise if the procer- coid of/), talismani is not stimulated or is unable to penetrate the gut wall of tuna species other than yellowfin tuna, or that species of tuna other than yellowfin are capable of immune rejection (Orr et al. 1969). The reasons for the dramatic decrease in inci- dence of dorsal aorta infection in large (>45 kg) yellowfin tuna are unknown. Possible reasons in- clude procercoids ingested less frequently by larger animals, host destruction of the parasite, increased mortality of parasitized fish, or move- ment of the parasite out of the dorsal aorta into other major arterial vessels. Of these alterna- tives, increased mortality of infected fish seems unlikely. While we did not demonstrate directly the pre- 773 FISHERY BULLETIN: VOL. 85, NO. 4 ence of antibodies against the parasite, several histological cross sections show worms in stages of degeneration, and host destruction of larval and adult cestodes has been shown in other teleosts (Kennedy and Walker 1969; Smith 1973; MacKenzie 1975). Examination of a small num- ber of pyloric caeca, liver, spleen, lateral arteries, and stomach vasculature of large yellowfm tuna (>45 kg) showed >85% infection, suggesting 1) that the parasite may move out of the dorsal aorta into other large arteries as yellowfin tuna grow, 2) that the host response to the parasite may be less vigorous in vessels other than the dorsal aorta, and/or 3) that the parasite may pref- erentially select other vessels in larger fish. Baudin Laurencin (1971) found a decrease in branchial artery infection of yellowfin tuna with increasing body size, but no change in rates of infection of abdominal vessels. The change in the rate of infection seen in Au- gust and September 1985 in small yellowfin tuna (Fig. 1) is, we believe, due to a large influx of uninfected fish into the Hawaiian fishery. Al- though we have no direct corroborating evidence, such as increased catch per unit effort for this size yellowfin tuna at that time. Tester and Naka- mura (1957) have shown that there are repeated influxes of small yellowfin tuna into areas near the main Hawaiian Islands during late summer and early fall. Furthermore, the dramatic differ- ences seen in relative condition factors, relative liver weights, and relative heart weights between unparasitized fish caught during August- September 1985, and the remaining unpara- sitized fish, clearly imply that the former group had a different history. We have no evidence nor are we hypothesizing that these groups come from genetically isolated subpopulations. We do believe, however, that the two groups were separate most likely since hatch- ing. We do not know the maximum lifespan of the dorsal aorta parasite, but one yellowfin tuna killed after 172 days in captivity at the Kewalo Research Facility was parasitized. Since fish in captivity are fed only frozen food and their tanks are supplied with filtered seawater (Queenth and Brill^), it is unlikely this fish became infected after capture. Yellowfin tuna of the 1-3 kg size range are about 270-360 days old (Uchiyama and SQueenth, M.K. K., and R.W. Brill. 1983. Operations and procedures manual for visiting scientists at the Kewalo Re- search Facility. Southwest Fish. Cent. Admin. Rep. H-83-7, 16 p. Natl. Mar. Fish. Serv., NOAA, Honolulu, HI. Struhsaker 1981) and therefore could have car- ried the parasite most of their lives. The slow increase in prevalence from October 1985 through March 1986 remains to be ex- plained, but could be due to emigration of the new group of yellowfin tuna out of the Hawaiian fish- ery or slowly increasing infection of the new group. This latter explanation implies higher prevalence of the parasite around islands which could be due either to a greater number of final (shark) or intermediate hosts around islands. Pathology The severe enarteritis associated with D. talis- mani infection suggests that the parasite is recog- nized by the fish's defense system as foreign ma- terial in the dorsal aorta. Dead worms within the inflammatory tissue imply that the parasite is not well adapted for survival in this location. Pre- sumably, D. talismani would elicit, and be at- tacked by, a similar inflammatory response irre- spective of its location within the vasculature. This response was not observed in other vessels and additional work is needed to clarify the site specificity of the immune response. Our findings also suggest that cellular elements are responsi- ble for the destruction of the larval cestodes when located in the dorsal aorta. Effect of the Infection of the Dorsal Aorta on Natural Mortality We found no evidence to support our original hypothesis that infected yellowfin tuna are activ- ity limited and therefore less able to secure food. When the data from the unparasitized fish caught August-September 1985 are excluded, there are no differences in relative condition factors, rela- tive liver weights, or relative heart weights be- tween parasitized and unparasitized fish. When parasitized and unparasitized fish from the single school caught 15 January 1985 are compared, no significant differences in these parameters, mean short-term (i.e., RNA/DNA ratios), or mean long- term (i.e., relative otolith weights) growth are ev- ident. Because parasite emboli appear to cause almost complete occlusion of the anterior dorsal aorta (the only blood vessel supplying the viscera and swimming muscles), the lack of differences be- tween infected and uninfected fish was not ex- pected. Overstreet (1977), investigating the ef- fects of plerocercoid infection on sciaenid fishes in 774 BRILL ET AL.: INFECTION OF DORSAL AORTA IN YELLOWFIN TUNA the Gulf of Mexico, found no apparent detrimen- tal effect on the host, but these parasites were found encysted in the muscle, not the vascula- ture. Although the ability of parasitized yellowfin tuna to function as predators are apparently not affected, their ability to escape predation remains to be tested. It is possible that infected yellowfin tuna are subjected to differential predation, as has been shown to be for roach, Rutilus rutilus, infected in the coelomic cavity by the plerocercoid of Ligula sp. (Van Dobben 1952). Pressure Flow Relationships in the Dorsal Aorta The unphysiologically high pressures required to pump saline through the dorsal aortas of mod- erately and heavily infected fish remain to be ex- plained in light of apparent lack of effects of the parasite on other measures of the fish's condition, including the absence of cardiac hypertrophy. It is possible that the dorsal aorta, because of its thick muscular wall (J. Brock, unpubl. obs.), becomes significantly less compliant postmortem. Such changes could require that higher pressures be generated to create a given degree of expansion. Therefore higher pressures would be required postmortum to push a given flow rate of saline through the vessel. In summary, D. talismani appears to have no significant adverse effects on the physiological fitness and natural mortality of small yellowfin tuna in spite of apparent vascular blockage. How these fish are able to cope with dorsal aorta infec- tion requires further investigation. Use of Dasyrhynchus talismani as a Natural Tag for Tracing Movements of Small Yellowfin Tuna We feel that this parasite offers excellent po- tential as a natural marker for tracking the movements of separate groups of small yellowfin tuna between or into specific fisheries. (For a re- view of the use of parasites to delineate stocks for management purposes, see MacKenzie 1983.) Dasyrhynchus talismani does not fulfill all seven requirements for an ideal natural tag listed by Sindermann (1983), but it does appear to meet the requirements of 1) having significant differences of geographic prevalence, 2) being easily detected, 3) being able to be definitively identified, 4) hav- ing minimum effect on host survival, and 5) sur- viving in the host for long periods. Data on preva- lence ofD. talismani could be combined with data on prevalence of other parasites, as has been shown by Lester et al. (1985) for skipjack tuna, or combined with data on relative condition factor, relative heart weight, and relative liver weight to provide information on fish movements. ACKNOWLEDGMENTS The authors wish to express their appreciation to David Jones (University of British Columbia) for drawing our attention to this problem, and to Frank Goto and Brooks Takanaka of the United Fishing Agency (Honolulu, HI) for allowing us to inspect and take samples fi-om yellowfin tuna on the auction floor. We also thank Greg Webber for his data-collecting work, Theresa Villanueva for reviewing the literature, Dave Karl and Chris Winn (University of Hawaii) for helping with the RNA/DNA analyses, Michael Hadfield (Univer- sity of Hawaii) for allowing us to use his Cahn microbalance, and Carol Hopper (Waikiki Aquar- ium) and Terry Foreman (Inter- American Tropi- cal Tuna Commission) for reviewing drafts of this manuscript. LITERATURE CITED Baudin Laurencin, F. 1971. Crustaces et helminthes parasites de I'Albacore (Thunnus albacares) du Golfe de Guinee - Note prelimi- naire. Doc. Sci. Cent. Rech. Oceanogr. Abidijan 2:11-30. BOEHLERT, G W. 1985. Using objective criteria and multiple regression models for age determination in fishes. Fish. Bull., U.S. 83:103-117. BULOW, F J. 1970. RNA-DNA ratios as indicators of recent grovsrth rates in fish. J. Fish. Res. Board Can. 27:2343-2349. BuLOw, F J., M E Zeman, J. R Winningham. and W F Hudson 1981. Seasonal variations in RNA-DNA ratios as indica- tors of feeding, reproduction, energy storage, and condi- tion in a population of bluegill, Lepomis macrochirus Rafinesque. J. Fish Biol. 18:237-244. BUSSIERAS, J., AND J. F. ALDRIN. 1965. Une tetrahynchose vasculaire des thons du golfe de Guinee due aux larvae plerocerus de Dasyrhynchus talis- mani R. Ph. Dollfus 1935. Rev. Elev. Med. Vet. Pays Trop. 18:137-143. Chen, C.J and R -T Yang 1973. Parasites of yellowfin tuna in the waters of south- west Taiwan. Acta Oceanogr. Taiwan. 33:181-198. Deardorff, T L , R. B. Raybourne. and T E Mattis 1984. Infections with trypanorhynch plerocerci (Cestoda) in Hawaiian fishes of commercial importance. Univ. Hawaii Sea Grant Coll. 6(3):l-6, DOLLFUS, R P 1935. Sur quelques parasites de poissons recoltes a Cas- tiglione (Algerie). Trav. Publ. Stn. Agric. Peche Cas- tiglione, Annu. 1933, Vol. 2, p. 19-279. 775 FISHERY BULLETIN: VOL. 85, NO. 4 Heinz, M. L., and M. D. Dailey. 1974. The Trypanorhyncha (Cestoda) of elasmobranch fishes from southern Cahfornia and northern Mex- ico. Proc. Helminthol. Soc. Wash. 41:161-169. Kennedy. C. R , and P J Walker 1969. Evidence for an immune response by dace, Leucis- cus leuciscus , to infections by the cestode Caryophyllaeus laticeps. J. Parasitol. 55:579-582. King, J E.. and I I Ikehara. 1956. Comparative study of the food of bigeye and yel- lowfin tunas in the central Pacific. U.S. Fish Wildl. Serv., Fish. Bull. 57:61-85. KiSHINOUYE, K. 1923. Contributions to the comparative study of the so- called scombroid fishes. J. Coll. Agric. Imp. Univ. Tokyo 8:295-475. Lester. R. J. G . A Barnes, and G Habib 1985. Parasites of skipjack tuna, Katsuwonus pelamis: fishery implications. Fish. Bull., U.S. 83:343-356. Mackenzie, K. 1975. Some aspects of the biology of the plerocercoid of Gilquinia squali Fabricius 1794 (Cestode: Trypanorhyn- cha). J. Fish Biol. 7:321-327. 1983. Parasites as biological tags in fish population stud- ies. Adv. Appl. Biol. 7:251-331. MuRNo, H N . AND A Fleck 1966. Recent developments in the measurement of nucleic acids in biological materials. Analyst 91:78-88. Orr, T S. C, C. a. Hopkins, and G H Charles. 1969. Host specificity and rejection of Schistocephalus solidus. Parasitology 59:683-690. Overstreet, R. M. 1977. Poecilancistnum caryphyllum and other try- panorhynch cestode plerocercoids from the musculature of Cynoscion nebulosus and other sciaenid fishes in the Gulf of Mexico. J. Parasitol. 5:780-789. Pollard, D. A 1971. The biology of a landlocked form of the normally catadromous salmoniform fish Galaxias maculatus (Jenyns). I. Life cycle and origin. Aust. J. Mar. Fresh- water Res. 22:91-123. 1972. The biology of a landlocked form of the normally catadromous salmoniform fish Galaxias maculatus (Jenyns). IV. Nutritional cycle. Aust. J. Mar. Fresh- water Res. 23:39-48. 1973. The biology of a landlocked form of the normally catadromous salmoniform fish Galaxias maculatus (Jenyns). VI. Effects of cestode and nematode para- sites. Aust. J. Mar. Freshwater Res. 24:281-295. PouPA, O , and B Ostadal. 1969. Experimental cardiomegalies and "cardiomegalies" in free-living animals. Ann. N.Y. Acad. Sci. 156:445- 468. Sindermann, C J 1983. Parasites as natural tags for marine fish: a re- view. NAFO Sci. Coun. Studies 6:63-71. Smith, H D 1973. Observations on the cestode Eubothrium salvelini in juvenile sockeye salmon (Oncorhynchus nerka) at Babine Lake, British Columbia. J. Fish. Res. Board Can. 30:947-964. Tester, A L., and E L Nakamura. 1957. Catch rate, size, sex, and food of tunas and other pelagic fishes taken by trolling off Oahu, Hawaii 1951- 55. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 250, 25 p. UCHIYAMA, J. H„ and P. STRUHSAKER. 1981. Age and growth of skipjack tuna, Katsuwonus pelamis. and yellowfin tuna, Thunnus albacares, as indi- cated by daily growth increments of sagittae. Fish. Bull., U.S. 79:151-162. Van Dobben, W H 1952. The food of the cormorant in the Nether- lands. Andrea 40:1-63. Waldron, K D.. and J E King. 1963. Food of skipjack tuna in the central Pacific. InH. Rosa, Jr. (editor). Proceedings of the World Scientific Meeting on the Biology of Tunas and Related Species, La Jolla, July 2-14, 1962, p. 1431-1457. FAG Fish. Rep. 6, Vol. 3. Ward, J. W. 1962. Helminth parasites of some marine animals, with special reference to those from the yellow-fin tuna, Thun- nus albacares (Bonnaterre). J. Parasitol. 48:155. 776 AGE AND GROWTH OF SPANISH MACKEREL, SCOMBEROMORUS MACULATUS , FROM FLORIDA AND THE GULF OF MEXICO William A. Fable, Jr., Allyn G. Johnson, and Lyman E Barger' ABSTRACT Otoliths from 1,787 Spanish mackerel, Scomberomorus maculatus , were used to estimate age and growth rates of this species from Florida and the Gulf of Mexico. There was a wide range of lengths within an age group: the oldest male was 7 years old, while the oldest female was 9 years old. Length at age was significantly different for sexes, sampling areas, and collection gear. The von Bertalanffy growth equations were as follows: males (all cireas combined) /, = 794 (1 - e~0.24« + 0.94); females (all areas combined) /, = 739 (1 - e-0 33(< + 0 99); males (Florida only) If = 776 (1 - e-0.27(( + 0.73); females (Florida only) /, = 731 (1 - e-0.38« + 0.73)^ where / = fork length (mm) and t = years. Spanish mackerel, Scomberomorus maculatus, are found in the western Atlantic Ocean from the Gulf of Maine to the Yucatan Peninsula (Collette et al. 1978), and have their center of abundance off Florida (Trent and Anthony 1978). They sup- port extensive commercial and recreational fish- eries in the U.S. south Atlantic and Gulf of Mex- ico. In 1985, U.S. commercial landings totaled 5.8 million pounds (2,631 t) (U.S. Department of Commerce 1986a) while recreational landings were estimated to be 2.1 million pounds (953 t) (U.S. Department of Commerce 1986b). Informa- tion on Spanish mackerel published prior to 1978 actually concerned two species, .S . maculatus and S. brasiliensis (Collette et al. 1978). Collette et al. (1978) determined that Spanish mackerel south of the Yucatan Peninsula (on the Cen- tral and South American Atlantic coasts) are S. brasiliensis, and those along U.S. coasts are S. maculatus. There is disagreement in the literature on the interpretation of annuli on otoliths of Spanish mackerel. The first information on age and growth of S. maculatus was from fish collected in southeast Florida (Klima 1959). Later, Mendoza (1968) gave some limited age and growth infor- mation on S. maculatus from Veracruz, Mexico, and Powell (1975) provided the most recent infor- mation on Spanish mackerel age, growth, and ^Southeast Fisheries Center Panama City Laboratory, Na- tional Marine Fisheries Service, NOAA, 3500 Delwood Beach Road, Panama City, FL 32407-7499. reproduction in Florida. Powell interpreted an- nuli on Spanish mackerel otoliths differently than did Klima, and the different age determina- tions yielded different growth estimates. Men- doza (1968) did not estimate growth except by presenting his data in tabular form. We undertook this investigation to resolve these uncertainties in the literature and to derive more current age and growth parameters. This information will provide a better basis for ra- tional management of this species. STUDY AREA AND METHODS We collected 1,929 Spanish mackerel from 1977 through 1981 from the south Atlantic and Gulf of Mexico coasts of the United States. Most (1,422) of the fish came from northwest Florida and only 10 came from north of south Florida on the At- lantic coast (Table 1). Fork length (FL) of each Table 1. — Numbers of Spanish mackerel collected for age and growth study. Year Area 1977 1978 1979 1980 1981 Total Texas — 48 — 48 Mississippi/ Louisiana 41 79 — 23 — 143 Northwest Florida 59 377 31 955 — 1,422 South Florida — 87 31 59 129 306 Georgia — 10 — — — 10 Total 100 601 62 1,037 129 1.929 Manuscript accepted July 1987. FISHERY BULLETIN: VOL 85, NO. 4, 1987. 777 mackerel was measured to the nearest millime- ter. Sagittal otoliths were removed, washed, and stored dry. The clearest, most legible otolith from each fish (based on visual observation) was exam- ined to estimate age and growth. Whole otoliths were placed in a black-bottomed watch glass containing 100% glycerin and exam- ined with a binocular microscope at 28 x using reflected light. Otolith radius (OR) was measured in ocular micrometer units (1 unit = 0.0363 mm) on the posterior surface from the focus to the dis- tal margin along the axis of the sulcus acousticus (Powell 1975). Growth marks were counted and measured from the focus along the radius to their distal edge. The marks were opaque (light) under reflected light, while the interspaces were hya- line or translucent (dark). Otoliths were classified into age groups based on the number of opaque nonmarginal marks (Powell 1975). A mark was considered complete when a hyaline (dark) interspace or margin was visible from successive growth. Three readers in- dependently examined each of 520 otoliths to test the precision of our ageing technique. This infor- mation was analyzed using the method of Beamish and Fournier (1981). All other otoliths were independently examined by two readers; if their results did not agree, the data were not used. To compare age estimates based on surface (whole) and internal (sectional) examination, we sectioned 70 otoliths which had been previously examined on the surface (2-10 otoliths from dif- ferent fish from each age 0+ through 8-1- ), follow- ing the methods of Johnson et al. (1983). We determined time of annulus formation and validated our ageing technique by comparing monthly percentage frequencies of otoliths with opaque margins. A high percentage frequency (>45%) indicated recent annulus formation. We used a chi-square test to compare the monthly frequencies. The relationship between otolith radius and fork length was determined and used to back cal- culate fork lengths at earlier ages (Tesch 1971; Ricker 1975; Everhart et al. 1975). We used anal- ysis of covariance (ANCOVA) with age as the co- variate to test for differences in growth rates (lengths at age) of fish collected in different loca- tions, by different gears, and of different sexes. Mean back-calculated lengths were used to calcu- late von Bertalanffy (1938) growth parameters, employing a computer program developed by Abramson (1971). FISHERY BULLETIN: VOL. 85, NO. 4 RESULTS AND DISCUSSION Validation Age validation has often been overlooked in the age and growth literature (Beamish and McFar- lane 1983). Although there are numerous meth- ods available to establish the annual nature of otolith growth rings, we applied marginal incre- ment analysis, because it was the only practical method to use on this migratory, pelagic species. Annulus formation occurred in March, April, or May (Fig. 1). A chi-square test (x^ = 338.47, df = 1, P < 0.001) showed a highly significant dif- ference between the occurrence of otoliths with opaque margins in these months versus the other nine months of the year. Our findings are in agreement with Powell (1975) in that the main period of opaque mark formation was in the spring or early summer. He reported mark forma- tion in May, June, and July by examination of marginal increments. Previously Klima (1959) described both summer and winter growth rings and evaluated the marginal condition to decide that marks were deposited annually. Our obser- vations on the appearance of annuli in Spanish mackerel otoliths agreed with Powell (1975), in that we also were unable to discern the "first win- ter mark" that Klima (1959) described. Age To estimate the precision of our ageing, we compared sections to whole otoliths and evalua- tions by different readers. Examination of 70 sec- tioned otoliths provided a 97.4% agreement with previous surface examination of the same otoliths. Surface age determinations of three readers on 520 otoliths had a 97.7% agreement. Using the technique of Beamish and Fournier (1981), the index of average percent error was 0.3273, which we think is excellent. Of 1,929 Spanish mackerel examined, 1,787 (92.6%), ranging from 148 to 802 mm FL, were aged. The oldest female was 9 years old, while the oldest male was 7 years old. Powell's (1975) oldest fish, a female, was 8 years old, while Klima's (1959) oldest males and females were both 6 years old. These data and the data presented in Tables 2 and 3 indicate that females live longer than males. We found a wide range of lengths within an age group for both sexes (Tables 2, 3), as did Powell, with some Spanish mackerel of age 0 through 5 in 778 FABLE ET AL.: AGE AND GROWTH OF MACKEREL N =1.929 JAN. FEB MAR. APR. MAY JUN. JUL. AUG. SEP. OCT. NOV. MONTHS [Numbers in parentheses are numbers of fish examined] Figure 1. — Monthly percentage frequencies of Spanish mackerel otoliths with opaque margins. DEC. the same size interval. In the closely related king mackerel, S. caualla, Johnson et al. (1983) re- ported a similar situation. Our results substanti- ate wide variation in growth rates of individual Spanish mackerel. Growth Otolith radius (OR) was closely correlated with fish length (FL). The curvilinear relation FL = 1.5091 ORi 2639 (^ = 0.944) had a slightly better fit than the linear equation FL = -102.8061 + 6.1295 OR (r = 0.936). We used the former equa- tion to back calculate lengths at former ages for 949 fish that had at least one annulus (838 fish had no annuli and were classified as age 0). Neither Klima nor Powell reported any equations for an OR versus FL relationship. The mean back-calculated annual increments of fork lengths for male and female Spanish mackerel from all areas and years combined (Ta- bles 4, 5) indicate that growth rates were rapid until age 5 in females and to age 6 in males (the age 6 increment in males was based on one fish). After these ages, growth rates slowed apprecia- bly. Early growth was more rapid in females than males (first annual increment 123.6 as compared to 98.7). However, males maintained a higher growth rate through age 6, except for age 5, when the female annual increment was 55.3 mm versus 47.9 mm in males. Our back-calculations for Spanish mackerel showed variation in mean fork lengths at age be- tween sexes, areas, and years (Table 6). Females from south Florida grew faster than any other group and males from there grew faster than any other males. For Spanish mackerel from north- west Florida, where the largest number of fish were collected, analysis of covariance (ANCOVA) indicated significant differences in growth (length-at-age) between sexes and collecting gears (Table 7). ANCOVA was also used to test the significance of growth differences among geographic areas 779 FISHERY BULLETIN: VOL. 85, NO. 4 Table 2. — Fork length (mm) composition, in percent, of male Spanish mackerel by age group (locations combined). Length group Age in years Total number of fish 175-199 100.0 200-224 100.0 225-249 100.0 250-274 100.0 275-299 68.3 31.7 300-324 76.5 23.5 325-349 51.0 47.1 1.9 350-374 47.7 47.7 4.5 375-399 20.7 59.8 17.1 1.2 400-424 5.5 78.2 14.5 425-449 22.2 33.3 22.2 22.2 450-474 9.1 54.5 18.2 9.1 475-499 50.0 50.0 500-524 16.7 33.3 33.3 16.7 525-549 40.0 40.0 20.0 550-574 50.0 575-599 33.3 66.6 600-624 625-649 650-674 675-699 700-724 9.1 50.0 100.0 100.0 100.0 Total 1 1 5 2 41 102 155 88 82 55 9 11 2 6 5 2 3 0 3 0 1 1 575 Table 3. — Fork length (mm) composition, in percent, of female Spanish mackerel by age group (locations combined). Length group Age in years Total number of fish 0 1 2 3 4 5 6 7 8 9 175-199 100.0 1 200-224 100.0 2 225-249 80.0 20.0 5 250-274 100.0 3 275-299 96.7 3.3 30 300-324 84.8 13.6 1.5 66 325-349 77.0 23.0 152 350-374 52.4 46.4 1.2 166 375-399 46.7 51.1 2.2 137 400-424 41.1 52.5 5.7 0.7 141 425-449 22,8 62.4 12.9 2.0 101 450-474 25.3 49.3 20.0 2.7 1.3 1.3 75 475-499 8.1 48.4 40.3 3.2 62 500-524 5.3 52.6 31.6 8.8 1.8 57 525-549 1.9 39.6 32.1 17.0 9.4 53 550-574 25.0 22.5 37.5 15.0 40 575-599 5.0 30,0 40.0 2.5 40 600-624 5.9 23.5 41.2 23.5 5.9 17 625-649 13.6 45.5 22.7 13.6 4.5 22 650-674 6.3 25.0 25.0 25.0 6.3 6.3 6.3 16 675-699 16.7 33.3 25.0 8.3 16.7 12 700-724 42.9 14.3 42.9 7 725-749 16.7 33.3 16.7 16.7 16.7 6 750-774 0 775-799 100.0 1 Total 1,212 780 FABLE ET AL.: AGE AND GROWTH OF MACKEREL Table 4. — Mean back-calculated fork lengths (mm) at age for male Spanish mackerel from all areas, 1977-81. Age group XFL at capture Average back-calculated PL at age N 1 2 3 4 5 6 7 1 363.9 237 296.9 II 413.7 33 306.4 382.2 III 488.1 7 318.2 415.4 458.2 IV 536.8 4 353.8 423.5 483.9 529.8 V VI VII 605.0 5 0 1 374.9 448.2 522.4 567.1 596.3 679.0 342.0 521.8 570.9 606.5 642.5 287 300.8 399.5 489.8 556.1 604.0 98.7 90.3 66.3 47.9 657.1 671.7 657.1 671.7 53.1 14.3 Weighted mean Annual increment Table 5. — Mean back-calculated fork lengths (mm) at age for female Spanish mackerel from all areas, 1977-81. Age group XFL at capture Average back-calculated FL at age /V 1 2 3 4 5 6 7 8 9 1 420.1 437 344.6 II 503.6 113 340.8 463.1 III 580.9 62 349.2 480.2 5508 IV 596.7 30 356.9 471.1 529.2 5800 V 682.0 11 359.9 471.0 565.1 625.4 666.6 VI 683.0 3 405.1 472.9 546.5 602.5 643.7 673.7 VII 654.7 3 324.1 431.8 485.4 529.2 572.3 617.3 645.7 VIII 696.0 2 329.7 458.9 521.0 557.6 615.2 649.6 670.5 688.0 IX 737.0 1 399.0 470.0 529.0 596.7 653.4 685.3 704.6 717.5 730.5 662 345.4 469.0 543.8 587 9 643.2 650.8 663.8 697.8 730.5 Weighted mean Annual Increment 123.6 74.8 44.1 55.3 7.6 13.0 34.0 32.7 Table 6. — Weighted means of back-calculated fork lengths (mm) for male and female Spanish mackerel from all areas and years having appreciable numbers (over 100) of mackerels sampled. Age group All locations Northwest Florida Louisiana All years South Flonda 1981 All years All Florida 1978 1980 1981 1978 1980 All years All years Males 1 285 303 356 281 300 293 321 384 332 299 II 356 403 465 347 392 380 384 483 438 399 III 1448 474 531 1470 1460 1463 1440 558 508 494 IV 538 584 1529 1529 1479 607 566 561 V 1561 652 1561 1561 1454 1652 1654 1631 VI 1657 1657 1657 1657 VII 1672 1672 1672 1672 Females 1 325 347 371 326 346 342 334 366 364 348 II 428 465 507 434 466 454 436 509 500 475 III 492 526 573 486 536 517 500 574 573 557 IV '564 518 614 1555 1542 1548 518 615 614 607 V 1485 655 1536 655 654 654 VI 1540 665 1540 665 665 665 VII 1572 682 1572 682 682 682 VIII 1698 1698 1698 1693 IX 1730 1730 1730 1730 'Lengths based on less than 5 fish 781 FISHERY BULLETIN: VOL. 85, NO. 4 Table 7. — Results of analysis of covariance for growth differences observed in Spanish mackerel collected in northwest Florida, and fish collected in all areas by recreational hook and line, and gill net. Sum of Mean Tail prob- Source squares df square P ability Gill net Gear 2,499.15 2 1,249.57 13.04 0.00 Sex 5,275.31 1 5,275.31 55.03 0.00 Gear x sex 136.80 2 68.40 0.71 0.49 Age 41,384.91 1 41,384.91 431.73 0.00 Error 75,440.13 787 95.86 n = 794 NW Florida Area 3,792.09 2 1,896.05 22.83 0.00 Sex 476.13 1 476.13 5.73 0.02 Area x sex 689.00 2 344.50 4.15 0.02 Age 42,623.18 1 42,623.18 513.16 0.00 Error 38,955.39 469 83.06 n =476 Recreational hook and line Area 1,132.44 2 566.22 5.42 0.00 Sex 2,673.83 1 2,673.83 25.59 0.00 Area x sex 183.17 2 91.58 0.88 0.42 Age 124,338.42 1 124,338.42 1,190.00 0.00 Error 78,886.76 755 104.49 n = 762 (sex, area x sex, and age were also included in the covariance model) for recreational hook and line samples and gill net samples. Area differences were highly significant for both gear types, and sex differences were highly significant for gill net-caught fish, but somewhat less so for hook and line samples (Table 7). The area x sex inter- action was significant for hook and line, but not gill net samples. These ANCOVA results demonstrate that fe- males grew significantly faster than males. The significant differences between sampling gears are no doubt due to gear selectivity, i.e., hook and line selecting for larger fish of a given age and gill nets selecting for a specific size fish. Significant differences between sampling areas (consistent for both sampling gears) substantiate faster growth in south Florida (fish were larger at a given age) than in northwest Florida or Louisi- ana. We compared back-calculated lengths-at-age of Spanish mackerel (from all areas and from Flor- ida alone) with those of Powell (1975); lengths at ages 1 and 2 for both sexes were shorter, while those for ages 3-5 were increasingly longer (Table 8). There was a greater discrepancy between our data and Powell's for males than for females. Florida males from our study were 38 mm shorter than Powell's at age 1, but by age 5 they were 120 mm longer. Florida females from our study were Table 8. — Mean back-calculated fork lengths (mm) at age by sex for Spanish mackerel from Powell (1975) and this study. Powell's data were transformed from standard length by his formula FL= 1.0728 SL + 2.4267. Powell Fable et al. Age group Florida All areas Males Females Males Females Males Females 1 337 373 299 348 301 345 II 421 481 399 475 400 469 III 459 542 494 557 490 544 IV 489 580 561 607 556 588 V 511 621 631 654 604 643 VI 657 665 657 651 VII 672 682 672 664 VIII 698 698 IX 730 731 25 mm shorter than Powell's at age 1, but by age 5 they were 33 mm longer. Some of this discrepancy can be explained by the fact that Powell used the direct proportion method for his back-calculations, whereas the program by Abramson (1971) employs the regression method. Carlander (1981) pointed out potential problems with this method, but they primarily concern the fact that when using the scales for ageing, not all scales on a fish are the same size. This problem is of lesser importance when ageing is done from otoliths. Our estimates of the von Bertalanffy growth coefficient (k) are smaller, and our asymptotic 782 FABLE ET AL,: AGE AND GROWTH OF MACKEREL lengths (Lx) are larger (especially for males) than those derived by Powell (1975) and Nomura (1967) (Table 9). Nomura used Klima's (1959) data to compute growth curves for Florida fish. Our Lx estimates are much closer to the maxi- mum observed lengths in our samples (802 mm FL female and 723 mm FL male) than were the estimates from other authors. The differences be- tween our estimates and Powell's (1975) are eas- ily explained because we included the oldest fish in our back-calculations, whereas Powell only in- cluded fish up to 5 years old, forcing his growth coefficient (k ) to be higher. Therefore, we believe our growth parameters are a more accurate re- flection of population growth and more appropri- ate to use in assessment of the status of the stock. EVERHART. W. H., A. W. ElPPER, AND W. D. YOUNGS. 1975. Principles of fishery science. Cornell Univ. Press, Ithaca, N.Y., 288 p. Johnson, A. G,, W. A. Fable, Jr.. M. L. Williams, and L. E. Barger 1983. Age, growth and mortality of king mackerel, Scomberomorus cavalla, from the southeastern United States. Fish. Bull., U.S. 81:97-106. Klima, E F 1959. Aspects of the biology and the fishery for Spanish mackerel, Scomberomorus maculatus (Mitchill), of south- em Florida. Fla. Board Conserv. Mar. Res. Lab. Tech. Ser. No. 27, 30 p. Mendoza, a. 1968. Aspects of the biology and the fishery of the Spanish mackerel, Scomberomorus maculatus (Mitchill), in the state of Veracruz. Bios 1(2), 22 p. Nomura, H. 1967. Dados biologicos sombre a serra, Scomberomorus Table 9. — Von Bertalanffy growth parameters for Spanish mackerel. Males Females Author K L^ (FL mm) ^0 (years) K U (FL mm) to (years) Fable et al. all areas combined 0.24 794 -0.94 0.33 739 -0.99 Fable et al. Florida 0.27 776 -0.73 0.38 731 -0.73 Powell (1975) 0.48 555 -1.12 0.45 694 -0.78 Nomura (1967) using Klima's (1959) data 0.40 607 +0.15 0.40 720 + 0.28 ACKNOWLEDGMENTS We thank Richard Condrey, Churchill Grimes, and Edward Houde for their constructive reviews of this manuscript. Appreciation is also extended to Gerald Scott for his assistance in the statistical analyses. LITERATURE CITED Abramson, N J 1971. Computer programs for fish stock assessment. FAO Fish. Tech. Pap. 101. Rome, Italy. BEA.MISH, R J., and D a. FOURNIER 1981. A method for comparing the precision of a set of age determinations. Can. J. Fish. Aquat. Sci. 38:982-983. Bea.mish. R J.. AND G A. McFarlane. 1983. The forgotten requirement for age validation in fisheries biology. Trans. Am. Fish. See. 112:735-743. Carlander, K. D. 1981. Caution on the use of the regression method of back-calculating lengths from scale measurements. Fisheries 6(l):2-4. Collette, B. B , J. L. Russo, and L, A. Zavala-Camin. 1978. Scomberomorus brasiliensis , a new species of Spanish mackerel from the western Atlantic. Fish. Bull., U.S. 76:273-280. maculatus (Mitchill), das aguas cearenses. Arq. Estac. Biol. Mar. Univ. Fed. Ceara 7(l):29-39. Powell, D 1975. Age, growth, and reproduction in Florida stocks of Spanish mackerel, Scomberomorus maculatus. FDNR, Fla. Mar. Res. Publ. No. 5, 21 p. RiCKER, W E. 1975. Computation and interpretation of biological statis- tics offish populations. Fish. Res. Board Can. Bull. 191, 382 p. Tesch. R. W. 1971. Age and growth. /« W. E. Ricker (editor). Methods of assessment of fish production in fi-esh waters, p. 98- 130. Blackwell Scientific Publ., Oxf Trent, L., and E. A. Anthony 1978. Commercial and recreational fisheries for Spanish mackerel, Scom.beromorus maculatus. In E. L. Naka- mura and H. R. Bullis, Jr. (editors). Proceedings of the Mackerel Colloquium, p. 17-32. Gulf States Mar. Fish. Comm. No. 4. U.S. DEP-iVRTMENT OF COMMERCE. 1986a. Fisheries of the United States, 1985. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Curr. Fish. Stat. 8380, 121 p. 1986b. Marine recreational fishery statistics survey, At- lantic and Gulf coasts, 1985. U.S. Dep. Commer., Natl. Mar. Fish. Serv., Curr. Fish. Stat. 8327, 130 p. VoN Bertalanffy L. 1938. A quantitative theory of organic growth (inquiries on growth laws. 11). Hum. Biol. 19(2):181-213. 783 FEEDING HABITATS OF SPOT, LEIOSTOMUS XANTHURUS , IN POLYHALINE VERSUS MESO-OLIGOHALINE TIDAL CREEKS AND SHOALS^ Steven P O'Neil^ and Michael P. Weinstein^ ABSTRACT Young-of-year spot, Leiostomus xanthurus, were collected by otter trawl within tidal creeks and on adjacent shoals in polyhaline and meso-oligohaline zones of the York River, Virginia. Total densities of spot at Blevins Creek, a polyhaline system, were twice that of the meso-oligohaline Goalders Creek. Stomach content analysis confirmed previous studies of the generally opportunistic feeding strat- egy of juvenile spot. However, distinct differences in food utilization were observed between creeks and among creek and shoal stations. In addition, prey utilization differences due to habitat generally paralleled seasonal distribution patterns of dominant macrobenthos reported for the area. Two major ontogenetic groups were distinguished. Small spot {<30 mm SL) consumed more plank- tonic food items (calanoid copepods) than the larger size classes, which fed on more benthic prey and displayed greater overlap in diet. Small spot tended to be selective; larger spot were more opportunis- tic. Tidal salt marshes and their associated drainages are recognized primary nurseries for spot, Leio- stomus xanthurus, (Herke 1971; Parker 1971; Weinstein 1979; Currin et al. 1984). Shortly after recruitment, young spot tend to concentrate in tidal creeks, and by late spring densities in these creeks are often several times higher than in nearby seagrass habitats or shoal areas (Wein- stein and Brooks 1983; Smith et al. 1984). Once recruited to tidal creeks, spot seem to take up residence, w^ith limited movement out of (or be- tween) marshes until the fall mass exodus (Wein- stein 1983; Weinstein and Brooks 1983; Currin et al. 1984; Weinstein et al. 1984; Weinstein and O'Neil 1986). The role of marsh nurseries as predation refuges versus feeding areas is currently under debate (Boesch and Turner 1984). As suggested by the studies of Vince et al. (1976), it is likely that the marsh serves in both capacities. Qualita- tive and quantitative data on food availability and quality and on differences among habitats will be necessary to resolve the food versus refuge question. Ultimately, these data should be sup- ported by experimental studies on growth rates iVirgmia Institute of Marine Science Contribution No. 1419. 2Virginia Institute of Marine Science, Gloucester Point, VA 23062. 3Virginia Institute of Marine Science, Gloucester Point, VA 23062; present address: Lawler, Matusky & Skelly Engineers, One Blue Hill Plaza, Pearl River, NY 10965. Manuscript accepted July 1987. FISHERY BULLETIN: VOL 85, NO 4, 1987. versus the quality of food resources in different habitats (Weisberg and Lotrich 1982). We report here on one of the steps in the process, a descrip- tive comparison of gut contents of spot collected in tidal creeks and shoal areas in marshes of two salinity regimes, meso-oligohaline and poly- haline. Although the food habits of spot have been pre- viously studied, most investigators captured spot in openwater habitats, not in the primary nurs- eries (Parker 1971; Stickney et al. 1975; Chao and Musick 1977; Sheridan 1979). Only Hodson et al. (1981) studied food utilization of spot in tidal creeks. Their population, however, was restricted mainly to small fish (<40 mm) capable of exploit- ing the small creek rivulets and susceptible to capture by block net. This study expands the ef- fort of Hodson et al. (1981), and includes the en- tire seasonal residency period for spot in tidal creeks of the York River estuary, VA. A survey of food utilization was conducted in 1983 for all young-of-year size classes occupying two tidal creeks and nearby river shoals at widely sepa- rated salinities. Specific objectives of this effort were to 1 ) describe food utilization of juvenile spot in each habitat, 2) document any sequential onto- genetic changes in food utilization, and 3) com- pare the overall food utilization of spot residing in tidal creeks or adjacent shoals dissimilar in salin- ity. It was anticipated that feeding differences would reflect the availability and types of food in 785 FISHERY BULLETIN: VOL. 85, NO. 4 the two salinity regimes and microhabitats con- stituting the creek and shoal sites. STUDY AREA AND METHODS The York River estuary, a subestuary of the Virginia portion of the Chesapeake Bay (Fig. 1), covers about 208 km^ and extends 46 km from Tue Marsh Light to West Point, where it is formed by the confluence of the Pamunkey and Mattaponi Rivers. At two localities within the estuary, tidal creeks similar in physical dimen- sions (O'Neil 1983), but differing in salinity regimes, were selected as study sites: Goalders Creek, a meso-oligohaline site (sensu Remane 1934 and the Venice System of classification), and Blevins Creek, a polyhaline creek in the Guinea Marshes near the mouth of the river (Fig. 1). Field Methods Within each locality three stations were estab- lished: 1) in each creek approximately 1,500- 2,000 m upstream (where trawling was still possi- ble), 2) immediately inside the creek mouth, and 3) at shoal stations positioned approximately 200 m offshore in the York River proper in approxi- mately 3 m of water. Monthly collections (March-October 1982) with a 4.9 m semiballoon otter trawl with wings and body of 19 mm mesh and a 6.3 mm mesh cod end liner were made during daylight hours as close to high tide as possible. Four 2-min tows at about 1 m s"^ were made at each station. To reduce the chances of regurgitation, speci- mens were initially anesthetized in a mixture of seawater and 0.02 mL quinaldine (mixed in 10 mL acetone). Buffered formalin (10%) was then added for preservation. The abdominal cavities of large fish (>80 mm) were pierced to allow suffi- cient preservation of food items in the stomach. Water temperature and salinity were recorded prior to trawling at each station. Laboratory Methods In the laboratory, spot from each collection were sorted and counted. Individual standard lengths (SL) were measured; when more than 50 spot were captured in a single collection, a ran- dom subsample of 30 fish was used for length measurements. XJb.:$2^ vsv^ YORKTOWN Figure 1. — York River, VA, and relative locations of tidal creeks examined. A Creek. Goalders Creek, B = Blevins 786 O'NEIL and WEINSTEIN: FEEDING HABITATS OF SPOT For gut content analysis, fish from each of the four trawl samples representing a given station were pooled and then divided into several size classes. Initially, 5 mm size increments were used in order to corroborate the findings of others con- cerning an ontongenetic shift in feeding habits of spot. When mean standard lengths exceeded 20 mm, 10 mm size classes were adopted. Initially, up to 20 stomachs were removed from randomly selected individuals in each size class. Later, based on prey item diversity (Hurtubia 1973) comparisons for the June samples, 12 stom- achs per size class was set as the upper limit (O'Neil 1983). Stomach contents were pooled within size classes and analyzed using the Carr and Adams (1972) sieve fraction technique. After washing stomach contents from each sieve (2, 0.85, 0.425, 0.25, 0.15, and 0.075 mm meshes) into a small fingerbowl, a random subsample of approxi- mately 5 mL was removed. The subsample was placed in a labeled vial and the remainder was filtered onto a preweighed 55 mm filter pad and dried for 24 hours at 60°C. On the assumption that food particles of roughly the same size have approximately the same weight (Carr and Adams 1972), the total dry weight for each sieve fraction was proportioned among the prey types identified from its subsample. The Carr and Adams tech- nique provided for rapid, accurate identification of food items from a large number of stomachs and has been used successfully by several investiga- tors (Sheridan 1979; Stoner 1980; Livingston 1982; Lucas 1982). Statistical Analysis Dietary differences among various ontogenetic groups, between creeks, between stations within creeks, or for each month examined were com- pared using "normal" classification methods (Clifford and Stephenson 1975). Overlap of prey utilization was then determined using the com- plement of the Bray-Curtis dissimilarity mea- sure: 2 K-^2,i _J. n 2 ^^lj+^2j) and X2 are the values of the jth attribute for any pair of entities (size, station, month). Separate matrices were constructed for each comparison from untransformed, pooled monthly data using COMPAH (Boesch 1977). The data in each matrix were then clustered by the group- average method (Lance and Williams 1967). Diet information was based on dry weights of the 30 prey taxa categories, all of which were mutually exclusive except for the unidentified (UID) and miscellaneous (MISC) categories (Table 1). Prey items contributing <0.1 mg of total dry weight per size class were eliminated prior to the analy- sis. The miscellaneous category contains the total of all food items individually representing <2% of the final dry weight. In addition to the clustering procedure, recipro- cal averaging ordination (Guinochet 1973; Hill 1973) was used to provide independent verifica- Table 1 . — Prey categories used for tropic com- parisons. All but unidentified (UID) and miscel- laneous (t^ISC) are mutually exclusive feeding categories. where n is the number of attributes (prey) and Xj AMP Amphipoda BIV Bivalves BRA Branchipoda CAL Calanoids CAP Caprellidae CLS Clam siphons CHI Chironomidae CHL Chlorophyta COR Corophiidae Crs Crangon septemspinosa CRZ Crab zoea Cs Callinectes sapidus CYA Cyathura DET Detritus Eh Eteone heteropoda Et Edotea tribola FOR Foraminifera GAM Gammaridae HA, Harpacticoid 1 HA2 Harpacticoid 2 La Leucon americanus Lp Leptocheirus plumulosus MAC Macoma sp MAL Maldanidae Me Monoculodes edwardsi MISC Miscellaneous Na Neomysis americana NEM Nematoda NER Nereidae OLI Oligochaeta OST Ostracods PAL Palaemonidae PI Polydora ligni PLA Plant matter POL Polychaeta SPI Spionidae TEL Teleosteii UID Unidentified remains XAN Xanthidae 787 FISHERY BULLETIN: VOL. 85, NO. 4 tion of the dendrogram results. Reciprocal aver- aging is an eigenanalysis that ordinates both food type and habitat (or size class) variables simulta- neously and defines axes such that the variance of the scores on each axis is maximized. The first axis, therefore, represents the path of maximum variance, the second axis the next greatest, and so forth. This analysis was performed with ORDI- FLEX (Gauch 1977). RESULTS Physical Parameters With the exception of April and May, tempera- tures were slightly cooler at Blevins Creek than at Goalders Creek (Table 2). Salinity within Goalders Creek was reasonably stable consider- ing its meso-oligohaline location (Table 2). Except for a brief period in spring, Blevins Creek was polyhaline during the period of spot residence (salinity range 18-22%ci). Salinity in Goalders Creek was always at least 4%o lower than Blevins Creek and reached a maximum difference of 14%c during April. Such variations in tidal creeks is typical of the estuarine salinity gradient with dis- tance from the head of the estuary (Weinstein 1979; Weinstein et al. 1980). There were no dis- tinct salinity differences observed between either creek and its adjacent shoal station. Temporal Abundance and Distribution Monthly abundance and distribution patterns for spot in each creek system and adjacent shoals are shown in Figure 2. Overall, numbers of spot captured within the tidal creeks were similar, 2,355 versus 2,802 in Goalders and Blevins Creeks, respectively. Temporal distributions of spot within each locality were further compared by computing creek/shoal ratios. Spot were not encountered during the first sam- pling trip during late March 1982, but postlarvae and juveniles appeared in small numbers in April. At that time, spot were more abundant at the shoal stations than in the creeks (creek/shoal ratio of 0.28 for Goalders and 0.16 at Blevins). Young-of-year spot reached their maximum abundance in May, with 1,047 specimens taken up-estuary at Goalders Creek and 2,110 individu- als sampled from Blevins Creek. Spot at Goalders Creek were then more numerous at the stations within the creek (ratio 20.5), but still more preva- lent on the shoal down-estuary at Blevins Creek (ratio 0.52). From June to September, however, spot were clearly more abundant in the creeks of both systems. By the end of the investigation (Oc- tober 1982) spot once again dominated the shoal at Blevins Creek, but remained more abundant in the creek at Goalders. Monthly size distributions of spot in the two tidal creeks and adjacent shoals were examined by dividing the samples taken at each station into 5 mm SL size classes and comparing their relative frequencies among stations and locations. With the exception of a short period during recruitment (May) when more small fish were collected in Goalders Creek than at the nearby shoal station, none of the size-frequency comparisons differed significantly (Friedman's ANOVA, P < 0.05; O'Neil 1983). Table 2.— Monthly temperature (°C), salinity (%o), and vali jes and sediment estuary, 1982. analysis (% total dry weight) by trawl station 1, York River Goalders Creek Blevins Creek Month Upstream Downstream Shoal Upstream Downstream Shoal Mar. 11.0 (2.0) 11.0 (2.0) 11.0 (2.0) 9.5 (16.0) 9.5 (16.0) Apr. 13.5 (5.0) 13.5 (5.0) 15.0 (7.5) 16.0 (18.0) 16.0 (18.0) 17.0 (19.0) May 20.5 (10.0) 20.0 (10.0) 21.0 (11.0) 24.0 (16.0) 24.0 (20.0) 24.0 (18.0) June 26.0 (7.0) 26.0 (7.0) 26.0 (8.0) 25.5 (20.0) 25.0 (19.0) 25.0 (18.0) July 29.0 (10.0) 29.0 (13.0) 29.0 (13.0) 28.0 (22.0) 28.0 (20.0) 29.0 (22.0) Aug. 28.0 (11.0) 28.0 (11.5) 28.5 (10.0) 27.0 (22.0) 27.0 (19.5) 27.0 (20.0) Sept. 26.0 (13.0) 26.0 (14.5) 26.0 (16.0) 25.0 (21.0) 26.0 (21 0) 25.0 (20.0) Oct. 16.5 (11.0) 17.0 (11.0) 17.0 (12.0) 14.0 (20.0) 14.5 (22.0) 15.0 (20.0) Sediments (Sample cores taken in May) Sand and gravel 83.45 29.86 11.06 52.21 59.07 93.07 Silt 7.72 27.29 43.87 33.83 27.92 3.05 Clay 8.83 42.85 45.07 13.96 13.01 3.83 Organics 9.12 15.96 10.79 4.09 5.16 0.74 788 O'NEIL and WEINSTEIN: FEEDING HABITATS OF SPOT BLEVINS CK ■ CREEK □ SHOAL Figure 2. — Relative densities of spot at tidal creeks (values shown are monthly means of both creek stations) and shoal sampling localities. Asterisk indicates that May values for Blevins Creek are drawn to half scale. Values above histograms are ratios of creek to shoal densities. Trophic Analysis During this study, over 1,750 spot stomachs were removed and analyzed. In both creeks, spot underwent size-related, as well as temporal and spatial, changes in food utilization. Food utiliza- tion differences owing to size-related (ontoge- netic) changes were examined by cluster analysis (Fig. 3). Calanoid copepods were the dominant prey of the smallest spot size classes (Fig. 4). The 26-30 mm size class had begun to consume more substrate-oriented prey (polychaetes and ne- matodes). All the spot examined between 40 and 100 mm SL had considerable overlap in a wide variety of food items. The great majority, how- ever, were benthic organisms, e.g., maldanid and nereid polychaetes, Leptocheirus amphipods, free- living nematodes, and oligochaetes. Spot over 101 mm were clustered separately because of Leucon americanus in the diet. It thus appears that onto- genetic changes in spot diet shifted from a spe- cialist mode when small to a more opportunistic strategy in larger size classes. Size-class data were also subjected to reciprocal averaging ordination (Fig. 4). Results closely par- allel those in the numerical classification. Axis 1, accounting for 49% of the variance, defined the small, planktonic size classes, which consumed mostly calanoid copepods. The spot over 101 mm were separated along Axis 2, with Leucon ameri- canus and Monoculodes edwardsi the dominant food items. The remaining size classes lay in the plane of Axes 2 and 3 in association with a large variety of benthic prey. The dendrogram representing the differences between stations for all size classes of spot pooled (Fig. 5) indicated that there are two main clusters that correspond to the food distinctions between the two creeks. In addition, both shoal stations clustered as distinct outliers. Dominant prey items at the Goalders Creek sites included nereid polychaetes, clam siphons, a gammarid amphipod (Leptocheirus plumulo- sus), and harpacticoid copepods. At Blevins Creek, spot utilized proportionately more ne- matodes, maldanid polychaetes, and oligochaetes. At both locations spot made significant use only of specific parts of some prey items, i.e., clam siphons and tails of maldanid polychaetes (Currin et al. 1984). Prey utilization differences were also noted be- tween the creek stations and the adjacent shoal. At Goalders Creek the amphipod Monoculodes edwardsi, which dominated feeding on the shoal, was partially responsible for the separation noted in the dendrogram (Fig. 5). In the polyhaline sys- 789 FISHERY BULLETIN: VOL. 85, NO. 4 0- io: ^ 20 : 30 : > 40 : 50 : < -I 60 : z en 70 : 80; 90 : 16-20 CAL PI CS PLA urn: :-OLi 21-25 CAL HA^ CS HA2 NEM PI MISC -UID NER 26-30 OLI NEM NER HA2 HAi DET 130 24 JMISC 31-40 NEM OLI HA2 HAi CS MAL NER m^ 4/211 ^ufi 41-50 MAL HA- NEM NER CS Eh . OLI . MISC HAi OST 51-60 HA2 NEM OLI MAL NER T5sr m. MISC Me PLA 61-80 HA. Lp NER NEM CS DET MAL PLA Hsr: MISC ^Eh :-Cs Me 3/333 2/260 3/388 81-100 NER CHL NEM HA-, OST PLA CS MAL \k MISC 1-DET ^T '100 Me PLA La HA- GS NER NE ffiT MISC -SPI -FOR OST •AMP ^Na 14/264 5/61 Figure 3. — Cluster analysis of prey similarity among Leiostomus size classes for the York River estuary, 1982. Prey abbreviations are listed in Table 1. Ratios at bottom of each column represent number of empty stomachs/ total sample size. X < 100 CL GM 01 31 40'+' 16-20 21 25 r#; 41 50 0 UT •101 -?r-Axis2 erao Na Et 81-100 Figure 4. — Reciprocal averaging of prey and spot size class. Prey abbreviations: CH = Chlorophyta, CL = Calanoid copepods, Cs = Crangon septemspinosa, Eh = Eteone heleropoda, Ei = Edotea triloba, FR = Foraminifera, GM = Gam- maridae, H] = Small harpacticoid copepods, H2 = Large harpacticoid copepods, La = Leucon americanus , Me = Monoculodes edwardsi , ML = Maldanidae, Na = Neomysis americana , OL = Oligochaeta, OR = Orbiniidae, OT = Ostracods, PL = Plant matter. 790 ONEIL and WEINSTEIN: FEEDING HABITATS OF SPOT 10_ _ 20_ > 30_ t- §40: i 50_ 60_ 1- 1 1- z NER Ill Lp > o OST 1- z HA2 o NEM a CS PLA Cp 1/273 GU -DET MISC CS Lp HA. NEM OST HAi NER PLA SPI MISC 4/380 GD HAo Me PLA Lp La CS HAf ZSsZ MISC J-Et GA NEM NER HA, CHL OLI MAL usr MISC -PLA -Eh 2/228 GS 2/284 BU NEM MAL HA, CS CHL OLI PLA DET "PST' MISC -LF ■Eh -NER -HA, 5/277 BD MAL OLI NEM HA, CS Eh SPI "HTTT FOR MISC -ORI ■LF 17/311 BS Figure 5. — Cluster analysis of prey similarity among habitats for spot from a polyhaline (BD = Blevins downstream, BU = creek, VA, 1982. Prey abbreviations cire listed in Blevins upstream, BS = Blevins shoal) and a meso-oligohaline (GU = Goalders upstream, etc.) tidal Table 1. Ratios at the bottoms of each column represent number of empty stomachs/total sample. tern, however, there were few clear differences caused by presence or absence of particular prey types. Instead, it was more a question of which food item was dominant. Spot appeared to eat more nematodes within the creek and more mal- danid polychaetes on the shoal. To confirm the results of the classification anal- ysis, reciprocal averaging was used on the same food-habitat matrix (Fig. 6). Prey items located near a given station "lollipop" are the dominant food items utilized by spot at that station. Axis 1, accounting for 45% of the data variance, clearly separated the low and high salinity creek sys- tems. Axis 2 (28% of the variance) isolated the shoal stations relative to the intracreek sites. Ne- matodes and maldanid polychaetes were again closely associated with the Blevins Creek sites. Nereids, Leptocheirus, and Monoculodes were dominant at the Goalders Creek habitats. To compare seasonal patterns in food utiliza- tion between habitats, classification dendro- grams were also constructed using monthly data for each creek (Figs. 7, 8). At Goalders Creek there was little overlap of prey utilized in April compared to all other months (Fig. 7). The main reason for this appears to be the large proportion of calanoid copepods consumed in April. August and October were grouped together because of the amount of nereids eaten, and the remaining months were added to this cluster individually, depending on their overall dissimilarity. April was also an outlier at Blevins Creek, be- cause of the dominance of calanoids in the diet of young spot (Fig. 8). May and June were clustered together because of the similarity in the consumption of maldanid polychaetes and nematodes. August and September were similar in the proportions of four prey items utilized: mal- danids, nematodes, nereids, and harpacticoid copepods. Although these food items were proba- bly incidental in their diet, July was a separate group because of the large amount of Chloro- phyta present in the stomachs examined from that month; October was isolated because Foraminifera became an important addition to the diet. 791 FISHERY BULLETIN: VOL. 85, NO. 4 AXIS 2 Figure 6. — Reciprocal averaging of prey and habitat for spot col- lected from tidal creeks and adjacent shoals of the York River, VA, 1982. Station abbreviations are the same as Figure 4. Prey abbrevia- tions: Cp = Cyathura polita , CS = Clam Siphons, Cs = Crangon septemspinosa , GM = Gammaridae, Lp = Leptocheirus plumulosus , Me = Monoculodes edwardsi, ML = Maldanidae, NM = Nema- toda, NR = Nereidae, OL = Oligochaeta, PL = Plant Matter, TE = Teleostei. 0 10 20 30 40 50 60 70 80 90 100 I o > a Q H- Z UJ u a m a. - 4* > OC < 1 _^ET -GAM —AMP u> CAL HA2 Lp NER NER ,HAt ^AMP La HA2 NER HAj CS PLA Lp CS Me Lp NEM SPI NEM CS HAi PLA OST OLI OST PLA OST OST PLA HA2 CS NEM Me HA2 HA2 NEM CS 6§T CS PLA NEM 100 m were off Cape Hatteras, North Carolina. Despite the presence of horseshoe crabs in estuaries as far north as Maine, New York is the northward limit on the shelf. This suggests that inshore populations in New England may be relatively isolated from each other and from the large Middle Atlantic shelf population. Generalizations about the natural history, behav- ior, and ecological importance of horseshoe crabs, Limulus polyphemus (L.), are primarily based on studies of the shallow-water phase of its life cycle (Shuster 1979, 1982; Wells et al. 1983). The popu- lations in the mid-Atlantic region are most acces- sible in late spring and early summer, when adults spawn en masse on sandy estuarine beaches. Knowledge of behavior (Shuster 1950; Rudloe 1980; Barlow et al. 1982; Cohen and Brockmann 1983), orientation (Rudloe and Herrnkind 1976; Botton and Loveland in press), morphometries (Shuster 1955; Riska 1981), sedi- ment disturbance (Woodin 1978, 1981), and pre- dation (Smith and Chin 1951; Smith et al. 1955; Botton 1984a, b) is all based on studies of shallow- water or intertidal individuals. Population esti- mates have been restricted to shallow-water adults (Baptist et al. 1957; Sokoloff 1978; Rudloe 1980; Shuster and Botton 1985), with the excep- tion of Botton and Haskin (1984), who surveyed the population on the inshore New Jersey conti- nental shelf. Perhaps because of the spectacular intertidal ^Division of Science and Mathematics, Fordham University, College at Lincoln Center, 113 West 60th Street, New York, NY 10023. ^Northeast Fisheries Center Woods Hole Laboratory, Na- tional Marine Fisheries Service, NOAA, Woods Hole, MA 02543. Manuscript accepted July 1987 FISHERY BULLETIN: VOL. 85, NO 4, 1987. mass spawning phenomenon and accessibility, es- tuarine populations have received a dispropor- tionate amount of attention by ecologists. In con- trast, most of the animal's life is spent sublittorally. An adult female may spawn com- pletely over several successive high tides; in gen- eral, repeated breeding is more characteristic of males (Rudloe 1980). Juveniles, during their first and second summer, are abundant on intertidal flats (Shuster 1955, 1979), but the remainder of the species' 14-19 year life span (Ropes 1961) is spent subtidally except for the annual spawning migration. This report summarizes latitudinal and bathy- metric distributions of horseshoe crabs on the northwestern Atlantic continental shelf. North- east Fisheries Center (NEFC) bottom trawl and ocean clam surveys during the past two decades have provided extensive data on the abundance and distribution of horseshoe crabs, principally north of Cape Hatteras, NC (Ropes et al. 1982; NEFC unpubl. data). Seasonal and annual trends in abundance are also discussed in the present report. Concern for evaluating the general popu- lation characteristics of horseshoe crabs parallels expanding commercial exploitation of the species. Animals are presently harvested to extract blood for the Limulus amoebocyte lysate (LAD test, and as bait in eel (Anguilla rostrata), conch (Busycon sp.), and other fisheries (Pearson and Weary 1980); the vast majority of the fishing ef- 805 FISHERY BULLETIN: VOL. 85, NO. 4 fort is concentrated between Virginia and New Jersey (Botton and Ropes in press). MATERIALS AND METHODS NEFC bottom trawl surveys ranged from Cape Fear, NC, north to the Scotian Shelf; clam sur- veys cover the Middle Atlantic region north to Georges Bank. Both are general purpose survey programs, collecting data for standing stock as- sessments for many finfish and shellfish species, as well as collecting specimens for age determina- tion, dietary analysis, and many other purposes (Grosslein 1969; Clark 1979). During the period covered by this study (1975- 83), the #36 and #41 Yankee otter trawls were used as sampling gear during bottom trawl sur- veys. Both nets have a mesh size of 5 inches throughout the wings and body, and 4.5 inches in the cod end. A liner of 0.5-in nylon mesh is em- ployed at the aft end of the top belly and the cod end. Trawls are equipped with roller gear to facil- itate use over rough bottoms. All tows were 30 minutes in duration at a vessel speed of 3.5 knots; stations were located using loran. An average of 1,129 stations per year was sampled (range, 711 stations in 1975 to 1,547 stations in 1979). Cruises were conducted during fall (September through early December) and spring (March through May) in all years, with five additional surveys during the summers (July through Au- gust) of 1977-81 and two during the winters (Jan- uary through February) of 1978 and 1983. Sam- pling was conducted both day and night. After sorting the catch, personnel recorded the number of horseshoe crabs taken in each tow and their total wet weight to the nearest 0.1 kg. A stratified random sampling design was used in the surveys (Grosslein 1969). The region was divided into several strata based primarily on depth. Stations were allocated to strata roughly in proportion to the area of each stratum and as- signed to specific locations within strata at ran- dom (Clark 1979). For the purpose of this paper, stations of 9-27 m (5-15 fm) depth were defined as "inshore" and those deeper than 27 m as "offshore". Preliminary inspection of the catch data compiled by Ropes et al. (1982) showed the bulk of the horseshoe crab population to be lo- cated between northern New Jersey and southern Virginia. Within this region, there were 27 in- shore and 16 offshore strata, based on depth and location. Stratified mean number per tow and biomass per tow of horseshoe crabs were con- verted into estimates of standing stock by using the "area swept" by a standard survey trawl in relation to catch as an estimate of minimum abso- lute density. Tows within strata were used to cal- culate variances around the means. Total popula- tions for the inshore and offshore regions were estimated by expanding the average stratified mean catch per tow by the ratio of total area sur- veyed to the area sampled by an average tow. Further details of statistical methods are found in Clark (1979). Ocean clam surveys used commercial style hy- draulic dredges, towed for 5 minutes at 1.5 knots. The design and performance of this sampling gear is discussed by Meyer et al. (1981). Cruises from 1965 to 1978 used either a 30-in (91 cm) or 48-in (122 cm) knife-width dredge; cruises from 1979 to 1983 used a 60-in (152 cm) dredge. There were no ocean clam surveys in 1968, 1971-73, and 1975. Station locations were selected using a stratified random sampling design; the average number of stations sampled per year was 370 (range, 139 in 76 74 72 70 42 40 38 36 34 r Sp ring 1 ' 1 ■ \ O *\ 0 • or '^- _^ ' i -d • / h i (} • / Ches Bay ir© 4 t-1 4 ^v' \ i. • >5.0 0 1.0-4.9 • 0.1 - 0.9 • 0.01-0.0 9 o 0.00 / / /o Figure l. — Mean number of horseshoe crabs per tow, 806 BOTTON and ROPES: POPULATIONS OF HORSESHOE CRABS 1979 to 655 in 1966). Catches were not expanded into stock estimates, as for the bottom trawl sur- vey data. To analyze broad latitudinal variations in abundance and frequency of occurrence, catch and sampling effort were grouped by 1-degree lat- itude and longitude blocks. For example, stations at lat. 38°06'N, long. 74°02'W and 38°45'N, 74°50'W were both grouped together as 38°N, 74°W. Bottom trawl and ocean clam survey data were analyzed separately, because strati- fication schemes differed and the efficiencies of the two types of sampling gear cannot readily be compared. RESULTS Latitudinal Distribution During groundfish surveys, 7,035 crabs were taken at 983 stations distributed from 33°N to 41°N; 75^^ of the crabs were caught between 37°N and 40°N (Fig. 1, App. Table 1). Highest abun- dance and frequency of occurrence was found on the shelf nearest the mouth of Delaware Bay. The maximum number of individuals per tow, 99, was obtained on 22 March 1976, from a station located off Assateague Island, VA, at 38°00'N, 75°14'W, in 13 m of water. Mean number per tow generally decreased with increasing distance from shore (Fig. 1). Horseshoe crab abundance decreased both north of 40°N and south of 37°N (Fig. 1). Crabs were absent northeast of Montauk Point, Long Island (41°N, 72°W), and no crabs were found on or north of Georges Bank despite intensive sam- pling effort. Fewer than 2% of all horseshoe crabs collected were found south of 35°N (Cape Hat- teras), and only one animal was caught south of 34°N. The observed latitudinal distribution of horse- shoe crabs in ocean clam surveys paralleled the above trends. A total of 1,640 animals was taken at 535 stations clustered primarily between Vir- FlGURE 1.— Continued ~hy T latitude and longitude blocks, based on NEFC groundfish trawl data from 1975 to 1983. 807 Table 1. — Distribution of horseshoe crabs in clam surveys, 1965-83; 1,640 animals enumerated at 535 stations. Number of crabs over (number of sta- tions with one or more individuals) in each row. FISHERY BULLETIN: VOL 85, NO. 4 Inshore Offshore Latitude (N) Longitude (W) 75° 74° 73° 72° 40° 4 (3) 23 (14) 4 (4) 39° 115 (46) 3 (3) 38° 237 (77) 351 (103) 3r 671 (176) 30 (16) 36° 153 (76) 1 (1) 35° 48 (16) ginia and New Jersey (Table 1). Thus, horseshoe crabs were most abundant along the continental shelf from Virginia to southern New Jersey in both groundfish and ocean clam surveys. Seasonal and Annual Variations in Standing Stock On the New Jersey-Virginia continental shelf, average abundance and biomass were highest during spring and fall cruises, lower in summer, and lower still (based on limited sampling) in winter (Fig. 1, Table 2). Horseshoe crabs were present in more than half the inshore tows throughout the year. Frequency of occurrence did not fluctuate as widely as abundance or biomass, and was consistently higher inshore than off- shore. Population estimates show considerable an- nual variation, with highest recorded catches in Table 2. — Average of seasonal variation in horseshoe crab abun- dance and biomass on the continental shelf, northern New Jersey to southern Virginia, based on groundfish surveys from 1975 to 1983. Strata shallower than 27 m were defined as inshore, and strata deeper than 27 m were defined as offshore. Winter Spring Summer Fall Inshore Mean no. per trawl 0.41 7.44 3.26 4.54 % occurrence 56 76 61 78 Population (X 106) 0.410 3.103 1.258 1.881 Biomass (metric tons) 93 3,626 2,075 2,969 Offshore Mean no. per trawl ' — 1.03 0.10 0.42 % occurrence — 29 30 22 Population (X 106) — i,445 0.129 0.530 Biomass (metric tons) — 1,975 231 972 1975 1976 1977 1978 1979 1980 1981 1982 1983 F S F S Su F W s Su F S Su F S Su F S Su F S F W s F -• ' ' I ' I I I I I I I'l [IT 2 4 6 8 10 12 14 16 18 0 2 4 6 8 Mean No. per Tow Figure 2. — Mean number of horseshoe crabs per tow (with 95% confidence Hmits) on the continental shelf from New Jersey to Virginia, based on NEFC groundfish trawl data from 1975 to 1983. W = winter; S = spring; Su = summer; F = fall; In- shore = all tows within sampling strata shallower than 27 m; Offshore = all tows within sampling strata deeper than 27 m. Where no confidence limit is shown, calculated limit was smaller than the width of the datum point. spring of 1976 and 1981 (Fig. 2). However, no clear trends in standing stock between 1975 and 1983 were evident. We estimate, based on the more complete fall and spring surveys, a mini- mum inshore population ranging from 1.8 to 3.1 million individuals (2,969 to 3,626 t), and a mini- mum offshore population ranging from 0.5 to 1.4 million individuals (972 to 1,975 t) (Table 2), for a total of some 2.3-4.5 million individuals. Coeffi- cients of variation, based on individual stratum catches, ranged from 11.6 to 41.6 for individual inshore survey estimates, and from 23.6 to 87.5 for offshore estimates. Bathymetric Distribution Horseshoe crabs were taken at stations be- tween the inshore sampling limit, 9 m, and 290 m depth (Fig. 3). Seventy-four percent of the total number caught in bottom trawl surveys were taken from stations shallower than 20 m; and 92% were caught at depths <30 m. This trend was not an artifact of sampling effort. Offshore sta- tions (>27 m) comprised approximately 73% of the sampling effort but produced <10% of the 808 BOTTON and ROPES: POPULATIONS OF HORSESHOE CRABS catch. Mean abundance and biomass were nearly an order of magnitude higher at the inshore strata than the offshore strata, and horseshoe crabs were at least twice as likely to be caught inshore than offshore during all seasons (Table 2). The operation of the hydraulic clam dredge was limited to waters <80 m, and within this range, crabs were again most abundant at shallower depths. Sixty-two percent of the total catch was found shallower than 20 m, and 90% was found shallower than 30 m (Fig. 3). Ninety-six animals were caught in 19 tows from 100 to 199 m depth, while 53 animals were taken in 7 tows below 200 m (Table 3). These deep-water individuals were caught between 34°21'N and 37°42'N (North Carolina to southern Virginia), with the majority from Cape Hatteras, south. o Trawl Dredge 150 >200 Depth (m) Figure 3. — Bathymetric distribution of horseshoe crabs, as per- cent of total catch, based on bottom trawl data (solid line) and ocean clam dredge data (dashed line). DISCUSSION Horseshoe crabs on the northwestern Atlantic continental shelf were most abundant between Virginia and New Jersey. On this section of the shelf, the population was estimated to be some 2.3-4.5 million individuals and was relatively constant between 1975 and 1983. The estimate of inshore abundance was necessarily conservative, because large survey vessels could not operate in shallow water. Horseshoe crabs may be abundant in areas inshore of the NEFC surveys; for exam- ple, in New Jersey State surveys, stations <1.8 km from shore frequently have over 10 horseshoe crabs/5-min hydraulic dredge tow in the Cape May area (Botton and Haskin 1984), and in areas of Narragansett Bay, 20-min otter trawls con- ducted by the State of Rhode Island have caught up to 40 horseshoe crabs (R. Sisson'^). Both in- shore and offshore stock estimates may also be conservative because the trawls, which are equipped with roller gear, may be <100% effi- cient in sampling horseshoe crabs, particularly if the animals are burrowed. Seasonal surveys of the middle Atlantic conti- nental shelf indicate a decline in horseshoe crab abundance during summer (July-August), which is consistent with the hypothesis that the shelf animals have seasonal inshore spawning migra- tions (Shuster and Botton 1985). More animals were caught during spring (April-early May) and fall (September-early December) periods, repre- senting the prespawning and postspawning sea- sons, respectively. Therefore, the Virginia to New Jersey shelf population probably consist largely of individuals which spawn during the spring and early summer in the Chesapeake and Delaware Bays, and disperse offshore. The range of horseshoe crabs on the continental 3R. Sisson, Division of Fish and Wildlife, Rhode Island De- partment of Environmental Management, Wakefield, RI 02879, pers. commun. December 1983. Table 3. — Occurrence of all horseshoe crabs below 100 m depth on the continental shelf in groundfish surveys from 1975 to 1983. Depth Latitude Longitude Number (m) (N) (W) Date of crabs 102 34°25' 75°53' 3/79 1 113 35°29' 74°50' 3/82 25 117 35°48' 74°52' 9/81 4 118 35°46' 74°51 ' 3/77 3 118 36°27' 74°47' 9/77 2 120 36''20' 74°48' 3/77 7 120 36°20' 74°48' 3/78 1 120 37^42 ■ 74°15' 9/79 1 120 36°10' 74°48' 9/80 2 125 36°20' 74°55' 3/79 4 133 34°56' 75°21 ' 3/79 2 135 35°00' 75°16' 9/79 4 137 35°33' 74°49' 3/76 21 170 36^1 0' 74°47' 3/76 1 173 35°51' 74°52' 3/82 4 183 35^41 ' 74°49' 9/77 8 186 34°21 ■ 75°55' 3/82 1 188 34°52' 74°24' 3/79 1 189 35M0' 74°48' 3/79 7 1189 35°39' 74°48' 9/81 7 205 35°53' 74°50' 3/76 18 205 35°37' 74°49' 10/76 12 220 35°50' 74=51' 9/79 2 228 34°27' 75°44' 9/79 1 246 36°25' 74°46' 3/76 10 290 35°20' 74°54' 3/80 3 809 FISHERY BULLETIN: VOL. 85, NO. 4 shelf is more limited than its estuarine distribu- tion. Although substantial shelf populations are restricted to the south and west of Long Island, NY, breeding populations are found in estuaries as far north as Hog Bay, ME (44°35'N) (Born 1977). Populations in Narragansett Bay, Barn- stable Harbor, Buzzards Bay, Cape Cod Bay, and Nantucket Harbor are large enough to be com- mercially exploited for Limulus lysate and/or bait (Botton and Ropes in press). Why such popula- tions remain close to shore is unclear. However, it is consistent with the data of Baptist et al. (1957). They showed that individuals in Plum Island Sound, MA, remained in the local area 3 years after tagging. The more northerly horseshoe crabs may be more discrete and isolated estuarine populations than those from North Carolina to New York. The small number of crabs on the southern New Eng- land shelf suggests that migrations of crabs be- tween estuaries may be limited, although such populations may be occurring at depths too shal- low to be sampled by large vessels. However, in the September 1985 trawl survey of the territo- rial waters of Massachusetts, only 34 horseshoe crabs were caught at 16 of the 94 stations sampled (mean depth 28 m, range 6-76 m) with similar numbers recorded during other recent surveys (B. Kelly'*). If, in fact, these New England popula- tions are isolated from the large Virginia-New Jersey stock, overexploitation may have serious detrimental effects. Although horseshoe crab lar- vae are weak swimmers, they are not commonly found in the plankton. Dispersal between discon- tinuous New England estuaries therefore de- pends on migration of juveniles or adults. How- ever, the issue of stock identity may require further study. Shuster (1979) argued, based on morphometric data, that horseshoe crabs formed discrete populations throughout the geographic range. On the other hand, Saunders et al. (1986) found no evidence for genetic divergence between New England and middle Atlantic populations, based on their analysis of mitochondrial DNA. The most noteworthy feature of the bathymet- ric distribution was the presence of horseshoe crabs at the edge of the continental shelf, at depths to 290 m. These animals were concen- trated off North Carolina, where the continental slope is much closer to shore than at any other location in the Middle Atlantic Bight. This sug- gests that distance from shore, rather than depth per se, limits the dispersal of crabs on the conti- nental shelf Horseshoe crabs are eurythermal, tolerating temperatures from -1.1° to over 40°C (Mayer 1914; Fraenkel 1960); neither of these ex- tremes are likely on the northwestern Atlantic continental shelf Laboratory animals in an elec- tronic shuttlebox arrangement voluntarily occu- pied temperatures from 15° to 40°C (Reynolds and Casterlin 1979), but the avoidance of cooler water may not apply to all populations, as all experi- mental animals were indigenous to the Gulf of Mexico. Our depth record at 290 m exceeds the 200 m record of Wolff (1977) but is not the maxi- mum depth attained by this species. A sub- mersible camera operated by the Duke Uni- veristy Marine Laboratory photographed a horseshoe crab at 1,097 m depth at 32°38'N, 76°33'W (D. BuntingS). The potential orientation cues directing such deep-water animals to and from estuarine spawn- ing beaches are of interest. Rudloe and Herrnkind (1976) showed that wave surge was important in determining the orientation of crabs in shallow waters near breeding sites, while Barlow et al. (1982) found that visual cues are important in the selection of cement "female models" by spawning males. Horseshoe crab eyes are sensitive to polar- ized light (Waterman 1950) and to low levels of visible light, and there are a variety of endoge- nous morphological changes that may permit photoreceptors to have high light sensitivity (Barlow et al. 1980). Whether such physiological properties are ecologically significant in enabling crabs to orient from the edge of the continental shelf to the estuarine spawning beaches is not yet known. Much remains to be learned about the ecologi- cal relationships between horseshoe crabs and other shelf fauna. Botton and Haskin (1984) found that adult horseshoe crabs were dietary generalists off the New Jersey coast, both in terms of taxa and sizes of food items selected. Predation by horseshoe crabs in Delaware Bay affects bivalve abundance, size-frequency pat- terns, and spatial distributions (Botton 1984b, c). Significant commercial fisheries for surf clams, Spisula solidissima , and ocean quahaugs, Arctica islandica , overlap the range of horseshoe crabs on the northwestern Atlantic continental shelf A study of horseshoe crab stomach contents is in *B. Kelly, Massachusetts Department of Marine Fisheries, East Sandwich, MA 02537, pers. commun. October 1985. 5D. Bunting, Duke University Marine Laboratory, Beaufort, NC 28516, pers. commun. April 1985. 810 BOTTON and ROPES: POPULATIONS OF HORSESHOE CRABS progress which will evaluate the importance of horseshoe crab predation to these bivalves. ACKNOWLEDGMENTS We are indebted to the scores of individuals who enumerated and weighed horseshoe crabs at sea. Special thanks are given to Ambrose Jearld for encouraging this study and to Linda Patanjo, Eva Monteiro, Ralph Mayo, and Loretta O'Brien for making data logs accessible and comprehend- able. Steve Murawski, Joe Idoine, and Gary Shep- herd assisted with computer analysis. Brian Kelly (Commonwealth of Massachusetts, Divi- sion of Marine Fisheries) and Richard Sisson (Rhode Island Department of Environmental Management, Division of Fish and Wildlife) have kindly collected and allowed us to use data on the occurrence of horseshoe crabs in their trawl sur- veys. David Bunting provided us with the deep- water photograph obtained on RV Eastward cruise E-9-66. The manuscript was improved by the critical reviews of Steve Murawski, Mike Fog- arty, Steve Clark, and Carl Shuster. The senior author is grateful for the support provided by the Fordham University Research Council Biomedi- cal Support Program, and the National Marine Fisheries Service. LITERATURE CITED Baptist, J P , O. R. Smith, and J. W Ropes. 1957. Migrations of the horseshoe crab, Limulus polyphemus, in Plum Island Sound, Massachu- setts. U.S. Fish. Wildl. Serv,, Spec, Rep, Fish,, No. 220, 15 p. Barlow, R. B., S. C. Chamberlain, and J. Z. Levinson. 1980. Limulus brain modulates the structure and func- tion of the lateral eyes. Science 210:1037-1039. Barlow, R B , L C Ireland, and L Kass. 1982. Vision has a role in Limulus mating behav- ior. Nature 296:65-66. Born, J W 1977. Significant breeding sites of the horseshoe crab {Limulus polyphemus) in Maine and their relevance to the critical areas program of the state planning of- fice. Planning Rep. 28, 44 p, Maine Critical Areas Program, State Planning Office, Augusta, ME. Botton. M L 1984a, Diet and food preferences of the adult horseshoe crah Limulus polyphemus , in Delaware Bay, New Jersey, USA. Mar. Biol. 81:199-207. 1984b. The importance of predation by horseshoe crabs, Limulus polyphemus, to an intertidal sand flat commu- nity. J. Mar, Res, 42:139-161, 1984c. Spatial distribution of three species of bivalves on an intertidal flat: the interaction of life-history strategy with predation and disturbance, Veliger 26:282-287, Botton, M, L., and H. H. Haskin. 1984. Distribution and feeding of the horseshoe crab, Limulus polyphemus , on the continental shelf off New Jersey, Fish, Bull,, U,S, 82:383-389, Botton, M L., and R, E. Loveland. In press. Orientation of the horseshoe crab, Limulus polyphemus , on a sandy beach, Biol, Bull, Botton, M. L., and J. W. Ropes. In press. The horseshoe crab fishery and resource in the United States, Mar, Fish. Rev. Clark, S. H, 1979, Application of bottom-trawl survey data to fish stock assessment. Fisheries 4(3):9-15, Cohen, J. A., and H. J. Brockmann. 1983. Breeding activity and mate selection in the horse- shoe crab, Limulus polyphemus . Bull, Mar, Sci. 33:275- 281, Fraenkel, G. 1960. Lethal high temperatures for three marine inverte- brates: Limulus polyphemus , Littorina littorea and Pagu- rus longicarpus . Oikos 11:171-182. Grosslein, M. D. 1969, Groundfish survey program of BCF Woods Hole. Comm. Fish, Rev, 38(8-9):22-30, Mayer, A. 1914, The effects of temperature upon tropical marine an- imals. Pap, Tortugas Lab, Carnegie Inst. Publ. 183, 6:1- 24. (Cited in Shuster 1979.) Meyer, T. L., R. A. Cooper, and K. J. Pecci. 1981, The performance and environmental effects of a hy- draulic clam dredge. Mar. Fish, Rev, 43(9): 14-22, Pearson, F., and M. Weary. 1980. The Limulus amoebocyte lysate test for endo- toxin, BioScience 30:461-464, Reynolds, W. W., and M. E. Casterlin. 1979. Thermoregulatory behaviour of the primitive arthropod Limulus polyphemus in an electronic shuttle- box. J, Thermal Biol, 4:165-166. RiSKA. B 1981, Morphological variation in the horseshoe crab Lim- ulus polyphem.us . Evolution 35:647-658. Ropes, J. W. 1961. Longevity of the horseshoe crab, Limulus polyphemus (L,), Trans, Am, Fish, Soc, 90:79-80. Ropes, J. W., C. N. Shuster, Jr., L. O'Brien, and R. Mayo. 1982. Data on the occurrence of horseshoe crabs, Limulus polyphemus (L.), in NMFS-NEFC survey sam- ples. Woods Hole Lab. Ref Doc. No. 82-23, 40 p. Rudloe. a. 1980, The breeding behavior and patterns of movement of horseshoe crabs, Limulus polyphemus , in Apalachee Bay, Florida, Estuaries 3:177-183. Rudloe, A., and W. F Herrnkind. 1976. Orientation of Limulus polyphemus in the vicinity of breeding beaches. Mar. Behav. Physiol. 4:75-89, Saunders, N. C, L. G. Kessler, and J C Avise. 1986, Genetic variation and geographic differentiation in mitochondrial DNA of the horseshoe crab, Limulus polyphemus . Genetics 112:613-627. Shuster, C. N., Jr. 1950. Observations on the natural history of the Ameri- can horseshoe crab, Limulus polyphemus . Woods Hole Oceanogr, Inst, Contrib. 564, p. 10-23. 1955. On morphometric and serological relationships within the Limulidae. Ph.D Thesis, New York Univer- sity, N.Y., 287 p. 811 FISHERY BULLETIN: VOL. 85, NO. 4 1979. Distribution of the American horseshoe "crab," Limulus polyphemus (L.). In E. Cohen (editor). Biomed- ical applications of the horseshoe crab (Limulidae), p. 3- 26. Liss, N.Y. 1982. A pictorial review of the natural history and ecology of the horseshoe crab Limulus polyphemus, with refer- ence to other Limulidae. In J. Bonaventura, C. Bonaventura, and S. Tesh (editors). Physiology and biol- ogy of horseshoe crabs, p. 1-52. Liss. N.Y. Shuster, C. N.. Jr., and M. L. Botton 1985. A contribution to the population biology of horse- shoe crabs, Limulus polyphemus, in Delaware Bay. Estuaries 8:363-372. Smith, O. R., and E Chin. 1951. The effects of predation on soft clams, Mya are- naria. Natl. Shellfish Assoc. Conv. Address, p. 37-44. Smith. O R.. J P Baptist, and E Chin 1955. Experimental farming of the soft-shell clam, Mya arenaria, in Massachusetts, 1949-1953. Comm. Fish. Rev. 17(6):1-16. SOKOLOFF, A. 1978. Observations on populations of the horseshoe crab Limulus iXiphosura) polyphemus. Res. Popul. Ecol. 19:222-236. Waterman, T H. 1950. A light polarization analyzer in the compound eye of Limulus. Science 111:252-254. Wells, S M , R M Pyle, and N. M Collins 1983. The invertebrate red data book. lUCN, Gland (Switzerland). Wolff, T 1977. The horseshoe crab {Limulus polyphemus ) in North European waters. Vidensk. Meddr. Dansk. Naturh. Foren. 140:39-52. WOODIN. S. A 1978. Refuges, disturbance and community structure: a marine soft-bottom example. Ecology 59:274-284. 1981. Disturbance and community structure in a shallow water sand flat. Ecology 62:1052-1066. Appendix table 1 . — Abundance and frequency of occurrence of horseshoe crabs on the middle Atlantic continental shelf. Based on National Marine Fisheries Service trawl surveys from all seasons, 1975 to 1983. No. No. Latitude/ stations stations No. % occur- longitude sampled with crabs crabs rence x/trawl 4rN, 72°W 2 0 0 0.0 0.00 AVN, 71 °W 187 0 0 0.0 0.00 41 °N, 70°W 181 3 10 1.6 0.06 40°N, 74'W 13 4 9 30.8 0.69 40°N, 73°W 396 122 355 30.8 0.90 40°N, 72°W 399 39 72 9.8 0.18 40°N, 71 °W 318 0 0 0.0 0.00 40°N, 70°W 303 0 0 0.0 0.00 39°N, 74°W 269 134 999 49.8 3.71 39°N, 73°W 250 5 9 2.0 0.04 39°N, 72 'W 255 0 0 0.0 0.00 39°N, 71°W 61 0 0 0.0 0.00 39"N, 70°W 23 0 0 0.0 0.00 38°N, 75°W 92 74 1,208 80.4 13.13 38°N, 74°W 390 167 1,986 42.8 5.09 38"N, 73"W 233 1 1 0.4 0.04 38°N, 72°W 34 0 0 0.0 0.00 37°N, 75°W 323 174 1,085 53.9 3.36 37°N, 74°W 264 16 21 6.1 0.08 37°N, 73°W 7 0 0 0.0 0.00 36°N, 75°W 297 80 456 26.9 1.54 36°N, 74°W 183 17 54 9.3 0.29 35 N, 76"W 2 1 1 50.0 0.50 35°N, 75'W 247 77 413 31.2 1.67 35°N, 74°W 78 31 234 39.7 3.00 34°N, 77°W 102 4 4 3.9 0.04 34''N, 76°W 218 30 114 13.8 0.52 34 N, 75°W 75 3 3 4.0 0.04 33°N, 77°W 144 1 1 0.7 0.01 33°N, 76°W 58 0 0 0.0 0.00 812 NOTES ANALYSIS OF SEA TURTLE CAPTURES AND MORTALITIES DURING COMMERCIAL SHRIMP TRAWLING Five species of sea turtles occur in coastal United States waters of the southern North Atlantic and the Gulf of Mexico and are listed and protected under the Endangered Species Act (1973). These are the Kemp's ridley turtle, Lepidochelys kempi; hawksbill turtle, Eretmochelys imbricata; leatherback turtle, Dermochelys coriacea; green turtle, Chelonia mydas; and loggerhead turtle, Caretto caretta. Each of these species are cap- tured by commercial shrimp trawlers, and these incidental captures have been identified as a source of sea turtle mortalities (Hopkins and Richardson 1984). Several prior studies have attempted to quan- tify turtle catch rates and mortalities by trawlers through interviews with vessel captains (Anony- mous 1976,1 19772; Cox and Mauerman 1976; Ra- balais and Rabalais 1980) and through direct ob- servations by observers during commercial shrimp trawling (Hillestad et al. 1978; Ulrich 19783; Roithmayr and Henwood 1982^). While these studies provide estimates of capture and mortality rates, more specific information is re- quired to effectively protect the stocks. In particu- lar, managers need to know when and where tur- tle captures occur, which species are impacted, at what depths the majority of captures occur, and how many turtles are captured and killed. This report provides a preliminary analysis of existing data collected by fisheries observers dur- ing commercial U.S. shrimp trawling. Data from three National Marine Fisheries Service (NMFS) observer projects were used for analysis of turtle catch per unit effort (CPUE) and mortality rates. A brief description of the projects follow: •Anonymous. 1976. Incidental capture of sea turtles by shrimp fishermen in Florida. Prelimmary report of the Flor- ida West Coast Survey, University of Florida Marine Advisory Program, 3 p. ^Anonymous. 1977. Alabama shrimp fishermen inter- views for 1977-1978. Marine Resources Office, Alabama Coop- erative Extension Service, 1 p. 3Ulnch, G. F. 1978. Incidental catch of loggerhead turtles by South Carolina commercial fisheries. Report of the Na- tional Marine Fisheries Service, Contract No. 03-7-042-35151, 33 p. 4Roithmayr, C, and T. Henwood. 1982. Incidental catch and mortality report. Final report to Southeast Fisheries Cen- ter. National Marine Fisheries Service, NOAA, 75 Virginia Beach Drive, Miami, FL 33149, 20 p. FISHERY BULLETIN: VOL. 85, NO. 4, 1987. 1) The sea turtle incidental catch and mortality project was instituted to provide information on the incidental capture and associated mor- tality of sea turtles off the southeastern United States. Trained fishery observers were placed aboard commercial shrimp vessels op- erating on the major grounds in the Gulf of Mexico and southern North Atlantic from 1979 through 1981. 2) The goal of the excluder trawl project was to design an apparatus for use with existing shrimping gear which would effectively pre- vent the incidental capture of sea turtles. Ini- tial design and testing of prototype models were conducted during 1977, and field trials were continued through 1984. Fishery observ- ers aboard cooperative and chartered shrimp trawlers began data collection in 1978. Data collection procedures were similar to those of the incidental catch project except that data records were maintained for each net. In this manner, the performance of excluder nets could be compared with that of standard trawls. 3) The objectives of the shrimp fleet discards proj- ect were to estimate the magnitude and spe- cies composition of incidental fish captures by the Gulf shrimp fleet. Data were collected through contractual arrangements with state agencies from 1973 through 1978. These agen- cies placed observers on commercial vessels to obtain at-sea sampling off their respective coasts. Data records similar to those of the other two projects were completed for each tow. In estimating turtle CPUE and mortalities by species, we restricted our analyses to loggerhead, Kemp's ridley, and green turtles. Leatherback and hawksbill turtles were also captured in shrimp trawls, but the infrequency of captures made predictions of CPUE for these species im- precise. In predictions of CPUE for all species combined, these capture records were included. Data Analyses For estimations of turtle CPUE and mortali- ties, the three observer projects were merged. For each data set, effort (E) was standardized to re- 813 fleet hours towed with a single, 30.5 m headrope length net using the formula E = (nets * length = 30.5 m) * (min = 60) where nets = number of nets towed, length = headrope length of a net (meters), min = minutes fished. Turtle CPUE (R ) and 95% confidence interval (C.I.) were calculated according to methods de- scribed in Snedecor and Cochran (1967) using the formulae R = ^tJ^e, 1 = 1 i = l 95% C.I. on/? =/? + 1.96 (l/S) RE,)'^ln(n-l) where R = CPUE (turtles/30.5 m net hour), R = estimated CPUE, T, = number of individuals (turtles), E, = effort (30.5 m net hour), n = sample size (number of tows). The data were stratified by species, season, depth, and statistical zone (corresponding to those used by NMFS for reporting shrimp landings). For each zone, turtle CPUE, mean depth of cap- ture, mean length of tow, and mortality were computed. In summarizing the data, the Gulf of Mexico was subdivided into eastern (NMFS statistical zones 1-7, corresponding to the Flori- da west coast excluding the panhandle), central (NMFS statistical zones 8-17, corresponding to the Florida panhandle through Louisiana), and western (NMFS statistical zones 18-21, corre- sponding to the Texas coast) areas. The southern North Atlantic area included the east coast of the United States from Florida to North Carolina, statistical zones 24-33. Part of zone 28, the Cape Canaveral ship channel and adjacent shrimping grounds (lat. 28°15'N to 28°30'N) was excluded to avoid positively biasing CPUE estimates. This habitat harbors large concentrations of turtles throughout the year, and high turtle catch rates (0.3643 ± 0.0045 turtles/hour)^ do not reflect those occurring on the shrimping grounds outside the Canaveral area. Exclusion of these data is not expected to cause an underestimate of mortalities for the southern North Atlantic because commer- cial shrimping effort near Cape Canaveral is re- stricted to three or four vessels during most of the year. Estimates of shrimp fishing effort for the off- shore Gulf of Mexico shrinip fishery were ob- tained from the NMFS Galveston Laboratory (E. Klima^). The shrimp fishing effort was corrected for relative amounts of effort by single rigged, double rigged, and quad rigged vessels and then standardized to 30.5 m net hours. The Atlantic shrimp fishing effort was based on an effort esti- mate developed in 1983 (Anonymous 1983^). Be- cause the data were being updated, more current Atlantic shrimp fishing effort data will be avail- able at a later time. Percent mortality of the total catch was esti- mated by a least squares linear regression using percent mortality as dependent upon minutes fished which yielded the relationship of Y = 0.00165X - 0.03. The average mortality over 30-min increments of tow length was calcu- lated, and 10 unweighted means were regressed on minutes fished. Although this approach may violate the assumption of homogeneity in regres- sion, it was assumed to be the most appropriate means of describing this relationship, since the dependence of mortality on tow time is strongly statistically significant (r = 0.98; P< 0.001). Percent mortality was multiplied by turtle cap- tures ±95% upper and lower confidence bounds of turtle captures to estimate the number of turtles killed. Results and Discussion Turtle captures and mortality by statistical zone and season with associated trawling effort data were analyzed. While the total observer ef- fort in the Gulf of Mexico (16,771 hours) was greater than the southern North Atlantic (9,943 hours), 482 turtles were captured in the southern North Atlantic and only 52 were captured in the '■^Means ± the 95% confidence interval will be used through- out the paper. 6E. Klima, Southeast Fisheries Center Galveston Laboratory, National Marine Fisheries Service, NOAA, 4700 Avenue U, Galveston, TX 77550, pers. commun. Summer 1986. '^Anonymous. 1983. Environmental assessment of a pro- gram to reduce the incidental take of sea turtles by the commer- cial shrimp fishery in the southeast United States. U.S. De- partment of Commerce, National Marine Fisheries Service, 9450 Koger Blvd., St. Petersburg, FL 33702. 814 Gulf of Mexico (Table 1). This indicates that per unit effort, 16 turtles were captured in the At- lantic for every one turtle captured in the Gulf. An attempt was made to compare mean depth and duration of tow for turtle captures with the mean depth and duration of tow for all effort by area with and without turtle captures. The mean depth of fishing and mean length of tow were computed from effort data for each statistical zone and for tows in which loggerhead, Kemp's ridley, or green turtles were captured. In most cases (particularly the Gulf of Mexico) sample sizes were small, and no patterns or consistency were evident. We suggest that despite some apparent statistical differences which we attribute to small sample sizes, average depth and tow duration of turtle captures were probably not different from that of the effort. Summary information on observer effort, CPUE, shrimping effort, estimated captures, and estimated mortality in the Gulf of Mexico and southern North Atlantic are presented for logger- head, Kemp's ridley, and green turtles (Table 1). Estimated CPUE for all turtles in the Gulf of Mexico (zones 1-21) was 0.0031 ± 0.0008 turtles/ net hour, and CPUE for the southern North At- lantic (zones 24-33) was 0.0487 ± 0.0041 turtles/ net hour. The calculation of estimated mortality used minutes fished as a means of estimating the per- cent of the turtles captured that are killed. Based on mean tow times from our effort data, the over- all mortality rate for the Gulf of Mexico is 29%. The eastern Gulf mortality rate is 34%, the cen- tral Gulf rate is 22%, and the western Gulf rate is 38%. For the Atlantic coast, the rate is 21% re- flecting the shorter average duration of trawl tows on this coast. The mortality rates based on minutes fished do not distinguish among species. This is because of the small numbers of captures for species other than loggerhead turtles. If there are differences in the ability of the other turtle species to survive long periods of immersion and the stress involved in being captured in a trawl, the differences are not measurable from these data. In using minutes fished to estimate mortality, the data did not conform to expected models over the range of our observations. In tows of <60-min duration, mortality rates were <1% suggesting that the logistic model might be most appropriate to describe the relationship. However, of logistic, 2d and 3d order polynomial and linear models, the best fit over the range of tow times observed in these studies was provided by the linear model. In tows of <60-min duration and in tows longer than 360 minutes, the linear model is probably inappropriate; mortality is negligible in very Table 1 . — Observer effort, turtle captures, CPUE, shrimping effort, estimated captures and estimated mortality of loggerhead, Kemp's ridley, and green turtles in the Gulf of Mexico and the southern North Atlantic. NMFS Annual observer Number CPUE + 95% shrimping Estimated Estimated effort (net of C.I. on CPUE effort (net captures mortality Area hours) turtles (turtles/net hour) hours) 1 (turtles/yr) (turtles/yr) Loggerhead turtles, ( Oaretta caretta Atlantic 9,943 453 0.0456 ± 0.0039 704,376 32,120 ±2,747 6,745 ± 577 Gulf of Mexico eastern 2,589 12 0.0046 ± 0.0026 611,530 2,813 ± 1,590 956 ± 541 central 6,353 14 0.0022 ±0.0012 2,391,498 5,261 ± 2,870 1,157 ±631 western 7,829 16 0.0020 ±0.0010 1,312,670 2,625 ± 1,313 998 ± 499 overall 16,771 42 0.0025 ± 0.0008 4,315,698 10,789 ±3,453 3,129 ± 1,001 Kemp's ndley turtles. , Lepidochelys kempi Atlantic 9,943 18 0.0018 + 0.0008 704,376 1 ,268 ± 564 266 ± 119 Gulf of Mexico eastern 2,589 0 0 611,530 2245 ± 245 83 ±83 central 6,353 2 0.0003 ± 0.0004 2,391,498 717 ±957 158 ±210 western 7,829 4 0.0005 ± 0.0005 1,312,670 656 ± 656 249 ± 249 overall 16.771 6 0.0004 ± 0.0004 4,315,698 1,726 ±1,726 501 ± 501 Green turtle, Chelonia mydas Atlantic 9.943 7 0.0007 ± 0.0003 704,376 493 ± 21 1 104 ±44 Gulf of Mexico eastern 2,589 0 0 611,530 261 ± 122 21 ±41 central 6,353 2 0.0003 ± 0.0003 2,391,498 717 ± 717 158 ± 158 western 7,829 0 0 1,312,670 2131 ±262 50 ± 100 overall 16,771 2 0.0001 ± 0.0002 4,315,698 432 ± 863 125 ±250 'Gulf of Mexcio effort estimates provided by NMFS. Galveston Laboratories (E. Kiima text footnote 5) and soutfiern North Atlantic effort based on estimates from Anonymous 1983 2Based on CPUE for ttie overall Gulf of Mexico. 815 short tows and never reaches 100% because tur- tles may be captured at any time during the tow and will survive if captured in the latter stages. Tows shorter than 1 hour and longer than 6 hours, however, are relatively uncommon in com- mercial shrimping operations. In the southern North Atlantic, the CPUE for all turtles was strongly dependent on depth (Fig. 1). In depths >10 fathoms, turtle captures were rare, even though, based on aerial surveys (Fritts et al. 1983), turtles are distributed well offshore in waters considerably deeper than 10 fathoms. The strong depth dependency of CPUE may re- flect the fact that the continental shelf is rela- tively narrow along the southeastern seaboard, and the fact that most shrimping occurs in waters <10 fathoms. In the Gulf of Mexico, CPUE ap- peared to be relatively constant over all depths (Fig. 1). These estimates are conservative because only offshore (outside the barrier islands) effort and turtle captures were considered. It should be emphasized that trawl related tur- tle mortalities are not confined to U.S. waters, but occur on a worldwide basis. The same turtle populations impacted in U.S. waters are also im- pacted in territorial waters of other countries. In the case of the Kemp's ridley which is believed to be equally distributed in United States and Mex- ican waters, Mexican trawlers may account for mortalities similar to those of U.S. trawlers. To effectively protect sea turtles, international coop- eration is essential. Acknowledgments We thank all individuals who participated in the collection of data aboard commercial vessels UJ Q. o 0.16 -■ 0.15 - 0.14 - 0.13 - 0.12 - 0.11 - 0.10 - 0.09 - 0.08 - 0.07- 0.06 - 0.05 - 0.04 - 0.03 - 0.02 - 0.01 - 0.0 - a SOUTHERN NORTH ATLANTIC D GULF OF MEXICO n n ^ JZL 6 7 8 9 10 DEPTH IN FATHOMS 11 12 13 14 15 15 + Figure 1. — Catch per unit effort (turtles/net hour) as a function of depth for captures in the southern North Atlantic and the Gulf of Mexico. Conclusions From our analyses, it is evident that significant numbers of sea turtles are captured by commer- cial trawlers in both the Gulf of Mexico and the southern North Atlantic, and that over 20% of these turtles are drowned in the trawl. We esti- mate that 9,874 loggerhead, 767 Kemp's ridley, and 229 green turtles may be killed annually. and those persons who managed each of the projects. In particular, we appreciate the contri- butions of Frederick Berry, Andrew Kemmerer, Walter Nelson, Wilber Seidel, John Watson, Charles McVea, Charles Roithmayr, and Butch Pellegrin. Rick Minkler and Mark McDuff pro- vided computer programming support, Velda Harris tj^jed the manuscript, and Arvind Shah provided statisical expertise. 816 Literature Cited Cox. B A . AND R G Mauerman, 1976. Incidental catch and disposition by the Brownsville- Port Isabel Gulf shrimp fleet. Cameron Co. Ext. Ser., San Benito, TX. and Texas Shrimp Assoc, Brownsville TX, 55 p. Fritts.T H . a B Irvine. R D Jennings, L A Collum. W Hoff- man. AND M A McGehee. 1983. Turtles, birds and mammals in the northern Gulf of Mexico and nearby Atlantic waters. U.S. Fish Wildl. Serv. Div. Biol. Ser., Wash., D.C., 455 p. HiLLESTAD, H O , J I Richardson, and G K Williamson. 1978. Incidental capture of sea turtles by shrimp trawler- men m Georgia. Proc. Ann. Conf Southeast. Assoc. Fish Wildl. Agencies, p. 167-178. Hopkins, S., and J Richardson. 1984. A recovery plan for marine turtles. U.S. Gov. Print. Off., 281 p. Rabalais. S C, and N. N Rabalais. 1980. The occurrence of sea turtles on the south Texas coast. Contrib. Mar. Sci. 23:123-129. Snedecor. G W . AND W G Cochran. 1967. Stastical methods. [6th ed.] Iowa State Univ. Press, Ames, lA, p. 536-537. Tyrrell A Henwood Warren E Stuntz National Marine Fisheries Service, NOAA Southeast Fisheries Center Mississippi Laboratories Pascagoula Facility Pascagoula, MS 39568-1207 THE RELATIONSHIP BETWEEN LUNAR PHASE AND GULF BUTTERFISH, PEPRILUS BURT/, CATCH RATE Through the joint efforts of Japan and the United States, a research program was conducted in fall 1984 and spring 1985 to identify squid resources in the northern Gulf of Mexico (Grace 1984, 1985). Although large concentrations of squid were not located, commercial quantities of gulf butterfish, Peprilus burti, were encountered. Maximum sustainable yield (MSY) estimates from the spring data indicated annual potential catches of 50,000 t with a projected ex-vessel value of $19 million (GledhilP). Although gulf butterfish are sufficiently abundant to support a fishery, critical gaps of information on gulf but- iGledhill, C. T. 1985. A preliminary estimate of gulf but- terfish [Peprilus burti) MSY and economic yield. Unpubl. manuscr., 66 p. Southeast Fisheries Center, Mississippi Labo- ratories, National Marine Fisheries Service, NOAA, Pascagoula, MS 39568-1207. terfish distribution and location exist which are needed in order to harvest this resource effi- ciently. Preliminary data from the U.S. -Japan joint surveys indicated that gulf butterfish catch rates were greatest at bottom temperatures of 15°-19°C. Subsequent scientific and commercial efforts at targeting gulf butterfish based upon bottom temperature have produced catches rang- ing from few individuals to many tons. In a recent study, we found that fishing success for gulf but- terfish was often high for several days followed by periods of low success (Allen et al. 1986). This phenomena parallels catch patterns encountered by east coast gulf butterfish fishermen (Amos^), who suggest that lunar phase affects catch rates. We analyzed the effect of lunar phase on catch rates. The purpose of this paper is to present evi- dence that bottom trawling success for gulf but- terfish is related to lunar phase. Methods Gulf butterfish catches from the two U.S.- Japanese joint surveys and from an additional gulf butterfish survey conducted by SEAMAP (August 1985) were examined. Initially, catch rates per hour of individual trawls were calcu- lated per calendar day. A lunar day value (1-29) was assigned to each calendar day of trawling during the three cruises. Lunar day 1 was as- signed to the third calendar day proceeding the new moon on through day 29 falling on the third calendar day following the last quarter moon phase. Mean catch (kg/hour per lunar day) was then calculated and plotted. Catches from trawled stations outside of the depth range in which gulf butterfish were caught during each trip (i.e., < minimum depth or > maximum depth) were not included when calculating mean catch/hour per lunar day. The effects of moon phase and trip on natural log catch rates (ln(a: + 1), where x = kg/hour per individual trawl) of gulf butterfish were investi- gated, using the general linear model (GLM) pro- cedures (SAS) Institute (1982). Type III sums of squares were used for the analysis due to unequal number of observations in each subclass. Each observation from each trip was assigned into a lunar phase period (1-4). Mean catch (ln(x + 1)/ hour) and number of trawls sampled during each trip and lunar phase are presented in Table 1. An ^Duncan Amos, Georgia Marine Extension Program, P.O. Box Z, Brunswick, GA 31523, pers. commun. July 1986. FISHERY BULLETIN: VOL 85, NO 4. 1987. 817 Table 1. — Mean catch (ln(x + 1)/hour) of gulf butterfish and number of trawls sampled during each trip and lunar phase. Trip Phase Number Mean catch 1 1 1 2 1 3 1 4 2 1 2 2 2 3 2 4 3 1 3 2 3 3 3 4 6 1.13 24 2.57 13 1.58 31 2.31 24 2.07 47 3.24 9 1.29 21 2.16 39 0.49 35 0.58 21 0.26 62 0.40 analysis of variance (ANOVA) model was devel- oped to test for the effect of trip, lunar phase, and the interaction between trip and lunar phase. Scheffe's test was used to contrast each lunar phase with the other three phases. Results Peak catch rate was observed to occur in the first quarter moon phase following the new moon (Fig. 1). There was a highly significant difference 800 700 600 among trips and lunar phases (Model 1, Table 2). The interaction between trip and lunar phase was not significant iP = 0.33) and was therefore dropped from the model resulting in Model 2 (Table 3). In model 2, there was a highly signifi- cant difference among trips and a significant dif- ference among lunar phases. A comparison of means using Scheffe's test for each moon phase (Table 4) indicated that catch rate during the first quarter moon phase was significantly greater than catch rates during the last quarter, new, and full moon phases. Table 2. — Analysis of variance table testing the effects of moon phase (M), research trip (T), and the interaction between trip and moon phase (T * M) on gulf butterfish catch rates during three butter- fish research surveys in the Gulf of Mexico 1984- 85. O) I o < o z < UJ 500 400 300 200 100 O O LUNAR DAY C Source df MODEL 1 ss F-ratio Pr>F Model Error 11 320 384.864 1,130.662 9.90 O.ooor* Variable df TYPE III ss F-ratio Pr>F T M T*M 2 3 6 175.758 43.110 24.353 24.87 4.07 1.15 0.0001" 0.0074** 0.3338 ■"Significant effect at P < 0.01 . Table 3. — Analysis of vanance table testing the effects of moon phase (M) and research trip (T) on gulf butterfish catch rates during three butterfish research surveys in the Gulf of Mexico 1984-85. Source df MODEL 2 ss F-ratIo Pr>F Model Error 5 326 360.511 1,155016 20.35 0.0001" Variable df TYPE III ss F-ratio Pr>F T M 2 3 272.183 37.648 38.41 3.54 O.ooor- 0.0149* ■"Significant effect at P < 0.01. "Significant effect at P < 0.05. Table 4. — Mean catch rate per hour (ln(x + 1)/hour) by lunar phase Lunar phase Catch rate/hour Figure 1. — Graph of mean catch (kg/hour I of gulf butterfish by lunar day. New Moon 1.09 First Quarter 2.21* Full Moon 088 Last Quarter 1.24 *P < 0.05. 818 Discussion Although lunar rhythmicity in marine organ- isms, particularly marine invertebrates, has long been recognized (Palmer 1974), lunar rhythms in which a single peak of activity occurs each month in fishes appear to be rare (Gibson 1978). Most accounts of variations in catch rate of commer- cially important species which correlate with moon phase refer to clupeids (Gibson 1978). Blax- ter and Holliday (1963) suggested several possi- ble explanations for the apparent lunar rhythmic- ity of clupeid catches including: 1) intensity of moonlight, 2) effect of tides, and 3) fishermen be- havior. Gulf butterfish are normally trawled during daylight when they concentrate near bottom fol- lowing nocturnal vertical migration. However, this migration is difficult to describe because con- ventional echo sounding equipment poorly tracks gulf butterfish movement owing to atrophy of the swim bladder in gulf butterfish over 100 mm standard length (Horn 1970). Differences in catch rates between lunar phases may be attributed to changing vertical movements of gulf butterfish in the water column. The lunar pattern is probably not due to onshore-offshore movement out of the fishery's area of operation. In the three research cruises, sampling was stratified by bottom depth (36-585 m) and data do not suggest horizontal movements of gulf butterfish outside these depths. In conclusion, further work on lunar rhythmic- ity relationships of gulf butterfish is needed. Re- sults may greatly enhance commercial and scien- tific efforts in harvesting and surveying gulf butterfish, respectively, by identifying alternate fishing methods (e.g., midwater trawling) that successfully target gulf butterfish during all moon phases. Literature Cited Allen. R L., J H Render. A W Liebzeit, and G W Bane 1986. Biology, ecology, and economics of butterfish and squid in the northern Gulf of Mexico. La. State Univ., Coastal Fish. Inst. Publ. 86-30, 169 p. Blaxter. J. H S , AND F. G T Holliday 1963. The behavior and physiology of herring and other clupeids. In F. S. Russell (editor). Advances in marine biology, Vol. 1, p. 262-393. Acad. Press, Lond. Gibson, R N. 1978. Lunar and tidal rhythms in fish. In J. E. Thorpe (editor!. Rhythmic activity of fishes, p. 201-213. Acad. Press, Lond. Grace. M. 1984. U.S.-Japan squid survey, 10/11-11/17/84, Nisshin Maru No. 201. U.S. Dep. Commer. Natl. Oceanic At- mos. Adm., Natl. Mar. Fish. Serv., Cruise Rep., 18 p. 1985. U.S.-Japan squid survey, 4/18-5/29/85, Nisshin Maru No. 201. U.S. Dep. Commer., Natl. Oceanic At- mos. Adm., Natl. Mar. Fish. Serv., Cruise Rep., 18 p. Horn, M. H. 1970. Systematics and biology of the stromateid fishes of the Genus Peprilus. Bull. Mus. Comp. Zool. 140:164- 271. Palmer, J. D. 1974. Biological clocks in marine organisms. Wiley, Lond., 173 p. SAS Institute, Inc. 1982. SAS user's guide: Statistics. 1982 ed. SAS Institute, Gary NC, 584 p. Jeffrey H. Render Robert L. Allen Coastal Fisheries Institute Louisiana State University Baton Rouge, LA 70803-7503 MOVEMENTS OF COHO, ONCORHYNCHUS FJSUTCH, AND CHINOOK, O. TSHAWYTSCHA, SALMON TAGGED AT SEA OFF OREGON, WASHINGTON, AND VANCOUVER ISLAND DURING THE SUMMERS 1982-85 Knowledge of the migration patterns of salmonids in the ocean is an important consider- ation in developing fishery management plans. Catches of coded- wire tagged salmon in the ocean have yielded much information on general distri- bution patterns of different stocks and species of salmon (see for example Hunter [1985], Garrison [1985], and Howell et al. [1985]). Other studies have dealt with movements of salmon tagged in offshore waters of the northern North Pacific Ocean (Hartt 1962, 1966; French et al. 1975; God- frey 1965; Godfrey et al. 1975) and in coastal waters of British Columbia, Washinton, Oregon, and California (Milne 1957; Vernon et al. 1964; Kauffman 1951; Van Hyning 1951; Fry and Hughes 1951). Movements of juvenile salmon in coastal waters of the Gulf of Alaska were studied by Hartt and Dell (1986); in Georgia Strait, British Columbia, by Healey (1980); and in coastal waters off Oregon and Washington by Pearcy and Fisher (unpubl. manuscr.)^ ^W. C. Pearcy and J. P. Fisher. Migration of coho salmon (Oncorhynchus kisutch ) during their first summer in the oceans. Unpubl. manuscr. College of Oceanography Oregon State University, Corvallis, OR 97331. FISHERY BULLETIN: VOL. 85, NO. 4, 1987. 819 Movements of individual maturing salmon off Oregon and Washington are still poorly known. In this paper we examine migration after tagging of salmonids collected during purse seine cruises off the Oregon and Washington coasts from 1982 to 1985 and off the west coast of Vancouver Is- land, B.C., in 1984. Methods Maturing and juvenile salmon were collected by purse seine during May 1982, 1983, and 1985; June 1982-85; July 1984; and September 1982-84. Coho salmon, Oncorhynchus kisutch, were classi- fied as maturing or juvenile, based on the length- frequency distribution of the catch in each month. The distribution was usually bimodal and the di- vision between juvenile and maturing coho salmon was about 300 mm FL in May and June, 360 mm FL in July, and 420 mm FL in August and September. Chinook salmon, O. tshawytscha, <400 mm in all months were arbitrarily classi- fied as juveniles. Numbered orange Floy^ tags were applied with a Dennison Mark II tagging gun between the pterygiophores just below the dorsal fin of fish anesthetized with MS-222. Fish were allowed to recover for a few minutes in tanks with circulat- ing saltwater and then were released into the ocean. Date and location of release was recorded for each tagged fish. Condition of the fish after handling varied, but most swam vigorously in the recovery tank and rapidly swam away when re- leased. However, some scale loss almost always occurred and for some individuals was extensive. Information on movements of coho and chinook salmon was obtained from subsequent recoveries in ocean and terminal fisheries and on spawning grounds or at hatcheries. No reward was offered 2Reference to trade names does not imply endorsement by the National Marine Fisheries Service, NOAA. Table 1. — Summary of mark and recovery data for coho and chinook salmon tagged in the ocean off Oregon, Washington, and the west coast of Vancouver Island. featuring coho Juvenile coho Matunng Chinook Juvenile Chinook No. Year tagged No. rec. (%) No. tagged No. rec. (%) No. tagged No. rec. (%) No. tagged No. rec. (%) 1982 194 21 (10.8) 0 — 73 3 (4.1) 0 — 1983 142 17 (12.0) 0 — 5 1 (20.0) 0 — 1984 162 1985 215 10 (6.2) 13 (6.0) 86 18 3 (3.5) 0 (0) 56 37 4 (7.1) 8 (21.6) 27 46 0 (0) 1 (2.1) All years 713 61 (8.6) 104 3 (2.9) 171 16 (9.4) 73 1 (1.4) Maturing coho Maturing Chinook Other studies: No. tagged No. rec. (%) No. tagged No. rec. (%) Fry and Hughes (1951) Tagging off California in 1939-42, 1948, 1949 954 26 (2.7) 6,144 483 (7.9) Boydston (unpubl.) Tagging off California in 1971 and 1972 3,341 409 (12.2) Van Hyning (1951) Tagging off Oregon in 1948 and 1949 506 29 (5.7) 221 11 (5.0) Kauffman (1951) Tagging off Washington W. coast Vancouver Is 1948 and 1949 and >. in 65 16 (24.6) 635 33 (5.2) Milne (1957) 5,458 476 (8.7) 7,194 970 (135) 820 for return of tags. The straight Hne distance be- tween release and recovery locations indicated the minimum distance travelled (called "net movement") for fish recovered in the ocean. A series of connected straight line tracks were used to estimate net movement of fish recovered in locations where a single line could not be used (e.g., recoveries in Puget Sound). Straight line distance travelled in the ocean was added to dis- tance travelled upstream to estimate net move- ment for fish recovered in river systems. Approx- imate latitudinal change was used to estimate net movement of fish for which an accurate recovery location was not known (e.g., "recovered off Coos Bay"). Net migration rate was estimated by divid- ing net movement by days between release and recovery. Results and Discussion Numbers of fish tagged and percentages recov- ered are summarized in Table 1 for coho and Chi- nook salmon released in different years. Recovery rates were similar for maturing coho (mean 8.6%, range 6-12%) and chinook (mean 9.4%, range 4- 22%) salmon. These are similar rates to those found for these two species in other studies (Table 1). Numbers of fish recovered in different areas from releases off Oregon, Washington, and Van- couver Island are given in Table 2. Simplified migration patterns are shown in Fig. 1. Recover- ies of coho salmon released off Oregon were mainly (81%) from the Columbia River and Ore- gon. Only 11% were recovered in the Strait of Juan de Fuca or Puget Sound. This distribution differs from Van Hyning's (1951) finding that 47% of coho salmon tagged between June and Au- gust from Cape Lookout, OR, to the Columbia River were recovered in Puget Sound. Recoveries of coho salmon released off Washington were more widely distributed and 46% were recovered from the Columbia River to Cape Flattery, 20% in Oregon, 23% in the Strait of Juan de Fuca or Puget Sound areas, and 11% in British Columbia. Estimated net migration of coho salmon be- tween release and recovery (including upstream migration for those recovered in freshwater) aver- aged 181 km and ranged from 7 to 657 km (Fig. 2A.) Most coho salmon were recovered within 150 days of release. The two fish recoverd after 330 and 380 days were released as juveniles and recovered the following year as adults. Net migration rates of the maturing coho salmon tagged in coastal waters were generally very low and ranged from 0.1 to 20.4 km/day with a mean rate of 3.6 km/day (Fig. 2B). Coho salmon recovered in the open ocean off Oregon, Washing- ton, or Vancouver Island (circles) had only slightly higher mean rates of movement than those recovered in the Strait of Juan de Fuca or Puget Sound areas (triangles) or those recovered Table 2. — Recovery areas of fagged coho and chiinook salmon released off Oregon, Washington, and the west coast of Vancouver Island. Recovery areas en 1 •a (0 c ^ CO > 08 m m o8 c CO CO CO CO c (0 o 3 o Q) fl o c c c o 8 c 3 li- > E 3 o CD 01 c o a> 05 CO B) c CO > eu CO J3 0 0 z o S" > ir § CO CO § c CO E 3 JO CO c CO 1= o c: ro 6 TO to CO (0 E 3 O c CO CD 5 CO CO o c CO 3 "o CO 3 O CO 0 0 CO m Q) 0 Release areas o O o O cS O O o O o O o O CO 3 Q. m Coho Off Oregon 11 7 3 2 2 1 Off Washington 6 1 7 8 1 2 3 5 1 1 Off west Vancouver Island 1 Chinook Off Oregon 1 2 2 1 1 Off Washington 3 3 2 1 Off west Vancouver Island 2 821 COHO CHINOOK >iu Sir. of Juan ' Vancouver Island cJq Fuca Q /^ t, ■1 k WASH. V Columbi a Ri ver 50' 46* 44c 42< 40* 38' 36' ^:. '^^ ^i5£r. of Jum Vancouver Island ^q Pijca Q r^ Flattery'- yA f; ;{^S Paget Sound !♦ . ^... WASH t ^^fci^ Columbia River 50* 48* 46' .Coos Bay 'tape Mendocino 44< 42* 40' — 38* 36^ Figure 1. — Migration patterns of tagged echo and chinook salmon released at sea. For fish recovered in the ocean ofF Oregon, Washington, or Vancouver Island, release latitude is indicated by the tail of the arrow and recovery latitude by the head of the arrow. For fish recovered in inland waters or river systems the head of the arrow points to the system in which the fish was recovered. Solid dots indicate fish released and recovered at the same latitude. Numbers offish are approximately proportional to thickness of arrows. Most releases and recoveries were within 50 km of the coast and the positions of arrows do not represent true distances from shore. in coastal bays or river systems, including the Columbia River (squares) (4.4, 3.5, and 2.7 km/ day, respectively, Fig. 2B). Similar low net move- ment rates were found for coho salmon in coastal waters by Van Hyning (1951) off Oregon (3.0 km/ day), by Kauffman (1951) off Washington and Vancouver Island (3.9 km/day), and by Milne (1957) also off Vancouver Island (9 km/day). In all studies the stresses related to capture and tagging may cause some mortality and weaken some surviving fish, affecting speed of migration. Hartt (1966) suggested that tagging retards mi- gration by at least 1 day. However, movements of fish immediately after release in sonic tagging experiments were often rapid (Madison et al. 1972; Stasko et al. 1973). 822 COHO CHINOOK E 4J c a E a > o a 700 600 500 400 300 200 100 O O o "O a ML 700 r 600 500 400 300 200 100 ^ 200 300 400 100 200 300 400 500 600 700 -o \ E a □ tr •P c a E a > o 4J a z 22 20 18 16 14 12 10 3 6 4 2 0 100 300 10 9 8 7 6 5 4 3 2 1 0 B Do -»- 100 200 300 400 500 600 700 Days Release to Recovery 20 18 16 - 14 12 10 8 6 - 4 - 2 - 0 o o .J°l □ 8 25 50 100 125 10 A 9 8 7 6 5 0 D 4 - 3 0 2 - 0 1 0 C 0 D 0 0 0 A ] 25 50 75 100 125 15 RqIqqsq Data (May 1 - Day 0) E r v^ 400 0 K 0 / 300 01 300 200 A 0 100 0 200 t- 0 A 8 0 u a 100 0 -100 y a 0 0 2 ° ■•'.s -100 -200 -300 •0 D 4-1 K o8 ^0*^ o 0 □ o *J -200 h 0 o 0 -400 U -500 _J 4J -300 L ^0 -600 a z -400 4 . . ° . 5 . -700 3 44 45 40 47 48 4B 50 ^-. 43 44 45 46 47 49 50 Release Latitude Figure 2. — A) net movement vs. days between release and recovery, B) net movement rate vs. days between release and recovery, C) net movement rate vs. release date, and D) net latitude change (+ = north and - = south) vs. release latitude for tagged echo and chinook salmon. Recoveries in inland waters (Strait of Juan de Fuca, Puget Sound, Georgia Strait, and associated river systems) are indicated by triangles, those in the op)en ocean from the west coast of Vancouver Island to northern California by circles, and those in coastal bays or river systems by squares. For fish released off Oregon or Washington and later recovered in Puget Sound, net latitudinal change (D) is given as the change between the release location and Cape Flattery. 823 In contrast to the low movement rates observed for coho salmon in coastal waters off Oregon, Washington, and Vancouver Island, net migra- tion rates reported for salmon tagged in offshore waters of the North Pacific were generally much higher. Godfrey et al. (1975) calculated an aver- age rate of 24 km/day for tagged coho salmon recovered in the Japanese high seas fishery and 30 km/day for coho recovered in coastal waters. Hartt (1962, 1966) estimated that the migration rates of sockeye salmon into Bristol Bay averaged 44-50 km/day, whereas those of maturing sockeye salmon caught on the high seas averaged about 32 km/day. Chum salmon had migration rates at sea of 23-50 km/day; pink salmon had average rates of 43 km/day for coastal returns and 50 km/ day for high seas returns. Rapid migrations in coastal waters of British Columbia and Washington were also found for pink salmon (Vernon et al. 1964; Stasko et al. 1973) and for sockeye salmon (Madison et al. 1972) and off the Kurile Islands for chum salmon (Ichihara et al. 1975). However, migration rates slowed greatly as fish neared their home river systems (Vernon et al. 1964; Groot et al. 1975). Because net migration rates of coho salmon in coastal waters off Oregon, Washington, and Van- couver Island are so much lower than movement rates found for other salmon stocks, these coho are probably spending less time migrating in a single direction compared to meandering and feeding. Similarly, Milne (1957) concluded that coho salmon in coastal waters of British Colum- bia probably meander during both feeding and spawning migrations. Slow, feeding movements off Oregon and Washington are also suggested by the long time period (3-4 months) during which individual stocks are available in the ocean fish- eries (Hunter 1985). The relatively fast net mi- gration rates observed for some coho salmon re- captured within 33 days of release (Fig. 2B) suggest that actual movement rates over short time periods may be quite high but that meander- ing courses over time produce low net migration rates. Higher migration rates for fish tagged in late summer, to be expected if movements were changing from predominantly feeding to homing, were not apparent (Fig. 2C). Roughly equal numbers of coho salmon were recovered to the north (27) and to the south (35) of release sites, although most (8 of 11) coho re- leased from lat. 45°N and south were recovered to the north (Fig. 2D). Van Hyning (1951) also found that most coho tagged south of 45°N travelled to the north after release; however, he found that most coho released off northern Oregon and the Columbia River (46°15'N) were recovered to the north as well. Fry et al. (1951) and L.B. Boydston (California Department of Fish and Game, un- publ. data) reported that most recoveries of ma- turing coho salmon tagged off northern California were to the north, off Oregon or Washington. Northward migration by most of the maturing coho salmon tagged at sea south of 45°N during their final summer in the ocean is consistent with the distributional patterns of coastal Oregon and early run Columbia River stocks in the ocean fisheries. Peak catches of coastal Oregon coho salmon stocks are off northern California in May and June and shift to waters off Coos Bay in July and August (compiled from Hunter 1985). Rela- tively high percentages (24-37%) of the ocean catch of coastal Oregon coho salmon stocks (all combined) are off northern California (Garrison 1985; Hunter 1985; Oregon Department of Fish and Wildlife 1982). Similarly, between 62 and 65% of early run Columbia River stocks are caught south of the Columbia River (Hunter 1985; Howell et al. 1985; Oregon Department of Fish and Wildlife 1982). Therefore, many fish from these two stock groups, which make up a substantial fraction of the coho catch off Califor- nia and Oregon, migrate south and then migrate north sometime later during the summer to re- turn to their natal systems. Southward migration into waters off northern California and southern Oregon may be advantageous to these coho salmon stocks because of the potentially high food production fueled by strong coastal upwelling during the summer in this area. Other stocks of coho salmon are caught to the north of where they entered the ocean during their final summer in the ocean. About 47% of the late run Columbia River coho are caught north of the Columbia River in Washington and British Columbia (Howell et al. 1985). Smaller, but sig- nificant percentages of other stock groups from the south (early Columbia River, private hatch- ery, and other coastal Oregon groups) are also caught as maturing adults north of their natal streams. Thus, these fish would eventually have to migrate to the south to return to their natal streams. Therefore, the subsequent southward movement of many of the maturing coho salmon we tagged north of 45°N (Fig. 2D) is not surpris- ing. The slow net migration rates, prolonged resi- dence in coastal waters, and mixed north and 824 south net movements suggest that maturing echo salmon in coastal waters of Oregon and Washing- ton, unlike stocks of salmon from the Gulf of Alaska and Bering Sea regions, are not highly migratory with precisely directed and timed movements. Many juvenile coho salmon off Ore- gon and Washington also reside in coastal waters and do not appear to undertake rapid or long mi- grations out of this region (Pearcy and Fisher fn. 1). Of the 7 recoveries of chinook salmon released off Oregon, 5 were off Oregon, in the Columbia River, or in coastal Oregon rivers; 1 was off north- ern California; and 1 was off Washington. Of the 9 recoveries of chinook salmon released off Wash- ington, 3 were in the Columbia River, 3 off Ore- gon, 2 off Washington, and 1 in British Columbia. Two chinook salmon tagged off the west coast of Vancouver Island were recovered in the Strait of Juan de Fuca (Fig. 1, Table 2). Estimated net migration of chinook salmon av- eraged 201 km and ranged from 0 to 685 km (Fig. 2A). Unlike coho salmon, which spend only 1 year in the ocean and which were mostly recovered within 150 days of release, almost half of the chi- nook salmon, which may spend several years in the ocean, were recovered after more than 200 days at liberty. Mean net migration rate of chi- nook salmon was 1.9 km/day (n = 16, Fig. 2B). As was found for coho salmon, net migration rates of chinook salmon were many times lower than rates found for salmon tagged in offshore waters of the North Pacific Ocean. Therefore, some chi- nook salmon also appear to undertake meander- ing feeding movements in coastal waters off Ore- gon, Washington, and Vancouver Island. There was no evidence for acceleration of migration rate late in summer (Fig. IC). Tagged maturing chinook salmon differed from coho salmon in that most moved to the south after release (Fig. 2D). Columbia River and many coastal Oregon stocks of chinook salmon are caught in the ocean fisheries predominantly to the north of Oregon, i.e., north of their natal sys- tems (Wahle et al. 1981; Garrison 1985). Some of the maturing chinook salmon that we tagged may have been moving slowly toward their natal sys- tems from the north. One chinook salmon was recovered 319 days after release over 656 km to the south, off northern California (Figs. 1, 2D). Other species of salmonids tagged off Oregon and Washington were recovered only in very low numbers. There were only 2 recoveries from 164 tagged pink and 36 tagged chum salmon. The greatest net movement was by a chum salmon tagged on 1 June 1985 off Seaside, OR, just south of the Columbia River and recovered on 8 August 1985 in Hecate Strait, B.C. (great circle distance 830 km). This fish was at liberty for 68 days and its minimum movement rate as 12.2 km/day. Acknowledgments We thank all those fishermen that returned tags to us and personnel of the Oregon Depart- ment of Fish and Wildlife, Washington Depart- ment of Fisheries, and Canadian Department of Fisheries and Oceans for their help in recovering tags. Tom Quinn, Bob Garrison, and an anony- mous reviewer provided helpful comments on the manuscript. This research was sponosored by Na- tional Marine Fisheries Service, NOAA (NA-85- ABH-00025) and Office of Sea Grant, Department of Commerce (Grant No. NA-81-AA-0-00086, R/ OPF-17). Literature Cited French, R R , R G. Bakkala, and D. F. Sutherland 1975. Ocean distributions of stocks of Pacific salmon, Oncorhynchus spp., and steelhead trout, Salmo gaird- nerii, as shown by tagging experiments. U.S. Dep. Com- mer., NOAA Tech. Rep. NMFS, SSRF-689, 89 p. Fry, D H , and E P Hughes 1951. The California salmon troll fishery. Pac. Mar. Fish. Comm. Bull, 2:8-42, Garrison, R L 1985. Stock assessment of anadromous salmonids. Oreg. Dep. Fish. Wildl,, Fish Div., Annu. Prog. Rep., Portland, OR, 28 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. N. Pac. Fish. Comm. Bull. 16, p. 1-39. 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. N. Pac. Fish. Comm. Bull. 31, 80 p. Groot, C , K Simpson, I. Todd, P D Murray, and G A Buxton. 1975. Movements of sockeye salmon {Oncorhynchus nerka ) in the Skeena River estuary as revealed by ultra- sonic tracking. J, Fish. Res. Board Can. 32:233-242. Healey.M C 1980. The ecology of juvenile salmon in Georgia Strait, British Columbia. In W. J. McNeil and D. C. Himsworth (editors), Saimonid ecosystems of the North Pacific, p. 203-229. Oregon Sate Univ. Press, Corvallis. Hartt, a C 1962, Movement of salmon in the North Pacific Ocean and Bering Sea as determined by tagging, 1956-1958. Int. N. Pac. Fish. Comm, Bull. 6, 157 p. 1966. Migrations of salmon in the North Pacific Ocean and Bering Sea as determined by seining and tagging, 825 1959-1960. Int. N. Pac. Fish. Comm. Bull. 19, 141 p. Hartt. a. C, and M B. Dell. 1986. Early oceanic migrations and growth of juvenile Pacific salmon and steelhead trout. Int. N. Pac. Fish. Comm. Bull. 46, 105 p. Howell, P., K. Jones, D. Scarnecchia, L. LaVorg, W. Kendra, AND D. ORTMANN. 1985. Stock assessment of Columbia River salmonids. Vol. 1: chinook coho, chum and sockeye salmon stock summaries. U.S. Dep. Energy, Bonneville Power Admin., Div. Fish Wildl. Portland, OR 558 p. Hunter, M. A. 1985. The 1976-1978 brood coho model. Wash. Dep. Fish. Prog. Rep. No. 222, 146 p. Ichihara, T., T. Yonemorl and H. Asal 1975. Swimming behavior of a chum salmon Oncorhynchus keta, on the southern migration off Eto- rofu Island, the southern Kurile Islands. Bull. Far Seas Fish. Res. Lab. 13:63-77. Kauffman, D E 1951. Research report on the Washington State offshore troll fishery. Pac. Mar. Fish. Comm. Bull. 2:77-91. MlLNE, D. J. 1957. Recent British Columbia spring and coho salmon tagging experiments, and a comparison with those con- ducted from 1925 to 1930. Fish. Res. Board Can. Bull. 113, 56 p. Madison, D M., R. M Horrall, A B Stasko, and A. D. Hasler. 1972. Migratory movement of adult sockeye salmon (Oncorhynchus nerka ) in coastal British Columbia as re- vealed by ultrasonic tracking. J. Fish. Res. Board Can. 29:1025-1033. Oregon Department of Fish and Wildlife 1982. Ocean catch distribution of OPI coho stocks in 1980 and 1981. Ocean Salmon Manage. Oreg. Dep. Fish Wildl., Newport, OR. Stasko, A. B., R. M. Horrall, A. D. Hasler, and D Stasko 1973. Coastal movements of mature Eraser River pink salmon (Oncorhynchus gorbuscha ) as revealed by ultra- sonic tracking. J. Fish. Res. Board Can. 30:1309-1316. Van Hyning, J. M. 1951. The ocean troll fishery off Oregon. Pac. Mar. Fish. Comm. Bull. 2:43-76. Vernon, E. H., A. S. Hourston, and G. A. Holland. 1964. The migration and exploitation of pink salmon runs in and adjacent to the Eraser River Convention Area in 1959. Int. Pac. Salmon Fish. Comm. 15, 196 p. Wahle, R J , E Chaney, and R E Pearson. 1981. Areal distribution of marked Columbia River Basin spring chinook salmon recovered in fisheries and at par- ent hatcheries. Mar. Fish. Rev. 43(121:1-9. J. P. Fisher W C Pearcy College of Oceanography Oregon State University Corvallis, OR 97331 DAILY GROWTH INCREMENTS IN OTOLITHS OF JUVENILE BLACK ROCKFISH, SEBASTES MELANOPS: AN EVALUATION OF AUTORADIOGRAPHY AS A NEW METHOD OF VALIDATION Investigations into the temporal periodicity of growth increment formation in otoliths of larval and juvenile fishes have produced conflicting ac- counts. Taubert and Coble (1977), Barkman (1978), Wild and Foreman (1980), and Campana and Neilson (1982), among others, have con- firmed daily increment formation in otoliths from various species of larval and juvenile fishes. There have been a few studies, however, in which increment counts were not representative of ac- tual age of the fish (Wild and Foreman 1980; Gef- fen 1982; Neilson and Geen 1982). Nondaily in- crement formation has been explained by the inclusion of subdaily rings in age estimates as well as by methodological errors in preparing and viewing the otoliths (Campana 1983a; Campana and Neilson 1985). Since size and age offish, food limitations, and environmental conditions have been suggested to affect increment formation, validation is necessary in each study where fish age is estimated. Several techniques have been used to validate daily growth increments in larval and juvenile fish otoliths. Fish of known age, raised from fer- tilization or birth under controlled laboratory conditions, provide the best material to determine frequency of increment formation (Taubert and Coble 1977; Barkman 1978; Tanaka et al. 1981; Miller and Storck 1982). For many species, how- ever, rearing the larvae from birth through the juvenile stage is difficult or impossible. An alter- nate method of age validation introduces a chem- ical mark onto those calcified structures which exhibit periodic growth zones, such as otoliths, scales, and spines. The antibiotic oxytetracycline hydrochloride (OTC) has been used most success- fully in this manner (Wild and Foreman 1980; Campana and Neilson 1982; Ralston and Miyamoto 1983; Dabrowski and Tsukamoto 1986). The OTC is taken up at the site of calcifica- tion and fluoresces bright yellow under ultravio- let light, compared with the blue autofluores- cence of normal tissue. Most recently, stable strontium has been used to demonstrate daily in- crement formation in squid statoliths (Hurley et al. 1985) and in mass marking of coho salmon (Yamada et al. 1979). For some species, a time- mark may also be induced on the otolith by stress, 826 FISHERY BULLETIN: VOL. 85, NO. 4, 1987. such as cold shock (Mugiya and Muramatsu 1982), or by simply bringing field-captured fish into the laboratory (Boehlert and Yoklavich 1985). Comparing increment counts with number of days following the time-mark accurately esti- mates frequency of occurrence of the growth in- crements. Our study evaluates the commonly used OTC and an alternate chemical, the ra- dioisotope calcium-45, in terms of their success as time-markers to validate daily growth increment formation in the otoliths of juvenile black rock- fish, Sebastes melanops. Materials and Methods Young-of-the-year black rockfish, Sebastes melanops, ranging from about 2 to 5 g wet body weight and 47 to 64 mm standard length (SL), were collected from a rocky, intertidal area 8 km south of Newport, OR, in July 1982 and from Yaquina Bay, OR, in July 1983. Fish were held in 200 L tanks under ambient water temperature conditions which fluctuated between 13° and 17°C; a ration of ground squid and shrimp was offered ad libitum and photoperiod was main- tained at 13 h light and 11 h darkness. After at least 10 days of acclimation to laboratory condi- tions, fish were anesthetized with MS-222 and injected intramuscularly (midbody below dorsal fin) with a solution of either OTC or calcium-45. Fish continued to feed immediately following in- jection and handling. Calcium-45 Fish were injected with a solution of low- calcium physiological saline and calcium-45 (^^CaCls dissolved in HCl; New England Nu- clear).^ Through preliminary experiments, a dose of 0.1 ixCi '*^Ca/g wet body weight proved to be optimum for isotope uptake and retention. Four fish each were sacrificed at 1, 4, 12, 24, 48, 72, 96, 120, 144, 168 hours and at subsequent 4-d inter- vals for 63 days following injection. Three fish were sacrificed after having maintained good health and growth for 1 year following injection. Four nonradioactive fish were sacrificed on the first day of the experiment and used as blanks or controls in determining activity levels of the in- jected fish. At the time intervals specified above, fish were anesthetized, blotted dry, measured (nearest mm, SL), and weighed (nearest 0.01 g). Both sagittal otoliths were removed from each fish, rinsed thor- oughly in water to remove surface contamination of calcium-45, and stored dry for liquid scintilla- tion counting (LSC) or autoradiography. One otolith from each fish, with the exception of the three 1-yr-olds, was weighed, dissolved in 0.1 mL concentrated HCl, diluted with 10 mL of Beck- man Ready-solv EP liquid scintillation cocktail, and assayed for calcium-45 activity in a Beckman LS 8000 liquid scintillation counter. Activity was corrected for decay and quench and expressed as disintegrations per minute (DPM) per mg of sam- ple. The perceived decrease in radioactivity due to the increase in weight of otolith over the experi- mental period was corrected using the following equation: Corrected activity = DPM 1 References tx) trade names do not imply endorsement by the National Marine Fisheries Service, NOAA. mg tissue^. weight tissue^, mean weight tissue^ where tf is time at sacrifice and ti is time of exper- iment initiation. Mean weight of otolith at i, was obtained from the 4 fish sacrificed prior to injec- tion; since all fish were of similar length at the onset of experimentation, these 4 fish adequately represented size of injected fish. Four otoliths from time interval 1 hour and two otoliths from each of intervals 4 and 12 hours and 1, 4, 11, 19, 39, 55, and 63 days were prepared for autoradiography. The right otolith of each pair was affixed to a microscope slide with histological mounting medium. The proximal surface of the otolith was ground with 600 grit carborundum paper on a rotating wheel until the focus was just visible and most of the curvature of the otolith was removed. The mounting medium was gently heated and the otolith was turned to expose the distal surface. Grinding was continued until most of the mounting medium was removed from the margins of the otolith. The external surface was polished using jeweler's rouge (3 ^JLm) and the whole slide was immersed in an ultrasonic cleaner to remove all loose particles from the otolith surface. The resulting sagittal section was coated with Kodak NTB3 nuclear emulsion and exposed in a light-tight box for 8 days at 4°C. The autoradiographs were developed in Kodak D-19 developer for 2 minutes, fixed for 5 minutes in 827 Kodak Fixer, rinsed for 20 minutes in distilled water, and viewed under transmitted light with a compound microscope at 400 x magnification to determine presence and location of exposed silver grains. Growth increments were enumerated from the time-mark to the otolith margin. In most of the otoliths, a check (or exceptionally dense band) was noted prior to the deposition of the radioactive mark. The location of this check, in terms of numbers of increments from the time- mark, was also determined. Oxytetracycline A stock solution was prepared using 25 mg OTC (Sigma Chemicals Co.) in 5 mL of physiolog- ical saline. Each fish received a dosage of 0.5 mg OTC or 0.1 mL of stock solution. This approxi- mates the dosage reported by Mugiya and Mura- matsu (1982) for goldfish and Weber and Ridgway (1962) for sockeye salmon smolts. Fish were sacri- ficed 21 days after injection, weighed and mea- sured, and both sagittal otoliths were removed, cleaned, and stored dry in the dark. A sagittal section of the right otolith was pre- pared as previously described. Sections were viewed at 160 x magnification, using a compound light microscope equipped with ultraviolet illu- mination. The fluorescent mark was located with an ocular marker. Increments were enumerated from this mark to the outer margins of the otolith using visible light. Results and Discussion Calcium-45 One hundred and three black rockfish were in- jected with the radioisotope, calcium-45; there were no mortalities during the 63-d postinjection sampling period. Over the course of the experi- ment, average fish length increased from 52.5 mm (SD = 1.29, N = 4) on day zero to 70.5 mm (SD = 8.23, N = 4) on day 63; average total body weight increased from 2.3 g (SD = 0.13, N = 4) to 8.10 g(SD = 2.60, A^ = 4). LSC demonstrated that calcium-45 was taken up and retained in the sagittal otoliths of all fish. Incorporation of calcium-45 into the otolith oc- curred as early as 1 hour following injection, which was the initial sampling interval; mean activity at this time was 1,377 (SE = 329) DPM/ mg otolith (Fig. 1). Similar activity values and uptake patterns of calcium-45, up to 72 hours ID.UU- 14.00 .. •• 12.00- -. " 1 E 10.00- . ■■ T ^ - •■ J i a. o 8.00 "o 6.00 i ft 4.00- i 2.00- 0.00- i 1 1 —1 1 1 r 0.00 10.00 20 00 30.00 40.00 50.00 60.00 70,00 DAYS AFTER INJECTION Figure 1. — Accumulation and retention of ^sCa in the otolith of black rockfish. Mean activity and 1 SE are indicated. N = 2 to 4 fish per time interval. postinjection, were observed in the otoliths of rainbow trout, although times of maximum incor- poration and retention were not assessed over longer periods (Mugiya 1974). Radioactivity in the rockfish otoliths increased sharply for the first 15 days (up to an average of 10,279 DPM/mg, SE = 581), followed by a gradual increase to an apparent asymptote (Fig. 1). The rapid uptake and retention of calcium-45 in liver, muscle, and epidermis of .S. melanops, and the gradual elimi- nation of the isotope from these tissues (Yok- lavich and Boehlert unpubl. data), could con- tribute to the increase in otolith activity over time; presumably, calcium-45 is transported from these tissues to the otolith via the blood (Mugiya 1974). The lack of decrease in activity in the otolith substantiates the findings of Ichii and Mugiya (1983) and Campana (1983b), which sug- gest that calcium deposited in otoliths of goldfish and stressed coho salmon, respectively, remains immobilized and is not resorbed. Although no data were presented, it had been suggested ear- lier by Pannella (1980) that resorption of calcium occurs in the otoliths of some tropical fish species, possibly invalidating ages based on increment counts. Our data show no evidence of resorption, lending support to the usefulness of increment counts in reliably estimating age. Scattered exposed silver grains were evident along the interface of otolith section and mount- ing resin in the autoradiographs of otoliths from the earliest sampling periods (1 hour to 1 day), although positive association of the grains with a 828 distinct site of isotope incorporation into the otolith was not discernible. Since the radioactive mark was on the edge of the otoliths from these early sampling periods, it was more difficult to identify than the mark left on otoliths offish sac- rificed later in the experiment. Distinct bands of silver grains, designating the site of isotope up- take, were evident in all but four otoliths sampled from day 4 through day 63 (Fig. 2). A less dense background of grains spanned 7 to 10 bands around the site of uptake; postinjection increment counts were made from the site of densest grain f Jl ^A -^' ^ ,- *.-*^ '1 s* Figure 2. — Example of the silver grains produced by the autoradiograph of "^^Ca in the otolith of a 69 mm juvenile black rockfish. Arrow indicates band of densest grains; for scale, arrow = 35 ixm. This photo is from the anterodorsal region of the otolith. Note the clear increments present on the left side of the figure, represent- ing the increments distal to the time-mark. 829 accumulation to the edge of the otoHth. The pos- terodorsal area of the sagittal sections showed the heaviest accumulation of the isotope and was also the easiest area in which to count increments. This observation is consistent with Irie (1960) and Mugiya (1974), who concluded that high cal- cium uptake occurred in the dorsal region, as well as in the anterior and posterior tips of otoliths; these are the regions of fastest otolith growth. The number of growth increments from the band of densest accumulation of silver grains to the edge of the otolith section closely approxi- mated the number of days the fish were held in the laboratory following injection (Table 1), thereby validating the occurrence of daily growth increments in these juvenile black rockfish. Vali- dation of the frequency of growth increment for- mation, obtained from fish held under optimal growth conditions in the laboratory, and the ap- parent lack of otolith resorption, as demonstrated by the increasing retention of calcium-45 with time, suggest that daily increments on otoliths could provide accurate representation of age and growth for field-caught juvenile Sebastes melanops. Daily increments have recently been suggested to occur on otoliths of early larvae of S. marinus (Penny and Evans 1985). In previous work on juvenile Sebastes, growth increments had been counted but not validated (Moser and Ahlstrom 1978; Boehlert 1981). Our study is thus the first confirmation of daily increments on otoliths of juvenile Sebastes. Checks, or exceptionally dense and dark bands Table 1 . — Age validation using otoliths from black rockfish marked with calcium-45 or oxytetracycline. Number of days from capture to injection is compared with number of growth increments from cap- ture check to time-mark; number of days from injection to sacrifice is compared with number of increments from time-mark to margin of otolith. Number of days N/lean number Treatment/ from increme Check- nts from sample Capture- Injection- Mark- size injection sacrifice mark margin 45Ca/4 10 0.04 10 10 /2 10 0.17 10 10 /2 10 0.5 10 10 /2 10 1 10 10 /1 10 4 10 4 /1 10 11 10 11 /I 10 19 10 19 /2 10 39 10 36 /I 10 55 10 55 /2 10 63 10 63 OTC/5 21 21 1 Scattered silver grains associated with margin. 830 deposited as daily increments, were observed in otoliths from fish used in the calcium-45 experi- ments. One check preceded the radioactive time- mark and another check was associated with the time-mark itself Ten growth increments were noted from the earliest check to the time-mark in each otolith (Table 1). Formations of checks in otoliths have been documented for many species, including coho salmon (Campana 1983b), goldfish (Mugiya and Muramatsu 1982), and several trop- ical species (Brothers et al. 1983). Such checks have been attributed to periods of physiological stress to the fish due to collection, migration, change in feeding or habitat, temperature, life history stages, or anything else that disrupts growth. In the present study, the time from fish collection to injection of the isotope marker was 10 days (Table 1). It seems clear that the observed checks were produced as a consequence of stress encountered during capture and transport to lab- oratory conditions and can be used as additional evidence of daily deposition of growth increments. If such checks are reliably produced, they may be better than chemical time-markers for validation studies such as these. Oxytetracycline The OTC was incorporated into the otoliths of each of the 15 fish injected and produced a distinct fluorescent band. The growth increments follow- ing injection of OTC, however, were less distinct on most of the otoliths and consequently all otoliths could not be used to validate daily incre- ment formation. Weak increment definition fol- lowing OTC incorporation in otoliths of larval spot and pinfish has been reported by Hettler (1984). Although Hettler suggested that the lack of distinct increments resulted from experimental stress, postinjection increments were clearly visi- ble in otoliths from the juvenile rockfish which were injected with calcium-45 and held under lab- oratory conditions similar to the OTC experi- ments. Five of the rockfish otoliths did display clear growth bands following the fluorescent time-mark; enumeration of these increments is summarized in Table 1. These otoliths show the same results as those from calcium-45 treat- ments, demonstrating the daily periodicity of growth increment formation in juvenile black rockfish. It is unclear, however, why 67% of the otoliths failed to produce prominent daily incre- ments after OTC incorporation. The fluorescent OTC mark was more intense and easier to identify than the exposed silver grains in the autoradiographs of most otoliths. OTC is less hazardous to handle in the laboratory and can be detected in the otolith for much longer periods than calcium-45. The OTC was still visi- ble in the otolith at least 3 years following injec- tion and has in fact been used in studies for mass- marking offish for identification purposes, where time at liberty may be even longer (Tsukamoto 1985). The activity of calcium-45 was evident in autoradiographs of a few otoliths which were de- veloped 2 years following injection; amount of ac- tivity depends upon the effective half-life of the isotope (164 days for calcium-45) and the initial amount of activity in the tissue. An autoradio- graph of a transverse section through the otolith of one of the fish injected and held for 1 year revealed a discontinuous band of very faint, ex- posed silver grains, dispersed primarily over the internal and dorsal areas of the otolith. Associa- tion of the isotope with an annular band was not observed. Autoradiographs are difficult to pro- duce, expensive, and time consuming. On the other hand, OTC is simply observed in the otolith section under ultraviolet light. Our recommenda- tions for validating the daily formation of growth increments in juvenile rockfishes are 1) the use of OTC, if growth increments can be routinely iden- tified following injection, or 2) stress marks, which are induced either when transferring fish from the field to laboratory or by subjecting fish to abrupt environmental changes. Where this type of induced stress is not appropriate, as in environ- mentally controlled laboratory studies, and OTC marking is unsuccessful, marks could reliably be produced with the calcium-45 technique de- scribed in this paper. Acknowledgments We thank S. L. Boehlert for generously offering her expertise in preparation of the autoradio- graphs. We also appreciate the helpful comments of S. E. Campana, S. Ralston, and an anonymous reviewer on earlier drafts of this manuscript. This research was supported by NOAA, National Marine Fisheries Service, Northwest and Alaska Fisheries Center, Seattle, WA, through contract 81-ABC-00192-PR6. Literature Cited Barkman, R C 1978. The use of otolith growth rings to age young At- lantic silversides, Menidia menidia. Trans. Am. Fish. Soc. 107:790-793. Boehlert, G W. 1981 . The effects of photoperiod and temperature on labo- ratory growth of juvenile Sebastes diploproa and a com- parison with growth in the field. Fish. Bull., U.S. 79:789-794. 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. Brothers, E B , D. McB Williams, and P. F. Sale 1983. Length of larval life in twelve families of fishes at "One Tree Lagoon", Great Barrier Reef, Aus- tralia. Mar. Biol. 76:319-324. Campana, S. E. 1983a. Feeding periodicity and the production of daily growth increments in otoliths of steelhead trout (Salmo gairdneri) and starry flounder (Platichthys stella- tus). Can. J. Zool. 61:1591-1597. 1983b. Calcium deposition and otolith check formation during periods of stress in coho salmon, Oncorhynchus kisutch. Comp. Biochem. Physiol. 75A:215-220. 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. DaBROWSKI, K , AND K. TSUKAMOTO. 1986. Tetracycline tagging in coregonid embryos and lar- vae. J. Fish Biol. 29:691-698. Geffen, A J 1982. Otolith ring deposition in relation to growth rate in herring (Clupea harengus) and turbot (Scophthalmus maximus) larvae. Mar. Biol. 71:317-326. Hettler, W F 1984. Marking otoliths by immersion of marine fish lar- vae in tetracycline. Trans. Am. Fish. Soc. 113:370-373. Hurley, G V , P. H. Odense, R K O'dor, and E G Da we. 1985. Strontium labelling for verifying daily growth in- crements in the statolith of the short-finned squid (Illex illecebrosus). Can. J. Fish. Aquat. Sci. 42:380-383. ICHII, T , AND Y. MUGIYA 1983. Comparative aspects of calcium dynamics in calci- fied tissues in the goldfish Carassius auratus. Bull. Jpn. Soc. Sci. Fish. 49:1039-1044. IRIE, T 1960. The growth of the fish otolith. J. Fac. Fish. Anim. Hush. Hiroshima Univ. 3:203-229. Miller. S J , and T. Storck. 1982. Daily growth rings in otoliths of young-of-the-year largemouth bass. Trans. Am. Fish. Soc. 111:527-530. MOSER, H G , AND E. H AHLSTROM. 1978. Larvae and pelagic juveniles of blackgill rockfish, Sebastes melanostomus , taken in midwater trawls off southern California and Baja California. J. Fish. Res. Board Can. 35:981-996. MUGIYA, Y 1974. Calcium-45 behavior at the level of the otolithic organs of rainbow trout. Bull. Jpn. Soc. Sci. Fish. 40:457-463. MUGIYA. Y , AND J MURAMATSU 1982. Time-marking methods for scanning electron mi- croscopy in goldfish otoliths. Bull. Jpn. Soc. Sci. Fish. 831 48:1225-1232. Neilson. J D., AND G H Geen. 1982. Otoliths of chinook salmon (Oncorhynchus tshwytscha ): daily growth increments and factors influ- encing their production. Can. J. Fish. Aquat. Sci. 39:1340-1347. Pannella, G 1980. Growth patterns in fish sagittae. In D. C. Rhoads and R. A. Lutz (editors), Skeletal growth of aquatic or- ganisms: biological records of environmental change, p. 519-560. Plenum Press, N.Y. Penny, R W , and G T Evans 1985. Growth histories of larval redfish (.Sebastes spp.) on an offshore Atlantic fishing bank determined by otolith increment analysis. Can. J. Fish. Aquat. Sci. 42:1452- 1464. Ralston, S . and G. T Miyamoto 1983. Analyzing the width of daily otolith increments to age the Hawaiian snapper, Pristipomoides filamento- sus. Fish. Bull., U.S. 81:523-535. Tanaka. K., 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. Taubert, B D., and D W Coble 1977. Daily rings in otoliths of three species of Lepomis and Tilapia mossambica. J. Fish. Res. Board Can. 34:332-340. TSUKAMOTO. K 1985. Mass-marking of ayu eggs and larvae by tetracy- cline-tagging of otoliths. Bull. Jpn. Soc. Sci. Fish. 51:903-911. Weber. D D., and G J Ridgway. 1962. The deposition of tetracycline drugs in bones and scales of fish and its possible use for marking. Prog. Fish-Cult. 150-155. Wild, A , and T J. Foreman. 1980. The relationship between otolith increments and time for yellowfin and skipjack tuna marked with tetra- cycline. Inter-Am. Trop. Tuna Comm. Bull. 17:507-541. Yamada. S B . T J Mulligan, and S J Fairchild 1979. Strontium marking of hatchery-reared coho salmon (Oncorhynchus kisutch, Walbaum). J. Fish. Biol. 14:267-275. Mary M. Yoklavich Mark O. Hatfield Marine Science Center Oregon State University Newport. OR 97365 Present address: Northwest and Alaska Fisheries Center National Marine Fisheries Service, NOAA 7600 Sand Point Way NE, BIN CI5700 Seattle, WA 98115 George W. Boehlert Southwest Fisheries Center Honolulu Laboratory National Marine Fisheries Service, NOAA 2570 Dole St. Honolulu, HI 96822-2396 832 INDEX Fishery Bulletin: Vol. 85, Nos. 1-4 ABLE, K. W., D. C. TWICHELL, C. B. GRIMES, and R. S. JONES, "Sidescan sonar as a tool for detection of demersal fish habitats," 725 ABLE, K. W— see WILSON et al. "Age and growth reproductive cycle, and histochemical tests for heavy metals in hard clams, Mercenaria mercenaria , from Rar- itan Bay, 1974-75," by John Ropes, 653 "Age and growth of Spanish mackerel, Scomberomorus macula- tus, from Florida and the Gulf of Mexico." by William A. Fable, Jr., Allyn G. Johnson, and Lyman E. Barger, 777 "Age determination of Pacific cod, Gadus macrocephalus , using five ageing methods," by Han-Lin Lai, Donald R. Gunderson, and Loh Lee Low, 713 Ageing studies allocation of age-length keys, 179 bass, striped, 171 clams, hard, 653 cod. Pacific, 713 flounder, witch, 611 mackerel. Spanish. 777 scad, round, 251 scorpionfish. 99 shark, tiger. 269 AHRENHOLZ, DEAN W.. WALTER R. NELSON, and SHERYAN P. EPPERLY, "Population and fishery character- istics of Atlantic menhaden, Brevoortia tyrannus" 569 Albacore fleet interaction affecting catch, 703 ALLEN, ROBERT L.— see RENDER and ALLEN Allen's recruitment rate bias and variance. 117 AllocUnus holderi — see Kelpfish. island Allozyme data larval drift, 1 Ammodytes sp. — see Sand lance "Analysis of sea turtle captures and mortalities during commer- cial shrimp trawling," by Tyrrell A. Henwood and Warren E. Stuntz, 813 Anchovy Hawaiian, fecundity and spawning. 127 northern, feeding and growth. 213 Anoplopoma fimbria — see Sablefish APPELDOORN, RICHARD S., "Assessment of mortality in an offshore population of queen conch, Strombus gigas L., in southwest Puerto Rico," 797 Argopecten irradians — see Scallops, bay ARMETTA, THERESE M., and BRADLEY G. STEVENS, "Aspects of the biology of the hair crab, Erimacrus isenbeckii, in the eastern Bering Sea," 523 "Aspects of the biology of the hair crab, Erimacrus isenbeckii , in the eastern Bering Sea," by Therese M. Armetta and Bradley G. Stevens, 523 "Assessment of interaction between North Pacific albacore, Thunnus alalunga, fisheries by use of a simulation model," by P. Kleiber and B. Baker, 703 "Assessment of mortality in an offshore population of queen conch, Strombus gigas L., in southwest Puerto Rico," by Richard S. Appeldoorn, 797 Autoradiography evaluation of method, 826 AXELL, BRITA— see LOVE et al. Bahamas, Berry Island conch, queen, 299 BAIRD, RONALD C— see GARTNER et al. BAKER, B.— see KLEIBER and BAKER BANNEROT, SCOTT P.— see IVERSEN et al. BARGER, LYMAN E.— see FABLE et al. BARNETT, HAROLD J.— see KUDO et al. Bass kelp, larval drift, 1 striped otolith ageing technique, 171 population structure, 167 BECKER, D. SCOTT, and KENNETH K. CHEW, "Predation on Capitella spp. by small-mouthed pleuronectids in Puget Sound, Washington," 471 Behavior studies lobster, spiny, 45 Benthic macroinvertebrates impact of clam harvesting, 281 Benthosema suborbitale — see Lantemfishes BEST, PETER B., "Estimates of the landed catch of right (and other whalebone) whales in the American fishery, 1805- 1909," 403 833 "Bias and variance in Allen's recruitment rate method," by J.W. Horwood, 117 Billfish larval studies, 757 "Biological data on Berry Islands (Bahamas) queen conchs, Strombus gigas, with mariculture and fisheries management implications," by Edwin S. Iversen, Edward S. Rutherford, Scott P. Bannerot, and Darryl E. Jory, 299 Blacksmith larval drift, 1 life history summary, 4 BOEHLERT, G. W— see KENDALL et al. BOEHLERT, GEORGE W— see YOKLAVICH and BOEH- LERT BOGDANOWICZ, S. M — see KORNFIELD and BOGDANO- WICZ BOROWSKY, BETTY, "Laboratory studies of the pattern of reproduction of the isopod crustacean Idotea baltica" 377 BOTTON, MARK L., and JOHN W. ROPES, "Populations of horseshoe crabs, Limulus polyphemus , on the northwestern At- lantic continental shelf," 805 BOURKE, ROBERT— see BRILL et al. BOWERING, W. R., "Distribution of witch flounder, Glypto- cephalus cynoglossus , in the southern Labrador and eastern Newfoundland area and changes in certain biological parame- ters after 20 years of exploitation," 611 BRANSTETTER, STEVEN, J. A. MUSICK, and J. A. COLVO- CORESSES, "A comparison of the age and growth of the tiger shark, Galeocerdo cuvieri , from off Virginia and from the north- western Gulf of Mexico," 269 Brevoortia patronus — see Menhaden, gulf Brevoortia tyrannus — see Menhaden, Atlantic BRILL, RICHARD W., "On the standard metabolic rates of tropical tunas, including the effect of body size and acute tem- perature change," 25 BRILL, RICHARD W., ROBERT BOURKE, JAMES A. BROCK, and MURRAY D. DAILEY, "Prevalence and effects of infection of the dorsal aorta in yellowfin tuna, Thunnus albacares , by the larval cestode, Dasyrhynchus talis- mani," 767 BROCK, JAMES A— see BRILL et al. BROOKS, ANDREW— see LOVE et al. Brosme brosme — see Cusk BROTHERS, EDWARD B.— see JONES and BROTHERS BUTLER, RICHARD W., WALTER A. NELSON, and. TYRRELL A. HENWOOD, "A trawl survey method for estimat- ing loggerhead turtle, Caretta caretta , abundance in five east- em Florida channels and inlets," 447 Butterfish lunar phase and catch, 817 CALDER, DANIEL R— see GREELEY et al. CALKINS, DONALD G — see MERRICK et al. Callinectes sapidus — see Crab, blue Callorhinus ursinus — see Seal, northern fur Cancer magister — see Crab, Dungeness Capitella spp. — see Polychaetes Carcinonemertes errans — see nemertean Dungeness, crab egg predation Caretta caretta — see Turtles Caulolatilus princeps — see Whitefish, ocean Caulolatilus — see Tilefishes Ceratoscopelus warmingii — see Lanternfishes Cestodes infection in yellowfin tuna, 767 "Changes in the population structure of male striped bass. Mo- rone saxatilis , spawning in the three areas of the Chesapeake Bay from 1984 to 1986," by Robert W. Chapman, 167 CHAPMAN, ROBERT W., "Changes in the population struc- ture of male striped bass, Morone saxatilis, spawning in the three areas of the Chesapeake Bay from 1984 to 1986," 167 Chelonia my das — see Turtles, sea CHEW, KENNETH K — see BECKER and CHEW Chomis punctipinnis — see Blacksmith Clams, hard age, growth, and reproduction, 653 histochemical testing, 653 mechanical harvesting, 281 CLARKE, M. E.— see KENDALL et al. CLARKE, THOMAS A., "Fecundity and spawning frequency of the Hawaiian anchovy or nehu, Encrasicholina pur- purea," 127 Clinocottus analis — see Sculpin, wooly Clupea harengus — see Herring, Atlantic •God, Pacific • age determination, 713 ' allocation jof age-length keys, 179 834 COLEY, TRAVIS C— see McCABE et al. R. Loughlin, and Donald G. Calkins, 351 COLLINS, MARK R., C. WAYNE WALTZ, WILLIAM A. ROU- MILLAT, and DARYL L. STUBBS, "Contribution to the life history and reproductive biology of gag, Mycteroperca mi- crotepis (Serranidael, in the South Atlantic Bight," 648 COLLINS, ROBSON— see LOVE et al. COLOGNE, JOHN B— see HOLT et al. COLVOCORESSES, J. A.— see BRANSTETTER et al. "A comparison of the age and growth of the tiger shark, Gale- ocerdo cumeri , from off Virginia and from the northwestern Gulf of Mexico," by Steven Branstetter, J. A. Musick, and J. A. Colvocoresses, 269 Conch, queen biological data for mariculture and management, mortality rates, 797 299 "Contribution to the life history and reproductive biology of gag, Mycteroperca microlepis (Serranidae), in the South Atlantic Bight," by Mark R. Collins, C. Wayne Waltz, William A. Rou- millat, and Daryl L. Stubbs, 648 Crab blue, survival in eelgrass, 53 Dungeness, egg predation, 140 golden, exploratory trapping, 547 hair, biology, 523 horseshoe, population studies, 805 "Crater wounds on northern elephant seals: the cookiecutter shark strikes again," by Burney J. LE BOEUF, JOHN E. McCOSKER, and JOHN HEWITT, 387 Croaker, Atlantic density and distribution, 601 predation on brown shrimp, 59 Crustacean, isopod reproduction, 377 Cusk habitat detection using sonar, 725 DAILEY, MURRAY D.— see BRILL et al. "Daily growth increments in otoliths of juvenile black rockfish, Sebastes melanops: An evaluation of autoradiography as a new method of validation," by Mary M. Yoklavich and George W. Boehlert, 826 DALLEY, E. L., and G. H. WINTERS, "Early life history of sand lance (Ammodytes), with evidence for spawning of A. du- bius in Fortune Bay, Newfoundland," 631 Dasyrhynchus talismani — see Cestodes Decapterus punctatus — see Scad, round "Decline in abundance of the northern sea lion, Eumetopias jubatus, in Alaska, 1956-86," by Richard L Merrick, Thomas "Density and depth distribution of larval gulf menhaden, Bre- voortia patronus, Atlantic croaker, Micropogonias undulatus, and spot Leiostomus xanthurus , in the northern Gulf of Mex- ico," by Susan M. Sogard, Donald E. Hoss, and John J. Govoni, 601 Dermochelys coriacea — see Turtles, sea "Development of the eggs and larvae of the yellowchin sculpin, Icelinus quadriseriatus (Pisces; Cottidae)," by Richard F. Feeney, 201 Diaphus dimerilii — see Lantemfishes "Differentiation of mitochondrial DNA in Atlantic herring, Clupea harengus ," by I. Kornfield and S. M. Bogdanowicz, 561 "Digestion rates and gastric evacuation times in relation to temperature of the Sacramento squawfish, Ptychocheilus gran- dis," by Bruce Vondracek, 159 "Distribution, abundance, reproduction, food habits, age, and growth of round scad, Decapterus punctatus, in South Atlantic Bight," by L. Stanton Hales, Jr., 251 "Distribution and abundance of billfish larvae (Pisces: Istio- phoridae) in the Great Barrier Reef Lagoon and Coral Sea near Lizard Island, Australia," by Jeffrey M. Leis, Barry Goldman, and Shoji Ueyanagi, 757 "Distribution and yield of the deepwater shrimp Heterocarpus resource in the Marianas," by Robert B. MofTit and Jeffrey J. Polovina, 339 "Distribution, feeding, and growth of larval walleye pollock, Theragra chalcogramma , from Shelikof Strait, Gulf of Alaska," by A. W. Kendall, Jr., M. E. Clarke, M. M. Yoklavich, and G. W. Boehlert, 499 "Distribution of witch flounder, Glyptocephalus cynoglossus , in the southern Labrador and eastern Newfoundland area and changes in certain biological parameters after 20 years of ex- ploitation," by W. R. Bowering, 611 DNA, mitochondrial herring, Atlantic, 561 Dolphin schools, density estimates, 419 spotted, abundance surveys, 435 Dorosoma cepedianum — see Isopod crustacean Dover sole predation on polychaetes, 471 DUNBAR-COOPER, ANN— see LUTZ and DUNBAR- COOPER "Early life history of sand lance (Ammodytes), with evidence for spaviTiing of A. dubius in Fortune Bay, Newfoundland," by E. L. Dalley and G. H. Winters, 631 835 "Ecological consequences of mechanical harvesting of clams," Charles H. Peterson, Henry C. Summerson, and Stephen R. Fegley, 281 Ecology mechanical harvesting, 281 Economic studies allocation of age-length keys, 179 crab, hair, 523 whale products, 1805-1909, 403 whiting. Pacific, 745 Eelgrass survival of blue crab, 53 "Effect of a river-dominated estuary on the prevalence of Car- cinonemertes errans , an egg predator of Dungeness crab, Cancer magister" by George T. McCabe, Jr., Robert L. Emmett, Travis C. Coley, and Robert J. McConnell, 140 "Effects of air exposure on desiccation rate, hemolymph chem- istry, and escape behavior of the spiny lobster, Panulirus argus," by Gregory K. Vermeer, 45 "Effects of an El Niiio event on the food habits of larval sablefish, Anoplopoma fimbria, off Oregon and Washington," by Jill J. Grover and Bori L. 011a, 71 "The effects of bottom trawling on American lobsters, Homarus americanus, in Long Island Sound," by Eric M. Smith and Penelope T. Howell, 737 El Nino effects on sablefish, 71 Electrophoresis genetic variation in salmon, 681 Embiotoca jacksoni — see Perch, black EMMETT, ROBERT L.— see McCABE et al. Encrasicholina purpurea — see Anchovy, Hawaiian English sole predation on polychaetes, 471 Engraulis mordax — see Anchovy, northern Environmental studies butterfish, 817 mechanical clam harvesting, 281 EPPERLY, SHERYAN P.— see AHRENHOLZ et al. Eretmochelys imbricata — see Turtles, sea Erimacrus isenbeckii — see Hair crab "Estimates of the landed catch of right (and other whalebone) whales in the American fishery, 1805-1909," by Peter B. Best, 403 "Estimating density of dolphin schools in the eastern tropical Pacific Ocean by line transect methods," by Rennie S. Holt, 419 Estuarine studies crab egg predation, 140 mechanical clam harvesting, 281 predation on crab eggs, 140 shrimp, brown, 59 Eumetopias jubatus — see Sea lion, northern Euthynnus affinis — see Kawakawa Evolutionary studies larval drift, 1 "Exploration for golden crab, Geryon fenneri, in the South At- lantic Bight: distribution, population structure, and gear as- sessment," by Elizabeth L. Wenner, Glenn F. Ulrich, and John B. Wise, 547 FABLE, WILLIAM A., Jr., ALLYN G. JOHNSON, and LYMAN E. BARGER, "Age and growth of Spanish mackerel, Scomberomorus maculatus , from Florida and the Gulf of Mex- ico," 777 "Factors affecting cooked texture quality of Pacific whiting, Merluccius productus , fillets with particular emphasis on the effects of infection by the Myxosporeans Kudoa paniformis and K. thyrsitis." by George Kudo, Harold J. Barnett, and Richard W. Nelson, 745 "Fecundity and spawning frequency of the Hawaiian anchovy or nehu, Encrasicholina purpurea," by Thomas A. Clarke, 127 "Feeding ecology and growth energetics of larval northern an- chovy, Engraulis mordax," by Gail H. Theilacker, 213 "Feeding habitats of spot, Leiostomus xanthurus , in polyhaline versus meso-oligohaline tidal creeks and shoals," by Steven P. O'Neil and Michael P. Weinstein, 785 FEENEY, RICHARD F., "Development of the eggs and larvae of the yellowchin sculpin, Icelinus quadriseriatus (Pisces: Cotti- dae)," 201 FEGLEY, STEPHEN R.— see PETERSON et al. Fish habitats sonar detection, 725 Fish, oceanic tagging and injection method. 645 "Fish predation on juvenile brown shrimp, Penaeus aztecus Ives: effects of turbidity and substratum on predation rates," by Thomas J. Minello, Roger J. Zimmerman, and Eduardo X. Mar- tinez, 59 FISHER, J. P., and W. P. PEARCY, "Movements of coho, Oncorhynchus kisutch , and chinook, O. tshawytscha, salmon tagged at sea off Oregon, Washington, and Vancouver Island during summers 1982-85," 819 Fishery albacore, 703 836 commercial habitat detection using sonar, 725 flounder, witch, 611 lobster, 737 management queen conchs, 299 menhaden, Atlantic, 569 salmon, chinook, 13 scorpionfish, 99 shrimp sea turtle capture, 813 troll sockeye salmon migration, 455 tuna incidental dolphin mortality, 435 whale, 1805-1909, 403 whiting. Pacific, 745 Fishes, marine shore larval drift, 1 Flatfish predation on polychaetes, 471 Flounder, southern, shrimp predation, 59 witch distribution and exploitation, 611 habitat, 147 Food habits anchovy, northern, 213 sablefish, 71 scad, round, 251 scorpion fish, 99 spot, 785 squawfish, 159 FOREMAN, TERRY, "A method of simultaneously tagging large oceanic fish and injecting them with tetracycline," 645 Fourier series (FS) density estimates dolphin school, 419 estimates of dolphin density, 419 Gadus macrocephalus — see Cod, Pacific Gag life history and reproduction. 648 Galeocerdo cuvieri — see Shark, tiger GARTNER, JOHN V., Jr., THOMAS L. HOPKINS, RONALD C. BAIRD, and DEAN M. MILLIKEN, "The lanternfishes (Pisces: Myctophidae) of the eastern Gulf of Mexico," 81 Gear flounder habitat surveys. 147 "Genetic estimates of stock compositions of 1983 chinook salmon, Oncorhynchus tshawytscha , harvests off the Washing- ton coast and the Columbia River," by Fred Utter, David Teel, George Milner, and Donald Mclsaac, 13 Genetic stock identification salmon, chinook, 13 Genetic studies bass, striped, 167 herring, Atlantic, 561 larval drift, 1 salmon, chinook, 13, 681 salmon, coho, 681 "Genetic variation in chinook, Oncorhynchus tshawytscha , and coho, O. kisutch, salmon from the north coast of Washington," by R. R. Reisenbichler and S. R. Phelps, 681 GERRODETTE, TIM— see HOLT et al. Geryon fenneri — see Crab, golden Girella nigricans — see Opaleye Glyptocephalus cynoglossus — see Flounder, witch Glyptocephalus zachirus — see Rex sole Goby, bluebanded larval drift, 1 life history summary, 4 GOLDMAN, BARRY— see LEIS et al. GOVONI, JOHN J.— see SOGARD et al. GREELEY, MARK S., Jr., DANIEL R. CALDER, and ROBIN A. WALLACE, "Oocyte growth and development in the striped mullet, Mugil cephalus, during seasonal ovarian re- crudescence: relationship to fecundity and size at matu- rity," 187 GRIMES, C. B.— see ABLE et al. GROOT, C, and T. P QUINN, "Homing migration of sockeye salmon, Oncorhynchus nerka, to the Eraser River," 455 Gross conversion efficiency new expression, 139 GROVER, JILL J., and BORI L. OLLA, "Effects of an El Niiio event on the food habits of larval sablefish, Anoplopoma fim- bria, off Oregon and VJashington" 71 Growth studies anchovy, northern, 213 clams, hard, 653 conch, queen, 299 crab, hair, 523 flounder, witch, 611 gross conversion efficiency, 139 mackerel, Spanish, 777 mullet striped, 187 pollock, walleye, 499 sand lance, 631 scad, round, 251 scorpionfish, 99 sculpin, yellowchin, 201 shark, tiger, 269 shrimp, deepwater, 339 837 Growth studies — Continued squid, 163 trout, rainbow, 395 GSI — see Genetic stock identification Gulf of Mexico lantemfishes, 81 GUNDERSON, DONALD R— see LAI et al. "Habitat partitioning by size in witch flounder, Glypto- cephalus cynoglossus: a reevaluation with additional data and adjustments for gear selectivity," by S. J. Walsh, 147 Habitat studies flounder, witch, 147 spot, 785 HALES, L. STANTON, Jr., "Distribution, abundance, re- production, food habits, age, and growth of round scad, De- capterus punctatus, in South Atlantic Bight," 251 Halfmoon larval drift, 1 life history summary, 4 HANLON, ROGER T., PHILIP E. TURK, PHILLIP G. LEE, and WON TACK YANG, "Laboratory rearing of the squid Loligo pealei to the juvenile stage: growth comparisons with fishery data," 163 Harvesting studies clams, 281 conch, queen, 299 Hatchery studies conch, queen, 299 "Heart and gill ventilatory activity in the lobster, Homarus americanus , at various temperatures," by Renee Mercaldo- Allen and Frederick P. Thurberg, 643 HECK, K. L., Jr.— see WILSON et al. KENWOOD, TYRRELL A., and WARREN E. STUNTZ, "Analysis of sea turtle captures and mortalities during commercial shrimp trawling," 813 HENWOOD, TYRRELL A.— see BUTLER et al. Herring, Atlantic mitochondrial DNA, 561 Heterocarpus ensifer — see Shrimp, deepwater Heterocarpus laevigatas — see Shrimp, deepwater Heterocarpus longirostrus — see Shrimp, deepwater HEWITT, JOHN— see LE BOEUF et al. Hibernation studies sea turtle, 34 HINCKLEY, SARAH, "The reproductive biology of walleye pollock, Theragra chalcogramma , in the Bering Sea, with refer- ence to spawning stock structure," 481 Histochemical testing clams, hard, 653 HOLT, RENNIE S., "Estimating density of dolphin schools in the eastern tropical Pacific Ocean by line transect meth- ods," 419 HOLT, RENNIE S., TIM GERRODETTE, and JOHN B. COLOGNE, "Research vessel survey design for monitoring dol- phin abundance in the eastern tropical Pacific," 435 Homarus americanus — see Lobster, American "Homing migration of sockeye salmon, Oncorhynchus nerka, to the Eraser River," by C. Groot and T. P. Quinn, 455 HOPKINS, THOMAS L.— see GARTNER et al. HORWOOD, J. W., "Bias and variance in Allen's recruitment rate method," 117 HOSS, DONALD E.— see SOGARD et al. HOWELL, PENELOPE T.— see SMITH and HOWELL Icelinus quadriseriatus — see Sculpin, yelllowchin Isistius brasiliensis — see Shark, cookiecutter Isopod, crustacean reproduction, 377 IVERSEN, EDWIN S., EDWARD S. RUTHERFORD, SCOTT P. BANNEROT, and DARRYL E. JORY, "Biological data on Berry Islands (Bahamas) queen conchs, Strombus gigas, with mariculture and fisheries management implications," 299 JOHNSON, ALLYN G.— see FABLE et al. JONES, CYNTHIA, and EDWARD B. BROTHERS, "Validation of the otolith increment aging technique for striped bass, Morone saxatilis, larvae reared under subopti- mal feeding conditions," 171 JONES, R. S.— see ABLE et al. JORY, DARRYL E.— see IVERSEN et al. "Juvenile blue crab, Callinectes sapidus , survival: an evalua- tion of eelgrass, Zostera marina, as refuge," by K. A. Wilson, K. L. Heck, Jr., and K. W. Able, 53 Katsuwonus pelamis — see Tuna, skipjack Kawakawa second specimen, 650 standard metabolic rates, 25 838 Kelp bass larval drift. 1 life history summary, 4 Kelpfish, island larval drift, 1 life history summary, 4 KENDALL. A. W. Jr.. M. E. CLARKE. M. M. YOKLAVICH, and G. W. BOEHLERT, "Distribution, feeding, and growth of larval walleye pollock, Theragra chalcogramma , from Shelikof Strait, Gulf of Alaska," 499 KLEIBER, P., and B. BAKER. "Assessment of interaction be- tween North Pacific albacore, Thunnus alalunga, fisheries by use of a simulation model." 703 KORNFIELD, I., and S. M. BOGDANOWICZ. "Differentia- tion of mitochondrial DNA in Atlantic herring, Clupea haren- gus," 561 KOZLOFF. PATRICK— see YORK and KOZLOFF Larval studies anchovy, northern, 213 bass, striped, 171 billfish, 757 cestodes, 767 crab, blue, 53 croaker, Atlantic, 601 drift patterns, 1 marlin black, 757 blue, 757 menhaden, gulf, 601 pollock, walleye, 499 sablefish, 71 sailfish, 757 sand lance, 631 sculpin, yellowchin, 201 spot, 601 LE BOEUF, BURNEY J., JOHN E. McCOSKER, and JOHN HEWITT, "Crater wounds on northern elephant seals: the cookiecutter shark strikes again," 387 KUDO. GEORGE. HAROLD J. BARNETT, and RICHARD W. NELSON, "Factors aftecting cooked texture quality of Pacific whiting, Merluccius productus , fillets with particular emphasis on the effects of infection by the Myxosporeans Kudoa paniformis and K. thyrsitis , 745 Kudoa paniformis — see Myxosporean Kudoa thrysitis — see Myxosporean KYNARD, BOYD, and JOHN P. WARNER, "Spring and sum- mer movements of subadult striped bass, Morone saxatilis, in the Connecticut River." 143 "Laboratory rearing of the squid Loligo pealei to the juvenile stage; growth comparisons with fishery data." by Roger T. Han- Ion, Philip E. Turk, Phillip G. Lee, and Won Tack Yang, 163 "Laboratory studies of the pattern of reproduction of the isopod crustacean Idotea baltica ," by Betty Borowsky. 377 Lagodon rhomboides — see Pinfish LAI, HAN-LIN, "Optimum allocation for estimating age com- position using age-length key." 179 LAI, HAN-LIN, DONALD R. GUNDERSON, and LOH LEE LOW, "Age determination of Pacific cod, Gadus macro- cephalus, using five ageing methods," 713 Lampanyctus alatus — see Lanternfishes Lanternfishes Gulf of Mexico species, 81 'The lanternfishes (Pisces: Myctophidae) of the eastern Gulf of Mexico," by John V. Gartner. Jr., Thomas L. Hopkins, Ronald C. Baird, and Dean M. Milliken, 81 LaRIVIERE, M.— see MATHEWS and LaRIVIERE LEE, PHILLIP G.— see HANLON et al. Leiostomus xanthurus — see SPOT LEIS, JEFFREY M., BARRY GOLDMAN, and SHOJI UEYANAGI, "Distribution and abundance of billfish larvae (Pisces: Istiophoridae) in the Great Barrier Reef Lagoon and Coral Sea near Lizard Island, Australia," 757 Lepholatilus — see Tilefishes Lepidochelys kempi — see Turtles, sea Lepidophanes guentheri — see Lanternfishes "Life history and fishery of the California scorpionfish, Scor- paena guttata , within the Southern California Bight," by Milton S. Love, Brita Axell, Pamela Morris, Robson Collins, and An- drew Brooks, 99 Life history studies gag, 648 sand lance, 631 scorpionfish, 99 Limulus polyphemus — see Crabs, horseshoe Line transect theory (LT) density of dolphin schools, 419 Lingcod movement in the Pacific Northwest, 153 Lobster American effects of trawling, 737 habitat detection using sonar, 725 ventilatory activity, 643 spiny effects of air exposure, 45 Loligo pealei — see Squid 839 Longlining albacore fleet interaction, 703 LOUGHLIN, THOMAS R.— see MERRICK et al. LOVE, MILTON S., BRITA AXELL, PAMELA MORRIS, ROB- SON COLLINS, and ANDREW BROOKS, "Life history and fishery of the California scorpionfish, Scorpaena guttata , within the Southern California Bight," 99 LOW, LOH LEE— see LAI et al. LUTZ, PETER L., and ANN DUNBAR-COOPER, "Variations in the blood chemistry of the loggerhead sea turtle, Caretta caretta ," 37 Lythrypnus dalli — see Goby, bluebanded Mackerel, Spanish age and growth, 777 Macrozoarces americanus — see Pout, ^cjan Management — see Fisheries, management Mariculture conch, queen, 299 Marlin black, 757 blue, 757 MARTINEZ, EDUARDO X — see MINELLO et al. MATHEWS, S. B., and M. LaRIVIERE, "Movements of tagged lingcod, Ophiodon elongatus, in the Pacific North- west," 153 McCABE, GEORGE T., Jr., ROBERT L. EMMETT, TRAVIS C. COLEY, and ROBERT J. McCONNELL, "Efi'ect of a river- dominated estuary on the prevalence of Carcinonemertes er- rans, an egg predator of the Dungeness crab, Cancer magister," 140 McCONNELL, ROBERT J.— see McCABE et al. McCOSKER, JOHN E.— see LE BOEUF et al. McISAAC, DONALD— see UTTER et al. Medialuna californiensis — see Halfmoon Menhaden Atlantic population and fishery characteristics, 569 gulf density and distribution, 601 MERCALDO-ALLEN, RENEE, and FREDERICK P. THUR- BERG, "Heart and gill ventilatory activity in the lobster, Homarus americanus , at various temperatures," 643 Mercenaria mercenaria — see Clam, hard Merluccius productus — see Whiting, Pacific 840 MERRICK, RICHARD L., THOMAS R. LOUGHLIN, and DON- ALD G. CALKINS, "Decline in abundance of the northern sea lion, Eumetopias jubatus , in Alaska, 1956-86," 351 "A method of simultaneously tagging large oceanic fish and injecting them with tetracycline," by Terry Foreman, 645 Micropoganias undulatus — see Croaker, Atlantic Microstomus pacificus — see Dover sole Migration studies crab, hair, 523 salmon, 819 salmon, sockeye, 455 shad, gizzard, 380 MILLIKEN, DEAN M — see GARTNER et al. MILNER, GEORGE— see UTTER et al. MINELLO, THOMAS J., ROGER J. ZIMMERMAN, and EDU- ARDO X. MARTINEZ, "Fish predation on juvenile brown shrimp, Penaeus aztecus Ives: effects of turbidity and substra- tum on predation rates," 59 Mirounga angustirostris — see Seal, northern elephant Models simulation of fleet interaction affecting catch, 703 MOFFIT, ROBERT B., and JEFFREY J. POLOVINA, "Distri- bution and yield of the deepwater shrimp Heterocarpus resource in the Marianas," 339 Morone saxatilis — see Bass, striped Morphometric variation perch. Pacific ocean, 663 "Morphometric variation of Pacific ocean perch, Sebastes alu- tus, off western North America," by Jay C. Quast, 663 MORRIS, PAMELA— see LOVE et al. Mortality rates conch, queen, 797 "Movements of coho, Oncorhynchus kisutch , and chinook, O. tshawytscha , salmon tagged at sea off Oregon, Washington, and Vancouver Island during summers 1982-85," by J. P. Fisher and W. C. Pearcy, 819 "Movements of tagged lingcod, Ophiodon elongatus, in the Pacific Northwest," by S. B. Mathews and M. LaRiviere, 153 Mugil cephalus — see Mullet, striped MUGIYA, YASUO, "Phase difference between calcification and organic matrix formation in the diurnal growth of otoliths in the rainbow trout, Salmo gairdneri ," 395 Mullet, striped growth and development of oocytes, 187 MUSICK, J. A.— see BRANSTETTER et al. Mycteroperca microlepis — see Gag Myctophum affine — see Lanternfishes Myxosporeans infection in Pacific whiting, 745 Nehu — see Anchovy, Hawaiian NELSON, RICHARD W— see KUDO et al. NELSON, WALTER A— see BUTLER et aL NELSON, WALTER R.— see AHRENHOLZ et al. Nemertean crab egg predation, 140 predation on crab eggs, 140 Notolychnus valdiviae — see Lanternfishes O'LEARY, JOHN, and DOUGLAS G. SMITH, "Occurrence of the first freshwater migration of the gizzard shad, Dorosoma cepedianum , in the Connecticut River, Massachusetts," 380 O'NEIL, STEVEN P., and MICHAEL P. WEINSTEIN, "Feed- ing habitats of spot, Leiostomus xanthurus , in polyhaline versus meso-oligohaline tidal creeks and shoals," 785 "Occurrence of the first freshwater migration of the gizzard shad, Dorosoma cepedianum , in the Connecticut River, Massa- chusetts," by John O'Leary and Douglas G. Smith, 380 OLLA, BORI L— see GROVER and OLLA "On the compatibility of a new expression for gross conversion efficiency with the von Bertalanffy growth equation," by W. Silvert and D. Pauly, 139 "On the estimation of the numbers of northern fur seal, Cal- lorhinus ursinus , pups born on St. Paul Island, 1980-86," by Anne E. York and Patrick Kozloff, 367 "On the standard metabolic rates of tropical tunas, including the effect of body size and acute temperature change," by Richard W. Brill, 25 Oncorhynchus kisutch — see Salmon, coho Oncorhynchus nerka — see Salmon, sockeye Oncorhynchus tshawytscha — see Salmon, chinook "Oocyte growth and development in the striped mullet, Mugil cephalus, during seasonal ovarian recrudescence: relationship to fecundity and size at maturity," by Mark S. Greeley, Jr., Daniel R. Calder, and Robin A. Wallace, 187 Oocytes mullet striped, 187 Opaleye larval drift, 1 life history summary, 4 Ophiodon elongatus — see Lingcod "Optimum allocation for estimating age composition using age- length key," by Han-Lin, Lai, 179 Otolith studies bass, striped, 171 mackerel, Spanish, 777 rockfish, 383 rockfish, black, 826 scad, round, 251 trout, rainbow, 395 Panulirus argus — see Lobster, spiny Paralabrax clathratus — see Kelpbass Paralichthys lethostigma — see Flounder, southern Parasite studies cestode infection in tuna, 767 Myxosporean infection in Pacific whiting, 745 Parophrys vetulus — see English sole "Patterns of larval drift in southern California marine shore fishes inferred from allozyme data," by Robin S. Waples and Richard H. Rosenblatt, 1 PAULY, D.— see SILVERT and PAULY PEARCY, W. P.— see FISHER and PEARCY Pelagic dispersal patterns of larval drift, 1 Penaeus aztecus — see Shrimp, brown Peprilus burti — see Butterfish Perch black larval drift, 1 life history summary, 4 Pacific ocean morphometric variation, 663 PEREZ FARFANTE, ISABEL, "Revisions of the gamba prawn genus Pseudaristeus , with description of two new species (Crustacea: Decapoda: Penaeoidea)," 311 PETERSON, CHARLES H., HENRY C. SUMMERSON, and STEPHEN R. FEGLEY, "Ecological consequences of mechani- cal harvesting of clams," 281 "Phase difference between calcification and organic matrix for- mation in the diurnal growth of otoliths in the rainbow trout, Salmo gairdneri," by Yasuo Mugiya, 395 PHELPS, S. R.— see REISENBICHLER and PHELPS Physiological studies lobster, American 841 Pinfish predation on brown shrimp, 59 Pleuronectids — see Flatfish Pollock, walleye allocation of age-length keys, 179 distribution, feeding, and growth, 499 reproductive biology, 481 POLOVINA, JEFFREY J— see MOFFIT and POLOVINA Purse seine fishery tuna, yellowfin incidental dolphin mortality, 435 QUAST, JAY C, "Morphometric variation of Pacific ocean perch, Sebastes alutus, off western North America," 663 QUINN, T. P.— see GROOT and QUINN Polyacrylamide gel electrophoresis oocyte size determination, 187 Polychaetes predation by flatfishes, 471 "Population and fishery characteristics of Atlantic menhaden, Brevoortia tyrannus," by Dean W. Ahrenholz, Walter R. Nelson, and Sheryan P. Epperly, 569 Population studies bass, striped, 167 Chinook salmon, 13 conch, queen, 299 crabs, horseshoe, 805 dolphin, spotted, 435 dolphins, 419 larval drift, 1 menhaden, Atlantic, 569 recruitment rate method, 117 salmon, chinook, 13 sea lion, northern 351 seal, northern fur, 367 turtles, loggerhead, 447 "Populations of horseshoe crabs, Limulus polyphemus , on the northwestern Atlantic continental shelf," by Mark L. Botton and John W. Ropes, 805 725 Pout, ocean habitat detection using sonar. Prawn gamba, 311 Predation shrimp, brown. 59 "Predation on Capitella spp. by small-mouthed pleuronectids in Puget Sound, Washington," by D. Scott Becker and Kenneth K. Chew, 471 "Prevalence and effects of infection of the dorsal aorta in yel- lowfin tuna, Thunnus albacares, by the larval cestode, Dasyrhynchus talismani ," by Richard W. Brill, Robert Bourke, James A. Brock, and Murray D. Dailey, 767 Pseudaristeus — see Prawn Pterygiophore annul i aging scorpionfish. 99 Ptychocheilus grandis — see Squawfish 842 Red drum predation experiments. 67 REISENBICHLER, R. R,, and S. R. PHELPS, "Genetic varia- tion in chinook, Oncorhynchus tshawytscha , and coho, O. kisutch, salmon from the north coast of Washington," 681 "The relationship between lunar phase and gulf butterfish, Peprilus burti, catch rate," by Jeffrey H. Render and Robert L. Allen, 817 "Relationship of otolith length to total length in rockfishes from northern and central California," by Tina Echeverria Wyl- lie, 383 RENDER, JEFFREY, H., and ROBERT L. ALLEN, "The rela- tionship between lunar phase and gulf butterfish, Peprilus burti, catch rate," 817 Reproductive biology anchovy, Hawaiian, 127 clams, hard, 653 crab, golden, 547 crab, hair, 523 flounder, witch, 611 gag, 648 isopod crustacean, 377 pollock, walleye, 481 rockfishes, 229 sand lance, 631 scad, round, 251 scorpionfish, 99 sculpin, yellowchin, 201 shrimp, deepwater, 339 "The reproductive biology of walleye pollock, Theragra chalco- gramma, in the Bering Sea, with reference to spawning stock structure," by Sarah Hinckley, 481 "Research vessel survey design for monitoring dolphin abun- dance in the eastern tropical Pacific," by Rennie S. Holt, Tim Gerrodette, and John B. Cologne, 435 "Revisions of the gamba prawn genus Pseudaristeus , with de- scription of two new species (Crustacea: Decapoda: Pe- naeoidea)," by Isabel Perez Farfante, 311 Rex sole predation on polychaetes, 471 Rockfish maturity and reproduction, 229 otolith length, 383 Rockfish, black growth increments, 826 ROPES, JOHN, "Age and growth reproductive cycle, and his- tochemical tests for heavy metals in hard clams, Mercenaria mercenaria, from Raritan Bay, 1974-75," 653 ROPES, JOHN W— see BOTTON and ROPES ROSENBLATT, RICHARD H— see WAPLES and ROSEN- BLATT ROUMILLAT, WILLIAM A.— see COLLINS et al. RUTHERFORD, EDWARD S.— see IVERSEN et al. Sablefish allocation of age-length keys, 179 Sailfish larval studies, 757 Salmo gairdneri — see Trout, rainbow Salmon chinook genetic estimates of stock compositions, 13 genetic variation, 681 migration patterns, 819 stock compositions, 13 coho genetic variation, 681 migration patterns, 819 sockeye homing migration, 455 Sampling techniques allocation of age-length keys, 179 Scad, round biology in the South Atlantic Bight, 251 Scallops, bay impact of clam harvesting, 281 SCHAEFER, KURT M., "Second record of the Kawakawa, Euthynnus affinis , from the eastern Pacific Ocean," 647 Sciaenops ocellatus — see Red drum Scomberomorus maculatus — see Mackerel, Spanish Scorpionfish lift history and fishery, 99 Sculpin wooly larval drift, 1 life history summary, 4 yellowchin egg and larval development, 201 Sea lion, northern decline in abundance, 351 Sea turtle, loggerhead blood chemistry and hibernation, 34 Seagrass beds clam harvesting, 281 Seal, northern fur estimation of pups bom, 367 Seals, northern elephant crater wounds, 387 Sebastes alutus — see Perch, Pacific ocean Sebastes melanops — see Rockfish, black Sebastes spp. — see Rockfishes "Second record of the Kawakawa, Euthynnus affinis, from the eastern Pacific Ocean," by Kurt M. Schaefer, 647 Semicossyphus pulcher — see Sheephead Shark, cookiecutter crater wounds on northern elephant seals, 387 Sheephead larval drift, 1 life history summary, 4 Shrimp brown fish predation, 59 deepwater distribution and yield, 339 gamba, 311 turtle mortality during trawling, 813 "Sidescan sonar as a tool for detection of demersal fish habi- tats," by K. W. Able, D. C. Twichell, C. B. Grimes, and R. S. Jones, 725 SILVERT, W., and D. PAULY, "On the compatibility of a new expression for gross conversion efficiency with the von Berta- lanffy growth equation," 139 SMITH, DOUGLAS G.— see O'LEARY and SMITH SMITH, ERIC M., and PENELOPE T. HOWELL, "The effects of bottom trawling on American lobsters, Homarus americanus , in Long Island Sound," 737 SMR — see Standard metabolic rates SOGARD, SUSAN M., DONALD E. HOSS, and JOHN J. GOV- ONI, "Density and depth distribution of larval gulf men- haden, Brevoortia patronus, Atlantic croaker, Micropogonias undulatus, and spot, Leiostomus xanthurus , in the northern Gulf of Mexico," 601 Sonar, sidescan fish habitat detection, 725 Spawning, stock structure pollock, walleye, 481 843 Spot density and distribution, 601 feeding habitats, 785 "Spring and summer movements of subadult striped bass, Mo- rone saxatilis , in the Connecticut River," by Boyd Kynard and John P. Warner, 143 Squawfish digestion rates, 159 Squid laboratory rearing, 163 Standard metabolic rates tuna, tropical, 25 Statistical models allocation of age-length keys, 179 bias and variance in Allen's recruitment rate method, 117 Stenella attenuata — see Dolphin, spotted STEVENS, BRADLEY G.— see ARMETTA and STEVENS Stock assessments menhaden, Atlantic, 569 Strombus gigas — see Conch, queen STUBBS, DARYL L— see COLLINS et al. STUNTZ, WARREN E.— see KENWOOD and STUNTZ SUMMERSON, HENRY C— see PETERSON et al. Surveys, aerial ship dolphin school density, 419 Surveys, research crab, golden, 547 dolphin abundance, 435 trawl turtles, loggerhead abundance, 447 Tagging studies fish, oceanic, 645 Tapeworms infection in yellowfin tuna, 767 TEEL, DAVID— see UTTER et al. THEILACKER, GAIL H., "Feeding ecology and growth en- ergetics of larval northern anchovy, Engraulis mordax," 213 Theragra chalcogramma — see Pollock, walleye "Thirty-four species of California rockfishes: maturity and sea- sonality of reproduction," by Tina Wyllie Echeverria, 229 Thunnus albacares — see Tuna, yellowfin THURBERG, FREDERICK P.— see MERCALDO-ALLEN and THURBERG Tilefishes habitat detection using sonar, 725 "A trawl survey method for estimating loggerhead turtle, Caretta caretta , abundance in five eastern Florida channels and inlets," by Richard W. Butler, Walter A. Nelson, and Tyrrell A. Henwood, 447 Trawling lobster, American, 737 shrimp sea turtle capture, 813 Trout, rainbow growth of otoliths, 395 Tuna skipjack standard metabolic rate, 25 yellowfin cestode infection, 767 incidental dolphin mortality, 435 standard metabolic rates, 25 TURK, PHILIP E.— see HANLON et al. Turtles loggerhead, abundance estimates, 447 sea, capture during shrimp trawling, 813 TWICHELL, D. C— see ABLE et al. UEYANAGI, SHOJI— see LEIS et al. ULRICH, GLENN F.— see WENNER et al. UTTER, FRED, DAVID TEEL, GEORGE MILNER, and DON- ALD McISAAC, "Genetic estimates of stock compositions of 1983 chinook salmon, Oncorhynchus tshawytscha , harvests off the Washington coast and the Columbia River," 13 "Validation of the otolith increment aging technique for striped bass, Morone saxatilis, larvae reared under suboptimal feeding conditions," by Cynthia Jones and Edward B. Brothers, 171 "Variations in the blood chemistry of the loggerhead sea turtle, Caretta caretta ," by Peter L. Lutz and Ann Dunbar-Cooper, 37 VERMEER, GREGORY K., "Effects of air exposure on desic- cation rate, hemolymph chemistry, and escape behavior of the spiny lobster, PanuUrus argus," 45 Vertebral centra ageing method for tiger sharks, 269 Von Bertalanffy growth equation compatibility of new expression, 139 VONDRACEK, BRUCE, "Digestion rates and gastric evacua- tion times in relation to temperature of the Scramento squaw- fish, Ptychocheilus grandis" 159 WALLACE, ROBIN A.— see GREELEY et al. WALSH, S. J., "Habitat partitioning by size in witch flounder, 844 Glyptocephalus cynoglossus: a reevaluation with additional data and adjustments for gear selectivity," 147 WALTZ. C. WAYNE— see COLLINS et al. WAPLES, ROBIN S., and RICHARD H. ROSENBLATT, "Patterns of larval drift in southern California marine shore fishes inferred from allozyme data," 1 WARNER, JOHN P.— see KYNARD and WARNER WEINSTEIN, MICHAEL P.— see O'NEIL and WEINSTEIN WENNER. ELIZABETH L., GLENN F. ULRICH, and JOHN B. WISE, "Exploration for golden crab, Geryon fenneri in the South Atlantic Bight: distribution, population structure, and gear assessment," 547 Whales bowhead, catch and product estimates, 1805-1909, 403 gray, catch and product estimates, 1805-1909, 403 humpback, catch and product estimates, 1805-1909, 403 right catch and product estimates, 1805-1909, 403 Whitefish, ocean larval drift, 1 life history, 4 Whiting, Pacific cooked quality, 745 WILSON, K. A., K. L. HECK, Jr., and K. W. ABLE, "Juvenile blue crab, Callinectes sapidus, survival: an evaluation of eel- grass, Zostera marina, as refuge," 53 WINTERS, G. H— see DALLEY and WINTERS WISE, JOHN B— see WENNER et al. WYLLIE ECHEVERRIA, TINA, "Relationship of otolith length to total length in rockfishes from northern and central California," 383 WYLLIE ECHEVERRIA, TINA, "Thirty-four species of Cali- fornia rockfishes: maturity and seasonality of reproduction," 229 YANG, WON TACK— see HANLON et al. YOKLAVICH, M. M — see KENDALL et al. YOKLAVICH, MARY M., and GEORGE W. BOEHLERT, "Daily growth increments in otoliths of juvenile black rockfish, Sebastes melanops: An evaluation of autoradiography as a new method of validation," 826 YORK, ANNE E., and PATRICK KOZLOFF, "On the estima- tion of the numbers of northern fur seal, Callorhinus ursinus, pups born on St. Paul Island, 1980-86," 367 ZIMMERMAN, ROGER J.— see MINELLO et al. Zostera marina — see Eelgrass 845 NOTICES NOAA Technical Reports NMFS published during first 6 months of 1987 Technical Report NMFS 47. Reproduction, maturation, and seed production of cultured species. Proceedings of the Twelfth U.S.- Japan Meeting on Aquaculture, Baton Rouge, Louisi- ana, October 25-29, 1983. By Carl J. Sindermann (editor). February 1987, iii + 73 p. [13 papers.] 48. Widow rockfish. Proceedings of a workshop, Tiburon, California, December 11-12, 1980. By William H. Lenarz and Donald R. Gunderson (edi- tors). January 1987, iii + 57 p. [11 papers.] 49. Reproduction, movements, and population dynamics of the southern kingfish, Menticirrhus americanus, in the northwestern Gulf of Mexico. By Stephen M. Harding and Mark E. Chittenden, Jr. March 1987, iii + 21 p., 8 tables, 10 figs. 50. Preparation of acetate peels of valves from the ocean quahog, Arctica islandica, for age determina- tions. By John W. Ropes. March 1987, iii + 5 p., 4 figs., 1 table. 51. Status, biology, and ecology of fur seals. Proceed- ings of an international symposium and workshop, Cambridge, England, 23-27 April 1984. By John P. Croxall and Roger L. Gentry (editors). June 1987, V + 212 p. [14 Species Summaries; 12 Contributed Papers; 5 Rapporteurs' Report; 1 Bibliography.] 52. Limited access alternatives for the Pacific groundfish fishery. By Daniel D. Huppert (editor). May 1987, iii + 45 p. [8 papers.] 53. Ecology of east Florida sea turtles. Proceedings of the Cape Canaveral, Florida, sea turtle workshop, Miami, Florida, February 26-27, 1985. By Wayne N. Witzell (convener and editor). May 1987, iii + 80 p. [11 papers.] 54. Proximate and fatty acid composition of 40 south- eastern U.S. finfish species. By Janet A. Gooch, Malcolm B. Hale, Thomas Brown, Jr., James C. Bon- net, Cheryl G. Brand, and Lloyd W. Regier June 1987, iii + 23 p., 43 tables. Some NOAA publications are available by purchase from the Superinten- dent of Documents, U.S. Government Printing Office, Washington, DC 20402. 846 u S Poiiai Sarv.ca STATEMENT OF OWNERSHIP, MANAGEMENT AND CIRCULATION I A. TITLE OF PUBLICATION FISHERY BULLETIN IB PUBLICATION NO -370 2 OAT£ OF FILING 1 OCTOBER 1987 3 FREQUENCY OF ISSU£ QUARTERLY 3A NO OF ISSUES PUBLISHED 3B ANNUAL SUBSCRIPTION ANNUALLY , | PRICE SI 6. 00 rv. County, State and ZIP*4 Code/ (Not pnniers) 7600 SANO POINT WAY NE, BIN C15700 (BLDG. 1), SEATTLE, WA 93115 5 COMPLETE MAILING ADDRESS OF THE HEADQUARTERS OF GENERAL BUSINESS OFFICES OF THE PUBLISHER fNotphnier) NATIONAL MARINE FISHERIES SERVICE, NOAA , DOC UNIVERSAL BUILDING, WASHINGTON D.C. 20235 6 FULL NAMES AND COMPLETE MAILING ADDRESS OF PUBLISHER. EDITOR. AND MANAGING EDITOR (This iitm MUST NOT be blank) PUBLISHER fName and Complete Mailing Addrea) JACK MCCORMICK, SCIENTIFIC PUBLICATIONS OFFICE, NMFS 7600 SAND POINT WAY NE . BIN C15700 (BLDG. 1). SEATTLE. WA 98II3 EDITOR IName and Compute Mailins Addrea) MARY FUKUYAMA. SCIENTIFIC PUBLICATIONS OFFICE, NMFS 7600 SAND POINT WAY NE, BIN C15700 (BLDG. l), SEATTLE. WA 98115 MANAGING EDITOR (Name and Complete Mailing Addreu) MARY FUKUYAMA, SCIENTIFIC PUBLICATIONS OFFICE, NMFS 7600 SAND POINT WAY NC . BIN C1570O (BLDG. 1], SEATTLE. WA 93115 OWNER Federal income tei pur poies iCheck one/ pxi HAS NOT CHANGED DURING ! ! PRECFDiNG 12 WONTHG Di llf changed. publiS' change wtth this stx EXTENT AND NATURE OF CIRCULATION (See Insiructioru on reverse side) AVERAGE NO COPIES EACH ISSUE DURING PRECEDING U MONTHS ACTUAL NO COPIES OF SINGLE ISSUE PUBLISHED NEAREST TO FILING DATE A TOTAL NO COPIES ^iVei Prea Run) 2,070 2,076 6 PAID AND/OR REQUESTED CIRCULATION (handled by U.S GPO Sale* ihrcHJBh dealen end carr.en. )tf»eiven(Jon end counter tales Was l. D.C. 202^0) 2 Mail Sub*cr\'-A*-,'>'^\*i 20-4 Mill \N 11(11 l.lllHAHl UH niilE J